The orchid mantis exhibits high ontogenetic colouration variety and intersexual life history differences

Masquerade, the resemblance of animals to inedible or inanimate objects, exists universally throughout the animal kingdom, especially in arthropods. However, masquerade has received little attention from biologists and is often misinterpreted as mimicry by the public and even by scientists, as a consequence of the lack of systematic biological information for masqueraders. Therefore, using the orchid mantis Hymenopus coronatus (Insecta: Mantodea), a classic masquerader, as the study species, we asked: (1) what is the population abundance and life cycle? (2) whether it closely coexists with specific plant or not? (3) how its colour morph changes across life stages? and (4) whether and how the key environmental factors affect its life cycle? Results suggested that the orchid mantis (1) had an extremely low wild population abundance; (2) did not coexist with specific plant; (3) exhibited colour morph diversity; (4) might match its reproductive and nymph developmental period with local seasonal fluctuations of temperature and precipitation. We then compared its life history differences between the two sexes. The results indicated that the two sexes can be remarkably different in development duration, growth rate, lifespan and body weight. This study is the first systematic investigation of the orchid mantis; the results provide useful natural history data for ecologists and evolutionary biologists to understand the adaptive strategies of elusive appearance of masqueraders.


Introduction
Masquerade is a form of camouflage in which the masqueraders resemble inedible or inanimate objects to render themselves sensorily detectable but cognitively misclassified by predators or/and prey, thereby increasing their survivorship by avoiding predation or/and gaining success to hunting (Endler 1981;Skelhorn et al 2010a;Skelhorn 2015). This strategy evolves when predators and prey perceive the similarity between a masquerader and its model or models, consequently increasing the fitness of the masquerader (Skelhorn et al 2010b). The masquerader is often visually comparable to the model in both colouration and morphology. Predator deception (protective masquerade) and prey deception (aggressive masquerade) have been documented as the most common functions of masquerade (Skelhorn 2015). Whereas herbivores usually favour predator deception, top predators often favour prey deception (Skelhorn et al 2010a). However, small carnivores (such as spiders and mantids) are often required to evolve both protective and aggressive masquerade to be successful masqueraders, since they have to manage to survive under evolutionary pressures from both predator avoidance and prey attraction. The bird-dropping-resembling crab spider Phrynarachne ceylonica, which conceals itself visually to avoid predators such as birds, but chemically attracts prey items such as flies (Yu et al 2021), is a good example. Most masqueraders, such as the ghost mantis and the dead leaf butterfly (Skelhorn 2015), hide themselves against the background. However, a few species are highly contrasted against the background in order to lure prey (Cuthill 2019), we define this type of masquerade as conspicuous masquerade. Theoretically, conspicuous masquerade would be favoured by small carnivores when both their predators and their prey rely heavily on visual signals. Furthermore, conspicuous masqueraders often present ontogenetic colouration and morphological polymorphisms among different life stages and between the sexes, which are hypothesised to expand the masqueraders' distribution and hence enhance their adaptation (Higginson and Ruxton 2010;Caro et al 2016;Yu et al 2022). Although the phenomenon of conspicuous masquerade has been noted for decades, no species has been systematically studied for its field abundance, colour morph diversity and flexibility, microhabitats and key environmental factors that may affect its life cycle, life history and intersexual variations. The lack of this information restricts our understanding of the evolution and adaptive significance of conspicuous masquerade. To fill this research gap, we selected the orchid mantis, Hymenopus coronatus (Insecta: Mantodea), as the first species for systematic study.
H. coronatus is the only masquerader that resembles an entire blooming flower, which makes it an excellent model of conspicuous masquerade. Its flower-like appearance may functionally integrate pollinator attraction and predator avoidance (O'Hanlon 2016) and is the basis for its vernacular name of orchid mantis for over 200 years. However, all descriptions and interpretations of the orchid mantis were based on individual observations rather than empirical studies until 2013. James and his colleagues experimentally quantified the colouration and morphology of the orchid mantis (O'Hanlon et al 2013) and then performed spectral and morphological comparisons with the potential model plant species . They proposed a contrary opinion to the traditional interpretation of mimicry, as they showed that the orchid mantis did not mimic orchids or any specific flower. Hence, the orchid mantis was interpreted as a generalist rather than a specialist mimic . However, this interpretation was not confirmed due to lack of field investigation, such as population abundance, life cycle, how environmental factors affect its life cycle, its microhabitats and colour pattern changes across life stages. Behavioural experiments were then performed to examine the pollinator-deception function of 1 3 the orchid mantis, and they concluded that the orchid mantis could attract prey (insect pollinators) by its body surface colour, whereas the existence or symmetry of the femoral lobes did not attract pollinators (O'Hanlon 2014). In addition, Mizuno et al (2014) found that the orchid mantis could lure honeybees by emitting the chemical ingredients of honeybee communication pheromones, and hence they concluded that H. coronatus employed a chemical aggressive mimicry strategy. Although this evidence enhanced the hypothesis that the function of the resemblance of H. coronatus to a flower was to deceive pollinators, these results were based on only a few individuals or on three-dimensional printed models.
A systematic survey of H. coronatus will be vital for ecologists and evolutionary biologists to interpret how it survives and why its flower-like appearance has evolved (O'Hanlon 2016). Hence, we conducted this survey by integration of field investigations and laboratory rearing records of the orchid mantis. Specifically, we tried to answer: (1) What are the population abundance and life-cycle patterns in the field? (2) How is the life cycle affected by key environmental factors (temperature and precipitation)? (3) Is the orchid mantis closely associated with a specific microhabitat? (4) What is the life history of H. coronatus and what are the intersexual differences in life history? (5) What are the diversity and flexibility of the colour morphs, and what are the variations of the colour morphs among and within life stages? Answers to these questions would provide the first systematic biological information on this classic conspicuous masquerader, which might help ecologists and evolutionary biologists to interpret the evolution and adaptation of its special appearance.

