S29.5: The importance of non-reproductive functions of bird colouration, especially anti-predator adaptations

Department of Zoology, University of Göteborg, Box 463, SE 405 30 Göteborg, Sweden, e-mail f.gotmark@zool.gu.se

It is not surprising that most papers on the function and evolution of bird colouration are concerned with the bright, ornamental colours (or 'signals') that characterise many bird species, especially male birds in the breeding season. Many yellow, red or blue birds are beautiful and this beauty demands an explanation. The present strong interest in bird colours has been stimulated by exciting theories in the field of sexual selection (Andersson 1994). However, most bird species are dull-coloured and dominated by the most common colours of nature; green, grey and brown. I believe that most ornithologists, if asked, would guess that these dull colours make the birds cryptic to predators, and so minimise the risk of predation. This obvious assumption is probably the main reason why there are so few studies of crypsis. Another reason is that crypsis is difficult to study; predation can rarely be directly observed in nature, and colours, patterns and backgrounds are not easy to quantify (Endler 1990).

I hope my review will show that bird colours and their relations to predation are fascinating and important to study if we are to fully understand the evolution of bird colouration, and indeed bright colours. Some important questions are: What ecological and behavioural traits favour crypsis in birds? Could predation favour bright and contrasting plumages, through warning colouration or disruptive crypsis in certain habitats? Could predation play a role in shaping the often different plumages of males and females? I present a review based on some recent work, especially concerned with the effects of predators on adult birds and adult plumages. I also briefly summarise the evidence for other known non-reproductive functions of bird colouration. A considerable part of the text, especially regarding predation and its effects concerns my own work, which reflects the fact that few people have studied these aspects (my own work, especially predation papers between 1992 and 1997, has not been summarised elsewhere). I apologise for overlooked relevant papers, and welcome all comments. One topic that is not treated here, but which also deserves more study, is egg colouration (e.g. Götmark 1992a, Westmoreland & Kiltie 1996, and references therein). For excellent reviews of bird colouration that also treat non-reproductive functions, see also Burtt (1981), Butcher & Rohwer (1989) and Savalli (1995).

Several studies have suggested that presence of pigments (melanin) reduces wear of feathers. This was shown in experiments with artificial wear (through particles), in which pigmented feathers (melanin) of American Wood Warblers Phylloscopus sibilatrix were more resistant than non-pigmented feathers (Burtt 1979, see also Bonser 1995). In the Yellow-rumped Warbler Dendroica coronata, albino wing feathers on one wing of a specimen were clearly more worn (and shorter) than normal ones on the other wing (Barrowclough & Sibley 1980). These results provide one explanation for why wing and tail feathers in many species are generally dark and contain melanin, although it is not clear if feathers with high content of melanin gain better protection than those with intermediate concentration of pigments.

The effect of colouration and pigmentation on thermoregulation is complicated (review in Walsberg 1988) and needs more empirical work. Detailed studies of males in the Phainopepla Phainopepla nitens, a North American bird that nests in hot, arid environments, showed that the effect of black feathers on solar heat gain was small compared with the effects of behaviour and postural changes; black male plumage might be selected for communication, for crypsis, or for both (Walsberg 1982). Pigmentation of body parts may also affect vision in birds, in surprising ways. In the Willow Flycatcher Empidonax traillii and in many other birds, the upper mandible projects into the visual field and so potentially may affect vision. Experiments where the normal dark mandible was painted white showed that manipulated birds had lower foraging success and spent more time in shaded habitat than controls (Burtt 1984). Comparative studies of wood warblers had shown that species foraging in the open usually have dark bills, while those using shaded habitats more often had light bills, also suggesting a cost of producing pigments in these birds. This fascinating study shows that colouration also of minor body parts may be adaptive.

In general, few studies of the effects of colouration on physiology and related aspects have been published in recent years. More work is needed to clarify the influence of colouration on wear of feathers and patterns of moult. Ideally, in such studies tests should be conducted of alternative hypothesis (see below).

The underparts of many seabirds are light, which may improve foraging success, since the birds would be more difficult to detect for aquatic prey against a bright sky ('hunting camouflage'; see review in Simmons 1972). There is experimental support for this idea, in a study where Black-headed Gulls Larus ridibundus were painted black on the underparts. In an indoor pool, normal control birds caught more fish per time unit than did black gulls (Götmark 1987).

Another related idea is that white or contrasting (black-and-white) plumage of seabirds, especially on the upper parts, facilitates detection, flocking and group-feeding, which would be beneficial if individual hunting success increases with flock size (Simmons 1972). Using the same indoor pool and Black-headed Gulls, Götmark et al. (1986) found that individual fishing success increased with flock size, from solitary birds to flocks of six birds. Larger flocks split the school and the gulls managed to catch more fish. In seabirds, feeding groups composed of several species with different hunting tactic might also improve feeding success (Hoffman et al. 1981). Beside foraging, factors such as competition and predation risk (Bretagnolle 1993), sexual selection (Jones & Hunter 1993), and species recognition (Pierotti 1987) have probably influenced the evolution of seabird plumage colouration. Thus, even in this relatively homogeneous group (as regards colour patterns), several or many selective pressures probably influence colouration.

