Chapter 1

INTRODUCTION AND CLASSIFICATION OF MIMICRY SYSTEMS

It is hardly an exaggeration to say, that whilst reading and reflecting on the various facts given in this Memoir, we feel to be as near witnesses, as we can ever hope to be, of the creation of a new species on this earth.

Charles Darwin (1863) referring to Henry Bates’ 1862 account of mimicry in Brazil

COPYRIGHTED MATERIAL

Mimicry, Crypsis, Masquerade and other Adaptive Resemblances, First Edition. Donald L. J. Quicke. © 2017 John Wiley & Sons Ltd. Published 2017 by John Wiley & Sons Ltd.

1

0003114056.INDD 1 7/14/2017 12:48:07 PM 2 Donald L. J. Quicke

A BRIEF HISTORY at Oxford University. He describes the results of extensive experiments in which insects were presented to a captive The first clear definition of biological mimicry was that of monkey and its responses observed. The article is over 100 Henry Walter Bates (1825–92), a British naturalist who pages long and in the foreword he notes that a lot of the spent some 11 years collecting and researching in the observations are tabulated rather than given seriatim Amazonas region of Brazil (Bates 1862, 1864, 1981, G. because of the “great increase in the cost of printing”. Woodcock 1969). However, as pointed out by Stearn Nevertheless, such observations are essential first steps in (1981), Bates’ concept of the evolution of mimicry would understanding whether species are models or mimics or quite possibly have gone unnoticed were it not for Darwin’s have unsuspected defences. review of his book in The Natural History Review of 1863. Around the middle of the nineteenth century, another Bates’ observations of remarkable similarity between Englishman, Alfred Russel Wallace (1823–1913), an intrepid ­butterflies belonging to different families led him to ponder traveller, natural historian and thinker, was coming up with what might be the reason for this. He concluded that there important notions concerned with mimicry and aposema- must be some advantage, for example, for a ‘white butterfly’, tism (Wallace 1867). He had earlier travelled to Brazil and Dismorphia theucharila (Pieridae), to depart from the typical collected with Henry Bates and later went on to explore form and colouration of the family, and instead to resemble South‐East Asia. Indeed, he came up with the idea of evolu- unpalatable Heliconius species.1 He also noticed that in all tion by natural selection more or less contemporaneously the bright and conspicuous butterfly colour pattern com- with Charles Darwin, though unlike Darwin he had little for- plexes there was at least one species that was distasteful to mal education (H.W. Greene & McDiarmid 2005). His early predators of butterflies (Sheppard 1959). Bates was also appreciation of the nature of aposematism and thoughts on ahead of his time in his estimation of the huge and largely poisonous snake mimicry are particularly pertinent­ here. undescribed diversity of the Neotropical insect fauna. Mimicry and adaptive colouration have long been popular During his time in Amazonia he estimated that he had col- topics that have grabbed the imagination of both the public lected some 14,712 species, of which approximately 8000 and academic biologists due to the incredible detail in many were new, a number that seemed utterly implausible to most resemblances. Good early treatments include those of entomologists working in the UK at that time (Stearn 1981). Poulton (1890), G.D.H. Carpenter & Ford (1933) and Cott Some groups of insects seem to have an enormous pro- (1940), all of which document numerous natural history pensity for evolving mimicry, and within apparently closely observations and interesting ideas. Wolfgang Wickler’s related groups can have evolved to resemble models of a (1968) popular book on mimicry in plants and animals wide range of colour patterns, shapes and sizes, such as, for with many fine illustrations by H. Kacher no doubt fired example, the day‐flying, chalcosiine zygaenid moths, which many people (including myself) with enthusiasm for the are no doubt mostly or entirely Müllerian mimics (Yen et al. topic. Komárek (2003) provides an excellent and more bio- 2005), or the day‐flying Epicipeiidae moths which, with graphic description of the arguments, ideas and personali- only 20 or so species, collectively mimic various papillionid, ties that shaped our understanding of crypsis and mimicry pierid, geometrid, zygaenid and lymantriid butterfly and up until 1955 (with some comments on subsequent works moth models. No wonder this astonishing potential for up to 1990). Other good general books include Pasteur ­variation has fascinated entomologists for years. (1972), D.F. Owen (1980), Forbes (2011) and J. Diamond & A lot of early research involved the collection and Bond (2013), as well as more academic works such as ­publication of field observations and relatively simple Ruxton et al. (2004a), Stevens & Merilaita (2011) and ­experiments, such as feeding various insects to predators Stevens (2016). The book by Ruxton et al. provides a critical and observing reactions (fine examples include G.A.K. review of many experiments, models and arguments to do Marshall & Poulton 1902, Swynnerton 1915b, R.T. Young with anti‐predator adaptations in general, not just mimicry 1916, Carpenter 1942). A rather lovely, if quaint, example and camouflage, but there is a great deal of overlap. is that of G.D.H. Carpenter (1921), a medical doctor by Many arguments, often heated, were also involved in the ­profession who was based in Uganda for some time early discussions of mimicry. Some of the examples show before becoming Hope Professor of Zoology (Entomology) such perfect matching of detail that many scientists found it hard to believe that they could have resulted from natural selection for progressively more similar forms from dispa- 1. Butterfly systematics has progressed since Bates’ time and many of the species he collected and referred to as Heliconiidae are rate starting points. Some thought that only major muta- now placed in the tribe Ithomiini in the nymphalid subfamily tions could be involved rather than Darwinian gradual Danainae, while his Heliconini are now classified as a tribe of accumulation of small changes. This led to hearty debate Nymphalidae. about how natural selection and genetics work, for example

