Selection for predation, not female fecundity, explains sexual size dimorphism in the orchid mantises

Abstract

Here we reconstruct the evolutionary shift towards floral simulation in orchid mantises and suggest female predatory selection as the likely driving force behind the development of extreme sexual size dimorphism. Through analysis of body size data and phylogenetic modelling of trait evolution, we recovered an ancestral shift towards sexual dimorphisms in both size and appearance in a lineage of flower-associated praying mantises. Sedentary female flower mantises dramatically increased in size prior to a transition from camouflaged, ambush predation to a floral simulation strategy, gaining access to, and visually attracting, a novel resource: large pollinating insects. Male flower mantises, however, remained small and mobile to facilitate mate-finding and reproductive success, consistent with ancestral male life strategy. Although moderate sexual size dimorphisms are common in many arthropod lineages, the predominant explanation is female size increase for increased fecundity. However, sex-dependent selective pressures acting outside of female fecundity have been suggested as mechanisms behind niche dimorphisms. Our hypothesised role of predatory selection acting on females to generate both extreme sexual size dimorphism coupled with niche dimorphism is novel among arthropods.

Evolutionary diversification in the raptorial forelegs of Mantodea: Relations to body size and depth perception

“…All the traits scaled allometrically with pronotum length (Figure 3, Table 1), a common measure of body size among arthropods (Svenson et al, 2016). These results are in contrast to an intraspecific study on Stagmomantis theophila that showed isometric growth of the forelegs in relation to body mass (Sutton, Doroshenko, Cullen, & Burrows, 2016).…”

“…When possible, I included both males and females, measuring 1–3 individuals per sex, per species. While there is dimorphism in some mantis species (Svenson et al, 2016), I averaged males and females as none of the specimens measured represented known major dimorphic species. Furthermore, because some species had only one sex to measure, or the sex was unclear, averaging the sexes better represented the interspecific diversity.…”

“…The coevolution of shorter forelegs may then be a result of selection on narrow bodies and depth perception and not a particular prey type per se. Conversely, other groups have independently evolved wider heads, and thus farther depth perception, with proportionately longer forelegs, including the flower mimics such as the orchid mantis ( Hymenopus coronatus ), who have also been shown to be under selection for predation (Svenson et al, 2016), and leaf mimics such as the dead leaf mantis ( Deroplatys desiccata ). The longer forelegs likely increases the displacement of the forelegs, increasing their reach, which may be beneficial for these habitat mimics (Anderson et al, 2014).…”

“…These three segments act as simple mechanical levers, therefore variation in the length of any one segment may alter feeding performance (e.g., displacement of the segment). Variation in length of foreleg segments may also reflect functional specializations related to the ecology of the species, similar to terrestrial vertebrates, such as habitat mimicry or cursorial versus ambush predation (Svenson, Brannoch, Rodrigues, O'Hanlon, & Wieland, 2016; Svenson & Whiting, 2004, 2009).…”

An elusive Neotropical giant, Hondurantemna chespiritoi gen. n. & sp. n. (Antemninae, Mantidae): a new lineage of mantises exhibiting an ontogenetic change in cryptic strategy

“…Representatives of Chaeteessidae and Mantoididae were included to root the phylogeny based on their consistent recovery as the two earliest branches of Mantodea (Svenson and Whiting 2009, Wieland 2013, Svenson et al 2015). We assembled a molecular dataset from previously published works (Svenson and Whiting 2004, 2009, Svenson et al 2016) and newly generated sequence data (see Table 1 for GenBank accession numbers). The molecular dataset included four genes: the mitochondrial cytochrome oxidase I (COI) and NADH dehydrogenase subunit 4 (ND4) and the nuclear Histone subunits 2 and 3 (H2A and H3).…”