Introduction

The terms “gynandromorph” and “intersex” describe individuals simultaneously displaying both male and female sex characteristics. Although typically rare, their occurrence is documented throughout the animal kingdom, particularly in insects (Bergerard 1972). While the terms are often used interchangeably, gynandromorphs (also referred to as sexual mosaics) and intersexes are not the same. Gynandromorphs are individuals composed of a mosaic of genetically male and female tissues corresponding to distinct male and female characters with clear boundaries between the genetically different tissues (Brust 1966; Morgan 1916). In contrast, intersexes are composed of a single sexual genotype but exhibit both male and female sex characteristics which often appear as a blend or intermediate of the two sex characters (Goldschmidt 1916; Sturtevant 1920).

There are various possible origins of intersexuality and gynandromorphism, and the mechanisms of many are still not understood. Gynandromorphism is thought to generally originate from a number of developmental and fertilization errors, which may result in such conditions as double fertilization, delayed syngamy, or the loss or addition of a sex chromosome (Cooper 1959; Homsher and Yunker 1981; White 1973). Intersex, on the other hand, likely originates from interference with one or more of the regulating factors involved in the sex determining process of the individual, often caused by mutations, polyploidy or parasitism (Beukeboom and Kamping 2006; Cline 1984; Rigaud and Juchault 1998). While gynandromorphs tend to exhibit discrete patterns such as bilateral symmetry, anterior–posterior division, and transverse patterning and characters with clear demarcation between male and female (Cooper 1959), intersexes often display a range of intermediates between male and female types and characters (Sassaman and Fugate 1997; Vance 1996).

Within the haplodiploid insect order, Hymenoptera, intersexuality and gynandromorphism have been recorded more frequently in the Aculeata (wasps and bees) than other groups. They are documented in more than 40 ant species (Yoshizawa et al. 2009), and a recent review of gynandromorphism in bees lists specimens from 64 species (Wcislo et al. 2004). However, many of the causal mechanisms and conditions required for gynandromorphism remain unknown and may vary between taxa. For instance, in the pharoah's ant, Monomorium pharonis, gynandromorphs only result from heat shock (Berndt and Kremer 1982), whereas in the Malaysian ant, Cardiocondyla kagutsuchi, gynandromorphs appear spontaneously at normal temperatures (Yoshizawa et al. 2009). Similarly, gynandromorphs in the honeybee can be induced with exposure to extreme temperature (Drescher and Rothenbuhler 1963) but also appear spontaneously in certain populations, resulting from the entry of more than one sperm into the egg (Laidlaw and Tucker 1964).

Gynandromorphs and intersexes have been documented less frequently among the Parasitica, noted in only eight of the more than 40 families of parasitic wasps. This lack of documentation does not necessarily indicate that aberrant forms are rarer in this group, but may reflect a size and/or study bias. Members of this group are often very small and aberrant forms are likely to be overlooked. However, multiple aberrant individuals have been documented in the two most intensively studied taxa in the superfamily, Chalcidoidea, which contains the smallest specimens.

A variety of gynandromorphic types appear in the parasitic wasp Bracon hebetor (=Habrobracon), arising from multiple origins such as division of accessory nuclei, postcleavage fertilization and polyspermy (Clark and Gould 1972; Greb 1933; Whiting and Anderson 1932). Two types of intersexes have also appeared spontaneously in lab cultures of B. hebetor, both associated with genetic mutations (Whiting 1943b; Whiting et al. 1934). For decades, B. hebetor served as a model system for the study of hymenopteran development and genetics, and the valuable data obtained from the appearance of gynandromorphs and intersexes ultimately led to the discovery of single-locus complementary sex determination (sl-CSD). Sl-CSD is the widespread and ancestral mechanism of hymenopteran sex determination in which sex is determined by multiple alleles at a single sex determining locus—individuals which are heterozygous at the sex determining locus develop as females, while those which are hemizygous or homozygous at this locus develop as males (Whiting 1943a).

