Sexual coevolution in the traumatically inseminating plant bug genus Coridromius


Nikolai J. Tatarnic, Evolution & Ecology Research Centre, University of New South Wales, Sydney 2052, Australia.
Tel.: 61 2 9385 8565; fax: 61 2 9385 1558; E-mail:


Sexual conflict has recently been proposed as a driving force behind the rapid diversification of genitalia among sexually reproducing organisms. In traumatically inseminating insects, males stab females in the side of the body with needle-like genitalia, ejaculating into their body cavity. Such mating is costly to females and has led to the evolution of cost-reducing ‘paragenitalia’ in some species. Whereas some consider this evidence of sexually antagonistic coevolution, others remain unconvinced. Variation in the reproductive morphology of both sexes – particularly males – is alleged to be negligible, contradicting the expectations of a coevolutionary arms race. Here, we use a phylogeny of the traumatically inseminating plant bug genus Coridromius to show that external female paragenitalia have evolved multiply across the genus and are correlated with changes in male genital shape. This pattern is characteristic of an evolutionary arms race driven by sexual conflict.


Mating sometimes involves conflict between males and females, with each sex vying to promote its own interests, often at a cost to its partner (Chapman et al., 2003; Arnqvist & Rowe, 2005). This sexual conflict can result in the evolution of sexually antagonistic structures and behaviours, with any adaptive advantage gained by one sex soon matched by counter-adaptations in the other, generating an evolutionary arms race (Parker et al., 1979; Alexander et al., 1997; Rice & Holland, 1997; Arnqvist & Rowe, 2005). Sexually antagonistic coevolution (SAC) has been posited as a driving force behind morphological diversification (Holland & Rice, 1998; Arnqvist & Rowe, 2005; Rowe & Westlake, 2006) and as a catalyst for speciation (Arnqvist et al., 2000; Gavrilets & Waxman, 2002; Martin & Hosken, 2003), yet few compelling examples have been documented, in part because of the difficulty in identifying complementary antagonistic traits in both sexes.

One of the most overt manifestations of sexual conflict is traumatic insemination, where males use piercing genitalia to inseminate females by injecting sperm into the female’s body cavity through the abdominal wall (Carayon, 1959, 1966), circumventing the female reproductive tract. Traumatic insemination has evolved several times in invertebrates but is most widespread in the true bug infraorder Cimicomorpha, where it has evolved independently three times (Tatarnic et al., 2006), and is best known in bedbugs (Carayon, 1959, 1966; Stutt & Siva-Jothy, 2001; Morrow & Arnqvist, 2003; Reinhardt et al., 2003; Siva-Jothy, 2006; Reinhardt & Siva-Jothy, 2007). Traumatic insemination in true bugs is believed to have arisen either in response to female pre- or post-mating resistance (Arnqvist & Rowe, 2005) or as a by-product of male-male conflict over sperm precedence (Arnqvist & Rowe, 2005; Řezáč, 2009). Regardless of its origins, traumatic insemination is costly to females, resulting in direct physical damage (Morrow & Arnqvist, 2003) and increased risk of infection (Reinhardt et al., 2003). To mitigate the harm caused by traumatic insemination, females of many taxa have evolved de novo internal and external paragenital modifications at the site of intromission, collectively known as the spermalege (Carayon, 1959, 1966). Spermalege morphology of females varies across related taxa, ranging from complete absence to the presence of specialized internal and external structures (Carayon, 1959, 1966; Tatarnic et al., 2006; Tatarnic & Cassis, 2008). In contrast, male genitalia in traumatically inseminating species are assumed to be more or less invariant across species (Hosken & Stockley, 2004; Eberhard, 2006). This apparent lack of variation in males, and subsequent lack of covariation between the sexes, contradicts the expectations of SAC theory (Morrow & Arnqvist, 2003) and has been used to contest the significance of sexual conflict in the evolution of traumatic insemination (Hosken & Stockley, 2004; Eberhard, 2006). To date, no assessment of male variability and no phylogeny of a traumatically inseminating lineage have been published to assess this.

