The mechanisms underlying convergent evolution in the plumage patterns of birds Thanh-Lan Gluckman Pembroke College December 2014 This dissertation is submitted for the degree of Doctor of Philosophy 08 Fall The mechanisms underlying convergent evolution in the plumage patterns of birds i Declaration This dissertation is my own work and contains nothing that is the outcome of work done in collaboration with others, except as specified in the text and acknowledgements. It is not substantially the same as any that I have submitted, or, is being concurrently submitted for a degree or diploma or other qualification at the University of Cambridge or any other University or similar institution except as declared in the Preface and specified in the text. I further state that no substantial part of my dissertation has already been submitted, or, is being concurrently submitted for any such degree, diploma or other qualification at the University of Cambridge or any other University of similar institution except as declared in the Preface and specified in the text. This dissertation does not exceed the prescribed 60,000 word limit as prescribed by the Degree Committee. Thanh-Lan Gluckman The mechanisms underlying convergent evolution in the plumage patterns of birds ii Summary Convergent evolution is a central theme in biology. Birds are an ideal system to examine the mechanisms underlying convergent evolution. Although bird patterning is diverse, within-feather patterns have repeatedly converged on the same four types: mottled patterns, scales, bars and spots. Other avian patterns occur, e.g. stripes, but are rare. In my thesis I examine the four main mechanisms underlying convergent evolution in plumage patterns: evolutionary genetics, evolutionary development, natural selection for signaling and camouflage. Japanese quail (Coturnix japonica) is a model system in developmental biology. Examining the developmental basis of pattern formation using molecular techniques, the dorsal patterning of embryonic quail is likely due to activation of the melanocortin-1 receptor, which is a highly conserved pathway in vertebrates. I examined whether a reaction-diffusion based theoretical model of pattern formation may predict developmental constraint in two groups that have different lifestyles and spectacular patterns: waterfowl (Anseriformes) and gamebirds (Galliformes). Tracing the evolutionary trajectory of pattern evolution with Bayesian comparative modeling there was evidence for developmental constraint in pattern evolution. Adaptive explanations may also result in convergence. Cuckoo-hawk mimicry has been demonstrated in the common cuckoo (Cuculus canorus) and the Eurasian sparrowhawk (Accipiter nisus), but may be prevalent in Old World cuckoos. Randomly selecting a parasitic cuckoo from each genera of Old World cuckoos and <8 sympatric raptors, I quantified their barred patterns using digital image analysis and found that parasitism can explain convergent evolution in the patterns of parasitic cuckoos and raptors. Patterns may have evolved due to ecological selection. Examining the patterns of 80% of all avian species worldwide, I found that habitat does not predict patterning, and that all four patterns are found in all habitats. These results demonstrate that the mechanisms of convergent evolution are diverse, and that development and natural selection have contributed to pattern evolution. The mechanisms underlying convergent evolution in the plumage patterns of birds iii Acknowledgements First and foremost I must thank my PhD supervisor, Dr. Nick Mundy. Thank you for seeing the potential of my ideas, and allowing them to come to life. Not having had experience in molecular and developmental biology was quite the challenge, but it has opened a whole new world that I could not have imagined before. Thank you also for your patience through my incessant questioning (not always well thought out) and controversies (same again). Having had the highest caliber of supervision I can only wish the same for anyone else wishing to pursue a PhD and that they will also be able to explore previously unimagined worlds. I would also like to thank my advisors, Prof. Nick Davies and Dr. Claire Spottiswoode for thoughtful comments and discussion. It may have been many years since we worked together, but I still have thanks for Dr. Gonçalo Cardoso for allowing me to survey plumage patterns in all birds worldwide. The six months that I spent categorizing patterns has had a lasting influence, hence the subject of this thesis. Disaster can be a wonderful thing. It allows your community to open new doors for you. For that reason I am indebted to the following people and organisations (in no particular order): Howard Bayliss, Pembroke college, the department of Zoology, the Board of Graduate Studies, Becky Coombs and Linda Wheatley. For camaraderie, I have praise and thanks for many: Kirsty MacLeod, Marj (Saucisson) Sorenson, Hannah Rowland, John Davey, Hope Usieta, Dareen AlMojil, Patricio Salazar, Amy Backhouse, Carolina Diaz, Alecia Carter and James Gilbert (it was nice to speak Bayesian gibberish with you). For much needed technical advice, training and support I am grateful to Arielle Bennett for teaching me the ins and outs of RT-PCR and to all of the members of the Baker lab (Connie Rich, Sophie Miller, Dr. Melinda Modrell, Dr. Perrine Barraud, Dr. Clare Baker) thank you for your patience, advice and technical prowess on cloning and in situ hybridization techniques. I would also like to thank my collaborators Marius Somveille and Kate Marshall, as well as the fantastic team at The mechanisms underlying convergent evolution in the plumage patterns of birds iv the Natural History Museum at Tring, in particular Douglas Russell and Hein van Gruow, as well as Matt Lowe and Mike Brooke from the Cambridge Zoology museum, Clair Castle and Jane Acred from the Balfour Library, Stuart Rankin at the Darwin supercomputing cluster Cambridge, Cassie Stoddard, and especially Aditya Saxena for continuously inspiring me with other worlds. For helpful comments on manuscripts and ideas I would like to thank Rose Thorogood, Will Feeney, Rick Prum, John Endler, Paul Brakefield, Richard Wallbank, Jack Green, and Joseph Hanly. To my family and friends, your support was an important ingredient. In particular I would like to thank Amael Paillex, Charlotte Chang (I still miss your colourful language), Lucy Robinson (Darl!), Renad Mansour, Eli Harji, Oskar Brattstrom, Lina Arenas, Penny Jones, Wendy Gu and Danika Parikh. Finally, my PhD would not have been possible without the support of my grant funding bodies: the Cambridge Commonwealth, European and International Trust, the John Stanley Gardiner Fund, and the Pembroke college graduate research fund. The mechanisms underlying convergent evolution in the plumage patterns of birds v Table of Contents Chapter 1: Introduction .................................................................................................. 1   Summary ...................................................................................................................... 3   Introduction ................................................................................................................. 5   Evo-devo of plumage patterns ................................................................................ 10   Building avian phenotypes ............................................................................................... 10   Building a feather .............................................................................................................. 12   Melanin synthesis and deposition .................................................................................... 14   Models of within-feather pattern formation ....................................................................... 17   Developmental constraint ................................................................................................. 21   Natural selection on plumage patterns .................................................................... 21   Patterns as signals ........................................................................................................... 22   A camouflage function of plumage patterns ..................................................................... 25   Adaptation and convergent evolution ............................................................................... 27   Conclusions .............................................................................................................. 28   Thesis outline ............................................................................................................ 28   Chapter 2: Activation or inhibition, MC1R and the developmental basis of within-feather pattern formation ................................................................................. 31   Abstract ..................................................................................................................... 33   Introduction ............................................................................................................... 35   Materials and Methods ............................................................................................. 45   RT-PCR analysis ..................................................................................................... 46   Immunohistochemistry ............................................................................................ 45   Results ....................................................................................................................... 47   Melanocyte distribution ........................................................................................... 47   Genes inhibiting MC1R in melanocytes .................................................................. 48   Genes activating MC1R in melanocytes ................................................................. 57   Discussion ................................................................................................................. 61   Chapter 3: Evolutionary pathways to convergence in plumage pattern phenotypes .................................................................................................................... 68   Abstract ..................................................................................................................... 70   The mechanisms underlying convergent evolution in the plumage patterns of birds vi Introduction ............................................................................................................... 72   Materials and Methods ............................................................................................. 79   Phylogenies ............................................................................................................. 79   Data collection and coding ...................................................................................... 80   Modelling of plumage pattern evolution .................................................................. 82   Model priors and modelling parameters .................................................................. 84   Statistical analysis ................................................................................................... 85   Results ....................................................................................................................... 87   Taxonomic distribution of patterns .......................................................................... 87   Evolution within patches of plumage ....................................................................... 88   Regional modularity in plumage pattern evolution .................................................. 89   Global model of plumage pattern evolution ............................................................. 91   Discussion ................................................................................................................. 95   Chapter 4: Cuckoos in raptors’ clothing, barred plumage illuminates a fundamental principle of Batesian mimicry ............................................................. 102   Abstract ................................................................................................................... 104   Introduction ............................................................................................................. 106   Materials and Methods ........................................................................................... 109   Results ..................................................................................................................... 117   Discussion ............................................................................................................... 127   Chapter 5: Ecological selection for bird plumage patterns worldwide ................. 132   Abstract ................................................................................................................... 134   Introduction ............................................................................................................. 136   Methods and materials ........................................................................................... 143   Data collection ....................................................................................................... 143   Measuring phylogenetic signal .............................................................................. 144   Ecological selection for plumage patterns ............................................................. 145   Results ..................................................................................................................... 149   Plumage pattern evolution due to shared ancestry ........................................................ 149   Ecological selection for plumage patterns ...................................................................... 152   Ecological selection for mottled, scaled, barred or spotted patterns .............................. 155   A comparison of ecological selection between closed and open habitats ...................... 156   Discussion ............................................................................................................... 160   The mechanisms underlying convergent evolution in the plumage patterns of birds vii Chapter 6: Discussion and Conclusion .................................................................... 168   Concluding thoughts .............................................................................................. 170   Developmental explanations ................................................................................. 172   Developmental constraint ...................................................................................... 174   Adaptive explanations ........................................................................................... 175   References .................................................................................................................. 178   Supplementary material ............................................................................................. 214   Table S2.1. Primer sequences tested for RT-PCR. ........................................................ 214   Fig. S3.1a. Distribution of plumage patterns per patch of plumage in Anseriformes. ..... 217   Fig. S3.1b. Distribution of plumage patterns per patch of plumage in Galliformes. ........ 218   Fig. S3.2. Local pattern evolution within patches, in Anseriformes and Galliformes. ..... 221   Fig. S3.3. Regional pattern evolution within regions in Anseriformes and Galliformes. . 224   Fig. S3.4. Pattern evolution over the whole body, in Anseriformes and Galliformes. ..... 226   Table S3.1. Prior probability of encountering models of <12 parameters. ...................... 227   Table S3.2. Prior probability of encountering models of <6 parameters. ........................ 228   Table S3.3. The number of unique models as well as the ancestral state of patterns per body region in Anseriformes and Galliformes. ......................................................... 229   Fig. S4.1. Granularity spectra of the ventral and flank area of each species sampled. .. 232   Fig. S4.2a. Median pattern attributes for Eudynamys scolopacea. ................................ 233   Fig. S4.2b. Median plumage pattern attributes for Chrysococcyx flavigularis. ............... 234   Fig. S4.2c. Median plumage pattern attributes for Cacomantis sonneratii. ................... 235   Fig. S4.2d. Median plumage pattern attributes for Cercococcyx mechowi. .................... 236   Fig. S4.2e. Median plumage pattern attributes for Cuculus saturates. ........................... 237   Table S4.1. Comparison of distribution overlap from a taxonomic source ..................... 238   Table S4.2. Cuckoo and raptor host prey species .......................................................... 239   Table S4.3a. The individual types of habitat inhabited by cuckoos and sympatric raptors of Africa .............................................................................................................. 242   Table S4.3b. The individual types of habitat inhabited by cuckoos and sympatric raptors of Oceania .......................................................................................................... 244   Table. S5.1. A representative sample of literature of pattern type and pattern function in different contexts (stationary or moving) in animals. .................................................. 246   References for Table S5.1 .............................................................................................. 261   Table S5.2. The number of bird species with plumage patterns in the 16 habitat types of the world across the class Aves. ................................................................................ 271   The mechanisms underlying convergent evolution in the plumage patterns of birds 1 Chapter 1: Introduction “To create is to recombine” (François Jacob, 1977). The mechanisms underlying convergent evolution in the plumage patterns of birds 2 The mechanisms underlying convergent evolution in the plumage patterns of birds 3 Summary The remarkable diversity of the plumage of birds has captivated observers since ancient times. From the time of Charles Darwin and Alfred Wallace, avian phenotypes are classically considered to have evolved due to natural selection. Within-feather patterns, such as bars and spots, frequently occur in birds. In spite of the diversity in avian phenotypes, bird plumage patterns have predominantly converged on just four types of patterns: mottled, scaled, barred and spotted patterns. Other patterns have evolved in birds but are comparatively rare, e.g. stripes, squares, and triangles. The main mechanisms underlying convergent evolution in bird plumage patterns are currently unknown. The prevailing view is that plumage patterns function in camouflage and/or communication. However, that just four patterns have repeatedly evolved in birds implies that there may be developmental constraint in plumage pattern evolution. Developmental constraint as an evolutionary force is gaining momentum with proponents advocating an equivalent importance to natural selection. The relatively new field of evolutionary developmental biology (evo-devo) focuses on the means by which developmental processes may bias the evolution of morphological diversity. Evo-devo has the possibility to enhance our understanding of diversity by combining empirically derived principles of cellular and tissue morphogenesis to inform evolutionary processes. Natural selection in bird plumage has been studied widely. However, the majority of studies have focused on uniform patches of colouration, whereas the number of studies on plumage patterns is comparatively few. The main mechanisms underlying convergent evolution span the new field of evolutionary development, via modularity and developmental constraint, as well as neo-Darwinian theory, which spans natural selection for signaling as well as camouflage. In my PhD thesis I investigate how each of these major themes has contributed to convergent evolution in the plumage patterns of birds. The mechanisms underlying convergent evolution in the plumage patterns of birds 4 The mechanisms underlying convergent evolution in the plumage patterns of birds 5 Introduction The plumage of birds has captured the imagination since at least the time of Aristotle (384-322 BC) and has caused many an observer to ponder the existence of a diversity in the phenotypes of animals (Hunter 1780; Davies 2011). In particular, Charles Darwin and Alfred Russell Wallace emphasized the importance of natural selection in the evolutionary history of avian plumage. However, Darwin and Wallace broadly disagreed in key aspects of natural selection. Darwin emphasized the importance of extravagant traits such as the bright spotted train of the male peacock as a result of sexual selection (Darwin 1871), whereas Wallace stressed the importance of less brightly coloured “drab” plumage for camouflage (Darwin and Wallace 1858). Later, Wallace would deviate from his strictly adaptationist approach to argue that bright coloration may not in fact be adaptive, but may result from aspects of development: “Colour may be looked upon as a necessary result of the highly complex chemical constitution of animal tissues and fluids” (Wallace 1889). Since Darwin and Wallace, studies of natural selection have attracted a wide range of research attention in many animal groups, and birds in particular have been studied widely (Andersson 1994). Melanin-based pigmentation is of particular interest as the mechanism of melanin synthesis is highly conserved in vertebrate animals. In birds, melanin can be uniform or spatially variable within feathers, and a range of hypotheses have been proposed to explain the function and evolution of avian melanin based traits (e.g. Prum and Williamson 2002; Majerus and Mundy 2003; Roulin et al. 2003; Mundy et al. 2004; Mundy 2005; Hill and McGraw 2006a; Hill and McGraw 2006b; Gluckman and Cardoso 2010; Emaresi et al. 2013; Emaresi et al. 2014). Recently, both empirical and macroevolutionary approaches have begun to demonstrate that our understanding of the process of natural selection is not always as imagined. For example, some studies have failed to find sexual selection on sexually dimorphic traits nor heritable genetic variation, that are key components of natural selection theory (e.g. Griffith, Owens, and Burke 1999; Hadfield et al. 2006; Westneat 2006). The mechanisms underlying convergent evolution in the plumage patterns of birds 6 The theory of Darwinian evolution is based on the principles of variation, heredity and natural selection (Darwin and Wallace 1858; Darwin 1868; Darwin 1869; Darwin 1871; Wallace 1889). At the time of the inception of Darwinian evolution, the modern founder of the science of genetics, Gregor Johann Mendel, had begun unraveling the rules of heredity (Mendel 1866) but his works remained largely unnoticed until their rediscovery in 1900 (Henig 2001). Genetics explained variation and heredity and its incorporation into Darwin’s theory of evolution resulted in what is now known as the modern synthesis (Huxley 1942; Kutschera and Niklas 2004), which is the prevailing paradigm in evolutionary biology. In this current union of ideas, natural selection is regarded as the main mechanism of change that acts on the phenotype of populations of species’ in the context of their environment. Under this model of evolution, the direct units that selection is acting upon are the genes involved in phenotypic change. For example, the rate of amino acid change in the melanocortin-1 receptor (MC1R) locus in birds is correlated with sexual dichromatism in galliform birds (Nadeau, Burke, and Mundy 2007) and changes at this locus are also involved in polymorphism which is a sexually selected trait in some avian species (Mundy et al. 2004). However, these same traits may be correlated with body size across birds, which implies a link between evolution and development in melanin-based avian plumage (Riegner 2008). The field of evolutionary developmental biology (evo-devo) is a relatively new field of biology that studies the evolution of developmental processes and its contribution to morphological diversity. Evo-devo focuses on developmental plasticity, modularity in evolution and the regulatory mechanisms by which genes can be selectively turned on and off (West-Eberhard 2003; Klingenberg 2008). Seemingly disparate organisms may use the same genes, but these genes can be regulated differently between organisms to produce variation in dissimilar aspects of morphology. For example, bone morphogen proteins (BMPs) function in diverse aspects of morphology such as teeth and teeth suppression in mice and birds (Chen et al. 2000), beak morphology (Fritz et al. 2014) and human melanocortin systems (Giraldi et al. 2011). In the field of evo-devo, empirical genetic approaches applied at the level of cell and tissue morphogenesis are used to illuminate general principles of The mechanisms underlying convergent evolution in the plumage patterns of birds 7 development (Haag and Lenski 2011). For example, there are similarities in the developmental pathways that produce variation in mouse and avian phenotypes (Nadeau et al. 2008; Manceau et al. 2011) which is consistent with the idea of molecular economy, where evolution alters developmental processes to create novel structures from old or existing gene networks. Fig. 1.1. The prevalent within-feather patterns of birds. In spite of the diversity of avian phenotypes, avian plumage patterns have repeatedly converged on the same four patterns: Irregular - a) mottled plumage in a female sharp-tailed grouse (Tympanachus phasianellus); Regular - b) barred plumage in a male Andean goose (Chloephaga melanoptera), c) scaled plumage in a male falcated duck (Anas falcata), d) spotted plumage in a male great argus (Argusianus argus). Figure taken from Marshall and Gluckman, in review. Visual patterns, such as bars and spots, are common throughout the animal kingdom and are composed of a motif (sub-pattern) that is reiterated within and/or between feathers to create a patch of patterning (Fig. 1.1) (Prum and Williamson 2002; Kenward et al. 2004; Riegner 2008; Gluckman 2014). In birds, within-feather patterns are formed via differential control of melanin (Hill and McGraw 2006b) and The mechanisms underlying convergent evolution in the plumage patterns of birds 8 have predominantly converged on the same four motifs: mottled, scaled, barred, or spotted (Fig. 1.1). There are few other regularly repeating patterns in birds with the exception of stripes, checkered patterning, and triangles (Fig. 1.2), which are relatively rare across the class Aves (T-L. Gluckman unpublished data). Therefore, birds can make other types of patterns but instead repeatedly converge on the same four motifs. It is currently unknown why bird plumage patterns have repeatedly converged on these patterns and this is the subject of this PhD thesis. A) B) Fig. 1.2. Plumage patterns that appear to only occur in one species. A) The breeding checkerboard plumage pattern on the dorsal surface of the common loon (Gavia immer), and B) the triangles in the breast of the male Wood duck (Aix sponsa). The mechanisms by which closely and distantly related species converge on the same phenotype are a central theme in evolutionary biology. Traditionally, the mechanisms underlying a similarity in animal phenotypes have been split into parallelism and convergence on the basis of whether the species’ in question are closely or distantly related (Brakefield 2006; Arendt and Reznick 2008). Under parallelism, closely related species evolve a similar phenotype due to similar mechanisms. For example, the observed patches of melanin based pigmentation in leaf warbler species of the genus Phylloscopus can be explained by developmental changes (Price and Pavelka 1996). In contrast, convergence has historically been applied to similar phenotypes in distantly related species arising from different The mechanisms underlying convergent evolution in the plumage patterns of birds 9 genetic mechanisms, for example, winged flight in birds, bats and insects. However, it is becoming clear that there is no consistent way of defining convergence when mechanisms are taken into account (Brakefield 2006). For example, the same genetic mechanism responsible for melanism in birds has also been implicated in lizards, cats and woolly mammoths (Majerus and Mundy 2003; Mundy et al. 2004; Arendt and Reznick 2008). Additionally, whereas the same locus is responsible for pale coloration in some populations of beach mice (Peromyscus polionotus), in other populations of the same species the mechanism underlying pale colour is unknown (Hoekstra et al. 2006). Here I use “convergent” as a flexible term to describe the independent evolution of similar phenotypic traits (Arendt and Reznick 2008). Research into convergent evolution via evo-devo and natural selection is required to understand how these two processes lead to the evolution of similar phenotypes. The evolutionary developmental biology of bird plumage has largely focused on the growth and formation of the feather itself (e.g. Noramly and Morgan 1998; Prum 1999; Chen and Chuong 2000; Fliniaux, Viallet, and Dhouailly 2004; Jiang et al. 2004) rather than within-feather patterning (but see Prum and Williamson 2002). An understanding of the dynamics of within-feather patterning could clarify the function and evolution of this little studied but common avian plumage trait. The prevailing view of the function and evolution of plumage patterns has remained largely rooted in the perspective of adaptationist ideas, but has seldom received research attention. A broad scale analysis of bird plumage patterns from an evolutionary and macroevolutionary perspective is overdue. In my thesis, I attempt to understand the evolutionary developmental biology of within-feather patterns and natural selection for patterning to examine how each contributes to convergent evolution. In the first half of my thesis I examine the developmental mechanism of patterning and whether the development biology of pattern formation may shape plumage evolution. In the second half of my thesis I focus on how natural selection has shaped plumage phenotypes to converge on the The mechanisms underlying convergent evolution in the plumage patterns of birds 10 same patterns for signaling and/or camouflage as per Darwin and Wallace’s hypotheses, respectively. In this introductory chapter, I discuss what is known of the evolutionary developmental biology of within-feather patterns and natural selection on bird plumage. I then discuss how I will study these processes in birds. “We can see why characters derived from the embryo should be of equal importance with those derived from the adult, for a natural classification of course includes all ages” (Darwin 1869). Evo-devo of plumage patterns Building avian phenotypes Plumage coloration can vary over the body and is coordinated into species- typical phenotypes by individual patches containing up to thousands of feathers (Prum and Dyck 2003). For example, the feathers of the peacock’s (Pavo cristatus) train are modified covert feathers from the upper tail that are pigmented with spots whereas the wing feathers have barred patterns. Patches of feathers that covary in coloration generally correspond to seven feather tracts (pterylae) over the body, which vary little in anatomical position between birds (Fig. 1.3; Lucas and Stettenheim 1972). The mechanisms underlying convergent evolution in the plumage patterns of birds 11 Fig. 1.3. The seven major individual feather tracts in birds. Field guide terminology, and the corresponding feather tracts (pterylae; Lucas and Stettenheim 1972) are as follows: a) Nape: interscapular tract, b) Wing (scapular, wing coverts, tertials, primaries and secondaries): humeral tract, upper marginal coverts of prepatagium and upper wing covert tract, c) Rump and uppertail coverts: dorsopelvic tract and dorsal caudal tract, d) Tail: upper major tail covert, upper median tail covert and rectrices tract, e) Breast: ventral cervical tract, f) Flanks or side: pecterosternal or pectoral tract, g) Vent and undertail coverts: abdominal or lateral and medial abdominal tracts. The species illustrated is the Natal francolin (Pternistes francolinus). Plumage patterns appear to be determined by feather follicle-level processes (Takeuchi et al. 1996; Prum and Williamson 2002; Yoshihara et al. 2011; Oribe et al. 2012; Yoshihara et al. 2012). The potential to form feather follicles is distributed throughout the ectoderm (Harris, Fallon, and Prum 2002). Numerous interactions between the epithelium and mesenchyme signal pteryla formation (Chuong 1993; Prum 1999; Chuong et al. 2000; Dhouailly et al. 2004; Lin et al. 2013). Epithelial- mesenchymal recombination in quail-chicken chimeras shows that the dermal cells of the ventral and dorsal surface have different origins: the dorsal dermis is of neural crest origin whereas the ventral dermis originates from lateral plate mesoderm The mechanisms underlying convergent evolution in the plumage patterns of birds 12 (Dhouailly et al. 2004; Fliniaux, Viallet, and Dhouailly 2004; Lin et al. 2006). From what is currently understood of plumage development, covariation of feather pigmentation within patches is indicative of local shared developmental mechanisms (Lucas and Stettenheim 1972; Prum and Dyck 2003; Lin et al. 2006; Wagner, Pavlicev, and Cheverud 2007). In contrast, it is unknown whether the development of pigmentation within one patch of plumage influences the evolution of pigmentation in another patch of plumage. For example, the pink-eared duck (Malacorhyncus membranaceus), has barred plumage on the flanks and the breast. Perhaps barred plumage evolved on the breast first, and the mechanism of pattern formation was subsequently recruited by the flanks, or vice versa. An interesting issue is whether plumage pattern evolution is segmented and follows an anteroposterior gradient or whether plumage pattern evolution is constrained to a dorsoventral gradient because the dermal cells of the ventral and dorsal surface are of different origins. At the level of the whole body, evolution of a novel plumage pattern may occur within the same patch or involve recruitment from other patches, or other modules, and might therefore appear relatively unconstrained. These considerations indicate that there may be modularity within a larger developmental hierarchy from the level of individual patches, to regions, and possibly over the whole body. Building a feather Feathers vary in type and coloration between chick, juvenile and adult phases (Fig. 1.4), as well as over the body and between the sexes. For example, chicks hatch with neoptile natal down feathers that are unbranched structures whereas adult plumage is predominantly comprised of pennaceous barbed feathers interspersed with neoptile down feathers (Prum 1999; Oribe et al. 2012). Both neoptile and pennaceous feathers are complex branched keratin structures. The entire plumage is replaced by moulting at least once a year The mechanisms underlying convergent evolution in the plumage patterns of birds 13 Fig. 1.4. Feathers vary in type between chick, juvenile and adult phases (Yu et al. 2004). During embryogenesis, and at hatching, chicks have unbranched feathers known as down feathers. During the first juvenile moult, barbed pennaceous feathers emerge and are similar in structure but smaller in comparison to the adult feathers that are depicted here are as the third generation of feathers. The development of neoptile feathers begins early in embryogenesis and is indicated by a thickening of the ectoderm by the elongation of epidermal cells, forming the epidermal placode (Noramly and Morgan 1998; Jiang et al. 1999). An aggregation of dermal cells forms below the epidermal placode that subsequently develops into the feather germ and follicle. Elongation of the placode leads to the formation of a cylinder of dermis enclosed in a tubular epidermal structure, which is the first feather germ, or short bud. The basal epidermis of the feather bud invaginates forming the feather follicle (Prum and Dyck 2003). From epidermal differentiation of the feather bud, barb ridges of the first natal down are formed. The mechanisms underlying convergent evolution in the plumage patterns of birds 14 In adulthood, feathers continue to grow from feather follicles but change dramatically in shape. Similar to the development of neoptile feathers, nutrients for the pennaceous feather are delivered by the feather pulp, located between and under the feather follicle (Fig. 1.5). In contrast with neoptile feathers, the developing barbs at the anterior side of the feather follicle of pennaceous feathers fuse to create the rachis. The individual barb ridges of the feather are formed by the keratin producing cells of feathers (keratinocytes) which take up melanin pigmentation, and is a central component of within-feather pattern formation (Prum and Williamson 2002; Prum and Dyck 2003). Fig. 1.5. Helical pennaceous feather growth taken from Lucas and Stettenheim (1972). A unique aspect of avian integument structure is the geometry of helical feather growth, which is an important consideration in within-feather pattern formation. Melanin synthesis and deposition From what is currently known, melanin is the most important type of pigment The mechanisms underlying convergent evolution in the plumage patterns of birds 15 involved in within-feather patterning (Hill and McGraw 2006a). The same types of melanin are synthesised in mammals and birds: eumelanin, resulting in dark brown or black coloration, and pheomelanin, resulting in yellow to buff brown (Richardson, Hornbruch, and Wolpert 1989; Takeuchi and Takahashi 1998; Tobita-Teramoto et al. 2000; Hill and McGraw 2006b). In addition, white coloration is due to an absence of melanin (Mundy et al. 2004). Within-feather patterning is predominantly comprised of combinations of eumelanin, pheomelanin and an absence of pigmentation. Alternate combinations do occur, such as in the barred plumage of the budgerigar (Melopsittacus undulatus), which is additionally pigmented with psittacofulvins that are specific to Psittaciformes, but these alternate combinations are rare (Hill and McGraw 2006b; T-L. Gluckman unpublished data). The melanins of feathers are synthesised endogenously and are derived from melanocyte pigmentation cells. Feather melanocytes are of neural crest origin. From the dermal pulp of the feather germ, melanocytes migrate into the tubular epidermis of the developing feather (Fig. 1.6). The organelles within melanocytes that are responsible for the synthesis of melanin are melanosomes. Melanocytes transfer melanosomes to the feather keratinocytes via pseudopodia, and melanosomes are transferred into the keratinocytes by phagocytosis (Lucas and Stettenheim 1972; Prum and Williamson 2002). The melanocyte-specific enzyme tyrosinase (TYR) catalyses the first steps of both eumelanin and pheomelanin synthesis. Eumelanin synthesis additionally requires the tyrosinase-related protein 1 (TYRP1) and DOPAchrome tautomerase (DCT) whereas pheomelanin appears to require few other proteins, but additionally requires cysteine (Ito, Wakamatsu, and Ozeki 2000; Hill and McGraw 2006b; Dessinioti et al. 2009; Galván, Ghanem, and Møller 2012). Each type of melanin has its own particular type of melanosome – eumelanosomes for eumelanin, and pheomelanosomes for pheomelanin – and it is currently thought that only one type of melanosome is contained within each melanocyte. Therefore, melanocyte type and distribution in feathers is a key component of within-feather pattern formation. However, little is known of the mechanisms underlying the spatial distribution of melanocytes. The mechanisms underlying convergent evolution in the plumage patterns of birds 16 Fig. 1.6. Melanosome transfer of keratinocytes during feather formation taken from Prum and Williamson (2002). The differential uptake of melanosomes into feather keratinocytes is of crucial importance in within-feather patterning. Melanin type synthesized in melanocytes is due to activity of the melanocortin-1 receptor (MC1R) which is a G protein coupled receptor on the cell surface (Ling et al. 2003; Takeuchi, Takahashi and Okimoto 2003) (Oribe et al. 2012). The MC1R locus is important in the evolution of avian coloration, and changes at this locus can have a widespread effect on melanin based coloration in birds (Ollmann et al. 1998; Kerje et al. 2003; Mundy et al. 2004; Roulin 2004; Mundy 2005; Nadeau 2006; Nadeau, Minvielle, and Mundy 2006; Nadeau, Burke, and Mundy 2007; Ducrest, Keller, and Roulin 2008; Dorshorst and Ashwell 2009; Vidal et al. 2010; Emaresi et al. 2013). Expression of eumelanin and pheomelanin is due to The mechanisms underlying convergent evolution in the plumage patterns of birds 17 differential stimulation of MC1R, which regulates the cyclic adenosine monophosphate (cAMP) signaling pathway. MC1R is activated by the binding of melanocortin peptides to MC1 receptors on the cell surface. Melanocortin peptides are generated by post-translational processing of the pro-opiomelanocortin peptide (POMC) that is encoded by the POMC gene (Benjannet et al. 1991; Takeuchi, Teshigawara, and Takahashi 1999; Yoshihara et al. 2011).The most common melanocortins that have a function in animal pigmentation are the adrenocorticotropic hormone (ACTH) and α-MSH (Takeuchi et al. 2003; Takahashi et al. 2006; Roulin et al. 2011; Yoshihara et al. 2011; Roulin and Ducrest 2013). Binding of melanocortin peptides to MC1R elevates intracellular cAMP levels which in turn stimulates the synthesis of eumelanin. Conversely, the agouti signaling protein (ASIP) acts as an inverse agonist of MC1R lowering intracellular cAMP levels leading to pheomelanin synthesis (Nadeau et al. 2008; Oribe et al. 2012; Yoshihara et al. 2012). ASIP has been implicated in mammalian (i) dorsoventral patterning, (ii) temporal-specific regulation of pigmentation during hair growth leading to banding patterns (Bultman, Michaud, and Woychik 1992; Vrieling et al. 1994; Barsh 1996; Fontanesi et al. 2010; Kaelin et al. 2012), and (iii) inhibition of melanocyte differentiation (Aberdam et al. 1998; Linnen et al. 2009). In chicken and quail, ASIP may function in dorsoventral patterning as well as within-feather patterning (Nadeau et al. 2008; Yoshihara et al. 2012). Two mechanisms of within-feather pattern formation have been demonstrated: melanocyte distribution/differentiation and pigment-type switching via the melanocortin-1 receptor (MC1R). For example, the distribution of melanocytes can lead to bars and spots in certain breeds of chicken (Lin et al. 2013). The second mechanism acts via differential stimulation of the MC1 receptors (Yoshihara et al. 2012). Models of within-feather pattern formation The proposed organizing mechanism for within-feather pattern formation in birds is reaction-diffusion, based on Alan Turing’s (1952) original proposition The mechanisms underlying convergent evolution in the plumage patterns of birds 18 describing morphogenesis (Turing 1952; Oster and Alberch 1982; Oster et al. 1988; Prum and Williamson 2002). Reaction-diffusion models are comprised of an activating morphogen, and an inhibiting morphogen, which is produced by the activator. The basis of the reaction-diffusion mechanism is that it interacts with a hypothetical gene product level, or morphogen, activating colour-specific enzymes (Oster and Alberch 1982; Oster et al. 1988). Reaction-diffusion models have been used to model patterns over the body in a variety of taxa including Lepidopterans and Zebras. For example, reaction-diffusion can accurately model animal patterning by incorporating the geometry and size of the domain on which the initiating reaction- diffusion mechanism resides, during embryonic development (Oster and Alberch 1982; Oster et al. 1988). Reaction-diffusion models can accurately simulate pattern changes over the body. For example, where the body and the tail meet, “notch” markings are apparent on the tails of striped zebras whereas spotted cheetah’s have striped tails at the periphery (Oster et al. 1988). It is thought that these biological examples demonstrate developmental constraint in the mechanism of pattern formation: spotted animals can have striped tails but striped animals cannot have spotted tails. Although barred, or striped, bodies with spotted tails are rare in the avian world (T-L. Gluckman unpublished data), a few species exhibit this pattern distribution. For example, peacocks have barred wings and spotted upper tail covert feathers. However, patches of homogeneous colouration separate each patch of the peacock’s plumage patterns, thereby separating localised within-feather plumage patterning mechanisms. It is thought that where two adjacent domains with independent reaction-diffusion events meet, they will interact to produce a different pattern. For example, where the limbs of zebras meet the body, a scapular stripe occurs, which is often described as “chevrons” in avian literature (Oster and Alberch 1982; Oster et al. 1988; Oster and Murray 2005). Domain size may also be important in animal phenotypes (Oster and Alberch 1982; Oster et al. 1988; Oster and Murray 2005). For example, a domain size that is too small or too large may result in an absence of patterning and a domain of The mechanisms underlying convergent evolution in the plumage patterns of birds 19 intermediate size is thought to facilitate reaction-diffusion based pigmentation patterns (Murray 1981). This is because when the spatial scale of the domain is restricted, the spatial structure between the morphogens cannot exist, but that when the spatial scale is large enough, the distribution of morphogens has the possibility to become uniform again. The repeating absence of patterns in large and small birds in avian morphospace supports these model based predictions (Murray 1981; Riegner 2008). However, reaction-diffusion models also predict a phenotype that shows variable pigmentation along the anteroposterior axis, e.g. white/black anterior section with a white/black posterior section. This is not observed in birds, which have pigmentation differences on the dorsoventral axis, e.g. countershading (Riegner 2008). In contrast with models of pattern formation in animals, Prum and Williamson’s (2002) reaction-diffusion model incorporates the unique considerations of feather growth. By modeling reaction-diffusion in the context of helical feather growth, Prum and Williamson (2002) were able to model the formation of all avian regular plumage patterns, i.e. scales, bars and spots, but not irregular mottled patterns (Fig. 1.1). This model makes explicit predictions as to which plumage patterns require the most stringent regulation: the regulation of scales is the least stringent, bars require a specific amount of morphogen, and spots have the most stringent formation parameters. This may have important implications for plumage pattern evolution as it implies that some patterns may be harder to make than others. In addition, perhaps the size of the patch of patterning may further constrain the mechanism of pattern formation, as some patterns may need smaller or larger feather tracts on which to form. Analyses of avian morphospace imply that with decreasing body size birds tend to evolve patches of black and white, then bars, then spots, before mottled plumage, and finally countershading (Riegner 2008). That avian morphospace correlates well with a transition in Prum and Williamson’s model (2002) could reflect developmental constraint. However, the previous analysis of avian morphospace must be treated cautiously as scales are not included, and phylogeny is not accounted for in the statistical methods (Riegner 2008). The mechanisms underlying convergent evolution in the plumage patterns of birds 20 Reaction-diffusion based models provide testable hypotheses with which to explore plumage pattern development and evolution. Currently it is unknown if reaction-diffusion controls pattern formation within feathers. The few experimental manipulations of plumage patterns in poultry have shown that barred plumage patterns can be transplanted or altered by transplanting melanocytes from barred hosts into the feather ectoderm of chickens (Willier and Rawles 1938; Willier and Rawles 1940). It is potentially significant for reaction-diffusion models that MC1R is differentially stimulated by MSH and ASIP resulting in variation in pigmentation within-feathers, but it is currently unknown whether MSH and ASIP interact to form plumage patterns, and whether reaction-diffusion is the mechanism underlying pattern formation. The development of the eyes on the tail-feathers of a peacock (Beebe 1922). The mechanisms underlying convergent evolution in the plumage patterns of birds 21 Developmental constraint The possibility to evolve plumage patterns may be constrained by a difficulty in developing a particular kind of pattern. For example, spotted patterns may have the most stringent regulation parameters (Prum and Williamson 2002) and might be harder to make than others. Therefore, development may constraint plumage pattern evolution. Price and Pavelka (1996) proposed that the repeated evolution of similar avian phenotypes may be explained by developmental constraint with two possible consequences: 1) among-species diversity due to the magnification of slight differences in ancestors, 2) parallelism (convergence) (Price and Pavelka 1996). The former is a plausible explanation, resulting in minor variation to a particular ancestral motif. For example, ducks exhibit a diversity of barred plumage, e.g. thick bars in the pink-eared duck (Malacorhyncus membranaceus), as well as thin bars (vermiculated), as seen in the tufted duck (Aythya collaris). However, diversity due to the magnification of slight differences in ancestors does not explain the diversity of pattern types in closely related species such as bars, scales and spots in Anatidae, or recursion of the same patterns in distantly related families. Therefore, convergence due to developmental constraint may better explain similar plumage patterns in distantly related taxa, such as the barred plumage of cuckoos, woodpeckers and galliform species, and potentially that of closely related species. Natural selection on plumage patterns Natural selection acts on the phenotype of organisms and a heritable reproductive advantage may lead to the selected phenotype becoming prevalent in different populations depending on their environment, which includes conspecifics and predators (Darwin and Wallace 1858; Darwin 1869; Wallace 1889). Over time, the process of natural selection gives rise to populations that are adapted to their particular ecological niche. On the basis of an adaptive advantage animal visual signals should evolve to be effective for the function that they serve, which can be in The mechanisms underlying convergent evolution in the plumage patterns of birds 22 communication and/or camouflage (Endler 1978; Endler 1992; Bradbury and Vehrencamp 1998; Kenward et al. 2004). It is generally thought that patterning is beneficial because patterns allow animals to be less conspicuous against heterogeneous backgrounds. The general principle underlying the function of patterns in camouflage is that patterns resemble aspects of the background in order to evade detection by predators or prey. In the context of signaling, patterns must visually diverge from the background to stand out to conspecifics. For example, bold patterns on a plain background make an animal standout, having a similar pattern to the background allows the animal to blend in, whereas patterns that oppose the geometric pattern of the background are conspicuous (Bradbury and Vehrencamp 1998). The communication signals of animals, especially sexually selected traits, are often visually striking and stand out from the background in which the animal is viewed (Andersson 1994). In contrast, the purpose of camouflage is to conceal the presence of an animal and this is often achieved by matching the background (Poulton 1890; Thayer 1909; Cott 1940). As a consequence of these opposing selection pressures there is a functional compromise between signaling and camouflage (Andersson 1994). A growing number of studies of animal behavior show that patterns can function in communication. In comparison, the number of studies of the function avian plumage patterns in camouflage is few, but the similarity of the patterns found in birds to other animal groups may be revealing of their protective benefit. For example, irregular pigmentation such as mottled patterns (Fig. 1.1) is common in many animal groups and has been shown to have a protective benefit in invertebrates (Thayer 1909; Cott 1940; Bradbury and Vehrencamp 1998; Stevens and Merilaita 2009a; Stevens and Merilaita 2009b). Patterns as signals In order for traits to serve as a social signal they must elicit a response in the receiver and in the context of sexual ornamentation, must be reliable indicators of The mechanisms underlying convergent evolution in the plumage patterns of birds 23 individual condition, such as through condition dependence (Andersson 1994; Bradbury and Vehrencamp 1998), or females that choose those ornaments benefit through indirect genetic benefits to their offspring and or improved access to resources (Darwin 1871; Fisher 1923). Empirical experiments demonstrate that patterns can elicit a social response in a range of species (e.g. Hasson 1991; Swaddle and Cuthill 1994; Roulin 1999a; Roulin et al. 2003; Bortolotti et al. 2006). For example, in a classical study of sexual selection and visual signalling, the number of eyespots on the tail/train of the male peacock (Pavo cristatus) is positively associated with mating success demonstrating that females select for males with more eyespots (Petrie, Halliday, and Sanders 1991). In addition, patterning can be a reliable indicator of individual condition. For example, the size of barred throat patterns is positively associated with body condition in female barred buttonquails (Turnix suscitator) (Muck and Goymann 2011) and adaptive sex-ratio biases are related to the plumage spottiness of parent barn owls (Tyto alba) (Roulin et al. 2010). Alternatively, pleiotropic effects, whereby one gene influences multiple phenotypic traits (e.g. via metabolic pathways), may also result in correlated responses to selection in melanin based traits (Ducrest, Keller, and Roulin 2008). For example, female eumelanin based traits in barns owls is correlated with offspring quality, life history attributes, physiology and morphology, and this interaction is thought to be due to pleiotropic effects of melanin production (Roulin 2004; Ducrest, Keller, and Roulin 2008; Emaresi et al. 2013). The mechanisms underlying convergent evolution in the plumage patterns of birds 24 Types de Plumes Larouss pour tous, Adolphe Philippe Millot (1907). Barred plumage patterns are particularly common in many different species and may function in both communication and camouflage (Riegner 2008; Gluckman and Mundy 2013). For example, compared to irregular mottled patterns, barred plumage is more likely to evolve on the ventral surface of males at sexual maturity, which is a more likely location for avian communication (Gluckman and Cardoso 2010). In contrast with irregular mottled patterns, the regularity of barred patterns could perhaps act as a signal of individual quality by making irregularities within the pattern