The range and quality of prey species differ greatly among closely related species of predators. However, the factors responsible for this diversified niche utilization are unclear. This is because the predation and resource competition do not always prevent species coexistence. In this paper, we present evidence in support of reproductive interference as a driver of niche partitioning, focusing on aphidophagous insect. Firstly, we present closely related generalist and specialist species pairs in aphidophagous lacewings to compare the reproductive interference hypothesis with two other hypotheses that have been proposed to explain niche partitioning in lacewings and sympatric speciation through host race formation and sexual selection. Secondly, we present a case study that shows how reproductive interference can drive niche partitioning in sibling ladybird species. Thirdly, we show that many ladybird genera include species inhabiting the same region but having different food and habitat preferences, raising the possibility that reproductive interference might occur in these groups. Finally, we show that intraguild predation cannot always explain the niche partitioning in aphidophagous insects including hoverflies and parasitoids. On the basis of the evidence presented, we urge that future studies investigating predator communities should take account of the role of reproductive interference.

1. Introduction

In nature, closely related species often occupy niches that diverge with respect to both type and breadth, and such niche divergence is observed not only in herbivores but also in many predators [14]. A frequently proposed explanation of niche differentiation is the trade-off hypothesis, according to which adaptations of a species that allow it to exploit one resource decrease its fitness on other resources, and this trade-off leads to different niche utilization by different species [5]. In predators, the main driving force of food specialization may be morphological and behavioral adaptations that enhance a species’ prey capture performance against one prey species but simultaneously reduce it against another [6]. However, the trade-off hypothesis is inadequate to explain differences in food choice among some closely related predator species. In particular, some species that possess traits that make them highly efficient foragers specialize in a highly defended prey species, even though they could potentially utilize a wide variety of food items, including less well-defended prey (e.g., [79]). For example, large-mouthed marine cottid fish species have been shown in a laboratory setting to have a high feeding performance on both slow-moving prey (e.g., crabs, isopods, and gastropods) and more mobile, and thus more elusive, prey (e.g., fishes and shrimp), but in nature they feed predominantly on elusive prey [10]. Such examples raise the possibility that negative interspecific interaction, rather than a species’ own resource-use traits, can restrict its host range in nature.

Examples of negative interspecific interactions that might promote niche partitioning among predators include intraguild predation and parasitism, where species competing for a shared resource also prey on or parasitize one another [11, 12]. Predation by predators on predators has been widely reported under both laboratory and natural conditions [13, 14]. Moreover, multiple species involved in intraguild predation have been observed to persist in the same locality in field studies (see the “Aphidophagous Guilds”). Therefore, the idea that intraguild predation, which combines predation with resource competition, is sufficient to drive niche partitioning in predators is still controversial. Thus, we would propose that it should be fruitful to consider alternative interspecific interactions that might explain the divergent food choices of carnivorous species.

One reported type of interspecific interaction between closely related predator species is heterospecific mating interactions (e.g., [1517]). For example, in damselflies (Nehalennia), males frequently attempt to form a tandem pair with a heterospecific female, a behavior that may waste the time and energy of both species even if hybridization does not occur [18]. Moreover, in a molecular study, Fitzpatrick et al. [19] detected genetic introgression between specialist and generalist garter snakes, which is indirect evidence that interspecific copulation occurs in nature. In fact, such reproductive interference is a general mechanism that is applicable to various species assemblages with overlapping mating signals, regardless of trophic level, if at least one species engages in sexual reproduction [20, 21]. However, we think that the potential effect of reproductive interference on the spatial distribution and resource use by predator species has been underestimated, possibly because researchers have focused on mechanisms that are specific to predator-prey systems, such as intraguild predation and strong resource competition [22]. Therefore, we assert that reproductive interference should be given greater consideration in the search for a mechanism to explain ecological specialization in predators.

