Speciation depends on the establishment of reproductive isolation between populations of the same species. Many mechanisms for the evolution of such isolation have been proposed (Otte and Endler 1989), which for animals can be divided into three broad categories. Isolation evolves as an incidental effect of: (1) gradual accumulation of incompatibility by genetic drift in allopatric populations; (2) adaptive divergence in response to environmental variation or sexual selection; and (3) rapid genetic changes associated with founder events, including both extreme drift and changes in pattern of selection due to population bottlenecks (Tregenza et al. 2000).

Classic models of speciation (Dobzhansky 1937; Mayr 1942) recognized that reproductive isolation, and subsequent speciation, could be generated by differences in sexual traits (including behaviors). Divergence for sexual traits between allopatric populations was considered to result either from drift, pleiotropy, or adaptation to environmental conditions, or following secondary contact, because individuals benefited by avoiding heterospecific matings (i.e., by reinforcement). However, it became clear that changes between populations in sexual traits could also result from sexual selection and this might represent a distinct process of speciation (West-Eberhard 1983; Coyne 1992; Wu 2001; Coyne and Orr 2004). Sexual selection has the potential to lead to rapid divergence between populations and possibly to generate reproductive isolation because it may have a direct effect on traits involved in mate recognition. It is important to point out that the rapid change between populations as a result of sexual selection can also play an indirect role in speciation by increasing the overall rate of change within isolated populations (Alahiotis 1976; Kilias et al. 1980; Ligon 1999). Sexual isolation results from different mating success among individuals within a population. Competition for fertilization occurs through direct competition between members of the same sex (e.g., male-male competition and sperm competition) or through the attraction of one sex to the other (e.g., female choice). Although long recognized as important in intrapopulation evolution, sexual isolation has more recently been invoked as a driving force behind speciation (Panhuis et al. 2001).

To identify the causes of speciation, we need to find which factors act before speciation to bring it about, rather than confusing them with processes that may have happened afterward (Etges 2002a,b). Comparisons within species have the advantage that observed divergence has occurred prior to speciation and thus may subsequently contribute to the process of speciation itself. Sexual selection changes gene frequencies in populations and produces microevolution, whereas sexual isolation might be directly involved in speciation (Lewontin et al. 1968). In the laboratory, sexual selection effects can originate via intrasexual competition (mate propensity or mate fight) or intersexual choice (mate choice), whereas sexual isolation effects are mainly caused by intersexual choice. In the wild, many different biological mechanisms may contribute to these two mating components (Rolán-Alvarez et al. 1999).

The partial reproductive isolation of populations from the same species is thought to be an initial step towards speciation (Etges 2002a,b). Experiments using laboratory strains have provided evidence of this, even though the development of this isolation in natural conditions is difficult to observe (Rice and Hostert 1993). A review of many studies on Drosophila species indicates that prezygotic isolation is likely to evolve faster than postzygotic isolation between sympatric than between allopatric pairs of species (Coyne and Orr 1997). Hence, analysis of speciation should include a systematic study of traits involved in prezygotic isolation (Tregenza and Bridle 1997), including coevolutionary phenomena (Cordero and Eberhard 2003; Kokko et al. 2003; Pizari and Snook 2003) at least in complex animals such as Drosophila. Systematic studies on sexual isolation have been done in our laboratory using a unique experimental system of D. melanogaster laboratory populations established since 1973 and additional ones since 1985. A high sexual isolation index (SII) has been observed between different combinations of the aforementioned populations, SII about 0.34 (Kilias et al. 1980). Previous genetic studies showed that both cytoplasm and the nuclear genome influences the level of mate discrimination (Kilias and Alahiotis 1982), and that multiple genetic factors for discrimination are distributed over all the chromosomes (i.e., the trait is polygenic).

Bacteria in the genus Wolbachia are cytoplasmically inherited rickettsiae that are found in a wide range of arthropods as well as in filarial nematodes (reviewed in Werren 1997; Stouthamer et al. 1999; Stevens et al. 2001; Bourtzis and Miller 2003). Several surveys have found these bacteria in over 16% of insect species, including each of the major insect orders (Werren et al. 1995a,b; Jeyaprakash and Hoy 2000). These bacteria cause a number of reproductive alterations in their hosts, including cytoplasmic incompatibility (CI) (review in Bourtzis et al. 2003), parthenogenesis induction (PI) (reviewed in Bourtzis and Miller 2003; Koivisto and Braig 2003), male killing (reviewed in Hurst et al. 2003), and feminization of genetic males (Rigaud et al. 1997). These modifications of host reproduction impart a selective advantage for the bacteria (Werren 1997).