Field investigation
The photographic records for 11 years (2011-2021) of field observations were collected in and around Xishuangbanna Tropical Botanical Garden (XTBG, 21° 55′ N, 101° 16′ E, 1125 hm 2 in area), Yunnan, China. XTBG is a tourist destination with dozens of tourist guides and hundreds of staff, who frequently explore the garden and always take photos and spread the news among colleagues when they see an orchid mantis as it is a famous star creature but rare. We collected photographs mainly from the tour guides and staff, and six were from the authors of this manuscript. To ensure accurate information, we selected photographs with necessary information, including recording dates, microhabitats, and specific locations. From these photographs, we first recorded the dates of observation, life stages (ootheca, first instar, flower-like juveniles and adults), colour morphs and the plants they remained on for further analysis of their life cycles, colour morph diversity and microhabitats. Second, to analyse the correlations between key environmental factors and the orchid mantis life cycle, we obtained eight-year (2010-2018) average daily temperatures (minimum, mean, maximum) and 10-year (2010-2020) daily rainfall records from Xishuangbanna Station for Tropical Rain Forest Ecosystem Studies, which locates inside XTBG.

Indoor rearing
To survey the indoor life history and intersexual variations in development and lifespan of the orchid mantis, we monitored nine oothecae (laid by four females) from being laid until hatched, then reared the 428 hatched orchid mantises until they die in the laboratory (located inside XTBG) from April 2020 to August 2021. We recorded the date and body weight (0.0001 g accuracy, electric analytic balance HZK-FA110S, Huazhi (Fujian) Electronic Technology Co., LTD) of each individual soon after hatching and each molt before feeding.
To study colour morph flexibility, we reared another 140 individuals and recorded the colour morphs of each individual for its entire life. The sex of each reared individual was determined when possible, according to the females' last slightly upturned-edge sternum and fewer coxosternites (Brannoch et al 2017).
Our rearing condition (light cycle, humidity and temperature) was not controlled, because our laboratory was located within the natural distribution region of the orchid mantis. The mantises had been kept individually in transparent plastic containers from the hatching stage. Here the newly hatched (first instar) and early life-stage nymphs (the second to the fourth instars) were kept in smaller containers (length: width: height = 5.7 × 5.7 × 4.2 cm), and later life-stage nymphs (fifth instar and above) were kept in larger containers (upper diameter: bottom diameter: height = 17.5 × 12 × 13.5 cm). Food was supplied according to life stage. For the first and second instar nymphs, lab reared fruit flies were supplied for convenience and as the tiny mantis cannot predate on larger preys like house flies; for the third instar nymphs onward, house flies were provided as the main diet, mixed with cockroaches, crickets and bees to fulfil the nutritional needs; for adult females, field-collected moths were offered as the main food, especially during the reproductive stage, as they could breed healthy and do normal reproduction only under frequent consumption of field collected moths and/or butterflies.

