Artificial neural networks and the study of evolution of prey coloration

In this paper, I investigate the use of artificial neural networks in the study of prey coloration. I briefly review the anti-predator functions of prey coloration and describe both in general terms and with help of two studies as specific examples the use of neural network models in the research on prey coloration. The first example investigates the effect of visual complexity of background on evolution of camouflage. The second example deals with the evolutionary choice of defence strategy, crypsis or aposematism. I conclude that visual information processing by predators is central in evolution of prey coloration. Therefore, the capability to process patterns as well as to imitate aspects of predator's information processing and responses to visual information makes neural networks a well-suited modelling approach for the study of prey coloration. In addition, their suitability for evolutionary simulations is an advantage when complex or dynamic interactions are modelled. Since not all behaviours of neural network models are necessarily biologically relevant, it is important to validate a neural network model with empirical data. Bringing together knowledge about neural networks with knowledge about topics of prey coloration would provide a potential way to deepen our understanding of the specific appearances of prey coloration.     

Linking the evolution and form of warning coloration in nature

Many animals are toxic or unpalatable and signal this to predators with warning signals (aposematism). Aposematic appearance has long been a classical system to study predator–prey interactions, communication and signalling, and animal behaviour and learning. The area has received considerable empirical and theoretical investigation. However, most research has centred on understanding the initial evolution of aposematism, despite the fact that these studies often tell us little about the form and diversity of real warning signals in nature. In contrast, less attention has been given to the mechanistic basis of aposematic markings; that is, ‘what makes an effective warning signal?’, and the efficacy of warning signals has been neglected. Furthermore, unlike other areas of adaptive coloration research (such as camouflage and mate choice), studies of warning coloration have often been slow to address predator vision and psychology. Here, we review the current understanding of warning signal form, with an aim to comprehend the diversity of warning signals in nature. We present hypotheses and suggestions for future work regarding our current understanding of several inter-related questions covering the form of warning signals and their relationship with predator vision, learning, and links to broader issues in evolutionary ecology such as mate choice and speciation.     

The protective value of conspicuous signals is not impaired by shape, size, or position asymmetry

Various conspicuous signals in nature promote initial and learnedavoidance by predators. It is widely thought that such signalsare most effective when highly symmetrical in features suchas size and shape, supported by recent laboratory experimentswith domestic chicks and artificial prey. However, no studyhas investigated the effect of asymmetry on conspicuous signalsin a natural setting, where viewing distances, angles, predatorspecies, and light conditions vary and where predators encounterprey sequentially rather than simultaneously. We undertook 2field experiments with artificial gray-scale prey, marked witha pair of white markings presented to wild avian predators,to test the effect of asymmetry on the survival value of conspicuoussignals in the field. Experiment 1 had treatments with symmetricalspots or with spots asymmetrical in area between 5 and 50%.All marked treatments survived better than unmarked controls,but there was no benefit of being symmetrical. Experiment 2tested the effect of possessing markings asymmetrical for shapeor position and any additive effect of these 2 features. Again,symmetry conferred no benefit and targets with markings asymmetricalfor position and/or shape survived equally well as those withsymmetrical arrangements. These findings indicate that asymmetryin warning signals may not be costly to prey in nature or beof less importance compared with other features of the signal,such as color and overall size.     