Cytoplasmic incompatibility (CI) is the most widespread and, perhaps, the most prominent feature that Wolbachia endosymbionts impose on their hosts (Bourtzis et al. 2003). Cytoplasmic incompatibility is a sperm-egg incompatibility that typically results in embryonic death due to disruption of male-derived chromosomes in the first mitosis. It can be either unidirectional or bidirectional. Unidirectional CI is typically expressed when an infected male is crossed with an uninfected female. The reciprocal cross is fully compatible, as are crosses between infected individuals. Bidirectional CI usually occurs in crosses between infected individuals harboring different strains of Wolbachia. Levels of incompatibility can vary with the type of Wolbachia infection, genetic background of the host, age of the males, or environmental conditions. For example, partial unidirectional CI usually occurs in Drosophila melanogaster, whereas CI is nearly 100 percent in the parasitoid wasp Nasonia vitripennis.

Cytoplasmic-incompatible Wolbachia have been implicated as a possible mechanism promoting speciation, particularly in arthropods where these bacteria are common and widespread (Laven 1959, Werren 1998, Bordenstein et al. 2003). The basic idea is that, by preventing or severely reducing gene flow between populations (or sibling species), CI Wolbachia will increase the probability of genetic divergence and speciation. In particular, CI induced by Wolbachia may be important in preventing gene flow between incipient species in two ways. First, bidirectional CI may directly prevent introgression between populations or species harboring different strains of Wolbachia (Werren 1998). Second, Wolbachia-induced CI may contribute to reproductive isolation by serving as one of several reproductive isolating mechanisms, which together prevent or greatly restrict interspecific gene flow (Werren 1998).

Indirect evidence for a possible role of Wolbachia in speciation includes studies showing bidirectional CI (bi-CI) (Bordenstein et al. 2001) and unidirectional CI (uni-CI) (Shoemaker et al. 2000) contributing to reproductive incompatibility between closely related species. Theoretical studies show that both uni- and bi-CI can promote divergence between populations (Telschow et al. 2002a,b) and probably reinforce premating isolation (Telschow et al. 2005a,b) in the face of substantial gene flow. However, the view that CI Wolbachia promotes speciation is controversial (Weeks et al. 2002; Coyne and Orr 2004). Among the arguments against is the view that bi-CI is expected to be uncommon between populations, that bi-CI is unlikely to be stably maintained (but see Keeling et al. 2003; Telschow et al. 2005b), and that CI can be incomplete and therefore may not be an effective barrier to gene flow. Empirical and theoretical studies are needed to resolve these issues.

Wolbachia could also contribute to speciation by other means (Werren 1998; Bordenstein 2003), such as accelerating rates of evolution of genes involved in gametogenesis (due to residence of these bacteria in gonads). Relatively unconsidered is the possibility Wolbachia can influence levels of mate discrimination in diverging populations, not indirectly by selecting for reinforcement but by contributing to cues involved in mate choice. In the present study, we show that antibiotic removal of Wolbachia decreases mate discrimination by approximately 50% in long-term experimental populations that have been subjected to different selective pressures for food source (metals versus ethanol) and that have diverged in mate discrimination. The effect of antibiotic curing persists for over 75 generations following treatment, is not observed in crosses between populations uninfected with Wolbachia, and is not due to genetic changes following antibiotic treatment. Results implicate Wolbachia (or an associated undetected bacterium) as a factor in mate discrimination in these experimental populations. The results suggest that Wolbachia may also play a role in premating isolation between diverging populations and incipient species.

Materials and Methods

Measuring Sexual Isolation

The sexual isolation index (SII) was measured as in previous studies using multiple choice experiments (Kilias et al. 1980) to allow the comparison of results. Thus, a total of 24 pairs (12 virgin females and males from one population and 12 virgin females and males from a second one) were used for every set of random mating experiments. This is a multiple mate choice design in which, in each cross, an equal number of males and females from each strain was present. The number of homogametic and heterogametic matings was recorded for 90 min by direct observation in mating chambers described by Elens and Wattiaux (1964). All flies were three- to four-day-old virgins. Flies from one population had the tip of one wing clipped to allow identification. The marking was altered between stocks. The clipping was performed on lightly anaesthetized flies using CO₂ gas, at the time they were “sexed” after emergence. It must be noted that the clipping does not affect mating propensity or discrimination (Petit et al. 1976; Kilias et al. 1980; Dodd and Powell 1985). Random mating was tested by chi-squared and the joint isolation index of Malogolowkin-Cohen (Malogolowkin-Cohen et al. 1965) was calculated as follows:

where: X_AA, X_BB, X_AB, X_BA stand for the four types of mating, ♀A × ♂A, ♀B × ♂B, ♀A × ♂B and ♀B × ♂A respectively.