Intersexes are well described in the parasitoid wasp, Nasonia vitripennis, with more than 60 arising from a single female in a lab culture due to a spontaneous genetic mutation at a single locus (Saul 1962). In 2006, Beukeboom and Kamping reported the appearance of “gynandromorphs” in a polyploid mutant strain of N. vitripennis which, like other wasps in the superfamily Chalcidoidea, does not exhibit CSD. Both haploid and diploid gynandromorphs were produced by unmated triploid females, all displaying a clear anterior to posterior gradient of female to male expression, opposite that of previously reported intersexes. Since the gynandromorphic specimens arose from unfertilized eggs, precluding fertilization errors from determining the phenotype and it is unlikely that multiple developmental mishaps occurred to allow gynandromorphs to develop from eggs with both haploid and diploid nucei, it is likely that the “gynandromorphs” described in this study are, in fact, intersex specimens. More recently, haploid intersexes were reported from unfertilized eggs in several field-collected strains of N. vitripennis (Kamping et al. 2007). Like those from the polyploid mutant strain, the intersexes displayed anterior to posterior feminization ranging from slightly feminized to appearing completely female. Also, like those from the polyploid mutant strain, the specimens were referred to as “gynandromorphs”. However, cytological examinations revealed these individuals to be haploid and not mosaics of haploid and diploid tissues, determining them to be intersexes (Beukeboom et al. 2007; Kamping et al. 2007). The intersex phenotype was shown to have heritable nuclear and cytoplasmic components sensitive to heat exposure and may be the result of a temperature-sensitive maternal effect sex determining factor produced in egg (Kamping et al. 2007).

Intersex and/or gynandromorph production appears to be quite common among the tiny chalcidoid wasps of the genus, Trichogramma. Many of these egg parasitoids are used for the biological control of lepidopteran pests (Smith 1996). The wasps typically exhibit arrhenotokous reproduction whereby diploid female offspring arise from fertilized eggs and haploid male offspring arise parthenogenetically from unfertilized eggs. However, populations of some Trichogramma spp. also display thelytokous reproduction associated with infection by Wolbachia, a maternally transmitted, intracellular reproductive parasite within the alpha-proteobacteria (Rousset et al. 1992; Stouthamer et al. 1993). Wolbachia induce thelytoky (the parthenogenetic production of diploid, female offspring) in their Trichogramma hosts through a mechanism of gamete duplication in which unfertilized eggs undergo diploidization through a modified mitotic division in which the maternal chromosome set is duplicated but fails to separate at anaphase. Mitotic divisions following this initial event are normal, resulting in the development of completely homozygous, Wolbachia-infected, female offspring (Stouthamer and Kazmer 1994). This is possible because, like Nasonia, Trichogramma have a mechanism of sex determination that is different from CSD, which allows homozygous diploid individuals to develop as functional females.

Sexually aberrant individuals (variously referred to as intersexes, sexual mosaics and gynandromorphs) have been documented in Wolbachia-infected, thelytokous populations of five Trichogramma species: T. chilonis, T. cordubensis, T. deion, T. oleae, and T. pretiosum (Bowen and Stern 1966; Cabello and Vargas 1985; Chen et al. 1992; Pintureau and Bolland 2001; Pintureau et al. 2002) and in several species of the Encyrtidae (Stouthamer 1997). Specimens are generally described as having female genitalia with one or two male antennae, but in two publications providing more detailed descriptions, individuals belonging to T. deion (reported as semifumatum) and T. chilonis were described as exhibiting anterior–posterior masculinization of varying degrees. Also, both studies reported a very small number of wasps displaying anterior–posterior feminization (Bowen and Stern 1966; Chen et al. 1992). Sexually aberrant forms have not been documented in Trichogramma in the absence of Wolbachia infection.

