Two main transitory mating systems have been recognized: gynodioecy (females/hermaphrodites) and androdioecy (males/hermaphrodites). In plants, gynodioecy is more prevalent than androdioecy, a trend predicted from theory [4,7,8] and confirmed empirically [9]. According to theoretical predictions, hermaphroditic populations can become gynodioecious when a recessive male-sterility gene spreads in the population [4]. Selection for genetically-fixed gynodioecy is thought to occur as an adaptive response to inbreeding depression of selfing hermaphrodites. The gradual accumulation of subsequent mutations that suppress female functions of hermaphrodites can then result in a dioecious population. Androdioecy is predicted to be a rarer transitory mating system because mutants conferring maleness must achieve at least twice as many outcross fertilizations as hermaphrodites [4,5]. Only in very specific conditions, namely large male advantage, low male frequencies and self-incompatibility, is androdioecy expected to be maintained [10].

In animals, contrary to plants, androdioecy is more common than gynodioecy [11]. This may indicate that there are fundamental differences in the evolution of mating systems in animals as compared to plants. In fact, in most cases it seems that that androdioecy is derived from dioecy [12,13]. Very few detailed models specifying intermediate stages between the evolution of dioecy and hermaphroditism have been proposed, although hermaphroditism has arisen multiple times during animal evolution [14,15]. Selection for reproductive assurance has been invoked as the most plausible explanation for the evolution of hermaphroditism from dioecy.

Nematodes are emerging as a promising model system to understand the conditions and consequences of mating type evolution (eg., [16,17]). The ease of culturing them in the laboratory and the availability of genetic tools for a number of species allows experimental manipulation of breeding types [18-20]. The success of Caenorhabditis elegans as genetic model system for developmental biology encouraged the adoption of other hermaphroditic nematodes for comparative studies that include the rhabditids C. briggsae [21-23]and Oscheius tipulae [24,25], and the more distantly related diplogasterid Pristionchus pacificus [26-28].

For rhabditids and diplogasterids androdioecious nematodes, hermaphrodites are basically modified females that produce a limited number of sperm. In contrast to plants, hermaphroditic nematodes cannot fertilize other hermaphrodites. Thus, nematode males are the only outcrossing agents, which give them an advantage over males of androdioecious plant species. Previous models [7,8], which assumed that hermaphrodites cross-fertilize, are therefore not applicable to nematodes.

Because males are nonessential for reproduction in C. elegans, models have been developed to explain the maintenance of males [18,29,30]. C. elegans males are formed either as a result from rare non-disjunction events during hermaphrodite gametogenesis, or as cross-progeny from hermaphrodite/male crosses. Typically, C. elegans males are rarely found in nature and are present as 0.1% of the population in laboratory cultures [31,32]. According to theoretical models, high inbreeding depression, high male mating ability, and high a rate of production of males by hermaphrodite self-fertilization are factors required for maintenance of androdioecy in C. elegans [18]. Experiments on the C. elegans N2, however, a strain that has been kept in the laboratory for many thousand generations, contradict the theoretical expectations for equilibrium frequency of males [18,29]. Furthermore, C. elegans isolates derived recently from the wild seem to counterselect males due to outbreeding depression [33], low non-disjunction rates and relatively poor mating ability [32].

The purpose of this study is to characterize life-history traits of P. pacificus, a nematode that evolved androdioecy independently. P. pacificus and C. elegans are cosmopolitan species with overlapping geographic distribution, but different ecological niches. C. elegans has been found mostly in decomposing organic matter (compost piles, farmland and garden soil, and various invertebrates close to decomposing matter) [31,34], while P. pacificus has been mostly found in soil samples and in association with scarab beetles [35,36].

In C. elegans, temperature has an obvious effect on the rate of X chromosome non-disjunction [32]. To determine if this was also the case for P. pacificus, we allowed a subset of strains to self at two additional temperatures, 15°C and 25°C. Strains were selected according to their rates of non-disjunction at 20°C and geographical location from where they were collected. It is clear that the highest proportion of males is produced at intermediate temperatures (Fig. ​(Fig.2).2). Strains isolated from geographical locations with similar climatic conditions presented very distinct non-disjunction rates (eg, Japanese strains RS5212 and RS5207) at 20°C, but more similar at lower temperatures. To determine whether there is interaction between environmental and genetic effects, an ANOVA test was performed (Table ​(Table2).2). The results indicate that there is a significant interaction between strain and temperature, which preclude the interpretation of the among strain and temperature differences.