0003114056.INDD 2 7/14/2017 12:48:07 PM Introduction and classification 3

Punnett (1915) and R.B. Goldschmidt (1945) on the side of The past 30 or so years have seen an enormous resur- major mutational leaps versus R.A. Fisher (1927, 1930), gence in research on adaptive colouration and mimicry L.P. Brower et al. (1971) and, more or less, de Ruiter (1958) (Guilford 1990b, Komárek 2003), both experimental and leaning towards gradualism. The current consensus is a theoretical, as can be seen by a quick scan of the dates of the combination, with an initial mutation that causes a large articles cited here. Computer‐generated graphics, usually phenotypic shift followed by subsequent evolutionary but not always in conjunction with human subjects, have refinement, called the ‘two‐step hypothesis’, most probably played an increasingly large role in investigations. being the major route, though gradualism might be Nevertheless, much is still being achieved with low‐tech ­sufficient in some circumstances (see Chapters 4 and 5). solutions, such as pastry model caterpillars exposed to pre- As J.R.G. Turner (1983) notes, in the complicated dation by garden birds, or baited triangular shapes that Heliconius system some quite large jumps in phenotype can roughly resemble moths resting on tree trunks exposed to occur as a result of simple genetic changes. woodland birds. Increasing awareness of the visual capa- When it comes to camouflage, much credit should be bilities of predators, or in some cases of potential mates, is given not to a scientist, but instead to the American portrait, leading to quite a lot of more carefully controlled work, but animal and landscape artist Abbott Handerson Thayer there is still room for greater awareness. It is all too easy to (1849–1921), who discovered the principle of concealment think that because a model looks life‐like to the experi- by countershading, discussed disruptive colouration, and menter, it will also appear life‐like to a bird. Some birds can dazzle markings and distractive features, and even tried to see well into the UV part of the spectrum, and if the signal help the military in disguising troops and ships (J. Diamond receiver is an insect, it is important to understand that & Bond 2013). Interestingly, many of his suggestions came although insects can see UV light, most cannot see much at under attack from many naturalists and even hunters. the red end of the spectrum. While not all of his suggestions might have been correct, Sometimes biologists get it wrong. For example, for a long and indeed he probably went over the top in trying to explain while the North American viceroy butterfly, Limenitis archip- all animal colouration as having some concealing function, pus (Nymphalidae), was thought to be a Batesian mimic of the argumentation employed on both sides is of interest. the monarch butterfly, Danaus plexippus (e.g. J.V.Z. Brower People such as United States president Theodore Roosevelt, 1958a). Now it is known to be actually unpalatable itself who was an enthusiastic hunter,2 dismissed Thayer’s claim (Ritland & Brower 1991) (see Chapter 4, section Plant‐ that a zebra’s stripes acts to help conceal it (Roosevelt derived toxins), and more recently it has been shown most 1911). Thayer’s counter‐argument was that just because probably to be a Müllerian mimic (S.B. Malcolm 1990, someone saw something, it did not mean that they saw eve- Guilford 1991, Ritland 1991, Ritland & Brower 1991, rything, because they do not know what they failed to Rothschild 1991) and to contain phenolic glucosides (sali- notice. In an amusing section, Thayer wrote: cortin and tremuloidin) (Prudic et al. 2007b) sequestered from its Salix food plant, though these are rather different Forty years of daily meeting the poacher at the post office does from those of the monarch and have different physiological not strengthen his credit. And forty years of Roosevelt’s seeing zebras not hidden by their costume, and failing to guess what effects. the animal’s stripes are for, are just as little to the point. Another often neglected aspect is the need for correct identification. D.F. Owen et al. (1994), for example, discov- Kingsland (1978) wrote a very nice discussion of Thayer’s ered that anomalous findings in an African butterfly mim- work and how it was received, and two of Thayer’s oil paint- icry system were resolved once it was realised that one of ings illustrating camouflage are reproduced in J. Diamond & the mimic species was actually a pair of different but closely Bond (2013, pp. 44 and 45); Ruxton et al. (2004b) repro- similar (cryptic) species. duce Thayer’s photographs of a dead grouse positioned as ‘in nature’ and with its underside dyed darker, illustrating the effectiveness of countershading (see Chapter 2). ON DEFINITIONS OF ‘MIMICRY’ AND ADAPTIVE RESEMBLANCE

2. The large game and some 11,000 other specimens that Mimicry can be defined in many ways but in most of these Roosevelt and other expedition members shot or collected on the there exists the concept that some subject forms a model for 1909 Smithsonian–Roosevelt African Expedition became a major part of the collections of both the Smithsonian Institution the resemblance of another, the mimic. In biological exam- (Washington D.C.) and the American Museum of Natural ples, this resemblance also carries with it the notion that History (New York). Some of the diaramas in the latter are real it serves to deceive another organism, though this is not as masterpieces of natural history display. clear cut as it may seem and certainly deception is not

0003114056.INDD 3 7/14/2017 12:48:07 PM 4 Donald L. J. Quicke

Table 1.1 Some selected definitions of mimicry and adaptive resemblance. (Adapted from Endler 1981 with permission from John Wiley & Sons.)

Publication Definition

Cott 1940 ‘In the former [protective resemblance or crypsis] an animal resembles some object which is of no interest to its enemy, and in so doing is concealed; in the latter [protective mimicry] an animal resembles an object which is well known and avoided by its enemy, and in so doing becomes conspicuous.’ Wickler 1968 ‘If a signal of interest to the signal receiver is imitated, then this is a case of mimicry, whereas if the general uninteresting background or substrate is imitated, then camouflage (or mimesis) is involved.’ Wiens 1978 ‘… the process whereby the sensory systems of one animal (operator) are unable to discriminate consistently a second organism or parts thereof (mimic) from either another organism or the physical environment (the models), thereby increasing the fitness of the mimic.’ Vane‐Wright ‘Mimicry involves an organism (the mimic) which simulates signal properties of a second living organism 1980 (the model) which are perceived as signals of interest by a third living organism (the operator) such that the mimic gains in fitness as a result of the operator identifying it as an example of the model.’ M.H. Robinson ‘Mimicry involves an organism (the mimic) which simulates signal properties of another organism (the 1981 model) so that the two are confused by a third living organism and the mimic gains protection, food, a mating advantage (or whatever else we can think of that is testable) as a consequence of the confusion.’ Maran 2005 ‘Proceeding from semiotics the essence of mimicry is the presence of two living beings (object) who have different applicability (interpretant) to the receiver (interpreter) and who because of the similarity of their messages or cues (representamen) are at least partly undistinctable for the receiver.’ Grim 2013 ‘Mimicry refers to functional ‘model–mimic–selecting agent’ trinity (with varying number of species involved) when the selecting agent (i.e. signal receiver) responds similarly to mimic and model to the advantage of the mimic.’ Speed 2014 ‘In its most general form mimicry refers to phenotypes of an organism that are adaptively modified to resemble living or nonliving components of its environment.” Dalziell et al. ‘…a [signal] is mimetic if the behaviour of the receiver changes after perceiving the … resemblance 2015 between the mimic and the model, and the behavioural change confers a selective advantage on the mimic.’

­necessary for a large set of adaptive resemblances. A num- where the model was definable. In present parlance, that ber of influential definitions of mimicry were cited by Endler includes masquerade (i.e. resemblance of an organism to a (1981) and his Table 1 is reproduced here with a few addi- definite object of no interest to a predator or herbivore) tions (Table 1.1). Von Beeren et al. (2012) provided a sum- (Endler 1981), along with all classical protective, reproduc- mary of how various authors have applied terms related to tive, dispersal and social mimicry cases. In a lot of older lit- camouflage and mimicry in relation to how the operator erature, masquerade is called mimesis. (dupe, signal receiver) perceives them and I have extended it That arguments have raged for years over the precise with further examples in Table 1.2. Pasteur (1982, Fig. 1) meaning of the term mimicry indicates that there are cate- presents a neat timeline showing where various authors gories of relationships between organisms, or between drew the distinction between mimicry (homotypy in his organisms and inanimate subjects, which some people see ­terminology) and crypsis. The range is considerable – some as hard and fast examples of mimicry while others do not. workers restricted use of the term mimicry to anti‐predator Such grey areas serve to highlight what for more than a resemblances with defined models (essentially the Batesian– hundred years has been a fermenting, and sometimes Müllerian spectrum with unpalatable models), while at the ­acrimonious, debate. Probably the main reasons why there other extreme (e.g. Bates 1862, Turner 1970) all forms of has been such a long history of debate on this issue is camouflage were included under the definition as well. A ­precisely because of some people’s notions that ‘mimicry’ number of people drew the line between crypsis due to must involve deceit and that it must involve a definite model. background matching, countershading, etc. and cases Thus Müllerian mimicry, which is the convergence in

0003114056.INDD 4 7/14/2017 12:48:07 PM Introduction and classification 5

Table 1.2 Relationships proposed between various terms used in anti‐predator mimicry and camouflage in relation to the predator’s response to the prey, as employed by various authors. (Adapted from von Beeren et al. 2012 under the terms of the Creative Commons Attribution Licence CC BY 3.0.)