In terrestrial birds, foraging-related plumage colouration, such as hunting camouflage, has been little studied. For predators like Accipiters, it is likely that colouration partly has evolved to reduce the risk of being detected by the prey, both for perched and flying hawks. In addition, in some passerine birds with bright wing and tail markings, colour patterns may facilitate flushing of insect prey (Jablonski 1996).

In some birds species, nestlings or chicks have ornaments with bright colours. In coots for instance, the chicks have modified, conspicuous orange feathers on the front of the body and bright red patterns on the bill (Harrison 1985). Recently, Lyon et al. (1994) showed that American Coot F. americana parents prefer to feed ornamented chicks over non-ornamented ones (chicks with trimmed orange feathers), leading to higher growth rates and greater survival for ornamented chicks. Lyon et al. (1994) suggested that parental choice in this way can select for ornamental traits in offspring.

Nestlings and chicks are mostly cryptic in appearance, and therefore this interesting idea may have limited applicability in birds. However, one common bright signal in nestling songbirds is a yellow, orange or red mouth, displayed as the nestling beg for food from the parents. Götmark & Alström (1997) examined the hypothesis that the bright mouths of chicks of songbirds have been favoured by feeding preferences in parents. In a field experiment, the orange-yellow mouth of Great Tit Parus major nestlings was dyed brightly red, and the feeding response of parents recorded. In nest boxes with extra daylight through a window, experimental chicks were on average given twice as much food (biomass) as control chicks (sham-dyed). In normal nest boxes, the tendency was similar, but not significant. Thus, at least in good light, great tit parents preferred to feed young with red mouth, a preference for colourfulness that helps explain the evolution of bright gapes in chicks of passerine birds. Similar results were independently reported by Kilner (1997, 1998) for Canary Serinus canaria nestlings, studied in a different theoretical context, but manipulated in a similar way (by food-dye). Great Tit nestlings have more or less fixed mouth colouration (Götmark & Alström 1997), while begging Canary nestlings are able to increase the saturation of the mouth colour through blood flow in the integument of the mouth (Kilner 1997, 1998). Thus, a similar trait may have evolved in different ways in different species. However, there is considerable variation between species in mouth colouration and more work is needed to clarify this variation, as well as to explore alternative forms of colour communication between parents and their young (see Booth 1990, Kilner 1997).

To explain variation in plumage ornaments in wintering sparrows, Rohwer (1975) proposed that the size of ‘badges’, such black bibs, is positively related to individual dominance status. This interesting but controversial idea seems at least partly supported by a number of experiments with dyed birds (see Senar 1999 for a review of the evidence).

Warning colouration is a well-studied adaptation in invertebrates, especially in the diurnal butterflies (Guilford 1990, Endler 1991). A combination of bright or contrasting colours and distastefulness in invertebrate prey has probably been selected through reduced predation risk. Predators learn to avoid distasteful or poisonous prey, or they do not kill such prey after capture, and the prey may escape (demonstrated in several studies).

The idea of aposematism is more than 100 years old. About 50 years ago, it was suggested to be true also for brightly coloured birds (Cott 1947, Cott & Benson 1970). While preparing museum specimens in Egypt during the Second World War, Cott noted that hornets ate carcasses of some species, but avoided others, which he found had more bright or contrasting plumages. After five years of tests with hornets and cats as predators, Cott (1947) presented evidence suggesting that a number of bright species such as the Pied Kingfisher Ceryle rudis and Hoopoe Upupa epops were aposematic. Later, a human tasting panel was used for a test of 200 South African species, with the same conclusion as regards colouration and palatability (Cott & Benson 1970).

These results are not often cited and seem controversial, partly because Cott did not use the most relevant predators - Accipiters - as test subjects (Pough 1988). However, both Baker & Parker (1979) and Krebs (1979) thought that the evidence presented by Cott was convincing. But several problems remain; for instance, Cott scored ‘visibility’ of female plumage only, and used only those scores in his analyses. A re-analysis of his original data, with scoring of both female and male breeding plumages, confirmed the trend for the 30 European and north African species (Götmark 1994a). For the larger South African data set, an association between ‘visibility’ (and also ‘conspicuousness’; Götmark 1994a) and palatability was found for 87 non-passerine birds, but for 105 passerines, a correlation existed for females only, and not for males. Although this re-analysis partly supports Cott's work, the finding that conspicuousness of breeding plumage of male passerines was not correlated with palatability makes it difficult to draw firm conclusions. Passerine birds are interesting, as they are relatively small and should suffer comparatively high adult predation rate. If a correlation between plumage and palatability occurs only during the non-breeding season, learning of a general association is rendered more difficult for the predators, especially since the birds are brighter in the breeding season. Clearly, more work is needed to resolve these problems (for more discussion, see Götmark 1994a).

Recently, possible cases of chemical defence and aposematism were discovered in New Guinea; three species of the genus Pitohui contained the poisonous alkaloid homobatrachotoxin, also known from neotropical poison-dart frogs (Dumbacher et al. 1992, Dumbacher & Pruett-Jones 1996). The concentration of toxin was high in skin and feathers, which in fact is what one would expect for chemical defence in birds: after capture of prey, raptors usually start to remove feathers before eating (and killing) the prey (e.g., Newton 1986). In addition, the pitohuis emit a strong sour odour, which might help to repel predators. Experiments with natural predators are needed to demonstrate whether the pitohuis are aposematic, which seems plausible. The study of Dumbacher et al. also suggests that Cott made one mistake: he should not have skinned the birds and tested only the flesh, but instead tasted them directly after capture, including the feathers! Dumbacher et al. (1992) found that muscles and other organs of pitohuis were much less toxic than skin and feathers.