Population abundance and life-cycle patterns in the field
Population abundance was evaluated by population density calculated by the number of records per square kilometres per year. To analyse the life-cycle pattern, the field photographic records of H. coronatus were carefully checked to define life-cycle stages, including oothecae, newly hatched (first instar), flower-like stage (second instar to the last instar) and adult (male or female), then all records of each life-cycle stage were pooled monthly to visualise the monthly distribution with the package ggplot2 (Wickham 2016).

The relationship between life cycle and key environmental factors
Monthly temperature and rainfall were visualised then compared with the life-cycle plots visually to analyse the relation.

Microhabitats of orchid mantis in the field
All orchid mantises were recorded on plants, so microhabitats in our study refers to (1) where they were found on the plant i.e., on flower or leaf or other part of the plant, and (2) host plant diversity of the mantis. Therefore, we carefully checked and identified the host plants, the nomenclature followed (Duocet Group 2016), and then counted the number of host plant species, and analysed mantis number on each host plant species.

Life history of orchid mantis
We used laboratory records to analyse the life history of orchid mantis. Firstly, we analysed the duration of ootheca period (days from being laid until hatched), hatching duration (hours from the first nymph emerged until the last nymph dropped out) and hatching pattern (synchronous or asynchronous). Then based on the records of the 91 reared individuals (38 females and 53 males) that reached adulthood and had complete data records, we, firstly, calculated six life history aspects for each orchid mantis individual, i.e., (1) duration of each instar (days between two sequential molts), (2) number of molts before adulthood, (3) duration of development or length of nymphal period (days from hatching to the last molt), (4) duration of adulthood (days from the last molt until died), (5) whole life span (days from hatching until died), and (6) growth rate of each instar (between-two-sequential-molts body weight increase/duration); secondly, for both sexes, we (1) visualised the average duration of each instar and adulthood by stacked barplots to show a general developmental pattern of each sex; (2) visualised the inter-sexual differences of four main life history aspects with boxplots, i.e., the (a) number of molts, (b) adult body weight (body weights soon after the last molts), (c) duration of development, and (d) the whole life span; then the inter-sexual differences were statistically tested by Mann-Whitney-Wilcoxon tests (MWW), a powerful nonparametric statistic method for when the data does not fit the t-test assumptions (Kloke and McKean 2015). We then (3) visualised the body weights and growth rates of each instar and each sex during developmental period with scatterplots to show the growth pattern of nymphal orchid mantis, followed by detailed statistical tests of (a) body weights between two sexes at each event, i.e., when they just hatched, and every time they molted, (b) the growth rates between two sexes during each instar, and (c) development time of each life stage between two sexes; again we used MWW tests as some of our data does not fit the t-test assumptions, which were tested with the package performance (Lüdecke et al 2021).
Data visualisations were done with the package ggplot2 (Wickham 2016), all visualisations and statistical tests were performed in R 4.2.1 through RStudio 2022.07.1 + 554 (R Core Team 2022; RStudio Team 2022).

Colour morph diversity and flexibility along and within life stages
Firstly, we checked and classified the field photographic recorded of H. coronatus to different colour morphs and counted the frequency of each morph. The colour morphs were decided artificially with the same standard: the colour that dominates the entire body was recognized as the individual colour morph.
Then the field recorded morph diversity was verified by the reared individuals in our laboratory. We first defined the general colour morph pattern in the first filial generation with the same methods in field investigation and perceived that colour morph flexibility started to show from the fourth-instar stage. Then we repeated rearing conditions in the second filial generation for life history records, and successfully identified and assessed colour morph data for each life stage of 97 individuals who survived at least to the fourth-instar stage. The proportion of colour morph(s) in each flower-resemblance stage was calculated. To clarify the extent of colour morph shift in a series of life stages, we also counted individuals that shift colour morph from fourth-instar stage and survived to 1 3 maturity and obtained the rate of colour morph shift. These data were analysed and visualised by Excel Office 16.