Predator’s reaction to potential prey

Not detected as a Detected as an Detected as an interesting discrete entity uninteresting entity entity (causing a reaction (causing no reaction) (causing no reaction) beneficial to the mimic) Reference(s)

Crypsis Masquerade Mimicry Endler 1981, 1988 Eucrypsis Mimesis Homotypy Pasteur 1982 Eucrypsis Plant‐part mimicry Mimicry M.H. Robinson 1981 Crypsis Masquerade Mimicry Ruxton et al. 2004a Cryptic resemblance Cryptic resemblance Sematic resemblance Starrett 1993 Crypsis Masquerade — Stevens & Merilaita 2009b Crypsis Crypsis Mimicry Vane‐Wright 1976, 1980 Camouflage or mimesis Camouflage or mimesis Mimicry Wickler 1968 Crypsis Masquerade Mimicry Endler 1981, 1988 Eucrypsis Mimesis Homotypy Pasteur 1982 Eucrypsis Plant‐part mimicry Mimicry M.H. Robinson 1981 Crypsis Masquerade Mimicry Ruxton et al. 2004a, Ruxton 2009 Cryptic resemblance Cryptic resemblance Sematic resemblance Starrett 1993 Crypsis Masquerade — Stevens & Merilaita 2009b

appearance between members of a guild of unpalatable understanding of the systems of interest. Edmunds consid- ­species to mutual benefit, has often been excluded because ers various sea‐slugs (nudibranch molluscs) that closely the signal displayed by members of the guild does not resemble, both in colouration and in texture, the sponges deceive a predator. and hydroids upon which they feed (see Fig. 2.2). Whether Similarly, the form of camouflage called ‘crypsis’, in which such cases are mimetic or cryptic depends crucially on organisms show a general resemblance to background whether a potential predator, say a fish, ignores the sponges properties such as colour, luminance and texture but not to or hydroids because they are not suitable food, or actively any definite model, fails to satisfy the definitions of mimicry avoids them because they pose a threat. Both situations are provided by many authors (Wickler 1968, Wiens 1978), quite possible and may be expected to occur. Thus in these whereas a camouflaging resemblance of an organism to a and perhaps many more instances, an adaptive resem- defined model, maybe a caterpillar to a twig, does. Examples blance can be both mimetic and cryptic, depending on of camouflage with a definite (definable) model are now which particular type of predator one is considering. usually referred to as ‘masquerade’ (see Chapter 3) but Edmunds (1991) concluded that although some nudi- there is an inevitable grey area – after all, masquerading branch molluscs are very brightly coloured (see Fig. 4.32), animals have almost always evolved from cryptic ones by evidence that they are warningly coloured is rather scant gradual increases in similarity to a more precisely defined (see also Guilford & Cuthill 1991 and Chapter 4, section model. At some stage it becomes apparent that something is Evidence for individual selection). Nevertheless, it would masquerading, but there are many intermediate cases. The appear that some nudibranchs may act as Müllerian mod- differences may also depend on the perceptions of a given els for young fish (Randall & Emery 1971). observer. Early discussions of mimicry were almost entirely centred on Many people have tried to make a clear distinction examples at the extreme end of the spectrum of what are now between mimicry and camouflage, for example Pasteur considered to be mimetic resemblances (Rothschild 1981). (1982), but this is not always an easy matter and may in This had the effect of emphasising differences, whereas in fact be impossible in some cases. A particularly widespread nature there is often a continuum; indeed there must be problem when considering mimicry systems, alluded to by because of the way organisms evolve. That narrow view M. Edmunds (1974a), is our frequent lack of detailed also tended to draw attention away from how the species got

0003114056.INDD 5 7/14/2017 12:48:07 PM 6 Donald L. J. Quicke

there in the first place, and, as Rothschild points out, This seems rather nit‐picking to me though I agree that “­mimicry may be fluid, and the category we assign a species some borderlines are hard to define. Vane‐Wright (1981) to may be different according to time and place”. published a rejoinder to Edmunds and noted, among other Vane‐Wright (1980) created a flurry of interest, debate and things, that his “original attempt at a general definition was even dissent as a result of various aspects of his attempt to introduced in the context of trying to define 40 theoretically define mimicry (Table 1.1) in which he distinguished between different types of mimicry and to separate them from all models that are of no interest to the predator, such as twigs, possible types of crypsis.” dead leaves, etc., and those that are. Mimics of the former M. Edmunds (1974a) defined the occurrence of Batesian would then be regarded as cryptic by his definition. More mimicry as being when “a predaceous animal, which avoids than half of an issue of Biological Journal of the Linnean Society eating one animal producing a particular signal, is deceived (volume 16, part 1) was devoted to criticism and useful com- into avoiding a second animal which produces a similar but mentary on Vane‐Wright’s definition (Cloudsley‐Thompson counterfeit signal”. He went on to conclude that animals 1981, M. Edmunds 1981, Endler 1981, M.H. Robinson that precisely mimic other types of inedible objects, such as 1981, Rothschild 1981) and, of course, Vane‐Wright’s leaves, twigs, thorns, etc. in order to avoid predation, are (1981) reply. The issue also includes a shortened version of Batesian mimics. I do not think that such broad definitions Henry Bates’ famous 1862 paper which set out the basis of are useful because there are distinctly different effects on his thought on the mimicry that now bears his name. the models; if models are living then the more palatable Cloudsley‐Thompson (1981) felt that the term mimicry Batesian mimics there are, the more the model species should be restricted to resemblances of one animal by ­suffers, but it is hard for twigs or pebbles to suffer in the another and the word ‘disguise’ used for when an animal is same sense. “like a stick, lichen, bark, faeces, a stone or some other Similarly, Endler (1981) also found difficulty with Vane‐ ­inanimate object unattractive to potential predators”, Wright’s system when it comes to determining whether though he notes that he might have been unduly influenced some resemblances are mimetic or cryptic because of in this respect by his supervisor Hugh Cott, who considered the animate/inanimate distinction. For example, resem- that all animal resemblances to plants should be referred to blances involving sticks or twigs (i.e. parts of a plant) as crypsis. Cloudsley‐Thompson quotes from a written involve a potentially animate­ model, whereas resem- response to him from Dick Vane‐Wright concerning the blances to a stone or some dung do not. Neither of these nature of mimicry and it is worth repeating here: types of model is of interest to the predator, but Vane‐ Wright used the term mimicry for when the model “is an I’m afraid I stick to my contention that it is the ‘reference frame’ of the operator that determines whether or not crypsis or organism or part of one”, which therefore excludes cases mimicry should be employed. Tiger stripes ‘are’ aggressive with stones or dung as models. crypsis, leaf‐insects are defensively cryptic. But mantids which M.H. Robinson (1981) criticised Vane‐Wright’s definition look like flowers (e.g. Idolomantis diabolicum) are aggressive primarily because it requires interpretation of the phrase mimics, as their victims are (presumably) actively seeking out “of interest” and as a remedy he couched his definition only flowers and we imagine that most flowers have evolved their in terms of confusion. While Robinson’s definition ‘flashy’ signals to attract pollinators. I don’t see how you could (Table 1.1) overcomes some difficulties, as with Vane‐ define such a general word as disguise in an operational way Wright’s (1980) definition, it precludes resemblances to different to mimicry and crypsis as I have defined them, or as non‐living entities such as pebbles, bubbles, etc. Vane‐ others have tried to do so. It might be useful as a word for Wright (1981) points out that Robinson’s definition also lumping both phenomena together. ignores both evolution by natural selection and the M. Edmunds (1981) argued that Vane‐Wright’s definition ­receiver’s perception, as well as being in itself a gross of mimicry could be seen as too broad, and presented a ­over‐­simplification of Bates’ original notion. M.H. Robinson number of examples emphasising the fine borderline (1969), on the other hand, made an important contribu- between a ‘mimic’ being of no interest to a receiver and tion to classifying different types of crypsis, and particularly being of potential interest. For example, in reference to the distinguishing eucrypsis (homochromy, countershading blue jay/stick caterpillar experiments of de Ruiter (1952), and disruptive colouration) from ‘plant part mimicry’ Edmunds notes that: ­(contra Vane‐Wright 1980) (i.e. Cott’s (1940) ‘special [the] jays could not distinguish stick‐like caterpillars from twigs ­protective resemblance’ and Turner’s (1961) ‘disguise’). until they accidentally trod on one. The question then is, was Crypsis and hiding from predators are probably the basic the signal of the caterpillar of no interest at all (crypsis), or was (plesiomorphic) defensive strategies in most major groups it of possible interest as an object to grip with the claw, in which (M. Edmunds 1990), and masquerade, mimicry and case it becomes mimicry. ­aposematism may evolve when circumstances are right