There are some possible cases of mimicry in birds, involving an ‘aggressive’ or well-defended model, and a mimic resembling it in colouration. Examples include petrels that resemble jaegers and skuas, and do not get harassed by these kleptoparasitic birds (Spear & Ainley 1993). In addition, there seem to be examples among terrestrial birds (see Diamond 1994, Savalli 1995), which have not been studied in detail. The pitohui species are also possible examples (Dumbacher & Pruett-Jones 1996).

Cott (1947) and Baker & Parker (1979) suggested that predation has been the major selective pressure influencing bird colouration, explaining not only cryptic plumages but also the evolution of bright colouration. Their ideas are interesting, but controversial (mildly speaking). Most ornithologists and ecologists believe that intraspecific communication (e.g., sexual selection) is the main reason for bright colour patterns. Cott emphasised aposematism as the mechanism favouring bright plumages (see above), while Baker & Parker proposed that bright birds are agile and vigilant, and so difficult to catch. They suggested that predators would learn which species are hard to capture and then tend to avoid such prey (they also considered other predation-related hypotheses). A central prediction of the unprofitable prey hypothesis (UPH) is that experienced predators, when given a choice, should more often attack a cryptic than a conspicuous bird, even if the former is more difficult to detect. The alternative prediction, higher attack rate for the conspicuous bird, follows from the common assumption of a predation cost for bright plumages (due to higher detectability). Attempts to test these predictions have been done recently, by use of taxidermic mounts ('prey') placed at sites where Sparrowhawks Accipiter nisus and Goshawks A. gentilis pass on migration, or near nests of sparrowhawks (Götmark 1992b, 1993, 1995, Götmark & Unger 1994, Götmark 1997, Götmark et al. 1997). Many pairs of stuffed male and female birds, or pairs of cryptic and conspicuous species, were exposed for attacks by hawks. For stuffed male and female Pied Flycatchers (Ficedula hypoleuca), the surprising result is that sparrowhawks tend to avoid the black-and-white males and attack the cryptic, brown females more often (Götmark 1992b, 1993, 1995). However, it is difficult to interpret these results as evidence for the UPH; a novelty effect may be involved. The migratory pied flycatcher arrives late in the breeding season, and the males have a type of plumage which is rare (also lost during the moult in August) . Moreover, the hawks may not recognise the female flycatchers (mounts) as a separate category of prey, but confuse them with other relatively similar and profitable prey species. Hence, the interpretation of the results depends on assumptions about prey recognition in the hawks (see also Götmark 1995 and Slagsvold et al. 1995 for different results and discussion).

Studies of live prey and prey killed by predators are needed, especially since behaviour should be an important component of predation risk. In the Chaffinch Fringilla coelebs, (live) cryptic females are taken more often by Sparrowhawks than bright males in the breeding season (Götmark et al. 1997), contrary to the common assumption that singing and displaying colourful males suffer higher predation risk. In a simultaneous experiment with mounts at hawk nests, we showed that male and female Chaffinch mounts were attacked about equally often. Thus, the hawks did not avoid the males as predicted by the UPH, even though males were under-represented as prey. For this species at least, the UPH was clearly refuted. Females are probably taken more often by Sparrowhawks because they are more exposed to attacks and easier to catch; they spent more time foraging, were more active, and were closer to the ground (Götmark et al. 1997). Our results are important, as they demonstrate that adult predation selects for crypsis in females and therefore also may select for sexual dimorphism in breeding plumage (Wallace 1889, see also Martin & Badyaev 1996). This does not exclude an important role of sexual selection for bright plumage and sexual dichromatism (Andersson 1994, Götmark et al. 1997).

Our studies of attacks on mounts of different species gave limited (Götmark & Unger 1994) or no support for the UPH (Götmark 1997). Comparisons of attacks on mounted brown Jays Garrulus glandarius and black-and-white Magpies Pica pica indicated a predation cost of bright plumage, and also showed that the predation rate was frequency-dependent or 'apostatic' (Allen 1988); in years when one prey species was very common, predation rate increased for this species (Götmark 1997).

In short, there is little support for the idea that bright plumage signals unprofitability (Baker & Parker 1979), except perhaps for pitohuis and other toxic birds. Evidence cited by Baker & Parker from studies of insects, that effective escape can be the basis of prey unprofitability, has been criticised by Brower (1995). However, a related idea, for which there is some indirect evidence in birds, is 'pursuit deterrent' signals in prey (Woodland et al. 1980, Hasson 1991, referred to as 'perception advertisement' by Baker & Parker 1979). Prey that have detected the predator, especially at an early stage, are often difficult to catch for predators. In this situation, it may be beneficial for the prey to communicate or signal detection of the predator, for instance by means of colour patches on the tail (Alvarez 1993).