perceptually salient (Gluckman and Cardoso 2009). That regular patterns can facilitate communication of honest aspects of individual quality may select for repeated evolution of these types of patterns on the ventral surface of males (Gluckman and Cardoso 2010). The function of barred plumage patterns appears to be diverse and an alternative function of this type of pattern may be in Batesian mimicry (Davies and Brooke 1988; Davies 2000; Payne and Sorensen 2005; Davies and Welbergen 2008; Davies 2011; Welbergen and Davies 2011). In Batesian mimicry, a harmless species (the mimic) mimics a harmful species (the model) in order to avoid detection (Bates The mechanisms underlying convergent evolution in the plumage patterns of birds 25 1862). The common cuckoo (Cuculus canorus) is thought to mimic the Eurasian sparrowhawk (Accipiter nisus) in order to evade detection while laying its parasitic egg in the nest of other host species. Its striking resemblance is primarily due to the barred plumage patterns that appear to be remarkably similar to its model. Moreover, barred plumage patterns evolved with parasitism in Old World cuckoos in five genera that implies that cuckoo-hawk mimicry may drive cuckoos to converge on the barred plumage patterns of their model raptor species (Kruger, Davies, and Sorenson 2007). A camouflage function of plumage patterns In contrast with a signal function, patterns could allow animals to be less conspicuous against heterogeneous backgrounds (Poulton 1890; Thayer 1909; Cott 1940). Wallace (1889) was the first to consider that less conspicuous coloration may have an adaptive function in concealment. This idea was later taken up by Thayer (1909), who observed that countershading might counterbalance the effects of an animals own shadow thereby concealing its form, and Cott (1940), who thought that animals might modify their colour and pattern to conceal their form. In spite of the importance of camouflage in natural selection, understanding of the function of coloration and patterns in camouflage has progressed slowly since the time of Wallace, Poulton, Thayer and Cott. Recent years have seen a steady increase in the number of studies that test the function of colour and patterning in camouflage due to advances in techniques that allow the incorporation of the visual perspective of the receiver (Stevens and Merilaita 2009b). Most studies of camouflage are in non-avian animals and directly follow the hypotheses of the founding fathers of visual anti-predator defenses. However, the same patterns that have been studied in other animal groups, e.g. irregular mottled or stippled patterns as well as stripes and spots, are also exhibited in birds. Although there is some understanding of visual systems in predators, the main predators of any given avian species are often unknown. Given that there is some understanding of anti-predator defenses in other animals, it would seem that the best place to start in understanding The mechanisms underlying convergent evolution in the plumage patterns of birds 26 the camouflage properties of avian visual patterns, and whether camouflage drives convergent evolution in bird plumage patterns, is to review what is currently known of camouflage in other animal systems. Pattern-based anti-predator defenses can be diverse in their form and the way that they facilitate the evasion of detection by predators. Some types of camouflage conceal animals while stationary, e.g. background matching and disruptive markings, and other types of camouflage prevent capture during movement, e.g. by motion dazzle and flicker fusion (Poulton 1890; Thayer 1909; Cott 1940; Bradbury and Vehrencamp 1998). Irregular mottled plumage patterns (Fig. 1.1) are likely to function in stationary camouflage, where the animal is still, because irregular patterns can generally match the patterning of one or several background types via background matching. Irregular patterns may also facilitate stationary camouflage by seeming to create false sets of edges to prevent recognition by predators via disruptive camouflage (Thayer 1909; Cott 1940; Bradbury and Vehrencamp 1998; Stevens and Merilaita 2009a; Stevens and Merilaita 2009b). For example, in the non-breeding plumage of ducks (e.g. Aythya and Somateria), it has been suggested that irregular patterning may facilitate the evasion of detection by predators during the flightless period accompanying wing moult (Hohman and Richard 1994; Hohman 1996), but this has not been empirically tested. The other avian plumage patterns of scales, bars and spots (regular patterns) may facilitate the camouflage of animals as a second mode of camouflage during movement or flight via motion-dazzle and flicker-fusion (Thayer 1909; Jackson, Ingram, and Campbell 1976; Pough 1976; Endler 1978; Endler 1980; Brodie 1989; Brodie 1992; Brodie 1993; Madsen and Shine 1994; Lindell and Forsman 1996; Stevens and Merilaita 2009b; Scott-Samuel et al. 2011; Helversen, Schooler, and Czienskowski 2013; How and Zanker 2014). For example, experiments testing motion-dazzle using prey capture by human participants on computer screens have shown that striped/barred patterns make ‘capture’ of moving stimuli more difficult and produce more misdirected attacks than other forms of camouflage and coloration (Stevens et al. 2011; Helversen, Schooler, and Czienskowski 2013; How and Zanker The mechanisms underlying convergent evolution in the plumage patterns of birds 27 2014). The visual mechanism underlying the camouflage benefit of motion dazzle is currently unknown, however flicker-fusion may cause fatigue in the motion sensitive cells of the eye (Snowden 1998). Under flicker-fusion, regular patterns blur during movement and cannot be effectively perceived, for example adders (Vipera berus) that have regularly repeating patterns have higher survival rates (Jackson, Ingram, and Campbell 1976; Pough 1976; Endler 1978; Brodie 1989; Brodie 1992; Brodie 1993; Madsen and Shine 1994; Lindell and Forsman 1996). However, it is unknown whether this effect is via flicker-fusion or motion-dazzle. Adaptation and convergent evolution Although studies on the function of animal camouflage have been applied to many animal groups, the existing ideas on camouflage have not been applied to the function of bird plumage patterns. Studies of convergent evolution in plumage phenotypes have predominantly focused on patches of uniform coloration, in particular carotenoid coloration, with relatively fewer studies of melanin (Christidis, Schodde, and Baverstock 1988; Hackett and Rosenberg 1990; Price and Pavelka 1996; Kusmierski et al. 1997; Odeen and Bjorklund 2003; Andersson, Prager, and Johansson 2007; Bleiweiss 2007; Cardoso and Mota 2008; Jones and Kennedy 2008; Prager and Andersson 2010; Friedman, Kiere, and Omland 2011). The proposed adaptive mechanism of convergence in regional colouration has centered on ecology (Crochet, Bonhomme, and Lebreton 2000; Dumbacher and Fleischer 2001; Omland et al. 2006; Bleiweiss 2007; Weibel and Moore 2007), but several studies failed to find an ecological correlate (Omland 1997; Omland and Lanyon 2000; Omland et al. 2006). Additional suggested mechanisms of convergence are sexual selection (Omland and Lanyon 2000; Prager and Andersson 2010), and developmental constraint (Price and Pavelka 1996; Omland and Lanyon 2000; Majerus and Mundy 2003; West-Eberhard 2003; Prager and Andersson 2010). For example, examining the evolution of carotenoid coloration in widowbirds and bishops (Euplectes spp.), coupled with the underlying physiological mechanisms, Prager and Andersson (2010) found that the ancestral state of carotenoid based coloration is The mechanisms underlying convergent evolution in the plumage patterns of birds 28 yellow plumage, and that red coloration is a more derived state that has convergently evolved multiple times. However, convergent evolution in carotenoid based plumage in Euplectes is due to two different mechanisms: ketolase conversion or an increase in carotenoid concentration (Prager and Andersson 2010). Conclusions The prevailing view of plumage pattern function and evolution has focused on adaptationist ideas. There are multiple potential mechanisms underlying convergent evolution that are from the new field of evo-devo as well as the prevailing school of neo-darwinian theory. In my PhD thesis I explore the main mechanisms underlying convergent evolution, from evolutionary developmental biology, as well natural selection for signaling as well as camouflage in the plumage pattern of birds. To do this I used techniques spanning molecular biology, Bayesian comparative modeling, digital image analysis, and ecological selection. Thesis outline Chapter 2: An understanding of the molecular basis of within-feather pattern formation is lacking and in this chapter I investigate two mechanisms of pattern formation using molecular and developmental techniques. Chapter 3: Previous studies indicate that there may be developmental constraint in plumage pattern evolution. Therefore, I investigated whether particular aspects of plumage pattern development may shape the direction of plumage pattern evolution in two ecologically different groups of birds using Bayesian comparative modeling. Chapter 4: It has been demonstrated that plumage patterns can function in signaling in a number of species of birds. In Old World cuckoos, a signaling function may have been co-opted to facilitate deception via Batesian mimicry in many species of birds which I investigated using digital image analysis techniques. This chapter was The mechanisms underlying convergent evolution in the plumage patterns of birds 29 published in Animal Behaviour (2013). Chapter 5: The plumage patterns of birds have inspired many an observer to consider whether they may have evolved to facilitate concealment from predators. Given that patterns are viewed in the context of their habitat the natural extension of a camouflage benefit is that habitat selects for plumage patterns, which I investigated using ecological techniques in all birds worldwide in collaboration with Marius Somveille and Kate Marshall. Chapter 6: In my final chapter I summarise my findings of the mechanisms underlying convergent evolution in the plumage pattern of birds, I discuss the implications of my results and highlight future areas of research that I will pursue. The mechanisms underlying convergent evolution in the plumage patterns of birds 30 The mechanisms underlying convergent evolution in the plumage patterns of birds 31 Chapter 2: Activation or inhibition, MC1R and the developmental basis of within-feather pattern formation Japanese quail (Coturnix japonica) chick. The mechanisms underlying convergent evolution in the plumage patterns of birds 32 The mechanisms underlying convergent evolution in the plumage patterns of birds 33 Abstract Pattern development is an important question in biology as it requires precise spatiotemporal control. Melanin pigmentation is highly conserved in vertebrates and is the basis of feather patterning. Due to shared ancestry with mice, the focus of avian pattern formation has been melanocyte distribution/differentiation and inhibition of MC1R, whereas activation of MC1R has been relatively understudied. Japanese quail is a model system in development biology and has a phenotype comprised of pheomelanin on the ventral surface, whereas dorsal feathers occur in alternating eumelanin and pheomelanin stripes as well as within-feather patterns. In this study, we explored the evolution of regulatory mechanisms underlying differential deposition of eumelanin and pheomelanin via inhibition (ASIP and its paralogue AGRP) and activation (POMC, PC1 and PC2) of MC1R during quail embryogenesis for comparison with adult chicken and adult quail. In situ hybridization using Sox10 revealed that melanocytes are found in all ventral and dorsal feather follicles (FFs) at E8 and E12. RT-PCR results demonstrated some similarity in ASIP alternatively spliced transcripts of adult quail and chicken, but also revealed four novel ASIP transcripts. ASIP transcripts were variably expressed over the ventral and dorsal surfaces. In situ hybridization revealed that ASIP is strongly expressed in ventral FFs in E8 and E12, but not dorsally. Additionally, there was no support for a role of AGRP in regulating pigmentation. POMC transcripts shared only some similarity with the chicken. POMC FF expression was strong at E8 and E12 in both ventral and dorsal FFs. PC1 was not found within FFs. At E8 dorsoventral PC2 FF expression was faint, but at E12, PC2 was strongly expressed in ventral and dorsal FFs. Therefore, the pale-bellied phenotype of quail is due to inhibition of MC1R via ASIP, whereas POMC and PC2 probably contribute to dorsal pigmentation and within-feather patterning. However, it is unknown how pigment-type switching to pheomelanin is controlled dorsally. Nevertheless, our study further supports the role of ASIP in quail pale- bellied phenotype, and that activation of MC1R via POMC and PC2 may contribute to dorsal patterning. The mechanisms underlying convergent evolution in the plumage patterns of birds 34 The mechanisms underlying convergent evolution in the plumage patterns of birds 35 Introduction The developmental basis of plumage patterns is an interesting question, as it is the material that is shaped by natural selection, and the properties of developmental mechanisms have implications for the evolution of plumage phenotypes. Pigmentation patterns are common in mammals and birds, frequently differ between males and females, as well as adults and juveniles, and are derived from well-defined hierarchical developmental modules (Lu et al. 1994; Vrieling et al. 1994; Wilson et al. 1995; Jackson 1997; Krude and Grüters 2000; Prum and Williamson 2002; Prum and Dyck 2003; Steingrímsson, Copeland, and Jenkins 2006; Anistoroaei and Christensen 2007; Candille et al. 2007; Jackson et al. 2007; Fontanesi et al. 2010). The main focus of pattern evolution has been its functional significance in communication and camouflage, including background matching and aposematism (Ruxton, Speed, and Kelly 2004; Stevens and Merilaita 2009a; Allen et al. 2010; Gluckman and Cardoso 2010). Yet, the repeated convergence of these patterns within and between animal groups implies that the evolution of pattern phenotypes may be subject to developmental constraints (Murray 1981; Price and Pavelka 1996; Arendt and Reznick 2008). It is hypothesised that a Turing reaction- diffusion system of melanin synthesis may be responsible for within-feather patterning which relies on the combined action of two molecules (morphogens) (Prum and Williamson 2002). However, this has not been demonstrated empirically and it is currently unknown what these morphogens may be. Pigmentation patterns can be comprised of patches of feathers that have the same pigmentation within feathers, or can be alternately pigmented with different types of melanin within-feathers creating visual patterns such as bars and spots. To produce within-feather patterns requires both spatial and temporal control of pigmentation. The only known type of pigmentation that can be precisely controlled by spatiotemporal mechanisms to produce within-feather patterns is melanin, which is solely produced by melanocytes. The developmental origins of feather-based melanin begin with neural crest cells that follow the dorsolateral migratory pathway before entering the ectoderm and developing feather follicles as melanoblasts. Melanoblasts develop into mature pigment forming cells, melanocytes, that The mechanisms underlying convergent evolution in the plumage patterns of birds 36 synthesise melanin in melanosomes and transfer them to keratinocytes via pseudopodia (Le Douarin and Kalcheim 1999; Prum and Williamson 2002). Two mechanisms of within-feather pattern formation have been demonstrated: melanocyte distribution/differentiation and pigment-type switching via the melanocortin-1 receptor (MC1R). For example, the distribution of melanocytes can lead to complex patterns such as bars and spots, in certain chicken breeds (Lin et al. 2013). The second mechanism acts via differential stimulation of the MC1 receptors on the surface of melanocytes that determines the type of melanin synthesized: inhibition of MC1R leads to pheomelanin (pale yellow/red) whereas activation leads to eumelanin (black/brown) (Bilodeau et al. 2001; Tachibana et al. 2001; Gantz and Fong 2003; Hill and McGraw 2006b). The agouti signaling protein (ASIP) is an inverse agonist of MC1R which has been implicated in mammalian (i) dorsoventral patterning, (ii) temporal-specific regulation of pigment deposition during hair growth leading to banding patterns (Bultman, Michaud, and Woychik 1992; Vrieling et al. 1994; Barsh 1996; Fontanesi et al. 2010; Kaelin et al. 2012), and (iii) inhibition of melanocyte differentiation (Aberdam et al. 1998; Linnen et al. 2009). In chicken and quail, ASIP may function in dorsoventral patterning, within-feather patterning and sexual dichromatism (Nadeau et al. 2008; Oribe et al. 2012; Yoshihara et al. 2012). Due to the importance of ASIP in patterning in mice, coupled with the strong conservation of pigmentation genes in mammals and birds, previous studies on the mechanisms of feather patterning via differential stimulation of MC1R have focused on inhibition, whereas genes involved in activation have been largely overlooked (Yoshihara et al. 2011; Oribe et al. 2012; Yoshihara et al. 2012). The agonists of MC1R are melanocortin peptides. Melanocortins are a family of structurally related peptides that are generated by post-translational processing of the pro-opiomelanocortin peptide (POMC) that is encoded by the POMC gene (Benjannet et al. 1991; Takeuchi, Teshigawara, and Takahashi 1999; Yoshihara et al. 2011). The development of plumage pigmentation in the Japanese quail (Coturnix japonica) is a classic model system (Willier and Rawles 1938; Rawles 1939; Willier and Rawles 1940; Richardson, Hornbruch, and Wolpert 1989; Richardson, The mechanisms underlying convergent evolution in the plumage patterns of birds 37 Hornbruch, and Wolpert 1990; Richardson, Hornbruch, and Wolpert 1991). Quail embryos develop dorsal stripes of dark and pale feathers, with differential pigmentation of eumelanin and pheomelanin within the dark feathers, and a pale belly consisting of pheomelanin-containing feathers. In adulthood, Japanese quail have pronounced patterns on the dorsal surface and retain a pale belly. From early transplantation experiments between several host and donor species (e.g. quail, guinea fowl, and chicken breeds) it was thought that the capacity of a feather follicle to develop patterning was largely the property of the melanocyte (Willier and Rawles 1938; Rawles 1939; Willier and Rawles 1940). For example, the transfer of melanoblasts from donors with barred feather patterns into a non-barred host resulted in patterned feathers in the host, whereas the transfer of melanoblasts from a non-patterned donor into a patterned host did not result in patterns in the host (Willier and Rawles 1940). However, a recent landmark study demonstrated that within-feather patterning can occur either by alternating the presence/absence of melanocytes or by suppression of melanocyte differentiation which is likely due to extracellular factors (Lin et al. 2013). Although this represents an important step forward, Lin et al.’s (2013) study focused on plumage that is due to either black and/or white coloration whereas many plumage patterns, including that of quail, are largely due to differential deposition of eumelanin and pheomelanin. Melanocyte development and differentiation is controlled by the endothelin B receptor (EDNRB), as well as its ligand endothelin 3 (EDN3), and the melanocyte initiation transcription factor (MITF). The EDNRB receptor and EDN3 can cause large changes in the phenotype of vertebrates and chickens (e.g. White leghorn) and quail (Panda mutation) due to impaired differentiation and survival of neural crest cells (Jackson 1997; Niwa et al. 2002; Miwa et al. 2006; Miwa et al. 2007; Saldana- Caboverde and Kos 2010; Kinoshita et al. 2014). Mutations at the MITF locus cause widespread changes in quail and chicken coloration e.g. blue in chicken and silver in quail (Mochii, Ono, et al. 1998; Niwa et al. 2002; Minvielle et al. 2010; Lin et al. 2013). Silver homozygotes in quail are entirely white, whereas heterozygotes retain a reduced amount of within-feather patterning. The major pathways that influence MITF The mechanisms underlying convergent evolution in the plumage patterns of birds 38 transcription are the mitogen-active protein kinase (MAPK) pathway, Wnt signalling pathway and cyclic adenosine monophosphate pathway (cAMP), all of which are known to play a role in pigmentation in mice (Steingrímsson, Copeland, and Jenkins 2006; Poelstra et al. 2014). MAP protein kinases have diverse functions in cell proliferation and gene expression (Cuadrado and Nebreda 2010). However, it is unknown whether this pathway affects pigmentation patterns in birds. Transcriptional activation via the Wnt signalling pathway also serves diverse functions, and regulates transcription of target genes (Mosimann, Hausmann, and Basler 2009). It appears that the Wnt signalling pathway is important in feather formation, however little is known of its function in avian pigmentation (Chang et al. 2004). Defective transport of melanosomes via mutations in the melanophilin gene (MLPH) within follicular melanocytes of vertebrates can cause phenotypic changes leading to an overall white or pale coloration (e.g. silver mutation) (Matesic et al. 2001; Ishida et al. 2006; Anistoroaei and Christensen 2007; Welle et al. 2009). In adult Japanese quail, mutations in MLPH result in a paler phenotype, including dorsal within-feather patterns, but the pattern is still present (Bed’Hom et al. 2012). Melanin synthesis requires the melanocyte-specific enzyme Tyrosinase (TYR) to catalyse the first steps of the pathway. TYR catalyses the reaction of the amino acid tyrosine to dihydroxyphenylalanine (DOPA), and then catalyses the reaction of DOPA to form DOPAquinone (Ito, Wakamatsu, and Ozeki 2000). In chickens, mutations in the Tyrosinase gene have an albino phenotype (Tobita-Teramoto et al. 2000; Chang et al. 2006). Beyond these initial steps of melanogenesis, pheomelanogenesis and eumelanogenesis differ. Eumelanin synthesis requires tyrosinase-related protein 1 (TYRP1) and DOPAchrome tautomerase (DCT) (Ito, Wakamatsu, and Ozeki 2000). In contrast, pheomelanin appears to require few other proteins, but additionally requires cysteine for conjugation of dopaquinone (Hill and McGraw 2006b; Dessinioti et al. 2009; Galván, Ghanem, and Møller 2012). The main physiological reservoir of cysteine is the tripeptide glutathione (GSH) which is synthesised by glutathione S-transferase (GST) (Nataf et al. 1995). A quail mutant, black at hatch (Bh), has a phenotype that lacks between The mechanisms underlying convergent evolution in the plumage patterns of birds 39 stripe and within-feather pigmentation on the dorsal surface (Shiojiri et al. 1999). Both Bh heterozygotes and homozygotes lack dorsal patterning, which is due to an increase in the distribution of eumelanin or pheomelanin, respectively. To elucidate the mechanisms underlying phenotypic differences between wildtype and Bh mutants, Niwa et al. (2002), examined the expression pattern of genes involved in melanocyte development and pigment production, e.g. the Melanoblast/Melanocyte early antigen marker (MelEM) and the Melanosomal matrix protein 15 (Mmp115). MelEM is an Alpha class subunit of GST and marks early melanoblasts that are committed to synthesize eumelanin (Nataf et al. 1995). Mmp115 functions in melanosome production (Mochii, Agata, et al. 1988; Mochii, Takeuchi, et al. 1988; April, Jackson, and Kidson 1998). Most of the study genes involved in melanocyte development (Mitf, MelEm antigen, Kitl, Kit and EdnrB2) did not correlate with pigmentation patterns between wildtype and Bh mutants, except for the MelEm antigen. Therefore, MelEM may play a role in the dorsal pigmentation of quail, but its function in melanin synthesis is currently unknown (Nataf et al. 1995). Of the study genes involved in pigment production (Dct, Tyrp1, Tyr and Mmp115), Dct and Tyrp1 were down-regulated in the melanocytes of homozygote Bh mutants. Mitf actively controls Dct and Tyrp1 (Mochii, Mazaki, et al. 1998) and did not vary between wildtype or Bh mutants (Niwa et al. 2002). Therefore, the Bh locus may act upstream of Dct and Tyrp1, or the Bh locus might be involved in another signaling system that regulates the expression of eumelanogenic enzymes independent of Mitf. Two G-protein coupled receptors are required for normal melanocyte function: EDNRB and MC1R (Jackson 1997). There is currently no evidence to suggest that the EDNRB locus may have a function in within-feather patterning (Kinoshita et al. 2014). However, mutations at the MC1R locus underlie large variation in melanin distribution over the body in a wide variety of mammals and birds, including both mutants and wild-type polymorphisms (Harding et al. 2000; Kijas et al. 2001; Theron et al. 2001; Eizirik et al. 2003; Kerje et al. 2003; Majerus and Mundy 2003; Doucet et al. 2004; Rosenblum, Hoekstra, and Nachman 2004; Mundy 2005; Hoekstra et al. 2006; Nadeau, Burke, and Mundy 2007; Pointer and Mundy 2008; Vidal et al. 2010). The mechanisms underlying convergent evolution in the plumage patterns of birds 40 MC1R is expressed on the plasma membrane of follicular melanocytes in mammals and birds and differential control of this receptor results in either pheomelanin or eumelanin. As mentioned earlier, binding of ASIP to MC1R leads to a decrease in intracellular cAMP, leading to pheomelanin production, whereas binding of the melanocyte-stimulating hormone (MSH) to MC1R increases cAMP production leading to eumelanin synthesis. However, basal activity of MC1R can vary between species. For example, in mice the absence of MC1R stimulation in a POMC mutant does not result in dramatic changes in eumelanin (Slominski et al. 2005). In contrast, activation of MC1R is required for eumelanin synthesis in human hair, because MC1R is expressed ten times less in humans than in mice (Krude et al. 1998; Steingrímsson, Copeland, and Jenkins 2006; Jackson et al. 2007). Therefore, in humans POMC is required for eumelanin pigmentation, whereas in mice it is not. However, mice require ASIP to produce the pheomelanin-containing pale-bellied phenotype. Chicken MC1R expressed in vitro has low basal activity (measured by cAMP production) in the absence of ligands (Ling et al. 2003). Comparing binding affinities of ligands to chicken and human MC1R, the chicken receptor has a much lower affinity than in humans, e.g. binding affinity of α-MSH to chicken MC1R is 363 nM compared to 0.210 nM in humans (Ling et al. 2004). Thus, avian MC1R is highly likely to require activation although the density of melanocyte receptors in avian melanocytes is currently unknown. Two genes inhibit MC1R activity: the first to be discovered in birds was the agouti related protein (AGRP), followed by its paralogue ASIP. Early in vertebrate evolution, AGRP and ASIP arose by gene duplication. In the chicken, AGRP expression was first detected in skin tissue, before ASIP had been discovered, and is co-expressed with melanocortin peptides in the eye (Takeuchi, Teshigawara, and Takahashi 2000). AGRP is expressed in the nervous system of mammals and can have a paracrine effect on MCR function (Boswell, Li, and Takeuchi 2002; Klovins and Schiöth 2006; Nadeau et al. 2008). Given that AGRP can inhibit melanocortin receptors, it is possible that AGRP may play a role in inhibiting MC1R activity in avian pigmentation patterns, but this has not been investigated. The mechanisms underlying convergent evolution in the plumage patterns of birds 41 It has been reported that ASIP may function in phenotypic changes by inhibiting differentiation of melanoblasts (Aberdam et al. 1998; Linnen et al. 2009), e.g. Felidae, cattle, rabbit, quail and chicken (Eizirik et al. 2003; Girardot et al. 2005; Nadeau et al. 2008; Fontanesi et al. 2010; Oribe et al. 2012; Yoshihara et al. 2012). Unpigmented areas of plumage on the ventral surface of wings of Japanese quail chicks are DOPA-negative indicating an absence of melanoblast maturation (Richardson, Hornbruch, and Wolpert 1989). However, this only indicates whether pigment forming cells are capable of producing melanin, and does not indicate whether melanocytes are present or absent, and therefore, whether embryonic pale pigmentation in quail is due to melanocyte distribution and inhibited maturation. A more direct assessment of the expression of melanocyte distribution would be to trace neural crest cells within feather follicles. One such method is to use Sox10 as a marker, for cells of neural crest origin (Cheng et al. 2000). Alternatively, pale pigmentation and within-feather patterning may be due to regulation of MC1R activity without effects on melanocyte distribution/maturation. Different promoter sites of ASIP can lead to spatial and temporal control of pigmentation in laboratory mice and naturally occurring colour variants of the genus Peromyscus (Bultman, Michaud, and Woychik 1992; Vrieling et al. 1994; Fontanesi et al. 2010; Manceau et al. 2011). There are four murine ASIP alternatively spliced variants produced by two kinds of promoters – the distal promoter is ventral-specific and the other is a hair cycle-specific promoter that causes banded “agouti hairs” (Bultman, Michaud, and Woychik 1992; Vrieling et al. 1994; Manceau et al. 2011). In rabbits there are two ASIP alternatively spliced transcripts, each with a different promoter site (Fontanesi et al. 2010). Birds also have multiple ASIP alternatively spliced transcripts, with some evidence to suggest that they contribute to pattern formation and that the distal promoter site is also ventral specific (Nadeau et al. 2008; Yoshihara et al. 2012). Nadeau et al. (2008) were the first to document ASIP gene expression in skin samples of adult quail, quail mutant Yellow (which has a pale phenotype) and the chicken (Gallus gallus). The expression of ASIP was significantly higher on the The mechanisms underlying convergent evolution in the plumage patterns of birds 42 ventral surface in both phenotypes and species examined (Nadeau et al. 2008). Three ASIP alternatively spliced transcripts were reported: 1a is expressed on quail dorsal and ventral surfaces but in the chicken 1a is ventral specific, expression of 1b and 1c is the same in quail and chicken, where the former is ventral specific and the latter is expressed on both the dorsal and ventral surfaces (Nadeau et al. 2008). High variability in the expression of ASIP transcripts between the ventral and dorsal surfaces of quail and chicken was interpreted as variation due to within-feather patterning. An additional four transcripts were subsequently reported in the chicken, that corresponded well with the previously described adult quail and chicken ASIP transcripts, and are transcribed over three promoter sites (classes) within chicken feather follicles: Class 1 consists of five transcripts that have E1S as the non-coding leader exon, whereas class 2 has E4 as the leader exon, and class 3 has E5 (Yoshihara et al. 2012). Of these seven transcripts E1S was consistently expressed and all classes of ASIP were found in chicken dorsal feather follicles. In situ hybridization revealed that ASIP is present in the dorsal feathers of adult chicken where pheomelanin is present, but not eumelanin that indicates that ASIP may have a role in pheomelanin synthesis. However, the relative abundance of the different transcripts was not described and no comparison was made with ventral plumage. ASIP may also play a role in sexual dichromatism in chickens, and it was suggested that the most distal promoter (class 1) produces a pale-bellied phenotype that may have been conserved in mammals and birds (Oribe et al. 2012). Oestrogen treatment of adult male Okayama-Jidori chickens, that have slender barred ornamental feathers in the saddle, resulted in males molting into female like plumage, which consists of broader unbarred pennaceous feathers heavily pigmented with pheomelanin. Up-regulation of class 1 transcripts was correlated with this change in phenotype whereas the class 3 ASIP transcript was expressed on both surfaces, and class 2 could not be detected. However, barred patterns were still apparent in the female induced plumage in males, possibly because males were not castrated resulting in continued circulating testosterone. In addition, the electrophoretic figure for RT-PCR for class 1 mRNA showed a faint band on the dorsal surface of 1 week old chicks and 4 week old juveniles, albeit pale, and class 1 ASIP mRNA was also The mechanisms underlying convergent evolution in the plumage patterns of birds 43 found in dorsal feathers of the chicken (Yoshihara et al. 2012) demonstrating that class 1 is not ventral specific. The POMC gene is thought to have co-evolved with the five melanocortin receptors (MC1R - MC5R) 500 million years ago (Dores and Baron 2011; Dores 2013). The receptors serve in a diverse array of physiological functions: MC1R - pigmentation, MC2R – steroidogenesis, MC3R – homeostasis, MC4R – regulation of food intake as well as sexual function, MC5R – sebaceous gland secretions (Jackson 1997; Kerje et al. 2003; Hoggard 2004; Boswell 2005; Boswell and Takeuchi 2005). The most common melanocortins implicated in pigmentation are adrenocorticotropic hormone (ACTH) and α-MSH (Takeuchi et al. 2003; Takahashi et al. 2006; Roulin et al. 2011; Yoshihara et al. 2011; Roulin and Ducrest 2013). Chicken ACTH contains 39 residues and the first 13 amino acids of ACTH contain the α-MSH peptide (Hayashi, Imai, and Imai 1991). The endoproteases prohormone convertase 1 (PC1) and prohormone convertase 2 (PC2) are responsible for cleavage of POMC products. PC1 cleaves the pro-opiomelanocortin peptide to make ACTH as well as β-lipotropin whereas PC2 cleaves pro-opiomelanocortin or ACTH to make α-MSH or desacetyl-α-MSH (Benjannet et al. 1991; Lu et al. 1994; Ling et al. 2004). In vertebrates the main source of circulating MSH is the intermediate lobe of the pituitary, which is not well defined in humans and birds (Yoshihara et al. 2011). In humans, a lack of a well defined intermediate lobe of the pituitary may have led to the evolution of a cutaneous melanocortin system for integument pigmentation, or vice versa, and perhaps a similar situation has arisen in birds. ACTH and MSH coexist in the same cells in the cephalic half of the anterior lobe of the pituitary in the duck and chicken and have also been detected in chicken feather follicles (Yoshihara et al. 2011). In the chicken it has been found that ACTH binds with a higher affinity to MC1R than MSH and it has been proposed that ACTH may have a function in pigmentation patterns (Ling et al. 2004; Yoshihara et al. 2011). In a study on extracellular MC1R ligands in the silky and Okayama-Jidori breeds of chicken, it was discovered that POMC, PC1 and PC2 are expressed within The mechanisms underlying convergent evolution in the plumage patterns of birds 44 adult feather follicles (Yoshihara et al. 2011). Clear positive signals were found for ACTH using dot-blotting, but MSH was only expressed at trace levels. Investigating patterns of expression in adult feather follicles, four alternatively spliced POMC transcripts were detected that result from two promoter sites (Class a and b), with and without a non-coding exon (Yoshihara et al. 2011). Class a transcripts were only expressed in feather follicles whereas class b transcripts were expressed in the pituitary, hypothalamus, and peripheral tissues, in addition to feather follicles. Given that strong positive signals were only found for ACTH, and that ACTH binds with a higher affinity to MC1R than MSH, the authors suggested that ACTH might have a function in plumage pattern formation. However, the results of this study are difficult to interpret given that two different breeds of chickens that have very different phenotypes were studied (white plumage vs. barred plumage) yet no comparisons were made among the breeds and it is unclear which results related to which breed. Nevertheless, this study is the first to demonstrate expression of an avian POMC gene with alternatively spliced transcripts in chicken feather follicles. In tawny owls (Strix aluco), it was suggested that POMC/PC1/PC2 may mediate a correlation between melanin-based coloration and fitness (Roulin et al. 2011). Therefore, the loci involved in activation of MC1R may have an important contribution to within-feather patterns. The previous studies of ASIP and POMC function in the patterning of chicken and quail did not control for developmental stage (Nadeau et al. 2008; Yoshihara et al. 2011; Oribe et al. 2012; Yoshihara et al. 2012). Embryonic development provides the opportunity to strictly control for developmental stage in the mechanism of pattern formation. In this study, we aimed to identify how melanocyte differentiation or MC1R activity may influence patterning in wild-type quail embryos. In particular, we wanted to test if the species-typical plumage patterning of embryonic quail is due to a) melanocyte differentiation, b) MC1R inhibition, or c) MC1R stimulation, at two stages of feather development. If plumage patterns are due to melanocyte distribution, the Sox10 marker will show that melanocytes are not evenly spread within and between feather follicles on the dorsal surface and few melanocytes will be present in ventral feather follicles. In contrast, if plumage patterns are due to cAMP activity we would The mechanisms underlying convergent evolution in the plumage patterns of birds 45 expect the following: either ASIP or AGRP will be uniformly distributed in feather follicles on the ventral surface to control the pale-bellied phenotype and on the dorsal surface, ASIP/AGRP will be displayed between the dorsal stripes, and within-feathers that have banding patterns. In addition, we would expect ASIP cycle specific transcripts/promoter sites to be variable, whereas the ventral specific transcripts/promoter sites would be consistently expressed in all ventral samples. Finally, if POMC/PC1/PC2 does have a role in pigmentation patterns it will be expressed within the feather follicles, and perhaps these genes will be expressed where ASIP/AGRP are not, thereby completing the study of ASIP and pheomelanin in the chicken by Yoshihara et al. (2012), or perhaps one gene may be constantly expressed and the other varies in a temporal-specific manner. Materials and Methods Fertilized wild-type quail (Coturnix japonica) eggs were obtained from commercial sources. Plumage pigmentation is first visible between 8-9 days, the equivalent of chicken stages 35-36, and is fully developed by 11-12 days, or chicken stage 41-42 (Hamburger and Hamilton 1992; Ainsworth, Stanley, and Evans 2010). Therefore, we harvested embryos at E8 and E12. As there can be variation in developmental stage between individuals, each embryo was checked to ensure that it was at the required stage of development. Embryos where the eyelid had begun to overgrow the surface of the eyeball and had some feather pigmentation on the dorsal surface, but not on the forehead and crown, were considered to be representative of E8. Embryos that had prominent pigmentation and white feather germs around the eye were considered representative of E12. Some of the E12 embryos did not have pigmentation on the feet, a key spotting character. However, all specimens had fully developed plumage and were considered representative of well-developed embryonic plumage appropriate for E12. Tissue and embryos were harvested in phosphate buffered saline (PBS) containing diethylpyrocarbonate (DEPC) treated- distilled water. Samples of ventral and dorsal epidermal tissue were dissected and stored separately in RNAlater at 4 degrees overnight for increased tissue penetration, The mechanisms underlying convergent evolution in the plumage patterns of birds 46 and then at -20 until RNA extraction. For RT-PCR, we sampled three embryos of developmental stage E8 and three embryos of developmental stage E12, and for in situ hybridization we sampled 2-3 individuals of each of these developmental stages. RT-PCR analysis Total RNA was extracted from each sample with an RNeasy mini-kit (Qiagen), with a final elution in 30μl of RNase free distilled water. RNA integrity, purity and concentration (RIN values) were quantified using a BioAnalyser (Agilent). First strand cDNA syntheses were conducted with 1-3μg RNA and 1μl of 150ng/μl N6 primer in a total volume of 10μl using Superscript RT II (Invitrogen) following the manufacturer’s instructions. Reverse transcription polymerase chain reaction (RT-PCR) was carried out to examine which transcripts were expressed in each stage of development on samples from the ventral and dorsal surface of each embryo. RT-PCR was performed in 25μl total reactions containing 1.0 unit BIOTAQ polymerase (Bioline), 2.5μl 10x reaction buffer, 0.75μl 1.5mM MgCl2, 0.2μl 50mM dNTP, 1μl 10μM each primer and 1μl of first strand synthesis product. PCR reactions were performed in a DNA Engine (MJ Research, Watertown, MA), with the following cycling parameters: Heated lid 110°C, 94°C for 2 mins.; 40 x 94°C for 30 secs., 55- 64°C for 30-60 secs., 72°C for 1 min; 72°C for 5 mins. PCR products were visualised on a 1% agarose gel. Sequencing was performed with Sanger sequencing on both strands using the PCR primers. We searched for avian transcripts of our genes of interest on Genbank. Where possible, we used existing primers documented in other studies. To date, no study has demonstrated whether AGRP may have a function in avian phenotypes or that there are alternatively spliced transcripts of AGRP that have a function in plumage pigmentation. Therefore, we focused our sampling efforts The mechanisms underlying convergent evolution in the plumage patterns of birds 47 solely on the coding region of AGRP. Alternatively spliced transcripts and/or coding regions were amplified using published primers and primers designed from available quail and chicken ASIP and AGRP sequences from Genbank (Table 2.1; supplementary Table S2.1). Where available quail coding sequences were relatively short in comparison to the chicken, we designed primers using chicken transcripts to ensure increased sequence coverage, as these sequences are likely to be highly conserved between species. During the course of amplifying alternatively spliced transcripts of ASIP, previously undocumented transcripts emerged: Novel 8a, Novel 8b, Novel 9, Novel 10, Novel 11. For these new alternatively spliced ASIP transcripts we designed reverse internal primers from the 3’ end of the non-coding region to ensure full sequence coverage of the novel exons (Table 2.1). Subsequently, we conducted blast searches on Genbank to determine whether homologous sequences are present in the chicken genome. Similar to our approach to factors inhibiting MC1R, to document the role of activation of MC1R in avian pigmentation we conducted blast searches on Genbank to determine whether there are alternatively spliced transcripts of POMC. Several POMC alternatively spliced transcripts may have a function in pigmentation in chicken (Yoshihara et al. 2011). In addition, there are several unpublished quail POMC alternatively spliced transcripts available on Genbank (Table 2.1). Therefore, we designed primers to amplify the coding region of POMC as well as all chicken and quail alternatively spliced transcripts. For PC1 and PC2, there are no reported alternatively spliced transcripts that function in avian pigmentation. In addition, the previous study that demonstrated PC1 and PC2 expression in feather follicles focused solely on the chicken and did not demonstrate where these genes are being expressed within feather relative to pheomelanin and eumelanin. Therefore, we focused our efforts on testing whether PC1 and PC2 are expressed in quail feather follicles. The mechanisms underlying convergent evolution in the plumage patterns of birds 48 Sequences were aligned and edited using the SeqMan software (DNASTAR Inc., Madison WI, USA). Correspondence between quail and chicken alternatively spliced transcripts was examined by comparing and aligning sequences in Mega 6.06 (Tamura et al. 2013). Primers were subsequently optimized and redesigned as necessary. We used the housekeeping gene β-actin as a positive control for all RT-PCR reactions. Accession numbers of sequences from which the primers were designed, and amplicon size are listed in Table 2.1. The mechanisms underlying convergent evolution in the plumage patterns of birds 49 Table 2.1. Primers used for amplifying mRNA transcripts. The number of exons out of the total number of coding exons is provided as well as the expected sequence length, source of primers, and the accession number of each transcript. Target sequence Primer Primer sequence Non-coding exons (bp) Coding exons (bp) Expected amplicon length (bp) Primer source Accession β-actin Coding sequence ACT1F TGCGTGACATCAAGGAGAAG - Exon 2/2: 250 250 Nadeau 2006 AB199913.1 ACT1R CAGGTCCTTACGGATGTCCA Nadeau 2006 ASIP Coding sequence ASIPF2 TCATTTTCATGACAGTGGGATT - Exon 1-3/3: 465 450 Nadeau 2006 NM_001115079.1 ASIPR2.0 CCTTAACATGTTCCTCATTAGGTTTA Designed E1S E1SF TGAAAAGGAAGCAGAACCAGA 58 Exon 1-3/3: 465 523 Designed AB518061.1 ASIPR2.0 CCTTAACATGTTCCTCATTAGGTTTA Designed E1L E1LF AGTTTTGGAGGTTCATTTCTAATGT 405 Exon 1-3/3: 465 870 Designed AB518065.1 ASIPR2.0 CCTTAACATGTTCCTCATTAGGTTTA Designed E2 E2F TAAACACATTGATGGCATTAACAA 64 Exon 1-3/3: 465 529 Designed AB518062.1 ASIPR2.0 CCTTAACATGTTCCTCATTAGGTTTA Designed E3 E3F GAAGCAGGCAGTCTTCTTGG 72 Exon 1-3/3: 465 470 Designed AB518063.1 ASIPR2.0 CCTTAACATGTTCCTCATTAGGTTTA Designed The mechanisms underlying convergent evolution in the plumage patterns of birds 50 E4 ASIPF8 CCAGCATTTTCATATTTTCTGGA 111 Exon 1-3/3: 465 576 Nadeau 2006 AB518066.1 ASIPR2.0 CCTTAACATGTTCCTCATTAGGTTTA Designed E5 E5F TGAAATCAGTTGTGGCAGGAA 189 Exon 1-3/3: 465 654 Designed AB518067.1 ASIPR2.0 CCTTAACATGTTCCTCATTAGGTTTA Designed Novel 8a NewE1F GTGTGGTTGTGATGGTGATGG 71 Exon 1/3: 239 310 Designed TBA NewE1R GGGAGATCTGGGAGGTTCATT Designed Novel 8b NewE1F GTGTGGTTGTGATGGTGATGG 467 Exon 1/3: 239 706 Designed TBA NewE1R GGGAGATCTGGGAGGTTCATT Designed Novel 9 NewE4F GAGATCTTAAACAGCGCTGCA 372 Exon 3/3: 520 892 Designed TBA E2_R CAGCCTTAACATGTTCCTCATTA Designed Novel 10 NewE1F GTGTGGTTGTGATGGTGATGG 182 Exon 1/3: 239 421 Designed TBA NewE1R GGGAGATCTGGGAGGTTCATT Designed Novel 11 NewE3F TTTTTGGGAGCTGTTGTCCTC 184 Exon 3/3: 520 704 Designed TBA E2_R CAGCCTTAACATGTTCCTCATTA Designed AGRP Coding sequence AGRPF1 CCAGGACCATGCTGAAC - Exon 2/2: 449 449 Nadeau 2006 AB489990.1 AGRPR CAGGAAGATCAGCACCACCT Designed POMC Coding sequence POMCF CTGGGGCTGCTGCTGCTGTGT - Exon 2/2: 717 717 Designed NM_001031098.1 POMCR TGACCCTTCTTGTAGGCGCTTT Designed Promoter A APOMC CCCATAAGCGACTTGCCTTC 169 Exon 2/2: 711 880 Designed AB593424 The mechanisms underlying convergent evolution in the plumage patterns of birds 51 QER CAGAGTCATCAGCGGGGTCT Designed A-2; B-2 POMC B-2 CTCTCCCCCTGCAGCATC 143 Exon 2/2: 711 854 Designed AB593425; AB593427 QER CAGAGTCATCAGCGGGGTCT Designed Promoter B BF AGCGCTCCTCTGCAGTTTG 42 Exon 2/2: 711 753 Designed AB593426.1 BR CAGAGTCATCAGCGGGGTCT Designed T2 QEF2 ACTTCCAGCGTCTCCCAGAG 252 Exon 2/2: 711 963 Designed AB620012.1 QER CAGAGTCATCAGCGGGGTCT Designed T3 QEF1 GATTTCGGAGGCAAAGGATG 175 Exon 2/2: 711 886 Designed AB620013.1 QER CAGAGTCATCAGCGGGGTCT Designed PC1 Coding sequence PC1F CTACGCCAACTATGATCCAAGG - Exon 5-10/14: 847 847 Designed XM_003643060.2 PC1R TTTCCATCTTTTGGGATCAGC Designed PC2 Coding sequence PC2F GGGAGGGAAAGGAAGCATCT - Exon 9-11/12: 800 800 Designed XM_419332 PC2R GGTCTTCTCCCCAAGTGTGTG Designed The mechanisms underlying convergent evolution in the plumage patterns of birds 45 Immunohistochemistry Tissue preparation and procedures used DEPC-H20 and DEPC-PBS as necessary. For in situ hybridization, whole embryos were fixed in modified Carnoy’s (60% ethanol, 11.1% formaldehyde, 10% glacial acetic acid) for four hours, dehydrated in ethanol and cleared in Histosol (National Diagnostics). E12 embryos were cut along the sagittal plane to improve Histosol tissue penetration. Embryos were embedded in paraffin wax (Raymond Lamb) and transversely sectioned with a rotary microtome (Micron). E8 embryos were sectioned at 10μm, however due to the size difference and the extended time required to clear the E12 embryos in Histosol (3 weeks), the E12 embryos were more prone to crumbling and were sectioned at 14μm, which is still adequate for probe tissue penetration. Embryos were cut in serial sections and eight serial sections per embryo per slide were mounted in DEPC-H20 on Superfrost® Plus slides (VWR), and dried overnight at 37°C. All probes were initially tested on a sub-sample of each embryonic stage to test whether optimization was required e.g. probe dilution, temperature of in situ hybridization, length of MABT washes, and temperature in which the colour reaction was applied. All experiments represent a minimum of two replicates of E8 and E12 for each probe and the experiments were considered representative of development where most sections in each slide showed the same result – sections with background or abnormal staining that covered most of a section were excluded from the final interpretation or repeated if necessary. For each round of in situ hybridization experiments, at least one embryo was represented twice, one had the sense probe, and other replicate had the anti- sense probe. Other than the probe applied, the slide with the sense probe was experimentally tested in the same way as the corresponding slide with the anti-sense probe, e.g. the colour reaction was applied for the same amount of time at the same temperature. The sense probe is provided alongside the anti-sense probe from the corresponding tissue from the same section. Sox10 plasmids, containing the full-length chicken 2.2kb Sox10 sequence, were gifts of M. Bronner (Caltech, Pasadena, CA) to C.V.J Baker. For all other probes, primers for the coding sequences used to generate each probe are listed in Table 1. PCR product was ligated into One Shot Top10 chemically competent cells (Invitrogen) using a PCR cloning kit (Qiagen). PCR products were incubated with The mechanisms underlying convergent evolution in the plumage patterns of birds 46 cloning reagents and the vector overnight at 4°C in a total volume of 5 μl using the following amounts: 2x Ligation Master mix = 2.5μl, PCR product = 2μl, pDrive cloning vector = 0.5μl. Cells were transformed and grown according to the manufacturers instructions (Invitrogen). Plasmids were extracted using alkaline denaturation and stored at -20°C until probe synthesis (Birnboim and Doly 1979). To examine the direction of the insert relative to the promoter sites, plasmids were sequenced with Sanger sequencing. Probes were linearised using standard restriction enzymes and transcribed with DIG RNA labeling mix (Roche). The Sox10 anti-sense probe was linearized with HindIII and transcribed from T3. The ASIP, PC1, and PC2 anti-sense probes were linearized with NotI and transcribed from SP6, and the sense (negative control) probes were linearized with PstI and transcribed from T7. The AGRP and POMC anti-sense probes were linearized with PstI and transcribed from T7, and the sense probe was linearized with NotI and transcribed from SP6. In situ hybridization was performed on the paraffin embedded sections with Digoxigenin-labelled Sox10, ASIP, AGRP, POMC, PC1 and PC2 probes diluted in hybridization mix, separately. Each probe was hybridized overnight at 68-72°C in a Boekel slide incubator. After hybridization, slides were washed twice in 50% formamide, 50% 1xSSC and 0.1% Tween-20 at 65°C, then twice in MABT (0.1 M maleic acid, 150 mM NaCl, 0.1% Tween-20, pH 7.5) at room temperature. To block binding of nonspecific antibody, slides were incubated in 70% MABT, 20% natural sheep serum and 10% blocking reagent (Roche) for 2 hours. AP-conjugated anti- digoxigenin antibody (Roche) was diluted in this blocking solution at a dilution of 1/1500 and applied to slides, which were covered with parafilm and kept at room temperature overnight. Slides were subsequently rinsed in MABT 5 times for a minimum of 30 mins each. Slides hybridized with POMC probe were further rinsed in MABT overnight to reduce background. Slides were equilibrated in NTMT (100 mM NaCl, 50 mM MgCl2, 100 mM Tris pH 9.5, 0.1% Tween-20) twice for 10 mins each. To reveal colour, NBT-BCIP stock solution (Roche) was diluted in NTMT at a dilution of 1/1500 and applied to slides and covered with parafilm. The colour reaction of all probes, except for POMC, was performed at room temperature whereas the POMC colour reaction was conducted at 4°C. The mechanisms underlying convergent evolution in the plumage patterns of birds 47 The genes examined in our experiments have additional neuroendocrine functions as well as a putative function in pigmentation in the epidermis providing positive control of the anti-sense probe within sections (Takeuchi and Takahashi 1999; Takeuchi, Teshigawara, and Takahashi 2000; Ling et al. 2004; Nadeau et al. 2008; Yabuuchi et al. 2010; Roulin et al. 2011; Yoshihara et al. 2011; Saneyasu et al. 2013). Sections depicted in the figures are representative of staining between and within samples. Results Melanocyte distribution Eumelanin is present in dorsal feather follicles at E8 but is more pronounced at E12 whereas pheomelanin is present in ventral feather follicles, with a small number of feather follicles in the flanks pigmented with eumelanin. In situ hybridization with Sox10 revealed that melanocytes are present at E8 but that there is variation in distribution over the ventral and dorsal surface. On the ventral surface of E8 quail embryos, Sox10 staining was apparent only in some feather follicles but was present throughout all feather follicles on the dorsal surface. In addition, Sox10 staining was apparent in the epidermis on both ventral and dorsal surfaces (Fig. 2.1). At E12 Sox10 was consistently expressed throughout feather follicles on both the ventral and the dorsal surface, as well as in the epidermis. In feather follicles at both stages of development, there was Sox10 staining both where melanin is present, as well as where it is absent. The mechanisms underlying convergent evolution in the plumage patterns of birds 48 Fig. 2.1. Melanocyte distribution in Japanese quail feather follicles during embryonic development at E8 and E12. At E8 and E12 melanin is rarely present in ventral feather follicles (FFs) but is frequently observed in many FFs on the dorsal surface. Melanocytes are present within the epidermis and feather follicles. Staining of Sox10 occurs in areas where individual barbs are developing in feather follicles on both the ventral and the dorsal surface of E8 and E12 quail embryos. The distribution of melanocytes is not restricted to areas pigmented with melanin but also occurs in areas where melanin is absent. Genes inhibiting MC1R in melanocytes We investigated 2 loci that inhibit MC1R: ASIP and its paralogue AGRP. Previously, it has been reported that adult quail have three ASIP alternatively spliced transcripts (1a, 1b, 1c), whereas the chicken has seven ASIP isoforms that are distributed among three promoter sites (classes) (Nadeau et al. 2008; Yoshihara et al. 2012). Aligning the non-coding regions of these ASIP isoforms reveals that there is close correspondence (Fig. 2.2). Quail transcript 1a, although short, matches chicken E1L, quail 1b matches chicken E3, and quail 1c matches chicken E4. However, neither 1a nor 1b is an exact match to the transcripts in chicken, as E1S is not alternatively