A time-sequence functional analysis of mating behaviour and genital coupling in Drosophila: role of cryptic female choice and male sex-drive in the evolution of male genitalia


Rama S. Singh. Department of Biology, McMaster University, Hamilton, Ontario L8S 4K1. Canada.
Tel.: (905)525-9140 (ext 24378); fax: (905)522-6066;
e-mail: singh@mcmaster.ca


Male genitalia in Drosophila exemplify strikingly rapid and divergent evolution, whereas female genitalia are relatively invariable. Whereas precopulatory and post-copulatory sexual selection has been invoked to explain this trend, the functional significance of genital structures during copulation remains obscure. We used time-sequence analysis to study the functional significance of external genitalic structures during the course of copulation, between D. melanogaster and D. simulans. This functional analysis has provided new information that reveals the importance of male-driven copulatory mechanics and strategies in the rapid diversification of genitalia. The posterior process, which is a recently evolved sexual character and present only in males of the melanogaster clade, plays a crucial role in mounting as well as in genital coupling. Whereas there is ample evidence for precopulatory and/or post-copulatory female choice, we show here that during copulation there is little or no physical female choice, consequently, males determine copulation duration. We also found subtle differences in copulatory mechanics between very closely related species. We propose that variation in male usage of novel genitalic structures and shifts in copulatory behaviour have played an important role in the diversification of genitalia in species of the Drosophila subgroup.


Animal genitalia exemplify one of the most striking trends of morphological evolution. Male genitalic structures are complex and diversify rapidly whereas female genitalia remain relatively invariant (Eberhard, 1985). Despite this striking trend, the mechanisms underlying the rapid and divergent evolution of male genitalia remain poorly understood (Eberhard, 1985; Shapiro & Porter, 1989). Three major hypotheses have been proposed to account for the evolution of animal genitalia are the: (i) lock and key hypothesis posits that genitalia evolve through selection for pre-insemination isolation by mechanical means (Dufour, 1844), (ii) pleiotropy hypothesis maintains that evolution of genitalic structures is an indirect result of evolution of genetically related traits, i.e. through pleiotropic effects of genes that affect both genitalia and general morphology (Mayr, 1963) and (iii) sexual selection hypothesis contends that variation in male intromittent genitalic traits is related to male fertilization success and their rapid divergence is driven by sexual selection (Eberhard, 1985). The pleiotropy hypothesis is extremely difficult to test considering that we do not know all potential effects of any candidate gene(s) and their interactions. The lock and key theory has been a long-standing hypothesis but several studies have refuted its predictions (Shapiro & Porter, 1989; Arnqvist, 1997). Divergence of traits by sexual selection has become increasingly popular with growing evidence ranging from morphology (Eberhard, 1985, 2004) to sex-related molecules (see Civetta & Singh, 1999; Swanson & Vacquier, 2002). However, at both levels, deciphering how sexual selection drives the divergence of traits is under recurrent debate, particularly regarding the roles of female choice and male–female conflict (Andersson, 1994; Arnqvist, 1997, 2004; Gavrilets et al., 2001; Cordero & Eberhard, 2003; Hosken & Stockley, 2004).

Mating is an interactive process culminating in the transfer of sperm through copulation. Genitalia are primary copulatory traits and variation in male genital morphology and/or copulatory behaviour can present criteria for female choice. Traditional female choice invokes the ability of females to exercise choice based on titillatory/stimulatory faculty conveyed by male genital structures during copulation (Thornhill, 1983; Eberhard, 1996). If this is overridden by male coercion of copulation, females may be able to manipulate sperm usage (Eberhard, 1996; Miller & Pitnick, 2002; Snook & Hosken, 2004). Morphologically, this model of genitalia evolution explains the trend of divergent male and relatively invariable female genitalia, as female choice is based on existing male variation. The sexual conflict model, on the other hand, invokes a conflict of interests between the sexes that leads to a coevolutionary arms race. Coercive ability of males to control mating, often in conflict with females’ interests, can influence who sires offsprings (Arnqvist & Rowe, 1995; Alexander et al., 1996; Arnqvist et al., 2000). This sexually antagonistic coevolution model also supports the higher divergence of male genitalia but fails to explain why female genital morphology is relatively invariable (see Eberhard, 1985; Cordero & Eberhard, 2003). If females develop mechanisms/structural modifications to counter males’ coercive strategies, corresponding species-specific changes in female genitalia is expected, but this is not so (see Eberhard & Ramirez, 2004; Eberhard, 2004). Fundamental information about the function(s) of genitalic structures is lacking, leaving us to speculate how and why male genitalia evolve rapidly. Structure–function information is essential for further investigations into the mechanisms of sexual selection, allowing us to identify structures that have evolved to facilitate copulation or those which convey and/or receive mechanosensory/palpitatory stimuli in either sex.