0003114056.INDD 6 7/14/2017 12:48:07 PM Introduction and classification 7

Cyrtaucheniidae Migidae anachoresis/crypsis Actinopodidae Ctenizidae ldiopidae masquerade Nemesiidae Microstigmatidae Theraphosidae Batesian mimicry Paratropididae Barychelidae aposematism Hexathelidae Dipluridae Antrodiaetidae Atypidae Filistatidae Caponiidae Tetrablemmidae Dysderidae Oonopidae Segestriidae Pholcidae Diguetidae Plectreuridae Ochyroceratidae Leptonetidae Te lemidae Drymusidae Scytodidae Sicariidae Uloboridae Deinopidae Araneidae Tetragnathidae Nephilidae Mysmenidae Anapidae Linyphiidae Cyatholipidae Synotaxidae Theridiidae Nesticidae Agelenidae Desidae Amaurobiidae Zoropsidae Ctenidae Miturgidae Pisauridae Tr echaleidae Lycosidae Psechridae Senoculidae Oxyopidae Prodidomidae Gnaphosidae Lamponidae Cithaeronidae Ammoxenidae Gallieniellidae Tr ochanteriidae Salticidae Thomisidae Corinnidae Selenopidae Anyphaenidae Zoridae Sparassidae Philodromidae Clubionidae Homalonychidae Zodariidae Hahniidae Cybaiidae Dictynidae Phyxelididae Titanoecidae Nicodamidae Eresidae Hersiliidae Oecobiidae Palpimanidae Mimetidae Liphistiidae

Fig. 1.1 Modes of passive defence mapped onto independent phylogeny of families of spiders (pruned to show only those families for which hypotheses about modes of anti‐predator defence could be assessed). (Source: Adapted from Pekár 2014. Reproduced with permission from John Wiley & Sons.)

0003114056.INDD 7 7/14/2017 12:48:09 PM 8 Donald L. J. Quicke

(i.e. probably when the background is not uniform but The concept of ‘adaptive resemblance’ ­complex and contains many discrete or nearly discrete prey‐ sized items). Pekár (2014) has examined defensive strate- The term ‘adaptive resemblance’ (AR) was introduced by gies across the whole of the Arachnida (spiders) and shown Starrett (1993) as a “broad inclusive term” for a wide range that masquerade and mimicry have evolved in a restricted of mimetic and cryptic phenomena and was defined as: “any number of families in a very non‐random way (i.e. the strat- resemblance that has evolved or is maintained as a result egies are phylogenetically clustered) (Fig. 1.1). The most of selection for the resemblance.” It specifically excludes basal extant spiders, the liphistiids3 and mygalomorphs, are “incidental resemblance or convergence, which is due to generally large bodied, ground dwelling and nocturnal or common adaptive responses to functional requirements.” spend most of their time concealed in silk‐lined holes AR incorporates everything normally regarded as mimicry (­anachoresis), so it is not really surprising that they have as well as most things that are only occasionally regarded as not evolved bright colours (with rare exceptions in the mimicry. Starrett uses the term selective agent (SA) for what Theraphosidae, such as the red‐headed mouse spider, others have variously called the detectee, dupe, receiver, sig- Missulena occatoria) or mimetic resemblances to particular nal receiver or operator. AR includes both crypsis and mas- objects – the evolutionary advantage was not there. Orb‐ querade (both crypsis in Starrett’s definition) and makes no web weaving spiders such as aranaeids, tetragnathids and distinction about whether the model is alive, dead or inani- nephilids, however, are often fully visible in their webs in the mate. It [AR] encompasses Müllerian mimicry even though daytime, and among these bright patterns are much more some authors have strongly objected on the grounds that no prevalent, and may, in fact, be attractive to certain prey (see deception is involved, but it is clearly an adaptive resem- Chapter 10, section Alluring mimicries). A similar phyloge- blance. It also includes various topics rejected by Pasteur netic pattern of ancestral crypsis has recently been shown (1982), such as vocal mimicry (some bird song or alarm call for North American darkling beetles (Tenebrionidae: mimicry) or behavioural mimicry by primates. Pasteur’s rea- Asidini) by A.D. Smith et al. (2015), some of which are soning was that vocal mimicry was a ‘conscious imitation’ Batesian mimics of species of the related and well‐protected and not something that had been a direct result of natural genus Eleodes (see Chaper 6, section Experimental tests of selection, though of course the cognitive ability or propen- mimetic advantage). sity to do so was selected. It is far from clear which instances Schaefer & Ruxton (2009) coined the term ‘exploita- of such mimicries are consciously done. When a female tion of perceptual biases’ (EPB) to differentiate what they baboon goes off with a subordinate male for sex and they regarded as true mimicry from simple exploitation of hide from the dominant male to avoid his aggression – that is the sensory biases and loopholes of another organism. probably conscious, but when a bird imitates alarm calls of In their definition, the term mimicry only applies to conspecifics to deter a competing male, is that consciously cases where a receiver really misidentifies the mimic as a thought out or a behavioural pattern that has been directly specific model. selected to be part of the species’ repertoire? Starrett includes Here I have decided first to discuss camouflage (crypsis most if not all such cases under AR but also asks whether and masquerade), as despite differences of opinion about along the gradient of increasingly complex behavioural what constitutes mimicry and what does not, these evolu- complexity (cognitive ability) there is a line “beyond which tionary strategies differ from what is generally nowadays release from unconscious stereotyped behaviour allows dis- called mimicry, in that the mistakes made by operator cretionary situational and consciously deceptive imitation (dupe, receiver) have no effect on the model (Vane‐Wright that might not be subject to natural selection”. Surely there 1980, 1981, Endler 1981). Thereafter, I have largely is, but beyond that we enter the realm of psychology and ­followed a functional classification similar to that of that goes largely beyond the scope of this book and AR. S.B. Malcolm (1990) (Table 1.3), though with variations Nevertheless, I will discuss various human behaviours to on the terminology and a greater number of divisions. highlight the similarities between what evolution has done Because of the many biological roles that mimicry affects and how people deceive, sometimes unconsciously. there has long been a need to pigeon‐hole examples to This useful definition also makes for clearer thinking. make it easier to compare and gain better understanding Starrett pays particular attention to various other of different systems. “troublesome”­ cases. For example, the widespread occur- rence of bold black and white colouration in marine carni- vores such as penguins, killer whales and various dolphins 3. These South‐East Asian spiders, which retain a partially seg- (see Fig. 10.3a–d) is clearly not due to the different species mented abdomen and lack venom glands, are the sister group to evolving to resemble one another. At least in Spheniscus pen- all other living spiders. guins the colouration may be selected for because its