Rates of non-disjunction could affect brood sizes, since the production of each male generated by non-disjunction results in the production of one unviable or unfertile XXX hermaphrodite [44]. We selected 13 strains that span the range of X-chromosome non-disjunction rates (Fig. ​(Fig.3).3). The mean brood size of these inbred strains is 107 ± 15 SD progeny/hermaphrodite. There are significant genetic differences in brood size (F_12,100 = 2.20, P < 0.05) and embryonic mortality (F_12,98 = 5.72, P < 0.001) among different strains. However, no correlation was found between non-disjunction and embryonic mortality (r = 0.19, P = 0.52), or between non-disjunction and brood size (r = 0.10, P = 0.73). We also found no correlation between embryonic mortality and brood size (r = 0.17, P = 0.58).

In C. elegans, males of some strains are significantly more effective in mating than others [32,45,46]. We detected genetic variation in male fertilization ability among P. pacificus inbred lines as well (Fig. ​(Fig.4).4). Male fertilization ability was measured at two levels. First, we quantified male mating ability by scoring the proportion of matings that resulted in cross-progeny. Successful fertilization rates were equal to or higher than 90% for about half of the strains used for crosses. As an example of poor mating performance, males of strain RS5212 produced cross progeny with only 30% of the mated hermaphrodites. Second, we measured male fecundity by determining the proportion of male cross-progeny that resulted from successful fertilization events. P. pacificus showed significant variation of fecundity among strains (F_11,89 = 4.13, P < 0.001). RS5212 males produced some of the highest proportions (30%) of male cross-progeny among the strains analyzed, indicative of high fecundity. Most of the other strains (10/13) produced 10% or less male cross-progeny. We could not find any correlation between rates of X-chromosome non-disjunction and male mating ability (r = -0.47, P = 0.105), or between X-chromosome non-disjunction and male fecundity (r = 0.59, P = 0.08).

One of the factors that might influence efficient fertilization by males and sex ratios in cross progeny is sperm competition. Two types of sperm competition have been observed in nematodes: (1) between male and hermaphroditic sperm; (2) between X-bearing male sperm and male nullo-sperm. In C. elegans and C. briggsae, male sperm can efficiently outcompete hermaphroditic sperm [47-49], while in C. briggsae male X-bearing sperm has precedence over male nullo-sperm [49]. Precedence of X-bearing sperm results in bias towards hermaphroditic cross-progeny in the first days after mating. To identify cross-progeny and clarify the ratio of each sex after cross-fertilization in P. pacificus, we used a morphologically marked strain (see Materials & Methods). Hermaphrodites from strain PS312 pdl-9 (sy372), when mated to males, produced cross-progeny for about 4 days. Cross-progeny percentages peaked around the second to third day, accounting for approximately 65% of the brood (Fig. ​(Fig.5).5). Most animals died prematurely due to problems in egg-laying around the fifth day. Premature death may have precluded the usage of all male sperm stored in the hermaphrodite spermatheca and therefore the total proportion of male and hermaphrodite cross-progeny differed by 50% (X²₁ = 11.42; P > 0.05). It is clear, however, that X-bearing sperm of PS312 males has precedence over nullo-sperm: hermaphrodites are produced in higher proportions than male cross-progeny in the first three days (Fig. ​(Fig.5).5). For PS312 there is approximately equal chance of either hermaphroditic sperm or male sperm to fertilize an oocyte.

In the short term, high levels of inbreeding increase the chances for accumulation of detrimental recessive mutations [50-52]. The combination of high mutation rates and high level of inbreeding can potentially cause species extinction, making self-fertile organisms evolutionary "dead ends" [50,51]. Therefore, some self-fertilizing animals maintain high proportions of males in their populations because they show substantial inbreeding depression [53,54]. C. elegans, however, has no inbreeding depression for life span, brood size and various other life-history traits [29,33,55,56].

To determine if P. pacificus has inbreeding depression, we made four sets of crosses between different strains and determined the brood sizes of the hybrid F1s (Fig. ​(Fig.6).6). We selected four strains based on the criterion of being genetically divergent from each other, as determined by phylogenetic analysis of mitochondrial DNA sequences [37](Sommer, unpublished). Similar to C. elegans, we detected outbreeding depression, where the P. pacificus pure strains outperformed the hybrids [33]. Paired comparisons consistently resulted in negative heterosis values, which reflected the smaller brood sizes of hybrids when compared to the parental pure strains (Fig. ​(Fig.66).

To understand how significant the role of males is in distant relatives of C. elegans, we measured variation of various life-history traits important for outcrossing in strains of the nematode P. pacificus. The genus Pristionchus has at least four species that evolved androdioecy independently, including P. pacificus [39]. Although P. pacificus and C. elegans are distant relatives and have many developmental differences [26], they share some common features: the same type of sex determination (XX:XO), males are relatively rare [42], and the hermaphrodite is a modified female whose gonad first produces sperm and then shifts to produce oocytes [59]. Our results show that 23 P. pacificus strains produce males by X chromosome non-disjunction at comparable rates to C. elegans.