N = X_AA + X_BB + X_AB + X_BA.

The standard error of the sexual isolation index is:

The value of zero for the sexual isolation index indicates random mating; a value < 0, negative assortative and > 0, positive assortative mating. The experiments were performed at 25°C, 43 ± 4% RH, and 55 Lux, always at the same time of day (09.00–13.00). In all experiments the egg samples were allowed to develop at 25°C in a cornmeal-sugar-agar medium free of metals or ethanol.

The presence of assortative matings is tested using chi-squared, where the total number of the four types of matings (sum of replicates in each combination) were used. The presence of statistically significant differences in the SII for each combination before and after antibiotic treatment was tested using z-stat (Watanabe and Kawanishi 1979). Two-way ANOVA was used to examine if all combinations had the same response before and after treatment.

Results

The origin of the cage populations tested and the conditions they are maintained are shown in Figure 1. These long-term selection populations were originally established to investigate the genetics of Drosophila melanogaster sexual isolation (Alahiotis 1976). Those initial populations exhibited significant sexual isolation (Kilias et al. 1980) with cytoplasmic contribution as well as nuclear-cytoplasmic interactions affecting this character (Kilias and Alahiotis 1982). In previous studies, we have shown substantial levels of mate discrimination between the original M and E populations in our long-term selection experiments SII ≈ 0.387 (K. Koukou, G. Kilias, and S. Alahiotis, unpubl. data). We have subsequently found that the populations also differed in Wolbachia infection status (Fig. 1). By screening at least fifty individuals per population, we have shown that all individuals of cages CM2, DE1, and DM1 are Wolbachia infected (these strains are subsequently designated with a W superscript), and all individuals of populations CM1, CE2, and DE2 are uninfected (these strains are subsequently designated with a U superscript). Sequencing of the Wolbachia surface protein from DE1^W, CM2^W, CE1^W, DM1^W, and DM2^W lines show them to be the same as that previously reported for D. melanogaster. Because Wolbachia are cytoplasmically inherited, and we had previously detected cytoplasmic effects on levels of mate discrimination among these lines (Kilias and Alahiotis 1982), the investigation of whether Wolbachia infection status influences the level of assortative mating between these populations is of particular interest.

A subset of populations was antibiotically treated to cure them of Wolbachia. In addition, uninfected populations were also antibiotically treated to serve as controls for effect of treatment. Effectiveness of curing was then confirmed by polymerase chain reaction using Wolbachia specific primers. Six generations following antibiotic treatment, we then tested treated and untreated populations for levels of assortative mating using our standard behavioral assay (Elens and Wattiaux 1964; Kilias et al. 1980). Crosses were performed between four combinations of populations involving one infected and one uninfected population, and also between their corresponding treated populations. The population pairs that were chosen for experimentation followed a certain “philosophy”: they are all C × D populations (not sister populations) with one of the two reared in a food supplement of metals and the other in a food supplement of ethanol. This was done for the reason that these types of combinations presented the highest SII index in earlier studies that were carried out, a prerequisite necessary for testing the effect of Wolbachia on SII.