Population abundance and life-cycle patterns in the field
Fifty-four orchid mantis records were collected in total (52 were inside XTBG and two were outside but within 3 km from XTBG), including one ootheca, two first instars, 39 flower-like juveniles (second instars to the last instar nymphs) and twelve adults (ten females and two males). Calculated population abundance is 0.44 individual/km 2 /year which is very low, suggesting an extremely low population density.
Monthly distribution results suggested that the orchid mantis produced one generation per year in the investigated region, without overlap of generations. Oothecae, first-instar nymphs and adults were only recorded in summer (April-September) (Fig. 1A-D), especially from May to August, which suggested that summer was their reproductive period. The nymph developmental period was from July to the following May, since oothecae and first-instar nymphs were only seen in July and August and larger nymphs were only seen in April and May (Fig. 1A-C).

The relationship between life cycle and key environmental factors
Visual comparisons between the life-cycle pattern and the local monthly fluctuations of temperature and rainfall suggested that the life-cycle pattern of the orchid mantis might be an evolutionary response to temperature and rainfall, since the reproductive period coincided with the highest average monthly temperature (above 25.9 °C) and rainfall (above 219.5 mm) within a year (Fig. 1E). Only two males were seen among the 12 adults, both during the early reproductive period (April and May), which suggested that males may mature earlier and have a shorter lifespan than females.

Microhabitats of orchid mantis in the field
The recorded 54 orchid mantises were all on plants (except for one adult female that was attracted to the light in a washing room in a yard of a house outside XTBG) and were scattered in various habitats (roadside, managed shrub region, managed tree plantations, secondary forest, etc.) within an area of approximately 22 km 2 . All of the recorded individuals were distributed separately in space and time. Specifically, no two individuals were recorded within 100 m of each other at the same time. Twelve individuals were recorded on plants that could not be identified, and the remaining 42 individuals were recorded on 38 plant species (five monocots, 32 dicots and one fern) belonging to 34 genera and 29 families (Table S1). Among the 38 plant species that we identified, 35 species belonging to 27 families were recorded only once with a single orchid mantis, and three species Litsea glutinosa (Laraceae), Heptapleurum heptaphyllum (Araliaceae) and Glycosmis pentaphylla (Rutaceae) were recorded two or three times with a single orchid mantis. These results suggest that the orchid mantis was distributed in various microhabitats and seemed not highly associated with any specific plant. Only two individuals were recorded on flowers. This suggested that the orchid mantis mainly stayed on green plant leaves rather than any flowers, which highly contrasted with its body colour.

Intersexual differences in duration of development, body weight and lifespan
The oothecae hatched after 36.4 ± 4.07 days (mean ± SD), and all eggs in the same ootheca hatched synchronously (Supplementary material, Video 1) in around two to three hours. The average number of baby orchid mantis hatched from one ootheca was 48 ± 29.28. The mean body weight of the newly hatched nymphs was 5.51 ± 0.69 mg and no difference between females and males was detected (W = 979.0, P = 0.82, MWW test) (Table S2). However, compared with males, females had significantly greater numbers of molts to maturity (female: 7.97 ± 0.28 times; male: 5.72 ± 0.45 times; W = 2014, P < 0.01), greater body weights after the last molt (female: 1367.00 ± 281.00 mg; male: 241.00 ± 46.00 mg; W = 2014, P < 0.01) (Table S2), faster pre-adult growth rate (female: 8.34 ± 1.69 mg/day; male: 2.26 ± 0.30 mg/day; W = 2014, P < 0.01), longer duration of pre-adult development (female: 164.0 ± 13.1 days; male: 104.0 ± 18.3 days; W = 1972, P < 0.01) and much longer lifespans (female: 192.00 ± 28.80 days; male: 138.00 ± 37.50 days; W = 1351, P < 0.01) ( Fig. 2A-E; Table S2, S3). During the ontogenetic period, the average duration of development (12.20 ± 1.52 days) and the growth rate (0.56 ± 0.14 mg/day) of the first instars did not differ significantly between males and females (W = 1172.0, P = 0.17 for development duration; W = 1067.5, P = 0.63 for growth rate); however, from the second instar onwards, the growth rates and body weights after each molt of females began to be significantly faster and heavier, respectively, than that of males, although the duration of development within each instar and the adult lifespan were longer in males than in females (Fig. 2F, G; Table S2, S3). Furthermore, although not measured, it was obvious that females' body weight remarkably increased from the newly mature to the reproductive stage, whereas males' body weight remained relatively stable until death. The heavier body weight in females was due to their longer development time and faster growth rate, especially in the later developmental stages (Fig. 2 A, F, G; Table S2, S3).