In the Drosophila melanogaster subgroup, the posterior process of the genital arch varies dramatically in size and shape between very closely related species (Coyne & Kreitman, 1986, see Fig. 1a). The posterior process is often used to differentiate members of closely related species within the melanogaster subgroup, many of whom, are otherwise morphologically indistinguishable. A few studies have speculated that the posterior process is evolving directionally due to its role in sex and reproduction (Coyne & Kreitman, 1986; Liu et al., 1996; True et al., 1997; MacDonald & Goldstein, 1999). However, the actual function of the genital arch and most other genitalic structures in Drosophila is unknown. Robertson (1988) observed that the posterior process is inserted under the base of the ninth abdominal segment of the female during copulation. He conjectured that a correlation might exist between the size of the insertion points in the female and the size and shape of the posterior process. Therefore, mechanical obstruction may occur in interspecific copulations. However, there has been no empirical evidence of this correlation.

Figure 1.

 (a) Phylogeny of species of the melanogaster clade (from Powell, 1997), representing the extent of divergence of the posterior process. (b) External genitalia (male and female) of Drosophila melanogaster and D. simulans in resting position. Female genitalia are relatively invariable between these two species. Posterior process is highlighted in maroon and the cercus in green.

Very little is known about how genital structures are used during copulation (but see Eberhard & Ramirez 2004). Are there species-specific differences in genital coupling during copulation? If so, does divergence in genital morphology affect copulatory behaviour such as mounting strategies and copulation duration? In this study we systematically studied external genital coupling through the entire duration of copulation in Drosophila. We wanted to deduce the functional significance of external male genital structures, particularly the posterior process, in the sibling species D. melanogaster and D. simulans (diverged 2.5 MYA, Powell, 1997) (Fig. 1b). We specifically asked the following questions: (i) How do male genital structures, particularly, the posterior process function during copulation? (ii) Does mechanical isolation occur in interspecific copulations? (iii) Does variation in the shape and size of these structures affect copulation and copulatory behaviour and how? The deduced functions of these structures, for the first time, provide interesting information about their role during copulation. Species-specific differences in the mechanistic usage of male genital structures may have led to shifts in copulatory behaviours and consequently played an important role in the rapid divergence of genital structures in species of the melanogaster subgroup. We show here that the genital arch, posterior process and lateral plates are actively involved in establishing genital coupling. There are crucial differences in how genital structures are used during copulation by males from closely related species. We also find that copulation duration is male-driven regardless of their posterior process morphology or the identity of the female they mate with. We suspect that Drosophila males employ a ‘lock’-like genital coupling mechanism (not to be confused with the ‘lock and key’ theory), which determines the control of copulation duration by males. Finally, we observe a striking trend in genitalia evolution among species of the melanogaster subgroup; species of the melanogaster clade show modification and diversification of the genital arch whereas species of the yakuba clade show modifications and diversification of the lateral plates. This indicates that different sexual traits are under selection in different clades.

Materials and methods

Electron microscopic examination of male and female genitalia

To determine the resting positions of genitalic structures, adult males and females of D. melanogaster and D. simulans (roughly 8–9 days old) were flash frozen by dropping them into liquid nitrogen. Flies were then freeze-substituted in 80% ethanol at −20 °C for about 14 days. This procedure has been quite effective for studies of genital coupling (see Huber, 1993; Eberhard & Periera, 1996; Eberhard & Ramirez, 2004). Specimens were then air-dried, placed on an aluminium stub (with conductive adhesive), gold coated (3–4 nm) and examined in the Environmental Scanning Electron Microscope (ESEM). ESEM does not require gold coating but genitalia are quite complex with several fine structures. We found that gold coating significantly enhanced quality of visualization as well as the quality of micrographs.

Electron microscopic examination of genitalic coupling

Virgin females of each species were collected and retained in separate vials for 8–9 days. To obtain intraspecific mating pairs, four to six virgin females and six to eight conspecific males were introduced into 2 mL Eppendorf tubes. Both males and females were about 8–9 days old. Copulating pairs were carefully flash-frozen in liquid nitrogen at 5, 10, 15 and 20 min into copulation. For instance, once pairs initiated copulation, at 5 min, the entire tube was rapidly but carefully plunged into liquid nitrogen. Each time point was carried out separately to ensure that we did not freeze-copulating pairs at different time points in the same tube (e.g. 5 min pairs were collected first, followed by 10 min and so forth). Following flash-freezing, specimens were freeze-substituted and viewed under the ESEM as mentioned above. Similar procedures were used to obtain interspecific pairs in-copula. For conspecific pairs, a total of 10 pairs for each time point were obtained. Obtaining interspecific pairs at 5 and 17 min into copulation were extremely difficult because flies would very easily and rapidly disengage. We were only able to examine four pairs at 5 min, six pairs at 10 min, six pairs at 15 min and four pairs at 17-min stage.