0003114056.INDD 8 7/14/2017 12:48:09 PM Introduction and classification 9

Table 1.3 Simple classification of mimicry based on kind of interaction, trophic level and selective agent as summarised by S.B. Malcolm. (Adapted from Malcolm 1990 with permission from Elsevier.)

Mimicry Selective agent category Interaction (operator) Examples

Interspecific interactions Defensive Prey–predator Predator Müllerian, Batesian, ‘masquerade’, ‘predator mimicry’ Foraging mimicry Prey–predator Prey Aggressive and ‘masquerade’ (lampyrid beetles, bolas spiders, cleaner wrasse) Flower–pollinator Pollinator Flower mimicry (orchid mimicry of bees) Parasitic mimicry Host–parasite Host or vector Parasites, dispersion and mating (cuckoos, trematodes, fruit dispersal, orchid mimicry of bees)

Intraspecific non‐trophic interactions Sexual mimicry Male–female Same, or opposite, sex Mating lures (‘sneaky’ mating, egg or prey dummies) Social mimicry Any sex, juveniles Same, or opposite, sex Signals of hierarchy, deceptive alarm calls

conspicuousness causes schools of prey fish to ‘depolarise’, In terms of creating frameworks for classifying different making some prey individuals easier to see and capture (R.P. types of mimicry, three people in particularly have made Wilson et al. 1987). Why exactly these conspicuous mark- great advances – Wolfgang Wickler, Richard [Dick] Vane‐ ings might have this effect is unclear; maybe they ‘mimic’ Wright and Georges Pasteur. I will discuss their contribu- signals from other potential predators that might best be tions in some detail below. thwarted by a confused response. After all, over millions of years of evolutionary time, if the bold black and white pat- tern of piscivores leads to greater overall mortality one might Wickler’s system have expected schooling fish not to respond to it any longer. Wickler (1965) proposed a formal notation for mimicry as a communication network in which three players interact THE CLASSIFICATION OF MIMICRY and on which the costs or benefits of the signals to the par- SYSTEMS ticipants could be represented. Two of the parties he termed signal senders (S, called S1 and S2) and one the signal While the concept of mimicry is generally understood by receiver (R), which responds to the signals emitted by the everyone, mimetic relationships have evolved in relation to a senders. By convention, the model is represented by S1 and wide range of biological processes, which has made attempts the mimic by S2. Endler (1981) preferred to use the nota- to classify the different types complicated. Different workers tion P, S and R, referring to primary and secondary signal have also highlighted different factors as being important, generators and a receiver, but here I use Wickler’s system. and there is considerable overlap. While some authors are Wickler then provided the following formal definition of clearly willing to accept the idea that some resemblances mimicry as a system: have elements of more than one type of mimicry, for exam- ple many cases of crypsis, and sometimes masquerade, serve 1. comprising two signal senders (S1 + S2) which have to hide the aggressor simultaneously from potential preda- one or more receiver(s) in common, tors and from potential prey (see Figs 2.29 and 10.7d,e). 2. in which the receiver(s) respond similarly to the signals There may be other ‘grey areas’. Perhaps some authors have of the two senders, concentrated heavily on pigeon‐holing particular hard‐to‐ 3. in which it is advantageous for the receiver to respond place examples. While this can serve as a test of classifica- to one signal sender in a given way but disadvanta- tory systems, it is hard to imagine that any one simple system geous for it to respond to the other signal sender in the can accommodate all the results of evolution. same way.

