Introduction

Insect taxonomists have long been aware of the fact that adaptive genital asymmetries (i.e., asymmetries other than fluctuating asymmetry) are common and widespread in insects, but only recently have the relevant data been reviewed (Huber et al. 2007). In some insect orders genital asymmetries seem to be in the groundplan (e.g., Phasmida, Grylloblattodea, possibly also Embiidina, Mantophasmatodea), while in some orders the available data suggest several or up to dozens of independent origins (e.g., Heteroptera, Coleoptera, Diptera, Lepidoptera). The emphasis in Huber et al. (2007) was on the original data for individual insect orders, drawing information from the taxonomic, phylogenetic, ethological, and functional morphological primary literature. A relatively minor section was dedicated to evolutionary scenarios that may explain the observed patterns. About equal room was given to alternative scenarios, possibly blurring the fact that only one of them seems to best explain the majority of cases. In the present overview, I will concentrate on this scenario (which is largely built on ideas by Lamb 1922; Richards 1927; Ludwig 1932; Alexander 1964), on the evidence supporting each of its steps, and on ways in which it can be tested.

The basic idea

Figure 1 illustrates the main assumptions and evolutionary steps in the proposed scenario. The first assumption is a plesiomorphic symmetric female-above mating position. The first process involves a change of positions, in which the male moves to a position presumably giving him more control. The result is often some variant of male-above positions as shown in Fig. 1, but might also be a different position (belly-to-belly, side-by-side). At this stage, both male and female genitalia are still symmetric, and the position is individually random-sided, i.e., each individual male can mate from either side (or twist his abdomen or genitalia to either side). The inevitable and crucial consequence of most of these positional changes is asymmetric contact between male and female genitalia. This asymmetric contact results in different selective environments for each side, but only if an additional prerequisite is met: the position must become one-sided, i.e., an individual male must become restricted to mate from only one side. Different selective environments may then have a number of consequences, three of which are shown in Fig. 1: (1) right and left sides may change as to compensate for the mismatch resulting from asymmetric contact; (2) right and left sides may assume different functions; (3) one side may lose any function and become reduced. In some but not all of these processes, the female is expected to coevolve (not shown in Fig. 1), in rather conspicuous ways in an antagonistic coevolution scenario, in more subtle ways in a female choice scenario.

Fig. 1
figure 1

Simplified scheme for the evolution of insect genital asymmetries via changed mating positions; male abdomen shown in gray. a Plesiomorphic female-above position with symmetric genital contact and symmetric genitalia. b Derived mating position as a result of selection on males for more control; the position is individually random-sided (the male can mate on either side) and the genitalia are still symmetric. c Morphological asymmetry, correlated with shift to fixed one-sided mating position; (1) morphological compensation for mismatch; (2) division of labor between left and right structures; (3) one-sided reduction

Full size image

Evidence for the individual steps

Evidence for the female-above position and symmetric genitalia being plesiomorphic for Neoptera

Three types of evidence suggest that a symmetric female-above position and symmetric genitalia are plesiomorphic for neopteran insects: phylogenetic data, evidence from derived positions, and evidence from functional morphology.

A symmetric female-above position and symmetric genitalia are widespread in insects, including the Ephemeroptera (Morgan 1929; Snodgrass 1936; Despax 1949; Brinck 1957; Takemon 1990; Kluge 2003), the potential sister group of Neoptera (Ogden and Whiting 2003). I know of no formal cladistic analysis in insects including the character mating position, but a female-above position may by plesiomorphic for several orders including Blattaria, ‘Homoptera’, Mecoptera, Neuropterida, Orthoptera, Psocodea, and Siphonaptera (Huber et al. 2007). On the other hand, symmetric genitalia have been shown to be plesiomorphic in several formal analyses (Nelson 1984; Asche 1985; Landry 1991; Gielis 1993; Hodges 1998; Morse and Yang 2002; Yang and Morse 2002; Hebsgaard et al. 2004; Kaila 2004; Hsu and Powell 2005; Sihvonen 2005). In other cases, character mapping strongly suggests that asymmetry is derived (Orthoptera, Plecoptera, ‘Homoptera’, Heteroptera, Psocodea, Neuropterida, Trichoptera, Diptera, Lepidoptera; see Huber et al. 2007 and references therein). Change in the opposite direction, from asymmetry to symmetry, has apparently occurred too (Walker 1922 on Isoptera; Carayon 1977 on certain anthocorid and plokiophilid bugs; Hoch 2006 on Hawaiian Iolana planthoppers; Bickel 1987 and Sinclair and Cumming 2006 on certain dolichopodine flies; Gielis 1993 and Kaila 2004 on certain pterophorine and cosmopterigine moths; D. Ahrens, pers. comm. on some Sericini beetles), but when seen in a wider systematic context, all these cases seem to represent reversals to the original condition.