Production of males by non-disjunction comes at a cost, because for every XO male generated one XXX unviable hermaphrodite is produced. However, we did not find a correlation between high non-disjunction rates with lower brood sizes or higher rates of embryonic mortality. The percentage of embryonic mortality seems high for most strains, typically ranging between 10–30%. Similar rates were found for other selfing nematodes [60], but the causes for high mortality are unknown. It is possible that our results are slight overestimates of embryonic mortality, because we could not rule out the possibility that some hatched larvae died at the wall or edge of the plate before being scored. However, we repeated the assays and confirmed that strains scored as having high lethality displayed by many dead embryos.

Brood sizes in P. pacificus are about half of those of C. elegans [46,61] (this report). Although the difference might be partially explained by the higher embryonic lethality in P. pacificus, additional factors could also play a role. P. pacificus might be producing fewer sperm as an adaptation for earlier reproduction. Because spermatogenesis precedes oogenesis, the fewer sperm are produced, and the hermaphrodite can start reproducing earlier. C. elegans mutants that increase sperm number, for instance, were shown to be of disadvantage under natural conditions because the additional spermatogenesis delays oogenesis and therefore the minimum generation time [62]. Further work would be necessary to test whether the smaller brood size in P. pacificus is due to lesser sperm production.

Male mating is one of the most complex behaviors in nematodes [63,64], and many factors can influence successful mating: male sexual drive, ability of the male to locate the hermaphrodite, ability of the male to insert its spicules into the vulva, and amount of sperm transferred in each mating. Mating assays are typically performed in agar plates, where hermaphrodites and males are placed in a small spot of bacteria for a few hours or days [65,29,32,68]. We performed two tests for mating ability. The quantitative assay determined the proportion of hermaphrodites fertilized by 1–3 males in a 24-hour period. The qualitative assay tested male fecundity once they mated with hermaphrodites. Our results show that most hermaphrodites are fertilized by males in the above conditions. However, there is some variability of the number of progeny sired by each male. Some strains are clearly more fecund, such as RS5212 and RS5275. Because there is no standard nematode mating assay, it is not possible to compare our results with the ones performed in C. elegans. Given the benign conditions in which the crosses were performed, it is very likely that in more natural conditions male mating ability would be much lower. Further studies are required to access the effects of male nutritional status [69], habitat surface [70], health status [71] and longevity [72] on mating success.

The higher competitive ability of the male sperm to outperform the hermaphroditic sperm in fertilizing oocytes would be an indication for selection acting on males. This is well documented for C. elegans strain N2 [47], where mated hermaphrodites produce outcross progeny almost exclusively in the first 24 hours after mating. C. elegans male sperm can anchor to the wall of the spermatheca, preventing them from being swept away by the exiting eggs. In our study with P. pacificus PS312, we observed that male sperm is equally likely to fertilize an oocyte as a hermaphrodite sperm (Fig. ​(Fig.5).5). It is possible that P. pacificus male sperm lost the ability to anchor themselves, or that the shape of the spermatheca does not favor attachment. In fact, it has been observed that the shape of the P. pacificus spermatheca is very different from the one in C. elegans [73]. In C. elegans, the larger size of the male sperm has been reported to give a higher competitive advantage due to their faster crawling and ability to displace smaller sperm [74,75]. It remains to be determined whether the size of the sperm also plays a role in P. pacificus.

Differences in the competitive ability of P. pacificus X-bearing sperm over nullo-sperm were detected. Precedence in X-bearing sperm provides an advantage to faster reproduction and ability to colonize new habitats [49]. P. pacificus, which feeds on bacteria and fungi that grow on insect carcasses, must clearly rely on early progeny that are self-fertile to grow on such ephemeral habitats [76].

Theoretical models predict that males persist in androdioecious species with high levels of inbreeding depression [18,30]. Our results suggest that P. pacificus has outbreeding depression, where the hybrid has lower fitness than the pure strains. This might be caused by hybrid breakdown of co-adapted gene complexes, as found for C. elegans [33]. It should be noted, however, that we measured only one fitness trait in our assays for inbreeding depression (brood size). Different results could be produced for other ecologically-relevant traits, such as longevity, pathogen susceptibility [77], and ability to undergo dauer formation [78]. Other selfing species, such as C. briggsae, do not show outbreeding depression [79]. This indicates that this phenomenon is not particular to self-fertilizing nematodes.