Colour morph diversity and flexibility among and within life stages
In the field, the first-instar nymphs and adult orchid mantis exhibited stable colour patterns (Fig. 3A-D), yet individuals in the flower-like stages (second instar to the last instar) presented four main colour morphs (white, pink, purple and yellow) (Fig. 3E-H). Specifically, the first-instar nymphs all exhibited a black-red colour pattern (Fig. 3D), presumably as a mimicry of Reduviidae bugs (Hawkeswood and Sommung 2019), yet we believe that they may mimic nymphal Macroceroea grandis ( Figure S1) given the high population density in its living habitats (personal observations). Adults of both sexes presented a brown-white colour pattern (Fig. 3A, B). The 39 flower-like juveniles exhibited four different colour morphs, including eighteen white, eight pink, nine purple and four yellow individuals (Fig. 3E-H), which suggested a high colour morph diversity of the orchid mantis. Among individuals raised indoor, the orchid mantis exhibited high flexibility of colour morphs during the flower-like stages but stable colour morphs during the first instar and adult stages as well. The newly hatched first-instar nymphs and the adults had the same colouration patterns as field mantises. After the first molt, they started to show a flower-like appearance and exhibited a stable colour pattern, with a mixture of yellow, pink, purple and white ( Figure S2, A). This colour pattern was transformed into a stable white-purple pattern after the second molt ( Figure S2, B). Individual colour patterns showed differences from the third molt to the last molt ( Figure S2, D) and were dominated by three colours: white, pink and purple (details in Figure S3). In 94.5% (68/72) of individuals, the main body colour morphs changed from the fourth instar onward. These results suggested that the colour morphs became highly flexible during the flower-like life stages.