While there are clear genetic contributors to intersex and gynandromorph production in several taxa, a genetic component to the production of these aberrant phenotypes in Wolbachia-infected Trichogramma has not yet been established. However, several experiments demonstrate an association between aberrant phenotypes and maternal exposure to high temperatures (Bowen and Stern 1966; Cabello and Vargas 1985; Pintureau and Bolland 2001; Pintureau et al. 1999; Pintureau et al. 2002). It is unclear whether these individuals are intersexes or gynandromorphs. Only the study by Bowen and Stern (1966) using T. deion provides any details regarding the morphologies of the aberrant wasps which they conclude are true intersexes produced as intermediates when a temperature-sensitive sex-determining mechanism is exposed to intermediate temperatures. However, since discovering the role of Wolbachia in inducing thelytoky in many Trichogramma species, Stouthamer and Kazmer (1994) have suggested the aberrant specimens are, in fact, gynandromorphs composed of haploid and diploid tissues resulting from delayed gamete duplication allowing one or more mitotic divisions to occur before fusion of mitotic products. It is speculated that the phenotypes occur due to a heat-induced reduction in Wolbachia titer (Pintureau and Bolland 2001; Pintureau et al. 1999).

In this study, we present a detailed characterization of the morphologies associated with the sexually aberrant individuals produced by thelytokous Trichogramma kaykai (Pinto et al. 1997) and examine patterns in the production of these phenotypes. When viewed in light of the wasp's association with Wolbachia and additional knowledge about similar phenomena in other Trichogramma species, our examinations provide insight into the nature and origin of sexually aberrant phenotypes in Trichogramma, allowing us to determine with confidence that the deviant individuals occasionally appearing in Wolbachia-infected, thelytokous Trichogramma species are actually intersexes and not gynandromorphs or mosaics as they have been referred to in past literature. Identifying the nature of these aberrant forms will contribute toward elucidating the mechanism of sex determination in Trichogramma and is essential to understanding the methods by which Wolbachia is able to manipulate host sex in this system.

Materials and methods

Origin and maintenance of cultures

Two Wolbachia-infected, thelytokous Trichogramma lines were used in this study, each derived from a single female collected in its natural habitat. The wasps used to initiate the T. kaykai lines were collected in May, 2004 in the Kelso Dunes area (KD111) and Sidewinder mountains (SX58) of the Mojave Desert (San Bernardino County, California, USA) from eggs of the butterfly, Apodemia mormo, removed from the host plant, Eriogonum inflatum. Species identifications were confirmed using molecular protocols described in Stouthamer et al. (1999). Each line was maintained individually in the laboratory in 12 × 75 mm glass culture tubes stopped with cotton and incubated at 24°C, L:D = 16:8 and 50% relative humidity. Every 11 days, cultures were offered fresh honey and a surplus of irradiated eggs of Ephestia kuehniella adhered to small rectangular pieces of card stock using double-sided tape (egg cards).

Both lines in this study exhibited thelytokous reproduction under normal lab rearing conditions, indicating Wolbachia infection. The polymerase chain reaction (PCR) was used to confirm infection status (Werren et al. 1995).

Production of aberrant offspring

Under normal laboratory rearing conditions, Wolbachia-infected Trichogramma exhibit thelytokous reproduction, producing almost exclusively female offspring with extremely infrequent formation of males and aberrant types. Maternal rearing at high temperatures has been shown to elicit the formation of sexually aberrant offspring in Wolbachia-infected Trichogramma (Bowen and Stern 1966; Cabello and Vargas 1985). Bowen and Stern (1966) demonstrated that maternal exposure of T. deion to temperatures ranging from 25°C to 32°C induces the formation of “sexual mosaics,” while Cabello and Vargas (1985) found that these forms were produced in T. cordubensis only with maternal exposure to 29°C. To determine the optimal temperature for inducing the formation of aberrant forms in T. kaykai, mothers were reared at temperatures ranging from 28.5°C to 31.5°C. The proportion of aberrant offspring increased with increasing temperature. However, at the highest temperature, 31.5°C, total offspring dropped dramatically. Therefore, 30.5°C was selected for induction of aberrant forms.