Mating assays to determine copulation duration differences

Strains used: D. melanogaster– 0231.0 (Tucson, AZ, USA), D. simulans– 0251.2 (Tucson, AZ, USA), D. mauritiana (1986, St Martin, J. R. David, CNRS-Gif, France), D. sechellia (1985, Seychelle islands, J. R. David). Drosophila mauritiana and D. sechellia diverged from D. simulans about 0.5–0.6 MYA, Powell (1997). All flies were grown in standard cornmeal molasses medium at 23 °C. Five day old virgin males and females were collected for mating assays. Five females and seven males were introduced into vials for mating assays. Copulation durations were recorded until 25 complete copulations were obtained.

Genital morphology of species of the melanogaster subgroup

In addition to flies mentioned above, we used D. yakuba, D. santomea, D. teissieri and D. orena (obtained from J. R. David), D. pseudoobscura and D. virilis (Tucson, AZ, USA). We wanted to compare genitalia morphology from a phyloinformatic perspective to gain a better understanding of how male genitalia have evolved in the melanogaster subgroup.


Genital coupling in intraspecific matings

A time-sequence analysis of genital coupling provides valuable information about the functional significance of genital structures and their role in copulation. This information is crucial in understanding the underlying selective forces that may drive rapid evolution. Our results show that the genital arch, lateral plates, posterior process as well as the cercus actively participate in copulation (see Table 1 and Fig. 2 for a time-specific description of genital coupling behaviour and compare copulatory differences). In the resting position, the male's genital arch as well as the female's oviscape is retracted and concealed under the abdominal tergites in both species (Fig. 1b). In the first 5 min into copulation, the male genital arch as well as the female oviscape are extended outwards (Fig. 1b, lateral view vs. Fig. 3a). The lateral plates appear to be used like a pair of pincers to grasp the caudal end of the oviscape. At this point, it is likely that the surstyli also have a grasp on the valve and may be used to open/manipulate the valve (see Eberhard & Ramirez, 2004). Also at this point, the posterior process has grasped the oviscape ventrally (Fig. 3a). The alignment of genitalia is such that the cercus is in line to grasp the proximal end (proximal to abdominal tergites) of the oviscape (which occurs much later). In view of these observations, it appears that the posterior processes of the genital arch aids in grasping and holding the oviscape, whereas the surstyli operate on the valve (Fig. 3a). At the 5-min stage, the D. melanogaster posterior process has grasped the oviscape medially, away from the female's abdominal tergites (Fig. 3a). Strikingly, in comparison, the D. simulans posterior process is protracted much farther and graSPS the oviscape proximally and closer to the abdominal tergites. The distal end of the D. simulans posterior process is concealed under the ninth tergite of the female. These results indicate that D. simulans males may advance rather ‘quickly’ in the initial stages of copulation; for instance, they may mount females by thrusting their genital arch farther compared with D. melanogaster males. As a result, note that the cercus of D. simulans is much closer to the proximal end of the oviscape relative to D. melanagaster. These variations indicate differences in copulatory behaviours and strategies between the two species. Also, the cercus appears to open up at this stage relative to the resting position when it is tightly closed (Fig. 3a vs. Fig. 1b).

Table 1.   Summary of time-sequence genital coupling and behaviour in conspecific and interspecific matings.
Time (min)Genital coupling
D. melanogasterD. simulansmelmale vs. simfemalesimmale vs. melfemale
  1. mel, Drosophila melanogaster; sim, D. simulans.

5Posterior process: graSPS oviscape medially
Lateral plates: grasp oviscape distally
Cercus: open and aligned to grasp oviscape proximally
Posterior process: graSPS oviscape proximally (close to tergites)
Lateral plate: graSPS oviscape distally
Cercus: open and aligned to grasp oviscape proximally
Posterior process: graSPS oviscape medially
Lateral plates: grasp oviscape distally
Cercus: open but not in good alignment to grasp oviscape
Posterior process: graSPS oviscape proximally (close to tergites)
Lateral plate: graSPS oviscape distally
Cercus: open but not in good alignment to grasp oviscape
10Cercus: couples with oviscape distally
Genital coupling established
Posterior process: concealed under ninth tergite
Cercus: couples with oviscape distally
Genital coupling established
Posterior process: concealed under ninth tergite
Cercus: does not couple with oviscape
Posterior process: not yet concealed under ninth tergite
Genital coupling not established
Cercus: couples (maybe completely) with oviscape
Posterior process: largely concealed under ninth tergite
Genital coupling not established
15Genitalia tightly meshed in genital coupling
Posterior process: not visible
Genitalia tightly meshed in genital coupling
Posterior process: not visible
Cercus: still does not couple oviscape. But appears aligned to couple
Genital coupling not established
Cercus: appears coupled with oviscape
Genital coupling established
20Genitalia remain tightly meshedGenitalia remain tightly meshedCercus: has coupled with oviscape
Genital coupling established
Genitalia tightly coupled
Figure 2.