spliced in either transcript. In contrast, 1c and E4, which are The mechanisms underlying convergent evolution in the plumage patterns of birds 49 comprised of only one non-coding exon, is an exact match. Fig. 2.2. Correspondence of ASIP alternatively spliced transcripts between embryonic Japanese quail, adult quail and chicken. Empty boxes represent non- coding exons and solid boxes denote coding exons. Chicken promoter sites are underlined in green, chicken transcripts begin with “E”, and previously reported adult quail transcripts begin with “1” and are denoted with the corresponding chicken transcript with “==” (Nadeau et al. 2008; Yoshihara et al. 2012). Green lines and green boxes represent promoter sites and transcripts, respectively, that were successfully amplified in quail embryonic development, whereas red denotes transcripts that were not amplified. A blue line and blue empty boxes indicates a novel promoter site and novel alternatively spliced non-coding exons (begins with “N”), respectively, that have not been previously documented. We report four new ASIP alternatively spliced transcripts (Fig. 2.3). Although we found 5 new alternatively spliced transcripts (Novel 8a, 8b, 9, 10, 11), we were only successful in sequencing Novel 8a once, which is similar in the 5’ non-coding The mechanisms underlying convergent evolution in the plumage patterns of birds 50 sequence to Novel 8b. As such, we are uncertain if Novel 8a represents a new non- coding exon and discarded it from the subsequent results. The ASIP gene is located on chromosome 20. A refseq_genomic blastn search on Genbank against the Gallus gallus genome revealed that most of these novel transcripts are located upstream of ASIP. For Novel 8b, we did not find any significant similarity with genes on chromosome 20. Novel 9, 10 and 11 are located on the 5’ side of the ASIP locus. The distance from the non-coding exons to ASIP is 40kb for novel 10 and 11 (Fig. 2.2, 2.3). Novel 11 is comprised of a new leader non-coding exon, alternatively spliced with the previously described non-coding exon 2 found in chicken, and is located between E1L and E2 (Yoshihara et al. 2012). #Novel 8a; #Novel 8b CAACGACTTGAAGTCAGGGCCGTCTTCTTGGCTACATGGCTGAG CAACGAGCTACCTTCAGGTGGTGTGAGGCTGGCAATGGGGTGCAGATGGGGCCTCATATCTC TGAGTAAATGGAACAGGCAGATGAGAAGGGCACAGCTGTGCTCTTCCACCTCAGCATGCTCC CTGGACATCCATGCCTCTTGGCTGCCCACTAACCTCAGCTGGAGAGTGCACAGTGATTAATA TTAGTGAATCTGATCTAAGCCTATTGCAAACCATTGATCTGAGCGGGCTGCACTGCTCCAAA AATAGCCTCACAATATTACTGAGGCATAGCTCTTGGTGTGGTGAGCTGCTCAAGCTGCCTAA TCACATACAAGACCCCAGGCTCTGTATGTTCTCAATGCTTTTTTCCTGTGACCACTGCTCAT CACACTTGGCTACTGTGGGTTAATACAGCTATGTCATGTTCAGCCAAATCAGCTGTCACATC AGCTCTCCATTT #Novel 9 Cj CAGTGACTACATCTGGAAAAGAAAAAAGAAGAAACATGAAGATGACCATACGCCGTTGTA ||||||| ||||||||||||||||||| |||||||||||||||||| ||||| |||||| Gg CAGTGACAACATCTGGAAAAGAAAAAA--AGAAACATGAAGATGACCGTACGCTGTTGTA Cj CAGTGACTACATCTGGAAAAGAAAAAAGAAGAAACATGAAGATGACCATACGCCGTTGTA ||||||| ||||||||||||||||||| ||||||||||||||||||| ||||| |||||| Gg CAGTGACAACATCTGGAAAAGAAAAAA—-AGAAACATGAAGATGACCGTACGCTGTTGTA Cj TTTCCTTCTCAGATGAGTTATATTTTGAAAGCAGCAATAAAAAA-TTCTCAGATAAGGGA |||||||||||||||||||||||||||||||||||||||||||| ||||| | ||||||| Gg TTTCCTTCTCAGATGAGTTATATTTTGAAAGCAGCAATAAAAAACTTCTCGGGTAAGGGA Cj TTGGAGAAGTGATGGATTTGTGGCAGCGGCTGTCTTTTGGTCCTTCGTTGACTGCTTTGG |||||||||||||||| |||||||||||||||||||||||||||| | ||||| | ||| Gg TTGGAGAAGTGATGGAATTGTGGCAGCGGCTGTCTTTTGGTCCTTGGCTGACTATTCTGG The mechanisms underlying convergent evolution in the plumage patterns of birds 51 Cj GGTTGA--ATTTTTCTTTTTAATGTTCACATGTGTGTGCCTTTTTTGGGGGAAATGGGGC |||||| ||||||||||| ||||||||| |||||| |||||||| ||| ||||||||| Gg GGTTGATTTTTTTTCTTTTTGATGTTCACACGTGTGTACCTTTTTT-GGGAAAATGGGGC Cj CAGATTCCCCCCTGGCTGAGAGTCTACAGCTTCTTAAAGGTGAAGGCTGTTCCAGT |||||||||||||||||||||| ||||||||||||||||||||||||||||||||| Gg CAGATTCCCCCCTGGCTGAGAGGCTACAGCTTCTTAAAGGTGAAGGCTGTTCCAGT #Novel 10 Cj TTGTGGCAGCGGCTGTCTTTTGGTCCTTCATTGACTGCTTTGGGGTTGA--ATTTTTCTT |||||||||||||||||||||||||||| ||||| | ||||||||| |||||||| Gg TTGTGGCAGCGGCTGTCTTTTGGTCCTTGGCTGACTATTCTGGGGTTGATTTTTTTTCTT Cj TTTAATGTTCCACCATGTGTGTGCCTTTTTTGGGGGAAATGGGGCCAGATTCCCCCCTGG ||| |||||| || |||||| |||||||| ||| |||||||||||||||||||||||| Gg TTTGATGTTC--ACACGTGTGTACCTTTTTT-GGGAAAATGGGGCCAGATTCCCCCCTGG Cj CCTGAGAGTCTACAGCTTCTTAAAGGTGAAGGCTG-TCCAGT ||||||| |||||||||||||||||||||||||| |||||| Gg -CTGAGAGGCTACAGCTTCTTAAAGGTGAAGGCTGTTCCAGT #Novel 11 + exon 2 Cj TTCCAGGACCAGCAACTCTTATTTATAAATGCTATACAGCGTATATGTTAAGATTTATAC |||||| ||||||| |||||||||||||||| ||||| ||||||||||| |||||||||| Gg TTCCAGAACCAGCAGCTCTTATTTATAAATGTTATACCGCGTATATGTTCAGATTTATAC Cj AAGATGTTTTCATGGTACGCATAGAATTCTGAACCCTGTAAACACACTGATGGCATTAAC |||||||||||||||||||||||||||||||||||||||||||||| ||||||||||||| Gg AAGATGTTTTCATGGTACGCATAGAATTCTGAACCCTGTAAACACATTGATGGCATTAAC Cj AAGGACCAGATGTCAGTTGCCTTTCTCATAAAGAAAACAACAAG |||||||||||||||||||||||||||||| ||||||||||||| Gg AAGGACCAGATGTCAGTTGCCTTTCTCATATAGAAAACAACAAG Fig. 2.3. New ASIP alternatively spliced transcripts that have not been previously documented in the avian literature. Novel 8b had no significant similarity with the chicken genome. The first six base pairs of the 5’ non-coding sequence of Novel 8a (base pairs written in grey) and Novel 8b is the same as indicated by an underline. We were only able to sequence Novel 8a once and given its similarity to Novel 8b, Novel 8a may not represent a new alternatively spliced transcript. Novel 9-11 (Cj: The mechanisms underlying convergent evolution in the plumage patterns of birds 52 Coturnix japonica) are aligned to the chicken genome (Gg; Gallus gallus). A vertical bar - | - indicates matching base pairs and a space indicates no match, whereas dashes indicate indels. Novel 11 contains a new leader exon alternatively spliced with the existing chicken non-coding exon 2 (Yoshihara et al. 2012) which is highlighted in grey. RT-PCR analyses revealed that ASIP was always expressed on both the dorsal and ventral surfaces and in both stages of quail embryonic development (Table 2.2). Non-coding exon E3 as well as alternatively spliced transcripts that combine more than one non-coding exon, such as the leader exon E1S, could not be amplified despite multiple attempts with numerous primer pairs. Of the isoforms that we successfully amplified by designing primers based on previously reported transcripts, there was considerable variation in expression at both the temporal and spatial scale (Fig. 2.2; Table 2.1, 2.2; supplementary Table S2.1). Exon E1S was consistently expressed on the ventral surface in both stages of development, but was only present on the dorsal surface at E12. Exon E1L was only present on the ventral surface at E8 and was not present later in development on either surface (Table 2.2). E2 was consistently present on the ventral surface at both stages of development, and was variable on the dorsal surface at E8 but was absent on the dorsal surface of E12. E4 was variable in expression on the dorsal surface early in development but was consistently present on the ventral surface. At E12, E4 was also consistently present on the ventral surface, but was not expressed on the dorsal surface. Similarly, E5 was consistently present on the ventral surface at both early and late stages of development, and was consistently present on the dorsal surface early in development, but in E12, E5 was variable in expression on the dorsal surface. For the primer pairs targeting E1S, E2 or E3, we did not find multiple bands in agarose gels, indicating that these exons do not appear to be alternatively spliced with other non-coding exons during quail embryonic development, as occurs in the chicken. Similar to the existing ASIP alternatively spliced transcripts there was variation in the expression of ASIP Novel 8b and Novel 10 within and between developmental stages. In contrast, ASIP Novel 11 and Novel 9 were expressed on both the ventral and dorsal surfaces in all embryos at stages E8 and E12 (Fig. 2.4; Table 2.2). The mechanisms underlying convergent evolution in the plumage patterns of birds 53 Table 2.2. Patterns of exon expression for ASIP and AGRP in embryonic quail at developmental stage E8 and E12 determined by RT-PCR. “+” = present, “-“ = absent, “+/-“ = variable expression. Target  sequence   E8   E12   β-actin Dorsal   +   +       Ventral   +   +   ASIP  coding     exons  (6-­‐8)   Dorsal   +   +   Ventral   +   +   E1S   Dorsal   -­‐   +   Ventral   +   +   E1L   Dorsal   -­‐   -­‐   Ventral   +   -­‐   E2   Dorsal   +   -­‐   Ventral   +   +   E4   Dorsal   +/- -­‐   Ventral   +   +/- E5   Dorsal   +   +/-   Ventral   +   +   Novel  8b   Dorsal   +   +/-­‐     Ventral   +/-­‐   +   Novel  9   Dorsal   +   +       Ventral   +   +   Novel  10   Dorsal   +/-­‐   -­‐     Ventral   +/-­‐   +/-­‐   Novel  11   Dorsal   +   +     Ventral   +   +   AGRP  coding   exons  (1-­‐2)   Dorsal   +/- -­‐       Ventral   +/- +/- The mechanisms underlying convergent evolution in the plumage patterns of birds 54 Fig. 2.4. mRNA expression, detected by RT-PCR, of the previously undocumented ASIP alternatively spliced transcript Novel 11 in Japanese quail. The dorsoventral surface is indicated with brackets and the developmental stage is above the gel lanes: V = Ventral, D = Dorsal, N = Negative control. RT-PCR results are provided for ventral and dorsal surfaces for three samples of each stage, followed by a negative control. The second locus that can inhibit MC1R is the ASIP paralogue AGRP. As there is no previous evidence to suggest that AGRP may have a function in patterns, or that there are alternatively spliced transcripts that have a role in pigmentation, we focused on the expression patterns of the coding sequence. AGRP expression was detected a both developmental stages. Expression at E8 was variable on both the ventral and dorsal surfaces, whereas at E12, it was not expressed dorsally and had variable expression ventrally (Table 2.2). In situ hybridization confirmed that ASIP is expressed early in development at E8 and continued into later stages of development at E12 (Fig. 2.5). In both early and late stages of development, ASIP is strongly expressed within the pulp of developing feather follicles on the ventral surface and many ventral cells surrounding the dermal pulp in the epidermis exhibit ASIP nascent transcription. Some feather follicles on the flanks of the ventral surface near or on the wing have eumelanin pigmentation and a minority of these feather follicles had some staining for ASIP in the feather pulp both where eumelanin is present as well as where it is absent (Fig. 2.6). In contrast, in feather follicles on the dorsal surface of E8 and E12 embryos, there was no staining for ASIP in the dermal pulp but there does appear to be faint nascent transcription within some cells of the feather pulp and the epidermis. The pattern of expression of The mechanisms underlying convergent evolution in the plumage patterns of birds 55 ASIP on the dorsal surface did not vary between feather follicles that are pigmented with eumelanin and those that are not pigmented. Fig. 2.5. In situ hybridization results for ASIP and AGRP in quail feather follicles at E8 and E12. ASIP is expressed in the dermal pulp of ventral feather follicles in E8 and E12 embryos, but is not visible on the dorsal surface. AGRP was not expressed in the feather follicles in early or late development, either ventrally or dorsally. Representative anti-sense and sense probes are shown in the spinal cord of E12 in ASIP and E8 in AGRP. The mechanisms underlying convergent evolution in the plumage patterns of birds 56 Fig. 2.6. ASIP expression in a feather follicle with eumelanin on the flank of an E12 Japanese quail embryo. In some feather follicles in the flank or on the wing of E12 embryos ASIP expression was observed with eumelanin pigmentation. Staining of AGRP was less prevalent than ASIP (Fig. 2.5). Unlike ASIP, AGRP was not strongly expressed in the pulp of developing feather follicles on the dorsal or ventral surface, at either stage of development or in the epidermis. However, nascent transcription of AGRP was commonly observed on the ventral and dorsal surface of E8 as well as E12 in feather follicles that contain melanin as well as those that did not have melanin pigmentation (Fig. 2.7). Fig. 2.7. Representative nascent transcription of AGRP in ventral developing feather follicles in E8 Japanese quail embryos. Similar staining was found in feather follicles on the dorsal surface of quail E8, as well as both the ventral and dorsal surface of quail E12, and the epidermis. The mechanisms underlying convergent evolution in the plumage patterns of birds 57 Genes activating MC1R in melanocytes We investigated 3 loci that are involved in activation of MC1R: POMC, PC1 and PC2. Yoshihara et al. (2011) previously described four alternatively spliced transcripts of POMC in chicken encompassing two promoter sites that are transcribed with and without an additional non-coding exon: A-1, A-2, B-1, and B-2. The chicken distal promoter site A is located 14kb upstream in another gene, ANGPTL7, whereas promoter site B is located within the POMC gene. In addition, three unpublished adult quail POMC alternatively spliced transcripts are available on NCBI: T1, T2, and T3 (Fig. 2.8; Table 2.1). Aligning these sequences, there is no correspondence between the chicken POMC A-1 and the quail transcripts. Chicken POMC B-1 corresponds well with quail T1 indicating that adult quail possess promoter site B. Chicken POMC A-2 and B-2 partially correspond with quail T2 and T3 as each possesses a chicken POMC non-coding exon but does not possess the corresponding chicken leader exon (Yoshihara et al. 2011). Finally, quail POMC T2 and T3 are similar in possessing a 5’ quail non-coding exon, as well as the chicken non-coding exon, but T3 also possesses an additional 3’ quail non-coding exon before the POMC coding region (Fig. 2.8). The mechanisms underlying convergent evolution in the plumage patterns of birds 58 Fig. 2.8. Correspondence between alternatively spliced transcripts of POMC in embryonic Japanese quail and chicken. Adult chicken promoter sites and alternatively spliced transcripts are listed on the left, and the corresponding quail transcript is listed to the right of the transcript. Quail POMC transcripts that have no correspondence with chicken transcripts are listed with the novel promoter site (Yoshihara et al. 2011). Empty boxes represent non-coding exons and solid boxes denote coding exons. The non-coding exons that were amplified in quail embryonic development are represented with green, whereas red denotes the promoter site and non-coding exons that we were unable to amplify. RT-PCR analyses revealed that the coding exons of POMC are always expressed on both the dorsal and ventral surfaces in quail E8 and E12 (Table 2.3). The chicken POMC non-coding exon POMC A-1 and quail T2 could not be amplified despite multiple attempts with numerous primer pairs (Fig. 2.8; Table 2.1; supplementary Table S2.1). Transcripts arising from promoter site B were present in almost all ventral and dorsal samples, at both E8 and E12, the exception being one dorsal sample in E8. The 3’ quail non-coding POMC exon was also variable in its expression, as it was amplified in two samples from each of E8 and E12, and was expressed in both ventral and dorsal samples. The mechanisms underlying convergent evolution in the plumage patterns of birds 59 Table 2.3. Patterns of expression for POMC in quail at embryonic stage E8 and E12 determined by RT-PCR. “+” = present, and “+/-“ = variable expression. In situ hybridization revealed that POMC is expressed in the epidermis and feather follicles on the dorsal and ventral surface of quail E8 (Fig. 2.9). In addition, nascent transcription of POMC is frequently observed throughout feather follicles on both ventral and dorsal surfaces, in early and late stages of quail embryonic development. Within feather follicles, POMC is apparent in the feather cuticle as well as the feather pulp and this pattern of expression is consistent on the ventral and dorsal surface, both where melanin is present and where it is absent. This pattern of expression is the same in E12 embryos. The other loci that are involved in activation of the MC1R pathway are the endoproteases PC1 and PC2 that cleave POMC products to make ACTH and MSH, respectively. As there is no evidence to suggest that there are alternatively spliced transcripts of PC1 and PC2 that have a role in feather follicle pigmentation, we focused on the expression patterns of the coding sequence. RT-PCR revealed that PC1 was present in nearly all samples on the ventral and dorsal surface at E8 and E12, the exception being one ventral sample from E8 (Fig. 2.9). Similarly, PC2 was present in all dorsal and ventral samples, in both stages of development examined. In situ hybridization demonstrated that there is nascent transcription of PC1 Target   sequence   E8   E12   POMC Dorsal   +   +     Ventral   +   +   Promoter  B   Dorsal   +/- +     Ventral   +   +   3'  quail  non-­‐ coding  exon   Dorsal   +/-­‐   +/-­‐       Ventral   +/-­‐   +/-­‐   The mechanisms underlying convergent evolution in the plumage patterns of birds 60 within the feather pulp and the feather cuticle on the ventral and dorsal surface in E8 quail embryos (Fig. 2.9). At E12 the low level of expression of PC1 had not progressed further than nascent transcription. PC2 is faintly expressed in feather follicles and the epidermis at E8 on both ventral and dorsal surfaces. At E12, PC2 was also faintly expressed in the epidermis but there was strong expression of PC2 within the feather pulp as well as the feather cuticle (Fig. 2.9). The patterns of expression were consistent in feather follicles over both the ventral and dorsal surface where melanin is expressed as well as where it is absent. Fig. 2.9. POMC-, PC1- and PC2- Patterns of expression of POMC, PC1 and PC2 in feather follicles at E8 and E12. POMC is expressed in the dermal pulp of ventral and dorsal feather follicles in E8 and E12 embryos. Nascent transcription of PC1 was observed in developing feather follicles but did not develop further at E12. Similarly, there is nascent transcription of PC2 at E8 on both ventral and dorsal surfaces, but when feather development is more progressed (e.g. E12), PC2 is strongly expressed The mechanisms underlying convergent evolution in the plumage patterns of birds 61 in feather follicles on both the ventral and dorsal surface. Representative anti-sense and sense probes are shown in the dorsal epithelium of E12 in POMC and PC1, as well as in the spinal cord of E12 in PC2. Discussion Our data provide some of the first evidence that extracellular ligands involved in MC1R activation and inhibition are involved in producing phenotypic variation over the ventral and dorsal surfaces in quail embryogenesis. Sox10 staining demonstrated that the species-typical dorsal pigmentation in embryonic quail is not due to melanocyte distribution. We discovered novel quail ASIP transcripts, that have not been previously described that may indicate evolution in the mechanism of inhibition of MC1R (Nadeau et al. 2008; Yoshihara et al. 2012). Similar to previous studies, we found that the pale-bellied phenotype of quail is consistent with inhibition of MC1R via ASIP, and that this mechanism has been conserved between mammals and birds. In contrast with mice, dorsal temporal specific patterning within-feathers does not appear to be due to regulation of ASIP. Of the six POMC transcripts previously described, only one is conserved between quail and chicken whereas the other POMC transcript possessed a 3’ non-coding exon that may be quail specific. Both of these transcripts showed variable expression over the ventral and dorsal surface. Nevertheless, the POMC coding region is expressed within dorsal and ventral feather follicles demonstrating that activation of MC1R may be required to produce eumelanin. In quail embryonic development, there is little expression of PC1 within feather follicles whereas PC2 is expressed strongly in later quail embryonic development. Therefore, it may be that POMC products are directly cleaved by PC2 to make MSH. Together, this suggests that MC1R is differentially stimulated across the ventral and dorsal surface to create variation in phenotypes. There was considerable variation in the expression of the seven previously described alternatively spliced transcripts of ASIP, between adult quail, chicken and embryonic quail (Nadeau et al. 2008; Yoshihara et al. 2012). Of these seven, only three transcripts were observed in embryonic quail, and only one of these three has previously been reported in adult quail (E4). In rabbits and mice, the distal ASIP promoter leads to ventral specific coloration, and the proximal promoter performs The mechanisms underlying convergent evolution in the plumage patterns of birds 62 temporal control of coloration during hair development (Bultman, Michaud, and Woychik 1992; Vrieling et al. 1994; Fontanesi et al. 2010; Manceau et al. 2011). Two of the three ASIP transcripts identified in adult quail have a leader exon of E1S, which corresponds to the chicken distal promoter, and these transcripts are expressed both dorsally and ventrally in adult quail. In embryonic quail, only one alternatively spliced transcript out of nine has a leader exon comprised of E1S (Fig. 2.3). Of the alternatively spliced sequences expressed in quail embryonic development, similar to the chicken but dissimilar to adult quail, the ASIP non-coding exon E1L is ventral specific but lacks the leader exon E1S. However, unlike adult quail and chicken, ASIP E3 could not be amplified in embryos (Fig. 2.3). We report on four undescribed alternatively spliced transcripts of ASIP, one of which is alternatively spliced with E2 and three out of four of these transcripts are also on the chicken genome (Fig. 2.3, 2.4). The novel non-coding exon N9 is positioned next to E2, with which it is alternatively spliced, whereas two of the novel non-coding exons are located a further 40kb upstream of ASIP on chromosome 20 (N10 and N11) and the position of N8b is unknown. Annotation of the quail genome is likely to reveal a similarity in exon position in the genome and should improve the resolution of N8b. However, as our intention was to document the distribution of existing ASIP transcripts between embryonic quail, adult quail and the chicken, we did not perform 5’ RACE that would clarify the number of ASIP transcripts, and which promoter sites initiate transcription. Nevertheless, our findings indicate surprising variation in the expression of ASIP alternatively spliced transcripts between embryonic quail, adult quail, and chicken breeds which could indicate that there is variation between developmental stages in quail, as well as evolution within the mechanism of the pale-bellied phenotype between different species/breeds of chicken that may be lineage specific. We anticipate that there may be further variation in ASIP transcripts in other species of birds. The other locus we examined that may inhibit MC1R, AGRP, did not appear to influence ventral pigmentation or within-feather patterning (Fig. 2.5). RT-PCR results demonstrated that AGRP is variable in expression over the ventral and dorsal surface of developing quail embryos, which would be consistent with a temporal specific function (Table 2.2). Although nascent transcription of AGRP in E8 feather follicles was found, this did not progress further than nascent transcription at E12 The mechanisms underlying convergent evolution in the plumage patterns of birds 63 (Fig. 2.7). If AGRP has a ventral specific function in inhibiting MC1R the expectation would be that it would have been expressed in all ventral samples, but this was not the case (Table 2.2). In situ hybridization with the ASIP and AGRP probes failed to demonstrate expression of these genes on the dorsal surface of embryos, in contrast to RT-PCR results. This may be due to a difference in the sensitivity of these techniques (Fig. 2.7; Table 2.2). RT-PCR results demonstrated that ASIP is consistently expressed on both the dorsal and ventral surface of E8 and E12 embryos, yet in situ hybridization revealed that ASIP was strongly expressed within developing feather follicles only on the ventral surface of embryos of both developmental stages. Perhaps ASIP expression on the dorsal surface of quail embryos is important in pigmentation patterns but was not detected by in situ hybridization or perhaps ASIP has a more prevalent function in dorsal epidermis, which we did not control for in RT-PCR. However, a difference in the sensitivity between RT-PCR and in situ hybridization is unlikely to be the cause of a discrepancy between techniques in the case of AGRP given that the expression of the AGRP coding sequence was variable within and between samples (Table 2.2). The gene involved in activation of MC1R, POMC, appears to have few alternatively spliced transcripts (Fig. 2.2, 2.8; Table 2.1, 2.3). Similar to ASIP, we found that POMC transcripts varied in their patterns of expression between quail embryonic development and the chicken. Of the POMC transcripts expressed, we did not find any that are dorsal specific. However, this may be due to a lack of detailed 5’ information. A future study will investigate POMC alternatively spliced transcripts and promoter sites utilizing 5’ RACE in quail. To definitively test whether MC1R activation can occur without POMC peptides, experimental downregulation of POMC is required, but given the low basal activity of MC1R in chicken in comparison with mice and humans (Jackson 1997; Ling et al. 2003; Ling et al. 2004) it is likely that POMC is required for activation. In addition, an important follow up study would be to examine the distribution of MC1R within feather follicles. Our main finding, that POMC is transcribed in feather follicles on the dorsal surface of early and late quail embryonic development, indicates that the generation of within-feather patterning may require activation of MC1R. It has been reported that The mechanisms underlying convergent evolution in the plumage patterns of birds 64 ACTH has a higher binding affinity to MC1R than MSH (Ling et al. 2003). However, in the feather follicles of quail embryos in early and late stages of plumage development PC2, but not PC1, is expressed (Fig. 2.9). The patterns of expression of PC2 correlates well with pigmentation in embryogenesis: both pigmentation and PC2 are less prevalent at E8 whereas pigmentation and PC2 are strongly expressed at E12 (Fig. 2.9). Therefore MSH, not ACTH, is probably responsible for activation of MC1R to produce dorsal pigmentation patterns in quail. In contrast chicken breeds express PC1, PC2 and POMC within feather follicles (Yoshihara et al. 2011). However, the results of the chicken study are unclear given that the two breeds used in the study are apigmented (Silky chicken) and barred (Okayama-Jidori) and there was no comparison made between the patterns of gene expression between these phenotypes. Nevertheless, it is interesting that PC1 is expressed within chicken feather follicles and this suggests differences in the colour mechanism between wildtype quail and chicken breeds. But it is unclear if these differences are representative of evolution or artificial selection in the chicken. There are three avian MSH peptides, α-MSH, β-MSH and γ-MSH, which vary in similarity to human peptides (Ling et al. 2004). The amino acid residues of α-MSH are identical in human and chicken, whereas β-MSH varies in three positions, and γ- MSH is the most variable. It is thought that the ACTH peptide may have been an important ligand for all MCRs in the ancestral vertebrate but that during vertebrate evolution, the other MSH subtypes evolved specificity for MCR types. For example, α-MSH has evolved specificity for MC1R in humans. There are three forms of α- MSH: desacetyl-, monoacetyl-, and diacetyl-α-MSH and the MC1 receptors of chicken have a slightly higher affinity for the desacetylated peptide, which is the product that is directly cleaved from POMC peptides. Given that activation of MC1R may be required for melanin synthesis, similar to humans, perhaps the desacetylated α-MSH peptide functions in plumage pattern formation. Our major finding, that POMC and PC2 are expressed in feather follicles, contrasts with the emphasis of previous studies of within-feather patterning (Nadeau et al. 2008; Yoshihara et al. 2012; Lin et al. 2013). It is interesting that both ASIP and PC2 are expressed within feather follicles of quail embryos on the ventral surface. It is currently unknown what concentrations of ASIP and MSH are found in the ventral feather follicles of quail embryos. The affinity of MC1R for ACTH/MSH is lower in The mechanisms underlying convergent evolution in the plumage patterns of birds 65 birds than humans and mice indicating that a significantly higher concentration of ligands would be required for MC1R activation (Ling et al. 2004). We showed that activation of MC1R may be required for melanogenesis on the ventral and dorsal surface, but it is unknown what mechanism may inhibit MC1R to create the species typical dorsal stripes and within-feather pigmentation. There are several possible candidate genes that downregulate MC1R activity. Alternatively, perhaps the β-defensin ligand may down-regulate cAMP in the dorsal plumage between stripe and within-feather patterns in quail. β-defensin is a member of a family of secreted peptides that are structurally similar to agouti. In vertebrates, β- defensin is highly polymorphic in sequence and copy number and has the potential for extensive cross-talk with the melanocortin system but its function in avian plumage coloration has not been demonstrated (Candille et al. 2007). Several other candidate genes may inhibit MC1R to create quail dorsal stripes and within-feather patterning. In mice Corin appears to suppress the ASIP pathway and modulates temporal banding patterns. The Corin gene encodes a transmembrane protease that acts in the dermal papilla in mice (Enshell-Seijffers, Lindon, and Morgan 2007), but does not appear to have an effect in domestic and wild cats (Kaelin et al. 2012). Alternatively, attractin, a single-transmembrane-domain protein, interacts with melanocortin receptors to regulate energy metabolism and pigmentation (Gunn et al. 1999; Gunn and Barsh 2000). In mice, a lack of attractin produces little or no yellow pigmentation, and acts downstream of ASIP but upstream of MC1R, and prevents follicular melanocytes from responding to ASIP. Similarly, mice that lack intracellular ubiquitin ligase mahogunin also lack a pigmentary response to ASIP (He et al. 2003). Thus there are many different loci that could differentially inhibit the MC1R pathway to produce dorsal between stripe and within- feather patterning. It is thought that plumage patterns may be a result of reaction-diffusion dynamics (Prum and Williamson 2002) but given that reaction-diffusion relies on an activator that syntheses its own inhibitor, it is unclear how the molecular mechanisms of pigmentation might relate to this model. Our data are consistent with previous reports that, similar to mice, ASIP is required to produce a pale-bellied phenotype in quail whereas, we found little evidence for a pigmentation function of AGRP. In The mechanisms underlying convergent evolution in the plumage patterns of birds 66 contrast with previous research on avian pigmentation, we showed that POMC and PC2 are likely to play a greater role in pigment patterning than hitherto realised. The mechanisms underlying convergent evolution in the plumage patterns of birds 67 The mechanisms underlying convergent evolution in the plumage patterns of birds 68 Chapter 3: Evolutionary pathways to convergence in plumage pattern phenotypes The mechanisms underlying convergent evolution in the plumage patterns of birds 69 The mechanisms underlying convergent evolution in the plumage patterns of birds 70 Abstract Avian plumage is ideal for investigating phenotypic convergence because of repeated evolution of the same within-feather patterns. In birds, there are three major types of regular patterns within feathers: scales, bars and spots. Existing models of within-feather pattern development, suggest that scales have the simplest developmental mechanism, bars require more stringent regulation than scales, and spots have the strictest developmental parameters. We hypothesized that increasing stringency in the mechanism of pattern formation predicts the evolutionary trajectory of patterns, and hence scales evolve first, followed by bars and finally spots. Here, using Bayesian phylogenetic modeling we reconstructed pattern evolution in the most spectacularly patterned avian clades – aquatic waterfowl (Anseriformes) and terrestrial gamebirds (Galliformes). Independent analyses of seven feather patches reveal that spots evolve after bars and scales. However, bars evolve more frequently from an absence of patterns than scales, contradicting our predictions. Analyses of larger body regions support a dorsoventral axis of modularity. Over the whole body, many constraints are conserved, e.g. spots are derived. Overall there was remarkable similarity in the evolutionary trajectories of plumage pattern evolution in Galliformes and Anseriformes, suggesting that developmental constraint is similar in these two orders, despite large ecological differences. The mechanisms underlying convergent evolution in the plumage patterns of birds 71 The mechanisms underlying convergent evolution in the plumage patterns of birds 72 Introduction Comparative studies are a powerful tool for understanding the underlying processes behind similarity in animal forms and have revealed that the mechanisms underlying convergent evolution are diverse as well as surprising (Parra-Olea and Wake 2001; Prud'homme et al. 2006; Arendt and Reznick 2008; Fritz et al. 2014). Phenotypic convergence may arise from similar selective regimes, but may also be at least partly explained by developmental constraint. This latter view is gaining momentum with proponents advocating an equivalent importance to natural selection (Maynard Smith et al. 1985; Gould 2002). Under developmental constraint, some phenotypes may be developmentally more readily accessible than others, thereby biasing evolution to follow particular pathways. The spectacular plumage phenotypes of birds have been subject to considerable attention due to their diversity, functional significance and ease of study (Price and Pavelka 1996; Kimball and Ligon 1999; Chuong et al. 2000; Omland and Lanyon 2000; Badyaev et al. 2001; Price and Bontrager 2001; Harris, Fallon, and Prum 2002; Prum and Williamson 2002; Widelitz et al. 2003; Chang et al. 2004; Jiang et al. 2004; Harris et al. 2005; Badyaev and Landeen 2007; Pointer and Mundy 2008; Riegner 2008; Kimball, Mary, and Braun 2011). Plumage coloration can vary over the body and is coordinated into species-typical phenotypes by individual patches containing up to thousands of feathers (Prum and Dyck 2003). For example, the tail of the peacock (Pavo cristatus) has coloured spots whereas the wings have bars. Patches of feathers that covary in coloration generally correspond to feather tracts (pterylae), which vary little in anatomical position between birds, and are indicative of local shared developmental programs or modules (Lucas and Stettenheim 1972; Wagner, Pavlicev, and Cheverud 2007). Plumage patches may be comprised of feathers of uniform coloration or feathers that are patterned. For example, the distinct patches of coloration on the crown and rump in the genus Phylloscopus are made up of uniform pigmentation, with little variation within individual feathers (Price and Pavelka 1996). In contrast, the patch of spotted plumage on the flanks of the male Zebra finch (Taeniopygia guttata) is due to regular within-feather patterning, which is composed of spatially variant pigmentation within the vane of each feather (Prum and Williamson 2002; Kenward The mechanisms underlying convergent evolution in the plumage patterns of birds 73 et al. 2004; Riegner 2008; Gluckman and Cardoso 2009; Gluckman and Cardoso 2010). While plumage coloration has been extensively studied, the evolution of within-feather patterning has received less attention despite abundant interspecific variation (Fig. 3.1) (Riegner 2008; Gluckman and Cardoso 2010). Fig. 3.1. The most frequent regular plumage patterns found in birds and the hypothesis of developmental constraint in plumage pattern evolution on the basis of increasing complexity. a) Scales - king eider (Somateria spectabilis), b) Bars - snow partridge (Lerwa lerwa), c) Spots - great argus (Argusianus argus). If there is developmental constraint in plumage pattern evolution on the basis of increasing stringency, then perhaps scales evolve from the ancestral state of uniform coloration, followed by bars, and finally spots. Images were taken at the University Museum of Zoology, the University of Cambridge, by T-L. Gluckman and are copyright of the University Museum of Zoology. Within-feather patterning can be split into two types based on the distribution of pigmentation: irregular pigmentation (mottled plumage), where the vane is heterogeneously pigmented, and regular patterns, which are comprised of the same recurring motif (Gluckman and Cardoso 2010; Gluckman 2014). Regular within- feather patterns have largely converged on the same strikingly simple set of repeating geometric patterns in birds: scales - where the feather border is regularly pigmented with a different shade of melanin (Fig. 3.1a); bars - alternating bars of lighter and darker pigmentation perpendicular to the feather’s axis (Fig. 3.1b); and spots - one or more regular spots of pigmentation within feathers (Fig. 3.1c) (Fig. 1 in The mechanisms underlying convergent evolution in the plumage patterns of birds 74 Prum and Williamson 2002). There are a few other regular motifs in birds, including chevrons and stripes, that are relatively rare, and two types of patterning that apparently occur in single species: checkered patterning in the common loon (Gavia immer), and triangles in the breast of the wood duck (Aix sponsa). This demonstrates that although other types of patterns are possible plumage patterns repeatedly converge on the same three motifs. Bird plumage is predominantly pigmented with melanins and carotenoids (Hill and McGraw 2006b). Melanins are of key importance for feather patterning since they are the only pigment that can be differentially deposited in precise spatio- temporal sequence during feather growth to create within-feather patterns (Prum and Williamson 2002; Hill and McGraw 2006b). In contrast, carotenoid-based coloration is typically confined to feather tips and so only contributes to uniform patches of coloration over the body. In a comparative survey encompassing 90% of avian species worldwide there were no observed cases of plumage patterns where melanin appeared to be absent, and additional types of coloration in patterning (e.g. psittacofulvins and carotenoids) were rare (T-L. Gluckman unpublished data). Studies of convergent evolution in plumage phenotypes have largely focused on patches of uniform coloration, in particular carotenoid coloration, with fewer studies of melanin (Christidis, Schodde, and Baverstock 1988; Hackett and Rosenberg 1990; Price and Pavelka 1996; Kusmierski et al. 1997; Omland and Lanyon 2000; Odeen and Bjorklund 2003; Hofmann, Cronin, and Omland 2006; Andersson, Prager, and Johansson 2007; Bleiweiss 2007; Jones and Kennedy 2008; Cardoso and Mota 2010; Prager and Andersson 2010; Friedman, Kiere, and Omland 2011). The proposed mechanisms underlying convergence in uniform coloration have focused on ecological explanations (Crochet, Bonhomme, and Lebreton 2000; Dumbacher and Fleischer 2001; Bleiweiss 2007; Weibel and Moore 2007), although several studies failed to find an ecological correlate (Omland 1997; Omland and Lanyon 2000; Omland et al. 2006). Additional suggested mechanisms of convergence are sexual selection (Omland and Lanyon 2000; Prager and Andersson 2010), and developmental constraint (Price and Pavelka 1996; Omland and Lanyon 2000; Majerus and Mundy 2003; West-Eberhard 2003; Prager and Andersson 2010). The proposed mechanisms shaping the evolution of uniform plumage The mechanisms underlying convergent evolution in the plumage patterns of birds 75 coloration may also influence the evolution of regular within-feather patterning. In particular, ecological factors may select for convergence in phenotypes, for example to provide crypsis to evade detection by predators (Shine and Madsen 1994; Lindell and Forsman 1996; Marshall 2000; Stevens and Merilaita 2009b). Ecological selection for plumage patterns is likely to be dependent on habitat and how the patterns are perceived against their background (Bradbury and Vehrencamp 1998; Endler 1998). Therefore, selection for camouflage in avian phenotypes is likely to differ between species due to variation in habitat. However, different types of patterns are found in sympatric species of birds that occupy the same habitat demonstrating that ecological selection may not necessarily result in convergent evolution, e.g. the barred warbler (Sylvia nisoria) and the red-backed shrike (Lanius collurio) (Polak 2012). In addition, the same patterns are frequently found in avian orders that live in different habitats (Gluckman and Mundy 2013). For example, waterfowl, which are aquatic, frequently have barred plumage, but so do many non-aquatic birds, such as raptors. In several species of birds, plumage patterns have a social function, such as the spots of the male peacock (Pavo cristatus) and the barred plumage of the red- legged partridge (Alectoris rufa) (Petrie, Halliday, and Sanders 1991; Bortolotti et al. 2006; Muck and Goymann 2011; Pérez-Rodríguez, Jovani, and Mougeot 2013). More broadly, barred patterns appear to have generally evolved for camouflage, but have also evolved under sexual selection on the ventral surface of males (a more likely location for signalling to conspecifics) and may function as a social signal in many species of birds (Gluckman and Cardoso 2010). As a communication signal, the same type of plumage patterns may converge on the same motif due to sensory exploitation or cognitive receiver biases. For example, perhaps barred patterns have frequently evolved under sexual selection because they are conspicuous to conspecifics (Gluckman and Cardoso 2009; Gluckman and Cardoso 2010). Additionally, barred plumage may have evolved as a specialised adaptation to facilitate cuckoo-hawk mimicry in Old World cuckoos (Gluckman and Mundy 2013). However, it is thought that signal evolution is likely to be habitat and context dependent, which can vary considerably between species (Bradbury and Vehrencamp 1998; Endler 1998). Developmental constraint in uniform coloration has been studied in the genus The mechanisms underlying convergent evolution in the plumage patterns of birds 76 Phylloscopus by employing a Turing reaction-diffusion model (Price and Pavelka 1996). Turing reaction-diffusion systems form morphological patterns from the combined action of two molecules (morphogens), an activator and an inhibitor. Chemical gradients of the morphogens induce spatially explicit patterns that are controlled by four key parameters per morphogen: the rates of production, decay and diffusion, as well as the overall strength of the interaction (Turing 1952). Price and Pavelka (1996) showed that the evolution of patches of white plumage (unmelanized feathers) could be attributed to increasing and decreasing rates of morphogen production. For mammalian coloration, the application of Turing models could explain why mammals that have a spotted tail with a striped body are not found (Murray 1981), thus implying a form of developmental constraint, best considered as relative constraint (Price and Pavelka 1996). In a landmark study, all regular within-feather patterns were successfully simulated with a reaction-diffusion based model (Prum and Williamson 2002). By modeling differential pigment uptake by keratinoctyes during feather development, Prum and Williamson found that regular plumage patterns could be produced by manipulating spatial and temporal periodicity. According to this model, the production of scales has a low rate of morphogen decay and is governed by spatial periodicity of melanin uptake. The production of bars requires a higher rate of morphogen decay resulting in temporal periodicity of melanin uptake. Notably, the formation of spots is distinct as it is comprised of simultaneous spatial and temporal differentiation and has the narrowest range of parameters (Fig. 6 in Prum and Williamson 2002) From Prum and Williamson’s model (2002), we hypothesised that the mechanism of within-feather pattern formation may bias the production of pattern variation during evolution in a stepwise order from decreasing to increasing stringency, thereby acting as a constraint. Our interpretation of this model of within- feather pattern formation is that scales have the least stringent conditions, bars have more stringent parameters than scales, and spots have the narrowest range of parameters. Therefore, our hypothesis of within-feather pattern evolution is that scales evolved first, followed by bars whereas spots are the most derived pattern (Fig. 3.1). From what is currently understood of plumage development, covariation of The mechanisms underlying convergent evolution in the plumage patterns of birds 77 feather pigmentation within patches is indicative of local shared developmental mechanisms and hence patches are a logical focus to study evolutionary pathways (Fig. 3.2) (Lucas and Stettenheim 1972; Prum and Dyck 2003; Wagner, Pavlicev, and Cheverud 2007). In contrast, it is currently unknown whether the development of pigmentation within one patch of plumage influences the evolution of pigmentation in another patch of plumage, although it seems likely that this may be the case. For example, the pink-eared duck (Malacorhyncus membranaceus), has barred plumage on the flanks and the breast. Perhaps barred plumage evolved on the breast first, and the mechanism of pattern formation was subsequently recruited by the flanks, or vice versa. An interesting issue is whether such recruitment across patches may be more favoured within developmental compartments, or modules, which may in principle occur along the dorsoventral axis or anteroposterior axis. At the level of the whole body, evolution of a novel plumage pattern may occur within the same patch or involve recruitment from other patches, or other modules, and might therefore appear relatively unconstrained. Analysis of the whole body is confounded by the co- occurrence of multiple different pattern types, e.g. species in both Anseriformes (e.g. Hottentot teal, Anas hottentoti) and Galliformes (e.g. Elliot’s pheasant, Syrmaticus ellioti) have separate patches with all four pattern phenotypes considered here: absence, scales, bars and spots. Scoring of patterns over the whole body thus necessitates prioritizing particular patterns over others. These considerations lead to the development of a hierarchical approach in which we first consider evolution within patches, then evolution within regions, and finally evolution over the whole body. Evolutionary pathways within patches provide basic evidence for mechanisms occurring within a developmental unit. Directionality would demonstrate that some transitions are preferred over others and allow a direct test of our hypotheses. Similarity among patch models and particular regional models would indicate modularity in pattern development. Similarity at the level of the whole body would suggest an absence of other mechanisms at this level. The mechanisms underlying convergent evolution in the plumage patterns of birds 78 Fig. 3.2. The plumage patches sampled in this study and combinations of patches analysed to investigate regional modularity. (1) The seven individual patches. Field guide terminology, and the corresponding feather tracts (pterylae; Lucas and Stettenheim 1972) are as follows: a) Nape: interscapular tract, b) Wing (scapular, wing coverts, tertials, primaries and secondaries): humeral tract, upper marginal coverts of prepatagium and upper wing covert tract, c) Rump and uppertail coverts: dorsopelvic tract and dorsal caudal tract, d) Tail: upper major tail covert, upper median tail covert and rectrices tract, e) Breast: ventral cervical tract, f) Flanks or side: pecterosternal tract in Anseriformes or pectoral tract in Galliformes, g) Vent and undertail coverts: abdominal tract in Anseriformes or lateral and medial abdominal tracts in Galliformes. (2) In regional analyses, patches are grouped into regions that may represent developmental modules (similarly coloured patches form part of the same region), with either three regions arranged on an anteroposterior axis (2a) or two regions arranged on a dorsoventral axis (2a). (3) In the whole body analysis, all patches are analysed together. The species illustrated is the Natal francolin (Pternistes francolinus). The mechanisms underlying convergent evolution in the plumage patterns of birds 79 The avian orders with the most spectacular plumage patterns are the waterfowl (Anseriformes) and gamebirds (Galliformes), which together form a monophyletic group (Galloanserae) (Mindell et al. 1999; van Tuinen, Sibley, and Hedges 2000; Livezey and Zusi 2007; Morgan-Richards et al. 2008). Each order includes iconic examples of patterns such as the spotted plumage of the great argus (Argusianus argus; Galliformes, Fig. 3.1C). Anseriformes and Galliformes have dramatically different lifestyles, comprising waterbirds and landbirds, respectively (del Hoyo, Elliott, and Sargatal 1992), and are thus likely to be subject to a host of different selection pressures. Variation in selection pressure is ideal for testing our hypothesis of developmental constraint in these two orders because a similarity in developmental constraint should lead to similar evolutionary pathways. Here, we examine directionality in within-feather pattern evolution, using Bayesian phylogenetic modeling in Anseriformes and Galliformes separately, with patterning identified from museum skins. We traced pattern evolution in a hierarchical order to assess whether there may be generalities in these ecologically diverse orders and examine whether there is a) directionality in pattern evolution, b) whether the direction of evolution provides support for increasing complexity in within-feather patterning developmental mechanisms, c) whether convergence follows similar pathways in both orders, d) whether there is evidence for regional modularity, and e) whether global models of plumage pattern evolution differ from the developmental models of within patch and regional modularity. Materials and Methods Phylogenies We searched the literature for published species level relationships. The best available phylogenies on the basis of species coverage, inclusion of both mtDNA and nuclear DNA, and inclusion of branch length information, are as follows: Anseriformes, 188 spp. (73%) (Gonzalez, Düttmann, and Wink 2009); Galliformes, 170 spp. (59%) (Kimball, Mary, and Braun 2011). Together these phylogenies cover The mechanisms underlying convergent evolution in the plumage patterns of birds 80 all families and 63% of extant species across the two orders (Howard, Dickinson, and Moore 2003). We collected plumage pattern information from each species (nominate subspecies where applicable) represented in these phylogenies from museum skins at the Natural History Museum at Tring and the University Museum of Zoology, Cambridge. Data collection and coding Current developmental evidence suggests that the default plumage phenotype in males and females in Anseriformes and Galliformes is the male plumage (Owens and Short 1995). Therefore, we collected plumage pattern information for the seven patches of plumage over the body for the males of each sample species (Fig. 3.2). We assigned the character state of each of the seven feather patches as scales, bars, spots, or an absence of patterns, following the description by Prum and Williamson (2002). Some species exhibit what appear to be longitudinal stripes along feathers, but on closer inspection are an angular version of scales with a central pigment patch, and were scored as scales, e.g. the breast and nape plumage of the vulturine guineafowl (Acryllium vulturinum) (Fi.g 1a in Prum and Williamson 2002). A small number of species in this study have chevron patterns – Anseriformes – 2 spp., Galliformes – 5 spp.. Given that chevrons are rare in these orders and that they are similar to patterns made of bars, in that the borders do not meet to create a central pigment patch, we scored chevrons as bars (Fig. 1e and Fig. 6e in Prum and Williamson 2002). For most species sampled, the type of within-feather patterning across the vane of each feather, as well as between individuals of the same species, was the same for each patch considered. However, in a rare number of cases there was variation between individuals. To focus on the most developmentally relevant patterns in these rare cases we recorded the pattern that covered the majority of the feathers, in the patch under consideration, and where relevant, the predominant pattern in the majority of individuals sampled. For example, in the Natal francolin (Pternistes francolinus), the feathers in the flanks can have both bars and scales (Fig. 3.2). In the example depicted, bars cover the majority of the feathers in the flanks, and this individual would have been assigned as having bars. However, in The mechanisms underlying convergent evolution in the plumage patterns of birds 81 most individuals of the sample population of the Natal francolin, scales predominantly covered most of the feathers in the flanks, and were considered representative of this species. An additional type of pattern, mottled plumage, is present in many birds. It is currently unknown whether all mottled patterns can be considered homologous, or whether they may be classified into discrete types based on the size, shape and distribution of pigmentation across the vane of the feather. Therefore, mottled plumage was scored as unknown. To summarize our findings of local pattern evolution within all seven patches of plumage, we report on the frequency with which each transition between patterns occurs out of the total number of patches. The most probable model of evolution could then be compared with our hypothesis as well as the full (null) model of evolution. To investigate whether within-feather pattern evolution in one patch of plumage may precede and/or promote evolution of patterning in other patches, we conducted our analyses by combining patterns into body regions in three ways (Fig. 3.2). Each body region (e.g. ventral surface) was coded as the most derived pattern for each patch it contained as indicated by the summary model of local evolution. For example, if a body region consisted of patches containing both bars and spots, it was coded as having spots. In Anseriformes, in the summary model of local evolution within patches there is no conflict in the order of transitions and the most derived pattern is clear. However, in Galliformes, bars evolve from an absence of patterns and the next pattern to evolve from bars could be either scales or spots (see Results). We took this uncertainty into account by examining each possible trajectory for each body region separately for comparison in Galliformes. For example, males of the satyr tragopan (Tragopan satyra) have scales on the flanks and vent, and spots on the breast. In the analysis of the ventral region of Galliformes, where the flanks, breast and vent are collapsed into one character, we compared whether assigning either scales or spots as the most derived character created conflict in the analysis. The