Several details of derived mating positions appear best explained assuming a plesiomorphic female-above position (Alexander 1964; Alexander and Otte 1967). In the widespread ‘false male-above position’ (e.g., Caelifera, Phasmida, Mantodea, Thysanoptera, Plecoptera, Embiidina, Grylloblattodea, many Heteroptera, etc.), the male sits on top of the female, but his abdomen is bent around the female abdomen and his genitalia actually contact the female from below (as in a female-above position) (e.g., Baunacke 1912; Fedorov 1927; Schrader 1930; Weber 1930; Hase 1932; Friederichs 1934; Keilbach 1935; Jordan and Wendt 1938; Larsén 1938; Rawat 1939; Stefani 1953; Mika 1959; Ross 1970, 2000; Jordan 1972; Stewart and Stark 1977; Whitman and Loher 1984; Terry and Dyreson 1996). The same is usually the case in side-by-side positions (e.g., many ‘Homoptera’, Panorpidae, some Heteroptera, etc.) (e.g., Weber 1930; Cook 1963; Davis and Usinger, 1970; White 1970; Kaltenbach 1978). The ‘circumversion’ (360° rotation) of the male genitalia of cyclorrhaphan flies would seem mysterious under the assumption of a plesiomorphic male-above position, but is easily explained by hypothesizing an evolutionary 360° rotation of the male relative to the female (e.g., a horizontal rotation from female-above to end-to-end, and from there in the same direction to male-above). The same is true for end-to-end and belly-to-belly positions that usually involve rotation of the abdomen or the genitalia by 180° (e.g., Dermaptera, Heteroptera, strophandrous Hymenoptera, Bittacidae) (Kuhl 1928; Herter 1963; Popham 1965; Bornemissza 1966; Mickoleit and Mickoleit 1978; Briceño and Eberhard 1995; Schulmeister 2001; Walker and Fell 2001; Kamimura 2006). Symmetric male-above positions seem to be rare and restricted to derived taxa, and are thus considered derived (e.g., some gerromorph bugs: Ekblom 1926; Heming-van Battum and Heming 1989; Schuh and Slater 1995). They may equally (cf. cyclorrhaphan flies above) result from an evolutionary 360° rotation of the male relative to the female, in this case not horizontal but vertical. In Lachesilla (Psocodea) such a vertical 360° rotation happens during copulation (Klier 1956).

Data from functional morphology also argue for a plesiomorphic female-above position. There seems to be an overwhelming dominance of ‘inverse correlation’ of the genitalia, i.e., the dorsal side of the male contacts the ventral side of the female (Richards 1927; Weber 1930; Hennig 1973; McAlpine 1981; Cobben 1982; Asche 1985; Heming-van Battum and Heming 1989; Sforza and Bourgoin 1998; Soulier-Perkins and Bourgoin 1998; Schulmeister 2001). This character is extremely conservative, probably due to the complex fit of male and female genitalia that does not allow the genitalia to simply rotate against each other when positions are changed. In some taxa, the external appearance suggests ‘direct correlation’ (i.e. dorsal side of the male contacts the dorsal side of the female) but internal structures (like the phallus) may be inverted (Lepidoptera, Heteroptera; Huber et al. 2007). A few cases of apparent direct correlation (e.g., Singh-Pruthi 1925 and Kunze 1959 on Cicadellidae; Tobias 1972 and Statzner 1974 on Trichoptera) need confirmation. In theory, both a symmetric female-above and a symmetric male-above position would imply inverse correlation, but as argued above, symmetric male-above positions are rare and almost certainly derived.

Evidence for selection on males to manipulate females and for derived mating positions being superior for this purpose

Male signals during copulation that are best explained as courtship are common and widespread in insects (Eberhard 1991, 1994). A wide range of processes are known to affect male paternity even after being accepted as a partner in copulation (Eberhard 1996), and copulatory courtship constitutes a conservative indication that selection in the form of cryptic female choice may be acting on the males (Eberhard 1991). Many of the copulatory behavior patterns listed in Eberhard (1991, 1994) involve the male legs and mouthparts, and it seems obvious that male-above positions give the male a wider range of opportunities to use these body parts as signaling devices during copulation.