 Schematic diagram representing the time taken to establish genital coupling in Drosophila melanogaster, D. simulans and interspecific mating couples.

Figure 3.

 Genital coupling in conspecific matings. The posterior process (highlighted in maroon) is used to grasp the oviscape. Coupling of the cercus (highlighted in green) with the oviscape proximal to the abdominal tergite represents the last sequence before complete genital coupling is established. Time points are represented in the right-hand side of the figures. (a) Genital coupling at 5 min into copulation. (b) Genital coupling at 10 min into copulation. (c) Genital coupling at 15 min into copulation. (d) Genital coupling at 20 min into copulation. Panels of the left-hand side are Drosophila melanogaster matings. Panels of the right-hand side are D. simulans matings. All figures are 250–300 μM (GA, genital arch).

Observations at 10 min into copulation support the notion that the lateral plates, genital arch and posterior processes are used to grasp the oviscape. The posterior process maybe used to draw the female genitalia closer to the male genitalia (or, to enable the males to pull themselves into copulatory position). At this stage, D. melanogaster as well as D. simulans posterior processes are largely concealed under the ninth abdominal tergite of the female (as reported by Robertson, 1988; Fig. 3b). It must be noted that genital coupling from 5 to 10 min into copulation probably occurred through a sequence of movements, either due to contraction of the female oviscape or male-driven movements to establish genital coupling (or both). Figure 3c shows that the genital arch is contracted. Therefore, it is likely that between the 5- and 10-min stages, the posterior process is possibly released; edges further and graSPS a different region of the oviscape until it achieves a grasp on the proximal end of the oviscape under the ninth abdominal tergite. At this point it appears inserted under the ninth tergite, implying that there are no physical ‘points of insertion’. The cercus (previously opened up at the 5-min stage) has grasped the distal end of the oviscape at this stage, thereby establishing genital coupling (Fig. 3b).

At 15 min into copulation, in both species, male and female genitalia are pressed tightly against each other and have established a genital ‘mesh’ (Fig. 3c). This mesh can be compared with a ‘lock’-like state, in the sense that, copulating pairs at this stage do not disengage from each other easily, even if they fall off the walls of the vials. Copulating flies at this stage can even be picked up with a pair of forceps without being disturbed. This ‘lock’-like state may be a mechanism employed by males to ensure copulatory success, it would be important to study the nature of this ‘lock’ in detail (see Discussion). There are no particular differences between the 15- and 20-min stages of copulation. Male and female genitalia remain pressed tightly against each other, more so at the 20-min stage (Fig. 3d). It is important to note that at this stage, the genital arch has contracted considerably, supporting our notion that the sequence of genitalic movements between the 5- and 10-min stages may most likely be male-driven (Fig. 3a vs. Fig. 3c,d). This also means that such movements can convey palpitations that females can sense. We cannot, however, rule out the possibility of some degree of oviscape contractions by the female. Other genital structures such as the surstyli, aedagus, etc. are concealed but interesting information regarding how these may behave during this time is reported by Eberhard & Ramirez (2004).

Genitalic coupling in interspecific matings

Observations of genital coupling allow us to directly test if morphological differences in the size and shape of genitalic structures result in mechanical obstruction as predicted by the ‘lock and key’ hypothesis (Dufour, 1844). Samples at 5 and 17 min into copulation were extremely difficult to obtain, as flies would easily disengage, indicating a great degree of uncoupling.

In both interspecific crosses, at 5 min into copulation, we observed that the female oviscape is not as extended as in conspecific pairs. Males of both species grasp the oviscape with their lateral plates and posterior processes; however, the coupling appears rather awkwardly aligned compared with the conspecific pairs (Fig. 4a vs. Fig. 3a). Ten minutes into copulation, genital coupling is not achieved as in conspecific matings. However, strikingly, D. simulans males seem relatively more successful in grasping the oviscape relative to D. melanogaster males (Fig. 4b). This relative success is perhaps facilitated by the larger size of the posterior process and its extended initial thrust as mentioned above. This could be a strong indication of the effectiveness of a larger posterior process and why it may have been selected in D. simulans males. At the 15-min stage, males of both species have successfully managed to create genital coupling with the lateral plates as well as the posterior process (Fig. 4c). However, at this time, the cercus does not completely couple with the oviscape, indicating a certain degree of genital uncoupling. The cercus of both species eventually graSPS the oviscape at the 17-min stage and both interspecific pairs successfully achieve genital coupling (Fig. 4d).