In species of the traumatically inseminating plant bug genus Coridromius (Heteroptera: Miridae), males have a piercing intromittent organ very similar to that of bedbugs (Tatarnic et al., 2006; Tatarnic & Cassis, 2008). This organ is usually scythe-shaped but often varies in length, curvature, degree of twisting along its long axis, and thickness and angle of the apex (Figs 1i–l and 2a). In some species, the organ even takes the form of a tightly coiled corkscrew (e.g. Figs 1l and 2a, C. sommelieri). Similarly, females exhibit a wide range of external spermalege morphology (Figs 1a–h and 2a). In those species without an external spermalege, males inseminate females by stabbing them in the intersegmental membrane between abdominal segments (Fig. 1d). In other species, females have evolved hardened cuticular grooves (Figs 1e and 2a, C. hermosus) and/or copulatory tubes formed by invagination of the abdominal cuticle (Figs 1f–g and 2a, C. tahitiensis, C. sommelieri, C. nakatanii) to guide the male intromittent organ and restrict damage. These structures are sometimes accompanied by asymmetrical swelling and desclerotization of the abdominal tergites and laterotergites, which is thought to accommodate expansion of the underlying mesodermal component of the spermalege as it fills with sperm (Figs 1b–c and 2a, C. hermosus) (Tatarnic et al., 2006; Tatarnic & Cassis, 2008). Several modifications of nonabdominal structures are also implicated in mating. These include the presence of a plate-like lobe projecting from the posterior margin of the thorax to form a barrier partially obscuring the site of intromission and surrounding region (Figs 1e–f and 2a, C. hermosus), and modification of the lateral margin of the forewing to form a sclerotized tube to guide the male intromittent organ to the site of insemination (Fig. 1h). (For detailed images of all salient features see Tatarnic & Cassis, 2008.) Within Coridromius, males and females show more variation in reproductive morphology than is observed across all other traumatically inseminating true bugs.

Figure 1.

 Traits contributing to male and female complexity indices. Female paragenitalia and associated structures (a–h). In females, the right anterior dorsal laterotergites (yellow) may be unmodified (a), slightly swollen and desclerotized (b), or extremely swollen and desclerotized (c), with adjoining laterotergites sometimes fused together to form various dorsal or dorsolateral tubercles. When viewed laterally, females of some species exhibit no obvious site of insemination (d), whereas in others (e), the posterior margin of the second abdominal sternite (yellow) is flared and carinate, exposing a hardened intersegmental copulatory groove (grey), which is often combined with the thoracic metepimeral lobe (blue) extending caudally and partially obscuring the site of insemination. The copulatory groove sometimes leads to an invaginated copulatory tube (black) projecting into the abdominal cavity (f). Some species exhibit a shift in the site of insemination from between abdominal segments II and III to within segment II (g), or dorsolaterally (h). In the latter, the lateral margin of the hemelytron (red) is sometimes recurved and tube-shaped, acting as a copulatory guide. Male intromittent organ (i–l). In many species, the male intromittent organ (formed through the coupling of the aedeagus and left paramere) forms a scythe-like structure, with the following variants: the apex (green) is broad, the medial portion (purple) is evenly curved and of even thickness, and the basal portion (orange) is not significantly thickened. In its more derived form (j), the apex is thin and deflected away from the long axis, the medial portion is minimally curved and sometimes dorsally keeled (not shown), and the basal portion is thickened and/or laterally compressed. These traits may sometimes be combined with a single basal coil (k) or multiple coils running the length of the organ (l).

Figure 2.

 Phylogeny and distribution of Coridromius species. (a) Map of Coridromius species collection localities, generated through the Plant Bug Inventory website (http// North American records of the introduced Australian species C. chenopoderis are omitted. Yellow and red dots represent western and eastern clades, respectively. Insets show male and female reproductive traits of exemplar taxa, all of which exhibit correlated reproductive complexity indices. From left to right: C. nakatanii (Laos), C. hermosus (Papua New Guinea), C. tahitiensis (Tahiti) and C. sommelieri (Sabah). b: Phylogeny based on parsimony analysis of 67 morphological characters. Length = 135, consistency index = 0.62, retention index = 0.90. Values above branches are bootstrap/jackknife values for 100 replicates, and values below branches are Bremer support values for 1000 replicates. Branches lacking terminal squares indicate species known from only one sex. Branch colours indicate major clades, as shown on map. See Data S1 for full methods, analyses and matrices, and ref. 11 for detailed descriptions and images of all species and salient characters.

Materials and methods

In this study, we tested for correlation in the complexity of male and female reproductive morphology using a phylogenetic reconstruction of Coridromius and phylogenetically independent contrasts (Felsenstein, 1985). Because female response traits are expressed unevenly across the genus and involve modifications of various body parts, female complexity could not be measured using standard morphometric techniques. Instead, we quantified variation in female and male reproductive structures by generating complexity indices for both sexes of each species, following the methodology of Kuntner et al. (2009). These indices are the sum of the total number of salient characters expressed by the male or female of a species. Male complexity was based on aspects of intromittent organ shape, whereas female complexity was derived from spermalege form and associated abdominal, hemelytral and thoracic traits (see Data S1). A phylogeny of the genus based on 67 unordered morphological characters was constructed under maximum parsimony in TNT 1.1 (Goloboff et al., 2008), and the correlation of male and female complexity indices was then assessed using phylogenetically independent contrasts, as implemented in the PDAP plug-in (Midford et al., 2009) for Mesquite 2.6 (Maddison & Maddison, 2009). For further details see the Data S1, available online.