Savalli (1995) discuss some other hypotheses concerned with signalling and predation risk; image avoidance (relevant for polymorphic predators), predator deflection (by parents), and startle/confusion effects by brightly coloured prey.

Crypsis may be generally defined as the matching of an animal's colour pattern with the background (e.g. Savalli 1995). An alternative, more precise definition was provided in an important paper by Endler (1978): ‘a colour pattern is cryptic if it resembles a random sample of the background perceived by predators at the time and age, and in the microhabitat where the prey is most vulnerable to visually hunting predators’. This definition emphasises crypsis as seen by predators and field measurement of colour patterns and backgrounds. Endler (1978, p. 319) stated that the adaptive significance (survival value) of crypsis ‘has been borne out in numerous experimental studies’, but all the references concerned invertebrate or aquatic prey; Endler cited no studies concerned with avian prey. Savalli (1995) only devoted about half a page to crypsis and stated that the only experimental study of crypsis in birds was that of black-painted, fishing black-headed gulls (Götmark 1987, see above). This study concerned birds as predators and not as prey; it may be better to treat 'hunting camouflage' as a foraging adaptation (see above, but note that the citation by Savalli (1995) was consistent with his definition of crypsis).

A look in any field guide indicates that many or most birds are cryptic. However, crypsis may occur in different forms. There are a variety of apparently cryptic birds, with for instance streaked, barred, plain brown or green plumages, countershaded plumages, or cases of disruptive crypsis with more bold, contrasting patterns (see Cott 1940, and below). Endler's (1978) statement in his Introduction that ‘the determinants of colour patterns are poorly known’ is certainly true for cryptic birds. Crypsis may be most important for species and individuals that sit still for long periods, such as nightjars that rest during the day, and female ducks or grouse that incubate on nests in open habitats. These birds seem to blend well with their often fine-grained background. Many other apparently cryptic birds, such as warblers and flycatchers, have plain green, grey or brown plumages. They are mobile birds and I suggest that in these cases plain plumages confer equally or more effective crypsis in the foliage of trees and bushes than streaked and barred plumages.

Adult birds rely on two main behaviours to escape hunting predators: (1) remain immobile ('freeze') after detecting the predator or hearing alarm calls from other birds, or (2) seek shelter, including in a flock. In addition, some birds mob or attack predators. Although it is important for crypsis, freezing is virtually unstudied. For instance, one may ask whether it is related to plumage colouration in some way. Do both brightly coloured birds and cryptic birds freeze? Klump & Curio (1983) found that tits may freeze several minutes in response to a hawk model passing nearby (see also Shalter 1979). It is likely that immobile (freezing) birds are difficult to detect for predators; but if they are detected, they may be more easily taken than individuals that seek shelter. Therefore, birds that freeze are expected to be cryptic.

In his 1978 paper, Endler considered many factors that influence animal colour patterns and crypsis, including background matching (grain or patch size, colour, contrast, geometry, timing, and polymorphism), predation intensity, predator visual acuity and colour vision, prey-to-background distance, predator flicker fusion (movements that blend with the background), and sexual selection. Quantative studies of background matching are rare in birds. Endler (1990) emphasised and described more objective measurements of animal colour patches and background, by use of spectroradiometers and quantification of hue ('colour'), chroma (saturation), and brightness ('lightness'). The methods were recently practiced in a study of sexually dichromatic, lekking forest birds (Endler & Théry 1996). The authors found that the males of the three species were most conspicuous during the lek, due to choice of certain microhabitats for display. This should lead to effective signalling to conspecifics. At other times, conspicuousness of colour patches, and possibly also predation risk, was lower for the males (but foraging males might suffer higher predation than displaying males, at least in other species; see Götmark et al. 1997). Females were generally cryptic, as the colour patches were largely similar to those of the background.

Spectroradiometers will be important for more objective studies of animal colour patterns in the future, but one should be aware of their limitations (in present design). First, they give a sense of objectivity, but for both prey and predators we need to learn more about colour vision and the ecology of signals and light (e.g., Endler 1993) in order to interpret spectra from spectrometers. For instance, it is unclear to what extent Accipiter hawks differ from humans in colour vision, although it seems likely that they perceive UV, as do falcons and other birds (Bennett & Cuthill 1994, Viitala et al. 1995). Second, spectrometers are useful for measurements of animal colour patches and for comparisons with backgrounds (Endler & Théry 1996) but analyses of the full, complex pattern of colour patches of animals and backgrounds are still difficult to conduct. All or many of the patches in the visual field, and their interaction, should often be important for crypsis. Third, an important aspect of crypsis is the distance between prey and predator (Endler 1978), which often is much larger than that between conspecifics (prey). Changes in acuity, detail and patterns with distance are difficult to measure. Spectrometers may be most useful for quantification of colour patches used in intraspecific communication (but see Götmark 1996, 1997, Endler & Théry 1996).