Figure 4.

 Genital coupling in interspecific matings. (a) Genital coupling at 5 min into copulation. (b) Genital coupling at 10 min into copulation. (c) Genital coupling at 15 min into copulation. (d) Genital coupling at 17 min into copulation. Posterior process is highlighted in maroon and cercus in green (GA, genital arch).

Does male genital morphology influence copulatory behaviour?

Copulation duration is an important part of mating during which males must establish genital coupling and transfer sperm. It is generally assumed that males can use their genitalia to control copulation duration to secure copulatory success. There has been some evidence to shows that copulation duration is sex-specific (Coyne, 1993) but more importantly, we do not know if copulation duration is affected by male genital morphology (shape and size) in any way, i.e. if the posterior process was used to control copulation then we may expect to see males with a bigger posterior process control copulation durations more effectively (despite interspecific female rejective responses) than males with smaller posterior process. Armed with information on genital coupling and functional significance of the genital arch, posterior process and cercus, we specifically examined if the posterior process (size and shape) had any influence on copulation duration. This also allowed us to directly observe if female choice is operational during this stage of mating. Difference in copulation duration between D. melanogaster and D. simulans is small, therefore, we included two other related species, D. mauritiana and D. sechellia (diverged from D. simulans about 0.5–0.6 MYA, Powell 1997).

The results are striking and unambiguous (Fig. 5). First, posterior process morphology (size and shape) does not appear to influence copulation duration (Fig. 5), e.g. a bigger posterior process (in D. simulans) is not implemented to achieve longer (or shorter) copulation duration. Drosophila sechellia with a smaller posterior process has the longest average copulation duration (Fig. 5). Male-specific copulation duration is maintained, regardless of the size and shape of the posterior process and regardless of female's species identity. Copulation duration, therefore, is largely a species- and male-specific trait dictated by other factors. Secondly, if female behaviour (choice or conflict) had any effect we would expect, for instance, female D. mauritiana to terminate copulation or exhibit rejection response (e.g. struggle to dislodge males) after approximately 9–11 min. This is not so, indicating a remarkable compliance by females. Female choice based on genital morphology during copulation is therefore questionable. It must be noted, however, that a small but consistent reduction in copulation duration in interspecific crosses was seen. This is rather intriguing considering that, if males were unable to establish genital coupling efficiently and if males determined copulation duration, we would expect successful copulation to take longer. Why is it shorter? This reduction may reflect a female effect (average reduction: melanogaster = 13%, simulans = 14%, mauritiana = 19%, sechellia = 12%, Mann–Whitney sum rank test, P < 0.05 for all). Improper genital coupling in interspecific copulations leading to an inefficient lock is also a plausible cause for this reduction (see above). In any case, this reduction in copulation duration in interspecific pairs requires further investigation. It must be noted that considerable variation in copulation duration between strains have been observed (Coyne, 1993). However, results of copulation duration presented here largely reflect the general trend observed in other studies (Robertson, 1983; Coyne, 1993). Copulation durations were previously suggested to be male-dependent; for instance, when D. mauritiana males were crossed with D. simulans females, the heterospecific matings were most often short, following the D. mauritiana copulation duration (see Coyne, 1993). On the other hand, D. simulans males crossed with D. mauritiana females had generally longer copulation durations, following the average >20 min trend (Coyne, 1993).

Figure 5.

 Average copulation durations in within- and between-species matings (n = 25 copulating pairs for all crosses). Neither species-specific morphology of the male posterior process nor female choice influences copulation duration. Copulation duration is therefore a male-and species-specific trait dictated by other factors. Camera drawings of the posterior process are to scale.


Functional differences between D. melanogaster and D. simulans copulatory behaviour