Based on our phylogeny, Coridromius is divided into two major lineages, one western clade spanning Africa, subtropical Asia and Southeast Asia, and an eastern clade ranging from western Southeast Asia through Australia and the Pacific Islands (Fig. 2). Mapping male and female reproductive complexity onto the phylogeny reveals a pattern of repeated emergence and elaboration of female external paragenitalia (Fig. 3), coupled with increased male genitalic complexity, which occurs independently in both eastern and western clades. Male and female complexity are highly correlated (Fig. 4; 2-tailed t-test, t22 = 4.227–5.886, P < 0.001; see Data S1) – as female external reproductive morphology becomes more complex, this is matched by changes in the shape of the male intromittent organ. This coevolution is most obvious in species such as C. nakatanii and C. sommelieri, where males are endowed with an intromittent organ matching the shape and size of the copulatory tubes of females (in the former both are ‘C’ shaped, in the latter both are corkscrewed: Fig. 2a). In addition, across the phylogeny, there are multiple occurrences of male intromittent organ and female paragenital elaboration, with modifications in the male rarely unmatched by the female (Fig. 3, C. marmoreus, C. punctatus).

Figure 3.

 Male and female genital and paragenital complexity are correlated. Male (left) and female (right) complexity indices mapped onto the generic phylogeny. Taxa known only from one sex have been pruned from the tree.

Figure 4.

 Positivized independent contrasts of male vs. female complexity (number of contrasts = 23, degrees of freedom = 22, r2 = 0.616, t = 5.942, P < 0.001). All branch lengths set to zero. Regression lines: red/dark grey = Reduced Major Axis; green/light grey = Major Axis; black = Ordinary Least Squares. Results for other branch length estimates are found online in the Data S1.


An apparent lack of variation in male genitalia has been used as evidence to refute SAC as a driving force in traumatic insemination and paragenital evolution. Our results show that in Coridromius, the male intromittent organ exhibits intraspecific variation, which is strongly correlated to female spermalege complexity. Furthermore, diversification of male genital morphology has in some species been met with little or no corresponding external spermalege development of females. This suggests that in these species male specialization may precede that of females and is thus driving coevolution in this system. Whether superficially unmodified females are coevolving in ways other than through external morphology (e.g. via internal spermalege modification, physiological response, or resistance behaviour) remains to be tested.

Although our study reveals a clear pattern of morphological covariation, the functional significance of these physical changes is yet to be determined. It is not known, for example, how the various changes in male genital shape might alter costs in females (Reinhardt & Siva-Jothy, 2007), or conversely, how changes in female morphology might shape male genital form. In bedbugs, it has been speculated that variation in the curvature of the male intromittent organ may wound the female differently, or change the precision with which a male targets a specific site within the spermalege (Reinhardt et al., 2003). In Coridromius, some of the male genital modifications may function to brace and reinforce the organ (e.g. basal thickening, lateral compression), which we hypothesize as a coevolutionary response to female defensive traits (e.g., increasing sclerotization at the site of intromission).

Although our study does not rule out the possibility that other factors may be contributing to genital coevolution (e.g. cryptic female choice –Thornhill, 1983; Eberhard, 1996), in light of the strong evidence supporting sexual conflict in bedbugs (Stutt & Siva-Jothy, 2001; Morrow & Arnqvist, 2003; Reinhardt et al., 2003; Siva-Jothy, 2006), it seems most likely that such conflict is also a significant driver of evolution in this system.

The recent discoveries of traumatic insemination in Coridromius (Tatarnic et al., 2006), Drosophila (Kamimura, 2007) and spiders (Řezáč, 2009) reveal that this form of mating is more common than once thought, implying that it must provide some benefit, presumably to males (Pfiester et al., 2009), and that the underlying selective forces (e.g. sexual conflict) may be ubiquitous in gonochorist species. Given the natural replication of traumatic insemination in invertebrates, this reproductive behaviour provides a model system to test theories of sexual conflict and coevolution. Our phylogenetic approach provides a macroevolutionary perspective, which is so far lacking in all previous investigations of traumatic insemination, and near absent in sexual conflict studies.


We thank M. Kuntner for discussions regarding the methodology used in this analysis and R. Bonduriansky for reading an earlier version of the manuscript. This project was funded through an NSF Planetary Biodiversity Inventories Grant to R. T. Schuh and G. Cassis, and a grant from the Australia & Pacific Science Foundation to N. J. Tatarnic and G. Cassis.