In this paper, we focus on aphidophagous insects. First, we describe aphidophagous lacewing genera that include generalist and specialist predatory species pairs. In lacewings, it has been hypothesized that sympatric speciation has occurred through host race formation or sexual selection. We argue instead that the observed niche partitioning between closely related species may actually be a consequence of niche diversification through reproductive interference after allopatric speciation. Next, we present as a case study of niche partitioning by reproductive interference our own research on two species of ladybirds (Harmonia) in Japan. Then, we consider whether reproductive interference-driven niche partitioning can explain niche partitioning in other predatory ladybirds. Finally, we describe some aphidophagous insect communities in which intraguild predation is known to occur and discuss reproductive interference as a possible mechanism that can drive niche partitioning of intraguild predators. We hope that the arguments made in this paper will encourage researchers to pay more attention to interspecific mating behaviors and their ecological consequences in predator communities.

2. Specialist and Generalist Lacewings in North America

In this section we first examine the possible role of reproductive interference in the ecology of two sibling species (i.e., reproductively isolated species that are nearly identical in their appearance [25]) of the lacewings (Neuroptera: Chrysopidae) in North America, one in genus Chrysopa and another in Chrysoperla. The Chrysopa pair includes a strict specialist that uses a strange woolly aphid, and the Chrysoperla pair includes a conifer specialist. Then we compare the reproductive interference model with models of sympatric speciation based on host race formation and sexual selection, which have been considered the main drivers of genetic and habitat diversification in sympatric lacewings.

2.1. Chrysopa and Chrysoperla Species Complexes

Chrysopa quadripunctata is a generalist predator that preys upon a wide variety of aphids and other soft-bodied arthropods that live on various plant species, including apple, elm, goldenrod, hickory, maple, oak, and rose. In contrast, Chrysopa slossonae is a specialist predator that exclusively utilizes the woolly alder aphid, Prociphilus tessellatus [31]. Interestingly, the woolly alder aphid covers its body with secreted wax, which it uses to attract the ants that protect the aphid colonies from predators [31]. The wax may also serve as a physical defense against attacking predators [32]. However, C. slossonae larvae exhibit a sophisticated behavior in which they scrape the wax from their prey and attach it to their own back, thus preventing their detection by the attending ants [31]. In addition to this camouflage strategy, the specialized morphology (long legs and a large mandible) of C. slossonae larvae may enable them to consume the woolly alder aphid efficiently [33].

We argue that reproductive interference by C. quadripunctata may be responsible for the restricted food range of C. slossonae, for the following reasons. First, C. slossonae larvae can develop on less elusive prey that they do not utilize in nature, such as Myzus persicae and A. pisum [7, 34]. These observations are evidence that the fundamental niche of C. slossonae is wider than its realized niche. Moreover, interspecific copulation and hybridization occur between C. slossonae and C. quadripunctata, at least under laboratory conditions [34]. Although interspecific pairs of C. slossonae and C. quadripunctata can produce viable hybrids, the ecological traits of such hybrids show the intermediate values between their parent species [34]. This indicates that hybridization may hamper the local adaptation to each habitat. Therefore, it is possible that C. slossonae is forced to specialize on a highly defended prey to avoid heterospecific mating interactions with C. quadripunctata, which utilizes various preferred prey species in nature.

Chrysoperla carnea is a habitat generalist that lives in fields and meadows during the summer and migrates to deciduous trees in autumn. In contrast, its sibling species, Chrysoperla downesi, is a habitat specialist, living all year only in coniferous trees, in the region of eastern North America where the two species are sympatric [38]. It can be inferred that C. downesi specializes on conifer-associated prey items, but the information about the variety and quality of prey utilized by these species is still scarce. Each species, however, exhibits phenotypic traits that suit it to its preferred habitat. First, both species display cryptic adult body coloration that is specific to their habitats: C. carnea is bright green in spring and summer and reddish brown in autumn and winter, whereas C. downesi is dark green all year [39]. Second, each species has life-history traits, such as the critical day length for diapause induction (e.g., [40]), that suit it to its voltinism. Chrysoperla carnea can utilize a variety of food sources depending on the season and produces several generations each year, whereas C. downesi is restricted to a single generation in the spring, probably because the availability of their prey in coniferous trees is more seasonal. As in the Chrysopa species complex, hybridization can occur between C. carnea and C. downesi, at least under laboratory conditions [39], which suggests that reproductive interference might occur in nature.