Towards an ecology of protective coloration

The strategies underlying different forms of protective coloration are well understood but little attention has been paid to the ecological, life-history and behavioural circumstances under which they evolve. While some comparative studies have investigated the ecological correlates of aposematism, and background matching, the latter particularly in mammals, few have examined the ecological correlates of other types of protective coloration. Here, we first outline which types of defensive coloration strategies may be exhibited by the same individual; concluding that many protective coloration mechanisms can be employed simultaneously, particularly in conjunction with background matching. Second, we review the ecological predictions that have been made for each sort of protective coloration mechanism before systematically surveying phylogenetically controlled comparative studies linking ecological and social variables to antipredator defences that involve coloration. We find that some a priori predictions based on small-scale empirical studies and logical arguments are indeed supported by comparative data, especially in relation to how illumination affects both background matching and self-shadow concealment through countershading; how body size is associated with countershading, motion dazzle, flash coloration and aposematism, although only in selected taxa; how immobility may promote background matching in ambush predators; and how mobility may facilitate motion dazzle. Examination of nearly 120 comparative tests reveals that many focus on ecological variables that have little to do with predictions derived from antipredator defence theory, and that broad-scale ecological studies of defence strategies that incorporate phylogenetics are still very much in their infancy. We close by making recommendations for future evolutionary ecological research.     

Perspective: the evolution of warning coloration is not paradoxical

Animals that are brightly colored have intrigued scientists since the time of Darwin, because it seems surprising that prey should have evolved to be clearly visible to predators. Often this self-advertisement is explained by the prey being unprofitable in some way, with the conspicuous warning coloration helping to protect the prey because it signals to potential predators that the prey is unprofitable. However, such signals only work in this way once predators have learned to associate the conspicuous color with the unprofitability of the prey. The evolution of warning coloration is still widely considered to be a paradox, because it has traditionally been assumed that the very first brightly colored individuals would be at an immediate selective disadvantage because of their greater conspicuousness to predators that are naive to the meaning of the signal. As a result, it has been difficult to understand how a novel conspicuous color morph could ever avoid extinction for long enough for predators to become educated about the signal. Thus, the traditional view that the evolution of warning coloration is difficult to explain rests entirely on assumptions about the foraging behavior of predators. However, we review recent evidence from a range of studies of predator foraging decisions, which refute these established assumptions. These studies show that: (1) Many predators are so conservative in their food preferences that even very conspicuous novel prey morphs are not necessarily at a selective disadvantage. (2) The survival and spread of novel color morphs can be simulated in field and aviary experiments using real predators (birds) foraging on successive generations of artificial prey populations. This work demonstrates that the foraging preferences of predators can regularly (though not always) result in the increase to fixation of a novel morph appearing in a population of familiar-colored prey. Such fixation events occur even if both novel and familiar prey are fully palatable and despite the novel food being much more conspicuous than the familiar prey. These studies therefore provide strong empirical evidence that conspicuous coloration can evolve readily, and repeatedly, as a result of the conservative foraging decisions of predators.     

Empirical tests of the role of disruptive coloration in reducing detectability

Disruptive patterning is a potentially universal camouflage technique that is thought to enhance concealment by rendering the detection of body shapes more difficult. In a recent series of field experiments, artificial moths with markings that extended to the edges of their 'wings' survived at higher rates than moths with the same edge patterns inwardly displaced. While this result seemingly indicates a benefit to obscuring edges, it is possible that the higher density markings of the inwardly displaced patterns concomitantly reduced their extent of background matching. Likewise, it has been suggested that the mealworm baits placed on the artificial moths could have created differential contrasts with different moth patterns. To address these concerns, we conducted controlled trials in which human subjects searched for computer-generated moth images presented against images of oak trees. Moths with edge-extended disruptive markings survived at higher rates, and took longer to find, than all other moth types, whether presented sequentially or simultaneously. However, moths with no edge markings and reduced interior pattern density survived better than their high-density counterparts, indicating that background matching may have played a so-far unrecognized role in the earlier experiments. Our disruptively patterned non-background-matching moths also had the lowest overall survivorship, indicating that disruptive coloration alone may not provide significant protection from predators. Collectively, our results provide independent support for the survival value of disruptive markings and demonstrate that there are common features in human and avian perception of camouflage.     