Groups of twenty newly emerged, virgin Wolbachia-infected females from the SX58 and KD111 lines were each placed in cotton-stopped glass culturing tubes and allowed to oviposit at room temperature for 2 h on egg cards containing a surplus of E. kuehniella eggs. After 2 h, the wasps were removed and discarded, and the tubes containing the stung eggs were placed at 30.5°C. Approximately 12 h prior to emergence, 50 parasitized eggs were removed from each card using a fine tip paint brush. The eggs were placed into individual cotton-stopped glass culture tubes and placed at 24°C. Upon emergence, 15 females were chosen at random from each line and offered diluted honey and fresh egg cards daily for 7 days, the maximum amount of time preliminary experiments showed that most heat-exposed mothers survived and continued to lay eggs. Emerging offspring of these mothers were sexed daily and examined for abnormal sexual characteristics. Arcsine square root-transformed sex ratio data were analyzed using a mixed factorial General Linear Model with repeated measures in Minitab® 15 (State College, PA).

General characterization of aberrant individuals produced by T. kaykai

Subsequent to examining production patterns among aberrant offspring, additional aberrant offspring were induced by rearing the T. kaykai lines SX58 and KD111 at 30.5°C on three separate occasions. The aberrant wasps produced under these treatments were pooled with those produced by the same lines in the above experiment to obtain a total of 199 aberrant wasps which characters were examined and described considering the entire group.

Establishment of cured lines and reproductive ability of aberrant wasps

In order to provide mates for investigating the sexual function of aberrant wasps, corresponding arrhenotokous, Wolbachia-free cultures were obtained for each thelytokous line by feeding starved 1-day-old wasps a 0.5 mg/ml solution of the antibiotic, rifampicin, mixed with honey. The wasps were allowed to feed on the antibiotic honey for 24 h then given a fresh solution of antibiotic honey and an egg card. Wasps were allowed to parasitize host eggs for 24 h, and the egg cards were removed, isolated and incubated at the normal rearing temperature. Fresh egg cards and antibiotic honey solution were replaced daily for three additional days. Adult wasps emerging from the last egg card (day 4) were used to repeat this protocol for three more generations. Wasps from the fourth and 15th generations were tested and determined to be free of Wolbachia infection using PCR (Werren et al. 1995). Arrhenotokous wasps were then maintained more than 20 generations prior to experimental use.

The sexual function of individuals exhibiting mixed sex characters was investigated using three specimens with female genitalia belonging to categories C, D, and G and two specimens with male genitalia, one of category H and the other not described in enough detail to categorize the exact morphology. Wasps with male genitalia were paired with Wolbachia-free, arrhenotokous virgin females, and copulation was observed. The female was then offered fresh E. kuehniella eggs and her offspring examined for the presence of females, indicating fertilization had occurred. Specimens with female genitalia were not provided mates, but their ability to lay viable eggs was determined by providing them with fresh E. kuehniella eggs for oviposition and examining them for parasitization and offspring emergence. Few specimens were examined in this study because of the rarity of aberrant wasps and the resulting difficulty in coordinating their emergence with the availability of Wolbachia-free virgin wasps for mating.

Results

General characterization of aberrant individuals produced by T. kaykai

Patterns of external aberrant characters were classified into 11 morphological categories, A–K in order of increasing anterior to posterior masculinization (Table 1). More than half (57%) of the aberrant individuals examined were assigned to category G, having two male antennae, female coloring and female genitalia. The second and third most frequently occurring phenotypes fell into categories D and F, at 17.5% and 8.5% of the aberrant individuals. Categories E, H, I, J and K each contained only one or two specimens. Only four (2%) of the aberrant wasps examined had male genitalia, and there was no apparent pattern to when their posteriorly masculinized offspring were produced in the oviposition sequence, as they arose from eggs oviposited by their mothers on days 2, 3, 5, and 7 of the study (results not shown). Additionally, there was no observed correlation between early and late production and phenotypes, aberrant morphologies appear to be produced in a haphazard manner. The most common phenotype, in category G, was produced by females on every oviposition day between 1 and 10, with the exception of day 9. Also, among aberrant individuals displaying both male and female antennal features, there was no tendency toward right or left masculinization (Chi-square test, P > 0.05).