Prior to mounting, males arch their abdomen ventrally (they often pursue females in this position if females are moving around). Males then make several attempts with swift inward thrusts of their abdomen to grasp female's genitalia with their own (see Spieth 1952). Although males may use their forelegs to assume primary mounting posture (Coyne, 1993), they also rely on genital structures particularly the posterior processes or surstylli to grasp the female's genitalia during mounting. An elaborate posterior process, therefore, is a valuable accessory of the genital arch that males can use to facilitate mounting. Both D. melanogaster and D. simulans males use their posterior processes to grasp the oviscape. However, D. simulans males seem to thrust their genital arch and posterior process farther than D. melanogaster males in the first few minutes into copulation (Table 1, Fig. 3a,b). First, this is indicative of a difference in mounting effort and mechanism. Whether this difference is due to species-specific male mounting posture or female's alignment must be further investigated. Secondly and consequently, males of the two species grasp the female oviscape differently and at different regions in the early stages of copulation. This also means that in the subsequent 5- to 10-min stages, there is a difference in how genital coupling proceeds in the two species. Whether this induces differential responses from females must be tested. From the males’ perspective though, the larger and broader posterior process in concert with the extended thrust in D. simulans males may offer an advantage to mount and achieve genital coupling more rapidly relative to D. melanogaster. In addition, the larger shape and size of the posterior process in D. simulans males may be advantageous in creating a more efficient genital coupling. This notion is supported by the observation that D. simulans fare relatively better in interspecific copulations. These observations, however, must be further corroborated by intraspecific investigations to determine if indeed a bigger posterior process has an adaptive advantage in copulation. Nevertheless, it is apparent that the importance of the posterior process lies in the first 5–10 min of copulation when males must achieve strong genital coupling.

Failure to conform to the predictions of the ‘lock and key’

The lock and key theory posits that mechanical obstruction would occur due to morphological differences in male and female genitalia in interspecific copulations. First, it is important to note that female genital morphology in these two species is practically invariable (Fig. 1b), providing no basis for mechanical obstruction. Secondly, Robertson (1988) postulated that the mechanical obstruction possibly lay in the relationship between the size of the posterior process of males and the insertion points between the female's eighth and ninth abdominal tergites. Therefore, mechanical obstruction is expected in the case of D. simulans male vs. D. melanogaster female matings because of incompatibility in the size and/or shape of the posterior process, thereby leading to termination of copulation. The reverse cross (D. melanogaster male vs. D. simulans female) should enable pairs to create a better ‘fit’ due to the relatively smaller D. melanogaster posterior process. Our results show that both interspecific couples succeed in creating proper genital coupling (albeit with a delay), showing no indications of mechanical obstruction due to the size and shape of the posterior process. Therefore, the traditional predictions of the ‘lock and key’ do not hold (also see Mayr, 1963). There is, however, a certain degree of delay in genital coupling in interspecific crosses. There is incomplete coupling of the cercus with the oviscape in interspecific crosses within the 10- to 15-min stage relative to conspecific crosses (Table 1). A more detailed investigation of this issue is necessary to identify if the difference is due to male mounting mechanisms or female's responsive alignment prior to or during copulation.

Male–female conflict and female choice

Was the rapid diversification of the posterior process driven by female choice or the increased coercive ability of males? First, contrary to the predictions of male–female conflict, our study found no species-specific modifications in female genitalia that may have evolved to impede the males’ grasp or enable the female to gain control over copulation (see also Eberhard & Ramirez, 2004). This shows that there is no counter-evolutionary arms race in the divergence of external genital morphology in Drosophila, contrary to what has been found in some sexual structures of other insects (Arnqvist & Rowe, 2002). Drosophila females may use other means to counter male coercive manipulations such as wing fluttering or kicking (see Spieth 1952). However, our results show that such presumably rejective responses are not dramatic during copulation. Perhaps, then, the arms race in Drosophila may be predominantly post-copulatory (Miller & Pitnick, 2002; Garcia-Gonzalez & Simmons, 2005). Secondly, the design and function of the posterior process offers males a significant advantage to secure mounting and maintain copulation. Do they use this to control copulation? We have shown that morphologically, the posterior process does not influence copulation duration, i.e. a bigger posterior process is not used to prolong or shorten copulation duration presumably to suit the interests of the male (see below and Fig. 5). This implies at first that copulation duration is determined by other factors, which are evidently male-driven. Secondly, Drosophila males do not use their genitalia to control copulation, but mainly to expand their own chances of securing mounting, genital coupling and successful copulation. These efforts may not necessarily be in conflict to female's interest.

Alternatively, according to the theory of cryptic female choice, male traits that enhance copulatory success may be criteria for female choice. Mounting and remaining in copulatory position may be challenging for Drosophila males. Therefore, any modification that might enable males to achieve and maintain copulatory position will be advantageous. In the field, most males (of most Drosophila species) vigorously attempt to copulate with females. However, they are largely unsuccessful if females are not receptive (Spieth 1952). This indicates that female choice is strongly effective during courtship, when mate choice is relevant. More recently, Bertin & Fairbairn (2005) have shown that precopulatory choice acts strongly on genital morphology in water striders. In our study, anatomically, the first few minutes into copulation appears to be crucial in establishing genital coupling. Genital coupling is not complete until 10 min into copulation. Considering the fact that sperm transfer takes 5–7 min in Drosophila (Gilchrist & Partridge, 2000), it would be in the interests of the females to exercise choice during this time to reject males. If females accept males during this stage, the ‘lock’-like state that follows, may be inspected by females to ascertain the strength of the lock, which may be indicative of copulatory fitness (see also Eberhard & Ramirez, 2004). Strong genital coupling may be necessary to ensure efficient transfer of sperm. These possibilities must be tested. Nevertheless, our data show that female choice may not be manifested during copulation and may be restricted to precopulatory and post-copulatory stages. It must be noted, however, that we are still unaware of other crucial factors that may come into play during and between the stages of copulation represented in our study. We also do not know if the posterior process conveys mechanosensory and/or palpitatory stimuli, which would present a compelling case for operative female choice (Cordoba-Aguilar, 1999).