Endler & Théry (1996) did not quantify the effect of distance to observer (predator) on conspicuousness of the plumages of the lekking birds, but they discussed the possible effects (p. 446) and made clear that observation distance can have marked effects. The conclusion is the same from our studies of 'long-distance conspicuousness' or detectability of mounts in the field (attack) situation. For comparisons of prey types, we used photographs of the 'prey' in natural backgrounds, first presented to humans at distances where they could not see any bird, then successively closer, until a bird was detected (Götmark & Unger 1994, Götmark 1996, 1997). This simple approach has obvious limitations. Photographs are designed for human vision, and humans probably differ from hawks in colour vision. Contrary to common belief, however, visual acuity of raptors seems to be rather similar to that of humans, mainly due to our larger eyes which improve visual acuity (Reymond 1985, 1987). Moreover, we could include in our tests the effect of context (patterns of colour patches of prey and background) and observation distance. In addition, the ‘invisable’ UV component was to some extent taken into account by spectrometry (Götmark 1996, 1997).

We obtained several surprising results; for instance, the jay (mainly brown plumage) was about equally easy to detect as the magpie (contrasting black-and white plumage) for several (brown or green) natural backgrounds (Götmark 1997). The same was true for Meadow Pipit Anthus pratensis (streaked brown) and White Wagtail Motacilla alba (black, grey and white) in various backgrounds (Götmark & Unger 1994).

More realistic tests of detectability and crypsis were done in the field (instead of using photographs) with mounted birds placed in various natural backgrounds and with human observers asked to find (paired) mounts (Götmark & Hohlfält 1995). We used males and females of two prey species, the Pied Flycatcher and the Chaffinch. Males of the Chaffinch are quite colourful (blue, red-brown and green); females of both species are grey-brown. On the ground, male and female Chaffinches hardly differed in detection distance, but in trees the males were easier to detect (detected at longer distances), mainly due to their red-brown colour. For the Pied Flycatcher, the conclusion was the reverse; males and females did not differ in detectability in trees, but males were more conspicuous on the ground. As this species spend much time in trees, our study indicate selection for crypsis in the evolution of male colouration (see also Saetre et al. 1994 for the importance of sexual selection, and Dale & Slagsvold 1996 for other studies of conspicuousness in this species). Interestingly, in three black-and-white species (white wagtail, magpie and male pied flycatcher), experimental mounts were almost as difficult to detect as 'traditionally cryptic' counterparts; thus, these black-and-white birds may therefore be examples of disruptive crypsis (Cott 1940), and cases where predation has influenced colour evolution (Götmark & Hohlfält 1995). Red may be more conspicuous in plumages: mounts of European Blackbirds (Turdus merula) with a red (painted) wing patch were easier to detect than normal mounts (see Götmark 1996).

The word 'conspicuousness' is often confusing in discussions on bird colouration, sexual selection and predation risk. In many cases 'colourfulness' or perhaps 'brightness' are be more appropriate words; conspicuousness implies ‘easy to detect’ and this is rarely tested. We distinguished 'close quarter conspicuousness' (‘field guide appearance’) and 'long-distance conspicuousness' (Götmark & Unger 1994, Götmark 1996, 1997). de Repentigny et al. (1997) suggested a new approach to quantify conspicuousness, but it is unclear to what extent this method is useful for analysing crypsis or detectability of plumages. The authors first quantified chroma and brightness of colour patches of plumages, and by adding one (green) background in the analyses they stated that long-distance conspicuousness was taken into account. The background is, however, only one component of long-distance conspicuousness; one also needs to analyse how viewing distance affect detectability of the prey (bird), as seen in a natural background. The size of the bird changes with distance, as does the context (different background patches, and their connection to prey colour patches). This can easily be seen by placing and testing mounts in the field. In a simplified form, it may be possible to use the photographic method (Götmark & Unger 1994) also for comparative studies of many species, if relevant backgrounds are available.

One limitation of our studies of mounts (as well as the study of Endler & Théry 1996) is that movements by the birds are not taken into account in the analyses. It is possible or likely that bright or contrasting birds are much easier to detect than cryptic ones when they walk or fly. Birds are highly mobile animals and are probably often detected by predators through their movements (Götmark et al. 1997). However, the results should at least be applicable to birds that attempt to escape predators by 'freezing'. In addition, it is possible that certain mobile, bright or contrasting species are cryptic against contrasting backgrounds (see Endler 1978, for flicker fusion).

Beside descriptive studies of crypsis, or simple experiments like the ones described above, we need experimental studies of live birds, where colouration is manipulated. One interesting question is how bright colours (signals) initially may evolve, given selection for crypsis. Behavioural ecologists have usually assumed that conspicuous mutants suffer increased predation; however, since predation may be frequency-dependent or 'apostatic' (Allen 1988) a rare mutant might face lower predation rate and increase in frequency as long as the trait is rare in the population (see Endler 1991 for scenarios of initial evolution). Earlier work with mounts indicated that this latter possibility could be true: hawks attacked 'mutant' red-winged (red-painted) mounts of the European Blackbird less often than normal black or brown blackbird mounts (Götmark 1994b, 1996). For a more realistic test of this idea, we painted fledglings in broods of the Great Tit (Parus major) on underparts and head, either red (experimentals) or yellow (controls, this species is mainly black-and-yellow, and green on upper parts). As far as I know, this is the only study where plumage colouration of live birds has been manipulated to examine the influence of predation. The predators were nesting Sparrowhawks.