mechanisms underlying convergent evolution in the plumage patterns of birds 82 Similar to the summary model of local evolution within-patches, we derived models of regional modularity by reporting on the frequency of each transition between plumage patterns out of the total number of regions. The mode of regional modularity that is most plausible should contain the same transitions as the summary model of evolution within-patches of plumage. Finally, to derive a global model of plumage pattern evolution over the whole body, we used the same approach for scoring each species as used in the models of regional modularity, but incorporating information from all patches. Modelling of plumage pattern evolution We modeled plumage pattern evolution over the phylogeny to estimate which patterns evolve into one another, allowing us to derive a model of the probable evolutionary pathways to current plumage pattern phenotypes. Anseriformes and Galliformes have a wide distribution and live in different habitats, which may alter the evolutionary trajectory of each order (Gluckman 2014). Therefore, we examined each order separately to assess for similarity and differences in their evolutionary history. To estimate plumage pattern evolution in each order, we used the Reversible Jump Markov Chain Monte Carlo Multistate option in BayesTraits v.2 (Pagel, Meade, and Barker 2004; Pagel and Meade 2006). Markov Chain Monte Carlo (MCMC) is based on the proposition that traits can repeatedly evolve between any possible state on any branch of the tree. To estimate the rate of change between states, the Markov chain samples the plumage patterns at the internal nodes of the tree, in proportion to their probability, which is conditioned on the values at the tips. New rate parameter values are proposed in successive steps in the Markov chain resulting in a posterior sample distribution of rate coefficients and ancestral states. Each model of pattern evolution rate coefficients is visited in direct proportion to its posterior probability in the sample distribution (Pagel, Meade, and Barker 2004). Given that there are four pattern states, which in turn offer many parameters that describe evolution between plumage patterns, we used Reversible Jump MCMC (RJMCMC). The mechanisms underlying convergent evolution in the plumage patterns of birds 83 RJMCMC integrates rate restrictions by searching the posterior distributions of model parameters to avoid over parameterization. As such, we allowed BayesTraits to propose transition rates of plumage pattern evolution without restriction (e.g. we did not constrain any rate parameters to equal 0 based on a priori predictions) thereby making the analysis conditional on the data rather than our hypothesis (Gluckman 2014). For example, we hypothesized that spots are the most derived pattern as a consequence of having the strictest developmental parameters and therefore do not evolve directly from an absence of patterns (Fig. 3.1). In transition rate models that support this hypothesis, a rate parameter between an absence of patterns and spots equals 0, and therefore does not occur. This allows both incremental and non-sequential changes to occur in any direction and avoids imposing potentially false hypothesis based predictions. By making the analysis conditional on the data, this approach uses the best data currently available to assess the most probable models of plumage pattern evolution. Potential models of plumage pattern evolution visited by the Markov chain are distinct from the most probable model of plumage pattern evolution. The former describes the proposed models of plumage pattern evolution that make up the posterior sample distribution, whereas the latter is derived from statistically evaluating the posterior sample distribution. Each model of plumage pattern evolution is composed of a unique combination of transition rate parameters with values fixed to zero or are sampled as free parameters with positive values. Rate parameters fixed to zero were interpreted as an evolutionary transition that does not occur, and free rate parameters with a positive value were interpreted as evidence for an evolutionary transition that does occur. Therefore, qualitatively, each unique model of plumage pattern evolution is composed of transitions that do not occur, and transitions that do occur. Null model testing and model comparisons were conducted by assessing the posterior distribution of unconstrained models. If there were no developmental constraint, such as where natural selection drives plumage patterns to evolve in any direction, forward and backward evolutionary transitions between all pattern states would occur – the full (null) model. Therefore, if there were no directionality in plumage pattern evolution, the full model would be visited more frequently than expected by chance. Conversely, if sequential or non-sequential evolutionary The mechanisms underlying convergent evolution in the plumage patterns of birds 84 transitions were more probable, then models with these transitions would be most probable. In assessing the models of evolution without constraining any transitions, each unique model of pattern evolution is compared with every other possible model of pattern evolution (statistical methods are described in the next section). In BayesTraits we modeled the rates of plumage pattern evolution using a hyperprior with a gamma distribution defined by an empirical Bayes estimator (Pagel, Meade, and Barker 2004). For each analysis, we discarded the burn-in and the Markov chain was run until convergence across four independent runs (<1 lnHM). After convergence was reached, we thinned the sample distribution to remove autocorrelation. In analyses of within patch pattern evolution, we sampled 10,000 generations, per patch, per phylogeny. For the analyses of regional modularity, we sampled 2,000 generations, per region, per phylogeny. The average rate for each transition could not be averaged across feather tracts or body regions, as it is statistically incorrect. Instead, the average transition rates for each patch of plumage and body region are presented in the supplementary material. Model priors and modelling parameters The prior density on the free transition rate parameters were estimated using an empirical Bayes estimator (where the interval of the hyperprior is defined by the average and standard deviation of the maximum likelihood of all rate parameters) to reduce bias and uncertainty in choice of priors (Pagel, Meade, and Barker 2004). We used a hyperprior approach with a gamma distribution as our empirical Bayes estimator values had an intermediate range. The intervals were estimated for each analysis, for each patch of plumage, and for each region, in each group separately. For the analysis of independent evolution within patches of plumage, in each phylogeny, the Markov chain was run for 250 million generations sampling every 10,000th generation (7 x 2 = 14 individual analyses). The first 120,000 generations of RJMCMC (burn-in) were discarded to ensure parameter space was sufficiently explored. For the analysis of regional modularity the Markov chains were less stable and were run for 500 million generations sampling every 50,000th generation, discarding the first 150,000 generations as burnin (6 x 2 = 12 individual analyses). Each analysis of within-patches patches and modular feather regions, per order, was The mechanisms underlying convergent evolution in the plumage patterns of birds 85 run four times to ensure convergence had been reached within analyses as indicated by a stable harmonic mean of the log-likelihood that varied by <1 lnHM across all four runs. We checked for autocorrelation using the Ljung-box test statistic in SPSS v19.0 at lag 1 (IBM Corp.). A Ljung-box P > 0.05 was interpreted as indicating no autocorrelation. There was autocorrelation and we thinned the posterior sample distribution of models of plumage pattern evolution in each analysis, preserving the order of the models in which they were visited. For the analyses of both within patches and modular feather regions we thinned the sample distribution to every 100,000th model resulting in a posterior sample distribution of 10,000 models (2,500 models per independent run) for the analyses of within patches, and 2,000 models (500 models per independent run) for the analyses of modular feather regions. Statistical analysis The most probable models of plumage pattern evolution, each with their own most probable ancestral state of patterning, are visited in proportion to their Bayesian posterior probability. To qualitatively summarize whether each transition probably does not occur, or occurs, and account for model variation and uncertainty in the posterior sample distribution of models proposed, we employed multimodal inference (Burnham and Anderson 2002; Gluckman 2014). To assess which models of plumage pattern evolution are visited more frequently than expected by chance, we compared the prior odds of seeing each model of plumage pattern evolution, with the posterior odds derived from the posterior sample distribution. Final comparisons were made using Bayes Factors. Given that there are four states of patterning in this analysis, twelve possible evolutionary transitions can occur, but each can also have a transition rate of zero. To account for the varying numbers of zero and non-zero transitions, we calculated the prior odds of encountering each unique model of plumage pattern evolution in the sample distribution. Prior odds were calculated using binomial numbers for the number of transitions that do not occur, as well as bell numbers for transitions that occur, combined (supplementary Table S3.1, S3.2; see Currie et al. 2010 for a detailed explanation of calculations used). To assess the most probable model of plumage pattern evolution, we derived a top model set using a threshold approach of a BayesFactor of >=2, which is considered positive evidence (Kass and Raftery 1995; Burnham and Anderson 2002). The mechanisms underlying convergent evolution in the plumage patterns of birds 86 Due to variation in the total collection of models sampled in the posterior distribution, some ancestral states as well as rate parameters vary widely in whether they are fixed to zero, or sampled as free parameters with positive values. Therefore, we investigated what ancestral states and free parameters were favored in the top model subset of the posterior distribution, accounting for sampling variation. The marginal probability (MP) per ancestral pattern, as well as the unique model of pattern evolution, was calculated from the entire posterior sample distribution. For example, the ancestral for each type of pattern MP = (models in which this pattern is ancestral/10,000), and for each unique model of plumage pattern evolution MP = (Unique model/10,000). Fig. 3.3. The marginal probability of evolutionary transitions between plumage patterns in the feather tract of the breast in Anseriformes and Galliformes. Next to each transition is the marginal probability of a transition not occurring, followed by the marginal probability of it occurring. The width of the transition line is proportional to the rate of transition. Where the probability of a transition occurring is less than the probability of the transition not occurring, the transition line is grey indicating that it most probably does not occur. Conversely, where the marginal probability of it occurring is higher than not occurring, the transition line is black indicating that the transition probably occurs. The marginal probability of occurring and not occurring does not equal due to variation in the transitions represented in the top model set. The final marginal probability was calculated by cumulatively adding the MP of models in the top model set for each ancestral state of patterning and for each The mechanisms underlying convergent evolution in the plumage patterns of birds 87 evolutionary transition where it does not occur, as well as where it occurs, for comparison (Burnham and Anderson 2002). For example, in the breast of the galliform birds, the marginal probability (MP) of an absence of patterning not being the most probable ancestral state is 0.00 in the top model set whereas the MP of an absence of patterning being the most probable ancestral state is 0.89 (Table 3.1). In addition, the MP of scales, bars and spots not being the most probable ancestral state is 0.89 versus 0.00 of being the ancestral state. Assessing a transition from an absence of patterns to spots, the MP of the transition rate parameter describing it as not occurring is 0.87 and its MP of being non-zero is 0.01 (Fig. 3.3). Together this shows that an absence of patterns is most probably the ancestral state in the breast of galliform birds, and a transition from an absence of patterns to spots most probably does not occur. The marginal probability in the top model set accounts for variation in the entire posterior sample distribution, therefore the sum of the marginal probability of a transition not occurring and occurring rarely equals 1 as this requires every model in the posterior sample distribution to have the same result for that transition. Results Taxonomic distribution of patterns All of the different types of regular plumage patterns were represented in the seven plumage patches, with the exception of spots on the rump, nape and tail, as well as bars on the tail in Anseriformes, and scales on the tail in Galliformes (Table 3.1; see supplementary Fig. S3.1 for a taxonomic distribution of plumage patterns). In all body regions (dorsoventral and anteroposterior), and over the whole body, all plumage patterns were represented in both orders. We first present analyses of plumage pattern evolution within individual patches, followed by regions and finally the whole body. The mechanisms underlying convergent evolution in the plumage patterns of birds 88 Evolution within patches of plumage For individual patches of plumage that have more than two pattern states, there was variation in the number of unique models supported in the top model set (Table 3.1; supplementary Table S3.3). Plumage patches with only three pattern states (Anseriformes: nape, rump; Galliformes: tail) had less variation in the number of unique models in the top model set, than plumage patches with four pattern states. In the plumage patch with just two pattern states (Anseriformes: tail), only the full model was present in the entire posterior sample distribution. In contrast, in all models of patches of plumage with more than two pattern states, the full model was not visited more than expected by prior odds (Table 3.1; supplementary Table S3.3). In six out of seven patches of plumage in Anseriformes, and in all patches in Galliformes, the ancestral plumage was an absence of patterns (Table 3.1; supplementary Table S3.3). Both the average probability and the MP supported an absence of patterns in three patches in Anseriformes, and across all patches in Galliformes. Four patches in Anseriformes have equivocal support - rump, breast, vent and tail. However, the MP, which integrates model support, unlike the average probability, indicated that pattern absence is the most probable ancestral plumage in the rump, breast and vent. In only the tail of Anseriformes was there an equivocal ancestral state, which is probably due to having only two pattern states – barring and an absence of patterns (Table 3.1; supplementary Table S3.3). In all models of plumage evolution, except for the tail of Anseriformes, there was evidence of directionality as some transitions probably occurred and others did not (Fig. 3.4, supplementary Fig. S3.2). Examining the order of pattern evolution within-patches, bars evolve more frequently from an absence of patterns (Anseriformes: 4/6; Galliformes: 5/7) than scales (Anseriformes and Galliformes: 2/6) (Fig. 3.4). Indeed, in all models of plumage pattern evolution, the average transition rate from an absence of patterns to scales is low (supplementary Fig. S3.2). In both orders there are strong bidirectional transitions between scales and bars, and spots evolved more frequently from scales (Anseriformes: 5/5; Galliformes: 6/6) than bars (Anseriformes: 3/5; Galliformes: 5.5/6). The latter transition from bars to spots in Anseriformes includes equivocal transitions in two patches (supplementary Fig. S3.2). Finally, transitions from an absence of patterning to spots were rare and had The mechanisms underlying convergent evolution in the plumage patterns of birds 89 the lowest rate of transition where they occur (Anseriformes: 1.5/5; Galliformes: 1/7; supplementary Fig. S3.2). Therefore, within-patches the predominant order of plumage pattern evolution is bars, followed by scales and finally spots. Fig. 3.4. The summary model of local evolution within-patches across all seven plumage patches in Anseriformes and Galliformes. Next to each transition is the number of plumage patches in which the transition occurs out of the total number of plumage patches. The total number of plumage patches can vary from the maximum (seven) because in some patches particular patterns do not occur e.g. in Anseriformes no species have evolved spots on the tail or the rump, so the total number of plumage patches in which it can evolve is five. Where transition lines have an intermediate value, e.g. 1.5/5 for a transition from absence to spots in Anseriformes, this indicates that the transition was equivocal in one of the models of pattern evolution within plumage patches. The weight of each transition probably occurring is represented on a scale of pale grey (occurs rarely) to black (occurs in every plumage patch possible). Regional modularity in plumage pattern evolution Pattern evolution across body regions showed variation in the number of unique models supported in the top model set (Table 3.1; supplementary Table S3.3). For all body regions in both orders, the full model was not visited more than expected by prior odds, and, in some cases, was not visited in the entire posterior sample distribution (Table 3.1; supplementary Table S3.3). In three out of six body regions in Anseriformes, and in all analyses in Galliformes, both the average The mechanisms underlying convergent evolution in the plumage patterns of birds 90 probability and the MP support an absence of patterning as the ancestral state of plumage (Table 3.1; supplementary Table S3.3). The source of uncertainty in the ancestral state of Anseriformes is in three body regions - middle, posterior and ventral - however, the MP for each supports pattern absence as the most probable ancestral plumage. Incorporating uncertainty in the order of pattern evolution in Galliformes resulted in conflict in five transitions spanning four out of five body regions (Fig. 3.5; supplementary Fig. S3.3). There were varying degrees of similarity between the summary model of local evolution within-patches and all three modes of regional modularity. However, the mode of regional modularity that most resembled the summary model of within patch pattern evolution in both orders was a dorsoventral axis (Fig. 3.4, 3.5). Fig. 3.5. Regional models of plumage pattern evolution in the proposed modes of modularity in Anseriformes and Galliformes. Next to each transition is the number of The mechanisms underlying convergent evolution in the plumage patterns of birds 91 body regions in which the transition occurs out of the total number of body regions for that mode of modularity. The weight of each transition probably occurring is represented on a scale of pale grey (does not occur) to black (occurs in every modular feather region for that analysis). A dashed grey line indicates an equivocal transition - where the marginal probability was equally in favor of not occurring and occurring (0.5). Across both orders, there was disagreement between the summary model of within-patch evolution and the anteroposterior mode of modularity. In the Anseriformes anteroposterior model, scales evolve from an absence of patterns more frequently than bars, opposite to the summary of local patch evolution. Furthermore, in Galliformes a transition from spots to an absence of patterns is equivocal whereas in the summary model of location evolution it is a strong transition. In the dorsoventral axis model, in Anseriformes all transitions are qualitatively the same as the summary model of pattern evolution within-patches, the exception being a transition from scales to bars only occurring on one surface. In Galliformes, the dorsoventral axis model of pattern evolution does not conflict with the summary model of within-patch evolution (Fig. 3.4, 3.5). Together these results demonstrate support, for a regional dorsoventral axis of modularity. Global model of plumage pattern evolution The global models of plumage pattern evolution showed some important similarities with within-patch models. First, in both orders, a direct transition from an absence of patterns to spots probably does not occur. Second, there are strong bidirectional transitions between bars and scales in both orders (Fig. 3.6, supplementary Fig. S3.4). Third, in Galliformes bars evolve first. However, there were some differences between the global model and the summary model of evolution within-patches: in the Anseriformes global model, scales evolve first and the model lacks bidirectional transitions between bars and spots, while in Galliformes the whole body model does not have a transition from spots to an absence of patterns (Fig. 3.4, 3.6). The mechanisms underlying convergent evolution in the plumage patterns of birds 92 Fig. 3.6. The most probable evolutionary transitions between plumage patterns over the whole body in Anseriformes and Galliformes (where spots are derived – see Methods). Next to each transition is the marginal probability of a transition not occurring, followed by the marginal probability of it occurring. The width of the transition line is proportional to the rate of transition. A grey transition line indicates transitions that probably do not occur and black transition lines indicate transitions that probably occur. The marginal probability of occurring and not occurring does not equal 1 due to variation in the top model set. The mechanisms underlying convergent evolution in the plumage patterns of birds 93 Table 3.1. The frequency of the different types of patterns in the seven plumage patches over the body, the number of unique models in the entire posterior sample distribution as well as the top model set, and the average probability and marginal probability of the ancestral state of patterns, in Anseriformes and Galliformes. Ancestral state Plumage pattern frequency within plumage patches Posterior sample distribution Average probability Marginal probability Absence of patterns Scales Bars Spots Mottled Unique models Top model set (BF>=2) Full (null) model Absence of patterns Scales Bars Spots Absence of patterns Scales Bars Spots Anseriformes (N = 118) Nape 78 16 19 3 2 464 257 118: BF = 0.08 0.31 0.25 0.20 0.24 0.02; 0.78 0.80; 0.01 0.79; 0.01 0.80; 0.00 Wing 105 5 5 1 2 698 454 57: BF = 0.04 0.27 0.26 0.23 0.23 0.11; 0.67 0.75; 0.03 0.70; 0.08 0.77; 0.01 Rump 100 12 4 - 2 15 4 1435: BF = 0.72 0.33 0.33 0.33 N/A 0.30; 0.38 0.38; 0.29 0.68; 0.00 N/A Tail 109 7 - - 2 1 1 All 0.5 0.5 N/A N/A 0.50; 0.50 0.50; 0.50 N/A N/A Breast 74 14 12 13 5 379 218 207: BF = 0.14 0.25 0.27 0.22 0.26 0.44; 0.36 0.61; 0.19 0.77; 0.03 0.65; 0.15 Flanks 65 11 34 5 3 339 164 68: BF = 0.05 0.26 0.25 0.24 0.25 0.12; 0.76 0.88; 0.77; 0.88; 0.00 The mechanisms underlying convergent evolution in the plumage patterns of birds 94 0.00 0.11 Vent 86 9 14 4 5 379 229 106: BF = 0.07 0.2 0.25 0.29 0.26 0.45; 0.34 0.72; 0.07 0.50; 0.29 0.70; 0.09 Galliformes (N = 170) Nape 75 18 29 10 38 229 121 11: BF = 0.01 0.72 0.13 0.11 0.05 0.01; 0.93 0.94; 0.00 0.94; 0.00 0.94; 0.00 Wing 81 14 13 9 53 475 292 40: BF = 0.03 0.66 0.14 0.13 0.07 0.00; 0.87 0.87; 0.00 0.87; 0.00 0.87; 0.00 Rump 83 11 25 8 43 305 178 12: BF = 0.01 0.8 0.09 0.08 0.03 0.00; 0.93 0.93; 0.00 0.93; 0.00 0.93; 0.00 Tail 93 - 24 8 45 36 7 454: BF = 0.20 0.78 N/A 0.09 0.13 0.08; 0.63 N/A 0.71; 0.00 0.63; 0.08 Breast 105 16 17 13 19 265 134 34: BF = 0.02 0.48 0.19 0.2 0.13 0.00; 0.89 0.89; 0.00 0.89; 0.00 0.89; 0.00 Flanks 79 21 29 14 27 399 235 101: BF = 0.07 0.69 0.16 0.09 0.06 0.05; 0.76 0.78; 0.03 0.80; 0.01 0.81; 0.00 Vent 139 6 11 4 10 312 191 10: BF = 0.01 0.48 0.2 0.23 0.09 0.00; 0.95 0.95; 0.00 0.95; 0.00 0.95; 0.00 The mechanisms underlying convergent evolution in the plumage patterns of birds 95 Discussion Studies of phenotypic convergence in bird plumage have mostly focused on coloration, although regular patterns within feathers are widespread across the class Aves. Our analyses suggest that the ancestral state of plumage is an absence of patterns, a consistent finding in Galliformes, but with some variability in Anseriformes (Table 3.1). Contrary to our predictions of pattern evolution, bars largely evolved first rather than scales. However, our analysis confirms that spots are derived from other pre-existing patterns (scales), and therefore, as hypothesized, reaction-diffusion based spatiotemporal differentiation may constrain spots to evolve from an absence of patterning by a minimum of two transitions (Fig. 3.1, 3.4). Our analyses also suggest that the potential for patches of plumage to develop patterning is not equally spread over the body. Instead there is regional modularity over the dorsoventral axis, but not the anteroposterior axis, such that evolutionary trajectories of patches within the same dorsoventral region are similar. This occurs in two avian orders that have very different lifestyles, illustrating the importance of development in evolution. Finally, the models over the whole body demonstrate that many mechanisms are conserved from the level of patches, including the highly derived nature of spots, but also highlight some interesting differences. There is consistent support for directionality in plumage evolution at all levels of analysis, including all patches of plumage with more than two pattern states, all body regions and the whole body. (Fig. 3.4, 3.5, supplementary Fig. S3.2, S3.3). From an absence of patterning, bars predominantly evolved first, followed by scales, and finally spots, in both Anseriformes, and Galliformes (Fig. 3.4-3.6). The evolutionary trajectory of within-feather patterns demonstrates that with increasing complexity in the mechanism of pattern formation, different types of patterns become developmentally more accessible. These results are congruent with developmental constraint in this system, but the main pathway, that bars evolve first, does not follow our predictions and may indicate that the developmental basis of scales is more complex than that of bars. Some support for a relative lack of mutational constraint on bar formation comes from genetic studies. Several independent mutations, both autosomal and Z- The mechanisms underlying convergent evolution in the plumage patterns of birds 96 linked, can lead to bars from an absence of patterning (Muscovy duck - Hollander 1968; chicken - Crawford 1990). In the best studied case, the sex-linked barred mutation in chickens, controlled by the CDKN2A/B locus, is associated with pale bands devoid of melanocytes (Hellström et al. 2010). Thus a different locus, ASIP, controls temporally-related patterning in mammalian hairs (Tamate and Takeuchi 1984; Bultman, Michaud, and Woychik 1992; Barsh 1996; Kaelin et al. 2012), and is a potential candidate for within-feather patterning. Mutations at this locus in quail affects bar width in individual feathers (N.I. Mundy and F. Minvielle unpublished data), and ASIP expression in developing chicken feathers is spatially variable (Oribe et al. 2012; Yoshihara et al. 2012). Thus the evolutionary origin of bars may be more straightforward than inferred from the reaction diffusion model. Currently, a large gap in our understanding is a plausible mechanism for how these loci are involved in a reaction-diffusion mechanism. A transition from an absence of patterning to spots occurred in the flanks in both orders but in none of the other six patches. It has been demonstrated that spots can have a social function (Petrie, Halliday, and Sanders 1991; Kose, Mänd, and Moller 1999; Roulin 1999a; Roulin, Riols, and Dijkstra 2001) and sexually selected traits evolve quickly potentially masking a signal of constraint (Pomiankowski and Iwasa 1998). It therefore seems likely that this rare transition is a result of strong selection pressure circumventing developmental constraint. Results from regional models showed that constraint was still present at this level of analysis, and that the summary model of patch evolution showed greater similarity to the dorsoventral axis than anteroposterior axis models, in both orders (Fig. 3.5). This provides evidence that the dorsoventral regions are behaving as hierarchical developmental units or modules, each composed of multiple patches evolving in a similar manner. The similarity was particularly strong for Galliformes, while in Anseriformes there was one major difference – the dorsoventral model lacks a transition from bars to scales in the ventral region (Fig. 3.3, 3.5). The striking overall similarities between the summary model of local pattern evolution within-patches and the dorsoventral axis model of patterning imply an additional layer of developmental constraint in plumage pattern evolution. We currently know little about the interaction between pteryla formation and modularity in The mechanisms underlying convergent evolution in the plumage patterns of birds 97 phenotypes. Numerous interactions between the epithelium and mesenchyme are known to signal pterylae formation (Chuong 1993; Prum 1999; Chuong et al. 2000; Dhouailly 2004; Lin et al. 2006). Epithelial-mesenchymal transplant experiments in quail-chicken chimaeras show that there are important dorsoventral differences in integument development: the dorsal trunk dermis originates from the dermomyotome of the somites whereas the ventral dermis originates from lateral plate mesoderm (Mauger et al. 1982; Fliniaux, Viallet, and Dhouailly 2004). In addition, the feather dermal progenitors of the dorsal and ventral regions are specified by different signals (Lin et al. 2006). Therefore, the differential dorsoventral origins of dermis cells may have lasting developmental and evolutionary effects on the phenotype of birds. As for patches and regions, the whole body models showed evidence for directionality in plumage pattern evolution. If recruitment of patterning mechanisms across local patches and regions were common, this would lead to more transitions occurring in the whole body model. However, evidence for this is limited. For Galliformes it is striking that the main features of the patch and dorsoventral models, including a stronger transition from an absence of patterning to bars than scales, occur in the whole body model. For Anseriformes, the picture is mixed – whereas a transition from absence of patterns to scales occurs in the whole body, the transition from absence to bars does not, even though this is strongly supported in the within- patch models. Most strikingly, in both orders, spots can only evolve from pre-existing patterns and not from an absence of patterns for the whole body. Given the evidence for sexual selection acting on spots, this reinforces the view that there is overall developmental constraint on spot development. We have carried out a broad level analysis of two large sister taxa that are ecologically distinct, and the decision to analyse them separately receives post hoc justification from the finding of differences among them. It is of course possible that different transition rates of plumage patterns occur among different clades within these taxa. This is an interesting issue for future investigation. Although we analysed most extant species within Anseriformes and Galliformes (63% combined), 37% of species were excluded because of a lack of robust phylogenetic information. The effects of this on our results are unknown, but we note that the absence of these taxa is not different in principle to the absence of an unknowable number of extinct taxa in the dataset. Hybridization is common in both orders: an estimated 41.6% species of The mechanisms underlying convergent evolution in the plumage patterns of birds 98 Anseriformes and 21.5% species of Galliformes hybridize (Grant and Grant 1992). Hybridization can lead to rapid shifts in phenotype that could cause uncertainty in estimating model transitions. The effect of categorizing mottled plumage as missing data is also unknown. Given that mottled patterns do not appear to have a regular motif, categorizing these patterns is plagued by uncertainties. Having many categories for different types of mottled patterns would likely obscure a signal of directionality in evolution, whereas using a single category for a pattern that exists in many states, might overly constrain the model. Therefore, using a category of “unknown” is representative of what is currently known about plumage pattern formation, and using robust Bayesian based analyses based on multi-model inference should largely control for uncertainty (Pagel, Meade, and Barker 2004). An issue for future consideration will be the potential influence of female patterning on the evolution of patterning in males. In both orders studied, there is sexual dimorphism in plumage patterns (Gluckman 2014), which is estrogen- dependent (Owens and Short 1995; Kimball and Ligon 1999). As a consequence, it was thought that elaborate coloration initially evolved in both sexes via genetic correlation (Lande 1980; Owens and Hartley 1998; Kimball and Ligon 1999; Kraaijeveld, Kraaijeveld-Smit, and Komdeur 2007). However, currently there is little evidence to suggest that there is genetic correlation in plumage pattern evolution between males and females in Anseriformes and Galliformes (Gluckman 2014). Hence the possibility that particular patterns evolve first in females and are later acquired by males remains, and will be considered in future studies. Similar plumage patterns have evolved in many distantly and closely related species of birds (Riegner 2008; Stern and Orgogozo 2009; Gluckman and Cardoso 2010). We demonstrated directionality in plumage pattern evolution that is congruent with developmental constraint. Overall there was remarkable similarity in the trajectories of pattern evolution in Galliformes and Anseriformes, suggesting that the constraint is similar in the two orders, despite large ecological differences. As suggested by Price and Pavelka (1996) the role of natural selection may be “fine- tuning the appearance of the pattern, fixing and maintaining pattern elements at a given level of expression, and modifying behavioral and other features to maximize the patterns’ utility” on the basis of the order that patterns evolve. We suggest that directionality in plumage pattern evolution may be caused by the underlying The mechanisms underlying convergent evolution in the plumage patterns of birds 99 dynamics of the developmental system of patterning, which may be of general significance to birds. The mechanisms underlying convergent evolution in the plumage patterns of birds 100 The mechanisms underlying convergent evolution in the plumage patterns of birds 101 The mechanisms underlying convergent evolution in the plumage patterns of birds 102 Chapter 4: Cuckoos in raptors’ clothing, barred plumage illuminates a fundamental principle of Batesian mimicry This chapter was published in the journal of Animal Behaviour, DOI: 10.1016/j.anbehav.2013.09.020 Banded bay cuckoo Japanese sparrowhawk (Cacomantis sonneratii) (Accipiter gularis) The mechanisms underlying convergent evolution in the plumage patterns of birds 103 The mechanisms underlying convergent evolution in the plumage patterns of birds 104 Abstract A fundamental principle of Batesian mimicry is that it pays to look like a local harmful species that is recognizable to other local species (receivers). Mimicking an allopatric species confers no benefit, as it is not recognizable to local species. It is thought that the common cuckoo, Cuculus canorus, is a Batesian mimic of the Eurasian sparrowhawk, Accipiter nisus, predominantly via its barred plumage, which facilitates access to host nests to deposit eggs. Barring is widespread in five genera of Old World cuckoos, unlike nonparasitic cuckoos, and evolved after the evolution of parasitism. Although barred plumage is predominant in parasitic cuckoos, it is unclear whether it may have a widespread function in cuckoo-hawk mimicry. If widespread, there should be a visual similarity between all five genera of Old World parasitic cuckoos and sympatric raptors. In addition, given that it pays to look like a local harmful species, sympatry should predict the degree of similarity. We compared barred plumage from all five genera of parasitic Old World cuckoos and up to eight sympatric raptors using digital image analysis. Cuckoos predominantly matched most raptors for at least one pattern attribute. In addition, three out of five cuckoos closely resembled a sympatric raptor for all barred pattern attributes examined, and potential model species were not confined to sparrowhawks. Habitat did not appear to influence plumage pattern similarity in most species studied. Finally, the barred plumage of sympatric species was more similar in appearance than those in allopatry. Together this demonstrates that cuckoos look like a local harmful species, which is congruous with Batesian mimicry. The mechanisms underlying convergent evolution in the plumage patterns of birds 105 The mechanisms underlying convergent evolution in the plumage patterns of birds 106 Introduction In Batesian mimicry, a harmless species mimics an unpalatable or a harmful one. For example, Dismorphia butterflies (mimics) vary their colour pattern according to the local species of toxic Neotropical Heliconius butterflies (models; Bates 1862). By looking like a familiar unpalatable model, the mimic avoids detection. Therefore, a fundamental principle of Batesian mimicry is that there is an advantage to looking like a local unpalatable or harmful species that is recognizable. Parasitic cuckoos are an extraordinary example of mimicry with diverse strategies to trick hosts into rearing their young (Davies 2011). Some cuckoos closely mimic the eggs of their hosts (Brooke and Davies 1998; Moskát et al. 2008; Spottiswoode 2010; Stoddard and Stevens 2010), and others mimic host nestlings (Langmore et al. 2011). Studies of brood parasitism in cuckoos have predominantly focused on egg as well as chick mimicry, and have highlighted the drastic impact on host reproductive potential that has set the scene for a well-documented coevolutionary arms race (Davies, Brooke, and Kacelnik 1996; Davies 2000; Soler and Soler 2000; Grim 2006; Spottiswoode and Stevens 2011). However, given that parasitism begins with depositing eggs in a host nest, blocking access to the nest has the greatest potential to minimize reproductive costs (Moksnes et al. 2000; Davies and Welbergen 2009; Feeney, Welbergen, and Langmore 2012). As a consequence, brood-parasitic cuckoos appear to have evolved a range of strategies to gain access to host nests. For example, in the genera Clamator, Eudynamys and Scythrops it is reported that males elicit a mobbing response to distract hosts while females discreetly lay their eggs in host nests (Gaston 1976; Davies 2000). However, given the costs that hosts can impose on parasitic cuckoos, inconspicuousness should be favoured to evade detection (Davies and Brooke 1988; Davies and Welbergen 2008; Požgayová, Procházka, and Honza 2009; Langmore et al. 2011). Brood-parasitic cuckoos dupe hosts into treating the parasite as if it is something it is not by looking like either a harmless species (aggressive mimicry) or a harmful species (Batesian mimicry). Aggressive mimicry is suspected among drongo cuckoos (Surniculus spp.) that form foraging flocks with multiple species of small passerines, and contribute to the flock by acting as a predator sentinel (Feeney, Welbergen, and Langmore 2012). Batesian mimicry in cuckoos has been The mechanisms underlying convergent evolution in the plumage patterns of birds 107 suspected since the time of Aristotle owing to their remarkable resemblance to raptors, especially Accipiter hawks (Wallace 1889). Their striking visual similarity is derived from their yellow eyes and legs, their size and shape, flight patterns and barred underparts (Honza et al. 2006; Davies and Welbergen 2008; Payne 2010; Welbergen and Davies 2011; Trnka, Prokop, and Grim 2012). In addition, it has been suggested that the rufous morph of the common cuckoo, Cuculus canorus, might represent mimicry of the Eurasian kestrel, Falco tinnunculus (Trnka and Grim 2013). Some species additionally possess polymorphisms in the colour of their barred plumage to thwart hosts that can see past the cuckoos’ disguise. These polymorphic species are likely to possess multiple hawk-like features demonstrating that the host- parasite arms race probably also occurs in the adult phenotype of parasites (Thorogood and Davies 2012; Trnka and Grim 2013) and the alternative morph may represent frequency-dependent mimicry for an additional model, the Eurasian kestrel (Honza et al. 2006; Thorogood and Davies 2012; Trnka and Prokop 2012; Trnka and Grim 2013). In Old World cuckoos, barred plumage evolved within the context of host- parasite coevolution, suggesting that it is an adaptive strategy to facilitate access to host nests (Kruger, Davies, and Sorenson 2007). This type of plumage pattern covers most of the ventral surface and is composed of within-feather alternating light and dark pigmentation, transversal to the feather’s axis (Payne and Sorensen 2005; Bortolotti et al. 2006; Gluckman and Cardoso 2010; Payne 2010). Studies of Batesian mimicry in parasitic cuckoos have focused on the common cuckoo, which is thought to mimic the Eurasian sparrowhawk, Accipiter nisus (Davies and Welbergen 2008; Davies and Welbergen 2009; Welbergen and Davies 2011). Field experiments measuring host responses to models of the common cuckoo demonstrate that barring can constrain host aggression at close range and that polymorphisms in barred plumage coloration can thwart detection by hosts (Honza et al. 2004; Grim 2005; Moksnes et al. 2007; Thorogood and Davies 2012; Trnka and Prokop 2012; Trnka and Grim 2013). By looking like a harmful model, the common cuckoo can facilitate access to host nests in which to place its eggs and when hosts learn to discriminate, altering the phenotype can be a successful strategy. In Old World parasitic cuckoos (Cuculidae) 35 species out of 58 have barred plumage, which is also common in both sparrowhawks (Accipiter) and other raptor genera (Payne and Sorensen 2005; Christie and Ferguson-Lees 2010). Given that resembling a local The mechanisms underlying convergent evolution in the plumage patterns of birds 108 dangerous species has allowed the common cuckoo to flourish, Batesian mimicry via barred plumage patterns may be widespread in Old World cuckoos to facilitate access to host nests, and many types of raptors may be the dangerous model. Fig. 4.1. Barred plumage patterns. Although barred plumage is a common and easily identifiable pattern, it can vary extensively as shown on the same scale in (a) Cuculus saturatus and (b) Aviceda cuculoides. Photographs of plumage within the figure are copyright of the Natural History Museum and were taken by Thanh-Lan Gluckman. Barred plumage may have multiple functions in brood parasitism. It has been well documented that avian brood parasites monitor host nests discreetly from nearby perches (Alvarez 1993; Hauber and Russo 2000; Begum et al. 2011). At a distance, barred plumage may provide camouflage while the parasite is searching for and watching potential hosts, but when detected at close range it may contribute to hawk mimicry and constrain aggression; however, it has also been suggested that it may increase aggression by some hosts towards parasites (Honza et al. 2006; Welbergen and Davies 2008; Payne 2010; Welbergen and Davies 2011; Trnka and Prokop 2012). Barred plumage can vary extensively in size, spacing, contrast and The mechanisms underlying convergent evolution in the plumage patterns of birds 109 relative importance of the main marking size to the overall pattern (Fig. 4.1). If barred plumage patterns function in mimicry in brood-parasitic cuckoos, recognition of local raptors as predators by hosts (i.e. host discrimination and response) should drive a similarity in plumage pattern attributes to be geographically specific (Bates 1862; Briskie, Sealy, and Hobson 1992; Lindholm and Thomas 2000; Hale and Briskie 2007; Pfennig and Mullen 2010). Materials and Methods We sampled representatives of Old World parasitic cuckoos with barred plumage on the basis of taxonomy and geographical distribution. According to Howard and Moore (2003) there are six genera containing parasitic cuckoos with barred plumage: Cuculus, Chrysococcyx, Eudynamys, Cacomantis, Cercococcyx and Scythrops. The genus Scythrops has limited barring on the lower part of the ventral surface and, given its orientation when the bird is flying towards hosts, the barring is unlikely to function in mimicry; we thus removed this genus from further consideration. We focused our sampling efforts on cuckoos and raptors with a restricted distribution range in the tropics of Africa and Oceania (Asia and eastern Australia), where parasitic cuckoos are concentrated (Yom-Tov and Geffen 2005). North and South American species are not represented in this study as only a few brood-parasitic cuckoos are found on these continents and they do not have barred plumage (Payne and Sorensen 2005). To assess overlap of distributions we scanned maps from Raptors of the World (Christie and Ferguson-Lees 2010) and The Cuckoos (Payne and Sorensen 2005) for comparison by eye. For the purposes of this study it was important to establish approximate geographical range overlap, but not precise estimates of the extent of overlap. Therefore, we scored the overlapping distribution of raptors on the basis of up to 25%, 50%, 75% and 100% of that of cuckoos (Table 4.1). We aimed to sample all sympatric raptors for the cuckoos chosen. Where possible, most raptors that have a distribution overlap of greater than 0.25 were sampled. All raptors with the largest overlapping distribution category were sampled, except for Hieraaetus kienerii, which could not be sampled (Table 4.1). To ensure that these results are representative of geographical range overlap, we compared a subset of African species with a regional field guide and although there was some variation in the extent of overlap, both guides described sympatry for each The mechanisms underlying convergent evolution in the plumage patterns of birds 110 pair considered (supplementary Table S4.1; Sinclair and Ryan 2010). In this study we aimed to sample each genus, while controlling for pattern similarity resulting from phylogeny. Therefore, we randomly selected a species to represent one barred parasitic genus from Africa and similarly a barred parasitic species to represent a separate genus from Oceania, for which there was a minimum of five undamaged museum skins at the Natural History Museum at Tring, U.K. An additional cuckoo from a separate genus was then selected that had an overlapping distribution with the first cuckoo selected, to test whether cuckoos from separate genera resembled the same raptor. The first randomly selected cuckoo representing each genus had an adequate number of undamaged specimens. This sampling design also allowed us to model whether sympatry and allopatry predict pattern similarity and dissimilarity by comparing species within and between geographical areas. Where species exhibited multiple subspecies, for example Eudynamys scolopacea, we sampled the nominate subspecies. One cuckoo species in this study, Cuculus saturatus, is polymorphic and we focused our sampling efforts on the primary (more frequent) morph (Voipio 1953). We sampled sympatric raptors with subspecies and polymorphisms in the same way. Raptor species were selected on the basis of sympatry and the presence of barred plumage (Table 4.1). Only raptors that prey on live birds were included in the study - vultures were excluded as they are carrion eaters (supplementary Table S4.2). In some instances, raptors were illustrated with barred plumage in the books, but on inspection of specimens had little or negligible barring (i.e. barring only on the legs) and were thus removed from the study. In addition, some raptors could not be sampled owing to a lack of specimens and/or access (Table 4.1, species marked with an asterisk). This resulted in six raptors for the African cuckoos, and eight raptors for the Oceanian cuckoos, representing seven genera in total.       The mechanisms underlying convergent evolution in the plumage patterns of birds 111 Table 4.1. Approximate distribution overlap between cuckoos and their respective raptor. *Species not represented owing to a lack of access/intact specimens. Species Eudynamys scolopacea Chrysococcyx flavigularis Cacomantis sonneratii Cercococcyx mechowi Cuculus saturatus Aviceda cuculoides - <1 - <1 - *Aviceda subcristata <0.25 - - - <0.25 Aviceda leuphotes <0.50 - <0.50 - <0.50 *Henicopernis longicauda <0.25 - - - <0.25 Pernis ptilorhyncus <0.50 - <0.50 - <0.50 *Pernis celebensis <0.25 - - - <0.50 *Circaetus cinereus - <0.25 - <0.50 - Polyboroides typus - <1 - <1 - *Circus assimilis <0.25 - - - <0.25 Micronisus gabar - <0.25 - <0.50 - Accipiter trivirgatus <0.50 - <0.50 - <0.50 Accipiter tachiro - <1 - <0.75 - Accipiter soloensis <0.50 - <0.50 - <0.75 Accipiter fasciatus <0.25 - - - <0.25 Accipiter gularis <0.50 - <0.50 - <0.75 Accipiter virgatus <0.25 - <0.50 - <0.25 Accipiter ovampensis - <0.25 - <0.25 - *Accipiter melanoleucus - <0.75 - <0.75 - *Accipiter meyerianus <0.25 - - - <0.25 *Megatriorchis doriae <0.25 - - - <0.25 Kaupifalco monogrammicus - <1 - <1 - *Harpyopsis novaeguineae <0.25 - - - <0.25 *Hieraaetus morphnoides <0.25 - - - <0.25 *Hieraaetus kienerii <0.25 - <0.50 - <0.75 Spizaetus nanus <0.25 - <0.50 - <0.50 *Falco moluccensis <0.25 - <0.25 - <0.50 *Falco longipennis <0.25 - - - <0.25 The mechanisms underlying convergent evolution in the plumage patterns of birds 112 Currently, there is little information on how barred plumage functions in host- parasite interactions outside of the common cuckoo (Payne and Sorensen 2005). For example, it is unknown whether the male attracts a mobbing response from hosts while the female lays her eggs, or perhaps females approach the nest by themselves. Therefore, we sampled both males and females in species that have barred plumage (Table 4.2). Given that birds approach a nest facing forwards, it is more likely that the most important part of the patterning, where used in mimicry, is the ventral surface as well as the flanks. Therefore, we focused our sampling on these body regions, except in Chrysococcyx flavigularis for which we could sample only the ventral surface to avoid damage to the flanks/wing. Patterning can vary within barred plumage so we sampled from the top, middle and bottom of the ventral and left flank area of specimens, avoiding the brood patch, to assess pattern variation (Gluckman and Cardoso 2009). Following Stevens et al. (2007) and Stoddard and Stevens (2010), we collected digital images of museum specimens with a Fujifilm IS Pro UV-sensitive camera with known spectral sensitivity, equipped with a quartz CoastalOpt UV lens (Coastal Optical Systems). Images were captured with a human pass filter (Baader UV/IR cut filter: 400-700 nm) as well as a UV pass filter (Baader U filter: 300-400 nm) at a standard distance (173 cm) at eye level with constant illumination provided by a Kaiser RB260 digital lighting unit. All images were linearized to control for a nonlinear response to changes in radiance prior to image analysis (Stevens, Parraga, et al. 2007). Current evidence shows that avian pattern perception occurs via achromatic (luminance) vision, which is encoded in the double cones (Jones and Osorio 2004; Osorio and Vorobyev 2005). To assess which visual system is most representative of hosts we collated information on host and prey species for cuckoos and raptors of these regions. Although the information available is limited, from what is recorded, the hosts and prey species of all cuckoos and raptors sampled are predominantly passerines (supplementary Table S4.2). The blue tit, Cyanistes caeruleus, has the best-studied visual system of any passerine (Hart and Hunt 2006; Hart et al. 2012) and has been widely used as a model for passerine vision in previous studies (Spottiswoode and Stevens 2010; Stoddard and Stevens 2010; Spottiswoode and Stevens 2011). Therefore, to account for passerine vision, all images were transformed from camera colour space to the relative photon catches of the double cones of a blue tit (Stevens et al. 2006; Stevens, Parraga, et al. 2007). The mechanisms underlying convergent evolution in the plumage patterns of birds 113   Table 4. 