At present, the available laboratory findings on interspecific mating behavior are insufficient for the role of reproductive interference in these Chrysopa and Chrysoperla species to be evaluated. The hybrids were experimentally produced to clarify the existence of postmating reproductive isolation [34, 39] and to examine the genetic basis for the phenotypic traits [39, 40]. However, all possible behavioral mechanisms that can reduce adult reproductive success should be considered in investigations of the population dynamics of these two species pairs. For example, interspecific sexual harassment prior to interspecific copulation can reduce a female’s oviposition rate, longevity [41], and mating success [42], which might have population dynamics consequences and lead to subsequent niche separation. Thus, future studies of these sibling lacewing species should not only examine hybrid production but also quantify lifetime female reproductive success to evaluate the possibility that reproductive interference has driven the niche separation.

2.2. Sympatric Speciation through Host Race Formation

The difference in niche width between sympatric lacewings has been attributed to sympatric speciation [43, 44]. Sympatric speciation occurs when ecological adaptations to a different host (habitat) also produce nonrandom mating in a conspecific population, which eventually leads to a phenotypically and genetically divergent subpopulation, or species, even without geographical isolation [45]. In lacewings, courtship and copulation generally take place on plants that also serve as the feeding sites for both the adults and their offspring [46], indicating that adaptation and fidelity to a different habitat by each species may contribute to the development of reproductive isolation.

Tauber et al. [44] have argued that the comparative evidence from this Chrysopa species pair demonstrates for the first time a sympatric speciation mechanism in predatory insects. They proposed that a generalist ancestor of C. slossonae adopted a specialized foraging behavior and morphology as well as cryptic adult coloration that were well suited to a specific habitat. At the same time, a seasonal reproductive cycle, different from that of the ancestor, became established in the specialized subpopulation. This subpopulation then became reproductively isolated, both spatially and temporally, from the original population, which eventually led to its becoming a new specialist species by sympatric speciation.

Here we compare the two evolutionary scenarios, sympatric speciation and ecological specialization via reproductive interference, that have been proposed to explain the current niche partitioning between sibling lacewing species. The assumed order of ecological adaptation and genetic divergence differs between these two models [47] (Figure 1). In the sympatric speciation model, habitat differentiation is the primary impetus of genetic divergence in an originally uniform population. In this model, therefore, reproductive isolation is a by-product of adaptation to a particular habitat [48] (Figure 1(a)). In contrast, in the ecological specialization through reproductive interference model, genetic divergence (speciation) occurs allopatrically and reproductive interference promotes habitat differentiation when the two already distinct species come into secondary contact [49, 50]. In the reproductive interference mechanism of the second model, therefore, genetic incompatibility between the species is the cause of niche diversification (Figure 1(b)). Thus, the two models predict similar ecological consequences, namely, utilization of different habitats by two genetically close species, which prevents interspecific sexual interactions [47]. Accordingly, it is difficult to judge which model is applicable to a particular study system by examining current phenotypic traits and fitness.

The model of sympatric speciation in Chrysopa and Chrysoperla lacewings has some as yet unresolved problems. Most importantly, the generalist predator has been observed to visit the habitat on which its sibling species specializes, even though that habitat is less suitable for the generalist species. For example, C. quadripunctata is sometimes found on alder trees, where C. slossonae preys on the woolly alder aphid [34]. In addition, C. carnea visits not only meadows and deciduous trees but also coniferous trees, where C. downesi is found year-round [39]. This flexible behavior on the part of generalist predators may allow them to add a less suitable prey to their menu when the availability of more suitable prey is limited for some reason (e.g., [51]). Moreover, Henry [38] pointed out that the breeding periods of the sibling lacewing species greatly overlap, even though C. A. Tauber and M. J. Tauber [39] considered the seasonal mismatch to be an important mechanism of temporal reproductive isolation. Therefore, it is possible that observed niche overlap in space and time might be insufficient for sympatric genetic divergence to have occurred in the Chrysopa and Chrysoperla systems. Nevertheless, because genetic divergence sometimes occurs despite copulation and gene flow between taxa [5254], the apparent incomplete habitat isolation does not rule out sympatric speciation in these lacewings. At this time, therefore, it is not possible to rule out any particular view of speciation in the lacewing system.