Male sex-drive, an alternative to female choice and sexual conflict

Genitalia diversification has predominantly been explained by female choice or sexual conflict. Darwin (1871) had recognized that males play a predominantly active role in soliciting and initiating mating. Despite this observation, there has not been much emphasis on investigating the consequences of male's effort in mating, copulation and more importantly, in influencing female choice or conflict. For instance, the evolutionary fate of a novel trait (such as the posterior process, see below) can be greatly influenced by how males use this trait and also if females have a preference for this trait or evolve resistance to it. Our study brings to the forefront, the importance and relevance of males’ role in the mating ritual, particularly during copulation. We discuss below, the relevance of an important and new theory, male sex-drive (Singh & Kulathinal 2005), which provides an alternative to the choice and conflict theories, in explaining how genital structures can diversify rapidly.

Male sex-drive and control of copulation duration: do females have a choice?

Our results indicate that males actively use genital structures and develop mechanisms to facilitate mounting and copulation in an attempt to secure copulatory success. As a consequence of the relevant mechanisms, in this case the ‘lock-like’ mechanism, they determine copulation duration. The male sex-drive theory maintains that this active male dominance during courtship and copulation can be a major and important driving force in the evolution and diversification of sexual traits both morphology and molecular (see Singh & Kulathinal 2005). Male sex-drive is essentially the manifestations of male efforts (structural, physiological, behavioural) in courtship and mating (as Darwin, 1871 had recognized earlier). With respect to copulation, it is the males that must attempt and succeed in establishing genital coupling and transfer sperm. It is not unreasonable therefore to assume that copulatory behaviour is male-driven. How does male sex-drive explain the predominant male influence in determining copulation duration? We think that Drosophila have developed a ‘lock’-like genital coupling mechanism to facilitate copulation. This, however, is not to be confused with the traditional interpretations of the ‘lock and key’ theory that invokes a species-specific ‘lock’ and ‘key’ that is structurally designed (in both males and females) to create mechanical obstruction. To explain male control of copulation we propose two plausible reasons: (i) females cannot dislodge males due to a physical ‘lock’ employed by males, as a consequence of genital coupling, or (ii) female choice is not manifested physically, during copulation. We favour the first explanation and attribute male control of copulation duration to the employment of a ‘lock’-like mechanism by males to secure copulation, which, females are unable to break. Therefore, the control of copulation duration may be a consequence of the nature of genital coupling. Copulation duration per se may depend on the physiological nature of the sperm-transfer machinery in males (Gilchrist & Partridge, 2000) or efforts of males to physically remove stored sperm (Cordoba-Aguilar, 1999). From the perspective of females, attempts to break the male ‘lock’ during copulation may prove to be harmful. As an outcome of the preceding scenario, females may have developed mechanisms to manifest their choice at the precopulatory and post-copulatory stages (Cordoba-Aguilar, 1999; Snook & Hosken, 2004) and thereby avoid costly physical consequences of disrupting copulation. This is not to say, however, that females remain passive during copulation. Sensory stimuli received during copulation may convey signals of individual fitness and can be utilized by females (Eberhard, 1996).

Male sex-drive and the evolution of novel genitalic structures

The posterior process (highlighted in maroon, Fig. 6) is essentially a modification of the genital arch. Strikingly, we found that only species of the melanogaster clade, i.e. D. melanogaster, D. simulans, D. mauritiana and D. sechellia, have elaborate posterior processes (see Fig. 6). In comparison, species of the yakuba clade (D. yakuba, D. santomea, D. erecta, D. teisieri and D. orena) do not have posterior processes. In fact, none of these species have major modifications of the genital arch, but instead, most of them have modifications of the lateral plate (indicated by blue, Fig. 6). Drosophila yakuba and D. santomea have small processes near the apical end of the lateral plate (Fig. 6). Drosophila teissieri has a hook-like modification of the lateral plate. Drosophila orena and D. erecta do not have any such modifications. Comparing species outside of the melanogaster group, D. pseudoobscura has elaborate clasper-like modifications of the lateral plate. Drosophila virilis also has plate-like modifications but is difficult to ascertain if they are modifications of the genital arch or lateral plate. Remarkably, the female genitalia of all these species are relatively invariable and less complex (Fig. 6).