2. Female and male sample size for each cuckoo and raptor represented. Female Male Total Cuckoos Eudynamys scolopacea 11 N/A 11 Chrysococcyx flavigularis 3 2 5 Cacomantis sonneratii 5 5 10 Cercococcyx mechowi 5 5 10 Cuculus saturatus 5 5 10 Raptors Aviceda cuculoides 4 3 7 Aviceda leuphotes 5 5 10 Pernis ptilorhyncus N/A 4 4 Polyboroides typus 5 5 10 Micronisus gabar 5 5 10 Accipiter trivirgatus 5 5 10 Accipiter tachiro 5 5 10 Accipiter soloensis 5 N/A 5 Accipiter fasciatus 5 5 10 Accipiter gularis 10 10 20 Accipiter virgatus 4 5 9 Accipiter ovampensis 3 5 8 Kaupifalco monogrammicus 5 5 10 Spizaetus nanus N/A 5 5   Barred plumage is a relatively simple pattern that consists of the same regularly repeating motif, which is found in many species of birds worldwide (Kenward et al. 2004; Riegner 2008; Gluckman and Cardoso 2009; Gluckman and Cardoso 2010). As such, there is likely to be an inherent degree of similarity in plumage pattern attributes. Therefore, we employed a granularity analysis to objectively quantify multiple aspects of patterning in the transformed images of each specimen, following methods employed to quantify the camouflage of cuttlefish as The mechanisms underlying convergent evolution in the plumage patterns of birds 114 well as mimicry in egg patterning (Barbosa et al. 2008; Chiao et al. 2009; Spottiswoode 2010; Stoddard and Stevens 2010; Spottiswoode and Stevens 2011). A granularity analysis is akin to early-stage visual processing, which breaks down information into different spatial scales depending on the receptive fields (Godfrey, Lythgoe, and Rumball 1987). To simulate this process, we fast Fourier transformed each image with seven octave-wide isotropic band-pass filters. Each filter size, or granularity band, captures information at different spatial scales where large filter sizes correspond to small markings with high spatial frequency, and small filter sizes correspond to large markings with low spatial frequency. Information from each spatial scale is represented by the granularity spectra for an image, from which aspects of patterning can be measured (Fig. 4.2). Several absolute objective aspects of patterning were calculated from the granularity spectra (Fig. 4.2): the main marking size (the peak of the line), the importance of the main marking size to the overall pattern (proportion of power across all filter sizes that matches the peak value) and pattern contrast (overall amplitude of the spectrum across all spatial scales) (Stoddard and Stevens 2010). In addition, given that barred plumage is made up of a regularly repeating sub-pattern (Kenward et al. 2004; Gluckman and Cardoso 2009), we calculated the overall standard deviation of the main marking size over all spatial scales as a measure of pattern uniformity. Barred plumage attributes can vary within patches of patterning (Gluckman and Cardoso 2009). To assess whether the pattern attributes measured in this study varied within species, we compared patterns from the ventral surface and flank area, as well as the top, middle and bottom of each surface sampled. As the results were qualitatively similar within and between patches of plumage patterns within species (supplementary Fig. S4.1), we averaged the total values for each patch per individual for the final analysis. To ensure that the age of the museum specimens did not affect the integrity of the pattern data (Doucet and Hill 2009), we correlated each measure of pattern with year of collection (when available) where there was a minimum of eight individuals to ensure that a low sample size did not bias the correlation, resulting in a sample size of 79 individuals. The data were not normally distributed so we report the median rather than the mean. To estimate robust confidence intervals, we used 1000 bootstrap bias-corrected and accelerated simulations (BCA) using the The mechanisms underlying convergent evolution in the plumage patterns of birds 115 bootstrap module in PASW 20 (supplementary Fig. S4.2) (Ruscio and T. Mullen 2012). Fig. 4.2. Granularity spectra for an individual specimen of Aviceda cuculoides and Accipiter trivirgatus (representative samples of plumage are on the same scale). The granularity spectra of patterning are made up of the normalized energy at each filter size and indicate different aspects of patterning. For example, the granularity spectra peak at a marking filter size of two for A. trivirgatus, and four for A. cuculoides, which shows that A. trivirgatus has larger markings. The amplitude of the spectrum for the granularity spectra of A. cuculoides is much lower than that for A. trivirgatus indicating that A. trivirgatus has more contrast in its patterns than A. cuculoides. Photographs of plumage within the figure are copyright of the Natural History Museum and were taken by Thanh-Lan Gluckman. To compare similarity between the barred plumage of each cuckoo-raptor pair, as well as assess the extent of overlap in pattern attributes, we used a The mechanisms underlying convergent evolution in the plumage patterns of birds 116 nonparametric measure of effect size: probability of superiority (PS), also known as stochastic superiority (Nakagawa and Cuthill 2007; Ruscio and Mullen 2012). PS is the probability that an individual from species A has a higher scoring pattern variable (i.e. main marking size) than an individual from species B. A PS value between 0.5 and 0.71 indicates that the sample populations are highly similar for the trait measured (Grissom 2004). We used the slightly increased threshold of 0.73 as it is still within the realm of biological importance (Nakagawa and Cuthill 2007). For example, a PS value of 0.6 for main marking size in a cuckoo-raptor pair indicates that the two species being assessed have plumage patterns that overlap in size and match well for this attribute. To have a high degree of similarity the cuckoo-raptor pair must match in all four pattern attributes measured. Habitat may select for plumage pattern attributes to be similar in sympatric cuckoos and raptors to evade detection by prey and hosts alike. A posteriori, to examine whether habitat may be a confounding factor we compared the habitat of each species to assess its influence on cuckoo plumage patterns. If habitat shapes the plumage patterns of cuckoos, cuckoo-raptor pairs that have the highest number of overlapping habitat types should be most similar. Information on habitat was collated and tabulated for final comparison (Payne and Sorensen 2005; Christie and Ferguson-Lees 2010). No statistical tests were applied to these data owing to the small sample size per habitat type. Nevertheless, this represents the maximum available information for each cuckoo sampled and is indicative of the selection habitat may pose on plumage patterns. Finally, to assess whether cuckoos look more like a local dangerous model than an allopatric raptor, we modeled sympatry/allopatry as a predictor of pattern similarity. As mentioned previously, barring is a ubiquitous pattern that is likely to exhibit a basic degree of similarity in birds that express this trait because it is a relatively simple pattern. However, it is unlikely that barred plumage will vary in all four measures described simultaneously as a result of the inherent simplicity of the pattern. Therefore, we measured the four-dimensional Euclidean distance between each specimen of cuckoo and every individual raptor in plumage pattern phenotypic space, which also accounts for within-species variation (Spottiswoode and Stevens 2011). A large Euclidean distance between individuals indicates dissimilarity, and a small distance indicates similarity. In addition, given that some similarity in patterning The mechanisms underlying convergent evolution in the plumage patterns of birds 117 was expected, and that we sampled multiple cuckoo and raptor species, as well as individuals within species, we modeled these data using generalized linear mixed modeling using the glmer function in the lme4 package in R (R Development Core Team 2012). To model whether geographical overlap predicts pattern similarity/dissimilarity, we compared the relative fit of sympatry and allopatry as a predictor variable for comparison with a model without a predictor variable: glmer(Euc_distance~Symp_Allopatric+(Raptor|Raptor_ID)+(Cuckoo|Cuckoo_ID)), glmer(Euc_distance~+(Raptor|Raptor_ID)+(Cuckoo|Cuckoo_ID)), respectively. Together this represents the hypothesis that cuckoos look more similar to a local sympatric raptor and the alternative, that geographical overlap does not fit the data well and that there may be an alternative reason for a similarity in barred plumage between cuckoos and raptors. Finally, we used the Akaike information criterion (AIC) and the evidence ratio statistic (ER), to assess which model is a better fit for the data as well as assess the magnitude to which the better fitting model explains the variation (Burnham and Anderson 2002; Symonds and Moussalli 2010). We used AIC rather than AICc as the number of observations was more than 40 times the number of explanatory variables (Anderson et al. 2001). Results There was no correlation between year of collection and any measure of patterning between or within species (main marking size: 0.277 < R < 0.622, P > 0.055; proportion of power: 0.303 < R < 0.983, P > 0.116; contrast: 0.340 < R < 0.771, P > 0.091; marking size deviation: 0.342 < R < 0.821, P > 0.046). Granularity spectra differed between most of the cuckoo species sampled although there was some similarity between Cacomantis and Cercococcyx. In addition, there was substantial variation among the granularity spectra of raptors sampled (Fig. 4.3). Cuckoos predominantly matched sympatric raptors for between one and three pattern attributes. In three of the five cuckoos, there was matching for all four pattern attributes with at least one raptor (Fig. 4.3, Table 4.3). For most cuckoo species, there was no pattern matching in at least one raptor, except in the African Chrysococcyx, but for Eudynamys, there were two raptors with no pattern overlap (Fig. 4.3, Table 4.3). Two of the four closely matched The mechanisms underlying convergent evolution in the plumage patterns of birds 118 raptor species were Accipiter species. In addition, cuckoos closely resembled species of the genera Aviceda and Polyboroides (Table 4.3). There was variation in which pattern attribute was most matched across sympatric raptors: Cercococcyx matched most raptors for contrast, Chrysococcyx and Cacomantis predominantly matched raptors in main marking size and proportion of power, Cuculus matched most raptors for main marking size, whereas Eudynamys predominantly matched raptors for proportion of power (Table 4.3). In addition, there was no pattern matching between any allopatric cuckoo-raptor pair (Table 4.4). The mechanisms underlying convergent evolution in the plumage patterns of birds 119 Fig. 4.3. Granularity spectra for cuckoos and sympatric raptors, with pattern matching in four attributes on the basis of the probability of superiority. Spectra are shown for each cuckoo and its sympatric raptors: (a) African spp.; (b) Oceanian spp. The black line indicates the cuckoo. A grey dotted line indicates no pattern matching over any The mechanisms underlying convergent evolution in the plumage patterns of birds 120 attribute, a solid grey line indicates matching for one attribute, green for two attributes, blue for three attributes and red for all four pattern attributes. The measures of probability of superiority are in Table 4.3, and the median and sample sizes used for these calculations are in Table 4.2. The habitat types inhabited by each cuckoo and raptor were variable between species, as were the number of overlapping habitat types (Table 5, supplementary Table S4.3, S4.4). The number and type of habitats used by each pair predicted plumage pattern similarity only in Cacomantis and Accipiter fasciatus (Fig. 4.3, Table 4.5). The mechanisms underlying convergent evolution in the plumage patterns of birds 121 Table 4.3. Comparison of barring attributes of plumage patterns between each cuckoo and sympatric raptor on the basis of probability of superiority (PS). Cuckoo Sympatric raptor Main marking size Proportion of power Contrast Marking size deviation Number of matching attributes Eudynamys scolopacea Aviceda leuphotes 1 0.63 0.61 0.6 3 Pernis ptilorhyncus 0.65 0.98 0.65 0.8 2 Accipiter trivirgatus 1 0.57 0.6 0.52 3 Accipiter soloensis 0.9 1 1 1 0 Accipiter fasciatus 0.66 0.67 0.97 0.93 2 Accipiter gularis 0.81 0.63 0.94 0.91 1 Accipiter virgatus 0.93 0.92 0.94 0.96 0 Spizaetus nanus 0.82 0.54 0.7 0.7 3 Chrysococcyx flavigularis Aviceda cuculoides 0.6 0.6 1 1 2 Polyboroides typus 0.72 0.6 0.72 0.5 4 Micronisus gabar 0.58 0.6 1 0.98 2 Accipiter tachiro 0.52 0.6 1 0.98 2 Accipiter ovampensis 0.75 0.6 1 1 1 Kaupifalco monogrammicus 0.68 0.6 1 1 2 Cacomantis sonneratii Aviceda leuphotes 0.82 0.56 0.96 0.97 1 Pernis ptilorhyncus 0.58 0.93 0.9 0.6 2 Accipiter trivirgatus 0.82 0.62 0.9 0.89 1 Accipiter soloensis 0.87 1 1 1 0 Accipiter fasciatus 0.7 0.52 0.6 0.54 4 Accipiter gularis 0.54 0.65 0.55 0.65 4 Accipiter virgatus 0.69 0.9 0.53 0.59 3 Spizaetus nanus 0.6 0.58 1 1 2 Cercococcyx mechowi Aviceda cuculoides 0.6 0.73 0.57 0.52 4 Polyboroides typus 0.97 0.51 0.89 0.89 1 The mechanisms underlying convergent evolution in the plumage patterns of birds 122 Micronisus gabar 0.93 0.86 0.66 0.76 1 Accipiter tachiro 0.75 0.66 0.57 0.61 3 Accipiter ovampensis 0.99 0.98 0.73 0.78 1 Kaupifalco monogrammicus 0.99 0.99 0.72 0.77 1 Cuculus saturatus Aviceda leuphotes 0.65 0.9 0.72 0.65 3 Pernis ptilorhyncus 0.97 1 0.94 0.98 0 Accipiter trivirgatus 0.67 0.92 0.67 0.6 3 Accipiter soloensis 0.65 1 1 1 1 Accipiter fasciatus 0.97 0.7 1 1 1 Accipiter gularis 0.71 0.87 0.98 0.98 1 Accipiter virgatus 0.68 1 0.99 1 1 Spizaetus nanus 0.63 0.83 0.63 0.4 3 The mechanisms underlying convergent evolution in the plumage patterns of birds 123 Fig. 4.4. Box plot of the Euclidean distance in four-dimensional plumage pattern space between each individual cuckoo and raptor. Allopatric and sympatric cuckoos and raptors are presented in separate box plots with their individual mean, interquartile ranges and outliers. The 25th to 75th quartiles are encapsulated in the box and dashed lines indicate the lower 25th and upper 75th quartile ranges. The mechanisms underlying convergent evolution in the plumage patterns of birds 124 Table 4.4. Comparison of barring attributes of plumage patterns between each cuckoo and allopatric raptor on the basis of probability of superiority (PS). A range of PS values of 0.5–0.71 indicates overlap and therefore a high degree of similarity, therefore a PS value of 0.5 - 0.73 was considered to have an effect (Nakagawa and Cuthill 2007). Cuckoo Allopatric raptor Main marking size Proportion of power Contrast Marking size deviation Number of matching attributes Eudynamys scolopaceus Aviceda cuculoides 1.00 1.00 0.99 1.00 0 Polyboroides typus 1.00 1.00 1.00 1.00 0 Micronisus gabar 1.00 1.00 1.00 1.00 0 Accipiter tachiro 1.00 1.00 0.99 1.00 0 Accipiter ovampensis 1.00 1.00 1.00 1.00 0 Kaupifalco monogrammicus 1.00 1.00 1.00 1.00 0 Chrysococcyx flavigularis Aviceda leuphotes 1.00 1.00 1.00 1.00 0 Pernis ptilorhynchus 0.99 1.00 1.00 1.00 0 Accipiter trivirgatus 1.00 1.00 1.00 1.00 0 Accipiter soloensis 1.00 0.97 0.99 0.97 0 Accipiter fasciatus 1.00 1.00 1.00 1.00 0 Accipiter gularis 1.00 1.00 1.00 1.00 0 Accipiter virgatus 1.00 1.00 1.00 1.00 0 Spizaetus nanus 1.00 1.00 1.00 1.00 0 Cacomantis sonneratii Aviceda cuculoides 1.00 1.00 0.99 1.00 0 Polyboroides typus 1.00 1.00 0.99 1.00 0 Micronisus gabar 1.00 1.00 1.00 1.00 0 Accipiter tachiro 1.00 1.00 0.99 0.99 0 Accipiter 1.00 1.00 1.00 1.00 0 The mechanisms underlying convergent evolution in the plumage patterns of birds 125 ovampensis Kaupifalco monogrammicus 1.00 1.00 1.00 1.00 0 Cercococcyx mechowi Aviceda leuphotes 1.00 1.00 1.00 1.00 0 Pernis ptilorhynchus 0.98 1.00 1.00 0.99 0 Accipiter trivirgatus 1.00 1.00 1.00 1.00 0 Accipiter soloensis 1.00 1.00 1.00 1.00 0 Accipiter fasciatus 1.00 1.00 1.00 0.99 0 Accipiter gularis 1.00 1.00 1.00 1.00 0 Accipiter virgatus 1.00 0.99 0.99 0.99 0 Spizaetus nanus 0.99 1.00 1.00 1.00 0 Cuculus saturatus Aviceda cuculoides 0.99 1.00 1.00 1.00 0 Polyboroides typus 1.00 1.00 1.00 1.00 0 Micronisus gabar 1.00 1.00 0.99 1.00 0 Accipiter tachiro 1.00 1.00 1.00 1.00 0 Accipiter ovampensis 1.00 1.00 1.00 1.00 0 Kaupifalco monogrammicus 1.00 1.00 0.99 1.00 0 The mechanisms underlying convergent evolution in the plumage patterns of birds 126 Table 4.5. The number of overlapping habitat types between cuckoos and raptors. The number of habitats inhabited by each cuckoo is given after the species name. Raptor Eudynamys scolopacea, 13 Chrysococcyx flavigularis, 3 Cacomantis sonneratii, 7 Cercococcyx mechowi, 3 Cuculus saturatus, 7 Aviceda cuculoides - 1 - - - Aviceda leuphotes 6 - 5 - 2 Pernis ptilorhyncus 2 - 3 - 1 Polyboroides typus - 1 - 1 - Micronisus gabar - - - - - Accipiter trivirgatus 4 - 3 - 2 Accipiter tachiro - 3 - 1 - Accipiter soloensis 6 - 5 - 1 Accipiter fasciatus 6 - 5 - 1 Accipiter gularis 4 - 3 - 1 Accipiter virgatus 3 - 2 - 1 Accipiter ovampensis - - - - - Kaupifalco monogrammicus - 1 - 1 - Spizaetus nanus 3 - 3 - 1 The mechanisms underlying convergent evolution in the plumage patterns of birds 127 Discussion Previous work on the function of barred plumage in the common cuckoo revealed that it can reduce aggression in some hosts, which is congruent with Batesian mimicry (Davies and Welbergen 2008; Welbergen and Davies 2011; Thorogood and Davies 2012; Trnka and Prokop 2012; Trnka, Prokop, and Grim 2012; but see Honza et al. 2006). In Old World cuckoos, barring appears to have evolved with brood parasitism, but it was previously unknown whether this is perceptually relevant from the perspective of birds (Kruger, Davies, and Sorenson 2007). Although barring is common, it can vary greatly in a number of pattern attributes. Here, using objective digital image analysis we have shown that represen- tative species from all five genera of barred Old World parasitic cuckoos have barring that resembles that of raptors. Of these five cuckoo species, three are remarkably similar to a sympatric raptor; the potential models came from the genus Accipiter as well as Aviceda and Polyboroides. This demonstrates that the putative mimic need not be constrained to the genus Cuculus in Old World cuckoos and the model can be from multiple raptor genera. Finally, we found evidence that distribution overlap predicts plumage pattern similarity, suggesting that cuckoos resemble sympatric raptors more than allopatric raptors. Together these results demonstrate that cuckoos from all five genera of barred parasitic Old World cuckoos resemble a local dangerous model, conforming to a fundamental principle of Batesian mimicry. Therefore, barred plumage may function in cuckoo-hawk mimicry in many more species than the common cuckoo. Although barred plumage is an easily recognizable pattern, it can vary in multiple dimensions and the variation can be subtle (Fig. 4.3, supplementary Fig. S4.1). This underscores the importance of objectively quantifying patterns. Akin to other lines of the cuckoo-host coevolutionary arms race, if barred plumage broadly functions in cuckoo-hawk mimicry, there appears to be a range of mimicry strategies employed from matching multiple local models to specializing on one (Stoddard and Stevens 2010; Davies 2011). All cuckoos studied resemble a sympatric raptor, in some cases very closely (Fig. 4.3). Moreover, no allopatric cuckoo and raptor matched for any pattern attribute (Table 4.4). This indicates that Batesian mimicry, if indeed widespread, may be a well-developed strategy in the host-parasite arms race. There may not, however, be one clear strategy, similar to egg mimicry. For example, The mechanisms underlying convergent evolution in the plumage patterns of birds 128 C. flavigularis most closely resembles Polyboroides typus, whereas C. saturatus resembles multiple sympatric raptors over a range of pattern attributes (Fig. 4.3, Table 4.3). Perhaps in the former, hosts have evolved to become more discerning, whereas in the latter, cuckoo-hawk mimicry may be a relatively recent event akin to other stages of the host-parasite arms race (Stoddard and Stevens 2010; Davies 2011). Eudynamys scolopacea also resembles multiple raptors over a range of pattern attributes, which could reflect that males may use distraction displays to lure potential hosts from nests and perhaps in this system hosts are also less discerning than those of C. flavigularis. However, further study will be required to assess whether distraction displays are the predominant strategy in this species. We should also consider that Batesian mimicry might have evolved over a long period of time and that the present overlap of populations may not be representative of the historical distribution of each species. Although the current distribution of each species may not describe where and with which raptors and hosts cuckoo-hawk similarity evolved, once plumage patterns evolve they appear to be evolutionarily labile, which indicates that they may be easy to modify in relatively short periods of time (T-L. Gluckman and N.I. Mundy unpublished data). Therefore, although current distribution is not indicative of historical distribution overlap, discrimination behaviour by hosts could drive plumage pattern similarity in short periods of evolutionary time (Thorogood and Davies 2012; Trnka and Grim 2013) and in this study we could only analyse the best of what is currently known about their overlap. Hosts should be able to recognize broadly sympatric raptor species, which may not occur in precisely the same habitat on a finer scale (Pfennig and Mullen 2010). However, habitat may select for a similarity in barred plumage patterns (Thery 2001; Grim 2005; Endler 2007; Endler and Thery 2010). If this occurs, it would be expected that cuckoo-raptor pairs with the highest number of overlapping habitat types would be the most similar. However, this was not the case with the exception of Cacomantis sonneratii and A. fasciatus (Fig. 4.3, Table 4.4, 4.5, supplementary Table S4.3). These results, although crude in terms of characterization of habitat types, imply that selection for hawk mimicry, rather than camouflage, may be at least as important as habitat in driving the evolution of barred plumage in brood-parasitic cuckoos (Voipio 1953). Moreover, camouflage through barred plumage should The mechanisms underlying convergent evolution in the plumage patterns of birds 129 benefit both cuckoos and raptors, but similarly it should also benefit host and prey species, and other birds living in the same habitat. Yet many bird species sympatric with the study species do not have barred plumage, for example other raptors and passerine birds (supplementary Table S4.2). A link between barring and sedentary lifestyle in raptors has been proposed (Christie and Ferguson-Lees 2010), but there do not appear to be any studies showing statistical support for this, nor for a suggested link between barring and crypsis (Kruger, Davies, and Sorenson 2007). Future studies will need to characterize habitat attributes, or conduct field experiments, to elucidate the extent to which camouflage influences the evolution of barred plumage. Alternatively, a similarity in phenotypes may arise from phylogenetic constraint or random matching (Grim 2005). Convergence from phylogenetic constraint is unlikely given that barring is more prevalent in parasitic cuckoos (35/58) than in nonparasitic cuckoos (4/83) and evolved after the evolution of parasitism in the Cuculidae (Kruger, Davies, and Sorenson 2007). Random matching is also unlikely given that allopatry/sympatry better predicted plumage pattern similar- ity/dissimilarity than chance alone (Fig. 4.4). It has been suggested that variation in strategy is due to developmental constraint or progression in the arms race (Davies 2000; Stoddard and Stevens 2010). The current understanding of the mechanisms of plumage pattern formation is poor, and an added layer of complexity in cuckoo-hawk mimicry is that selection occurs via a third party, the host (Prum and Brush 2002). In this study we focused on barred plumage. However, other hawk-like features such as eye and leg coloration as well as polymorphisms are likely to represent important adaptations. Although these additional features may evolve via apostatic selection in some instances and selection for hawk mimicry in others, mimicry dynamics encompassing all features of hawk-like attributes is likely to be an important component of future research (Davies 2000; Thorogood and Davies 2012; Trnka and Prokop 2012; Thorogood and Davies 2013; Trnka and Grim 2013). In addition, this study encompasses only patterning but not colour, which will be investigated in a future study. Other important areas requiring investigation include selection on the host or prey to recognize barring as a threat in the systems we presented here. Our study has shown that there is a remarkable similarity of barred plumage in all genera of Old World barred parasitic The mechanisms underlying convergent evolution in the plumage patterns of birds 130 cuckoos and their sympatric raptors from the perspective of the host/prey. We hope this fosters further research beyond the common cuckoo and Eurasian sparrowhawk relationship. The mechanisms underlying convergent evolution in the plumage patterns of birds 131 The mechanisms underlying convergent evolution in the plumage patterns of birds 132 Chapter 5: Ecological selection for bird plumage patterns worldwide This project was conducted in collaboration with Marius Somveille and Kate Marshall, two current PhD students at the University of Cambridge. I conceived the project, collected all of the plumage pattern data, contributed to the literature review of the function of patterns in camouflage and communication, conducted all comparative testing as well as contributed some philosophical aspects of the ecological analysis. Marius designed and conducted the ecological analyses, and Kate contributed most of the literature review of the function of patterns in camouflage. Supplementary Table S5.3 is supplied on a CD-ROM on the back cover of this thesis. Abbott H. Thayer, Peacock in the Woods (1907). The mechanisms underlying convergent evolution in the plumage patterns of birds 133 The mechanisms underlying convergent evolution in the plumage patterns of birds 134 Abstract It is thought that animal coloration matches the background in which it is viewed to minimize detection by predators. In particular, patterns, e.g. bars and spots, may function in camouflage because patterns are less conspicuous against heterogeneous backgrounds. Avian plumage patterns have converged on just four types: irregular - mottled patterns; regular - scales, bars and spots. Little is known about the camouflage function of avian patterns. To infer the potential utility of bird plumage patterns we conducted a literature survey of patterns in animals. Our literature survey revealed remarkable similarity in pattern-based anti-predator defences in aquatic and terrestrial, and vertebrate and invertebrate animals: irregular patterns facilitate concealment when motionless in cluttered backgrounds, whereas regular patterns (e.g. bars) prevent capture during movement and probably require open habitats for movement. However, regular patterns also function in communication and are likely to evolve on the ventral surface of males and to be independent of habitat. There are 16 types of habitats worldwide that vary in structural composition, e.g. closed habitats are cluttered whereas open habitats are less visually noisy, and perhaps habitat may explain convergent evolution in bird plumage patterns. Alternatively, similar patterns may have evolved due to shared ancestry. We considered adult plumage patterns in each sex, including breeding and non-breeding plumages, in 8008 spp. of birds, and juvenile plumages of 2603 spp., to examine whether patterns a) evolved due to shared ancestry, and b) whether habitat has selected for convergent evolution in plumage patterns. Shared ancestry predominantly explained <30% of pattern evolution. All four patterns are found in all combinations of age/sex/breeding/dorsoventral plumages in all 16 habitats of the world. The rare significant association between patterns and the eco-regions of habitats was dominated by all patterns together (7.65% eco-regions) in contrast with individual patterns (1.69% eco-regions). Significant associations between eco- regions and patterns predominantly occurred in <25% of 15/16 habitats worldwide and patterns were not clustered in particular habitats. The rare significant associations between eco-regions and irregular and regular patterns were found in both closed and open habitats. A lack of association of the ventral breeding plumage of males and habitats is congruent with sexual selection. Therefore, habitat does not predict convergence in bird plumage patterns at the macroevolutionary scale. The mechanisms underlying convergent evolution in the plumage patterns of birds 135 The mechanisms underlying convergent evolution in the plumage patterns of birds 136 Introduction The prevailing view of signal evolution is that animal coloration matches the background in which it is viewed to minimize detection by predators (Wallace 1889; Poulton 1890; Bradbury and Vehrencamp 1998) but that sexually selected traits stand out (Darwin 1871; Andersson 1994). Animal phenotypes may be comprised of uniform coloration, such as the entirely black plumage of the common raven (Corvus corax), or spatially variant pigmentation, such as the barred and spotted patterns of the male Zebra Finch (Taeniopygia guttata). In comparison to uniform coloration, our understanding of the function of patterns is less well understood. It is generally thought that patterning is beneficial because patterns allow animals to be less conspicuous against heterogeneous backgrounds (Fig. 5.1). The general principle underlying the function of patterns in camouflage is that patterns resemble aspects of the background in order to evade detection by predators or prey. In the context of signaling, patterns must visually diverge from the background to stand out to conspecifics. A) B) C) Fig. 5.1. The hypothesis of animal pattern strategies. Considering animal patterns in the context of habitat, patterns can A) stand out, B) conceal the animal if its pattern is similar to its background, and C) stand out if the pattern is dissimilar to its background pattern (Fig. 8.16 in Bradbury and Vehrencamp 1998). The mechanisms underlying convergent evolution in the plumage patterns of birds 137 Signal evolution theory predicts that a visual signal will evolve to be effective for the function it serves (Endler 1978; Bradbury and Vehrencamp 1998; Kenward et al. 2004). It is thought that animals exploit the visual patterning of their background, including vegetation and the substrate, to either camouflage themselves or to stand out to conspecifics. For example, bold patterns on a plain background make an animal standout, having a similar pattern to the background allows the animal to blend in, whereas patterns that oppose the geometric pattern of the background are conspicuous (Fig. 5.1). Birds are an ideal system in which to test ecological selection for patterns as they have evolved multiple pattern types (Fig. 5.2) and inhabit a wide variety of habitats on all major landmasses. In spite of the diversity of avian phenotypes, plumage patterns have repeatedly converged on the same four types: Irregular - mottled plumage where the vane is heterogeneously pigmented; Regular – scales where pigmentation follows the edge of the vane and may be concentric, bars which are made of alternating dark and light pigmentation transversal to the feathers axis, and spots where one or more spots pigment each feather (Fig. 5.2). Feather patterns mostly consist of combinations of neutral colours comprising brown, buff, cream, black and sometimes grey (T-L. Gluckman unpublished data), which are all formed from melanin, as well as white, which is due to an absence of melanin (Mundy et al. 2004). More striking coloration, including green, red and blue are infrequent in the patterns of birds. Other types of patterns, such as the stripes of the vulturine guineafowl (Acryllium vulturinum), are comparatively rare. In addition, there is a pattern that is only found in one species of bird, the checkerboard pattern of the common loon (Gavia immer), and another pattern is found in less than ten species, the triangles in the breast of the male Wood duck (Aix sponsa) (T-L. Gluckman unpublished data). Therefore, birds can make other types of patterns but repeatedly converge on just the same four types (Fig. 5.2). Perhaps convergence in bird plumage patterns has arisen because these patterns are the most effective in camouflage and/or communication relative to the habitat in which each species’ lives (Poulton 1890; Thayer 1909; Cott 1940; Endler 1978). The mechanisms underlying convergent evolution in the plumage patterns of birds 138 Fig. 5.2. The prevalent feather patterns of birds. In spite of the diversity of avian phenotypes, feather patterns have repeatedly converged on the same four patterns: Irregular - a) mottled plumage in a female sharp-tailed grouse (Tympanachus phasianellus); Regular - b) barred plumage in a male Andean goose (Chloephaga melanoptera), c) scaled plumage in a male falcated duck (Anas falcata), d) spotted plumage in a male great argus (Argusianus argus). Figure taken from Marshall and Gluckman, in review. The main physical characteristics of visual signals are 1) intensity of the signal (brightness), 2) colour, 3) spatial characteristics, and 4) temporal variability of these three attributes (Bradbury and Vehrencamp 1998). Bird plumage patterns are dominantly composed of neutral colouration and current evidence suggests that texture (patterns) are detected by achromatic vision (Jones and Osorio 2004). Spatial and temporal characteristics of the pattern are likely to be important attributes in the context of the background in which they are viewed. For example, the relationship between spatial attributes of the pattern and the habitat changes in different combinations (Fig. 5.1) because pattern type can vary among juveniles and adults, and between breeding and non-breeding phenotypes (Björklund 1991; Gluckman and Cardoso 2010). Brightness and coloration are likely to have an important function in camouflage, but as a first attempt at ecological selection for plumage patterns we focused on the most prominent aspect of patterning: the spatial arrangement of pigmentation into the four types of patterning in relation to the spatial characteristics The mechanisms underlying convergent evolution in the plumage patterns of birds 139 of the background in which they are viewed (Fig. 5.2). Currently, there is little empirical evidence for the function of bird plumage patterns. To assess the function of animal patterns, we conducted a broad survey of the literature spanning over 80 studies from vertebrates and invertebrates, including aquatic and terrestrial animals. Our literature review illustrates that the function of patterns is type and context dependent, but is nonetheless similar in many animal groups (Table 5.1; supplementary S5.1). Pattern-based anti-predator defenses can be subdivided into those that conceal while stationary and those that prevent capture during movement (Poulton 1890; Thayer 1909; Cott 1940; Bradbury and Vehrencamp 1998; Stevens et al. 2006; Mäthger et al. 2008; Stevens, Yule, and Ruxton 2008; Hanlon et al. 2009; Stevens and Merilaita 2009a; Stevens and Merilaita 2009b; Stevens et al. 2011). Patterns used in communication are likely to repeat a motif (sub-pattern, e.g. bars and spots) because repetition increases the likelihood that a signal will be received (Kenward et al. 2004). Accordingly, there is a growing body of evidence demonstrating that regular patterns function in communication (Hasson 1991; Petrie, Halliday, and Sanders 1991; Swaddle and Cuthill 1994; Omland 1996; Roulin 1999a; Roulin 1999b; Gluckman and Cardoso 2010; Muck and Goymann 2011). Table 5.1. The number of species for which empirical, comparative and correlational evidence has demonstrated the function of irregular or regular patterns in camouflage and/or communication, spanning vertebrates and invertebrates, as well as terrestrial and aquatic species (see supplementary Table S5.1 for source studies). Table 5.1 and corresponding data taken from Marshall and Gluckman, in review. It is likely that irregular patterns function in the camouflage of stationary animals as they appear to generally match the irregular pattern of one or several Irregular Regular Camouflage 8 7 Communication 1 7 The mechanisms underlying convergent evolution in the plumage patterns of birds 140 background types (background-matching) and/or seem to create false sets of edges (disruptive camouflage) to evade detection by predators (Thayer 1909; Cott 1940; Endler 1978; Bradbury and Vehrencamp 1998; Stevens and Merilaita 2009a; Stevens and Merilaita 2009b). For example, it has been suggested that non-breeding plumage in ducks, which often consists of irregular patterning (e.g. the genera Aythya and Somateria), facilitates concealment of females during nesting and of both sexes during the flightless period accompanying wing moult (Batt 1992). Although there are few empirical studies demonstrating the role of irregular plumage patterns in the camouflage of birds, there is substantial evidence from studies in 8 species spanning amphibians, insects, crabs, cephalopods, and bird eggs (Table 5.1; supplementary Table S5.1). For instance, in colour-changing organisms, the irregular mottled, stippled and disruptive patterns of cuttlefish (Sepia officinalis) facultatively match their backgrounds for crypsis (Chiao, Kelman, and Hanlon 2004; Chiao, Chubb, and Hanlon 2007; Hanlon et al. 2007; Barbosa et al. 2008; Mäthger et al. 2008; Hanlon et al. 2009; Chiao et al. 2010; Chiao et al. 2013), octopuses exhibit irregular body patterns that imitate nearby background objects (Josef et al. 2012), and the colour- changing fiddler crab (Uca vomeris) exhibits irregular, mottled patterns under predation to be cryptic against their mud background (Hemmi et al. 2006). Thus, static camouflage patterns are likely to consist of irregular pigmentation. In birds irregular mottled plumage is common in many species (Fig. 5.2), and is more frequently found in females and juveniles in comparison to regular barred plumage patterns (Gluckman and Cardoso 2010). Current evidence suggests that regular patterns, such as bars and spots, facilitate camouflage during movement and therefore function as a secondary defence during the escape of prey by exploiting specific features of receiver visual acuity, e.g. motion-dazzle and flicker-fusion (Thayer 1909; Pough 1976; Endler 1978; Endler 1980; Brodie 1989; Brodie 1992; Brodie 1993; Shine 1994; Lindell and Forsman 1996; Stevens, Hopkins, et al. 2007; Stevens, Yule, and Ruxton 2008; Stevens and Merilaita 2009b; Scott-Samuel et al. 2011; Allen et al. 2013; Helversen, Schooler, and Czienskowski 2013; How and Zanker 2014). The underlying mechanisms of motion-based camouflage can be explained by the repetitive nature of regular patterns whereby repetition causes perceptual lag or fatigue in vision based motion-sensitive cells (Snowden 1998) or illusions of motion (Stevens, Yule, and Ruxton 2008). Evidence for motion-based camouflage via regular patterning has The mechanisms underlying convergent evolution in the plumage patterns of birds 141 been demonstrated in at least 7 species comprising fish, snakes, mammals and cephalopods (Table 5.1; supplementary Table S5.1). For example, bars or stripes have been shown to reduce the ability of predators to estimate the speed and direction of moving prey via motion-dazzle, allowing prey to evade capture after being detected (Stevens, Yule, and Ruxton 2008; Scott-Samuel et al. 2011; Stevens et al. 2011; How and Zanker 2014). Regular patterns may also facilitate motion- based background matching via flicker-fusion camouflage, where patterns blend during movement to match the background (Jackson, Ingram, and Campbell 1976; Pough 1976; Endler 1978; Brodie 1989; Brodie 1992; Brodie 1993; Shine 1994; Lindell and Forsman 1996). Evidence for flicker-fusion comes from behavioural and survivorship correlations between regular colour patterns and anti-predator escape behaviour in garter snakes (Thamnophis ordinoides) (Brodie 1989; Brodie 1992; Brodie 1993). However, regular patterns may have a dual function in visual communication with conspecifics at close range (Marshall 2000). Comparative and empirical experiments show that regular patterns function in communication in a range of species (Hasson 1991; Petrie, Halliday, and Sanders 1991; Swaddle and Cuthill 1994; Omland 1996; Roulin 1999a; Bortolotti et al. 2006; Gluckman and Cardoso 2010; Roulin et al. 2010; Muck and Goymann 2011). Out of the 14 species that exhibit regular patterns that have been investigated (Table 5.1), 7 use patterns in communication. Signal obstruction, via objects such as trees and branches in a forest, may select for regularly repeating patterns, because repetition of a message increases the chances that the signal is detected by their intended receivers (Kenward et al. 2004). A comparative study in birds showed that, compared to irregular mottled patterns, barred plumage is more likely to evolve on the ventral surface of sexually mature males – a likely location for visual communication (Gluckman and Cardoso 2010). Empirical evidence demonstrates that barred patterns are positively associated with body condition in female barred buttonquails (Turnix suscitator; Muck and Goymann 2011), and the barred flanks of the red- legged partridge (Alectoris rufa) appear to function as a social signal (Bortolotti et al. 2006), whereas spots on tail feathers of barn swallows (Hirundo rustica) appear to be a reliable signal of phenotypic quality in males (Kose, Mänd, and Moller 1999), and plumage spottiness in female barn owls (Tyto alba) appears to be a signal of mate quality (Roulin 1999b). Moreover, a classical study shows that the number of The mechanisms underlying convergent evolution in the plumage patterns of birds 142 eyespots on male peacock (Pavo cristatus) tails is positively associated with mating success (Petrie, Halliday, and Sanders 1991). There is currently little evidence to suggest that irregular patterns may function in motion-based camouflage (supplementary Table S5.1). In addition, given that motion-dazzle or flicker-fusion camouflage strategies appear to be reliant on repetition of a regular motif to cause illusions of motion (Snowden 1998; Stevens, Yule, and Ruxton 2008) it seems unlikely that irregular mottled patterns function in motion-based camouflage. Therefore, the function of bird plumage patterns in camouflage and communication is likely to be linked to whether the pattern is irregular or regular (Table 5.1). The distribution of plumage patterns on the basis of age class (juvenile or adult), gender and on which surface the pattern is distributed can be indicative of function (Björklund 1991). Dorsal patterns are likely to function in camouflage in adults of both sexes, and juveniles, as shown in birds and lizards (Stuart-Fox and Ord 2004; Gluckman and Cardoso 2010; Garcia, Rohr, and Dyer 2013). In contrast, a communication function is likely to evolve on the ventral surface of males (Gluckman and Cardoso 2010). Such an arrangement provides a solution to antagonistic selection pressures in the form of signal partitioning where the signal is separated from the camouflage component, such as in Bicyclus butterflies and agamid lizards (Endler 1992; Stuart-Fox and Ord 2004; Oliver, Robertson, and Monteiro 2009), but is likely to be dependent on which background the animal is viewed against (Fig. 5.1). Across the world the main visual difference between habitats is structural composition. Given the function of patterns (Table 5.1), and that patterns are viewed in the context of their environment (Fig. 5.1), pattern type and pattern location over the body may be associated with habitat. Static irregular camouflage patterns are harder to detect when viewed against more cluttered backgrounds with distractors and other moving conspecifics (Dimitrova and Merilaita 2009; Dimitrova and Merilaita 2012; Hall et al. 2013). Therefore, perhaps background-matching/disruptive camouflage occurs in more visually complex environments such as closed habitats, but probably not in open habitats such as aqueous environments and deserts (Fig. 5.1). The motion-based camouflage function of regular patterning has been demonstrated where prey/target is in motion (Stevens, Yule, and Ruxton 2008; Scott- Samuel et al. 2011; Stevens et al. 2011; How and Zanker 2014). Given that space is The mechanisms underlying convergent evolution in the plumage patterns of birds 143 required for movement, perhaps motion-based camouflage patterns are favoured in more open habitats such as sparse woodland (Fig. 5.1). Similar plumage patterns may occur in closely related species due to phylogenetic relatedness rather than for adaptive reasons (Felsenstein 1985; Revell, Harmon, and Collar 2008). Therefore, we tested whether plumage patterns may have been retained due to shared ancestry in 8009 spp. of adult birds and 2603 spp. of juveniles. We then tested whether ecology selects for convergent evolution in bird plumage patterns. The majority of avian species across the world are land birds. A small number of avian species are marine (0.02% - 217 spp.) and their distribution, especially during the non-breeding season when marine birds are at sea, is not well known (The Nature Conservancy, http://maps.tnc.org/gis_data.html). In addition, their distribution is coarse and covers a wide area making inferences from these species difficult. Therefore, we focussed solely on land birds. A camouflage function is likely on both the ventral and dorsal surface of females and juveniles, and the dorsal surface of males (Houston, Stevens, and Cuthill 2007) whereas regular patterns that have a social function are likely to be displayed on the ventral surface of males (Gluckman and Cardoso 2010). Given that cluttered habitats may also select for regularly repeating patterns (Kenward et al. 2004), but that regular patterns may also stand out on uniform backgrounds (Fig. 5.1), male ventral patterns may be independent of habitat type. Based on our literature review and a current understanding of natural selection we expect to find a significant association between closed and open habitats with irregular and regular patterns, respectively (Table 5.1). Methods and materials Data collection The best phylogenetic reconstructions for the class Aves are those by Jetz et al. (2012) on the basis of species coverage and breadth of data. 10,000 trees are available for this phylogeny, which is drawn from a single Bayesian analysis. To control for phylogenetic uncertainty we randomly selected 100 phylogenetic trees provided by Jetz et al (2012) on which to run our analyses. To collect plumage The mechanisms underlying convergent evolution in the plumage patterns of birds 144 pattern information we referred to field guides covering all major landmasses: North and Central America (Sibley 2000; van Perlo 2006), South America and Antarctica (De La Pena and Rumboll, 1998; Restall et al. 2007), sub-Saharan Africa and Madagascar (Langrand 1990; Sinclair and Ryan 2003), Europe, North Africa and Central Asia (Heinzel et al. 1995; Grimmett et al. 1999; Arlott 2007, 2009; Brazil 2009), East and South-East Asia (MacKinnon and Phillips 1993; Coates and Bishop 1997; Robson 2005), and Oceania (Beehler et al. 1986; Pratt et al. 1987; Simpson and Day 2004; Robertson and Heather 2005). Where multiple subspecies were present, we collected information on the nominate subspecies, resulting in a sample size of adults of 8008 spp.. Juvenile plumages are less frequently drawn in field guides resulting in a sample size of 2603 spp. We scored the plumage of both sexes as well as juveniles of each study species (where applicable) for an absence of patterns, mottled, scaled, barred and spotted patterns on the ventral and dorsal surface separately (Fig. 5.2). Some species have multiple patterns on either the ventral and/or dorsal surface, e.g., the male zebra finch (Taeniopygia guttata) has barred patterns on the breast and spotted flanks. In such species, all of the different types of plumage patterns on each surface were scored and each pattern type was analysed separately. Where species exhibited variable patterns between molts we collected the breeding and non- breeding plumage given that there may be variation in selection pressure on the different types of patterns exhibited. Squares, triangles and stripes also occur within the feathers of birds, but are comparatively rare (e.g. stripes - 43 spp.), and were excluded from the analysis. Measuring phylogenetic signal The tendency of species to resemble related species, more than randomly selected species from a tree can be measured as phylogenetic signal (Pagel 1999). The best statistical approach to estimating phylogenetic signal is an area of ongoing debate but simulation studies indicate that with a large number of species Pagel’s λ is the most robust test statistic (Münkemüller, Lavergne, and Bzeznik 2012). To examine whether plumage patterns have evolved due to phylogenetic inertia we estimated Pagel’s λ in the package phylosig (Revell 2011) using 100 optimization The mechanisms underlying convergent evolution in the plumage patterns of birds 145 iterations in R (R Development Core Team 2012). In Phylosig, Pagel’s λ is estimated as 0 or 1, where a value of 0 indicates that trait evolution is independent of the phylogeny, whereas 1 indicates a signal of phylogenetic inertia. We estimated phylogenetic inertia in each pattern, in each sex and age class, as well as breeding and non-breeding plumages, on the ventral and dorsal surface, over 100 randomly selected phylogenetic trees (see Data Collection; Jetz et al. 2012). We present the proportion of trees out of 100 in which λ = 1. Ecological selection for plumage patterns To examine whether habitat selects for plumage patterns at the global scale, we analysed the distribution of avian species across the main habitat types worldwide. We used a global dataset of the distribution of bird species which is described elsewhere (Somveille et al. 2013). Briefly, polygons representing the global distribution of 9783 non-marine bird species were obtained from BirdLife International and NatureServe (2012). Breeding distributions (polygons corresponding to the areas where a species is present during the breeding season) were estimated separately from non-breeding distributions (polygons where a species is present during the non- breeding season). Terrestrial eco-regions are defined as units of land with a distinct assemblage of natural communities and species, and are nested within biomes, which provide identification of habitats (Olson et al. 2001). To investigate how avian species are distributed in eco-regions, we used a global map of terrestrial eco- regions made available by The Nature Conservancy (http://maps.tnc.org/gis_data.html). After removing eco-regions that contained no avian species, 810 eco-regions remained. A species was classed as occurring in a given eco-region if its mapped range overlaps with any part of the eco-region. Given the coarseness of the species distribution data, this is not always true, but represents a good approximation of occurrence given the spatial resolution of the eco-regions (Hurlbert and Jetz 2007). Species richness was measured as the number of species occurring in a given eco-region (Figure 5.3a). The mechanisms underlying convergent evolution in the plumage patterns of birds 146 Fig. 5.3. Total avian species richness worldwide as well as the global distribution of terrestrial habitats containing avian species. A) The global distribution of avian species richness. B) Each habitat type encompasses multiple eco-regions (numbers in brackets). Each type of habitat may be classified as either closed or open as indicated by colour: green and blue = mostly closed; red, orange and yellow = mostly open. Mangroves are indicated with grey, as they cannot be classified as either open or closed. T = Temperate; and T/S - Tropical and Subtropical. Across the world there are 16 major types of habitats as follows: 1) Tropical and subtropical moist broadleaf forests; 2) Tropical and subtropical dry broadleaf forests; 3) Tropical and subtropical coniferous forests; 4) Temperate broadleaf and mixed forests; 5) Temperate coniferous forests; 6) Boreal forests/taiga; 7) Mediterranean forests, woodlands and scrub; 8) Tropical and subtropical grasslands, Global habitats The mechanisms underlying convergent evolution in the plumage patterns of birds 147 savannas and shrublands; 9) Temperate grasslands, savannas and shrublands; 10) Flooded grasslands and savannas; 11) Montane grasslands and shrublands; 12) Tundra; 13) Deserts and xeric shrublands; 14) Mangroves; 15) Inland water; and 16) Rock and ice. Each of these habitats occurs in multiple eco-regions. Given that a major hypothesis of this study is that irregular mottled patterns are associated with closed habitats whereas regular patterns, in particular barred patterns, may occur in more open habitats, we additionally categorized habitat type into open and closed. Closed habitats comprise habitats 1-7 described above which are dominated by dense forests. Open habitats are comprised of the other habitats (8-16) that typically consist of short vegetation that is often sparsely distributed. Mangroves cannot be easily classified as either closed or open and were removed from the comparison of closed and open habitats. Each of the analyses was conducted for a subset of data comprised of each combination of patterns (all patterns, mottled, scaled, barred and spotted) and gender/age/body location class: male/female/juvenile, breeding/non-breeding, ventral/dorsal (herein subset). Each habitat occurs in multiple eco-regions (Fig. 5.3b). To determine whether habitat has an effect on the distribution of each subset of pattern data, we examined the geographical variation in the proportion of species that have a) any type of pattern, and b) each individual type of pattern. For each pattern data subset we compared the proportion of patterned species within each eco-region to the global proportion of patterned species. If the proportion within eco-regions was significantly above or below the global proportion this was interpreted as selection for or against patterning, respectively. However, if the proportion at the eco-region level was the same as the global level this indicated support for a uniform or random distribution of the subset of pattern data. The corresponding null expectations in our experimental approach are a) the eco-region has the same proportion of patterned species as the global proportion of patterned species, e.g. if half of the species of birds in the world are patterned then half of the species in the local eco-region will be patterned, b) the eco-region has the same proportion of species with a particular subset of pattern data as the global proportion of species with that pattern subset, e.g. if half of the patterned species of birds in the world have a mottled pattern then half of the species in the eco-region will be mottled. To test whether the observed proportion of patterned species was The mechanisms underlying convergent evolution in the plumage patterns of birds 148 significantly higher or lower than the null hypothesis per eco-region we used a hypergeometric probability distribution around the null value in each eco-region. We obtained a cumulative probability distribution of the observed proportion of patterned species as well as an associated p-value for the observed value. If the p-value is significant and the cumulative probability is close to 0, this indicates that patterns appear less than expected by chance, and if the p-value is significant and the cumulative probability is close to 1, this indicates that patterns appear more than expected by chance. We allowed for some flexibility in variation of the cumulative probability and where the value was either <=0.02 or >=0.98, we considered this range of values as significant for selection against or for patterns, respectively. To correct for multiple testing, which can lead to pseudo-replication, we corrected the p- values of the eco-regions of each habitat using a multiple testing correction. We initially used a standard p-value significance threshold of 0.05 with a Bonferroni correction, which is the p-value threshold divided by the number of eco-regions. This correction can be overly stringent. For example, for 810 eco-regions this is 0.05/810 = 0.000062. Accordingly, a Bonferroni correction resulted in no significant results in any of our analyses (data not shown). Due to potential issues of power, we used a Benjamini-Hochberg 