2.3. Speciation through Sexual Selection

Henry [38] proposed that divergence in the courtship song, rather than habitat specialization, might be a primary driver of speciation in Chrysoperla lacewing species, whether allopatric or sympatric. In Chrysoperla, males produce substrate-transmitted calls on plants by vibrating their abdomen, and females respond to a potential partner by reciprocal signaling according to their own mating preference [55]. Henry [38] proposed that, based on the diversified and complex songs among close relatives, differentiation of courtship signals among populations should create ethological barriers to hybridization, thus accelerating the rate at which complete speciation can be achieved.

However, there is at least one problem with speciation through sexual selection. Chrysoperla females sometimes respond to heterospecific mating calls [55], which indicates that negative interspecific mating interactions can occur. In fact, under laboratory conditions copulation and hybridization can occur among populations (species) having different courtship songs [39]. These findings suggest that an ethological barrier alone might not ensure premating reproductive isolation.

This weak point of the speciation through sexual selection model (i.e., inevitable heterospecific mating interactions), however, supports the reproductive interference model. In general, the maintenance of incomplete species recognition abilities can be interpreted as a consequence of adaptive decision-making in reproductive animals. When the signals of high-quality conspecifics resemble those of heterospecifics, a trade-off is likely to exist between species recognition and mate-quality recognition within a species [56]. In this situation, strict species discrimination skill is unlikely to be maintained because individuals may also lose opportunities for acquiring high-quality conspecific mates. As a result, promiscuous mating occurs between the species. By this logic, reproductive interference is likely to be common in sibling species (because their mating signals are likely to be similar), and the reproductive interference might drive niche partitioning.

3. A Case Study in Harmonia

3.1. A Brief History of Studies of Harmonia Sibling Species

The multicolored Asian ladybird Harmonia axyridis is a common aphidophagous predator originally distributed in Russia, China, the Korean Peninsula, and Japan [24]. Because of its extreme polymorphism with regard to elytral color, H. axyridis has been extensively used for genetic studies. Theodosius Dobzhansky famously discovered geographic variation in the frequency of elytral color morphs across Siberia and East Asia [57]. Subsequently, extensive surveys established that elytral color morph frequencies change along a geographic (latitudinal) cline in Japan, where red (nonmelanistic) morphs at higher latitudes are gradually displaced by black (melanistic) morphs as latitude decreases [58]. Then, Hosino, who worked at an agricultural high school in Sanage, central Japan, discovered that elytral color morph frequencies were slightly different between Harmonia ladybirds collected from Japanese red pine (Pinus densiflora) and those collected from agricultural crops such as wheat, pear, and peach [59]. In addition, he reported that a ridge at the tip of the elytron, which is a genetic character exhibited by approximately 40% of H. axyridis individuals in the Sanage population, was seldom observed in individuals collected from pine trees [59]. For a few decades, Japanese entomologists were unable to account for these strange patterns.

Eventually, Hiroyuki Sasaji demonstrated the existence of a cryptic Harmonia species. He did so by experimentally demonstrating that a clear postmating reproductive barrier existed between H. axyridis individuals and those found on pine trees, even though they were observed to copulate together and produce eggs [60]. He also showed that larval coloration was distinct between these groups of individuals [61] (Figure 2), even though the adults were difficult to distinguish by their external and genital morphology [60], and, furthermore, he showed that the characteristic ridge at the tip of the elytra of H. axyridis adults never emerged in the pine-associated individuals [60]. On the basis of these findings, he concluded that the pine-associated individuals belonged to a cryptic species that could be clearly distinguished from H. axyridis by its postreproductive isolation and some phenotypic traits.