Apart from signaling, male-above positions may also confer the male an advantage with respect to physically restraining the female and give him more control over the termination of copulation. In some Psocodea, the female may terminate copulation by simply turning to the side (Mockford 1957). Obviously, a male-above position will impede this female move. Male legs are usually preadapted to grasp the female, male antennae and wings are not. It is illuminating that in some groups that have retained the female-above position, male antennae and wings have indeed been modified to perform grasping and holding functions (e.g., Mitzmain 1910; Holland 1955; Rothschild and Hinton 1968 on Siphonaptera; Carpenter 1936; Steiner 1937; Crampton 1940; Cooper 1974; Mickoleit and Mickoleit 1976 on Boreidae; Séguy 1944 and Schmutz 1955 on certain Phthiraptera). Male mayflies and certain phthirapteran lice have modified front legs that are used to grasp (and/or stimulate) the female from below (Nuttall 1917; Morgan 1929; Despax 1949; Brinck 1957; Kluge 2003).

Evidence for fixed, one-sided positions being derived from individually random-sided positions

As indicated above, information on mating positions is virtually absent in phylogenetic studies. However, indirect data strongly support behavioral flexibility (individually random-sided positions) to be plesiomorphic and fixed one-sided positions to be derived. This indirect data comes from apparent strict correlations of individually random-sided positions with symmetric genitalia and one-sided positions with asymmetric genitalia. This has been particularly well studied in the Heteroptera (references in Huber et al. 2007), but supporting evidence is widespread (Caelifera, Plecoptera, Embiidina, Bittacidae, Papilionidae, Mantophasmatodea, strophandrous Hymenoptera; Huber et al. 2007). Exceptions are extremely rare (e.g., Berthélemy 1979 on Brachypterainae) and need confirmation.

This strict correlation, together with the fact that symmetry appears plesiomorphic and asymmetry derived (see above) argues for random-sided positions to be plesiomorphic and one-sided positions to be derived. What could be the advantage of a one-sided over a random-sided position? From the male’s perspective, this change seems disadvantageous, as it reduces his options. The negative effect must be outweighed by some advantage of morphological asymmetry (see below), as the latter is always correlated with one-sided mating positions.

Evidence for selection favoring morphological asymmetry and one-sided positions

Genitalia are often considered multifunctional organs, and one way to achieve functional complexity is by specializing individual components for different tasks. Different tasks may be executed simultaneously or sequentially. In the moth Erynnis persius, for example, the right male valva has been observed to remain stationary, grasping the female while the left valva was used to scrape her sternum VII (Scott 1978). A similar division of labor is suggested by the terms “titillator” and “hook” for the left and right male genital structures respectively, in Zorotypus hubbardi (Walker 1922) but both structures may actually be derived from the same side (Walker 1922). In many other cases, conspicuous differences between left and right sides (e.g., Kuznetzov and Baryshnikova 2004 on Phyllonorycter moths) suggest functional segregation but no functional studies exist. In Periplaneta and Blatta, the tasks performed by right and left structures seem to be similar (to clasp the female) but one side acts during the initial phase of copulation while the other side assumes its function later (most cockroaches change position during mating) (Gupta 1947—cited in Scudder 1971; Khalifa 1950). In certain Corixa species and certain Miridae, one of the two parameres is used as a guide for the intromittent structure (Larsén 1938; Kullenberg 1947).

An alternative to division of labor between right and left sides is compensation for genital mismatch resulting from an asymmetric mating position. Evidence for such morphological adjustment to behavioral asymmetry is difficult to get. One prediction is correlation between the flexibility of the abdomen and the conspicuousness of the asymmetry. In taxa with less flexible abdomens the asymmetry should be more conspicuous because the abdomen cannot compensate for the mismatch. True bugs seem to support this prediction. Many true bugs have poorly flexible abdomens, and the asymmetries are often conspicuous (e.g., flat nepomorph bugs, Schuh and Slater 1995). In some taxa, however, the abdomen has secondarily acquired higher flexibility including asymmetric muscles, and their genitalia are symmetric (e.g., Keilbach 1935 and Larsén 1938 on Notonecta glauca).

A fixed, one sided position may result in one side of the male genitalia losing its function. If that is the case, it may either adopt a new function, be modified as to be able to continue perform the original function, or disappear. In the latter case one would expect to find secondarily unpaired male genital structures. Cimicidae and relatives may exemplify this route to morphological asymmetry (Carayon 1977). Other examples are the Eudermaptera where the left penis lobe is completely reduced (Hincks and Popham 1970; Haas 1995; Haas and Kukalova-Peck 2001), and the Ochteroidea where the left paramere is reduced or absent (Schuh and Slater 1995).