 correction per subset of patterns analysed, as it is the least stringent correction for multiple tests (Benjamini and Hochberg 2010). Our analysis depended on comparing the local proportion of patterned species within an eco-region with the global proportion of patterned species. In a small number of eco-regions, there were no species with any type of patterning and we could not calculate the proportions for our test statistic. However, the geographic distribution of eco-regions that have a complete absence of patterned species may be important. Therefore, we present the cases where there are no avian species with any kind of patterning separately. The mechanisms underlying convergent evolution in the plumage patterns of birds 149 Results Plumage pattern evolution due to shared ancestry All types of plumage patterns are found in all combinations of females, males and juveniles, for breeding and non-breeding plumage on both the ventral and dorsal surface (Table 5.2). Mottled plumage is the most prevalent type of pattern in birds and is found on the dorsal surface more frequently than on the ventral surface in all sex and age classes, and is least frequent in males. Barred patterns are also common and are frequently found on the ventral surface of breeding adult males, but also in juveniles. In comparison, scaled and spotted patterns are less prevalent and are frequently biased towards the dorsal surface of adults and the ventral surface of juveniles. However, spotted patterns are also frequently found on the ventral surface of breeding males and females. The mechanisms underlying convergent evolution in the plumage patterns of birds 150 Table 5.2. The presence of plumage patterns in the class Aves (adults: 8008 species, juveniles: 2603 species). Plumage pattern information was collected from field guides covering all major landmasses – see Methods for references. Mottled Barred Scaled Spotted Class Season Ventral Dorsal Both surfaces Ventral Dorsal Both surfaces Ventral Dorsal Both surfaces Ventral Dorsal Both surfaces Female Non-breeding 860 1081 1508 533 679 539 163 269 163 127 291 314 Female Breeding 864 1085 1514 500 624 978 162 202 321 300 157 427 Male Non-breeding 689 954 1288 530 876 535 153 277 162 122 385 314 Male Breeding 674 936 1261 514 478 975 161 156 320 305 104 427 Juvenile Juvenile 873 914 1368 886 325 627 274 158 203 385 223 157 The mechanisms underlying convergent evolution in the plumage patterns of birds 151 Table 5.3. The proportion of times that Lambda is estimated as 1 using 100 randomly selected phylogenetic trees for all types of plumage in 80% of avian species worldwide. In most cases, the proportion of times that λ = 1 was <0.3. Mottled Barred Scaled Spotted Class Season Ventral Dorsal Both surfaces Ventral Dorsal Both surfaces Ventral Dorsal Both surfaces Ventral Dorsal Both surfaces Female Non-breeding 0.28 0.26 0.28 0.25 0.33 0.20 0.28 0.29 0.27 0.22 0.22 0.20 Female Breeding 0.39 0.28 0.29 0.27 0.23 0.24 0.22 0.29 0.27 0.28 0.34 0.66 Male Non-breeding 0.32 0.25 0.21 0.26 0.26 0.18 0.29 0.20 0.28 0.25 0.22 0.17 Male Breeding 0.25 0.30 0.29 0.21 0.24 0.23 0.26 0.28 0.35 0.31 0.23 0.28 Juveniles Juveniles 0.23 0.27 0.25 0.27 0.27 0.32 0.28 0.21 0.22 0.24 0.28 0.26 The mechanisms underlying convergent evolution in the plumage patterns of birds 152 For most of the 60 combinations of plumage pattern, sex, gender and dorsoventral patterning, λ = 1 in less than 30/100 randomly selected phylogenetic trees. Seven out of the 60 combinations of pattern/sex/age/breeding class, λ = 1 in over 40/100 phylogenetic trees, and only one combination, female breeding plumage comprised of spots where both ventral and dorsal surfaces were combined, λ = 1 in over 50/100 (0.66; Table 5.3). Ecological selection for plumage patterns All four plumage patterns are found in all habitats. Each individual combination of pattern type and age/sex/breeding/dorsoventral subset occurs within each of the global 16 habitats regardless of whether they are closed or open (Fig. 5.3; supplementary Table S5.2). For example, mottled and barred patterns occur on the dorsal surface of juveniles and breeding females in open habitats such as Inland water as well as Rock and ice in addition to closed habitats such as Temperate broadleaf and mixed forests as well as Tropical and subtropical coniferous forests. In the analyses of an association between individual eco-regions and subset of plumage pattern data, the cumulative probability was 0 or 1 in 5.19% (414) and 53.02% (4229) of the 7976 individual analyses, respectively (eco-regions x subset of interest). 9.34% (745/7976) of the analyses had a significant association between plumage pattern subset and eco-region (where the cumulative probability was or close to 0 or 1 and the p-value was less than the adjusted significance threshold) (Table 5.4; supplementary Table S5.3 [CD-ROM]). Where there was a significant association with an eco-region, the results were dominated by a classification of all patterns together (7.65%) and an association with individual patterns was rare (1.23% + 0.01% + 0.39% + 0.06% = 1.69%). The mechanisms underlying convergent evolution in the plumage patterns of birds 153 Table 5.4. The patterns and number of eco-regions that is significantly different from the null expectation as well as the number of local patterned terrestrial species. All results are supplied in supplementary Table S5.3 [CD-ROM]. An association between all patterns together and eco-regions spans 15 out of the 16 habitats of the world across all age/sex/breeding classes (Table 5.5). In all age/sex/breeding classes, all plumage patterns appeared to be influenced by habitat on the ventral surface less than the dorsal surface. However, this significant association was predominantly in <25% of the eco-regions of each type of habitat. In only 4 habitats the number of eco-regions with a significant association with all patterns was >25%, and two of these habitats are comprised of only three eco- regions. Raw number of eco-regions that have significant results Proportion of eco- regions that have significant results % All patterns 610 7.65 Mottled 98 1.23 Scaled 1 0.01 Barred 31 0.39 Spotted 5 0.06 Total 745 9.34 The mechanisms underlying convergent evolution in the plumage patterns of birds 154 Table 5.5. The habitats and eco-regions that deviate from the null expectation with all patterns in land birds. The number of eco-regions is given in brackets per habitat as well as the maximum percentage of eco-regions within habitats that have a significant association per pattern data subset. T = Temperate; and T/S - Tropical and subtropical. Sex/Age Male Female Juvenile Season Breeding Non-breeding Breeding Non-breeding N/A Location Ventral Dorsal Ventral Dorsal Ventral Dorsal Ventral Dorsal Ventral Dorsal Total eco-regions 78   43   67   17   111   54   106   51   44   33   Boreal forest/taiga (22) <18% 4 - - - - 1 - 1 - - Desert and xeric shrubland (95) <5% - - 4 - 5 1 5 1 1 1 Flooded grassland and savanna (26) <12% 3 - 3 - 2 2 2 2 - - Inland water (3) <33% - - - - - - - - - 1 Mangrove (19) 2 - 1 - 4 1 4 - 1 - Mediterranean forest, woodland and scrub (39) <8% 3 - 1 - - - - - - - Montane grassland and shrubland (50) <22% 7 3 8 - 11 2 11 2 4 4 Rock and ice (3) <33% - - - - - - - - 1 - T broadleaf and mixed forests (88) <7% 6 3 5 1 4 3 4 2 4 3 Temperate conifer forest (48) <10% 5 4 4 2 3 2 3 2 3 1 T grassland, savanna and shrubland (48) <4% - - - - - 2 - 2 - 2 T/S coniferous forest (17) <6% 1 - 1 - 1 - 1 - - 1 T/S dry broadleaf forest (55) <16% 5 5 4 1 9 8 9 8 4 3 T/S grasslands, savannas and shrublands (48) <4% 1 - 1 - 8 3 8 3 4 2 The mechanisms underlying convergent evolution in the plumage patterns of birds 155 Ecological selection for mottled, scaled, barred or spotted patterns The individual patterns of mottled, scaled, barred and spotted rarely had a significant association with eco-regions (Table 5.4). In most cases, as the number of species increased, the number of eco-regions with a significant association decreased, e.g. where there is only one patterned species there are many eco- regions with a significant association, whereas few eco-regions had a significant association where the number of species >5 (Fig. 5.4). The rare cases of a significant association between patterns and eco-regions were not clustered in a particular plumage pattern subset and a particular habitat. For example, mottled breeding dorsal patterns in females occur more than expected by chance within the Amsterdam and Saint-Paul Islands temperate grasslands (cumulative p = 1, probability of cumulative p < 0.000063; supplementary Table S5.3 [CD-ROM]) that is one eco-region of the temperate grasslands, savannas and shrublands habitat. There are 39 other eco-regions within the temperate grasslands, savannas and shrublands habitat that have females with breeding dorsal mottled patterns and in none of these eco-regions does the species of this subset of data occur more (or less) than expected by chance. T/S moist broadleaf forests (230) <26% 41 28 35 13 64 29 59 28 22 15 Tundra (28) 0% - - - - - - - - - - The mechanisms underlying convergent evolution in the plumage patterns of birds 156 0 10 20 30 40 50 60 70 1 2 3 4 5 32 45 49 50 72 128 145 E co -r eg io ns Pattern species frequency Fig. 5.4. The number of eco-regions that deviate from the null expectation with either mottled, scaled, barred or spotted patterns plotted against the local number of patterned species in land birds. Across all permutations of sex/age/breeding/dorsoventral class per pattern, there were few significant associations and where there was a significant association, it was predominantly where the number of species <=5. A comparison of ecological selection between closed and open habitats Based on the prevailing view of camouflage and communication, the patterns that should demonstrate an association with closed and open habitats (Fig. 5.1, 5.3) are mottled and barred patterns (Gluckman and Cardoso 2010). The distribution of both mottled and barred patterns, on the dorsal and ventral surface of breeding and non-breeding females, as well as juveniles, is similar in the eco-regions of closed and open habitats and rarely differed from the null hypothesis (Fig. 5.5; supplementary Table S5.3 [CD-ROM]). Similarly, the proportion of species where males have patterns on the ventral surface, but also on the dorsal surface, in the breeding season was similar and did not differ from the null hypothesis in both closed and open habitats. The mechanisms underlying convergent evolution in the plumage patterns of birds 157 Fig. 5.5. Comparison of the proportion of species with mottled and barred patterns in breeding males and females, as well as juveniles, over the dorsal and ventral surface of land birds. Significant associations between mottled and barred patterns with eco- regions are rare (Table 5.4). The Patterned boxplots correspond to the proportion of patterned species in the pool of all avian species, the Mottled and Barred boxplots correspond to the proportion of these plumage patterns in the pool of patterned species for the indicated age/sex class. The boxplots in red correspond to closed forest habitat and the red boxplots correspond to open habitat. Mangroves (habitat 14; Fig. 5.3) do not easily fit into a closed or open classification and are not represented in this figure. Mottled and barred patterns rarely had significant associations with eco- regions (Table 5.4) and the patterns of association did not vary between open and closed habitats or subset of data. For example, the dorsal mottled patterns of breeding females had a significant association with the Amsterdam and Saint-Paul Islands temperate grasslands eco-region which is found in open Temperate grasslands, savannas and shrublands as well as the Bermuda subtropical conifer forests eco-region which is found in closed Tropical and subtropical coniferous forests (Fig. 5.3; supplementary Table S5.3 [CD-ROM]). In addition, the dorsal barred patterns of breeding females had a significant association with the Ascension scrub P ro p o rt io n o f p a tt e rn e d s p e c ie s D o r s a l V e n tr a l Male Female Juvenile Patterned Mottled Barred Patterned Mottled Barred Patterned Mottled Barred Closed habitat Open habitat The mechanisms underlying convergent evolution in the plumage patterns of birds 158 and grasslands eco-region which is found in open Tropical and subtropical grasslands, savannas and shrublands, as well as the San Félix-San Ambrosio Islands temperate forests eco-region which is found in the closed Temperate broadleaf and mixed forests habitat. Finally, in 6 out of the 16 habitats worldwide there are 17 eco-regions that do not have avian species with any type of pattern (Fig. 5.3; Table 5.6). However, four of these habitat types have <=5 total species. In the other two habitats, few eco-regions had significant associations: Temperate and subtropical moist broadleaf - 6 eco- regions out of 230 (2.6%); Tundra - 2 eco-regions out of 28 (7.1%). The mechanisms underlying convergent evolution in the plumage patterns of birds 159 Table 5.6. Eco-regions that do not have any avian species with any type of pattern in land birds. T = Temperate; and T/S - Tropical and subtropical. The number of eco-regions per habitat type is given brackets. Habitat Eco-region Avian species richness T grasslands, savannas and shrublands (40) Amsterdam and Saint-Paul islands temperate grasslands 1 Ascension scrub and grasslands 3 T/S coniferous forests (17) Bermuda subtropical conifer forests 2 T/S grasslands, savannas and shrublands (48) St. Helena scrub and woodlands 1 T/S moist broadleaf forests (230) Central Polynesian tropical moist forests 6 Fernando De Noronha-Atol das Rocas moist forests 5 Marquesas tropical moist forests 16 Rapa Nui subtropical broadleaf forests 2 Society Islands tropical moist forests 17 Trindade-Martin Vaz Islands tropical forests 1 Tuamotu tropical moist forests 18 Tubuai tropical moist forests 8 Western Polynesian tropical moist forests 8 Tundra (28) Marielandia Antarctic tundra 5 Maudlandia Antarctic desert 2 Scotia Sea Islands tundra 8 Southern Indian Ocean Islands tundra 6 The mechanisms underlying convergent evolution in the plumage patterns of birds 160 Discussion All types of plumage patterns have evolved in juveniles, females and males in non-breeding and breeding plumage, on both the ventral and dorsal surface (Table 5.1). In the majority of pattern types, shared ancestry only explained <30% of plumage pattern evolution (Table 5.3) indicating that convergent evolution in bird plumage patterns is unlikely to be predominantly due to phylogenetic relatedness. All types of patterns, in all age, breeding, sex and dorsoventral locations are found in all habitats (supplementary Table S5.2). Contrary to the existing hypotheses of the function of patterns in camouflage, the habitats in which patterns are viewed did not predict convergence in bird plumage patterns across either age or sex classes, breeding and non-breeding plumages, or the dorsoventral distribution of patterns (Fig. 5.4, 5.5; Table 5.4, 5.5). Congruent with the hypotheses of the function of plumage patterns in communication, habitat type did not predict the patterns found on the ventral surface of breeding males. However, the independence between plumage patterns and habitat was also found on the dorsal plumage of breeding males, as well as their non-breeding plumage. The rare significant associations between eco- regions and irregular and regular patterns were found in both closed and open habitats, irrespective of age/sex/breeding/dorsoventral class. In addition, there were few eco-regions where there are no patterned species and these were not clustered within a particular type of habitat (Table 5.6). Together this indicates that avian plumages appear to have converged on the same types of patterns independent of habitat (Fig. 5.5; Table 5.3). Our analyses indicate that phylogenetic inertia has contributed <30% of plumage pattern evolution in most cases. Although this represents a minority of the overall number of trees tested, this result was similar across all pattern data subsets indicating that there was variation in the estimation of lambda across trees (Table 5.3). That the variation was relatively similar between pattern subsets indicates that the source of this variation may be due to adaptive radiation in a specific part of the avian tree of life (Jetz et al. 2012). However, the results from the large majority of phylogenetic trees in this analysis indicate phylogenetic uncertainty is unlikely to be the cause of convergent evolution in bird plumage patterns. The only subset of data where there appeared to be phylogenetic inertia, female breeding spotted patterns over the whole body, may not be representative given that in the separate analyses The mechanisms underlying convergent evolution in the plumage patterns of birds 161 of ventral and dorsal female breeding spotted patterns λ = 1 in 28/100 and 34/100 trees, respectively. The 100 phylogenetic trees in this analysis were randomly chosen to control for bias as well as represent the uncertainty in a current understanding of avian phylogenetic relatedness. Also, the package Phylosig used to compute Pagel’s lambda is a binary statistic rather than continuous which may oversimplify the results, but given a limitation in the computational power required to run this analysis in other software packages such as BayesTraits, this approach represents the best available option. Adaptive radiation may cause an erroneous inflation of the test statistic. Conducting analyses at the genus level can control for erroneously inflated results, but would require an adequate sample size of genera within eco-regions to avoid losing statistical power. The species level results, combined with the number of species present within eco-regions that had a significant result, indicated that a genus level analysis was not required (Fig. 5.4; Table 5.4). Most of the significant results for individual patterns was where there were <=5 species within an eco- region. This indicates that there may have been inflation of the test statistic due to a low number of species in these cases and may not be representative (Table 5.6). Selection for any type of patterning was not clustered within eco-regions (Table 5.5). In addition, there did not appear to be an association between age, gender, breeding class or dorsoventral patterning with any type of habitat (Fig. 5.5; supplementary Table S5.3 [CD-ROM]). All of the different types of patterns were rarely found more or less than expected by chance on the ventral and dorsal surface of breeding and non-breeding females, as well as juveniles, per eco-region and per habitat (Table 5.4, 5.5; supplementary Table S5.3 [CD-ROM]). Similarly, all plumage patterns were found on the ventral and dorsal surface of the breeding and non- breeding plumages of males (p = > Benjamini-Hochberg corrected threshold; supplementary Table S5.3 [CD-ROM]). This could be due to low statistical power in our analyses. However, we used the least stringent correction possible to increase statistical sensitivity to ecological signal. In addition, regardless of the ecological techniques used, all of the different types of patterns, across all age, gender, breeding and dorsoventral classes are found in every type of habitat around the world, which is incongruent with camouflage-based predictions (Table 5.1; supplementary Table S5.2). Our analyses did not incorporate the extent to which a The mechanisms underlying convergent evolution in the plumage patterns of birds 162 pattern(s) may contribute to the overall phenotype of each avian species and as a result this analysis is necessarily coarse. However, this would not result in the prevalence of all pattern types in all habitats irrespective of age, gender, breeding/non-breeding and dorsal/ventral plumage. We could not statistically calculate whether a complete absence of any type of patterning might be significantly associated with habitats (Table 5.6). However, the distribution of the number of eco-regions without any species with patterning was not biased towards a particular type of habitat. For example, Tropical and subtropical moist broadleaf forests are a closed type of habitat and Tundra is an open type of habitat and each of these habitats have the most number of eco-regions that have species with no patterns (Fig. 5.3; Table 5.6). In addition, both of these habitats frequently have patterned species in other eco-regions (supplementary Table S5.2, S5.3 [CD-ROM]). Therefore, these rare cases where patterns are absent are unlikely to be representative of selection against patterns. If patterns evolved strictly for camouflage and/or signaling based on the way that the pattern is viewed in the context of its habitat, then it would be expected that patterns would be biased towards a particular gender and body surface. Under this expectation regular patterns in males should be independent of habitat on the ventral surface, but not necessarily on the dorsal surface, and patterning should have been biased towards the dorsal surface of females and juveniles, but were not (Fig. 5.1, 5.5; supplementary Table S5.2, S5.3). Comparing the mottled and barred patterns of breeding males and females as well as juveniles in closed and open habitats, there was little difference in their distribution between age and sex classes or the dorsal and ventral surface (Fig. 5.5). If patterns indeed make an animal stand out against its background then there should have been strong selection against regular patterns on the dorsal surface of females and juveniles in closed habitats, which was not the case. A lack of association in juvenile plumage patterns could be due to a low sample size (2603 spp.) but this would have resulted in either an absence of some types of plumage patterns, or an erroneous significance value of selection for or against some types of patterns in some habitats, which was not found (Table 5.2; supplementary Table S5.3 [CD-ROM]). Based on previous research and hypotheses, irregular patterns would be The mechanisms underlying convergent evolution in the plumage patterns of birds 163 expected in cluttered environments and regular patterns would be expected in open environments (Table 5.1, supplementary Table S5.1). However, irregular patterns were found throughout all types of habitats, including less noisy open habitats where it would be expected that they would stand out, e.g. deserts (Fig. 5.1; supplementary Table S5.2, S5.3). Moreover, regular patterns were also found throughout all types of habitats, including more noisy and closed habitats where it would be difficult to invoke motion-dazzle/flicker-fusion camouflage through movement. Breeding females with regular patterns, which would be constrained to take care of nests, as well as juveniles, were also found in closed habitats, which is contrary to motion- dazzle/flicker-fusion camouflage which is dependent on movement (Fig. 5.5). Alternatively, perhaps similar to reef fish, regular patterns might become blurred at a distance and match the background (Marshall 2000). A lack of association of the ventral surface of males with regular patterns, in particular barred patterns (Gluckman and Cardoso 2010), and habitats would be expected under sexual selection. However, visual signals should diverge in biological hotspots due to species recognition/sexual selection (Bradbury and Vehrencamp 1998). Habitat may influence whether the pattern stands out against its background, for example by standing out against a uniform background or opposing the geometric pattern of the background (Fig. 5.1). Other factors, such as whether the pattern is able to convey aspects of individual quality, would influence their evolution independent of the viewing background. This perhaps may explain why barred patterns have repeatedly evolved independently on the ventral surface of males (Gluckman and Cardoso 2010) (Fig. 5.5). Conceivably, the same forces of sexual selection/species recognition may have shaped the evolution of the other types of patterns in males, such as spotted patterns (Roulin et al. 2000; Roulin, Riols, and Dijkstra 2001; Petrie and Halliday 2008; Muck and Goymann 2011; Pérez-Rodríguez, Jovani, and Mougeot 2013). Perhaps the use of different strata in the vegetation, e.g. ground-dwelling vs arboreal, may present different visual microhabitat backgrounds that alter the selection for pattern type. For example, the tops of trees are more open and may allow for more movement required for motion-dazzle/flicker-fusion regular patterns whereas ground-dwelling species may benefit from background matching irregular patterning. However, it is difficult to reconcile this idea with the distribution of all The mechanisms underlying convergent evolution in the plumage patterns of birds 164 pattern types, in all age/sex/breeding/dorsoventral, in habitats that do not have strata, e.g. desert and tundra. The role of specific behaviours, such as whether a species is gregarious, combined with habitat use, may explain our results. Perhaps there is less selection for camouflage patterns in gregarious species because they share social information to evade predation rather than rely on visual patterning. For example, Artiodactyls that are solitary are significantly associated with dense forest habitats (Caro 2005). However, information on gregariousness, breeding habits as well as specific anti-predator behaviours are lacking in many avian species of the world and could not be controlled for in this analysis. Data on avian anti-predator and social behavior as well as major predators per species would be an invaluable contribution to the literature. Caro (2005) reports that although that there is a camouflage function in the patterns of many artiodactyl species, many studies did not appear to demonstrate an association with habitat. The reality is that we do not know how many studies have shown a negative result but remain unpublished. Perhaps a coarse granular approach to ecological selection for plumage patterns may explain our results. For example, many of the empirical studies that demonstrate a camouflage function in patterns conducted their analyses at short viewing distances and demonstrated an association at the scale of microhabitats (Table 5.1; supplementary Table S5.1). Therefore, perhaps an analysis at the habitat and eco-region scale contains too much variation in microhabitats. However, raptors, which are the main predators of birds, have excellent long range vision in many different types of habitats and it would seem unlikely that they would need short viewing distances to act as a selection pressure on the camouflage function of patterns (Tucker 2000; Christie and Ferguson-Lees 2010; O'Rourke et al. 2010). An alternative explanation is that microevolution is too labile to be captured at a macroevolutionary scale, unless the adaptation of the pattern has a long evolutionary history. Perhaps a better approach would be to consider patterning in the context of a coevolutionary process between signaler and receiver (Thompson and Cunningham 2002; Thompson, Nuismer, and Gomulkiewicz 2002; Thompson 2005). In particular, an alternative hypothesis is that perhaps four patterns are enough to provide a camouflage benefit to a community of species. The previous empirical studies of camouflage predominantly examined the interaction between the The mechanisms underlying convergent evolution in the plumage patterns of birds 165 visual form of a species in the context of one or two background types. However, in reality species live in communities of other species, regardless of whether they are gregarious or social. Recently it was demonstrated that there may be frequency- dependent selection on alternative morphs of Tetrix subulata grasshoppers, where each morph has varying amounts of mottled patterns in varying colours. In this study it was demonstrated that when all morphs are present in a population, all morphs benefit by providing variation in the search image of predators (Karpestam, Merilaita, and Forsman 2014). In the case of birds, perhaps the four different types of patterns, in the context of a community of avian species that may also comprise uniform coloration without patterns, may benefit all member species. Although intraspecific polymorphism is quite different to interspecific variation, perhaps intraspecific polymorphism may be the first step to speciation leading to interspecific variation. There are potentially three major reasons that may predict that the number of patterns increases with community size, assuming that phenotypes are mostly for crypsis: 1) More patterns reduces predation by increasing the number of predator search images, which may also favour polymorphisms within species, 2) niche differentiation via dietary specialisms and/or microhabitat use may lead to variation in the background in which predators view prey resulting in selection for different types of patterns, which may also have implications for polymorphisms, and 3) if some parts of the pattern phenotype are used in species recognition then there will be selection on patterns to diverge to facilitate recognition. In addition, it has been found that sexual dimorphism/monomorphism in avian phenotypes is not always indicative of breeding systems (Gluckman 2014) and perhaps these processes may also favour sexual dimorphism to reduce the chance of predator search images and/or facilitate gender recognition. Alternatively, all of these selection pressures may be operating on avian phenotypes simultaneously and this will be investigated in a future study. It may also be the case that developmental constraint may explain repeated convergence of the four types of plumage patterns independent of function. For example, based on a theoretical model of reaction-diffusion based plumage pattern formation (Prum and Williamson 2002), the evolutionary trajectory of plumage in Anseriformes and Galliformes followed the same trajectories within and between plumage patches of plumage over the body (see Chapter 4). Given that Anseriformes and Galliformes are diverse in life history attributes, this result warrants further The mechanisms underlying convergent evolution in the plumage patterns of birds 166 investigation at the macroevolutionary scale. In addition, it has been demonstrated that some plumage patterns shift along a trajectory correlated with body size in most avian species of the world (Riegner 2008). However, these results should be treated with caution, as the analysis did not control for phylogeny. Nevertheless, these studies indicate that there may be developmental constraints in plumage pattern evolution that may have implications for natural selection. A future study will investigate whether allometric relationships may predict plumage pattern evolution in a controlled comparative framework. Bradbury and Vehrencamp (1998) note that in regards to strategies for maximizing contrast in visual communication that “A serious problem… is that they are based on human perception”. Similarly, the predominant paradigm of the function of patterns in camouflage stem from a time when there were few, if any, objective approaches for quantifying patterning. Comparing the frequency of patterns per habitat type, we did not find that habitat predicts the presence of plumage patterns. The fact that all patterns are found in all habitats around the world is an intriguing result that implies that plumage pattern evolution does not conform to the prevailing views of selection for camouflage. The mechanisms underlying convergent evolution in the plumage patterns of birds 167 The mechanisms underlying convergent evolution in the plumage patterns of birds 168 Chapter 6: Discussion and Conclusion The mechanisms underlying convergent evolution in the plumage patterns of birds 169 The mechanisms underlying convergent evolution in the plumage patterns of birds 170 Concluding thoughts In my concluding chapter I bring together the findings of each chapter and highlight possible avenues for future research. My work on the development of pigmentation in embryonic Japanese quail has implications for evolutionary genetics and evolutionary-development. The finding that activation of MC1R is important in avian melanin based plumage colouration demonstrates that there is variation in pigmentation dynamics between mice and birds. This finding highlights the importance of considering evolution, in addition to conservation, of the mechanism of pigmentation between vertebrates and birds. Activation of MC1R via POMC and PC2 is a promising area of research in avian pigmentation dynamics and raises the possibility of directly investigating activation of MC1R in the developmental basis of plumage patterns in birds. Reaction-diffusion dynamics appear to have played a role in the evolution of within-feather patterning in Anseriformes and Galliformes and spots appear to be a derived trait. Given that patterns are known to have both a signaling and camouflage function, it may be that reaction-diffusion based systems have influenced the evolutionary developmental biology of trait diversity as well as speciation. Comparing the evolution and diversity of barred plumage patterns in parasitic cuckoos demonstrates that plumage patterns can be additionally refined for a specific signaling function. This underscores the importance of considering the visual perspective of the receiver as well as objective methods with which to analyse visual patterns. Although patterns are known to function in camouflage and communication, it is interesting rather than disappointing that ecology did not appear to predict plumage patterns. Perhaps this points to the reality that animals, regardless of whether they are social or not, live in a community of other animals which presents an exciting opportunity to re-examine what camouflage means, and also, whether evolutionary-development has had a further hand in plumage pattern evolution via allometric constraints. The mechanisms underlying convergent evolution in the plumage patterns of birds 171 The mechanisms underlying convergent evolution in the plumage patterns of birds 172 Developmental explanations Japanese quail is a model system in understanding developmental mechanisms. An interesting feature of the developmental biology of pigmentation in quail is the variation in the patterns of expression between embryos and adults. This implies that there may be evolution in the developmental basis of phenotypes between different developmental stages within a species. Future studies could examine whether the novel ASIP alternatively spliced transcripts discovered here are expressed in adult quail, and whether similar patterns of variation in alternatively spliced transcripts might be found in other avian species. It would seem plausible that there may be variation between embryonic and adult stages given the different types of plumage involved (neoptile vs pennaceous) and variation in selection pressure at different life stages, e.g. juvenile and adult. The finding that ASIP does not appear to be involved in dorsal patterns within-feathers demonstrates that there has been evolution in the developmental basis of phenotypes between mice and birds and highlights the importance of considering divergence in developmental mechanisms as well as conservation. Although the mechanism down-regulating dorsal within-feather and between-stripe pigmentation patterns in wildtype quail is currently unknown, there are several avenues of research (Walker and Gunn 2009; Walker and Gunn 2010). In particular, Corin affords an exciting prospect to examine divergence and conservation in the developmental basis of phenotypic evolution given that this loci, as well as ASIP and MC1R, explain variation in phenotypic evolution of two populations of Peromyscus mice (Manceau et al. 2010). This will be explored in the coming year. The major finding of the developmental basis of embryonic wildtype quail pigmentation is that activation of MC1R appears to be required for eumelanin pigmentation. This is in contrast with previous findings because existing research on the developmental basis of plumage patterns has been on understanding developmental similarity based on shared evolutionary history between mice and birds (Nadeau et al. 2008; Oribe et al. 2012; Yoshihara et al. 2012) despite hints that POMC and PC2 have a function in avian pigmentation (Yoshihara et al. 2011). However, there were dissimilarities between wildtype quail and chicken in that PC1 was only expressed in feather follicles of chicken breeds. This may be indicative of The mechanisms underlying convergent evolution in the plumage patterns of birds 173 the effects of artificial selection rather than natural selection but PC1 should nevertheless be tested alongside POMC and PC2 on other wildtype avian phenotypes such as the northern bobwhite (Colinus virginianus) and the common quail (Coturnix japonica), and will be presented in a future study. There does however appear to be confusion in the literature as to the possibility that PC2 can directly cleave POMC (e.g. Takeuchi et al. 2003; Yoshihara et al. 2011), even though it has been previously reported that PC2 directly cleaves POMC in mice (Benjannet et al. 1991). An important future step will be to understand the biochemistry of POMC, PC1 and PC2 products in birds. In addition, 5’ RACE of POMC transcripts may reveal promoter sites that are dorsoventral specific, and perhaps may be temporal-specific. Similarly perhaps there are PC2 alternative splice transcripts that may also be dorsoventral/temporal specific and this will also be explored in future studies. These results have opened several new avenues of research into the rich tapestry of the evolutionary developmental biology of avian phenotypes. However, the organizing mechanism of within-feather patterning is currently unknown. A 9-day old White Leghorn chick showing barred plumage on the right wing, produced by grafting limb-bud mesoderm from a Barred Plymouth Rock embryo (Willier and Rawles 1938) The mechanisms underlying convergent evolution in the plumage patterns of birds 174 Developmental constraint Although it is unknown whether reaction-diffusion based mechanisms govern within-feather pattern formation, the significant similarity of the trajectory of plumage pattern evolution in two avian groups that have different life history attributes is predictable from a hypothetical model of pattern formation (Prum and Williamson 2002). This affirms the power of biologically relevant models. Coupling biological modeling with sophisticated comparative modeling techniques demonstrates the possibility to evaluate the relevance of hypothetical models of developmental dynamics before the time consuming task of confirming that the model itself represents reality. Of course, this is fraught with potential false positives/negatives, but given that biological modeling as well as comparative modeling can be undertaken in a fraction of the time required to test molecular and developmental biology hypotheses in live systems, this solution provides a promising preliminary step in examining the potential of developmental dynamics. The similarity in the trajectory of plumage pattern evolution within and between patches in two different groups is remarkable and lends weight to “the following role for natural selection: fine-tuning the appearance of the pattern, fixing and maintaining pattern elements at a given level of expression, and modifying behavioral and other features to maximize the patterns’ utility” (Price and Pavelka 1996). In spite of the strong directionality in plumage pattern evolution in Anseriformes and Galliformes, there were frequent and strong evolutionary transitions to other types of patterns, similar to other studies (Gluckman 2014). It would therefore seem that once plumage patterns have evolved, patterns are a labile trait that can adapt quickly to environmental changes. This may have multiple important evolutionary consequences: i) if patterning has not evolved in an ancestral lineage then development may constrain trait evolution, ii) if patterning has evolved in an ancestral lineage of a species then it may be easier to adapt to changes in the environment, iii) together this implies that evolutionary-development may constrain or facilitate speciation. However, it is currently unknown whether this same evolutionary trajectory is the same in the most speciose part of the avian tree – Passeriformes. Additionally or alternatively, The mechanisms underlying convergent evolution in the plumage patterns of birds 175 allometric constraints may play a role in plumage pattern evolution (Riegner 2008) which may add an additional layer to mechanisms of plumage pattern speciation and evolution. A study examining whether plumage pattern evolution follows a similar evolutionary trajectory in all birds worldwide, whether evo-devo constrains or facilitates speciation, as well as whether there is an allometric relationship in avian pattern evolution (accounting for phylogeny and phylogenetic uncertainty) will be undertaken next year. Adaptive explanations Signal evolution in the barred plumage patterns of Old World parasitic cuckoos is apparent at both the macroevolutionary and microevolutionary scale (Kruger, Davies, and Sorenson 2007; Gluckman and Mundy 2013). In addition, parasitic cuckoos could mimic one or more models from multiple genera of raptors. This remarkable precision in sympatric similarity demonstrates that although it is possible that there may be developmental constraint in plumage pattern evolution, natural selection has adaptively modified this type of patterning to be parasite- specific. This has important implications for our understanding of signal evolution as a principle as it demonstrates that signals can be refined to be receiver specific, within a given type of signal. This underscores the importance of using objective based methods to examine visual traits. The common cuckoo (Cuculus canorus) and the Eurasian sparrowhawk (Accipiter nisus) The mechanisms underlying convergent evolution in the plumage patterns of birds 176 The study of cuckoo-hawk mimicry was representative of five genera of parasitic cuckoos but there are many other barred parasitic cuckoos as well as barred non-parasitic species (Payne and Sorensen 2005). An interesting future study would be to investigate whether these variations in barred plumage patterns are indeed specific to parasitic cuckoos compared to non-parasitic species. Barred plumage patterns can additionally function in communication (Swaddle and Cuthill 1994; Bortolotti et al. 2006; Gluckman and Cardoso 2010) and melanin based traits have pleiotropic effects in the Tawny owl (Roulin 2004; Ducrest, Keller, and Roulin 2008; Emaresi et al. 2013). It would be interesting to examine whether barred plumage patterns can additionally function in communication in parasitic species, as well as non-parasitic species. From a broader perspective, the prevailing view of the camouflage function of within-feather patterns did not predict where plumage patterns have evolved from a habitat perspective. In recent years there has been an increase in studies of the camouflage function of visual patterns in animals. It is an exciting finding that all pattern types are found in all habitat types across the world. This is congruent with the idea that patterns may fill visual niches, similar to ecological niches. To study this question would require examining all aspects of phenotypic diversity, including uniform colouration, in relation to habitat. In addition, as demonstrated in parasitic cuckoos, the same pattern may be refined in multiple ways, which may also add an additional layer of complexity. Therefore, such a study would require digital image analysis of all avian plumages worldwide, rather than scoring images from field guides. This type of question is timely given that the collection of avian plumages in the class Aves worldwide has begun using digital image analysis techniques, and is the part of a macroevolutionary project that I have begun managing. The mechanisms underlying convergent evolution in the plumage patterns of birds 177 The mechanisms underlying convergent evolution in the plumage patterns of birds 178 References Aberdam E, Bertolotto C, Sviderskaya EV, de Thillot V, Hemesath TJ, Fisher DE, Bennett DC, Ortonne JP, Ballotti R. 1998. Involvement of Microphthalmia in the Inhibition of Melanocyte Lineage Differentiation and of Melanogenesis by Agouti Signal Protein. Journal of Biological Chemistry 273:19560–19565. Ainsworth SJ, Stanley RL, Evans DJR. 2010. Developmental stages of the Japanese quail. Journal of Anatomy 216:3–15. Allen JJ, Mäthger LM, Barbosa A, Buresch KC, Sogin E, Schwartz J, Chubb C, Hanlon RT. 2010. Cuttlefish dynamic camouflage: responses to substrate choice and integration of multiple visual cues. Proceedings of the Royal Society B: Biological Sciences 277:1031–1039. Allen WL, Baddeley R, Scott-Samuel NE, Cuthill IC. 2013. The evolution and function of pattern diversity in snakes. Behavioral ecology 24:1237–1250. Alvarez F. 1993. Proximity of trees facilitates parasitism by Cuckoos Cuculus canorus on Rufous Warblers Cercotrichas galactotes. Ibis 331. Anderson DR, Link WA, Johnson DH, Burnham KP. 2001. Suggestions for Presenting the Results of Data Analyses. The Journal of Wildlife Management 65:373–378. Andersson MB. 1994. Sexual Selection. Princeton University Press. Andersson S, Prager M, Johansson E. 2007. Carotenoid content and reflectance of yellow and red nuptial plumages in widowbirds (Euplectes spp.). Functional Ecology 21:272–281. Anistoroaei R, Christensen K. 2007. Mapping of the silver gene in mink and its association with the dilution gene in dog. Cytogenet. Genome Research 116:316– 318. April CS, Jackson IJ, Kidson SH. 1998. Molecular cloning and sequence analysis of a chicken cDNA encoding tyrosinase-related protein-2/DOPAchrome tautomerase. Gene 219:45–53. The mechanisms underlying convergent evolution in the plumage patterns of birds 179 Arendt J, Reznick D. 2008. Convergence and parallelism reconsidered: what have we learned about the genetics of adaptation? Trends in Ecology and Evolution 23:26–32. Arlott N (2007) Birds of Europe, Russia, China, and Japan, Passerines. Princeton University Press, Princeton. Badyaev AV, Hill GE, Dunn PO, Glen JC. 2001. Plumage colour as a composite trait - developmental and fuctional integration of sexual ornamentation. American Naturalist 158:221–235. Badyaev AV, Landeen EA. 2007. Developmental evolution of sexual ornamentation: model and a test of feather growth and pigmentation. Integrative and Comparative Biology 47:221–233. Barbosa A, Mäthger LM, Buresch KC, Kelly J, Chubb C, Chiao C-C, Hanlon RT. 2008. Cuttlefish camouflage: The effects of substrate contrast and size in evoking uniform, mottle or disruptive body patterns. Vision Research 48:1242–1253. Barsh GS. 1996. The genetics of pigmentation: from fancy genes to complex traits. Trends in Genetics 12:299–305. Bates HW. 1862. Contributions to an insect fauna of the Amazon valley (Lepidoptera: Heliconidae). Transactions of the Linnean Society of London 23:495–56–. Batt BD. 1992. The Ecology and Management of Breeding Waterfowl. University of Minnessota Press. Bed’Hom B, Mohsen Vaez, Coville J-L, Gourichon D, Chastel O, Follett S, Burke T, Minvielle F. 2012. The lavender plumage colour in Japanese quail is associated with a complex mutation in the region of MLPH that is related to differences in growth, feed consumption and body temperature. BMC Genomics:442. Beebe W. 1922. A Monograph of the Pheasants. Dover Publications. Beehler BM, Pratt TK, Zimmerman DA (1986) Birds of New Guinea. Princeton University Press, Princeton The mechanisms underlying convergent evolution in the plumage patterns of birds 180 Begum S, Moksnes A, Røskaft E, Stokke BG. 2011. Factors influencing host nest use by the brood parasitic Asian Koel (Eudynamys scolopacea). Journal of Ornithology 152:793–800. Benjamini Y, Hochberg Y. 2010. Controlling the False Discovery Rate: A Practical and Powerful Approach to Multiple Testing. Journal of the Royal Statistical Society. Series B (Methodological) 57:289–300. Benjannet S, Rondeau N, Day R, Chretien M, Seidah NG. 1991. PCI and PC2 are proprotein convertases capable of cleaving proopiomelanocortin at distinct pairs of basic residues. Proceedings of the National Academy of Sciencess USA 88:3564– 3568. Birdlife International, NatureServe (2012) Bird species distribution maps of the world. BirdLife International, Cambridge, United Kingdom and NatureServe, Arlington, United States. Available: http://www.birdlife.org/datazone/info/spcdownload. Accessed 4th Jul 2013. Bilodeau ML, Greulich JD, Hullinger RL, Bertolotto C, Ballotti R, Andrisani OM. 2001. BMP-2 stimulates tyrosinase gene expression and melanogenesis in differentiated melanocytes. Pigment Cell Research 14:328–336. Birnboim HC, Doly J. 1979. A rapid alkaline extraction procedure for screening recombinant plasmid DNA. Nucleic Acids Research 7:1513–1523. Björklund M. 1991. Coming of age in fringillid birds: heterochrony in the ontogeny of secondary sexual characters. Journal of Evolutionary Biology Bleiweiss R. 2007. On the ecological basis of interspecific homoplasy in carotenoid- bearing signals. Evolution 61:2861–2878. Bortolotti GR, Blas J, Negro JJ, Tella JL. 2006. A complex plumage pattern as an honest social signal. Animal behaviour 72:423–430. Boswell T. 2005. Regulation of energy balance in birds by the neuroendocrine hypothalamus. The Journal of Poultry Science 42:161–181. Boswell T, Li Q, Takeuchi S. 2002. Neurons expressing neuropeptide Y mRNA in the infundibular hypothalamus of Japanese quail are activated by fasting and co-express The mechanisms underlying convergent evolution in the plumage patterns of birds 181 agouti-related protein mRNA. Molecular Brain Research. 100:31–42. Boswell Timothy, Takeuchi S. 2005. Recent developments in our understanding of the avian melanocortin system: Its involvement in the regulation of pigmentation and energy homeostasis. Peptides 26:1733–1743. Bradbury JW, Vehrencamp SL. 1998. Principles of Animal Communication. Sunderland MA: Sinauer Associates. Brakefield PM. 2006. Evo-devo and constraints on selection. Trends in Ecology and Evolution 21:362–368. Brazil M (2009) Birds of East Asia, China, Taiwan, Korea, Japan, and Russia. Princeton University Press, Princeton. Briskie JV, Sealy SG, Hobson KA. 1992. Behavioral Defenses against Avian Brood Parasitism in Sympatric and Allopatric Host Populations. Evolution 46:334–340. Brodie ED. 1989. Genetic correlations between morphology and antipredator behaviour in natural populations of the garter snake Thamnophis ordinoides. Nature 342:542–543. Brodie ED. 1992. Correlational Selection for Color Pattern and Antipredator Behavior in the Garter Snake Thamnophis ordinoides. Evolution 46:1284–1298. Brodie ED. 1993. Differential Avoidance of Coral Snake Banded Patterns by Free- Ranging Avian Predators in Costa Rica. Evolution 47:227–235. Brooke M de L, Davies NB. 1998. Egg mimicry by cuckoos Cuculus canorus in relation to discrimination by hosts. Nature 335:1–3. Bultman SJ, Michaud EJ, Woychik RP. 1992. Molecular characterization of the mouse agouti locus. Cell 71:1195–1204. Burnham KP, Anderson DR. 2002. Model Selection and Multi-Model Inference. New York: Springer. Candille SI, Kaelin CB, Cattanach BM, Yu B, Thompson DA, Nix MA, Kerns JA, Schmutz SM, Millhauser GL, Barsh GS. 2007. A Beta-Defensin Mutation Causes The mechanisms underlying convergent evolution in the plumage patterns of birds 182 Black Coat Color in Domestic Dogs. Science 318:1418–1423. Cardoso GC, Mota P. 2010. Evolution of female carotenoid coloration by sexual constraint in Carduelis finches. BMC Evolutionary Biology 10:82. Cardoso GC, Mota PG. 2008. Speciational evolution of coloration in the genus Carduelis. Evolution 62:753–762. Caro T. 2005. The Adaptive Significance of Coloration in Mammals. BioScience 55:125–136. Chang C-H, Jiang TX, Lin C-M, Burrus LW, Chuong C-M, Widelitz R. 2004. Distinct Wnt members regulate the hierarchical morphogenesis of skin regions (spinal tract) and individual feathers. Mechanisms of Development 121:157–171. Chang C-M, Coville J-L, Coquerelle G, Gourichon D, Oulmouden A, Tixier-Boichard M. 2006. Complete association between a retroviral insertion in the tyrosinase gene and the recessive white mutation in chickens. BMC Genomics 7:19. Chen CW, Chuong CM. 2000. Dynamic expression of lunatic fringe during feather morphogenesis: a switch from medial-lateral to anterior-posterior asymmetry. Mechanisms of Development 91:351–354. Chen Y, Zhang Y, Jiang T-X, Barlow AJ, Amand TRS, Hu Y, Heaney S, Francis-West P, Chuong C-M, Maas R. 2000. Conservation of Early Odontogenic Signaling Pathways in Aves. Proceedings of the National Academy of Sciences U.S.A. 97:10044–10049. Cheng Y-C, Cheung M, Abu-Elmagd MM, Alex Orme PJS. 2000. Chick Sox10, a transcription factor expressed in both early neuralcrest cells and central nervous system. Developmental Brain Research 121:233–241. Chiao C-C, Chubb C, Buresch K, Siemann L, Hanlon RT. 2009. The scaling effects of substrate texture on camouflage patterning in cuttlefish. Vision Research 49:1647– 1656. Chiao C-C, Chubb C, Buresch KC, Barbosa A, Allen JJ, Mäthger LM, Hanlon RT. 2010. Mottle camouflage patterns in cuttlefish: quantitative characterization and visual background stimuli that evoke them. Journal of Experimental Biology 213:187– The mechanisms underlying convergent evolution in the plumage patterns of birds 183 199. Chiao C-C, Chubb C, Hanlon RT. 2007. Interactive effects of size, contrast, intensity and configuration of background objects in evoking disruptive camouflage in cuttlefish. Vision Research 47:2223–2235. Chiao C-C, Kelman EJ, Hanlon RT. 2004. Disruptive Body Patterning of Cuttlefish (Sepia officinalis) Requires Visual Information on Edges and Brightness of Objects on Natural Substrate Backgrounds. The Biological Bulletin 207:153. Chiao C-C, Ulmer KM, Siemann LA, Buresch KC, Chubb C, Hanlon RT. 2013. How visual edge features influence cuttlefish camouflage patterning. Vision Research 83:40–47. Christidis L, Schodde R, Baverstock PR. 1988. Genetic and morphological differentiation and phylogeny in the Australo-Papuan scrubwrens (Sericornis, Acanthizidae). The Auk 105:616–629. Christie DA, Ferguson-Lees J. 2010. Raptors of the World. Bloomsbury Publishing. Chuong CM, Chodankar R, Widelitz RB, Jiang TX. 2000. Evo-devo of feathers and scales: building complex epithelial appendages. Current Opinion in Genetics and Development 10:449–456. Chuong CM. 1993. The making of a feather: homeoproteins, retinoids and adhesion molecules. Bioessays 15:513–521. Coates, BJ, Bishop DK (1997) A Guide to the Birds of Wallacea. Dove Publications, Alderley. Cott HB. 1940. Adaptive Coloration in Animals. Oxford: Oxford University Press. Crawford RD. 1990. Poultry Breeding and Genetics. Amseterdam: Elsevier. Crochet PA, Bonhomme F, Lebreton JD. 2000. Molecular phylogeny and plumage evolution in gulls (Larini). Journal of Evolutionary Biology 13:47–57. Cuadrado A, Nebreda AR. 2010. Mechanisms and functions of p38 MAPK signalling. Journal of Biochemsitry 429:403–417. The mechanisms underlying convergent evolution in the plumage patterns of birds 184 Currie TE, Greenhill SJ, Gray RD, Hasegawa T, Mace R. 2010. Rise and fall of political complexity in island South-East Asia and the Pacific. Nature 467:801–804. Darwin C, Wallace AR. 1858. On the tendency of species to form varieties: and on the perpetuation of varieties and species by natural means of selection. Zoological Journal of the Linnean Soceity. 