Table 1 Aberrant morphotypes produced by thelytokous Trichogramma kaykai
Full size table

Production of aberrant offspring

To characterize and compare patterns of daily production of aberrant offspring in thelytokous T. kaykai, 15 genetically identical Wolbachia-infected females from each of two genetically distinct lines, SX58 and KD111, were reared at 30.5°C. Three females in the SX58 line and two in the KD111 line perished prior to oviposition or did not lay eggs and were excluded from analysis. Aberrant individuals were rare in both lines with only 25 appearing among 520 total offspring produced in this experiment. In the SX58 line, only seven of the 12 ovipositing wasps produced aberrant offspring. Of the 304 offspring produced by these seven females only 11 were aberrant (3.5%). Similarly, in the KD111 line, only nine of the 13 ovipositing wasps produced any aberrant types, together producing 14 aberrant wasps among 216 total offspring (6.5%). Daily and total production of males and aberrant offspring were compared between the two T. kaykai lines. In general, the daily proportions of males and aberrant offspring produced increased with time from the onset of oviposition (Fig. 1). Mixed general linear model analysis with repeated measures shows that production increased significantly over the 7-day oviposition period in both lines (Table 2). Male production was affected by a significant interaction between line and day (Table 2), and the increase in male production was less apparent in SX58. However, that interaction was absent in the case of aberrant individuals. Also, there were no significant differences in the production of aberrant offspring between the two lines (Table 2). KD111 produced a significantly greater proportion of males over the course of the experiment (Table 2) due mainly to it producing a significantly higher proportion of males on each of days 3 through 7 (RM ANOVA, F 6 95 = 3.12, P ≤ 0.05).

Fig. 1
figure 1

Temporal variation in the mean production of males and aberrant offspring by thelytokous females from T. kaykai lines a SX58 and b KD111 reared at elevated temperature (30.5°C). Error bars represent ±1 s.e

Full size image
Table 2 Statistical results for GLM repeated measures ANOVA model testing the effects of line, day, and interaction on the production of males and aberrant offspring
Full size table

Reproductive ability of aberrant wasps

Five aberrant individuals exhibiting various morphologies were tested for reproductive function. Wasps with female genitalia were offered fresh E. kuehniella eggs for oviposition as virgins. All successfully parasitized host eggs and produced normal female offspring. Wasps with male genitalia were mated with cured, arrhenotokous T. kaykai females, copulation was observed and the mated female was given fresh eggs for oviposition. Both males tested appeared to copulate successfully, but the mated females produced only male offspring, indicating these very masculinized aberrant wasps were not sexually functional. It is unclear whether this failure was due to a problem with sperm production, sperm transfer or sperm quality.

Discussion

The aim of this study was to present a detailed characterization of the morphologies and patterns of production of sexually aberrant individuals appearing as a result of maternal exposure to high temperature in Wolbachia-infected, thelytokous T. kaykai. This information, when compared with additional knowledge about similar phenomena in other Trichogramma associated with Wolbachia, indicates an intersex origin, disproving the hypothesis presented by Stouthamer and Kazmer (1994) that the intermediates are gynandromorphs and suggesting that Wolbachia-induced thelytoky involves diploid restoration and feminization as independent processes.

Evidence for intersexuality

Sexually aberrant individuals displaying secondary sexual characteristics of both male and females have been documented in studies of several Trichogramma species infected with parthenogenesis-inducing (PI) Wolbachia and have been referred to interchangeably in these studies as intersexes, gynandromorphs and sexual mosaics. However, the true nature of these phenotypes was unknown. Based on knowledge of Wolbachia's role in the induction of thelytoky through gamete duplication in several parasitoid wasps, it has been hypothesized that these deviant individuals are gynandromorphic in nature. Stouthamer and Kazmer (1994) suggest the gynandromorphs are formed when Wolbachia-induced fusion of the mitotic products is delayed, occurring instead at some time after the first mitotic division. This delay results in the formation of a gynandromorph composed of both haploid and diploid tissues which express corresponding male and female phenotypes. However, our results, examined in light of previous studies of aberrant phenotypes in Trichogramma spp., suggest that these individuals are in fact intersexes composed of a single genetic constitution but displaying a combination of male and female secondary sex characteristics. There are several lines of evidence from which we draw this conclusion. The first is a clear inconsistency involving the proportions of male and female tissues that would be present if these were, in fact, gynandromorphs produced through delayed gamete duplication. The proportions of male and female tissues of gynandromorphs resulting from this mechanism should correspond to the timing of gamete duplication. For instance, if the fusion of mitotic products were to occur in the second mitotic division, the proportion of female to male cells in the embryo should be 1:2. Thus, according to this hypothesis, a gynandromorph produced by this mechanism would consist of at least two-third male tissue. This is not the case, particularly since the majority of morphologies we have documented display significantly more female characteristics, a phenomenon repeatedly documented in the literature. Stouthamer (1997) acknowledges this discrepancy, but invokes a differential rate of division between haploid and diploid cells as an explanation. However, there is no evidence for this mechanism. On the contrary, gynandromorphism has been shown in several other hymenopteran taxa, and no differential rate of division between haploid and diploid cells was found (Petters 1977; Yoshizawa et al. 2009).