Predator mimicry,not conspicuousness,explains the efficacy of butterfly eyespots

Large conspicuous eyespots on butterfly wings have been shown to deter predators. This has been traditionally explained by mimicry of vertebrate eyes, but recently the classic eye-mimicry hypothesis has been challenged. It is proposed that the conspicuousness of the eyespot, not mimicry, is what causes aversion due to sensory biases, neophobia or sensory overloads. We conducted an experiment to directly test whether the eye-mimicry or the conspicuousness hypothesis better explain eyespot efficacy. We used great tits (Parus major) as model predator, and tested their reaction towards animated images on a computer display. Birds were tested against images of butterflies without eyespots, with natural-looking eyespots, and manipulated spots with the same contrast but reduced resemblance to an eye, as well as images of predators (owls) with and without eyes. We found that mimetic eyespots were as effective as true eyes of owls and more efficient in eliciting an aversive response than modified, less mimetic but equally contrasting eyespots. We conclude that the eye-mimicry hypothesis explains our results better than the conspicuousness hypothesis and is thus likely to be an important mechanism behind the evolution of butterfly eyespots.     

Defining disruptive coloration and distinguishing its functions

Disruptive coloration breaks up the shape and destroys the outline of an object, hindering detection. The principle was first suggested approximately a century ago, but, although research has significantly increased, the field remains conceptually unstructured and no unambiguous definition exists. This has resulted in variable use of the term, making it difficult to formulate testable hypotheses that are comparable between studies, slowing down advancement in this field. Related to this, a range of studies do not effectively distinguish between disruption and other forms of camouflage. Here, we give a formal definition of disruptive coloration, reorganize a range of sub-principles involved in camouflage and argue that five in particular are specifically related to disruption: differential blending; maximum disruptive contrast; disruption of surface through false edges; disruptive marginal patterns; and coincident disruptive coloration. We discuss how disruptive coloration can be optimized, how it can relate to other forms of camouflage markings and where future work is particularly needed.     

Palatablility and escaping ability in Neotropical butterflies: tests with wild kingbirds (Tyrannus melancholicus, Tyrannidae)

The palatability and the ability of neotropical butterflies to escape after being detected, attacked and captured by wild kingbirds ( Tyrannus melancholicus Vieillot), was investigated by the release of 668 individuals of 98 butterfly species close to the birds, during their usual feeding activities. Most of the butterflies were attacked and eaten. Only the troidine swallowtails ( Parities and Battus ; Papilionidae) were consistently rejected on taste and elicited aversive behaviours in birds. Most other aposematic and/or mimetic species in the gehera Danaus and Lycorea (Danainae), Dione, Eueides and Heliconius (Heliconiinae), Hypothyris, Mechanitis and Melinaea (Ithomfinae), Biblis, Callicore and Diaethria (Limenitidinae) were generally eaten. Cryptic and non-mimetic species were always attacked and, if captured, they were also eaten. All Apaturinae, Charaxinae, Nymphalinae, Hesperidae, most Limenitidinae, Heliconiinag ( Agraulis, Dryas, Dryadula and Philaethria ) and Papilionidae ( Eurytides, Heraclides and Protesilaus ) were in this group. Results indicate that the learning process in kingbirds may demand a large mortality in prey populations, even among species generally accepted as unpalatable and aposematic. They also support the assertion that escaping ability and unpalatability evolved in butterflies as alternative strategies to avoid predation by birds. Mimetic relationships among several species are discussed. Evidence for the evolution of aposematism not related to unpalatability, but to escaping ability, was found for two hard-to-catch Morpho species.     

Neophobia and Dietary Conservatism:Two Distinct Processes?

Birds show distinct hesitation when approaching novel food and this has been termed ‘neophobia’. In laboratory-held birds like domestic chicks this effect lasts for a matter of a few minutes at most, but hesitant attack of novel foods can last for weeks or even months in wild birds. This effect, called ‘dietary conservatism’, seems to be a different type of learning process from neophobia as first described and has hitherto been largely overlooked. This paper presents some evidence for the view that the processes may be fundamentally different. We outline results from laboratory chicks that show neophobia to be easily deactivated by experience, which renders it unlikely to be an important force in wild birds. We also report evidence that the process of incorporation of novel food into the diet is not a simple one-stage process but includes at least four steps of assessment. The paper concludes with an outline of the importance of dietary conservatism in our understanding of the evolution of aposematism and the workings of mimicry. This revised version was published online in July 2006 with corrections to the Cover Date.     