Evidence for female asymmetry appearing after male asymmetry

The taxonomic literature seems to strongly suggest that genital asymmetries are much rarer in female that in male insects. In many groups there is no evidence for female asymmetry even though the male genitalia are strongly asymmetric (e.g., Phasmida, Grylloblattodea, Mantophasmatodea, Psocodea, references in Huber et al. 2007). In some taxa this may simply be due to the fact that female genitalia are barely studied (e.g. Embiidina, Zoraptera, Dermaptera), but in most taxa this seems to be a real trend. For example, morphological analyses of the female genitalia of Timema (Vickery 1993; Tilgner et al. 1999) revealed no asymmetry while the male genitalia are strongly asymmetric (Snodgrass 1937; Vickery 1993; Tilgner et al. 1999; Bradler 1999). Comparable in-depth studies are rare, but in several groups male and female genitalia are routinely illustrated by taxonomists (e.g. Trichoptera, Heteroptera, Lepidoptera, Diptera), and in these, character mapping generally resolves female asymmetry as being nested within taxa showing male asymmetry (Huber et al. 2007).

Testing the scenario

In contrast to the processes leading to asymmetry, the patterns resulting from the above scenario appear easy to test. Taxonomic studies need to be specific about the direction of asymmetry (directed, i.e., all males within a species are same-sided, versus bidirected); ethological studies need to be specific about the direction of mating positions (random sided versus one-sided); phylogenetic studies need to include both male and female asymmetries, as well as mating positions in the character matrix (provided there is variation within the studied taxon). Formal cladistic analyses including genital asymmetries are still rare (Nelson 1984; Asche 1985; Landry 1991; Gielis 1993; Hodges 1998; Morse and Yang 2002; Yang and Morse 2002; Hebsgaard et al. 2004; Kaila 2004; Ahrens 2005; Hsu and Powell 2005; Sihvonen 2005; Hoch 2006). One reason may be that asymmetries often evolve many times independently at or near terminal branches. Thus, they introduce noise rather than resolution, and excluding them from the analysis may result in a ‘better’ tree (or at least better tree statistics).

One field that appears in particular need of new data is functional morphology. In many groups, our understanding of functional details continues to be “almost unbelievably rudimentary” (Alexander and Otte 1967). Traditional studies tended to be not only typological but often relied on specimens that were by chance fixed during copulation. Only detailed observations of the positions and movements of right and left structures in asymmetric species may show if division of labor is a common phenomenon or not. In general, division of labor predicts a higher degree of specialization for different functions in asymmetric versus symmetric genitalia.

The case of termites suggests a possible test for the causal correlation between sexual selection and genital asymmetry. Termites have secondarily simple and symmetric genitalia, and this may be a result of relaxed sexual selection due to female monogamy (Eberhard 1985; Huber et al. 2007). Several further cases of reversal to symmetry have been reported (see above), and these might be correlated with similar changes in mating systems. On a similar line, one might predict changes to female-controlled positions in taxa with sex-role reversal. A possible example is Paravelia brachialis (Heteroptera, Veliidae), where the female mounts the male and stimulates him to copulate (Wilson 1958—cited in Heming-van Battum and Heming 1989). Depending on the mating position of the outgroup taxon, this might even select for female asymmetry.

At least in taxa with strongly asymmetric male genitalia, sexual selection theory would seem to predict certain levels of asymmetry in female genitalia. I propose that in an antagonistic coevolution scenario, where traits are supposed to evolve to overcome or defeat the other sexes’ resistance or manipulation, females should be expected to respond to conspicuous male asymmetries with similar (i.e., often conspicuous) morphological changes. In a female choice scenario, females are less expected to react conspicuously but to at least adapt their sensory system to male asymmetries. In general, the rareness of conspicuous female asymmetries does not seem to support antagonistic coevolution as being a widespread selective force shaping insect genitalia. Detailed studies of the female genital sensory system in species with strongly asymmetric male genitalia are largely missing but might shed light on the importance of cryptic female choice.

Finally, comparative evidence may come from groups with similar trends towards asymmetric genitalia. In crustaceans, for example, asymmetric genitalia are common in certain groups (e.g., Copepoda) but absent in others (Ludwig 1932). No recent review on asymmetric genitalia in crustaceans seems to exist.

Conclusion

The single most important first step towards insect genital asymmetry appears to be a switch of mating position from a symmetric female-above to an asymmetric male-above position. The driving force behind this switch is presumed to be selection on the male for more control over the female and over details of copulation. Once an asymmetric position is established, left and right genital structures may experience different selective environments and thus become asymmetric. The fact that genital asymmetries have evolved many times independently in a variety of insect orders opens the possibility to test this scenario and its components.