3:45–62. Darwin C. 1868. The Variation of Animals and Plants Under Domestication. London: John Murray. Darwin C. 1869. On the Origin of the Species by Means of Natural Selection, or, The Preservation of Favoured Races in the Struggle for Life. London: John Murray. Darwin C. 1871. The Descent of Man, and Selection in Relation to Sex. London: John Murray. Davies NB, Brooke M de L. 1988. Cuckoos versus reed warblers: adaptations and counteradaptations. Animal behaviour 36:262–284. Davies NB, Brooke MDL, Kacelnik A. 1996. Recognition Errors and Probability of Parasitism Determine Whether Reed Warblers Should Accept or Reject Mimetic Cuckoo Eggs. Proceedings of the Royal Society B: Biological Sciences 263:925–931. Davies NB, Welbergen JA. 2008. Cuckoo-hawk mimicry? An experimental test. Proceedings of the Royal Society B: Biological Sciences 275:1817–1822. Davies NB, Welbergen JA. 2009. Social transmission of a host defense against cuckoo parasitism. Science 324:1318–1320. Davies NB. 2000. Cuckoos, Cowbirds and Other Cheats. London: T. and A. D. Poyser. Davies NB. 2011. Cuckoo adaptations: trickery and tuning. Journal of Zoology 284:1– 14. De La Pena MR, Rumboll M (1998) Birds of Southern South America and Antarctica. Princeton University Press, Princeton. del Hoyo J, Elliott A, Sargatal J. 1992. Handbook of the Birds of the World: Ostrich to The mechanisms underlying convergent evolution in the plumage patterns of birds 185 Ducks. Barcelona: Lynx edicions. Dickinson EC (2003) The Howard and Moore Complete Checklist of the Birds of the World. Christopher Helm, London. Dessinioti C, Stratigos AJ, Rigopoulos D, Katsambas AD. 2009. A review of genetic disorders of hypopigmentation: lessons learned from the biology of melanocytes. Experimental Dermatology 18:741–749. Dhouailly D, Olivera-Martinez I, Fliniaux I, Missier S, Viallet JP, Thélu J. 2004. Skin field formation: morphogenetic events. International Journal of Developmental Biology 48:85–91. Dhouailly D. 2004. Danielle Dhouailly. Interviewed by Fiona M Watt. Journal of Cell Science 117:1873–1874. Dimitrova M, Merilaita S. 2009. Prey concealment: visual background complexity and prey contrast distribution. Behavioral ecology 21:176–181. Dimitrova M, Merilaita S. 2012. Prey pattern regularity and background complexity affect detectability of background-matching prey. Behavioral ecology 23:384–390. Dores RM, Baron AJ. 2011. Evolution of POMC: origin, phylogeny, posttranslational processing, and the melanocortins. Annals of the New York Academy of Sciences 1220:34–48. Dores RM. 2013. Observations on the evolution of the melanocortin receptor gene family: distinctive features of the melanocortin-2 receptor. Frontiers in Neuroscience 7:1–12. Dorshorst BJ, Ashwell CM. 2009. Genetic mapping of the sex-linked barring gene in the chicken. Poultry Science 88:1811–1817. Available from: Doucet SM, Hill GE. 2009. Do museum specimens accurately represent wild birds? A case study of carotenoid, melanin, and structural colours in long-tailed manakins Chiroxiphia linearis. Journal of Avian Biology 40:146–156. Doucet SM, Shawkey MD, Rathburn MK, H L Mays J, Montgomerie R. 2004. Concordant evolution of plumage colour, feather microstructure and a melanocortin The mechanisms underlying convergent evolution in the plumage patterns of birds 186 receptor gene between mainland and island populations of a fairy-wren. Proceedings of the Royal Society B: Biological Sciences 271:1663–1670. Ducrest A, Keller L, Roulin A. 2008. Pleiotropy in the melanocortin system, coloration and behavioural syndromes. Trends in Ecology and Evolution 23:502–510. Dumbacher JP, Fleischer RC. 2001. Phylogenetic evidence for colour pattern convergence in toxic pitohuis: Müllerian mimicry in birds? Proceedings of the Royal Society B: Biological Sciences 268:1971–1976. Eizirik E, Yuhki N, Johnson WE, Menotti-Raymond M, Hannah SS, O'Brien SJ. 2003. Molecular Genetics and Evolution of Melanism in the Cat Family. Current Biology 13:448–453. Emaresi G, Bize P, Altwegg R, Henry I, van den Brink V, Gasparini J, Roulin A. 2014. Melanin-specific life-history strategies. American Naturalist 183:269–280. Emaresi G, Ducrest A-L, Bize P, Richter H, Simon C, Roulin A. 2013. Pleiotropy in the melanocortin system: expression levels of this system are associated with melanogenesis and pigmentation in the tawny owl (Strix aluco). Molecular Ecology 22:4915–4930. Endler JA, Thery M. 2010. Interacting effects of lek placement, display behaviour, ambient light and colour patterns in three neotropical forest-dwelling birds. American Naturalist 421–452. Endler JA. 1978. Evolutionary Biology. (Hecht MK, Steere WC, Wallace B, editors.). Boston, MA: Springer US. Endler JA. 1980. Natural Selection on Color Patterns in Poecilia reticulata. Evolution 34:76–91. Endler JA. 1992. Signals, signal conditions, and the direction of evolution. American Naturalist S125–S153. Endler JA. 1998. Sensory ecology, receiver biases and sexual selection. Trends in Ecology and Evolution 13:415–420. Endler JA. 2007. The Color of Light in Forests and Its Implications. Ecological The mechanisms underlying convergent evolution in the plumage patterns of birds 187 Monographs 63:1–27. Available from: Enshell-Seijffers D, Lindon C, Morgan BA. 2007. The serine protease Corin is a novel modifier of the agouti pathway. Development 135:217–225. Feeney WE, Welbergen JA, Langmore NE. 2012. The frontline of avian brood parasite-host coevolution. Animal behaviour:1–10. Felsenstein J. 1985. Phylogenies and the Comparative Method. American Naturalist 125:1–15. Ferguson-Lees J, Christie DA (2001) Raptors of the world. Christopher Helm, London. Fisher RA. 1923. The Genetical Theory of Natural Selection. Nabu Press. Fliniaux I, Viallet JP, Dhouailly D. 2004. Signaling dynamics of feather tract formation from the chick somatopleure. Development 131:3955–3966. Fontanesi L, Forestier L, Allain D, Scotti E, Beretti F, Deretz-Picoulet S, Pecchioli E, Vernesi C, Robinson TJ, Malaney JL, et al. 2010. Characterization of the rabbit agouti signaling protein (ASIP) gene: Transcripts and phylogenetic analyses and identification of the causative mutation of the nonagouti black coat colour. Genomics 95:166–175. Friedman NR, Kiere LM, Omland KE. 2011. Convergent gains of red carotenoid- based coloration in the New World blackbirds. The Auk 128:678–687. Fritz JA, Brancale J, Tokita M, Burns KJ, Hawkins MB, Abzhanov A, Brenner MP. 2014. Shared developmental programme strongly constrains beak shape diversity in songbirds. Nature Communications 5:1–9. Galván I, Ghanem G, Møller AP. 2012. Has removal of excess cysteine led to the evolution of pheomelanin? Bioessays 34:565–568. Gantz I, Fong TM. 2003. The melanocortin system. American Journal of Physiology and Endocrinology. 284:468–74. Garcia JE, Rohr D, Dyer AG. 2013. Trade-off between camouflage and sexual The mechanisms underlying convergent evolution in the plumage patterns of birds 188 dimorphism revealed by UV digital imaging: the case of Australian Mallee dragons (Ctenophorus fordi). Journal of Experimental Biology 216:4290–4298. Gaston AJ. 1976. Brood Parasitism by the Pied Crested Cuckoo Clamator jacobinus. Journal of Animal Ecology 45:331–348. Girardot M, Martin J, Guibert S, Leveziel H, Julien R, Oulmouden A. 2005. Widespread expression of the bovine Agouti gene results from at least three alternative promoters. Pigment Cell Research 18:34–41. Gluckman T-L, Cardoso GC. 2009. A Method to Quantify the Regularity of Barred Plumage Patterns. Behavioural Ecology and Sociobiology 63:1837–1844. Gluckman T-L, Cardoso GC. 2010. The dual function of barred plumage in birds: camouflage and communication. Journal of Evolutionary Biology 23:2501–2506. Gluckman T-L, Mundy NI. 2013. Cuckoos in raptors' clothing: barred plumage illuminates a fundamental principle of Batesian mimicry. Animal behaviour 86:1165– 1181. Gluckman T-L. 2014. Pathways to elaboration of sexual dimorphism in bird plumage patterns. Biological Journal of the Linnean Society 111:262–273. Godfrey D, Lythgoe JN, Rumball DA. 1987. Zebra Stripes and Tiger Stripes: the Spatial Frequency Distribution of the Pattern Compared to That of the Background Is Significant in Display and Crypsis. Biological Journal the Linnean Society:427–433. Gonzalez J, Düttmann H, Wink M. 2009. Phylogenetic relationships based on two mitochondrial genes and hybridization patterns in Anatidae. Journal of Zoology 279:310–318. Gould SJ. 2002. The Structure of Evolutionary Theory. USA: Harvard University Press. Grant PR, Grant BR. 1992. Hybridization of bird species. Science 256:193–197. Griffith SC, Owens IP, Burke T. 1999. Environmental determination of a sexually selected trait. Nature 400:358–360. Grim Tomáš. 2005. Host Recognition of Brood Parasites: Implications for The mechanisms underlying convergent evolution in the plumage patterns of birds 189 Methodology in Studies of Enemy Recognition. The Auk 122:530–543. Grim Tomáš. 2006. The evolution of nestling discrimination by hosts of parasitic birds: why is rejection so rare? Evolutionary Ecology Research 8:785–802. Grissom RJ. 2004. Probability of the superior outcome of one treatment over another. Journal of Applied Psychology:1–3. Gunn TM, Barsh GS. 2000. Mahogany/Attractin: Enroute from Phenotype to Function. TCM 10:1–6. Gunn TM, Miller KA, He L, Hyman RW, Davis RW, Azarani A, Schlossman SF, Duke- Cohan JS, Barsh GS. 1999. The mouse mahogany locus encodes a transmembrane form of human attractin. Nature 398:152–156. Haag ES, Lenski RE. 2011. L'enfant terrible at 30: the maturation of evolutionary developmental biology. Development 138:2633–2637. Hackett SJ, Rosenberg KV. 1990. Comparison of phenotypic and genetic differentiation in South American antwrens (Formicariidae). The Auk 107:473–489. Hadfield JD, Burgess MD, Lord A, Phillimore AB, Clegg SM, Owens IPF. 2006. Direct versus indirect sexual selection: genetic basis of colour, size and recruitment in a wild bird. Proceedings of the Royal Society B: Biological Science 273:1347–1353. Hale K, V Briskie J. 2007. Response of introduced European birds in New Zealand to experimental brood parasitism. Journal of Avian Biology 38:198-204 Hall JR, Cuthill IC, Baddeley R, Shohet AJ, Scott-Samuel NE. 2013. Camouflage, detection and identification of moving targets. Proceedings of the Royal Society B: Biological Sciences 280:20130064. Hamburger V, Hamilton HL. 1992. A series of normal stages in the development of the chick embryo. Journal of morphology 195:231–272. Hanlon RT, Chiao CC, Mathger LM, Barbosa A, Buresch KC, Chubb C. 2009. Cephalopod dynamic camouflage: bridging the continuum between background matching and disruptive coloration. Philosophical Transactions of the Royal Society B: Biological Sciences 364:4717–4723. The mechanisms underlying convergent evolution in the plumage patterns of birds 190 Hanlon RT, Naud MJ, Forsythe JW, Hall K, Watson AC, McKechnie J. 2007. Adaptable Night Camouflage by Cuttlefish. American Naturalist 169:543–551. Harding RM, Healy E, Ray AJ, Ellis NS, Flanagan N, Todd C, Dixon C, Sajantila A, Jackson IJ, Birch-Machin MA, et al. 2000. Evidence for Variable Selective Pressures at MC1R. The American Journal of Human Genetics 66:1351–1361. Harris MP, Fallon JF, Prum RO. 2002. Shh-Bmp2 signaling module and the evolutionary origin and diversification of feathers. Journal of Experimental Zoology 294:160–176. Harris MP, Williamson S, Fallon JF, Meinhardt H, Prum RO. 2005. Molecular evidence for an activator-inhibitor mechanism in development of embryonic feather branching. In: Vol. 102. Proceedings of the National Academy of Sciences USA 11734–11739. Hart NS, Hunt DM. 2006. Avian visual pigments -characteristics, spectral tuning, and evolution. American Naturalist 1:7–26. Hart NS, Partridge JC, Cuthill IC, Bennett AT. 2012. Visual pigments, oil droplets, ocular media and cone photoreceptor distribution in two species of passerine bird: the blue tit (Parus caeruleus L.) and the blackbird (Turdus merula L.). Journal of Computational Physiology 186:375–387. Hasson O. 1991. Sexual displays as amplifiers: practical examples with an emphasis on feather decorations. Behavioral ecology 2:1–9. Hauber ME, Russo SA. 2000. Perch Proximity Correlates with Higher Rates of Cowbird Parasitism of Ground Nesting Song Sparrows. The Wilson Bulletin 112:150– 153. Hayashi H, Imai JK, Imai K. 1991. Characterization of Chicken ACTH and a-MSH: The Primary Sequence of Chicken ACTH Is More Similar to Xenopus ACTH than to Other Avian ACTH. General and comparative Endocrinology:434–443. He L, Eldridge PK, Gunn TM, Barsh GS. 2003. Accessory proteins for melanocortin signaling Attractin and Mahogunin. Annals of the New York Academy of Sciences 994:288–298. The mechanisms underlying convergent evolution in the plumage patterns of birds 191 Heinzel H, Fitter R, Parslow J (1995) Birds of Britain and Europe. Harper Collins Publishers, London. Hellström AR, Sundström E, Gunnarsson U, Bed’Hom B, Tixier-Boichard M, Honaker CF, Sahlqvist A-S, Jensen P, Kämpe O, Siegel PB, et al. 2010. Sex-linked barring in chickens is controlled by the CDKN2A /B tumour suppressor locus. Pigment Cell and Melanoma Research 23:521–530. Helversen von B, Schooler LJ, Czienskowski U. 2013. Are Stripes Beneficial? Dazzle Camouflage Influences Perceived Speed and Hit Rates. PLoS ONE 8:e61173. Hemmi JM, Marshall J, Pix W, Vorobyev M, Zeil J. 2006. The variable colours of the fiddler crab Uca vomeris and their relation to background and predation. Journal of Experimental Biology 209:4140–4153. Henig RM. 2001. The Monk in the Garden. Houghton Mifflin Harcourt. Hill GE, McGraw KJ. 2006a. Bird Coloration: Function and evolution. USA: Harvard University Press. Hill GE, McGraw KJ. 2006b. Bird Coloration: Mechanisms and measurements. USA: Harvard University Press. Hoekstra HE, Hirschmann RJ, Bundey RA, Insel PA, Crossland JP. 2006. A single amino acid mutation contributes to adaptive beach mouse color pattern. Science 313:101–104. Hofmann CM, Cronin TW, Omland KE. 2006. Using spectral data to reconstruct evolutionary changes in coloration: carotenoid color evolution in New World orioles. Evolution 60:1680–1691. Hoggard N. 2004. Regulation of adipose tissue leptin secretion by alpha-melanocyte- stimulating hormone and agouti-related protein: further evidence of an interaction between leptin and the melanocortin signalling system. Journal of Molecular Endocrinology 32:145–153. Hohman WL, Richard DM. 1994. Timing of Remigial Molt in Fulvous Whistling Ducks Nesting in Louisiana. The Southwestern Naturalist 39:190–192. The mechanisms underlying convergent evolution in the plumage patterns of birds 192 Hohman WL. 1996. Prevalence of double Wing Molt in Free-living Ruddy Ducks. The Southwestern Naturalist 41:195–198. Hollander WF. 1968. Brown-Rippled, A Recessive Mutant in the Muscovy Duck. Journal of Heredity 59:309–311. Honza M, Grim T, Capek M Jr, Moksnes A, R skaft E. 2004. Nest defence, enemy recognition and nest inspection behaviour of experimentally parasitized Reed Warblers Acrocephalus scirpaceus: Capsule Reed Warblers in a regularly parasitized population do not recognize Cuckoo Cuculus canorus as a special enemy and do not change their behaviour at nest immediately after being parasitized. Bird Study 51:256–263. Honza M, Šicha V, Procházka P, Ležalová R. 2006. Host nest defense against a color-dimorphic brood parasite: great reed warblers (Acrocephalus arundinaceus) versus common cuckoos (Cuculus canorus). Journal of Ornithology 147:629–637. Houston AI, Stevens M, Cuthill IC. 2007. Animal camouflage: compromise or specialize in a 2 patch-type environment? Behavioral ecology 18:769–775. How MJ, Zanker JM. 2014. Motion camouflage induced by zebra stripes. Zoology 117:163–170. Howard R, Dickinson EC, Moore A. 2003. The Howard and Moore Complete Checklist of the Birds of the World. UK: Christopher Helm. Hunter J. 1780. Account of an Extraordinary Pheasant. Philosophical Transactions of the Royal Society of London 70:527–535. Hurlbert AH, Jetz W. 2007. Species richness, hotspots, and the scale dependence of range maps in ecology and conservation. Proceedings of the National Academy of Sciences USA 104:13384–13389. Huxley J. 1942. Evolution. The Modern Synthesis. London: George Alien and Unwin Ltd. IBM Corporation. 2010. IBM SPSS Statistics for Mac, Version 19.0. IBM Corporation: USA. The mechanisms underlying convergent evolution in the plumage patterns of birds 193 Ishida Y, David VA, Eizirik E, Schäffer AA, Neelam BA, Roelke ME, Hannah SS, O'Brien SJ, Menotti-Raymond M. 2006. A homozygous single-base deletion in MLPH causes the dilute coat color phenotype in the domestic cat. Genomics 88:698–705. Ito S, Wakamatsu K, Ozeki H. 2000. Chemical analysis of melanins and its application to the study of the regulation of melanogenesis. Pigment Cell Research 8:103–109. Jackson IJ, Budd PS, Keighren M, McKie L. 2007. Humanized MC1R transgenic mice reveal human specific receptor function. Human Molecular Genetics 16:2341– 2348. Jackson IJ. 1997. Homologous pigmentation mutations in human, mouse and other model organisms. Human Molecular Genetics 6:1613–1624. Jackson JF, Ingram W III, Campbell HW. 1976. The dorsal pigmentation pattern of snakes as an antipredator strategy: a multivariate approach. American Naturalist1029–1053. Jetz W, Thomas GH, Joy JB, Hartmann K, Mooers AO. 2012. The global diversity of birds in space and time. Nature 491:444–448. Jiang TX, Jung HS, Widelitz RB, Chuong CM. 1999. Self-organization of periodic patterns by dissociated feather mesenchymal cells and the regulation of size, number and spacing of primordia. Development 126:4997–5009. Jiang TX, Widelitz RB, Shen W-M, Will P, Wu D-Y, Lin C-M, Jung H-S, Chuong C-M. 2004. Integument pattern formation involves genetic and epigenetic controls: feather arrays simulated by digital hormone models. International Journal of Developmental Biology 48:117–135. Jones AW, Kennedy RS. 2008. Plumage convergence and evolutionary history of the island thrush in the Philippines. Condor 110:35–44. Jones CD, Osorio D. 2004. Discrimination of oriented visual textures by poultry chicks. Vision Research 44:83–89. Josef N, Amodio P, Fiorito G, Shashar N. 2012. Camouflaging in a Complex Environment — Octopuses Use Specific Features of Their Surroundings for The mechanisms underlying convergent evolution in the plumage patterns of birds 194 Background Matching. PLoS ONE 7:e37579. Kaelin CB, Xu X, Hong LZ, David VA, McGowan KA, Schmidt-Kuntzel A, Roelke ME, Pino J, Pontius J, Cooper GM, et al. 2012. Specifying and Sustaining Pigmentation Patterns in Domestic and Wild Cats. Science 337:1536–1541. Karpestam E, Merilaita S, Forsman A. 2014. Natural levels of colour polymorphism reduce performance of visual predators searching for camouflaged prey. Biological Journal of the Linnean Society:1–10. Kass RE, Raftery AE. 1995. BayesFactors. Journal of the American Statistical Assocation 90:1–23. Kenward B, Wachtmeister C-A, Ghirlanda S, Enquist M. 2004. Spots and stripes: the evolution of repetition in visual signal form. Journal of Theoretical Biology 230:407– 419. Kerje S, Lind J, Schütz K, Jensen P, Andersson L. 2003. Melanocortin 1-receptor (MC1R) mutations are associated with plumage colour in chicken. Animal Genetics 34:241–248. Kijas JM, Moller M, Plastow G, Andersson L. 2001. A frameshift mutation in MC1R and a high frequency of somatic reversions cause black spotting in pigs. Genetics 158:779–785. Kimball RT, Ligon JD. 1999. Evolution of Avian Plumage Dichromatism from a Proximate Perspective. American Naturalist 154:182–193. Kimball RT, Mary CMS, Braun EL. 2011. A Macroevolutionary Perspective on Multiple Sexual Traits in the Phasianidae (Galliformes). International Journal of Evolutionary Biology 2011:1–16. Kinoshita K, Akiyama T, Mizutani M, Shinomiya A, Ishikawa A, Younis HH, Tsudzuki M, Namikawa T, Matsuda Y. 2014. Endothelin Receptor B2 (Ednrb2) Is Responsible for the Tyrosinase-Independent Recessive White (Mo(W)) And Mottled (mo) Plumage Phenotypes in the Chicken. PLoS ONE 9:86361. Klingenberg CP. 2008. Morphological Integration and Developmental Modularity. The Annual Review of Ecology, Evolution, and Systematics 39:115–132. The mechanisms underlying convergent evolution in the plumage patterns of birds 195 Klovins J, Schiöth HB. 2006. Agouti-Related Proteins (AGRPs) and Agouti-Signaling Peptide (ASIP) in Fish and Chicken. Annals of the New York Academy of Sciences 1040:363–367. Kose M, Mänd R, Moller AP. 1999. Sexual selection for white tail spots in the barn swallow in relation to habitat choice by feather lice. Animal behaviour 58:1201–1205. Kraaijeveld K, Kraaijeveld-Smit FJL, Komdeur J. 2007. The evolution of mutual ornamentation. Animal behaviour 74:657–677. Krude H, Biebermann H, Luck W, Horn R, Brabant G, Grüters A. 1998. Severe early- onset obesity, adrenal insufficiency and red hair pigmentation caused by POMC mutations in humans. Nature Genetics 19:155–157. Krude H, Grüters A. 2000. Implications of proopiomelanocortin (POMC) mutations in humans: the POMC deficiency syndrome. Trends in Endocrinology 11:15–22. Kruger O, Davies NB, Sorenson MD. 2007. The evolution of sexual dimorphism in parasitic cuckoos: sexual selection or coevolution? Proceedings of the Royal Society B: Biological Sciences 274:1553–1560. Kusmierski R, Borgia G, Uy A, Crozier RH. 1997. Labile evolution of display traits in bowerbirds indicates reduced effects of phylogenetic constraint. Proceedings of the Royal Society B: Biological Sciences 264:307–313. Kutschera U, Niklas K. 2004. The modern theory of biological evolution: an expanded synthesis. Naturwissenschaften 91:255–276. Lande R. 1980. Sexual Dimorphism, Sexual Selection, and Adaptation in Polygenic Characters. Evolution 34:292–305. Langmore NE, Stevens M, Maurer G, Heinsohn R, Hall ML, Peters A, Kilner RM. 2011. Visual mimicry of host nestlings by cuckoos. Proceedings of the Royal Society B: Biological Sciences 278:2455–2463. Langrand O (1990) A Guide to the Birds of Madagascar. Yale University Press, New Haven. Le Douarin N, Kalcheim C. 1999. The Neural Crest. Cambridge University Press. The mechanisms underlying convergent evolution in the plumage patterns of birds 196 Lin C-M, Jiang TX, Widelitz RB, Chuong C-M. 2006. Molecular signaling in feather morphogenesis. Current Opinion in Cell Biology 18:730–741. Lin SJ, Foley J, Jiang TX, Yeh CY, Wu P, Foley A, Yen CM, Huang YC, Cheng HC, Chen CF, et al. 2013. Topology of feather melanocyte progenitor niche allows complex pigment patterns to emerge. Science 340:1442–1445. Lindell LE, Forsman A. 1996. Sexual dichromatism in snakes: support for the flicker- fusion hypothesis. Canadian Journal of Zoology 74:2254–2256. Lindholm AK, Thomas RJ. 2000. Differences between Populations of Reed Warblers in Defences against Brood Parasitism. Behaviour 137:25–42. Ling MK, Hotta E, Kilianova Z, Haitina T, Ringholm A, Johansson L, Gallo-Payet N, Takeuchi S, Schiöth HB. 2004. The melanocortin receptor subtypes in chicken have high preference to ACTH-derived peptides. Br. J. Pharmacol. 143:626–637. Ling MK, Lagerström MC, Fredriksson R, Okimoto R, Mundy NI, Takeuchi S, Schiöth HB. 2003. Association of feather colour with constitutively active melanocortin 1 receptors in chicken. European Journal of Biochemistry 270:1441–1449. Linnen CR, Kingsley EP, Jensen JD, Hoekstra HE. 2009. On the Origin and Spread of an Adaptive Allele in Deer Mice. Science 325:1095–1098. Livezey BC, Zusi RL. 2007. Higher-order phylogeny of modern birds (Theropoda, Aves: Neornithes) based on comparative anatomy. II. Analysis and discussion. Zoological Journal of the Linnean Society 149:1–95. Lu D, Willard D, Patel IR, Kadwell S, Overton L, Kost T, Luther M, Chen W, Woychik RP, Wilkison WO. 1994. Agouti protein is an antagonist of the melanocyte- stimulating-hormone receptor. Nature 371:799–802. Lucas AM, Stettenheim PR. 1972. Avian anatomy: integument. Washington DC: US Department of Agriculture. MacKinnon J, Phillips K (1993) The Birds of Borneo, Sumatra, Java, and Bali. Oxford University Press, Oxford. The mechanisms underlying convergent evolution in the plumage patterns of birds 197 Madsen T, Shine R. 1994. Components of lifetime reproductive success in adders, Vipera berus. Journal of Animal Ecology. Majerus MEN, Mundy NI. 2003. Mammalian melanism: natural selection in black and white. Trends in Genetics 19:585–588. Manceau M, Domingues VS, Linnen CR, Rosenblum EB, Hoekstra HE. 2010. Convergence in pigmentation at multiple levels: mutations, genes and function. Philosophical Transactions of the Royal Society B: Biological Sciences 365:2439– 2450. Manceau M, Domingues VS, Mallarino R, Hoekstra HE. 2011. The developmental role of Agouti in color pattern evolution. Science 331:1062–1065. Marshall JN. 2000. Communication and Camouflage with the Same ‘Bright’ Colours in Reef Fishes. Philosophical Transactions: Biological Sciences 355:1243–1248. Matesic LE, Yip R, Reuss AE, Swing DA, O'Sullivan TN, Fletcher CF, Copeland NG, Jenkins NA. 2001. Mutations in Mlph, encoding a member of the Rab effector family, cause the melanosome transport defects observed in leaden mice. Proceedings of the National Academy of Sciences 98:10238–10243. Mauger A, Demarchez M, Herbage D, Grimaud JA, Druguet M, Hartmann D, Sengel P. 1982. Immunofluorescent localization of collagen types I and III, and of fibronectin during feather morphogenesis in the chick embryo. Developmental Biology 94:93– 105. Maynard Smith J, Burian R, Kauffman S, Alberch P, Campbell J, Goodwin B, Lande R, Raup D, Wolpert L. 1985. Developmental Constraints and Evolution: A Perspective from the Mountain Lake Conference on Development and Evolution. Quarterly Review of Biology 60:265–287. Mäthger LM, Chiao C-C, Barbosa A, Hanlon RT. 2008. Color matching on natural substrates in cuttlefish, Sepia officinalis. Journal of Computational Physiology A 194:577–585. Mendel JG. 1866. Versuche über Pflanzenhybriden Verhandlungen des naturforschenden Vereines in Brünn. Bd. IV für das Jahr, 1865 Abhandlungen:3–47. The mechanisms underlying convergent evolution in the plumage patterns of birds 198 Mindell DP, Sorenson MD, Dimcheff DE, Hasegawa M, Ast JC, Yuri T. 1999. Interordinal relationships of birds and other reptiles based on whole mitochondrial genomes. Systematic Biology 48:138–152. Minvielle F, Bed’Hom B, Coville J-L, Ito S, Inoue-Murayama M, Gourichon D. 2010. The "silver" Japanese quail and the MITF gene: causal mutation, associated traits and homology with the ‘blue’ chicken plumage. BMC Genetics 11:15. Miwa M, Inoue-Murayama M, Aoki H, Kunisada T, Hiragaki T, Mizutani M, Ito S. 2007. Endothelin receptor B2 (EDNRB2) is associated with the panda plumage colour mutation in Japanese quail. Animal Genetics 38:103–108. Miwa M, Inoue-Murayama M, Kobayashi N, Kayang BB, Mizutani M, Takahashi H, Ito S. 2006. Mapping of panda plumage color locus on the microsatellite linkage map of the Japanese quail. BMC Genetics 7:2. Mochii M, Agata K, Kobayashi H, Yamamoto TS, Eguchi G. 1988. Expression of gene coding for a melanosomal matrix protein transcriptionally regulated in the transdifferentiation of chick embryo pigmented epithelial cells. Cell Differentiation 67– 74. Mochii M, Mazaki Y, Mizuno N, Hayashi H, Eguchi G. 1998. Role of Mitf in differentiation and transdifferentiation of chicken pigmented epithelial cell. Developmental Biology 193:47–62. Mochii M, Ono T, Matsubara Y, Eguchi G. 1998. Spontaneous transdifferentiation of quail pigmented epithelial cell is accompanied by a mutation in the Mitf gene. Developmental Biology 196:145–159. Mochii M, Takeuchi T, Kodama R, Agata K, Eguchi G. 1988. The expression of melanosomal matrix protein in the transdifferentiation of pigmented epithelial cells into lens cells. Cell Differentiation 23:133–141. Moksnes A, Røskaft E, Braa AT, Korsnes L, Lampe HM, Pedersen HC. 2007. Behavioural responses of potential hosts towards artificial cuckoo eggs and dummies. Behaviour 116:64–89. Moksnes A, Røskaft E, Hagen LG, Honza M, Mork C, Olsen PH. 2000. Common Cuckoo Cuculus canorus and host behaviour at Reed Warbler Acrocephalus The mechanisms underlying convergent evolution in the plumage patterns of birds 199 scirpaceus nests. Ibis 142:247–258. Morgan-Richards M, Trewick SA, Bartosch-Härlid A, Kardailsky O, Phillips MJ, McLenachan PA, Penny D. 2008. Bird evolution: testing the Metaves clade with six new mitochondrial genomes. BMC Evolutionary Biology 8:20. Mosimann C, Hausmann G, Basler K. 2009. β-Catenin hits chromatin: regulation of Wnt target gene activation. Nature Reviews in Molecular Cell Biology 10:276–286. Moskát C, Székely T, Cuthill IC, Kisbenedek T. 2008. Hosts’ Responses to Parasitic Eggs: Which Cues Elicit Hosts’ Egg Discrimination? Ethology 114:186–194. Muck C, Goymann W. 2011. Throat patch size and darkness covaries with testosterone in females of a sex-role reversed species. Behavioral ecology 22:1312– 1319. Mundy NI, Badcock NS, Hart T, Scribner K, Janssen K, Nadeau NJ. 2004. Conserved genetic basis of a quantitative plumage trait involved in mate choice. Science 303:1870–1873. Mundy NI. 2005. A window on the genetics of evolution: MC1R and plumage colouration in birds. Proceedings of the Royal Society B: Biological Sciences 272:1633–1640. Murray JD. 1981. A pre-pattern formation mechanism for animal coat markings. Journal of Theoretical Biology:161–199. Münkemüller T, Lavergne S, Bzeznik B. 2012. How to measure and test phylogenetic signal. Methods in Ecology and Evolution 743–756. Nadeau NJ, Burke T, Mundy NI. 2007. Evolution of an Avian Pigmentation Gene Correlates with a Measure of Sexual Selection. Proceedings of the Royal Society: Biological Sciences 274:1807–1813. Nadeau NJ, Minvielle F, Ito S, Inoue-Murayama M, Gourichon D, Follett SA, Burke T, Mundy NI. 2008. Characterization of Japanese Quail yellow as a Genomic Deletion Upstream of the Avian Homolog of the Mammalian ASIP (agouti) Gene. Genetics 178:777–786. The mechanisms underlying convergent evolution in the plumage patterns of birds 200 Nadeau NJ, Minvielle F, Mundy NI. 2006. Association of a Glu92Lys substitution in MC1R with extended brown in Japanese quail (Coturnix japonica). Animal Genetics 37:287–289. Nadeau NJ. 2006. The evolutionary genetics of sexually selected plumage colour traits in the galliform birds. PhD thesis, Cambridge University. Nakagawa S, Cuthill IC. 2007. Effect size, confidence interval and statistical significance: a practical guide for biologists. Biological Reviews 82:591–605. Nataf V, Mercier P, De Néchaud B, Guillemot JC, Capdevielle J, Lapointe F, Le Douarin NM. 1995. Melanoblast/melanocyte early marker (MelEM) is a glutathione S- transferase subunit. Experimental Cell Research 218:394–400. Niwa T, Mochii M, Nakamura A, Shiojiri N. 2002. Plumage pigmentation and expression of its regulatory genes during quail development - histochemical analysis using Bh (black at hatch) mutants. Mechanisms of Development 118:139–146. Noramly S, Morgan BA. 1998. BMPs mediate lateral inhibition at successive stages in feather tract development. Development 125:3775–3787. O'Rourke CT, Hall MI, Pitlik T, Fernández-Juricic E. 2010. Hawk Eyes I: Diurnal Raptors Differ in Visual Fields and Degree of Eye Movement. PLoS ONE 5:e12802. Odeen A, Bjorklund M. 2003. Dynamics in the evolution of sexual traits: losses and gains, radiation and convergence in yellow wagtails (Motacilla flava). Molecular Ecology 12:2113–2130. Oliver JC, Robertson KA, Monteiro A. 2009. Accommodating natural and sexual selection in butterfly wing pattern evolution. Proceedings of the Royal Society B: Biological Sciences 276:2369–2375. Ollmann MM, Lamoreux ML, Wilson BD, Barsh GS. 1998. Interaction of Agouti protein with the melanocortin 1 receptor in vitro and in vivo. Genes and Development 12:316–330. Olson DM, Dinerstein E, Wikramanayake ED, Burgess ND, Powell GVN, Underwood EC, D'amico JA, Itoua I, Strand HE, Morrison JC, et al. 2001. Terrestrial Ecoregions of the World: A New Map of Life on Earth: A new global map of terrestrial ecoregions The mechanisms underlying convergent evolution in the plumage patterns of birds 201 provides an innovative tool for conserving biodiversity. BioScience 51:933–938. Omland KE, Hofmann CM, Hill GE, McGraw KJ. 2006. Adding color to the past: ancestral-state reconstruction of coloration. in Hill GE, McGraw KJ. 2006b. Bird Coloration: Mechanisms and measurements. USA: Harvard University Press. Omland KE, Lanyon SM. 2000. Reconstructing plumage evolution in Orioles (Icterus) - Repeated convergence and reversal in patterns. Evolution 54:2119–2133. Omland KE. 1996. Female mallard mating preferences for multiple male ornaments - experimental variation. Behavioural Ecology and Sociobiology 39:361–366. Omland KE. 1997. Examining two standard assumptions of ancestral reconstructions - repeated loss of dichromatism in dabbling ducks. Evolution 1–12. Oribe E, Fukao A, Yoshihara C, Mendori M, Rosal KG, Takahashi S, Takeuchi S. 2012. Conserved distal promoter of the agouti signaling protein (ASIP) gene controls sexual dichromatism in chickens. General and comparative Endocrinology 177:231– 237. Osorio D, Vorobyev M. 2005. Photoreceptor sectral sensitivities in terrestrial animals: adaptations for luminance and colour vision. Proceedings of the Royal Society B: Biological Sciences 272:1745–1752. Oster G, Alberch P. 1982. Evolution and bifurcation of developmental programs. Evolution 1–17. Oster GF, Murray JD. 2005. Pattern formation models and developmental constraints. Journal of Experimental Zoology 251:186–202. Oster GF, Shubin N, Murray JD, Alberch P. 1988. Evolution and morphogenetic rules - the shape of the vertebrate limb in ontogeny and phylogeny. Evolution 1–24. Owens I, Hartley IR. 1998. Sexual dimorphism in birds: why are there so many different forms of dimorphism? Proceedings of the Royal Society: Biological Science 265:397–407. Owens IP, Short RV. 1995. Hormonal basis of sexual dimorphism in birds: implications for new theories of sexual selection. Trends in Ecology and Evolution The mechanisms underlying convergent evolution in the plumage patterns of birds 202 10:44–47. Pagel M, Meade A, Barker D. 2004. Bayesian Estimation of Ancestral Character States on Phylogenies. Systematic Biology 53:673–684. Pagel M, Meade A. 2006. Bayesian analysis of correlated evolution of discrete character by reversiable jump MCMC. American Naturalist 167:808–825. Pagel M. 1999. Inferring the historical patterns of biological evolution. Nature 401:877–884. Parra-Olea G, Wake DB. 2001. Extreme morphological and ecological homoplasy in tropical salamanders. Proceedings of the National Academy of Sciences USA 98:7888–7891. Payne RB, Sorensen MD. 2005. The Cuckoos. Oxford University Press. Payne RB. 2010. Interspecific communication signals in parasitic birds. American Naturalist 363–375. Pecori Giraldi F, Cassarino F, Pagliardini L, Asnaghi V, Cavagnini F. 2011. The human POMC gene promoter: where do we stand? Journal of Endocrinological Investigation 34:454–460. van Perlo B (2006) Birds of Mexico and Central America. Princeton University Press, Princeton. Petrie M, Halliday T, Sanders C. 1991. Peahens prefer peacocks with elaborate trains. Animal behaviour 41:323–331. Petrie M, Halliday T. 2008. Experimental and natural changes in the peacocks (Pavo cristatus) train can affect mating success. Behavioural Ecology and Sociobiology 35:213–217. Pérez-Rodríguez L, Jovani R, Mougeot F. 2013. Fractal geometry of a complex plumage trait reveals bird's quality. Proceedings of the Royal Society B: Biological Sciences 280:20122783–20122783. Pfennig DW, Mullen SP. 2010. Mimics without models: causes and consequences of The mechanisms underlying convergent evolution in the plumage patterns of birds 203 allopatry in Batesian mimicry complexes. Proceedings of the Royal Society B: Biological Sciences 277:2577–2585. Poelstra JW, Vijay N, Bossu CM, Lantz H, Ryll B, Muller I, Baglione V, Unneberg P, Wikelski M, Grabherr MG, et al. 2014. The genomic landscape underlying phenotypic integrity in the face of gene flow in crows. Science 344:1410–1414. Pointer MA, Mundy NI. 2008. Testing whether macroevolution follows microevolution: Are colour differences among swans (Cygnus) attributable to variation at the MC1R locus? BMC Evolutionary Biology 8:249. Polak M. 2012. Habitat Preferences of the Sympatric Barred Warbler (Sylvia nisoria) and the Red-Backed Shrike (Lanius collurio) Breeding in Central Poland. Annales Zoologici Fennici 49:355-363. Pomiankowski A, Iwasa Y. 1998. Runaway Ornament Diversity Caused by Fisherian Sexual Selection. Proceedings of the National Academy of Sciences USA 95:5106– 5111. Pough FH. 1976. Multiple Cryptic Effects of Crossbanded and Ringed Patterns of Snakes. Copeia 1976:834–836. Poulton EB. 1890. The colours of animals. London: Kegan Paul, Trench, Trubner and Co. Ltd. Požgayová M, Procházka P, Honza M. 2009. Adjustment of incubation according to the threat posed: a further signal of enemy recognition in the Blackcap Sylvia atricapilla? Journal of Ornithology 150:569–576. Prager M, Andersson S. 2010. Convergent evolution of red carotenoid coloration in widowbirds and bishops (Euplectes spp.). Evolution 64:3609–3619. Price T, Bontrager A. 2001. Evolutionary genetics: The evolution of plumage patterns. Current Biology 11:R405–R408. Price T, Pavelka M. 1996. Evolution of a colour pattern: history, development, and selection. Journal of Evolutionary Biology 9:451–470. Prud'homme B, Gompel N, Rokas A, Kassner VA, Williams TM, Yeh S-D, True JR, The mechanisms underlying convergent evolution in the plumage patterns of birds 204 Carroll SB. 2006. Repeated morphological evolution through cis-regulatory changes in a pleiotropic gene. Nature 440:1050–1053. Prum RO, Brush AH. 2002. Review: The Evolutionary Origin and Diversification of Feathers. Quarterly Review of Biology 77:261–295. Prum RO, Dyck J. 2003. A hierarchical model of plumage: morphology, development, and evolution. Journal of Experimental Zoology (Mol Dev Evol) 298:73–90. Prum RO, Williamson S. 2002. Reaction-diffusion models of within-feather pigmentation patterning. Proceedings of the Royal Society B: Biological Sciences 269:781–792. Prum RO. 1999. Development and evolutionary origin of feathers. J. Exp. Zool. 285:291–306. R Development Core Team. 2012. R: A Language and Environment for Statistical Computing. Vienna: R Foundation for Statistical Computing. http://www.R-project.org. Rawles ME. 1939. The Pigment-Forming Potency of Early Chick Blastoderms. Proceedings of the National Academy of Science USA 1–9. Restall R, Rodner C, Lentino M (2007) Birds of Northern South America. Yale University Press, New Haven. Revell LJ, Harmon LJ, Collar DC. 2008. Phylogenetic signal, evolutionary process, and rate. Systematic Biology 57:591–601. Revell LJ. 2011. phytools: an R package for phylogenetic comparative biology (and other things). Methods in Ecology and Evolution 3:217–223. Richardson MK, Hornbruch A, Wolpert L. 1989. Pigment pattern expression in the plumage of the quail embryo and the quail-chick chimaera. Development 107:805– 818. Richardson MK, Hornbruch A, Wolpert L. 1990. Mechanisms of pigment pattern formation in the quail embryo. Development 109:81–89. Richardson MK, Hornbruch A, Wolpert L. 1991. Pigment patterns in neural crest The mechanisms underlying convergent evolution in the plumage patterns of birds 205 chimeras constructed from quail and guinea fowl embryos. Developmental Biology 143:309–319. Riegner MF. 2008. Parallel evolution of plumage pattern and coloration in birds: implications for defining avian morphospace. Condor 110:599–614. Robertson H, Heather B (2005) The Hand Guide to the Birds of New Zealand. Penguin Books, Auckland. Robson C (2005) Birds of Southeast Asia. Princeton University Press, Princeton. Rosenblum EB, Hoekstra HE, Nachman MW. 2004. Adaptive Reptile Color Variation and the Evolution of the MClR Gene. Evolution 58:1794–1808. Roulin A, Altwegg R, Jensen H, Steinsland I, Schaub M. 2010. Sex-dependent selection on an autosomal melanic female ornament promotes the evolution of sex ratio bias. Ecology 13:616–626. Roulin A, Ducrest A-L, Balloux F, Dijkstra C, Riols C. 2003. A female melanin ornament signals offspring fluctuating asymmetry in the barn owl. Proceedings of the Royal Society of London: Biological Science 270:167-171 Roulin A, Ducrest A-L. 2013. Genetics of colouration in birds. Seminars in Cell and Developmental Biology 24:594–608. Roulin A, Emaresi G, Bize P, Gasparini J, Piault R, Ducrest A-L. 2011. Pale and dark reddish melanic tawny owls differentially regulate the level of blood circulating POMC prohormone in relation to environmental conditions. Oecologia 166:913–921. Roulin A, Jungi TW, Pfister H, Dijkstra C. 2000. Female barn owls (Tyto alba) advertise good genes. Proceedings of the Royal Society of London: Biological Science 267:937–941. Roulin A, Riols C, Dijkstra C. 2001. Female plumage spottiness signals parasite resistance in the barn owl (Tyto alba). Behavioral Ecology 12:103-110 Roulin A. 1999a. Delayed maturation of plumage coloration and plumage spottedness in the Barn Owl (Tyto alba). Journal für Ornithologie 140:193-197 The mechanisms underlying convergent evolution in the plumage patterns of birds 206 Roulin A. 1999b. Nonrandom pairing by male barn owls (Tyto alba) with respect to a female plumage trait. Behavioral ecology 10:688–695. Roulin A. 2004. Proximate basis of the covariation between a melanin-based female ornament and offspring quality. Oecologia 140:668–675. Ruscio J, Mullen T. 2012. Confidence Intervals for the Probability of Superiority Effect Size Measure and the Area Under a Receiver Operating Characteristic Curve. Multivariate Behavioral Research 47:201–223. Ruxton GD, Speed MP, Kelly DJ. 2004. What, if anything, is the adaptive function of countershading? Animal behaviour 68:445–451. Saldana-Caboverde A, Kos L. 2010. Roles of endothelin signaling in melanocyte development and melanoma. Pigment Cell and Melanoma Research 23:160–170. Saneyasu T, Nakanishi K, Atsuta H, Ikura A, Kamisoyama H, Hasegawa S, Honda K. 2013. Age-Dependent Changes in the mRNA Levels of Neuropeptide Y, Proopiomelanocortin, and Corticotropin-Releasing Factor in the Hypothalamus in Growing Broiler Chicks. The Journal of Poultry Science 50:364–369. Scott-Samuel NE, Baddeley R, Palmer CE, Cuthill IC. 2011. Dazzle camouflage affects speed perception. PLoS ONE 6:20233. Shine R, Madsen T. 1994. Sexual Dichromatism in Snakes of the Genus Vipera: A Review and a New Evolutionary Hypothesis. Journal of Herpetology 28:114–117. Shine R. 1994. Allometric patterns in the ecology of Australian snakes. Copeia 4:851- 867. Shiojiri N, Niwa T, Wakamatsu K, Ito S, Nakamura A. 1999. Chemical analysis of melanin pigments in feather germs of Japanese Quail Bh (black at hatch) mutants. Pigment Cell Research 12:259–265. Sibley DA (2000) The Sibley Guide to Birds. Alfred Knopf, New York. Simpson K, Day N (2004) Birds of Australia, 7th edn. Princeton University Press, Princeton. The mechanisms underlying convergent evolution in the plumage patterns of birds 207 Sinclair I, Ryan P. 2010. Birds of Africa, South of the Sahara. Random House Struik. Slominski A, Plonka PM, Pisarchik A, Smart JL, Tolle V, Wortsman J, Low MJ. 2005. Preservation of Eumelanin Hair Pigmentation in Proopiomelanocortin-Deficient Mice on a Nonagouti (a/a) Genetic Background. Endocrinology 146:1245–1253. Snowden RJ. 1998. Shifts in perceived position following adaptation to visual motion. Current Biology 8:1343–1345. Soler M, Soler J. 2000. Brood-parasite interactions between great spotted cuckoos and magpies: a model system for studying coevolutionary relationships. Oecologia 125:309–320. Somveille M, Manica A, Butchart SHM, Rodrigues ASL. 2013. Mapping global diversity patterns for migratory birds. PLoS ONE 8:70907. Spottiswoode CN, Stevens M. 2010. Visual modeling shows that avian host parents use multiple visual cues in rejecting parasitic eggs. Proceedings of the National Academy of Sciences USA 107:8672–8676. Spottiswoode CN, Stevens M. 2011. How to evade a coevolving brood parasite: egg discrimination versus egg variability as host defences. Proceedings of the Royal Society B: Biological Sciences 278:3566–3573. Spottiswoode CN. 2010. The evolution of host-specific variation in cuckoo eggshell strength. Journal of Evolutionary Biology 23:1792–1799. Steingrímsson E, Copeland NG, Jenkins NA. 2006. Mouse coat color mutations: From fancy mice to functional genomics. Developmental Dynamics 235:2401–2411. Stern DL, Orgogozo V. 2009. Is genetic evolution predictable? Science 323:746–751. Stevens M, Cuthill IC, Windsor AMM, Walker HJ. 2006. Disruptive contrast in animal camouflage. Proceedings of the Royal Society of London: Biological Sciences 273:2433–2438. Stevens M, Hopkins E, Hinde W, Adcock A, Connolly Y, Troscianko T, Cuthill IC. 2007. Field experiments on the effectiveness of ‘eyespots’ as predator deterrents. Animal behaviour 74:1215–1227. The mechanisms underlying convergent evolution in the plumage patterns of birds 208 Stevens M, Merilaita S. 2009a. Defining disruptive coloration and distinguishing its functions. Philosophical Transactions of the Royal Society B: Biological Sciences 364:481–488. Stevens M, Merilaita S. 2009b. Animal camouflage: current issues and new perspectives. Philosophical Transactions of the Royal Society B: Biological Sciences 364:423–427. Stevens M, Parraga CA, Cuthill IC, Partridge JC, Troscianko TS. 2007. Using digital photography to study animal coloration. Biological Journal of the Linnean Society 1– 27. Stevens M, Searle WTL, Seymour JE, Marshall KL, Ruxton GD. 2011. Motion dazzle and camouflage as distinct anti-predator defenses. BMC Biology 9:81. Stevens M, Yule DH, Ruxton GD. 2008. A new look at Op art: towards a simple explanation of illusory motion. Proceedings of the Royal Society of London: Biological Sciences 275:2639–2643. Stoddard MC, Stevens M. 2010. Pattern mimicry of host eggs by the common cuckoo, as seen through a bird's eye. Proceedings of the Royal Society B: Biological Sciences 277:1387–1393. Stuart-Fox DM, Ord TJ. 2004. Sexual selection, natural selection and the evolution of dimorphic coloration and ornamentation in agamid lizards. Proceedings of the Royal Society B: Biological Sciences 271:2249–2255. Swaddle JP, Cuthill IC. 1994. Female Zebra Finches prefer males with symmetric chest plumage. Proceedings: Biological Sciences 258:267–271. Symonds MRE, Moussalli A. 2010. A brief guide to model selection, multimodel inference and model averaging in behavioural ecology using Akaike’s information criterion. Behaviouarl Ecology and Sociobiology 65:13–21. Tachibana T, Sugahara K, Ohgushi A, Ando R. 2001. Intracerebroventricular injection of agouti-related protein attenuates the anorexigenic effect of alpha-melanocyte stimulating hormone in neonatal chicks. Neuroscience Letters 305:131-134. Takahashi A, Amano M, Amiya N, Yamanome T, Yamamori K, Kawauchi H. 2006. The mechanisms underlying convergent evolution in the plumage patterns of birds 209 Expression of three proopiomelanocortin subtype genes and mass spectrometric identification of POMC-derived peptides in pars distalis and pars intermedia of barfin flounder pituitary. General and comparative Endocrinology 145:280–286. Takeuchi S, Suzuki H, Yabuuchi M, Takahashi S. 1996. A possible involvement of melanocortin 1-receptor in regulating feather color pigmentation in the chicken. Biochimica et Biophysica Acta 1308:164–168. Takeuchi S, Takahashi S, Okimoto R, Schiöth HB, Boswell T. 2003. Avian melanocortin system: alpha-MSH may act as an autocrine/paracrine hormone: a minireview. Annals of the New York Academy of Sciences 994:366–372. Takeuchi S, Takahashi S, Okimoto R. 2003. Avian Melanocortin System: α‐MSH May Act as an Autocrine/Paracrine Hormone. Annals of the New York Academy of Sciences 994:366-372. Takeuchi S, Takahashi S. 1998. Melanocortin receptor genes in the chicken-tissue distributions. General and comparative Endocrinology 112:220–231. Takeuchi S, Takahashi S. 1999. A possible involvement of melanocortin 3 receptor in the regulation of adrenal gland function in the chicken. Biochimica et Biophysica Acta 1448:512-518. Takeuchi S, Teshigawara K, Takahashi S. 1999. Molecular cloning and characterization of the chicken pro-opiomelanocortin (POMC) gene. Biochimica et Biophysica Acta 1450:452–459. Takeuchi S, Teshigawara K, Takahashi S. 2000. Expression of Agouti-related protein (AGRP) in the chicken: a possible involvement of AGRP in regulating peripheral melanocortin systems in the chicken. Biochimica et Biophysica Acta 1496:261–269. Tamate HB, Takeuchi T. 1984. Action of the e locus of mice in the response of phaeomelanic hair follicles to alpha-melanocyte-stimulating hormone in vitro. Science 224:1241–1242. Tamura K, Stecher G, Peterson D, Filipski A, Kumar S. 2013. MEGA6: Molecular Evolutionary Genetics Analysis version 6.0. Molecular Biology and Evolution 30:2725–2729. The mechanisms underlying convergent evolution in the plumage patterns of birds 210 Thayer GH. 1909. Concealing-Coloration in the Animal Kingdom; an Exposition of the Laws of Disguise Through Color and Pattern. London: Macmillan. Theron E, Hawkins K, Bermingham E, Ricklefs RE, Mundy NI. 2001. The molecular basis of an avian plumage polymorphism in the wild: a melanocortin-1-receptor point mutation is perfectly associated with the melanic plumage morph of the bananaquit, Coereba flaveola. Current Biology 11:550–557. Thery M. 2001. Forest light and its influence on habitat selection. Tropical Forest Canopies: Ecology and Management: 1–12. Thompson JN, Cunningham BM. 2002. Geographic structure and dynamics of coevolutionary selection. Nature 417:735–738. Thompson JN, Nuismer SL, Gomulkiewicz R. 2002. Coevolution and maladaptation. Integrative and Comparative Biology 42:381–387. Thompson JN. 2005. Coevolution: The Geographic Mosaic of Coevolutionary Arms Races. Current Biology 15:992–994. Thorogood R, Davies NB. 2012. Cuckoos Combat Socially Transmitted Defenses of Reed Warbler Hosts with a Plumage Polymorphism. Science 337:578–580. Thorogood R, Davies NB. 2013. Hawk mimicry and the evolution of polymorphic cuckoos. Chinese Birds 4:39–50. Tobita-Teramoto T, Jang GY, Kino K, Salter DW, Brumbaugh J, Akiyama T. 2000. Autosomal Albino Chicken Mutation (c a /c a ) Deletes Hexanucleotide (−∆GACTGG817) at a Copper-Binding Site of the Tyrosinase Gene. Poultry Science 79:46–50. Trnka A, Grim Tom. 2013. Color plumage polymorphism and predator mimicry in brood parasites. Frontiers in Zoology 10:1–1. Trnka A, Prokop P, Grim Tomáš. 2012. Uncovering Dangerous Cheats: How Do Avian Hosts Recognize Adult Brood Parasites? PLoS ONE 7:37445. Trnka A, Prokop P. 2012. The effectiveness of hawk mimicry in protecting cuckoos from aggressive hosts. Animal behaviour 83:263–268. The mechanisms underlying convergent evolution in the plumage patterns of birds 211 Tucker VA. 2000. The deep fovea, sideways vision and spiral flight paths in raptors. Journal of Experimental Biology 203:3745–3754. Turing AM. 1952. The Chemical Basis of Morphogenesis. Philosophical Transactions of the Royal Society B: Biological Sciences 237:37–72. van Tuinen M, Sibley CG, Hedges SB. 2000. The early history of modern birds inferred from DNA sequences of nuclear and mitochondrial ribosomal genes. Molecular Biology and Evolution 17:451–457. Vidal O, Araguas RM, Fernández R, Heras S, Sanz N, Pla C. 2010. Melanism in guinea fowl (Numida meleagris) is associated with a deletion of Phenylalanine-256 in the MC1R gene. Animal Genetics 41:656–658. Vrieling H, Duhl DM, Millar SE, Miller KA, Barsh GS. 1994. Differences in dorsal and ventral pigmentation result from regional expression of the mouse agouti gene. Proceedings of the National Academy of Sciences USA 91:5667–5671. Wagner GP, Pavlicev M, Cheverud JM. 2007. The road to modularity. Nature Review of Genetics 8:921–931. Walker WP, Gunn TM. 2009. Piecing together the pigment-type switching puzzle. Pigment Cell and Melanoma Research 23:4–6. Walker WP, Gunn TM. 2010. Shades of meaning: the pigment-type switching system as a tool for discovery. Pigment Cell and Melanoma Research 23:485–495. Wallace AR. 1889. Darwinism - An Exposition of the Theory of Natural Selection - With Some of Its Applications. London: Macmillan. Weibel AC, Moore WS. 2007. Plumage Convergence in Picoides Woodpeckers Based on a Molecular Phylogeny, with Emphasis on Convergence in Downy and Hairy Woodpeckers. Condor 107:797–809. Welbergen JA, Davies NB. 2008. Reed warblers discriminate cuckoos from sparrowhawks with graded alarm signals that attract mates and neighbours. Animal Behaviour 76:811–822. Welbergen JA, Davies NB. 2011. A parasite in wolf's clothing: hawk mimicry reduces The mechanisms underlying convergent evolution in the plumage patterns of birds 212 mobbing of cuckoos by hosts. Behavioral ecology 22:574–579. Welle M, Philipp U, Rufenacht S, Roosje P, Scharfenstein M, Schutz E, Brenig B, Linek M, Mecklenburg L, Grest P, B Drögemüller, B Haase, T Leeb, Drögemüller C. 2009. MLPH Genotype - Melanin Phenotype Correlation in Dilute Dogs. Journal of Heredity 100:S75–S79. West-Eberhard MJ. 2003. Developmental Plasticity and Evolution. New York: Oxford University Press. Westneat DF. 2006. No Evidence of Current Sexual Selection on Sexually Dimorphic Traits in a Bird with High Variance in Mating Success. American Naturalist 167:171– 189. Widelitz RB, Jiang TX, Yu M, Shen T. 2003. Molecular biology of feather morphogenesis - A testable model for evo-devo research. Journal of Experimental Zoology (Mol Dev Evol) 298B:109–122. Willier BH, Rawles ME. 1938. Feather Characterization as Studied in Host-Graft Combinations between Chick Embryos of Different Breeds. Proceedings of the National Academy of Sciences USA 24:446–452. Willier BH, Rawles ME. 1940. The Control of Feather Color Pattern by Melanophores Grafted from One Embryo to Another of a Different Breed of Fowl. Physiological Zoology 13:177–201. Wilson BD, Ollmann MM, Kang L, Stoffel M, Bell GI, Barsh GS. 1995. Structure and function of ASP, the human homolog of the mouse agouti gene. Human Molecular Genetics 4:223–230. Yabuuchi M, Bando K, Hiramatsu M, Takahashi S, Takeuchi S. 2010. Local agouti signaling protein/melanocortin signaling system that possibly regulates lipid metabolism in adipose tissues in chickens. Japan poultry science association 1–7. Yom-Tov Y, Geffen E. 2005. Host Specialization and Latitude among Cuckoos. Journal of Avian Biology 36:465–470. Yoshihara C, Fukao A, Ando K, Tashiro Y, Taniuchi S, Takahashi S, Takeuchi S. 2012. Elaborate color patterns of individual chicken feathers may be formed by the The mechanisms underlying convergent evolution in the plumage patterns of birds 213 agouti signaling protein. General and comparative Endocrinology 175:495–499. Yoshihara C, Tashiro Y, Taniuchi S, Katayama H, Takahashi S, Takeuchi S. 2011. Feather follicles express two classes of pro-opiomelanocortin (POMC) mRNA using alternative promoters in chickens. General and comparative Endocrinology 171:46– 51. Yu M, Yue Z, Wu P, Wu DY, Mayer JA. 2004. The developmental biology of feather follicles. Developmental Biology 1–11. Voipio P. 1953. The hepaticus variety and the juvenile plumage types of the Cuckoo. Omis Fennica, 30, 47. 