Dazzle coloration and prey movement

Many traits in animals reduce the rate of attack from visually hunting predators, including camouflage, warning signals and mimicry. In addition, some animal markings may reduce the likelihood that an attack ends in successful capture. These might include dazzle markings, high-contrast patterns that make the estimation of speed and trajectory difficult. However, until now, no study has experimentally tested whether some markings may achieve such an effect. We developed a computer 'game' where human 'predators' have to capture computer-generated prey moving across a background. In two experiments, we find that although uniform camouflaged targets were among the hardest to capture, so were a range of high-contrast conspicuous patterns, such as bands and zigzags. Prey were also more difficult to capture against more heterogeneous than uniform backgrounds, and at faster speeds of movement. As such, we find the first experimental evidence that conspicuous patterns, similar to those found in a wide range of real animals, make the capture of moving prey more challenging. Various anti-predator markings may work prey during motion, and some animals may combine such dazzle patterns with other functions, such as camouflage, thermoregulation, sexual and warning signals.     

Constrained camouflage facilitates the evolution of conspicuous warning coloration

The initial evolution of aposematic and mimetic antipredator signals is thought to be paradoxical because such coloration is expected to increase the risk of predation before reaching a stage when predators associate it effectively with a defense. We propose, however, that constraints associated with the alternative strategy, cryptic coloration, may facilitate the evolution of antipredator signals and thus provide a solution for the apparent paradox. We tested this hypothesis first using an evolutionary simulation to study the effect of a constraint due to habitat heterogeneity, and second using a phylogenetic comparison of the Lepidoptera to investigate the effect of a constraint due to prey motility. In the evolutionary simulation, antipredator warning coloration had an increased probability to invade the prey population when the evolution of camouflage was constrained by visual difference between microhabitats. The comparative study was done between day-active lepidopteran taxa, in which camouflage is constrained by motility, and night-active taxa, which rest during the day and are thus able to rely on camouflage. We compared each of seven phylogenetically independent day-active groups with a closely related nocturnal group and found that antipredator signals have evolved at least once in all the diurnal groups but in none of their nocturnal matches. Both studies lend support to our idea that constraints on crypsis may favor the evolution of antipredator warning signals.     

Warning coloration can be disruptive: aposematic marginal wing patterning in the wood tiger moth

Warning (aposematic) and cryptic colorations appear to be mutually incompatible because the primary function of the former is to increase detectability, whereas the function of the latter is to decrease it. Disruptive coloration is a type of crypsis in which the color pattern breaks up the outline of the prey, thus hindering its detection. This delusion can work even when the prey's pattern elements are highly contrasting; thus, it is possible for an animal's coloration to combine both warning and disruptive functions. The coloration of the wood tiger moth (Parasemia plantaginis) is such that the moth is conspicuous when it rests on vegetation, but when it feigns death and drops to the grass‐ and litter‐covered ground, it is hard to detect. This death‐feigning behavior therefore immediately switches the function of its coloration from signaling to camouflage. We experimentally tested whether the forewing patterning of wood tiger moths could function as disruptive coloration against certain backgrounds. Using actual forewing patterns of wood tiger moths, we crafted artificial paper moths and placed them on a background image resembling a natural litter and grass background. We manipulated the disruptiveness of the wing pattern so that all (marginal pattern) or none (nonmarginal pattern) of the markings extended to the edge of the wing. Paper moths, each with a hidden palatable food item, were offered to great tits (Parus major) in a large aviary where the birds could search for and attack the “moths” according to their detectability. The results showed that prey items with the disruptive marginal pattern were attacked less often than prey without it. However, the disruptive function was apparent only when the prey was brighter than the background. These results suggest that warning coloration and disruptive coloration can work in concert and that the moth, by feigning death, can switch the function of its coloration from warning to disruptive.     