0003114056.INDD 9 7/14/2017 12:48:09 PM 10 Donald L. J. Quicke

In terms of Wickler’s notation this is R:S1 negative R:S1 positive SR12S S1:S2 synergic from which the mimic in the system can be defined as the S1 + S2 S1 + S2 signal sender that elicits a response in the receiver that – – + – would be disadvantageous (negative) for the receiver, while R S1:S2 antergic R the other sender is the model (S1). In other words, the mimic S1 S2 S1 S2 is the party that emits counterfeit signals that elicit mala- – – dapted responses in the receiver. As noted by Wickler, the – – + – R:S2 negative R R mimic always has a selective advantage. This can be repre- I VIII VII II sented in terms of his notation as III S1 S2 S1 S2 IV SR12S – – – + + + R R Note that the mimic (S2) must always benefit because VI V ­otherwise the mimicry would not have been selected for in the first place. The receiver’s response, of course, does not S1 + S2 S1 + S2 have to be negative to the other sender and in the broader – + + + R:S2 positive context of adaptive resemblance it is possible for all the R R interactions to be positive, as for example in classical ‘Müllerian mimicry’, which would be represented as Fig. 1.2 The eight interactions that can occur in mimetic relationships with the signal receiver (R or dupe) and two signal SR12S transmitters (S1 and S2) showing how the signals and responses of each member of the triad are either beneficial (positive) or harmful (negative) to the others. The outer relationships are synergic, meaning that the existence of the mimic (S2) is Vane‐Wright’s system beneficial to the model (S1), whereas in the inner set of four the mimic has a detrimental effect on the model. (Source: Adapted Vane‐Wright (1976) provided a broader basis for the classi- from Vane‐Wright 1976 with permission from John Wiley & Sons.) fication of mimetic relationships, extending Wickler’s anal- ysis of interactions between the three components of the system, S1, S2 and R. Vane‐Wright’s (1976, 1981) defini- The three possible polar systems can be represented by tion differs from that of Wickler (1968) in that it does away S1 + R/S2, S2 + R/S1 and S1 + S2/R, as illustrated in Fig. 1.3. with the need for deception – thus it comfortably incorpo- Vane‐Wright’s classification was not intended to be a rates Müllerian mimicry in which the receiver is not complete treatment of adaptive resemblances as it only deceived, and single or mixed schools of fish in which one deals with cases in which the model is definable and ani- individual gains protection because of predator’s difficulty mate; his choice of criteria in fact prevents the inclusion in in visually separating it from all the other school members his classification of crypsis and masquerade. Nor does it (arithmetic mimicry). embrace ill‐defined features such as deflective markings and The first part of Vane‐Wright’s analysis considered dazzle, as these cannot be classified as being either synergic whether signals from the mimic (S2) are advantageous or or antergic and do not have a clearly defined model. disadvantageous to the receiver (R), and whether signals Out of the 40 conceivable types of mimetic relationship from the mimic (S2) are disadvantageous to the model (S1). allowed for in his classification (Table 1.4), Vane‐Wright Systems in which the mimic’s signal are not harmful to the was able to find non‐human examples for 21 with some model are referred to as ‘synergic’ and in the opposite case, certainty, a couple rather more speculatively, and human ‘antergic’. He depicted the full range of possibilities as examples (e.g. military decoys and spies) for a further shown in Fig. 1.2. three. Vane‐Wright notes that the bipolar S1/S2 + R The second part of Vane‐Wright’s system considered the ­system, in which the mimic and dupe belong to one species actual embodiment(s) of these three components (i.e. the while the model belongs to another, seemed to be com- species they belong to). While for many familiar cases of pletely empty, though he did suggest that some courtship mimicry, the model, mimic and receiver are each different activities of spiders and empidid flies which incorporate species, termed ‘disjunct’ systems, there are cases in which simulation of prey organisms could belong to this cate- all three are members of the same species (conjunct systems),­ gory. To these possibilities might be added the food lures or any two may belong to one species (bipolar systems). used by certain male fish that resemble invertebrates and

0003114056.INDD 10 7/14/2017 12:48:10 PM Introduction and classification 11

disjunct system neutral background elements to potential predators of Siphamia or objects to be avoided because of their spines? S1 S2 Probably both to different predators. Vane‐Wright thus accepts Miriam Rothschild’s (1975) R dual signals concept to explain how the nature of a signal (a) from a deceptive organism can change with distance (see Chapter 2, section Dual signals). He also notes, in response to three bipolar systems a criticism from Rothschild (1981) concerning how cuck- oos change during development from rejecting many typi- S1 S2 S1 S2 S1 S2 cally unpalatable insect food items to readily accepting them as adults, that “the terms model and mimic describe R R R functional entities in a schematised biological interaction: (b) (c) (d) they are not fixed properties of individuals” (Vane‐Wright 1981). Some resemblances might therefore be able to repre- conjunct system sent more than one of Vane‐Wright’s 40 categories simulta- neously, depending on the receiver. A given pair of butterfly S1 S2 species may be aposematic and unpalatable (i.e. Müllerian mimicry) to one predator, but one of them may be quite R ­palatable to a different predator (Batesian mimicry). (e)

Fig. 1.3 The different relationships at species level between the signal receiver (R or dupe) and two signal transmitters (S1 and S2) Georges Pasteur (1930–2015) with one disjunct tripolar system, three different disjunct bipolar systems and one conjunct system in which all three aspects of the Pasteur (1982) developed a largely three‐dimensional clas- mimetic relationship are confined to a single species. (Source: sification of mimicry, combining the concepts of functional- Adapted from Vane‐Wright 1976 with permission from John ity with the disjunct, bipolar and conjunct systems of Wiley & Sons.) Wickler (1965) and whether the dupe is indifferent to the model or whether it is of interest to the dupe, and in the lat- attract females (Kolm et al. 2012; see Chapter 11, section ter case, whether it is agreeable or forbidding to the dupe. Food dummies and sex), as well as human Native North Seven functional classes were recognised (Table 1.5), which American use of bird call mimicry to deceive European together with the five possible species compositions of a ­settlers and troops about their intent. mimicry system and three main classes of the dupe’s inter- Following Vane‐Wright’s explicit and objective classifica- est in the model gives a maximum of 105 different combi- tion of mimicry systems, several authors have pointed out nations, though Pasteur only found examples for a far that it will remain difficult to pigeon‐hole many examples smaller number. Pasteur was also responsible for naming because we usually do not know exactly what an operator many classes of mimicry after their discoverers or co‐discov- (e.g. a predator’s) perceptions of a given situation are, nor is erers, without due regard to the nature of the interactions there any reason why all operators should share the same involved, and this has tended to lead to confusion. perceptions or why any one operator should have the same Unfortunately, many of Pasteur’s categories still combine perception all of the time. Some interesting examples were examples of mimicries with very different functions. For cited by M. Edmunds (1981). For example, many sea‐slugs example, as discussed by Quicke et al. (1992), the term (Nudibranchia) resemble the food organisms on which they Kirbyan mimicry was employed to cover aggressive/repro- sit and feed not only in general colouration and texture, but ductive S1 + R/S2 systems involved in brood parasitism. sometimes also in more specific detail. Further, in some Within this seemingly tightly defined system, Pasteur instances, the sea‐slug’s prey/substrate may be unpalatable, included the egg mimicry of cuckoos and the mimicry of as with many sea anemones and sponges. Whether, then, Hymenoptera by robberflies (Diptera: Asilidae) whose lar- such examples should be classed as crypsis or as Batesian vae are predators of their model’s larvae (see Chapter 10, mimicry will depend upon whether the model is neutral to a section Cuckoldry, inquilines and brood parasitism). He also potential predator, such as a fish, or avoided by it. In a simi- suggested that this category might include the vocal mim- lar vein M. Edmunds cites Fricke’s (1970) description of the icry of host young by cuckoo chicks which resemble those fish Siphamia sp., which aggregate so as to resemble sea of their hosts (M.G. Anderson et al. 2009; see Chapter 10, urchins (Astropyga spp.). Should sea urchins be regarded as section Gentes and ‘cuckoo’ eggs). It is clear that several very

0003114056.INDD 11 7/14/2017 12:48:11 PM 0003114056.INDD 12

Table 1.4 Classification of ‘mimicry’ systems according to Vane‐Wright (1976) with a few extra examples added (note S1 = model, S2 = mimic, R = dupe). (Vane‐Wright 1976. Reproduced with permission from John Wiley & Sons.)