However, when Sasaji proposed denominating the new, pine-associated cryptic species, he discovered that Takizawa [62], at the dawn of insect taxonomy in Japan, had already described a species of ladybird collected from a pine tree and had named it Ptychanatis yedoensis (Ptychanatis is a currently unused synonym for genus Harmonia). Takizawa’s illustration of a larva of his new species (Figure 2(c)) clearly shows the larval characteristics of the cryptic species “discovered” by Sasaji. Sasaji [63], therefore, identified the cryptic species as Harmonia yedoensis Takizawa. In addition, deeply moved by Takizawa’s insight, Sasaji proposed Kurisaki-tento as the Japanese name of this species, because Takizawa had adopted the name “Kurisaki” when he became a Buddhist priest, and tento is the Japanese word for ladybird.

3.2. Host Specialization in H. yedoensis

The existence of these two species, H. yedoensis and H. axyridis, living in close proximity raises some questions; namely, what determines their habitat ranges? And how did two such phylogenetically close species with such a close morphological resemblance, at least in adults, come to occupy dramatically different niches? Harmonia axyridis is a truly generalist predator found in many different habitats, including scrub, orchards, grasslands, and coniferous and deciduous woodlands (e.g., [51]). In contrast, H. yedoensis is a specialist in central Japan that breeds exclusively on pine trees, where its larvae are regarded to prey mainly on the giant pine aphid (Cinara pini) [24] (Figure 3(a)). Because aphid colonies are spatially heterogeneous and temporally variable in terms of both quality and quantity, a generalist strategy that allows the ladybird to utilize multiple food patches has obvious advantages for maximizing lifetime fitness [51], as it does for other generalist insect species (e.g., [6466]). Therefore, how to explain host specialization, a strategy that appears to forego potentially available resources, has been an interesting problem in ecology and evolutionary biology.

It is clear from field surveys that H. yedoensis is much less abundant than H. axyridis. One result of this situation is that much more effort is required to collect enough H. yedoensis individuals to carry out research projects, compared with that required for H. axyridis. The rarity of H. yedoensis might be at least partly due to the low abundance of its prey item on pine trees. Although the yellow egg batches of H. yedoensis are easy to locate on dead or live pine tree needles, it is hard to locate nearby colonies of the giant pine aphid, which are usually very small (S. Noriyuki personal observation), unlike those of other aphid species, which tend to congregate on young shoots of their host plants. Syunsuke Shimamoto, a junior high school student studying the ecology of H. yedoensis, once remarked to one of the authors (S. Noriyuki), “I wonder why H. yedoensis larvae stay where there is so little food to eat, while H. axyridis larvae are on plants with lots of aphids.” We, too, ask this simple question.

Not only are giant pine aphids rare, they are also so mobile that ladybirds seem to have difficulty capturing them. In contrast to most arboreal aphids, which walk very slowly at best, the giant pine aphid has long legs and can walk fast (Figure 3(b)). We therefore experimentally evaluated the prey capture performance of H. yedoensis and H. axyridis hatchlings against the giant pine aphid and other aphid species [9]. As predicted, it was hard for H. axyridis hatchlings to capture the giant pine aphids. Interestingly, however, although H. yedoensis apparently specializes on the giant pine aphid, this species is not the easiest prey for H. yedoensis hatchlings to catch. Rather, we found that H. yedoensis hatchlings could capture other aphid species more easily, even though they never encounter them or prey upon them in nature. These results led us to hypothesize that H. yedoensis mothers deliberately choose a host that is less suitable for the prey capture activities of their offspring even though other suitable hosts are available in the region.

Harmonia yedoensis mothers help their offspring cope with their elusive aphid prey by investing more maternal resources in each hatchling. In aphidophagous ladybirds, the hatch rate of an egg batch is often less than 100% and the newly hatched larvae consume the unhatched sibling eggs. These unhatched eggs are not just an unavoidable developmental side effect; they are an adaptive maternal strategy known as trophic egg provisioning, because ladybird mothers are able to control the proportion of unhatched eggs in a clutch according to food availability [67]. In fact, trophic egg provisioning is more intense in H. yedoensis than in H. axyridis [26, 28]. Moreover, artificial manipulation of the trophic egg number provided to each hatchling has revealed that trophic egg consumption enhances the prey capture performance of H. yedoensis hatchlings against the giant pine aphid [9]. In addition, H. yedoensis eggs are significantly larger than H. axyridis eggs [26, 28]. Consequently, the amount of maternal resource invested per offspring is much higher in H. yedoensis than in H. axyridis. Importantly, however, the number of offspring that a mother can produce in her lifetime is lower in H. yedoensis than in H. axyridis, because of the trade-off between offspring size and number in the similar-sized species [26]. Therefore, H. yedoensis mothers apparently sacrifice reproductive success in order to supply more resources to each of her larvae, which are obligated to prey on the elusive aphid.