Predation on snakes of Argentina: Effects of coloration and ring pattern on coral and false coral snakes

The occurrence of coral snake coloration among unrelated venomous and non‐venomous snake species has often been explained in terms of warning coloration and mimicry. In Argentina, no field tests have been conducted to confirm this mimetic association between one venomous coral species (Micrurus phyrrocryptus, Elapidae) and two non‐venomous snake species with a similar color pattern (Lystrophis pulcher and Oxyrhopus rhombifer, Colubridae). The aims of this work were to test for the possible aposematic or cryptic function of the ring pattern and coloration of coral snakes and false coral snakes from central Argentina, and to analyse whether the pattern is effective throughout the year. Predation on snakes was estimated by using non‐toxic plasticine replicas of ringed venomous and non‐venomous snakes and unbanded green snakes placed along transects in their natural habitat during the dry and rainy season. Ringed color pattern was attacked by predators despite the background color. One of the replica types was attacked more than expected during the dry season, suggesting that both shape and width of rings may influence the choice by predators. The reaction of predators towards replicas that mimic snake species with ringed patterns is independent of the geographical region, and we can conclude that mimicry characteristics are quite general when the true models are present in the area.     

Camouflage and visual perception

How does an animal conceal itself from visual detection by other animals? This review paper seeks to identify general principles that may apply in this broad area. It considers mechanisms of visual encoding, of grouping and object encoding, and of search. In most cases, the evidence base comes from studies of humans or species whose vision approximates to that of humans. The effort is hampered by a relatively sparse literature on visual function in natural environments and with complex foraging tasks. However, some general constraints emerge as being potentially powerful principles in understanding concealment—a ‘constraint’ here means a set of simplifying assumptions. Strategies that disrupt the unambiguous encoding of discontinuities of intensity (edges), and of other key visual attributes, such as motion, are key here. Similar strategies may also defeat grouping and object-encoding mechanisms. Finally, the paper considers how we may understand the processes of search for complex targets in complex scenes. The aim is to provide a number of pointers towards issues, which may be of assistance in understanding camouflage and concealment, particularly with reference to how visual systems can detect the shape of complex, concealed objects.     

Aposematic coloration, luminance contrast, and the benefits of conspicuousness

Many organisms use warning, or aposematic, coloration to signaltheir unprofitability to potential predators. Aposematicallycolored prey are highly visually conspicuous. There is considerableempirical support that conspicuousness promotes the effectivenessof the aposematic signal. From these experiments, it is welldocumented that conspicuous, unprofitable prey are detectedsooner and aversion learned faster by the predator as comparedwith cryptic, unprofitable prey. Predators also retain memoryof the aversion longer when prey is conspicuous. The presentstudy focused on the elements of conspicuousness that conferthese benefits of aposematic coloration. Drawing on currentunderstanding of animal vision, we distinguish 2 features ofwarning coloration: high chromatic contrast and high brightness,or luminance, contrast. Previous investigations on aposematicsignal efficacy have focused mainly on the role of high chromaticcontrast between prey and background, whereas little researchhas investigated the role of high luminance contrast. Usingthe Chinese mantid as a model predator and gray-painted milkweedbugs as model prey, we found that increased prey luminance contrastincreased detection of prey, facilitated predator aversion learning,and increased predator memory retention of the aversive response.Our results suggest that the luminance contrast component ofaposematic coloration can be an effective warning signal betweenthe prey and predator. Thus, warning coloration can even evolveas an effective signal to color blind predators.     