Also, indicative of intersexuality is the range of morphologies documented in this study. We described a variety of intersexual morphotypes appearing in T. kaykai which we categorized into 11 groups ranging, anteriorly to posteriorly, from very female to very male. It has been suggested that intersexual characters in Trichogramma are produced with anterior–posterior male–female polarity (Stouthamer 1997), and this pattern is not uncommon in both gynandromorphs and intersexes (Dalla Torre and Friese 1899; Kamping et al. 2007; Mori and Perondini 1984). However, our results indicate that while the intersexes are predominantly femalelike, intersexual character combinations in T. kaykai appear with no directed pattern with regards to timing of production. A number of morphologies were produced along a continuum of intersexual morphotypes ranging from female to male as opposed to the clearly demarcated patterning characteristic of gynandromorphism.

Bowen and Stern (1966), the only other resource offering detailed information on the morphologies of aberrant individuals, also described a gradient of intermediate phenotypes. Not only was a range of female to male phenotypes produced among the specimens, but characters within individual specimens also exhibited intermediate degrees of male- or femaleness such as antennae with varying amounts of setae. This blend of characteristics is seen among the aberrant individuals produced in this study, and again, this lack of clear demarcation between affected tissues is indicative of intersexuality. Further examples of intermediate characters occur in T. deion and T. chilonis, which produce deviant phenotypes with aberrant genitalia that are often intermediates between male and female (Bowen and Stern 1966; Chen et al. 1992).

Another indication of intersexuality is Trichogramma's association with the reproductive parasite, Wolbachia. Many parasites exist which alter host reproduction and intersexuality is often associated with such parasitism. For instance, infection with mermithid nematodes is common in insects, and mermithid-induced intersexuality is well-studied in Diptera and Ephemeroptera (McKeever et al. 1997; Vance 1996). Microsporidia are known to induce intersexuality in various crustaceans, and infection with various dinoflagellates is associated with intersexuality in copepods (Bulnheim 1978; Ginsburger-Vogel 1991; Kimmerer and McKinnon 1987). Intersexes have also been seen in arthropods infected with thelytoky-inducing Cardinium such as Encarsia meritoria (=hispida) (Giorgini and Viggiani 1996) and Brevipalpus ovobatus (Groot and Breeuwer 2006).

Wolbachia is perhaps the most renowned reproductive parasite among terrestrial arthropods, and its association with intersexuality is well-documented. In the isopod, Armadillidum vulgare, feminizing Wolbachia induces a range of intersexual phenotypes (Bouchon et al. 2008). Incomplete feminization by Wolbachia also results in the production of intersexes in the leafhopper, Zyginidia pullela, and in the lepidopteran insects, Ostrinia scapulalis and Eurema hecabe, when treated with antibiotics (Kageyama and Traut 2004; Narita et al. 2009; Negri et al. 2006). Among Trichogramma species, heat-induced intersexuality is only documented in thelytokous, Wolbachia-infected Trichogramma, not in uninfected, arrhenotokous populations (Haile et al. 2002; Harrison et al. 1985; Pintureau and Bolland 2001), indicating that Wolbachia is somehow associated with intersex production.