As can be seen in Table 1 and Figure 2, there is a significant reduction in mate discrimination in the antibiotic treated pairs (two-way ANOVA, F_{untreated vs treated} = 58.677, P < 0.001, df = 1,70) with 45.6% of the variance in crosses being explained by treatment. Antibiotics resulted in approximately a 50% reduction in mate discrimination six generations following the treatment. Considering each pair individually, the SII declined from 0.40 ± 0.06 in combination CM1^U × DE1^W to 0.21 ± 0.05 in CM1^T × DE1^T (z = 2.292, P < 0.05). Similar significant changes were observed between untreated and treated populations of CM2^W × DE2^U (0.35 ± 0.06 to 0.14 ± 0.06; z = 2.5, P < 0.05) and CE2^U × DM1^W (0.38 ± 0.06 to 0.18 ± 0.06; z = 2.39, P < 0.05). An analogous ∼50% reduction was found for CE2^U × DE1^W (0.27 ± 0.06 to 0.15 ± 0.06; z = 1.49, P > 0.05), although this reduction is not statistically significant. In the latter case, the two populations shared the same ethanol selection regime. It must be noted that all treated populations still exhibit statistical significant assortative mating. No fundamental changes were observed in sexual isolation when the I_PSI index (Rolan-Alvarez and Caballero 2000; Perez-Figuerora et al. 2005) was used (e.g., for the combination CM1^U × DE1^W SII = 0.40, I_PSI = 0.405). Control crosses within populations give the expected low levels of SII indicative of random mating (Table 2). Chi-square tests showed no significant heterogeneity among replicates for any of the crosses. It has also to be noted that in these experiments, reciprocal matings between the strains occurred at similar frequencies.

To examine the reduction of the SII index as well as to rule out transient effects of antibiotics, we retested a subset of strains 45 to 75 generations after the tetracycline treatment. Figure 3 and Table 1 show these results. As can be seen, there is still a significant effect of antibiotic treatment on the level of mate discrimination when one or both lines are infected with Wolbachia (two-way ANOVA, F_{untreated vs treated} = 158.593, P < 0.001, df = 1,52), with 75.3% of variance explained by treatment. Considering the individual crosses, significant reductions due to treatment were again observed in CM1^U × DE1^W (SII = 0.38–SII = 0.23). We also performed crosses between the untreated CM1^U population and the treated DE1^T population that had been cured of its Wolbachia (not shown in Fig. 3, data in Table 1). These also showed a significant reduction in mate discrimination from 0.38 to 0.17 (z = 2.036, P < 0.05), indicating that treatment of the uninfected line was not a major contributor to changes in mate discrimination level but treatment of the Wolbachia infected line was.

When both strains were infected with Wolbachia (CM2^W × DE1^W and CM2^W × DM1^W), a similar in magnitude but nonsignificant reduction in SII was observed (from 0.38 to 0.21 and 0.24 to 0.13, respectively). In each cross Wolbachia again resulted in about a 50% reduction in the level of assortative mating, and when the two crosses are analyzed together a significant effect of treatment on SII is observed (two-way ANOVA, F_{untreated vs treated} = 106.930, P < 0.05, df = 1,38).

In addition, we conducted two crosses between populations that showed significant divergence in mate discrimination but that were not infected with Wolbachia (Table 1 and Fig. 3). In contrast to the previous results, there is no change in mate discrimination following antibiotic treatment when the diverged lines are uninfected with Wolbachia (Table 1 and Fig. 3). This occurs despite the fact that the two sets of tested lines show significant levels of mate discrimination. The data suggests that changes in mate discrimination are the result of Wolbachia curing, and not the consequence of changes in other bacterial associates present within the cultures (e.g., intracellular or surface bacteria). To strengthen this view we also screened infected populations with primers specific to three other intracellular bacterial groups (Cytophaga Group, Spiroplasmas, Rickettsias) known to be common in arthropods (Majerus et al. 1999; von der Schulenburg et al. 2000; Williams et al. 2001; Weeks et al. 2003); none was detected. Furthermore, to test whether some other bacterium was associated with the effect, we then used general prokaryotic 16S primers to amplify bacterial sequences from whole insects of the Wolbachia infected populations, cloned the product, and sequenced 24 clones per population. All clones yielded only Wolbachia 16S sequence, indicating once more that this is the most abundant bacterium within or on insects in these populations, and further implicating Wolbachia as the agent responsible for changes in SII.

Antibiotic treatment had no effect on levels of mate discrimination in crosses within populations, as was expected. Such within population controls exhibited nonsignificant isolation indices of −0.03 ± 0.07 to 0.07 ± 0.10 both before and following antibiotic treatments (Table 2). This finding clearly indicates that elimination of Wolbachia does not cause alterations in mate discrimination within these experimental populations, although it does alter level of mate discrimination between populations that have already diverged in mate choice. Therefore, antibiotic treatments are not simply causing alterations in some undetected bacteria that is associated with fruitflies and involved in mate discrimination. Instead, alteration of mate discrimination is correlated with elimination of Wolbachia from lines that have already diverged in mate preference.