Furthermore, we found by experiment that the giant pine aphid is nutritionally less suitable for the larval development of ladybirds. We froze some aphids, to exclude the effects of prey mobility on larval performance, and then fed them to larvae of H. yedoensis and of H. axyridis to evaluate their intrinsic suitability as food [27]. The result showed that in both ladybird species larval developmental performance was lower in larvae fed the giant pine aphid than in those fed the other prey species. Although we have not yet examined the proximate mechanism for the lower developmental performance, secondary compounds derived from the pine trees and stored in the aphid’s body (see [68, 69]), or simply an insufficient nutrient content, might be responsible.

Taken together, these findings show that the giant pine aphid is poor prey for both H. yedoensis and H. axyridis larvae with respect to abundance, capture difficulty, and its intrinsic suitability as food. Therefore, we conclude that food quality per se does not determine the food range of H. yedoensis. In contrast to the rarity and niche specialization of H. yedoensis, H. axyridis is abundant and utilizes various preferred prey species. These differences suggest that negative interactions between H. yedoensis and H. axyridis might greatly influence the food utilization of H. yedoensis in nature.

3.3. Reproductive Interference between the Two Harmonia Species

The results summarized in Section 3.2 led us to hypothesize that reproductive interference by H. axyridis against H. yedoensis is responsible for the specialization on less suitable prey by H. yedoensis. When H. yedoensis was first identified as a cryptic Harmonia species, it was already known that although interspecific copulation occurred, a postmating reproductive barrier existed between H. yedoensis and H. axyridis [60, 70]. Sasaji [60] also showed that the male genitalia were very similar between the two species with respect to both size and morphology, so there was no structural barrier to interspecific copulation. To examine the possible influence of reproductive interference on niche partitioning, it was necessary to quantify the effects of interspecific mating behaviors on reproductive success, which might affect the population dynamics of the two species.

Therefore, we performed some laboratory experiments to examine interspecific mating behaviors and their subsequent effect on reproductive success [29]. Both H. yedoensis and H. axyridis males attempted to mate with both conspecific and heterospecific females, and these attempts often resulted in interspecific copulation lasting a few hours or more. Females of both species can lay eggs even after interspecific copulation, but no viable offspring hatched. However, H. axyridis males tended to choose conspecific over heterospecific females, whereas H. yedoensis males promiscuously copulated with both conspecific and heterospecific females. We are unable to account for this difference in discrimination skill between H. yedoensis and H. axyridis, although chemical signals based on cuticular hydrocarbons or visual cues related to body size and color might be important for the interspecific communications. But, probably owing to this difference in discrimination skill, most H. yedoensis individuals failed to copulate with conspecifics, especially when the density of H. axyridis was high in the experimental arena. By contrast, most H. axyridis individuals successfully copulated with conspecifics regardless of whether H. yedoensis was present [29]. Here, it is important to note that the fitness reduction at the mating stage is asymmetric between these two species; that is, H. yedoensis is inferior to H. axyridis in terms of reproductive interference.

These results support the hypothesis that reproductive interference from H. axyridis has led to host specialization in H. yedoensis. In nature, H. yedoensis might suffer from a considerable fitness reduction if they utilize less elusive aphids in a patch occupied by H. axyridis. By specializing to the habitat of the giant pine aphid, which is especially elusive prey for H. axyridis, H. yedoensis can mitigate the cost of reproductive interference. Therefore, we argue that H. yedoensis mothers are obligated to choose lower quality prey for offspring performance where the ranges of these sibling species overlap.