Discussion
We investigated the biological features of the orchid mantis, including its field abundance, life cycle, microhabitats, colour morph diversity and the key environmental factors that may affect its life cycle, and then recorded its indoor life history and colour morph flexibility. The results suggest that the orchid mantis is extremely rare in the field which is consistent with experiences from previous experts (O'Hanlon 2016), and has highly diverse and flexible colour morphs which may help it to survive in variable microhabitats. Its intersexual differences in body size and lifespan may be evolutionary adaptations of sex role differences in a conspicuous masquerader for reproductive success.
Only 54 orchid mantis records were collected in 11 years. This suggested an extremely low abundance of the orchid mantis in the field and could reflected the adaptive difficulties of conspicuous masquerade under pressures from both predator deception and prey attraction, as occurs for the bird dropping spider Phrynarachne ceylonica (Yu et al 2021(Yu et al , 2022. The diversity and flexibility of the orchid mantids' body colour morphs may contribute to its fitness in various microhabitats, unlike successful mimics that often closely resemble a specific model and maintain the same spatiotemporal distribution as their models (Rettenmeyer 1970). Masqueraders resemble inedible and generally inanimate objects and often have highly flexible appearance Liu et al 2014), which may enhance the masquerader's adaptive success in multiple microhabitats (Skelhorn et al 2010a). Both the field and the indoor results showed that the orchid mantis exhibited diversity and flexibility of colour morphs. This suggests that different life stages may have different predator-prey interactions in orchid mantis and the colour morph ontogenetic variation may maximum its fitness. In addition, our results suggested that the orchid mantis did not coexist with a specific plant. These features implied that the orchid mantis was unlikely to resemble any specific flower but was likely to masquerade general flowers. This is consistent with the previous evidences Hawkeswood and Sommung 2019). Furthermore, the orchid mantis showed not only diverse colour morphs but also remarkable body size variations among life stages, which also obviously increased its diversity. This diversity and flexibility of appearance may be crucial for conspicuous masqueraders to survive in the tropical rainforest, which has a high diversity of microhabitats (Whitman and Agrawal 2009). In fact, according to the published literatures, most recorded conspicuous masqueraders are distributed in tropical rainforest (Liu et al 2014;Hawkeswood and Sommung 2019;Yu et al 2022). Adaptability to multiple microhabitats may be vital for conspicuous masqueraders, because, unlike herbivores that forage on motionless plants, the well-known conspicuous masqueraders are all tiny carnivores, such as spiders and mantises, which are thereby burdened more from high population density, such as fierce competition for prey, when they are distributed only within a specific habitat (Caro and Allen 2017;Wheatley et al 2020). In addition, we noticed that the shape and relative size of the femoral lobes changes enormously across instars but is stable within an instar, which suggest adaptive function of the femoral lobes. Here, we suggest the shape of the petal-shaped femoral lobes may have evolved under different selective pressures than those affecting its body colours.
The life-cycle pattern of the orchid mantis may be an evolutionary response to key local environmental factors (Hurd et al 2004). In our study area, the orchid mantis had a life-cycle pattern of one generation per year, without generational overlap. We suspected that its reproductive period was in the rainy season (May to August), which has high average temperatures and plentiful rainfall. There are three reasons for this conclusion. First, oothecae, newly hatched nymphs and all adults except two were only recorded during this period. Although two adults were observed in April and September (one in each month), we inferred that they were respectively newly matured and aged adults. Second, biological and abiotic factors in this season could meet the reproductive requirements of the orchid mantis. Female carnivorous insects need not only a large amount of food (Polis 1981) but also multinutrient foods (Krapu 1981) for incubating eggs and producing items such as oothecae (Boggs 1981). Various arthropods that served as food resources were much more abundant during the rainy season than during the rest of the year (Poulin et al 1992). Third, favourable temperature and humidity are vital for normal embryonic development 1 3 in insects (Howe 1967;Singh et al 2009), and both average monthly temperature (above 25.9 °C) and average monthly rainfall (above 219.5 mm) in XTBG remained at the maximum level from May to August. Therefore, reproduction during the rainy season should be an evolutionary adaptation to fluctuations in local environmental factors, whereas in the subhumid and dry seasons (from September to April) (Chen et al 2015), the lower temperature and rainfall may only satisfy the developmental demands of the nymphs.
The orchid mantis exhibited notable intersexual differences in body weight (heavier females), duration of development (protandry) and lifespan (females lived longer), which could be evolutionary responses for reproductive success. First, larger body size in females may be a result of selection for fecundity, which is universally documented in many arthropods, such as spiders (Head 1995) and insects (Honěk 1993;Berger et al 2008). Their larger body size enables females to produce greater numbers of offspring or larger offspring in a limited reproductive period (Pincheira-Donoso and Hunt 2017), which obviously benefits the orchid mantis by shortening its reproductive period while prolonging its embryo-hatching time for more oviposition trials. Second, we suggest that larger body size benefits adult females because they are motionless and hence are seldom noticed by predators. In contrast, the smaller body size of adult males may help them to escape from predators and search for mates, as being small may enhance their agility (Husak and Fox 2008). Intersexual variation in nymph developmental time (sexual bimaturism) may only be a by-product of the larger body size of female orchid mantises, as has been documented frequently in many taxa, especially in insects (Tammaru et al 2010; Morbey 2013). Protandry confers remarkable advantages by avoiding the drawbacks of inbreeding, guaranteeing the existence of intersexual reproductive synchrony, reducing virgin females' waiting costs, and increasing the number of male copulatory trials for reproductive success (Wiklund and Fagerström 1977;Morbey and Ydenberg 2001). The shorter duration of development and lifespan of males may explain why fewer adult males were recorded than females (two males and ten females) and why they were only recorded in the early reproductive stage (May) in the field. In addition, it needs to be investigated that whether this huge intersexual body size difference will cause reproductive difficulties, such as intersexual recognition, evaluation and mating body posture inefficiencies.
In conclusion, this is the first study to systematically investigate basic biological characteristics in a classic conspicuous masquerader, and it may provide crucial information for ecologists and evolutionary biologists to interpret how conspicuous masqueraders adapt and evolve. However, some essential questions are waiting to be addressed for further understanding of the masquerade phenomenon and the orchid mantis, including (1) Do intraspecific recognition and evaluation exist in conspicuous masquerader species, and if they do, how do they work? (2) What mechanism and factors cause the change and variety of colouration during the nymphal stage of orchid mantis? (3) How do males and females achieve reproductive synchrony with the huge dimorphism in development time and lifespan in the orchid mantis? Answers to these questions would contribute to further clarification of the evolutionary process and adaptive strategies of conspicuous masqueraders.