However, it is possible that antibiotic treatments resulted in nuclear genetic changes in the populations, for example, due to selection or sampling (drift). To test for this possibility, we created lines of the same nuclear genetic background by six generations of backcrossing. For this experiment, 100 males of each cured population were crossed with 100 females from the infected original population and 100 females from the respective treated population in each generation, to produce a matching cured and uncured population with a common genetic background. Backcrosses were performed for DE1, DE2, DM1, CE2, CM1, and CM2, and each resulting population is indicated by the additional superscript B. Again combinations using backcross populations show reduction in SII (Fig. 4). There is a significant reduction in SII associated with antibiotic treatment in both the original populations (ANOVA, F_{untreated vs treated} = 58.677, P < 0.001, df = 1,70, 45.6% of variance explained by treatment) and in the backcross populations (ANOVA, F_{untreated vs treated} = 85.722, P < 0.001, df = 1,74, 53.7% of variance explained by treatment). For example the combination CM1^UB × DE1^WB showed a reduction from 0.36 ± 0.06 to 0.20 ± 0.06 (z = 2.005, P < 0.05) (see Fig. 4). Similar significant changes were observed between untreated and treated populations of CM2^WB × DE2^UB (0.37 ± 0.05 to 0.20 ± 0.06; z = 2.089, P < 0.05) and CE2^UB × DM1^WB (0.35 ± 0.05 to 0.20 ± 0.06; z = 2.006, P < 0.05). These results show that the reduction in mate discrimination in crosses involving antibiotic treated Wolbachia infected lines is not due to nuclear genetic changes resulting from the treatment. Antibiotic curing of the associated Wolbachia is implicated as the cause of changes in mate discrimination in these populations.

Discussion

The unique system of our long-term cage populations, which developed strong sexual isolation about five years after their origin and have maintained it since, represents a good experimental tool for studies on reproductive isolation mechanisms (Kilias et al. 1980). Here we find that elimination of associated Wolbachia from the experimental populations that show strong sexual isolation reduces the level of isolation by around 50%. The finding that antibiotic removal of Wolbachia reduces levels of premating isolation is robust, is well replicated, and has been demonstrated in multiple crossing combinations and over an extended period of time (up to 75 generations post-treatment). It has also been shown that the effect is not due to nuclear genetic changes following antibiotic treatment but rather to a heritable component of the cytoplasm in Wolbachia infected populations. Sequencing part of the wsp gene of the associated Wolbachia strains suggested that all strains carry identical gene sequence to one another as well as to that of the wMel strain. However, it is still possible that there may be differences between the Wolbachia strains infecting the divergent D. melanogaster lines which could not be revealed with the sequencing of the wsp gene (see recent study by Riegler et al. 2005). If the Wolbachia strains have diverged at unsequenced loci then it is possible that the mating discrimination phenomena could be influenced by divergence in Wolbachia strains.

Our results indicate that, in long-term selection populations showing assortative mating, presence of Wolbachia contributes to the level of premating isolation between the populations. We can speculate on the causes of this. One possibility is that Wolbachia alter the pheromonal profiles of males and/or females, thus affecting mate preference. For example, cuticular hydrocarbons are known to affect mate preferences in Drosophila species (Coyne and Oyama 1995; Ishii et al. 2001). Bacteria are known to affect pheromomal production in some insects, such as gut bacteria in the locust (Dillon et al. 2000, 2002; Dillon and Charnley 2002) that produce the aggregation pheromone. A second possibility is that Wolbachia alter behavior in subtle ways that contribute to assortative mating. It is our view that the role of Wolbachia in premating isolation is likely to be a side-effect of the infection, rather than due to direct selection for the bacteria to induce mate discrimination. In other words, the effects of Wolbachia on mate cues were utilized during the nuclear genetic evolution of mate discrimination in these populations. Additional studies are needed to determine how the presence of Wolbachia alters mate preference in these populations. It is also apparent that removal of Wolbachia in crosses between populations not showing assortative mating does not induce it. Therefore, the effects of Wolbachia on pheromones or behavior must augment preexisting genetic biases in mate discrimination rather than being sufficient by themselves.