Synergic (i.e. mimicry is good or neutral for the model) Antergic (i.e. mimicry is bad for the model)

I II III IV V VI VII VIII Model and mimic Warning Aggressive Defensive Inviting Inviting Defensive Aggressive Warning

A (Disjunct, i.e. Müllerian, Angling with Arithmetic Useful weeds Batesian, Peckhamian, hawk ? 3 separate quasi‐Müllerian lures quasi‐Batesian, mimicry by species) hawk mimicry by cuckoo cuckoos B S1 + R/S2 Trophobionts ? Vine ‘eggs’, predator Cuckoos, monkey mimicry by prey, calls of big cats, monkey call mimicry pollinator sexual by cats, antigenic deception mimicry by parasites C S1 + S2/R Aposematism Alarm calls of Automimicry, Herding, Competition antshrike thanatosis schooling D S2 + R/S1 Food lure for courtship E Conjunct Military uniform Military decoys Bluff Egg dummies Sexual Bluff, appetitive use of Spies in fish competition alarm calls 7/14/2017 12:48:11 PM Introduction and classification 13

mimicry setting, Pasteur cites the case of some Oncidium Table 1.5 Pasteur’s (1982) seven functional classes of orchids whose flowers are so loosely attached that they mimicry. (Data from Pasteur 1982.) dance around in the slightest breeze and thus attract highly territorial male Centris bees, which attack almost anything Class Biological function flying by (Dodson & Frymire 1961). 1 Aggressive 2 Aggressive/reproductive 3 Reproductive Other approaches 4 Reproductive/mutualistic 5 Mutualistic Endler 6 Commensalist 7 Protective Endler (1981) considered that the result of a predator ­making a mistake should be taken into account as a basic characteristic of mimetic and cryptic systems (Table 1.6). Two criteria are considered: first, whether the mistake ­confuses the signal sender (mimic) with the background or different types of mimetic relationship are involved here. with a specific thing, and second, whether the mistake Cuckoo eggs usually have to be more or less accurate mim- affects the population dynamics of the model (if relevant). ics of host eggs in order to avoid rejection (Brooke & Davies Endler superimposed on his classification table the extents 1988, Lotem et al. 2009). of the definitions of mimicry applied by various other work- In another, perhaps even more extreme example, ers, and shows clearly that different authors have included Pasteur’s term Wicklerian–Barlowian mimicry was coined and excluded large categories from consideration. for conjunct reproductive mimicries. This includes cases where females mimic males (androchromatism; see Zabka & Tembrock Chapter 11, section Mate guarding through distracting other males) as well as the similarity between male and female Zabka & Tembrock (1986) proposed an alternative approach flowers in monoecious plants (see Chapter 12, section to the classification of mimicry systems based on “the aims” Flower automimicry – intraspecific food deception (Bakerian of the mimics and the behavioural responses involved. They mimicry)) and the mimicry of the erect male penis by the separate camouflage (crypsis in their terminology) from female’s clitoris in spotted hyaenas (see Chapter 13, section other mimicries on the basis that the model is irrelevant to The case of the spotted hyaena). Thus while Pasteur’s system the signal receiver, whereas it is relevant in true mimicries may allow most cases of mimicry to be classified, his choice and in those cases may cause an adverse or an appetitive of functional categories seems to be too general to lead to (i.e. feeding) reaction (Fig. 1.4, Table 1.7). To put it another groupings with much internal consistency, and certainly way, “mimicry is a phenomenon of the relevant environ- should not be followed without due caution. Similarly, the ment of the signal receiver”. Irrelevant models may be term Wasmannian mimicry has been employed as a either abiotic or biotic but relevant ones are almost certainly ­general term for mimicries of ants, irrespective of whether always going to be biotic. Zabka & Tembrock’s system can be it is Batesian or aggressive in nature. seen as a refinement of the broadly biological classification One distinction that Pasteur uses is between cases of of Wickler (1968) and Table 1.8 shows how they classified ­mimicry when there is a single model (or small tightly various of Wickler’s (1968) examples. It will be apparent defined set of models) whose species identity is obvious that Zabka & Tembrock’s classification of reproductive mim- (for example, sex pheromone mimicry by bolas spiders or icry does not make allowance for bipartite situations in sexually deceptive orchids or cases of Batesian mimicry in which the model and mimic belong to the same species (i.e. which a particular species is the unpalatable one), referred S1 + S2/R). However, such systems do exist, for example, in to as ‘concrete homotypies’, in contrast to ‘abstract homo- the eastern Mediterranean cucurbit Ecballium elaterium, the typy’ in which the model is only definable in a vague way, relatively rare female flowers, which do not produce nectar, such as a snake or a pair of vertebrate eyes. Semi‐abstract effectively mimic the commoner, nectiferous male flowers examples might include the vermiform lures of anglerfish on the same plant (Dukas 1987) (see Chapter 12, section and alligator snapping turtle. Even more extreme is when Flower automimicry – intraspecific food deception (Bakerian no particular type of model is represented, as in the case of mimicry)). However, their lumping under the category some deimatic startle displays, but the display is interpreted ‘reproductive mimicry’ of all systems in which some aspect as a possible source of danger. In a sexual (reproductive) of sexual behaviour of the dupe is mimicked seems to me

0003114056.INDD 13 7/14/2017 12:48:11 PM 14 Donald L. J. Quicke

Table 1.6 Effects of predator (P) making an error in discriminating model from mimic (S) and how different authors relate this to the distinction between mimicry and crypsis. (Adapted from Endler 1981 with permission from John Wiley & Sons.)

Mistake has no effect Mistake affects on population population dynamics or dynamics or Mistake affects evolution of more than evolution of P population dynamics 1 P species (model = inanimate of single P species (model = more than or background) (model = 1 species) 1 species)

Mistake depends on the Crypsis Polymorphism Convergence relationship between S Crypsis (=Eucrypsis) Polymorphism and Convergent evolution and background Aggressive mimicry (a) apostacy Mistake depends on Masquerade Batesian mimicry Müllerism similarity to specific Plant‐part mimicry Batesian mimicry Müllerian mimicry object or species, not Dung and stone Dispersal mimicry Mertensian mimicry background mimicry Reproductive mimicry Aggressive mimicry (b) Group mimicry Aggressive mimicry (c)

versus & criterion of Cott (1940) and Vane‐Wright (1980)