Our study showed higher predation rate for experimentals than controls; 5.5% vs 4.0%, respectively, although this difference was not significant when hawks pairs (n = 26) were used as sample units (Götmark & Olsson 1997). Yet, as the predation increased by 38% (4.0% vs 5.5%) in a sample of 1655 painted fledglings it seems likely that red colour led to higher risk (see Götmark & Olsson 1997 for discussion). Our study indicate selection for crypsis in red birds in woodland, at least for fledglings. Indeed, these are usually more cryptic than adults, and fledgling great tits in their dullish black-and-yellow plumage were often difficult to detect among the green leaves in the tree crowns (pers. obs.). With respect to the contrasting results for 'mutant' red-winged Blackbird mounts, one possibility is that they looked more 'aposematic', or novel, with their strongly contrasting black and red upperparts. Several woodland birds taken by nesting Sparrowhawks have reddish underparts, like the red-painted Great Tits, so these experimentals may have been less of a novelty (Götmark & Olsson 1997). Also, one cannot exclude the possibility that raptors treat mounts and live birds differently.

Selection for crypsis was also indicated by two other studies. First, Caldwell (1986) used model herons to study attacks by Common Black-hawks Buteogallus anthracinus on polymorphic Little Blue Herons Egretta caerulea. She found that the white, more conspicuous morph attracted more hawks. Slagsvold et al. (1995) used disappearance during the breeding season as measure of predation in Pied Flycatchers (sparrowhawks were known to nest in the area). Male flycatchers vary from brown (brown-and-white) to black (black-and-white) in plumage and the authors were able to estimate rate of disappearance for brown and black males. More black than brown males disappeared from the nesting areas, suggesting selection for crypsis.

It is possible, but remains to clarify whether 'countershading' is an important part of crypsis (Kiltie 1988). This concept describes the presence of dark dorsa and light ventra on many animals (common in birds), assumed to be an adaptive effect of shade-obliteration when illumination is from above. There seem to be no or few critical tests of the old concept of countershading (Kiltie 1988).

Non-reproductive functions of bird colours have received little attention, even though they probably are the major determinants of plumage colouration. Many or most birds have some signal component in their plumage or on their body, in many cases limited to a minor part, such as eye, bill, legs or tail (with overall cryptic appearance). A few species seem to be highly conspicuous. At present, we do not understand this variation, and have just begun quantitative studies of crypsis. Could it be that many 'bright' birds are relatively cryptic in natural habitats? Or do they represent cases where predation is a weak selective force? To understand the role of adult predation, we need basic studies of how predation risk varies in relation to ecology and behaviour in birds. For instance, in a study of 46 species of passerine birds preyed upon by sparrowhawks, Götmark & Post (1996) found highest predation risk for species foraging on or close to the ground. A prediction from this result would be that ground foraging passerines should be more cryptic than species foraging in trees (all else equal).

No research group has yet attempted to study all known functions of colouration in a single species (or in several species). For two species, our knowledge is quite extensive, covering reproductive as well as non-reproductive aspects of colouration (the great tit, e.g., Slagsvold & Lifjeld 1985, Norris 1990, 1993, Gosler 1993, references above; and the pied flycatcher, Lundberg & Alatalo 1992, references above). An interesting research programme would be to select bird species that represents various types of plumage colouration, and address the most important hypotheses and tests of them. Ideally, these species should be related, so that phylogyny and evolutionary aspects also can be studied. Such an integrated approach would be more valuable than studies of single functions of plumage colouration.

Allen, J.A. 1988. Frequency-dependent selection by predators. Philosophical Transactions of the Royal Society of London B319: 485-503.

Alvarez, F. 1993. Alertness signalling in two rail species. Animal Behaviour 46: 1229-1231.

Andersson, M. 1994. Sexual selection. Princeton Univ. Press, Princeton.

Baker, R.R. & Parker, G.A. 1979. The evolution of bird colouration. Philosophical Transactions of the Royal Society of London B 287: 63-130.

Barrowclough, G.F. & Sibley, F. C. 1980. Feather pigmentation and abrasion: test of a hypothesis. Auk 97: 881-883.

Bennett, A.T.D. & Cuthill, I.C. 1994. Ultraviolet vision in birds: what is its function. Vision Research 34: 1471-1478.

Bonser, R.H.C. 1995. Melanin and the abrasion resistance of feathers. Condor 97: 590-591.

Booth, C.L. 1990. Evolutionary significance of ontogenetic colour change in animals. Biological Journal of the Linnean Society 40: 25-163.

Bretagnolle, V. 1993. Adaptive significance of seabird colouration: the case of Procellariiformes. American Naturalist 142: 141-173.

Brower, A.V.Z. 1995. Locomotor mimicry in butterflies? A critical review of the evidence. Proceedings of the Royal Society of London B347: 413-425.

Burtt, E.H., Jr. 1979. Tips on wings and other things. In: Burtt, E.H., Jr (ed.) The behavioural significance of colour, pp. 75-100. Garland Press, New York.

Burtt, E.H., Jr. 1981. The adaptiveness of animal colours. BioScience 31: 723-729.

Burtt, E.H., Jr. 1984. Colour of the upper mandible: an adaptation to reduce reflectance. Animal Behaviour 32: 652-658.

Butcher, G.S. & Rohwer, S. 1989. The evolution of conspicuous and distinctive colouration for communication in birds. Current Ornithology 6: 51-108.