It could be argued that the long-term effects of antibiotic treatment on mate discrimination were due to elimination of some other bacterium. However, Wolbachia are implicated by several lines of evidence. First, the reduction in mate discrimination occurred following antibiotic treatment in crosses involving Wolbachia infected populations, but no reduction was observed in crosses between three uninfected populations, even though these also showed significant levels of mate discrimination. Because antibiotic treatment did not alter the level of mate discrimination in these lines, we would have to posit that the bacterium responsible must be in the Wolbachia infected lines but not in the three uninfected lines. This would seem to rule out the implication of a general surface or gut bacterium (e.g., one affecting cuticular hydrocarbons) common to all the Drosophila populations. Second, use of primers specific to common bacterial symbionts did not detect other bacteria, and sequencing of multiple clones from amplifications using general prokaryotic primers only yielded Wolbachia sequences. Furthermore, antibiotic treatments did not alter mate acceptance in populations that did not show divergence in mate discrimination (e.g., crosses between populations vs their antibiotically treated derived population). Therefore, this rules out a general (undetected) bacterium that is common to the populations and is involved in mate acceptance cues. In contrast, antibiotic treatment affects mate discrimination only when populations are infected with Wolbachia and show divergence in mate discrimination. Although our results do not formally prove that Wolbachia are responsible (this would require reintroduction of the bacterium into the cured populations by microinjection), Wolbachia are strongly implicated as the causative agent on the basis of these experiments.

The reduction in assortative mating is reciprocal in crosses between treated Wolbachia infected and uninfected populations. This suggests that the effect is not due simply to female mate discrimination against infected males, but involves an interaction between males and females. Clearly, more studies are needed to determine the behavioural, genetic and physiological basis of the observed nonrandom mating.

There is growing interest in the possible role of bacterial associates, such as Wolbachia, in host speciation. Cytoplasmic-incompatibile Wolbachia have been proposed to promote host speciation, due to their effects on levels of postmating reproductive incompatibility (Werren 1998; Shoemaker et al. 1999; Bordenstein 2003). In addition, models have been developed that predict Wolbachia will favor evolution of premating isolation (Randerson et al. 2000; Telschow et al 2005b). However, there are a number of counterarguments to the view that Wolbachia are involved in host speciation. For example, Coyne and Orr (2004) argue from comparisons of levels of postzygotic isolation in Drosophila species that there is little evidence of significant postzygotic isolation at very short genetic distances, which might be expected if Wolbachia was important in the early stages of speciation in this group. However the potential contribution of Wolbachia to isolation has not yet been directly assessed in those crosses, nor has the infection status of species in those studies been systematically determined. Therefore, it is currently difficult to assess the potential role of Wolbachia in that dataset. In contrast with the arguments above, Shoemaker et al. (1999) show that Wolbachia are a significant contributor to postzygotic isolation in sibling species of a pair of mushroom feeding Drosophila species; and Bordenstein et al. (2001) have shown that Wolbachia-induced incompatibility has arisen early in the speciation process in the parasitic wasp genus Nasonia. Moreover, Telschow et al. (2005b) show that CI Wolbachia can select for premating isolation more readily than the standard genic incompatibilities typically observed early in the evolution of postzygotic isolation. If this situation is common then detection of Wolbachia induced postzygotic incompatibility could be obscured by the rapid evolution of premating isolation. Clearly, there are contravening lines of evidence and arguments both supporting and disputing the role of Wolbachia in host speciation. Therefore, further investigation, rather than premature dismissal based on relatively weak arguments on both sides of the issue, is justified.

Within this frame we have shown yet another way that Wolbachia and possibly other associated bacteria could affect rates of speciation, by influencing the level of mate discrimination (premating isolation) in diverging populations. Even in the time course of experimental D. melanogaster populations subjected to divergent selection for toxins in food, partial premating isolation has evolved and Wolbachia (or an as yet undetected associated bacterium) contributes to the level of isolation.

In general, Wolbachia are incredibly abundant and widespread in arthropods (Werren 1997; Stouthamer et al. 1999;Werren and Windsor 2000; Stevens et al. 2001; Bourtzis and Miller 2003). It is worth noting here that a recent study shows that about 30% of D. melanogaster stocks in the stock center are Wolbachia-infected (Clark et al. 2005). Furthermore, a myriad of other parasitic and mutualistic microorganisms occurs in association with invertebrate and vertebrate species (O'Neill et al. 1997; Steinhert et al. 2000; Seckbach 2002; Bourtzis and Miller 2003). The discovery that symbiotic microbes such as Wolbachia can have strong influences on mate preference therefore may have very broad significance to eukaryotic speciation.