Sexual selection is predicted to favor the exaggeration of pre- and post-mating traits whenever the extent of their elaboration increases reproductive success through male-male competition or female choice (Darwin 1871; West-Eberhard 1983; Andersson 1994; Emlen 2008; Lüpold et al. 2016). Since sexual selection is stronger on males than on females in most species (Bateman 1948; Trivers 1972; Clutton-Brock and Parker 1992), traits acting as ornaments or armaments often evolve absolutely or relatively greater expression in males relative to females (i.e. male-biased sexual dimorphism). However, trait elaboration is typically costly in terms of energy and fitness (e.g., Somjee et al. 2018; Rometsch et al. 2021), and a male’s resource allocation to sexual traits —and thereby the net fitness benefits gained— depends on the somatic, genetic, or epigenetic condition. Condition dependence is a form of developmental plasticity where an individual’s available metabolic resources determine the extent of trait expression by optimizing the relative resource allocation between somatic maintenance and reproduction (Nur and Hasson 1984; Andersson 1986; Rowe and Houle 1996; Hill 2011). Although any trait has the potential to vary with condition, sexually selected traits are predicted to be particularly sensitive to condition due to their diversion of resources from somatic maintenance and survival (Rowe and Houle 1996; Cotton et al. 2004).
With stronger sexual selection on males than on females, resulting in the evolution of male-biased sexual dimorphism in condition-dependent traits, it follows that condition dependence itself should be sexually dimorphic. In other words, variation among males in the expression of sexually selected traits should be more tightly linked to their bearers’ underlying condition than that among females. Across multiple traits, then, those that are more sexually dimorphic should also show greater divergence in sex-specific condition dependence, mediated by sex-specific resource diversion from somatic maintenance and differential viability costs (Rowe and Houle 1996). Even though theory predicts such links between sexual selection, sexual dimorphism and condition-dependent trait expression, however, only few empirical studies have integrated some of these predictions (Bonduriansky 2007a,b; Oudin et al. 2015; Miller et al. 2016; Rohner and Blanckenhorn 2018).
In addition to the paucity of studies exploring condition dependence of traits in the context of sexual dimorphism, other limitations in interpreting the link between sexual selection and condition dependence are also prevalent. Specifically, many studies provide correlational rather than experimental evidence of condition dependence, and they are often limited to a focal trait without appropriate control traits or accounting for variation in body size (reviewed in Cotton et al. 2004; but see, Rohner and Blanckenhorn 2018; Fox et al. 2019; Cattelan et al. 2020, for some recent examples addressing these issues). Stronger and less biased evidence can come from comparisons across multiple traits that inform about their relative condition dependence within the same set of individuals, thereby placing the trait of interest in the context of general trait variation (Arnqvist and Thornhill 1998; Bonduriansky and Rowe 2005; Fairbairn 2005; Rohner and Blanckenhorn 2018). Additionally, since individuals can vary in their sensitivity to developmental stress, in their efficiency in turning acquired resources into growth and in their resource allocation strategy, studying condition dependence in a genetic context seems particularly important for sexually selected traits, in which trait elaboration is often assumed to signal genetic quality (Iwasa et al. 1991; Rowe and Houle 1996). Yet, the genetic contribution to condition dependence is rarely studied (but see, David et al. 2000; Kemp and Rutowski 2007; Hubbard et al. 2015).
In insects, as in many other taxa, the relative size of sexual traits is strongly influenced by the resource availability during juvenile development (David et al. 2000). In cyclorrhaphan Diptera, adult structures grow mostly during late larval and pupal development. Because late third-instar larvae and pupae cannot acquire more energy by feeding, the developmental precursors underlying different adult tissues develop in an energetically closed system wherein they directly compete for resources (Nijhout and Emlen 1998; Heming 2018). In adults with juvenile development under different nutritional conditions, shifts in relative resource allocation to trait growth directly relate to the dependency of trait expression on resource availability, i.e. condition (Rohner and Blanckenhorn 2018; Shingleton and Frankino 2018). Studying nutrition-dependent trait expression in flies using static allometries (i.e., the degree to which trait size changes with overall body size) hence permits testing whether trait exaggeration and sexual dimorphism relate to (sex-specific) condition dependence, and how the expression of one trait depends on the investment in others.
An ideal system to study how sexual selection drives sex-specific condition dependence is the fruit fly Drosophila prolongata. As an exception within the Drosophilidae, D. prolongata males are much larger than females (Kudo et al. 2015; Rohner et al. 2018a,b) and develop dramatically enlarged forelegs with conspicuous black and white stripes (Setoguchi et al. 2014; Fig. 1). Males use these forelegs to strike their opponents in dyadic fights (Kudo et al. 2015), and to wave at, or occasionally stimulate the abdomen of, the female during courtship (Setoguchi et al. 2014, 2015; Perdigón Ferreira and Lüpold 2022). These functions, combined with the male-biased sexual dimorphism in foreleg expression and overall body size, suggest intense premating sexual selection on males, with forelegs as a primary target. For example, Perdigón Ferreira and Lüpold (2022) showed that males stimulating the female’s abdomen during courtship through “leg vibration” (Setoguchi et al. 2014, 2015) had significantly higher mating success compared to males that did not show this behavior.
Here, using D. prolongata isofemale lines, we investigated the link between sexual dimorphism and sex-specific condition dependence (i.e., static allometries) in a genetic context. Overall, we tested the hypothesis that traits under sexual selection in males (e.g., exaggerated forelegs) show higher levels of male-biased sexual dimorphism and condition dependence than other traits (e.g., hindlegs not involved in mating behavior). Rather than focusing solely on total leg size, we also tested for differential patterns of condition dependence in different parts of the leg to infer a possible role of any such part in the unique male behaviors.