Caldwell, G.S. 1986. Predation as a selective force on foraging herons: Effects of plumage colour and flocking. Auk 103: 494-505.

Cott, H.B. 1940. Adaptive colouration in animals. Methuen, London.

Cott, H.B. 1947. The edibility of birds: illustrated by five years' experiments and observations (1941-1946) on the food preferences of the hornet, cat and man; and considered with special reference to the theories of adaptive colouration. Proceedings of the Zoological Society of London 116: 371-524.

Cott, H.B. & Benson, C.W. 1970. The palatability of birds, mainly based on observations of a tasting panel in Zambia. Ostrich Suppl 8: 357-384.

Dale, S. & Slagsvold, T. 1996. Plumage colouration and conspicuousness in birds: experiments with the pied flycatcher. Auk 113: 849-857.

Diamond, J.M. 1994. Two-faced mimicry. Nature, Lond. 367: 683-684.

Dumbacher, J.P., Beehler, B.M., Spande, T.F., Garraffo, H.M. & Daly, J.W. 1992. Homobatrachotoxin in the genus Pitohui: chemical defence in birds. Science 258: 799-801.

Dumbacher, J.P & Pruett-Jones, S. 1996. Avian chemical defense. Current Ornithology 13: 137-174.

Endler, J.A. 1978. A predator's view of animal colour patterns. Evolutionary Biology 11: 319-64.

Endler, J.A. 1990. On the measurement and classification of colour in studies of animal colour patterns. Biological Journal of the Linnean Society 41: 315-352.

Endler, J.A. 1991. Interactions between predators and prey. In: Krebs J.R., Davies N.B. (eds) Behavioural ecology: an evolutionary approach. Blackwell Scientific Publications, Oxford: 169-201.

Endler, J.A. 1993. The colour of light in forests and its implications. Ecological Monographs 63: 1-27.

Endler, J.A. & Théry, M. 1996. Interacting effects of lek placement, display behavior, ambient light, and colour patterns in three neotropical forest-dwelling birds. American Naturalist 148: 421-452.

Gosler, A. 1993. The great tit. London, Hamlyn.

Götmark, F. 1987. White underparts in gulls function as hunting camouflage. Animal Behaviour 35: 1786-1792.

Götmark, F. 1992a. Blue eggs do not reduce nest predation in the song thrush. Behavioural Ecology & Sociobiology 30: 245-252.

Götmark, F. 1992b. Anti-predator effect of conspicuous plumage in a male bird. Animal Behaviour 44: 51-55.

Götmark, F. 1993. Conspicuous colouration in male birds is favoured by predation in some species and disfavoured in others. Proceedings of the Royal Society of London, B253: 143-146.

Götmark, F. 1994a. Are bright birds distasteful? A reanalysis of H. B. Cott's data on the edibility of birds. Journal Avian Biology 25:184-197.

Götmark, F. 1994b. Does a novel bright colour patch increase or decrease predation? Red wings reduce predation risk in European blackbirds. Proceedings of the Royal Society of London B 256: 83-87.

Götmark, F. 1995. Black-and-white plumage in male pied flycatchers reduces the risk of predation from sparrowhawks during the breeding season. Behavioural Ecology 6: 22-26.

Götmark, F. 1996. Simulating a colour mutation: conspicuous red wings in the European blackbird reduce the risk of attacks by sparrowhawks. Funct. Ecol. 10: 355-359.

Götmark, F. 1997. Bright plumage in the magpie: does it reduce or increase the risk of attacks by goshawks? Behavioural Ecology & Sociobiology 40: 41-49.

Götmark, F. & Ahlström, M. 1997. Parental preference for red mouth of chicks in a songbird. Proceedings of the Royal Society of London B264: 959-962.

Götmark, F. & Hohlfält, A. 1995. Bright male plumage and predation risk in passerine birds: are males easier to detect than females? Oikos 74: 475-484

Götmark, F. & Olsson, J. 1997 Artificial colour mutation: do red-painted great tits experience increased or decreased predation? Animal Behaviour 53: 83-91.

Götmark, F. & Post, P. 1996. Prey selection by sparrowhawks Accipiter nisus: relative predation risk for breeding passerine birds in relation to their size, ecology, and behaviour. Philosophical Transactions of the Royal Society of London 351: 1559-1577.

Götmark, F., Post, P., Olsson, J. & Himmelmann, D. 1997. Natural selection and sexual dimorphism: sex-biased sparrowhawk predation favours crypsis in female chaffinches. Oikos 80: 540-548.

Götmark, F. & Unger, U. 1994. Are conspicuous birds unprofitable prey? Field experiments with hawks and stuffed prey species. Auk 111: 251-262.

Götmark, F., Winkler, D. W. & Andersson, M. 1986. Flock-feeding on fish schools increases individual success in gulls. Nature, Lond. 319: 589-591.

Guilford, T. 1990. The evolution of aposematism. In: Evans, D.L. & Schmidt, J.O. (eds) Insect defence: adaptive mechanisms and strategies of prey and predators. State University of New York Press, New York: 23-61.

Harrison, C. 1985. A field guide to the nests, eggs and nestlings of European birds. Collins, London.