The effects of pattern symmetry on detection of disruptive and background-matching coloration

Two, logically distinct but sometimes compatible, mechanismsof camouflage are background-matching and disruptive coloration.In the former, an animal's coloration comprises a random sampleof the background, and so target–background discriminationis impeded. In the latter, object or feature recognition iscompromised by placing bold, high-contrast colors so that theybreak up the prey's body into apparently unconnected objects.Recent experimental evidence for the utility of disruptive colors,above and beyond that conferred by background matching, hasbeen based on artificial prey with patterns lacking a planeof symmetry. However, it is plausible that the bilateral symmetrypresent in natural prey may compromise the efficiency of disruptivecoloration, on account of the potency of symmetry as a cue invisual search. In this study, we tested this prediction in thefield, by tracking the "survival" under bird predation of artificialmothlike targets placed on oak trees. These had background-matchingcolor patches placed either disruptively or nondisruptivelyand with or without bilateral symmetry. We found that symmetryreduced the effectiveness of both nondisruptive and disruptivebackground-matching coloration to a similar degree so that thenegative effects of symmetry on concealment are no greater fordisruptive than nondisruptive patterns.     

Testing the role of background matching and self‐shadow concealment in explaining countershading coloration in wild‐caught rainbowfish

Countershading, or dorsal pigmentary darkening (DPD), describes a form of vertically varying coloration, where an animal typically has a dark dorsal surface and a paler ventral side, and is widespread among mammals, birds, reptiles, fishes and insects. DPD is thought to confer concealment from predators and, in terrestrial systems, there is good evidence that the dark–light transition in body coloration acts to conceal the body's shadow. Surprisingly few studies of DPD have been conducted in aquatic environments, and thus it is not known whether the mechanisms of concealment are similar to those that operate in terrestrial habitats. In this study, we determined the role of the light environment and predation risk in determining DPD in wild‐caught populations of a freshwater fish, the western rainbowfish (Melanotaenia australis). We also examined the underlying mechanisms of DPD for concealment by testing the assumptions of background matching and self‐shadow concealment. In a subsequent experiment, we determined whether any observed variation in DPD was maintained when the visual background was manipulated in the laboratory (to induce a change in body coloration). We found that both the amount of downwelling irradiance and the level of predation risk at the collection site affected skin darkness (dorsal, ventral and overall), whereas the ratio of dorsoventral coloration (DPD) was not affected by the parameters considered. The laboratory experiment revealed that fish changed their body coloration to match their visual background, and did so by altering the relative ratio of dorsoventral skin darkness. In contrast with research on terrestrial animals, our findings suggest that the most likely method of achieving crypsis is through background matching, rather than self‐shadow concealment. It is thus possible that differences in the optical characteristics of terrestrial and aquatic environments, and/or variation in the angles at which prey are typically viewed and attacked, have resulted in divergent mechanisms of using DPD to attain crypsis. © 2015 The Linnean Society of London, Biological Journal of the Linnean Society, 2015, 114 , 915–928.     

Predator response to the coloured eyespots and defensive posture of Colombian four-eyed frogs

Deimatic displays, where sudden changes in prey appearance elicit aversive predator reactions, have been suggested to occur in many taxa. These (often only putative) displays frequently involve different components that may also serve antipredator functions via other mechanisms (e.g., mimicry, warning signalling, body inflation). The Colombian four-eyed frog, Pleurodema brachyops, has been suggested to gain protection against predation through putative deimatic displays where they inflate and elevate the posterior part of their body revealing eye-like colour markings. We exposed stationary artificial frogs to wild predators to test whether the two components (eyespot/colour markings, defensive posture) of their putative deimatic display, and their combination, provide protection from predation without the sudden change in appearance. We did not detect any obvious additive effect of defensive posture and eyespots/colour markings on predation risk, but found a marginally significant trend for model frogs in the resting posture to be less attacked when displaying eyespots/colour markings than when they were not, suggesting that the presence of colour markings/eyespots may provide some protection on its own. Additionally, we found that models in a resting posture were overall more frequently attacked on the head than models in a defensive posture, indicating that a defensive posture alone could help redirect predator attacks to non-vital parts of the body. The trends found in our study suggest that the different components of P. brachyops' coloration may serve different functions during a deimatic display, but further research is needed to elucidate the role of each component when accompanied by sudden prey movement.