30. The mechanisms underlying convergent evolution in the plumage patterns of birds 214 Supplementary material Table S2.1. Primer sequences tested for RT-PCR. Target sequence Name Sequence AGRP AGRP F1 ACA CAG GAC CAT GCT GAA CG AGRP AGRP F2 GAC ATC GAG GCA GAG CGA CT AGRP AGRP R1 CCT CTG CGG CTC CAA TAA AG AGRP AGRP R2 CCT CTG CGG CTC CAA TAA AG AGRP AGRPF1 CCC AGG ACC ATG CTG AAC AGRP AGRPF1 CCC AGG ACC ATG CTG AAC AGRP AGRPF16 GTG GAC CAT GAG CCT CCT CT AGRP AGRPF17 CCA GCT CCC CAG CAC TAA G AGRP AGRPF4 ATG CTG AAC GTG CTG CTG AGRP AGRPR1 CCA CAT GGG AAG GTG GTG AGRP AGRPR16 GGA TTC TCC TCT TCC CAT CC AGRP AGRPR17 TGA GCA CAA TGG ACC TAT GG AGRP AGRPR4 AGG TGG TGC TGA TCT TCC TG ASIP E2 ASIP E2_R CAG CCT TAA CAT GTT CCT CAT TA ASIP E3 ASIP E3R AGT AAA CAC TGG CAG ATT GTC TGA ASIP E4 ASIP E4R CAG CCT TAA CAT GTT CCT CAT TAG GTT ASIP NE1 ASIP NEW EXON 1F GGG AGA TCT GGG AGG TTC ATT ASIP ASIP R TTT GGG GGT GTC TTC AGT TC ASIP ASIP R2 CCT TAA CAT GTT CCT CAT TAG GTT TA ASIP ASIPF5 CCA ACA ATG AAA AGG AAG AAC C ASIP ASIPR5 GAT TTG GTT TAA CAC TTT GGG TTT ASIP ASIPF2 TCA TTT TCA TGA CAG TGG GAT T ASIP ASIPR6 TTT GGG GGT GTC TTC AGT TC ASIP E1L ASIP E1LF TCT CCT CGG CTA TAT GGC TGA G ASIP E1L ASIP E1LF AAG CCA GAA CTG GTG GTC AA ASIP E1L ASIP E1LF_2.0 AGT TTT GGA GGT TCA TTT CTA ATG T ASIP E2 ASIP E2F TCA TTT TCA TGA CAG TGG GAT T ASIP E2 ASIP E2F_2.0 TAA ACA CAT TGA TGG CAT TAA CAA ASIP E3 ASIP E3F GAA GCA GGC AGT CTT CTT GG ASIP E3 ASIP E3F TGA AAA GGA AGC AGA ACC AGA ASIP E4 ASIP E4F GTT CTT TTG GCT CAG TGG TAT CTC A ASIP E4 ASIP E4F CCA GCA TTT TCA TAT TTT CTG GA ASIP E5 ASIP E5F TGA AAT CAG TTG TGG CAG GAA The mechanisms underlying convergent evolution in the plumage patterns of birds 215 ASIP NEW EXON 1 ASIP NEW EXON 1F GTG TGG TTG TGA TGG TGA TGG ASIP INTERNAL 1 INT1 AGC AGC TCC AGC CTT TCA TC ASIP INTERNAL 2 INT2 TCT GGC GTA ACT GGA ACA GC PC1 PC1 F ACG GGC TGG AAA TTC AGG AT PC1 PC1 R AAC CCA AAT CGG CTG TTG AC PC1 PC1F.1 CTA CGC CAA CTA TGA TCC AAG G PC1 PC1F.2 GGG ACT CAC ACT GGG ACC TC PC1 PC1R.1 TTT CCA TCT TTT GGG ATC AGC PC1 PC1R.2 CCA AAT CCA AAT CGG CTG TT PC2 PC2F GGG AGG GAA AGG AAG CAT CT PC2 PC2F.1 GCT GGG ATA CAC AGG GAA GG PC2 PC2F.2 GGA GAG ACA TGC AGC ATC TGA PC2 PC2R GGT CTT CTC CCC AAG TGT GTG PC2 PC2R.1 CAT AGC TGG CTT TGG CAT TG PC2 PC2R.2 CCC AAC GCA GTG GAA TCT CT PC2 X2 PC2_X2F.1 CAG CCG TCT ACA CCA ACC AG PC2 X2 PC2_X2F.2 CCT TCG TCC TCC TCC TT PC2 X2 PC2_X2R.1 TGA ATG TGG AGG CTA GGG TTG PC2 X2 PC2_X2R.2 GGA AGC CCA GAC GTA GAT GC PC2 X3 PC2_X3F.1 GGC ATT GCC AAG GTC AGA A PC2 X3 PC2_X3F.2 CCT TCG TCC TCC TCC TCC TT PC2 X3 PC2_X3R.2 GCC CAG ACG TAG ATG CTT CC PC2 X3 PC2_X3 R TGT AGC CAC ACC AGC CTC TG PC2 X4 PC2_X4 F.3 GCA TTG CCA AGG TCA GAA GA PC2 X4 PC2_X4 R GGA AGC CCA GAC GTA GAT GC POMC POMCF CTG GGG CTG CTG CTG CTG TGT POMC POMCF.1 GTA TCC CAA TGG CGT GGA TG POMC POMCR TGA CCC TTC TTG TAG GCG CTT T POMC POMCR.1 CAG AGT CAT CAG CGG GGT CT POMC POMCR.2 CAT GGG GTA ACT CTC AGC CGA CT POMC A PROMOTER POMC A PROMOTER CCC ATA AGC GAC TTG CCT TC POMC A PROMOTER POMC A PROMOTER.FV2 CCC ATA AGC GAC TTG CCT TC POMC A PROMOTER POMC A-1 CAC CCC TCG CCA GTA GGT T POMC A PROMOTER R POMC AF CAA AGA ACT GAC CAT CCA CCA CAT T POMC A PROMOTER R POMC_AR.1 GGT ACG AGC CAC CAT CCT TC POMC A PROMOTER R POMC_AR.2 CCT CAC CCT CCT CCT CCT CT POMC A PROMOTER R POMC_AR.3 CCT CAC CCT CCT CCT CCT CT POMC B PROMOTER POMC B PROMOTER GGA GAC GGG GAA GGT GGT The mechanisms underlying convergent evolution in the plumage patterns of birds 216 POMC B PROMOTER POMC B-1 CAC CCC TCG CCA GTA GGT T POMC B PROMOTER POMC T1F.V2 GCC ACT GAG GCT GGA GTT TT POMC B PROMOTER POMC T1R CCC CTC ACT GAC CCT TCT TG POMC B PROMOTER F POMC BF GAA GGT GGT GGC TGC GCT CCA A POMC B PROMTER POMC T1 AGC GCT CCT CTG CAG TTT G POMC CHICKEN NCE POMC A-1F.V2 CAC CCC TCG CCA GTA GGT T POMC CHICKEN NCE POMC A-2 CTC AGG AGG GGC AGA AAT CC POMC CHICKEN NCE POMC A-2F.V2 CTC TCC CCC TGC AGC ATC POMC CHICKEN NCE POMC B-2 CTC TCC CCC TGC AGC ATC POMC QUAIL NCE 1 POMC QE R ATC TCC CTC CGG AAC TCC AT POMC QUAIL NCE 1 POMC QEF.2 ACT TCC AGC GTC TCC CAG AG POMC QUAIL NCE 1 POMC QEF.1 TCT CTT GCC TGT GGC TCT CA POMC QUAIL NCE 2 POMC QEF.3 GAT TTC GGA GGC AAA GGA TG The mechanisms underlying convergent evolution in the plumage patterns of birds 217 Fig. S3.1a. Distribution of plumage patterns per patch of plumage in Anseriformes. Empty boxes indicate that the type of pattern is unknown and/or is mottled plumage. Each type of pattern is found in extant species and although there is some variation in the most probable ancestral state, where there is support it is for an absence of patterns. Na pe Ru mp Ta il Br ea st Fla nk Ve nt Wi ng Na pe Ru mp Ta il Br ea st Fla nk Ve nt Wi ng The mechanisms underlying convergent evolution in the plumage patterns of birds 218 Fig. S3.1b. Distribution of plumage patterns per patch of plumage in Galliformes. Empty boxes indicate that the type of pattern is unknown and/or is mottled plumage. Each type of plumage pattern is found in extant species and the most probable ancestral state of plumage is an absence of patterns. Na pe Ru mp Ta il Br ea st Fla nk Ve nt Wi ng The mechanisms underlying convergent evolution in the plumage patterns of birds 219 GALLIFORMES ANSERIFORMES UPPERPARTS Nape Wing Rump The mechanisms underlying convergent evolution in the plumage patterns of birds 220 Tail UNDERPARTS Breast Flank The mechanisms underlying convergent evolution in the plumage patterns of birds 221 Vent Fig. S3.2. Local pattern evolution within patches, in Anseriformes and Galliformes. The width of each evolutionary step is proportional to the average rate in the top model set. Beside each evolutionary step is the marginal probability of each transition not occurring, followed by the marginal probability of it occurring. Where the transition probably does not occur, the transition line is grey. Conversely, where the transition probably does occur, the transition line is black. Equivocal transitions, where the marginal probability is =< 0.05 difference between not occurring and occurring, are indicated by a grey dashed line. The mechanisms underlying convergent evolution in the plumage patterns of birds 222 Scales are derived ANSERIFORMES GALLIFORMES ANTEROPOSTERIOR AXIS Anterior Middle Spots are derived The mechanisms underlying convergent evolution in the plumage patterns of birds 223 Scales are derived Scales are derived Posterior DORSOVENTRAL AXIS Dorsal Spots are derived The mechanisms underlying convergent evolution in the plumage patterns of birds 224 Scales are derived Ventral Spots are derived Fig. S3.3. Regional pattern evolution within regions in Anseriformes and Galliformes. To examine the effects of uncertainty in the order of plumage pattern evolution in Galliformes we modeled the effect of scales and spots being more derived. The width Spots are derived The mechanisms underlying convergent evolution in the plumage patterns of birds 225 of each evolutionary step is proportional to the average rate in the top model set. Beside each evolutionary step is the marginal probability of each transition not occurring, followed by the marginal probability of it occurring. Where the transition probably does not occur, the transition line is grey. Conversely, where the transition probably does occur, the transition line is black. Equivocal transitions, where the marginal probability is =< 0.05 difference between not occurring and occurring, are indicated by a grey dashed line. The mechanisms underlying convergent evolution in the plumage patterns of birds 226 Scales are derived WHOLE BODY ALL Fig. S3.4. Pattern evolution over the whole body, in Anseriformes and Galliformes. To examine the effects of uncertainty in the order of plumage pattern evolution in Galliformes we modeled the effect of scales or spots being more derived. The width of each evolutionary step is proportional to the average rate in the top model set. Beside each evolutionary step is the marginal probability of each transition not occurring, followed by the marginal probability of it occurring. Where the transition probably does not occur, the transition line is grey. Conversely, where the transition probably does occur, the transition line is black. Equivocal transitions, where the marginal probability is =< 0.05 difference between not occurring and occurring, are indicated by a grey dashed line. Spots are derived The mechanisms underlying convergent evolution in the plumage patterns of birds 227 Table S3.1. Prior probability of encountering models of <12 parameters. Calculations are comprised of binomial for Z (where Z = n parameters that are set to 0 i.e. do not occur) and Bell numbers for models with 12 possible transition rates. No. Zeros (z) Binomial for z 12 - z Bell number for 12 - z Binomial for z x bell number for 12 - z Prior 0 1 12 4213597 4213597 0.1524211599 1 12 11 678570 8142840 0.0245463499 2 66 10 115975 7654350 0.0041952384 3 220 9 21147 4652340 0.0007649641 4 495 8 4140 2049300 0.0001497589 5 792 7 877 694584 0.0000317243 6 924 6 203 187572 0.0000073432 7 792 5 52 41184 0.0000018810 8 495 4 15 7425 0.0000005426 9 220 3 5 1100 0.0000001809 10 66 2 2 132 0.0000000723 11 12 1 1 12 0.0000000362 The mechanisms underlying convergent evolution in the plumage patterns of birds 228 Table S3.2. Prior probability of encountering models of <6 parameters. Calculations were comprised of binomial for Z (where Z = n parameters that are set to 0 i.e. do not occur) and Bell numbers for models with three pattern states encompassing 6 possible transitions. No of zeroes Binomial for z 6-z Bell number for 6-z Binomial for z x bell number for 6-z Prior 0 1 6 203 203 0.23173516 1 6 5 52 312 0.04577465 2 15 4 15 225 0.01381216 3 20 3 5 100 0.00514933 4 15 2 2 30 0.00221239 5 6 1 1 6 0.00113507 The mechanisms underlying convergent evolution in the plumage patterns of birds 229 Table S3.3. The number of unique models as well as the ancestral state of patterns per body region in Anseriformes and Galliformes. Ancestral state Unique models Average probability Marginal probability Posterior sample Top model set (BF>=2) Full model Absence of patterns Scales Bars Spots Absence of patterns Scales Bars Spots Anseriformes                      Whole body 142 88 98: BF = 0.34 0.3 0.24 0.27 0.2 0.11; 0.75 0.84; 0.02 0.79; 0.08 0.85; 0.01 Anterior 219 158 35: BF = 0.12 0.3 0.25 0.22 0.23 0.23; 0.56 0.79; 0.00 0.72; 0.07 0.64; 0.15 Middle 216 144 13: BF = 0.04 0.25 0.25 0.25 0.25 0.31; 0.59 0.89; 0.01 0.62; 0.28 0.87; 0.03 Posterior 199 140 2: BF = 0.00 0.25 0.25 0.25 0.25 0.34; 0.59 0.92; 0.00 0.64; 0.29 0.88; 0.05 Ventral 181 190 43: BF = 0.14 0.23 0.27 0.22 0.28 0.37; 0.41 0.71; 0.07 0.74; 0.04 0.52; 0.26 Dorsal 273 191 23: BF = 0.08 0.28 0.25 0.22 0.24 0.05; 0.79 0.82; 0.02 0.82; 0.02 0.82; 0.02 Galliformes      Whole body (scales) 114 76 7: BF = 0.023 0.88 0.05 0.05 0.02 0.00; 0.92 0.92; 0.00 0.92; 0.00 0.92; 0.00 Whole body (spots) 95 63 6: BF = 0.020 0.84 0.09 0.06 0.01 0.00; 0.92 0.00; 0.00 0.00; 0.00 0.00; 0.00 Anterior (scales) 149 99 11: BF = 0.04 0.72 0.13 0.09 0.06 0.00; 0.92 0.00; 0.00 0.00; 0.00 0.00; 0.00 Anterior (spots) 139 89 12: BF = 0.04 0.71 0.14 0.1 0.05 0.00; 0.93 0.00; 0.00 0.00; 0.00 0.00; 0.00 Middle 129 82 14: BF = 0.05 0.84 0.06 0.07 0.03 0.00; 0.87 0.00; 0.00 0.00; 0.00 0.00; 0.00 The mechanisms underlying convergent evolution in the plumage patterns of birds 230 Posterior (scales) 145 112 1: BF = 0.003 0.82 0.09 0.07 0.02 0.01; 0.97 0.97; 0.01 0.97; 0.01 0.97; 0.01 Posterior (spots) 129 92 0 0.81 0.1 0.07 0.02 0.01; 0.96 0.96; 0.01 0.96; 0.01 0.96; 0.01 Ventral (scales) 186 113 112: BF = 0.38 0.76 0.11 0.07 0.06 0.00; 0.74 0.00; 0.00 0.00; 0.00 0.00; 0.00 Ventral (spots) 203 122 86: BF = 0.28 0.74 0.13 0.07 0.06 0.00; 0.74 0.00; 0.00 0.00; 0.00 0.00; 0.00 Dorsal (scales) 129 84 0 0.79 0.1 0.08 0.03 0.00; 0.93 0.93; 0.00 0.93; 0.00 0.93; 0.00 Dorsal (spots) 106 73 5: BF = 0.02 0.78 0.11 0.09 0.02 0.00; 0.94 0.94; 0.00 0.94; 0.00 0.94; 0.00 The mechanisms underlying convergent evolution in the plumage patterns of birds 231 No rm al is ed e ne rg y Filter size The mechanisms underlying convergent evolution in the plumage patterns of birds 232 Fig. S4.1. Granularity spectra of the ventral and flank area of each species sampled. Black lines indicate ventral patterns, and grey lines indicate the flank. To assess the variability of patterning within each patch of patterning sampled, we compared the top, middle and bottom section for the ventral and flank area respectively. A continuous line represents the top section, a dashed line represents the middle, and the dotted line represents the bottom section. Filter size No rm al is ed e ne rg y The mechanisms underlying convergent evolution in the plumage patterns of birds 233 Fig. S4.2a. Median pattern attributes for Eudynamys scolopacea. Confidence intervals are bootstrapped values and the grey line indicates the median of Eudynamys scolopacea. The mechanisms underlying convergent evolution in the plumage patterns of birds 234 Fig. S4.2b. Median plumage pattern attributes for Chrysococcyx flavigularis. Confidence intervals are bootstrapped values and the grey line indicates the median of Chrysococcyx flavigularis. The mechanisms underlying convergent evolution in the plumage patterns of birds 235 Fig. S4.2c. Median plumage pattern attributes for Cacomantis sonneratii. Confidence intervals are bootstrapped values and the grey line indicates the median of Cacomantis sonneratii. The mechanisms underlying convergent evolution in the plumage patterns of birds 236 Fig. S4.2d. Median plumage pattern attributes for Cercococcyx mechowi. Confidence intervals are bootstrapped values and the grey line indicates the median of Cercococcyx mechowi. The mechanisms underlying convergent evolution in the plumage patterns of birds 237 Fig. S4.2e. Median plumage pattern attributes for Cuculus saturates. Confidence intervals represent bootstrapped values and the grey line indicates the median of Cuculus saturatus. The mechanisms underlying convergent evolution in the plumage patterns of birds 238 Table S4.1. Comparison of distribution overlap from a taxonomic source (Payne and Sorensen 2005; Christie and Ferguson-Lees 2010) with a regional field guide (Sinclair and Ryan 2010). Chrysococcyx flavigularis Cercococcyx mechowi Taxonomic source Field guide Taxonomic source Field guide Aviceda cuculoides <1 <1 <1 <1 Polyboroides typus <1 <1 <1 <1 Micronisus gabar <0.25 <0.25 <0.50 <0.25 Accipiter tachiro <1 <0.25 <0.75 <0.25 Accipiter ovampensis <0.25 <0.25 <0.25 <0.25 Kaupifalco monogrammicus <1 <1 <1 <1 The mechanisms underlying convergent evolution in the plumage patterns of birds 239 Table S4.2. Cuckoo and raptor host prey species (Brown et al. 1982; Del Hoyo et al. 1994; Payne 2005). CUCKOOS Species Host/prey Eudynamys scolopacea Crows (Corvus splendens, C. macrorhynchos, C. enca, C. florensis, Urocissa erythrorhyncha) Drongo (Drongo macrocercus) Common starlings Mynas (Acridotheres tristis, A. grandis, Gracula religious) Chrysococcyx flavigularis Tit-flycatcher (Myioparus griseigularis) Cacomantis sonneratii Loras (Aegithina tiphia, A. viridissima) Bulbuls (Pycnonotus jocose) Yuhinia (Yuhinia zantholeuca) Cuckoo-shrike (Pericrococtus flames) Babblers (Stachyris) Cercococcyx mechowi Illadopsis (Trichastoma fulvescens) Forest robin (Stiphrornis erythrothorax) Monarch-flycatcher (Trochocercus nitens) Cuculus saturatus Warblers (genus Seicercus, Phylloscopus occipitals, P. reguloides) RAPTORS Aviceda cuculoides Small birds Aviceda leuphotes Asian palm swift (Cypsiurus batasiensis) Pernis ptilorhyncus Small birds (Passerines) Polyboroides typus Weavers Swifts Herons Barbet Roller Kingfisher Birds eggs - sparrow to darter (Anhinga melanogaster). Micronisus gabar Pipits Weaver birds Starlings The mechanisms underlying convergent evolution in the plumage patterns of birds 240 Barbets Thrush Coucals Francolins Accipiter trivirgatus Small birds Pigeon (Treron) Accipiter tachiro Doves (Streptopelia semitorquata) Cuckoo (Chrysococcyx klaas) Mousebirds Trogon Hornbills Accipiter soloensis Pigeon (Columba livia) Accipiter fasciatus Oriole (Oriolus chinensis) Pigeons Ducks Herons Rails Poultry Accipiter gularis Fruit doves (Ptilinopus) Sparrows (Passer montanus) Buntings (Emberiza) Tits (Parus) Warblers Nuthatches (Sitta) Magpie (Cyanopica cyanea) Accipiter virgatus Gamebirds (nudifigous young Gallus) Warblers Thrushes Barbets Accipiter ovampensis Doves Bee-eaters Hoopoes Woodpeckers Pipits Weaver birds The mechanisms underlying convergent evolution in the plumage patterns of birds 241 Prinia Kaupifalco monogrammicus Rarely small birds Spizaetus nanus Young or injured birds including Blackbirds (Turdus merula) Some birds The mechanisms underlying convergent evolution in the plumage patterns of birds 242 Table S4.3a. The individual types of habitat inhabited by cuckoos and sympatric raptors of Africa (Payne and Sorensen 2005; Christie and Ferguson-Lees 2010). Habitat Chrysococcyx flavigularis Cercococcyx mechowi Aviceda cuculoides Polyboroides typus Micronisus gabar Accipiter tachiro Accipiter ovampensis Kaupifalco monogrammicus Primary forest Yes - Yes - - Yes - Yes Secondary and gallery forest Yes - - Yes - Yes - - Lowlands Yes - - - - Yes - - Lowland mature forest - Yes - - - Yes - - Montane forest - Yes - - - - - - Forest edge - Yes - Yes - - - Yes Riverside gallery - - Yes Yes Yes Yes - Yes Savannah - - Yes Yes Yes Yes Yes Yes Eucalyptus and pine plantations - - Yes Yes - Yes Yes - Suburban gardens - - Yes - Yes Yes - Yes Dry woodland - - - - - Yes Yes Yes Mountain woodland - - - - - Yes - - The mechanisms underlying convergent evolution in the plumage patterns of birds 243 Tropical rainforest - - - - - Yes - - Mangroves - - - - - Yes - - Broadleaf woodland - - - - - - - Yes Open woodland - - - - Yes - - - Moist woodland - - - Yes - - - - Hill country - - - Yes - - - - The mechanisms underlying convergent evolution in the plumage patterns of birds 244 Table S4.3b. The individual types of habitat inhabited by cuckoos and sympatric raptors of Oceania (Ferguson-Lees and Christie 2001; Payne 2005). Habitat Eudynamys scolopacea Cacomantis sonneratii Cuculus saturatus Aviceda leuphotes Pernis ptilorhyncus Accipiter trivirgatus Accipiter soloensis Accipiter fasciatus Accipiter gularis Accipiter virgatus Spizaetus nanus Primary forest Yes Yes Yes Yes - Yes - Yes - - Yes Lowlands Yes Yes - - Yes Yes Yes - Yes - Yes Forest edge Yes Yes - Yes - - - Yes - - - Secondary forest Yes Yes - Yes - Yes - Yes Yes Yes Yes Remnant woodlands with large trees Yes - - - - - Yes - - - - Monsoon forest Yes - - - - - - - - - Riverine scrub Yes - - - - - Yes - - Yes - Cultivated/plantations Yes Yes - Yes Yes Yes Yes Yes Yes Yes - Heath forest Yes - - - - - - Yes - - - Gardens Yes Yes - Yes - - Yes Yes - - - Mangrove/swamp Yes - - Yes - - Yes - Yes - - Coniferous-deciduous forest - - Yes Yes - Yes - - Yes Yes - Larch taiga - - Yes - - - - - - - - Broad-leafed forest - Yes Yes - Yes - - - - - - The mechanisms underlying convergent evolution in the plumage patterns of birds 245 Thickets - - Yes - - - - - - - - Oak rhododendron forests - - Yes - - - - - - - - Subtropical woodlands - - Yes - - - Yes - - - - Mountain forest - - - - - Yes - - Yes Yes - Savannah - - - - - - - Yes - - - Rainforest - - - - Yes - - - - - - Coastal plains Yes - - - - - - - - - - Wooded hills Yes - - - - - - - - - - The mechanisms underlying convergent evolution in the plumage patterns of birds 246 Table. S5.1. A representative sample of literature of pattern type and pattern function in different contexts (stationary or moving) in animals. Literature cited covers birds, mammals, fish, reptiles, insects, cephalopods, crustaceans, and amphibians. Literature is categorized as 1) Hypotheses (theoretical hypotheses), 2) Experimental (direct empirical evidence), 3) Correlational/observational, 4) Comparative phylogenetic analyses, and 5) Methodological (proposing and/or testing new methods). Camouflage mechanisms cover the following: Stationary camouflage (Background matching [BM], disruptive camouflage [DC]) and motion camouflage (motion-dazzle [MD]), flicker-fusion [F-F]). Pattern type Behavioural context Camouflage Communication Camouflage mechanism Prey or signaller (where tested) Predator or signal receiver (where tested) References Type of literature/ evidence Irregular Stationary X BM N/A N/A (Poulton, 1890) Hypotheses Stationary X BM and DC N/A N/A (Thayer, 1909) Hypotheses Stationary X DC N/A N/A (Cott, 1940) Hypotheses Stationary X BM and DC Juvenile cuttlefish (Sepia officinalis) N/A (Hanlon & Messenger, 1988) Experimental Stationary X DC Frogs (Limnodynastes tasmaniensis) Garter snake (Thamnophis sirtalis) (Osorio & Srinivasan, 1991) Experimental Stationary X BM Aythya and Somateria ducks N/A (Hohman et al., 1992) Hypotheses The mechanisms underlying convergent evolution in the plumage patterns of birds 247 Pattern type Behavioural context Camouflage Communication Camouflage mechanism Prey or signaller (where tested) Predator or signal receiver (where tested) References Type of literature/ evidence Stationary X DC Marine isopod Idotea baltica (white- spotted phenotype albafusca) N/A (Merilaita, 1998) Experimental Stationary X BM Artificial paper moths Captive trained birds (Merilaita et al., 2001) Experimental Irregular Stationary Stationary X X BM and DC BM and DC Cuttlefish (Sepia officinalis) Cuttlefish (Sepia officinalis) Fish predators Fish predators (Akkaynak et al., 2013; Chiao et al., 2013; Hanlon et al., 2013) E.g. (Chiao & Hanlon, 2001; Barbosa et al., 2004; Chiao et al., 2005; Barbosa et al., 2007; Chiao et al., 2007; Mäthger et al., 2007; Shohet et al., 2007; Barbosa et al., 2008; Mäthger Experimental Experimental The mechanisms underlying convergent evolution in the plumage patterns of birds 248 Pattern type Behavioural context Camouflage Communication Camouflage mechanism Prey or signaller (where tested) Predator or signal receiver (where tested) References Type of literature/ evidence et al., 2008; Chiao et al., 2010; Chiao et al., 2011; Akkaynak et al., 2013; Chiao et al., 2013; Hanlon et al., 2013) S Stationary X BM Artificial paper moths Wild birds (Cuthill et al., 2005) Experimental Stationary X BM and DC Artificial paper moths Captive trained birds (Merilaita & Lind, 2005) Experimental Stationary X DC Artificial paper moths Wild birds (Stevens & Cuthill, 2006) Experimental Stationary X DC Artificial paper moths Wild birds (Stevens et al., 2006) Experimental Irregular Stationary X BM and DC Artificial paper Wild birds (Cuthill et al., 2006) Experimental The mechanisms underlying convergent evolution in the plumage patterns of birds 249 Pattern type Behavioural context Camouflage Communication Camouflage mechanism Prey or signaller (where tested) Predator or signal receiver (where tested) References Type of literature/ evidence moths Stationary X BM and DC Artificial paper moths (modelled on Thyatira batis) Wild birds (Schaefer & Stobbe, 2006) Experimental Stationary X BM Fiddler crabs (Uca vomeris) Wild and dummy birds (Hemmi et al., 2006) Experimental Stationary X DC Computer-generated moth images Humans (Fraser et al., 2007) Experimental Stationary X DC (surface disruption) Artificial paper moths Wild birds (Stevens et al., 2009) Experimental Moving and stationary X DC/MD Cuttlefish (Sepia officinalis) N/A (Zylinski et al., 2009) Experimental Stationary X BM and DC Artificial paper moths Captive trained birds (Dimitrova & Merilaita, 2010) Experimental Stationary X BM Octopuses (Octopus cyanea & O.vulgaris) N/A (Josef et al., 2012) Experimental The mechanisms underlying convergent evolution in the plumage patterns of birds 250 Pattern type Behavioural context Camouflage Communication Camouflage mechanism Prey or signaller (where tested) Predator or signal receiver (where tested) References Type of literature/ evidence Stationary X DC Computer-generated moth images Humans (Troscianko et al., 2013) Experimental Stationary X BM and DC Computer-generated moth images Humans Webster et al., 2013) Experimental Moving and stationary X BM and DC Computer-generated stimuli Humans (Hall et al., 2013) Experimental Irregular Stationary X BM Japanese quail (Coturnix japonica) N/A (Lovell et al., 2013) Experimental Stationary X BM and DC Moths (Jankowskia fuscaria) N/A (Kang et al., 2012, 2013a, 2013b; Kang et al., 2014) Experimental N/A X N/A Red-legged partridge (Alectoris rufa) Red-legged partridge (Alectoris rufa) (Pérez-Rodríguez et al., 2013) Experimental Stationary X BM Artificial paper Captive trained (DImitrova & Experimental The mechanisms underlying convergent evolution in the plumage patterns of birds 251 Pattern type Behavioural context Camouflage Communication Camouflage mechanism Prey or signaller (where tested) Predator or signal receiver (where tested) References Type of literature/ evidence moths birds Merilaita, 2014) Stationary X BM Artificial paper moths Captive trained birds (Merilaita & DImitrova, 2014) Experimental Irregular TOTAL Stationary = 42 Moving & stationary = 2 44 1 BM = 35 DC = 33 MD = 1 9 species: birds, insects, cephalopods, crustaceans, amphibians 4 groups: birds, fish, snakes, humans Ca. 45+ Experimental = 41/45 (91%) Hypotheses = 4/45 (8%) Pattern type Behavioural context Camouflage Communication Camouflage mechanism Prey or signaller (where tested) Predator or signal receiver (where tested) References Type of literature/ evidence Regular Moving X F-F Brightly-coloured snakes, e.g. coral snakes N/A (Pough, 1976) Hypotheses Moving X F-F North American N/A (Jackson et al., Correlational/ The mechanisms underlying convergent evolution in the plumage patterns of birds 252 Pattern type Behavioural context Camouflage Communication Camouflage mechanism Prey or signaller (where tested) Predator or signal receiver (where tested) References Type of literature/ evidence snakes (132 spp.) 1976) observational Regular Moving X F-F Guppies (Poecilia reticulata) N/A (Endler, 1980) Experimental Moving X MD and F-F Garter snakes (Thamnophis ordinoides) N/A (Brodie, 1989, 1992, 1993) but see Allen et al. 2013) Correlational/ observational N/A X BM Tiger (Panthera tigris) N/A (Godfrey et al., 1987) Experimental N/A X N/A Peafowl (Pavo cristatus) Peafowl (Pavo cristatus) (Petrie et al., 1991) Correlational/ observational N/A X BM Zebra, tigers N/A (Kiltie et al., 1994) Methodological Moving X F-F Vipera snakes N/A (Shine & Madsen, 1994) but see Allen et al. (2013) Hypotheses The mechanisms underlying convergent evolution in the plumage patterns of birds 253 Pattern type Behavioural context Camouflage Communication Camouflage mechanism Prey or signaller (where tested) Predator or signal receiver (where tested) References Type of literature/ evidence N/A X N/A Zebra finches (Taeniopygia guttata) Zebra finches (Taeniopygia guttata) (Swaddle & Cuthill, 1994) Experimental Moving X F-F Adders (Vipera berus) N/A (Lindell & Forsman, 1996) Correlational/ observational N/A X N/A Mallard ducks (Anas platyrhynchos) Mallard ducks (Anas platyrhynchos) (Omland, 1996) Correlational/ observational N/A X BM Mammalian carnivores N/A (Ortolani, 1999) Comparative phylogenetic Regular N/A X N/A Barn owls (Tyto alba) Barn owls (Tyto alba) (Roulin, 1999) Experimental N/A X N/A Artiodactyls N/A (Stoner et al., 2003) Comparative phylogenetic N/A X N/A N/A N/A (Kenward et al., 2004) Comparative phylogenetic Moving and X X Private UV Damselfish N/A (Siebeck, 2004) Experimental The mechanisms underlying convergent evolution in the plumage patterns of birds 254 Pattern type Behavioural context Camouflage Communication Camouflage mechanism Prey or signaller (where tested) Predator or signal receiver (where tested) References Type of literature/ evidence stationary signals (Pomacentrus amboinensis) Stationary X DC Clay models of adders (Vipera latastei gaditana) Wild birds (Niskanen & Mappes, 2005) Experimental N/A X N/A Red-legged partridge (Alectoris rufa) Red-legged partridge (Alectoris rufa) (Bortolotti et al., 2006) Correlational/ observational Moving X MD Cuttlefish (Sepia officinalis) N/A (Shohet et al., 2006) Experimental Moving X MD and F-F Computer-generated moving stimuli Humans (Stevens et al., 2008) Experimental Stationary X DC Artificial paper butterflies (modelled on Limenitis camilla) Wild birds (Stobbe & Schaefer, 2008) Experimental The mechanisms underlying convergent evolution in the plumage patterns of birds 255 Pattern type Behavioural context Camouflage Communication Camouflage mechanism Prey or signaller (where tested) Predator or signal receiver (where tested) References Type of literature/ evidence Stationary X Coincident disruptive camouflage Artificial paper moths/computer- generated stimuli Wild birds/humans (Cuthill & Székely, 2009) Experimental N/A X N/A N/A N/A (Gluckman & Cardoso, 2009) Methodological and Experimental Regular N/A X N/A Barn owls (Tyto alba) Barn owls (Tyto alba) (Roulin et al., 2010) Correlational/ observational, Experimental and Comparative phylogenetic N/A X BM Felidae N/A (Allen et al., 2011) Comparative phylogenetic N/A X N/A Barred buttonquails (Turnix suscitator) Barred buttonquails (Turnix suscitator) (Muck & Goymann, 2011) Correlational/ observational The mechanisms underlying convergent evolution in the plumage patterns of birds 256 Pattern type Behavioural context Camouflage Communication Camouflage mechanism Prey or signaller (where tested) Predator or signal receiver (where tested) References Type of literature/ evidence Moving X MD and F-F Computer-generated moving stimuli Humans (Scott-Samuel et al., 2011) Experimental Stationary X BM The least killfish (Heterandria formosa) Predatory fish (Kjernsmo & Merilaita, 2012) Experimental Moving X MD Computer-generated moving stimuli Humans (von Helversen et al., 2013) Experimental Moving X MD Computer-generated moving stimuli Locusts (Schistocerca gregaria) (Santer, 2013) Experimental Moving X MD Computer-generated moving stimuli Motion detection algorithm (How & Zanker, 2013) Experimental Regular TOTAL Stationary = 4 Moving = 14 24 10 F-F = 8 MD = 7 BM = 5 Ca. 13 species: birds, reptiles, cephalopods, fish, 4 groups: fish, insects, birds, humans Ca. 33+ Experimental = 17/33 (52%) Correlational The mechanisms underlying convergent evolution in the plumage patterns of birds 257 Pattern type Behavioural context Camouflage Communication Camouflage mechanism Prey or signaller (where tested) Predator or signal receiver (where tested) References Type of literature/ evidence Moving & stationary = 1 DC = 3 Private signals = 1 mammals comparative = 14/33 (42%) Hypotheses = 6% Irregular and regular N/A X BM Juvenile plaice (Pleuronectes platessa) N/A (Kelman et al., 2006) Experimental Moving and stationary X MD, F-F, BM, DC Computer-generated moving and static stimuli Humans (Stevens et al., 2011) Experimental Stationary X BM Artificial paper moths Captive trained birds (Dimitrova & Merilaita, 2012) Experimental N/A X BM and DC Galaxias nebula fish (Galaxiidae) N/A (Magellan & Swartz, 2013) Experimental Moving X MD, F-F, BM, distractive Computer-generated moving stimuli Humans (Hughes et al., 2014) Experimental The mechanisms underlying convergent evolution in the plumage patterns of birds 258 Pattern type Behavioural context Camouflage Communication Camouflage mechanism Prey or signaller (where tested) Predator or signal receiver (where tested) References Type of literature/ evidence markings Irregular and regular TOTAL Stationary = 2 Moving & stationary = 1 5 0 BM = 5 DC = 2 MD, F-F = 3 2 species (fish) 2 groups (birds and humans) Ca. 5+ Experimental = 5/5 (100%) Pattern type Behavioural context Camouflage Communication Camouflage mechanism Prey or signaller (where tested) Predator or signal receiver (where tested) References Type of literature/ evidence Bimodal Moving and stationary X X Camouflage vs. signal efficacy Guppies (Poecilia reticulata) Various aquatic predators (Endler, 1978) Hypotheses and Experimental Moving and stationary X X Predator avoidance vs. signal N/A N/A (Endler, 1987) Experimental The mechanisms underlying convergent evolution in the plumage patterns of birds 259 Pattern type Behavioural context Camouflage Communication Camouflage mechanism Prey or signaller (where tested) Predator or signal receiver (where tested) References Type of literature/ evidence efficacy Moving and stationary X X Distance- dependence Reef fish (e.g. Pygoplites diacanthus) Predatory fish (Marshall, 2000) Experimental Bimodal Stationary X X Signal partitioning Australian agamid lizards (Agamidae) N/A (Stuart-Fox & Ord, 2004) Comparative phylogenetic N/A X X Signal partitioning Bicyclus butterflies N/A (Oliver et al., 2009) Comparative phylogenetic N/A X X Signal partitioning Birds N/A (Gluckman & Cardoso, 2010) Comparative phylogenetic Stationary X X Background- matching vs. conspicuous signalling Giant cuttlefish (Sepia apama) N/A (Zylinski et al., 2011) Experimental N/A X X Predator avoidance vs. sexual Australian dragon lizards (Agamidae) N/A (Chen et al., 2012) Comparative phylogenetic The mechanisms underlying convergent evolution in the plumage patterns of birds 260 Pattern type Behavioural context Camouflage Communication Camouflage mechanism Prey or signaller (where tested) Predator or signal receiver (where tested) References Type of literature/ evidence dichromatis m Stationary X X Signal partitioning Australian mallee dragon lizards (Ctenophorus fordi) Avian predators (Garcia et al., 2013) Experimental Bimodal TOTAL Stationary = 3 Moving & stationary = 3 9 9 Multiple function = 5 Signal partitioning =4 Ca. 4+ species: reptiles, cuttlefish, birds, insects, fish 2 groups (fish, birds) Ca. 9+ Experimental = 5/9 (55%) Comparative = 4/9 (45%) The mechanisms underlying convergent evolution in the plumage patterns of birds 261 References for Table S5.1 Akkaynak, D., Allen, J. J., Mäthger, L. M., Chiao, C.-C., and Hanlon, R. T. (2013). Quantification of cuttlefish (Sepia officinalis) camouflage: a study of color and luminance using in situ spectrometry. Journal of Comparative Physiology A, 199(3), 211-225. Allen, W. L., Baddeley, R., Scott-Samuel, N. E., and Cuthill, I. C. (2013). The evolution and function of pattern diversity in snakes. Behavioral Ecology, 24(5), 1237-1250. Allen, W. L., Cuthill, I. C., Scott-Samuel, N. E., and Baddeley, R. (2011). Why the leopard got its spots: relating pattern development to ecology in felids. Proceedings of the Royal Society, Series B, 278(1710), 1373-1380. Barbosa, A., Florio, C. F., Chiao, C.-C., and Hanlon, R. T. (2004). Visual background features that elicit mottled patterns in cuttlefish, Sepia officinalis. The Biological Bulletin, 207(2), 154. Barbosa, A., Mäthger, L. M., Buresch, K. C., Kelly, J., Chubb, C., Chiao, C.-C., and Hanlon, R. T. (2008). Cuttlefish camouflage: The effects of substrate contrast and size in evoking uniform, mottle or disruptive body patterns. Vision Research, 48(10), 1242-1253. Barbosa, A., Mäthger, L. M., Chubb, C., Florio, C. F., Chiao, C.-C., and Hanlon, R. T. (2007). Disruptive coloration in cuttlefish: a visual perception mechanism that regulates ontogenetic adjustment of skin patterning. Journal of Experimental Biology, 210, 1139-1147. Bortolotti, G. R., Blas, J., Negro, J. J., and Tella, J. L. (2006). A complex plumage pattern as an honest social signal. Animal Behaviour, 72(2), 423-430. Brodie, E. D. (1989). Genetic correlations between morphology and antipredator behaviour in natural populations of the garter snake Thamnophis ordinoides. Nature, 342, 542-543. The mechanisms underlying convergent evolution in the plumage patterns of birds 262 Brodie, E. D. (1992). Correlational selection for color pattern and antipredator behavior in the garter snake Thamnophis ordinoides. Evolution, 46(5), 1284- 1298. Brodie, E. D. (1993). Differential avoidance of coral snake banded patterns by free- ranging avian predators in Costa Rica. Evolution, 47(1), 227-235. Chen, I.-P., Stuart-Fox, D. M., Hugall, A. F., and Symonds, M. R. E. (2012). Sexual selection and the evolution of complex color patterns in dragon lizards. Evolution, 66(11), 3605-3614. Chiao, C.-C., Chubb, C., Buresch, K. C., Barbosa, A., Allen, J. J., Mäthger, L. M., and Hanlon, R. T. (2010). Mottle camouflage patterns in cuttlefish: quantitative characterization and visual background stimuli that evoke them. Journal of Experimental Biology, 213, 187-199. Chiao, C.-C., Chubb, C., and Hanlon, R. T. (2007). Interactive effects of size, contrast, intensity and configuration of background objects in evoking disruptive camouflage in cuttlefish. Vision Research, 47(16), 2223-2235. Chiao, C.-C., and Hanlon, R. T. (2001). Cuttlefish camouflage: visual perception of size, contrast and number of white squares on artificial checkerboard substrata initiates disruptive coloration. Journal of Experimental Biology, 204, 2119-2125. Chiao, C.-C., Kelman, E. J., and Hanlon, R. T. (2005). Disruptive body patterning of cuttlefish (Sepia officinalis) requies visual information regarding edges and contrast of objects in natural substrate backgrounds. The Biological Bulletin, 208(1), 7-11. Chiao, C.-C., Ulmer, K. M., Siemann, L. A., Buresch, K. C., Chubb, C., and Hanlon, R. T. (2013). How visual edge features influence cuttlefish camouflage patterning. Vision Research, 83, 40-47. Chiao, C.-C., Wickiser, J. K., Allen, J. J., Genter, B., and Hanlon, R. T. (2011). Hyperspectral imaging of cuttlefish camouflage indicates good color match in the eyes of fish predators. Proceedings of the National Academy of Sciences USA, 108(22), 9148-9153. The mechanisms underlying convergent evolution in the plumage patterns of birds 263 Cott, H. B. (1940). Adaptive coloration in animals. London: Methuen and Co. Ltd. Cuthill, I. C., Stevens, M., Sheppard, J., Maddocks, T., Párraga, C. A., and Troscianko, T. S. (2005). Disruptive coloration and background pattern matching. Nature, 434, 72-74. Cuthill, I. C., Stevens, M., Windsor, A. M. M., and Walker, H. J. (2006). The effects of pattern symmetry on detection of disruptive and background-matching coloration. Behavioral Ecology, 17(5), 828-832. Cuthill, I. C., and Székely, A. (2009). Coincident disruptive coloration. Philosophical Transactions of the Royal Society B, 364(1516), 489-496. Dimitrova, M., and Merilaita, S. (2010). Prey concealment: visual background complexity and prey contrast distribution. Behavioral Ecology, 21(1), 176-181. Dimitrova, M., and Merilaita, S. (2012). Prey pattern regularity and background complexity affect detectability of background-matching prey. Behavioral Ecology, 23(2), 384-390. DImitrova, M., and Merilaita, S. (2014). Hide and seek: properties of prey and background patterns affect prey detection by blue tits. Behavioral Ecology, 25(2), 402-408. Endler, J. A. (1978). A predator's view of animal color patterns. Evolutionary Biology, 11, 319-364. Endler, J. A. (1980). Natural selection on color patterns in Poecilia reticulata. Evolution, 34(1), 76-91. Endler, J. A. (1987). Predation, light intensity and courtship behavior in Poecilia reticulata (Pisces: Poeciliidae). Animal Behaviour, 35(5), 1376-1385. Fraser, S., Callahan, A., Klassen, D., and Sherratt, T. N. (2007). Empirical tests of the role of disruptive coloration in reducing detectability. Proceedings of the Royal Society, Series B, 274(1615), 1325-1331. Garcia, J. E., Rohr, D., and Dyer, A. G. (2013). Trade-off between camouflage and sexual dimorphism revealed by UV digital imaging: the case of Australian The mechanisms underlying convergent evolution in the plumage patterns of birds 264 Mallee dragons (Ctenophorus fordi). Journal of Experimental Biology, 216, 4290-4298. Gluckman, T.-L., and Cardoso, G. C. (2009). A method to quantify the regularity of barred plumage patterns. Behavioral Ecology and Sociobiology, 63(12), 1837- 1844. Gluckman, T.-L., and Cardoso, G. C. (2010). The dual function of barred plumage in birds: camouflage and communication. Journal of Evolutionary Biology, 23(11), 2501-2506. Godfrey, D., Lythgoe, J. N., and Rumball, D. A. (1987). Zebra stripes and tiger stripes: the spatial frequency distribution of the pattern compared to that of the background is significant in display and crypsis. Biological Journal of the Linnean Society, 32(4), 427-433. Hall, J. R., Cuthill, I. C., Baddeley, R., Shohet, A. J., and Scott-Samuel, N. E. (2013). Camouflage, detection and identification of moving targets. Proceedings of the Royal Society, Series B, 280(1758), 20130064. Hanlon, R. T., Chiao, C.-C., Mäthger, L. M., and Marshall, J. (2013). A fish-eye view of cuttlefish camouflage using in situ spectrometry. Biological Journal of the Linnean Society, 109(3), 535-551. Hanlon, R. T., and Messenger, J. B. (1988). Adaptive coloration in young cuttlefish (Sepia officinalis L.): the morphology and development of body patterns and their relation to behaviour. Philosophical Transactions of the Royal Society B, 320(1200), 437-487. Hemmi, J. M., Marshall, J., Pix, W., Vorobyev, M., and Zeil, J. (2006). The variable colours of the fiddler crab Uca vomeris and their relation to background and predation. Journal of Experimental Biology, 209, 4140-4153. Hohman, W. L., Ankney, C. D., and Gordon, D. H. (1992). Ecology and management of postbreeding waterfowl. In B. D. J. Batt, A. D. Afton, M. G. Anderson, C. D. Ankney, D. H. Johnson, J. A. Kadlec and G. L. Krapu (Eds.), Ecology and management of breeding waterfowl (pp. 128-189). Minneapolis: University of Minnesota Press. The mechanisms underlying convergent evolution in the plumage patterns of birds 265 How, M. J., and Zanker, J. M. (2013). Motion camouflage induced by zebra stripes. Zoology, 117(3), 163-170. Hughes, A. E., Troscianko, J., and Stevens, M. (2014). Motion dazzle and the effects of target patterning on capture success. BMC Evolutionary Biology, 14, 201. Jackson, J. F., Ingram, W., and Campbell, H. W. (1976). The dorsal pigmentation pattern of snakes as an antipredator strategy: A multivariate approach. The American Naturalist, 110(976), 1029-1053. Josef, N., Amodio, P., Fiorito, G., and Shashar, N. (2012). Camouflaging in a complex environment-Octopuses use specific features of their surroundings for background matching. PLoS ONE, 7(5), e37579. Kang, C.-K., Moon, J.-Y., Lee, S.-I., and Jablonski, P. G. (2012). Camouflage through an active choice of a resting spot and body orientation in moths. Journal of Evolutionary Biology, 25, 1695-1702. Kang, C.-K., Moon, J.-Y., Lee, S.-I., and Jablonski, P. G. (2013a). Cryptically patterned moths perceive bark structure when choosing body orientations that match wing color pattern to the bark pattern. PLoS ONE, 8, e78117. Kang, C.-K., Moon, J.-Y., Lee, S.-I., and Jablonski, P. G. (2013b). Moths on tree trunks seek out more cryptic positions when their current crypticity is low. Animal Behaviour, 86(3), 587-594. Kang, C.-K., Stevens, M., Moon, J.-Y., Lee, S.-I., and Jablonski, P. G. (2014). Camouflage through behavior in moths: the role of background matching and disruptive coloration. Behavioral Ecology. doi: 10.1093/beheco/aru150 Kelman, E. J., Tiptus, P., and Osorio, D. (2006). Juvenile plaice (Pleuronectes platessa) produce camouflage by flexibly combining two separate patterns. Journal of Experimental Biology, 209, 3288-3292. Kenward, B., Wachtmeister, C. A., Ghirlanda, S., and Enquist, M. (2004). Spots and stripes: the evolution of repetition in visual signal form. Journal of Theoretical Biology, 230(3), 407-419. Kiltie, R. A., Fan, J., and Laine, A. F. (1994). A wavelet-based metric for visual The mechanisms underlying convergent evolution in the plumage patterns of birds 266 texture discrimination with applications in evolutionary ecology. Mathematical Biosciences, 126(1), 21-39. Kjernsmo, K., and Merilaita, S. (2012). Background choice as an anti-predator strategy: the roles of background matching and visual complexity in the habitat choice of the least killifish. Proceedings of the Royal Society, Series B, 279(1745), 4192-4198. Lindell, L. E., and Forsman, A. (1996). Density effects and snake predation: prey limitation and reduced growth rate of adders at high density of conspecifics. Canadian Journal of Zoology, 74(6), 1000-1007. Lovell, P. G., Ruxton, G. D., Langridge, K. V., and Spencer, K. A. (2013). Egg-laying substrate selection for optimal camouflage by quail. Current Biology, 23(3), 260-264. Magellan, K., and Swartz, E. R. (2013). Crypsis in a heterogeneous environment: relationships between changeable polymorphic colour patterns and behaviour in a galaxiid fish. Freshwater Biology, 58(4), 793-799. Marshall, J. (2000). Communication and camouflage with the same 'bright' colours in reef fishes. Philosophical Transactions of the Royal Society B, 355(1401), 1243-1248. Mäthger, L. M., Chiao, C.-C., Barbosa, A., Buresch, K. C., Kaye, S., and Hanlon, R. T. (2007). Disruptive coloration elicited on controlled natural substrates in cuttlefish, Sepia officinalis. Journal of Evolutionary Biology, 210, 2657-2666. Mäthger, L. M., Chiao, C.-C., Barbosa, A., and Hanlon, R. T. (2008). Color matching on natural substrates in cuttlefish, Sepia officinalis. Journal of Comparative Physiology A, 194(6), 577-585. Merilaita, S. (1998). Crypsis through disruptive coloration in an isopod. Proceedings of the Royal Society, Series B, 265(1401), 1059-1064. Merilaita, S., and DImitrova, M. (2014). Accuracy of background matching and prey detection: predation by blue tits indicates intense selection for highly matching prey colour pattern. Functional Ecology. doi: 10.1111/1365-2435.12248 The mechanisms underlying convergent evolution in the plumage patterns of birds 267 Merilaita, S., and Lind, J. (2005). Background-matching and disruptive coloration, and the evolution of cryptic coloration. Proceedings of the Royal Society, Series B, 272(1563), 665-670. Merilaita, S., Lyytinen, A., and Mappes, J. (2001). Selection for cryptic coloration in a visually heterogeneous habitat. Proceedings of the Royal Society, Series B, 268(1479), 1925-1929. Muck, C., and Goymann, W. (2011). Throat patch size and darkness covaries with testosterone in females of a sex-role reversed species. Behavioral Ecology, 22(6), 1312-1319. Niskanen, M., and Mappes, J. (2005). Significance of the dorsal zigzag pattern of Vipera latastei gaditana against avian predators. Journal of Animal Ecology, 74(6), 1091-1101. Oliver, J. C., Robertson, K. A., and Monteiro, A. (2009). Accommodating natural and sexual selection in butterfly wing pattern evolution. Proceedings of the Royal Society, Series B, 276(1666), 2369-2375. Omland, K. E. (1996). Female mallard mating preferences for multiple male ornaments. Behavioral Ecology and Sociobiology, 39(6), 353-360. Ortolani, A. (1999). Spots, stripes, tail tips and dark eyes: Predicting the function of carnivore colour patterns using the comparative method. Biological Journal of the Linnean Society, 67(4), 433-476. Osorio, D., and Srinivasan, M. V. (1991). Camouflage by edge enhancement in animal coloration patterns and its implications for visual mechanisms. Proceedings of the Royal Society, Series B, 244(1310), 81-85. Pérez-Rodríguez, L., Jovani, R., and Mougeot, F. (2013). Fractal geometry of a complex plumage trait reveals bird's quality. Proceedings of the Royal Society, Series B, 280(1755), 20122783. Petrie, M., Halliday, T., and Sanders, C. (1991). Peahens prefer peacocks with elaborate trains. Animal Behaviour, 41(2), 323-331. Pough, F. H. (1976). Multiple cryptic effects of crossbanded and ringed patterns of The mechanisms underlying convergent evolution in the plumage patterns of birds 268 snakes. Copeia, 1976(4), 834-836. Poulton, E. B. (1890). The colours of animals: their meaning and use. Especially considered in the case of insects. The International Scientific Series (2nd ed.). London: Kegan Paul, Trench Trübner and Co. Ltd. Roulin, A. (1999). Nonrandom pairing by male barn owls (Tyto alba) with respect to a female plumage trait. Behavioral Ecology, 10(6), 688-695. Roulin, A., Altwegg, R., Jensen, H., Steinsland, I., and Schaub, M. (2010). Sex- dependent selection on an autosomal melanic female ornament promotes the evolution of sex ratio bias. Ecology Letters, 13(5), 616-626. Santer, R. D. (2013). Motion dazzle: a locust's eye view. Biology Letters, 9(6), 20130811. Schaefer, M. H., and Stobbe, N. (2006). Disruptive coloration provides camouflage independent of background matching. Proceedings of the Royal Society, Series B, 273(1600), 2427-2432. Scott-Samuel, N. E., Baddeley, R., Palmer, C. E., and Cuthill, I. C. (2011). Dazzle camouflage affects speed perception. PLoS ONE, 6, e20233. Shine, R., and Madsen, T. (1994). Sexual dichromatism in snakes of the genus Vipera: A review and a new evolutionary hypothesis. Journal of Herpetology, 28, 114-117. Shohet, A. J., Baddeley, R., Anderson, J., and Osorio, D. (2007). Cuttlefish camouflage: a quantitative study of patterning. Biological Journal of the Linnean Society, 92(2), 335-345. Shohet, A. J., Baddeley, R., Anderson, J. C., Kelman, E. J., and Osorio, D. (2006). Cuttlefish responses to visual orientation of substrates, water flow and a model of motion camouflage. Journal of Experimental Biology, 209, 4717- 4723. Siebeck, U. E. (2004). Communication in coral reef fish: the role of ultraviolet colour patterns in damselfish territorial behaviour. Animal Behaviour, 68(2), 273-282. The mechanisms underlying convergent evolution in the plumage patterns of birds 269 Stevens, M., and Cuthill, I. C. (2006). Disruptive coloration, crypsis and edge detection in early visual processing. Proceedings of the Royal Society, Series B, 273(1598), 2141-2147. Stevens, M., Cuthill, I. C., Windsor, A. M. M., and Walker, H. J. (2006). Disruptive contrast in animal camouflage. Proceedings of the Royal Society, Series B, 273(1600), 2433-2438. Stevens, M., Searle, W. T., Seymour, J., Marshall, K. L. A., and Ruxton, G. D. (2011). Motion dazzle and camouflage as distinct anti-predator defenses. BMC Biology, 9, 81. Stevens, M., Winney, I. S., Cantor, A., and Graham, J. (2009). Outline and surface disruption in animal camouflage. Proceedings of the Royal Society, Series B, 276(1657), 781-786. Stevens, M., Yule, D. H., and Ruxton, G. D. (2008). Dazzle coloration and prey movement. Proceedings of the Royal Society, Series B, 275(1651), 2639- 2643. Stobbe, N., and Schaefer, M. H. (2008). Enhancement of chromatic contrast increases predation risk for striped butterflies. Proceedings of the Royal Society, Series B, 275(1642), 1535-1541. Stoner, C. J., Caro, T., and Graham, C. M. (2003). Ecological and behavioral correlates of coloration in artiodactyls: systematic analyses of conventional hypotheses. Behavioral Ecology, 14(6), 823-840. Stuart-Fox, D. M., and Ord, T. J. (2004). Sexual selection, natural selection and the evolution of dimorphic coloration and ornamentation in agamid lizards. Proceedings of the Royal Society, Series B, 271(1554), 2249-2255. Swaddle, J. P., and Cuthill, I. C. (1994). Female zebra finches prefer males with symmetric chest plumage. Proceedings of the Royal Society, Series B, 258(1353), 267-271. Thayer, G. H. (1909). Concealing-coloration in the animal kingdom: an exposition of the laws of disguise through color and pattern: being a summary of Abbott H. Thayer's discoveries. New York: Macmillan. The mechanisms underlying convergent evolution in the plumage patterns of birds 270 Troscianko, J., Lown, A. E., Hughes, A. E., and Stevens, M. (2013). Defeating crypsis: detection and learning of camouflage strategies. PLoS ONE, 8, e73733. von Helversen, B., Schooler, L. J., and Czienskowski, U. (2013). Are stripes beneficial? Dazzle camouflage influences perceived speed and hit rates. PLoS ONE, 8(4), e61173. Zylinski, S., How, M. J., Osorio, D., Hanlon, R. T., and Marshall, J. (2011). To be seen or to hide: visual characteristics of body patterns for camouflage and communication in the Australian giant cuttlefish Sepia apama. The American Naturalist, 177(5), 681-690. Zylinski, S., Osorio, D., and Shohet, A. J. (2009). Cuttlefish camouflage: context- dependent body pattern use during motion. Proceedings of the Royal Society, Series B, 276(1675), 3963-3969. The mechanisms underlying convergent evolution in the plumage patterns of birds 271 Table S5.2. The number of bird species with plumage patterns in the 16 habitat types of the world across the class Aves. Plumage patterns represent breeding and non-breeding plumage of males and females, as well as juveniles, on the ventral and dorsal surface. Species numbers are averaged across eco-regions for each habitat. T = Temperate; and T/S - Tropical and Subtropical. Sex/ age Season Location Pattern Boreal forest /taiga Desert and xeric shrubland Flooded grassland and savanna Inland water Mangrove Mediterranean forest, woodland and scrub Montane grassland and shrubland Rock and Ice T broadleaf and mixed forest T conifer forest T grassland savanna and shrubland T/S coniferous forest T/S dry broadleaf forest T/S grassland savanna and shrubland T/S moist broadleaf forests Tundra Male Breeding Ventral All 23.4 33.5 39.7 96.7 66.4 27.3 49.9 32.0 27.4 32.4 30.3 49.4 49.7 72.8 54.2 12.1 Mottled 11.9 16.6 19.1 40.3 32.2 13.1 23.0 15.7 14.0 16.5 15.1 23.6 22.3 34.1 25.3 5.4 Scaled 1.0 3.8 5.2 12.7 9.0 3.6 5.7 2.0 3.2 3.3 3.0 7.2 7.6 9.0 7.6 0.5 Barred 11.4 12.0 13.7 41.0 22.4 9.8 19.1 14.7 10.4 12.7 11.6 16.9 17.0 26.1 18.9 6.8 Spotted 1.6 3.3 4.0 8.3 6.1 3.3 4.7 2.3 1.9 2.9 2.4 5.4 5.6 6.9 5.1 0.8 Dorsal All 26.5 44.8 51.7 110.7 83.4 37.1 65.3 39.3 35.3 41.7 36.9 60.2 62.0 90.8 67.5 13.5 Mottled 14.5 24.9 29.6 60.0 47.8 22.2 37.4 23.0 21.0 23.6 21.2 31.1 34.3 52.8 38.0 7.6 Scaled 2.2 4.4 4.2 11.7 8.5 3.6 6.7 4.0 4.2 4.3 3.8 6.1 6.1 7.3 6.7 0.7 Barred 8.1 10.7 13.2 36.0 21.5 9.1 18.0 13.7 9.5 11.4 8.7 16.7 16.5 26.4 19.0 5.3 Spotted 6.5 10.4 10.8 18.3 15.5 8.1 12.3 7.3 6.6 8.6 7.8 13.9 12.1 15.4 12.0 3.0 Non- breeding Ventral All 23.7 33.7 39.9 96.7 66.9 26.9 49.8 32.0 27.5 32.8 30.3 50.1 50.0 72.2 54.4 12.3 Mottled 12.3 16.7 19.6 40.7 33.4 13.1 23.3 16.0 14.5 16.9 15.3 24.4 23.1 34.6 26.1 5.6 Scaled 0.9 3.5 4.9 12.3 8.2 3.3 5.3 1.3 2.9 3.1 2.8 7.2 7.0 8.6 7.0 0.5 Barred 11.2 12.5 14.0 41.3 23.2 9.8 19.3 15.7 10.5 13.2 11.6 18.1 18.0 25.7 19.4 6.7 Spotted 1.6 3.1 3.6 8.0 5.4 3.2 4.4 1.7 1.8 2.5 2.3 4.3 4.7 6.7 4.5 0.8 Dorsal All 27.1 45.8 52.5 112.7 86.5 37.2 66.2 40.0 35.9 42.6 37.3 64.2 64.9 91.5 69.4 13.8 Mottled 15.7 25.5 30.6 62.3 49.7 22.4 38.0 23.0 21.5 24.0 21.5 33.3 36.2 53.7 39.2 8.3 The mechanisms underlying convergent evolution in the plumage patterns of birds 272 Sex/ age Season Location Pattern Boreal forest /taiga Desert and xeric shrubland Flooded grassland and savanna Inland water Mangrove Mediterranean forest, woodland and scrub Montane grassland and shrubland Rock and Ice T broadleaf and mixed forest T conifer forest T grassland savanna and shrubland T/S coniferous forest T/S dry broadleaf forest T/S grassland savanna and shrubland T/S moist broadleaf forests Tundra Scaled 2.0 4.6 4.0 10.7 8.1 3.7 6.3 4.0 4.1 4.8 3.9 7.0 6.1 6.5 6.5 0.6 Barred 8.0 11.1 13.5 36.3 22.8 9.0 18.8 14.3 9.6 11.3 8.5 17.8 17.6 27.1 20.1 5.1 Spotted 6.5 10.4 11.1 18.7 16.1 8.2 12.6 7.3 6.7 8.7 7.9 14.0 12.2 16.0 12.1 3.0 Female Breeding Ventral All 32.4 43.3 49.5 117.0 81.3 37.7 61.9 43.3 36.6 43.8 40.0 61.1 59.5 89.4 66.9 17.3 Mottled 16.2 21.3 24.4 51.3 41.4 18.5 30.2 21.7 18.4 21.9 19.0 29.9 27.5 45.1 32.2 7.8 Scaled 4.2 5.4 7.0 15.7 10.4 5.1 7.1 3.3 4.9 4.9 5.1 8.6 8.8 10.8 9.1 2.2 Barred 13.2 14.7 15.8 45.3 26.2 12.5 21.5 18.0 12.9 16.6 14.4 20.4 20.4 29.1 23.0 8.2 Spotted 2.3 4.4 4.7 11.0 6.7 4.7 5.8 3.3 3.3 3.9 3.9 5.6 5.9 7.9 5.8 0.9 Dorsal All 29.2 49.9 56.8 121.7 94.5 40.6 71.8 45.0 39.1 46.5 39.6 71.4 70.1 99.1 75.7 15.2 Mottled 17.5 28.9 34.5 69.3 56.8 23.8 43.7 27.3 23.4 27.0 23.8 39.2 40.3 61.4 44.6 9.3 Scaled 1.6 4.4 4.3 11.0 8.1 4.1 5.9 3.7 3.7 4.1 3.1 6.9 6.4 7.1 6.7 0.4 Barred 10.5 12.7 14.3 36.7 23.7 11.1 18.2 15.3 12.5 14.2 10.7 19.8 18.9 26.1 20.9 6.2 Spotted 5.8 9.8 10.3 19.0 15.4 7.6 12.7 7.0 6.1 7.7 6.9 13.4 11.9 16.2 11.8 3.1 Non- breeding Ventral All 32.1 43.2 49.4 116.7 81.5 37.4 61.7 43.3 36.2 43.5 39.5 61.2 59.7 89.3 67.1 17.1 Mottled 15.9 21.2 24.2 51.3 41.5 18.4 30.2 21.7 18.2 21.7 18.8 30.1 27.8 45.0 32.3 7.6 Scaled 4.2 5.4 7.0 15.7 10.3 5.1 7.0 3.3 4.9 4.9 5.1 8.6 8.8 10.8 9.1 2.2 Barred 13.2 14.6 15.8 45.0 26.3 12.4 21.4 18.0 12.6 16.5 14.0 20.4 20.3 29.1 22.9 8.2 Spotted 2.3 4.4 4.7 11.0 6.7 4.7 5.8 3.3 3.3 3.9 3.9 5.6 5.9 7.9 5.8 0.9 Dorsal All 28.9 49.9 56.8 121.3 94.5 40.4 71.7 45.0 38.7 46.2 39.2 71.4 70.0 99.1 75.6 15.0 Mottled 17.2 28.9 34.6 69.3 56.8 23.7 43.7 27.3 23.2 26.8 23.6 39.2 40.3 61.4 44.6 9.1 Scaled 1.6 4.4 4.3 11.0 8.0 4.1 5.9 3.7 3.7 4.1 3.1 6.9 6.4 7.1 6.7 0.4 The mechanisms underlying convergent evolution in the plumage patterns of birds 273 Sex/ age Season Location Pattern Boreal forest /taiga Desert and xeric shrubland Flooded grassland and savanna Inland water Mangrove Mediterranean forest, woodland and scrub Montane grassland and shrubland Rock and Ice T broadleaf and mixed forest T conifer forest T grassland savanna and shrubland T/S coniferous forest T/S dry broadleaf forest T/S grassland savanna and shrubland T/S moist broadleaf forests Tundra Barred 10.5 12.6 14.2 36.3 23.8 10.9 18.1 15.3 12.3 14.1 10.3 19.8 18.8 25.9 20.9 6.2 Spotted 5.8 9.8 10.3 19.0 15.4 7.6 12.7 7.0 6.1 7.7 6.9 13.4 11.9 16.2 11.8 3.1 Juvenile N/A Ventral All 24.7 38.7 44.7 102.0 74.1 32.6 54.3 33.3 30.2 36.0 31.8 57.8 55.6 76.1 59.7 12.3 Mottled 15.6 23.6 27.2 59.3 45.1 17.7 31.7 19.3 19.0 22.8 19.8 36.0 31.8 43.7 33.7 7.3 Scaled 2.4 3.4 4.8 9.0 6.3 3.1 4.4 1.7 2.7 2.8 3.2 4.8 5.1 8.2 5.4 1.3 Barred 7.7 11.2 11.7 32.7 20.7 11.5 17.1 13.7 10.1 12.4 9.9 16.1 16.0 22.2 18.2 4.3 Spotted 1.6 2.7 3.2 7.0 5.0 2.8 3.5 1.0 1.5 1.9 1.8 4.2 4.7 5.1 4.5 0.9 Dorsal All 23.1 39.8 46.7 96.7 75.9 36.1 57.9 35.0 31.3 36.0 31.2 53.9 54.1 84.0 61.6 12.0 Mottled 14.8 25.4 31.6 62.7 49.4 21.2 38.1 22.0 20.2 23.4 20.9 36.7 35.7 53.0 38.3 7.7 Scaled 0.9 4.3 4.9 9.3 8.6 4.2 4.9 2.7 3.0 3.3 2.4 5.2 6.5 8.6 7.4 0.4 Barred 6.7 7.1 7.0 19.0 12.7 7.0 10.4 10.0 7.3 7.9 5.6 8.9 7.9 16.9 11.2 3.5 Spotted 3.6 6.0 6.3 12.7 9.4 6.9 8.3 4.7 4.2 4.6 5.0 6.6 6.9 10.9 8.0 1.7 The mechanisms underlying convergent evolution in the plumage patterns of birds 274 Dearest Africa, thank you for being my inspiration for change. Sorry I did not come back permanently, I ended up getting a PhD instead. Ngorongoro crater, 2002.