Hasson, O. 1991. Pursuit-deterrent signals: communication between prey and predator. Trends in Ecology and Evolution 6: 325-329.

Hoffman, W., Heinemann, D. & Wiens, J.A. 1981. The ecology of seabird feeding flocks in Alaska. Auk 98: 437-456.

Jablonski, P. 1996. Dark habitats and bright birds: warblers may use wing patches to flush prey. Oikos 75: 350-352.

Jones, I.L. & Hunter, F.M. 1993. Mutual sexual selection in a monogamous seabird. Nature, Lond. 362: 238-239.

Kilner, R. 1997. Mouth colour is an honest signal of need in begging canary nestlings. Proceedings of the Royal Society of London B264: 963-968.

Kilner, R. & Davies, N.B. 1998. Nestling mouth colour: ecological correlates of a begging signal. Animal Behaviour 56: 705-712.

Kiltie, R.A. 1988. Countershading: Universally deceptive or deceptively universal? Trends in Ecology and Evolution 3: 21-23.

Klump, G.M. & Curio, E. 1983. Reactions of blue tits to hawk models of different sizes. Bird Behaviour 4: 78-81.

Krebs, J.R. 1979. Bird colours. Nature, Lond. 282: 14-16.

Lundberg, A. & Alatalo, R.V. 1992. The pied flycatcher. London, Poyser.

Lyon, B.E., Eadie, J.M. & Hamilton, L.D. 1994. Parental choice selects for ornamental plumage in American coot chicks. Nature, Lond. 371: 240-243.

Martin, T.E. & Badyaev, A.V. 1996. Sexual dichromatism in birds: importance of nest predation and nest location for females versus males. Evolution 50: 2454-2460.

Newton, I. 1986. The sparrowhawk. Poyser.

Norris, K.J. 1990. Female choice and the evolution of the conspicuous plumage colouration of monogamous male great tits. Behavioural Ecology & Sociobiology 26: 129-138.

Norris, K.J. 1993. Heritable variation in a plumage indicator of viability in male great tits Parus major. Nature 362: 537-539.

Pierotti, R. 1987. Isolating mechanisms in seabirds. Evolution 41: 559-570.

Pough, F.H. 1988. Mimicry of vertebrates: are the rules different? American Naturalist 131: Suppl., 67-102.

de Repentigny, Y., Ouellet, H. & McNeil, R. 1997. Quantifying conspicuousness and sexual dimorphism of the plumage in birds: a new approach. Canadian Journal of Zoology 75: 1972-1981.

Reymond, L. 1985. Spatial visual acuity of the eagle Aquila audax: A behavioural, optical and anatomical investigation. Vision Res 25: 1477-1491.

Reymond, L. 1987. Spatial visual acuity of the falcon, Falco berigora: A behavioural, optical and anatomical investigation. Vision Res 27: 1859-1874.

Rohwer, S.A. 1975. The social significance of avian winter plumage variability. Evolution 29: 593-610.

Saetre, G.-P., Dale, S. & Slagsvold, T. 1994. Female pied flycatchers prefer brightly coloured males. Animal Behaviour 48: 1407-1416.

Savalli, U.M. 1995. The evolution of bird colouration and plumage elaboration. Current Ornithology 12:141-190.

Senar, J.C. 1999. Plumage colouration as a signal of social status. In: Adams, N.J. & Slotow, R.H. (eds) Proc.22 Int. Ornithol. Congr., Durban.

Shalter, M.D. 1979. Responses of nesting passerines to alarm calls. Ibis 121: 362-368.

Simmons, K.E.L. 1972. Some adaptive features of seabird plumage types. British Birds 65: 465-479, 510-521.

Slagsvold, T., Dale, S. and Kruszewics, A. 1995. Predation favours cryptic colouration in breeding male pied flycatchers (Ficedula hypoleuca). Animal Behaviour 50: 1109-1121.

Slagsvold, T. & Lifjeld, J. 1985. Variation in plumage colour of the Great tit Parus major in relation to habitat, season and food. Journal of Zoology 206: 321-328.

Spear, L. & Ainley, D. 1993. Kleptoparasitism by Kermadec Petrels, jaegers, and skuas in the eastern tropical Pacific: evidence of mimicry by two species of Pterodroma. Auk 110: 222-233.

Viitala, J., Korpimäki, E., Palokangas, P. & Kolvula, M. 1995. Attraction of kestrels to vole scent marks visible in ultraviolet light. Nature 373: 425-427.

Wallace, A. R. 1889. Darwinism. Macmillan, London.

Walsberg, G. E. 1982. Coat colour, solar heat gain, and conspicuousness in the phainopepla. Auk 99: 495-502.

Walsberg, G. E. 1988. Solar heat gain in birds: consequences of plumage colour, structure and optics. In: Ouellet, H. (ed.), Acta XIX Congressus Internationalis Ornithologici, Vol. 1. Univ. Ottawa Press, Ottawa.

Westmoreland, D. & Kiltie, R. A. 1996. Egg crypsis and cluch survival in three species of blackbirds (Icteridae). Biological Journal of the Linnean Society 58: 159-172.

Woodland, D.J., Jaafar, Z. & Knight, M.L. 1980. The 'pursuit deterrent' function of alarm signals. American Naturalist 115: 748-753.