Figure 6.

 External male genitalia of species of the melanogaster subgroup. Male genitalia (left-hand panel) are complex and diverse. Female genitalia are less complex and relatively invariable (right-hand panel). The posterior processes (highlighted in maroon) are essentially modifications of the genital arch and are recently derived structures, present only in species of the melanogaster clade (Drosophila melanogaster, D. simulans, D. mauritiana and D. sechellia; modifications highlighted in maroon). In comparison, species of the yakuba clade (D. yakuba, D. santomea, D. teisierii, D. erecta and D. orena) do not have posterior processes but have modifications of the lateral plates (modifications highlighted in blue). Interestingly, the cercus (highlighted in green) exhibits less divergence in morphology relative to modifications of the lateral plate and genital arch. The phylogeny (left) represents lineages that show modifications of the genital arch (thick maroon lines) and lateral plates (thick blue lines). Phylogeny and divergence times taken from Powell (1997).

Two striking trends emerge from this comparison: (i) the posterior process is a recently evolved novel character in the melanogaster clade and has diversified extensively in the past 2.5 MYA. It would appear that either males had a novel trait to use to their advantage to enhance their mating success, and/or, this new structure was also exposed to female choice or conflict. In the light of our results, the posterior process offered copulatory advantage for males of the melanogaster clade to enhance their copulatory success and therefore is subject to selection. (ii) In comparison, species without a posterior process must rely largely on their lateral plates and surstylli to grasp females, therefore, structural modifications of the lateral plates that may have offered copulatory advantage would be under selection. It is also interesting to note that the cercus (highlighted in green, Fig. 6), which also actively participates in establishing genital coupling, does not differ considerably between the species (Fig. 6). This highlights the interplay between the functional importance of genital structures and the selective pressures on these structures.

The underlying evolutionary pathways that lead to the divergence of morphological traits are invariably dependent on the diversity of mechanisms at play. If a ‘lock’-like system is universal in Drosophila or insects, variation in the usage of genital structures that facilitate this mechanism along with plasticity of those structures can rapidly influence its divergence. On the other hand, divergence in the mechanisms itself (mating/copulatory behaviour, strategies) can influence how structures are used and how they diversify. More importantly, we agree with Hosken & Stockley (2004) that mechanisms such as female choice and conflict, may not be mutually exclusive and may be operational at different stages of the mating ritual. Genitalia in insects are largely concealed and come into contact during copulation, it is important therefore to study how males use their genitalia during this stage and if (and how) females respond to it. From our results, it is likely that in species of the melanogaster clade (all of which have posterior processes), morphological diversification of genital structures can largely be attributed to differential copulatory abilities/mechanisms in mounting as well as establishing genital coupling during the initial stages of copulation (as our results have shown between D. melanogaster and D. simulans).


Predictions of both male sex-drive and cryptic female choice can adequately explain why only male genitalia are complex and diversify rapidly. A corresponding change in female genital morphology is not required as females are not necessarily harmed by structural variations that confer copulatory advantage to males. In addition, strong female choice during courtship reduces the likelihood of coercive mating, which is a costly strategy. An intricate interplay between male sex-drive and female choice is possible; variations in male genitalia that may or may not offer males copulatory advantage provide a substrate for cryptic female choice. Genitalia diversification has predominantly been explained by female choice or conflict. In the light of our results, we believe that choice and conflict may be predominantly operational during precopulatory and post-copulatory stages. However, with regard to the posterior process and its role during copulation, we find that female choice is not physically operative. This is either because females do not exercise choice during copulation, as they have already accepted males through precopulatory screening. Or, it could be because males, with their ‘lock’-like genital coupling render female choice ineffective. These are indications that variations in male efforts and strategies have a strong influence in the evolution of genitalic structures. The trend of diversification of genital structures in species of the Drosophila subgroup shows that different genitalic structures are under selection in different lineages/clades, indicating possible shifts in male genital usage and copulatory behaviour. It is important in future investigations to characterize the functions of genitalic structures and possibly differentiate between structures that confer copulatory advantage to males and those that are subject to female choice.


The authors would like to express their gratitude to W. G. Eberhard for offering extremely valuable guidance and help in experimental procedures as well as discussions during the course of this study. Also would like to thank A. Civetta and J. R. David for useful comments on the manuscript. Also thank J. N. A. Lott and K. Schultz of the Electron Microscope facility (Department of Biology) for their supervision, guidance and technical assistance. This study was supported by Natural Sciences and Engineering Council grant to R. S. S.