& versus criterion of Wickler (1968)

unhelpful (Table 1.9), and I prefer to use the term ‘sexual an ­argument against evolution. For example, the butterfly mimicry’ specifically for cases that have evolved to increase expert and geneticist Reginald Punnett (1915) was fertilisations/parenthood directly (see Chapter 11). ­fervently against the idea that gradualism could lead to the high‐fidelity mimicry he saw between many unrelated ­species of butterflies during his visit to Sri Lanka (formerly Maran Ceylon). Maran (2007, 2010) applies semiotic theory and notation, Mimicry has been particularly important when it comes essentially to Vane‐Wright’s (1981) classificatory system, to the genetic variation upon which evolution acts; does that allows for easy depiction of how different receivers (or evolution proceed mostly in small incremental steps in a perhaps the same receiver at different times or distances, or neo‐Darwinian framework, or do large leaps occur due to different parts of the mimic, etc.) perceives a mimetic system mutations in genes that have a large effect? The micro‐­ (Fig. 1.5). In Fig. 1.5a, for example, the mimic is an arbi- evolutionary paradigm has been largely accepted in most trary anglerfish (as in Fig. 10.7), but it uses two separate fields, yet it is clear that genes of large effect are involved in models for different purposes: the bulk of the fish’s body is many cases of aposematism and mimicry (though with cryptic, i.e. resembling the background encrusting organ- other modifier genes having lesser effect). As Orr & Coyne isms, while the lure is mimicking a worm‐like object. The (1992) stated, “[it] might be objected … that mimicry is not same receiver, the potential prey, responds to both. a ‘typical’ adaptation” because it so often seems to involve large phenotypic jumps. This question is still largely open to debate, despite modern genomics approaches shedding a lot MIMICRY AS DEMONSTRATION of new light on the subject, but the general consensus seems OF EVOLUTION to be that a two‐step process, with an initial mutation ­causing a large shift in phenotype followed by gradual fine‐ Mimicry has long been seen as important for understand- tuning, is probably involved in many cases, although a ing the process of natural selection, and indeed Batesian gradual shift is possible under certain circumstances and Müllerian mimicry are probably the best studied exam- (Lindström et al. 1999a, see Chapter 4). ples when it comes to natural populations (Orr & Coyne Nadeau & Jiggins (2010) suggest that the genetic 1992). Indeed the degree of perfection in many mimicry ­variation observed in natural populations that is responsi- systems as we perceive them led it, for a while, to be used as ble for many quantitative traits does not represent the

0003114056.INDD 14 7/14/2017 12:48:12 PM Classification of mimicry based on aims of mimics Signal receiver

No reaction

Is the model Appetitive relevant to reaction, e.g. eat the mimic? it, mate with it

Is the model good for the mimic?

Averse reaction, e.g. avoid it

Model

Abiotic models

Biotic models

Interference by resemblance Mimic

Crypsis

Other types of mimicry

Fig. 1.4 Diagrammatic representation showing how the presence of a mimic ‘interferes’ with the flow of information between the model and signal receiver and showing how the information can be interpreted by the receiver (dupe) as either relevant or irrelevant. (Source: Adapted from Zabka & Tembrock 1986.)

(a) (b) MO/R1 MO1 R

MI MO2 R2 MI Fig. 1.5 Depictions of two mimetic examples using semiotic notation. (a) This describes the relationship of an anglerfish (such as those in Fig. 10.7) which is the mimic (MI) but has two links to the receiver (R1 and R2), one via its worm‐like lure with an unspecified worm as model 1 (MO1) and the other through its cryptic resemblance to the background (MO2); (b) a depiction of relationships of a cuckoo bumblebee, Bombus (‘Psithyrus’) sp. (MI), that is a brood parasite and mimic of a normal bumblebee, Bombus sp. (MO) which is also (possibly: see Chapter 10: section Cuckoo bees and cuckoo wasps) one of the dupes of the resemblance (R1), the other receiver (R2) being a potential predator of both such as a bird. (Source: Adapted from Maran 2010. Reproduced with permission from Springer.)

0003114056.INDD 15 7/14/2017 12:48:12 PM 16 Donald L. J. Quicke

Table 1.7 Classification of mimetic and cryptic relationships involved in predator–prey interactions. (Adapted from Zabka & Tembrock 1986 with permission from Elsevier.)

Imitation of signals of the:

State of the mimic relevant environment irrelevant environment

Aversion (Protection) PROTECTIVE MIMICRY PROTECTIVE CRYPSIS Batesian and Müllerian mimicry; eyespots; Crypsis of prey; ‘fly mimicry’ of butterfly egg mimicry in plants; jamming beetles; thanatosis by prey; stone mimicry in plants (e.g. Lithops) Appetence (Feeding) AGGRESSIVE MIMICRY AGGRESSIVE CRYPSIS Angling; cleaner‐fish mimicry; perhaps some Crypsis of predators; death‐feigning by insectivorous plants e.g. Nepenthes (see predators; ‘vulture mimicry’ by Chapter 10, section Does aggressive mimicry zone‐tailed hawk occur in plants?)

Table 1.8 Zabka & Tembrock’s (1986) interpretation of the classification of mimicry systems suggested by Wickler (1968). (Adapted from Zabka & Tembrock 1986 with ­permission from Elsevier.)

Type of system Examples

I. CRYPSIS (mimesis & camouflage) II. MIMICRY 1.1. Batesian mimicry Wasp mimicry by hoverflies 1.2. Emsleyan mimicry Coral snake mimicry 2. Aggressive mimicry 2.1. Disjunct, i.e. with three or more participating species: Anglerfish; cleanerfish mimicry; carrion flowers; sporocyst of Leucochloridium macrostomum 2.2. Bipartite 2.2 a) Conspecific Non‐stinging male wasps mimicking stinging females 2.2 b) Conspecific Males and females of Corynopoma riisei 2.2 c) Conspecific Cuckoos; Ophrys orchids; female Photuris fireflies 2.3. Conjunct Egg mimicry by mouth‐breeding fish

Table 1.9 Overview of ‘reproductive mimicry’ as interpreted by Zabka & Tembrock (1986) but which contains such diverse ­functions as increasing fertilisations, obtaining food, and obtaining dispersal. (Adapted from Zabka & Tembrock 1986 with permission from Elsevier.)

Imitations of:

Signal‐receiver non‐communicative releaser communicative releasers

Interspecific receiver DISJUNCT, e.g. carrion flowers, sporocysts S1 + R/S2, cuckoos, brood parasitism in of Leucochloridium, male Photuris fireflies insects, orchid pollination by sexual deception Intraspecific receiver S1/S2 + R, e.g. prey mimicry by Corynopoma CONJUNCT, e.g. egg‐dummies in mouth‐ riisei male fish breeding cichlids, female mimicry by male scorpionflies to obtain food

0003114056.INDD 16 7/14/2017 12:48:12 PM Introduction and classification 17

spectrum of variation that leads to major adaptive traits, little to the improved insecticide resistance. The difference and that the latter includes mutations that are rare and is one of scale – the populations under selection in the lab perhaps not seen under normal circumstances in, say, lab- are likely to be a few thousands or tens of thousands of oratory selection experiments. For example, almost all insects, whereas in the field, blanket pesticide spraying is examples of resistance to insecticides observed in the field aimed at killing many millions of individuals and the selec- (such as mosquitoes to DDT) involve a single mutation that tion is much stronger. With each spray application the pest confers virtually complete­ protection. Artificial selection controllers are trying to kill all the insects, whereas in experiments in the laboratory almost inevitably lead to the lab the aim is to leave a few alive after each round polygenic changes, i.e. several genes, each contributing a of selection, from which to breed the next generation.

0003114056.INDD 17 7/14/2017 12:48:12 PM 0003114056.INDD 18 7/14/2017 12:48:12 PM