The origin of intersex in Trichogramma

Heat and antibiotic treatments have been shown to result in the production of males in Trichogramma infected with PI Wolbachia (Bowen and Stern 1966; Pintureau and Bolland 2001; Pintureau et al. 1999; Stouthamer et al. 1990). Antibiotic-induced male production in Muscidifurax uniraptor has been correlated with reduced Wolbachia density (Zchori-Fein et al. 2000). Bowen and Stern (1966) found in T. deion that rearing mothers at high temperatures resulted in increased male and intersex production in their broods and that intersexes produced after maternal exposure to higher temperatures displayed increasingly masculine characteristics. Thus, it has been suggested that intersex production may also be “dose-dependent”, appearing as a result of a reduction in Wolbachia titer in the developing egg (Pintureau et al. 1999). If Wolbachia titer affects the masculinity of intersexes, one would also expect the masculinity of intersexes produced later in a female's oviposition cycle to increase as Wolbachia titer is reduced in later produced eggs. This correlation was absent in our study. As a matter of fact, very few intersexes were produced displaying predominantly male characteristics. This could be indicative of a slight but additional feminizing role for Wolbachia after gamete duplication is induced. We did, however, observe a significant increase in the overall number of males and intersexes produced with increasing oviposition period (Table 2), which agrees with this hypothesis.

Reductions in Wolbachia titer have been associated with reduced penetrance in other reproductive phenotypes induced by the parasite such as male killing in Drosophila bifasciata (Hurst et al. 2000) and cytoplasmic incompatability in Aedes albopictus (Sinkins et al. 1995). Also, the antibiotic-induced males and intersexes appearing in the lepidopterans, Ostrinia scapulalis, and Eurema hecabe, are believed to result from insufficient numbers of bacteria required for complete feminization (Hiroki et al. 2002; Kageyama and Traut 2004).

The recent discovery of haploid females and intersexes in Nasonia (Beukeboom et al. 2007) demonstrates that ploidy is not as significant a determining factor in sex differentiation in non-CSD Hymenoptera as in those with CSD and indicates that an additional mechanism is involved in inducing a female phenotype. Encarsia hispida is another non-CSD wasp in which ploidy alone does not determine sexual differentiation. In this case, the endocellular reproductive parasite, Cardinium, appears to provide the additional mechanism required for the feminization of diploid individuals and, thus, production of females (Giorgini et al. 2009).

It is clear that Wolbachia is responsible for inducing the diploid state in unfertilized Trichogramma eggs through gamete duplication. However, it is probable that this process is not the only interaction required by Wolbachia to produce the female phenotype. If the mechanisms involved in thelytoky induction are dose-dependent requiring different threshold titers of Wolbachia, it is possible to obtain an egg with a high enough initial threshold of Wolbachia to induce gamete duplication, producing a diploid individual, but upon continual division of the cells (or due to uneven distribution within the cells), titer may be reduced below the threshold number of bacteria required for full feminization.

Our results and analyses clearly indicate that the sexually aberrant phenotypes appearing in offspring produced by heat-treated thelytokous, Wolbachia-infected T. kaykai are intersexual in nature. The reproductive function tests, demonstrating that wasps with female genitalia are able to produce normal offspring despite varying degrees of male secondary sex characteristics while femalelike wasps with male genitalia are unable to reproduce, strongly suggest that the wasps are diploid. However, cytological evidence documenting ploidy is needed to definitively confirm this conclusion. Also, the extent of Wolbachia's role in inducing the female phenotype after gamete duplication is unclear. It appears there is an association with Wolbachia infection and intersex production and that intersex is absent in studies using uninfected Trichogramma. However, to our knowledge, there have been no studies directly comparing intersex production in offspring of heat-treated Wolbachia-infected and uninfected Trichogramma. The tiny size of the wasp and great infrequency with which intersexes are produced makes it plausible that they are produced in the field or in experiments but are not noticed; thus, it remains a possibility that intersexes are produced as heat interferes with the sex determining processes of the diploid host and not Wolbachia. Experiments directed at distinguishing between the effects of heat on Wolbachia function versus host sex determining processes using infected and uninfected wasps would resolve this question. Based upon the results of such studies, the origin of the intersexual phenotype and, potentially, the current hypothesis of incomplete expression of thelytoky due to heat-reduced Wolbachia density can be investigated.