Biogeographical distribution patterns also suggest that reproductive interference plays a nonnegligible role in determining the food range of H. yedoensis. In the Ryukyu Islands, southern Japan, where H. axyridis is not found (Figure 4), H. yedoensis often visits various deciduous tree species [71, 72] (S. Noriyuki, personal observation). In contrast, on Hokkaido Island, northern Japan, where H. yedoensis does not occur, H. axyridis is a generalist predator, just as it is in regions of central Japan where the two species are sympatric. These patterns suggest that H. axyridis utilizes various preferred prey species without regard to the presence of H. yedoensis, whereas the negative impact of reproductive interference of H. axyridis on H. yedoensis forces the latter to become a specialist predator on less preferred prey where the two species coexist.

Comparison of Harmonia (ladybirds) and Chrysopa (lacewings) (Table 1) reveals many similarities in the relationships between these sympatric sibling species pairs. Both specialist species can potentially develop on a variety of food sources that are in nature utilized exclusively by their generalist counterpart. Moreover, the specialist species have morphological, behavioral, and physiological adaptations that enable them to handle their less suitable prey. Furthermore, interspecific copulation can occur between members of each pair, at least under laboratory conditions. Given these observations, it would be interesting to explore whether the mechanism of ecological specialization/niche differentiation is the same in these two aphidophagous taxa.

Of course, alternative interspecific interactions, not necessarily mutually exclusive with reproductive interference, must also be considered as possible mechanisms of the ecological specialization of H. yedoensis. In particular, intraguild predation may drive host specialization in H. yedoensis, because H. axyridis is known to be a strong intraguild predator against other ladybird species. However, the results of a laboratory experiment suggest that at the larval stage H. axyridis might not be dominant over H. yedoensis as an intraguild predator [73]. It is probable that the body size of the individual larvae, rather than their species identity, mainly determines the winner in interactions between these two predatory larvae, because the final instar larvae of both species feed on earlier instar larvae of the other species when they are put together in a Petri dish [73]. We think it is likely that intraguild predation and reproductive interference jointly contribute to niche partitioning between these two ladybird species, because both mechanisms can destabilize species coexistence locally. However, the direction of the food specialization, with H. yedoensis feeding on more elusive and H. axyridis feeding on less elusive prey, might be determined by the asymmetry of the reproductive interference. Nevertheless, the combined effect of predation and competition on individual fitness and species coexistence is predicted by mathematical models to produce complex outcomes [74, 75]. We therefore think that it would be very interesting to see how the incorporation of reproductive interference in a trophic module along with intraguild predation would alter population dynamics and community structure.

Now, about 100 years after Takizawa first described a new Harmonia ladybird species and several decades since its rediscovery by Sasaji, it is becoming clear that the mechanism of ecological specialization and generalization in H. yedoensis and H. axyridis involves negative interspecific interactions. This case study illustrates how the integration of taxonomy, ecology, and evolution can deepen our understanding of natural history and the great principles that underpin biological communities.

4. Other Specialist Ladybirds

Reproductive interference may have led to niche partitioning not only in Harmonia species but also in predatory ladybirds belonging to other genera. Therefore, we picked congeneric species pairs known to show different food (habitat) ranges or types in the same geographical region and examined the incidence of interspecific mating interactions in these congeneric pairs. Then we investigated conifer- (pine) associated specialist ladybirds in Britain and Japan to assess the generality of our hypothesis that reproductive interference can lead to niche partitioning in closely related species.

4.1. Niche Partitioning between Congeneric Ladybird Species Pairs

To examine niche partitioning in other ladybirds, we selected genera that contain at least two species with different food (habitat) types in the same region of Britain or Japan (Table 2). In both these countries, the natural history of ladybirds has been well studied by professional and amateur entomologists. Because, in this paper, our focus is on aphidophagous insects, we exclude phytophagous and mycophagous ladybirds from the list in the table, but this exclusion does not mean that reproductive interference might not be important in these functional groups [47]. We identified numerous closely related species in various ladybird genera that exhibited niche partitioning (Table 2). For example, Pseudoscymnus sylvaticus mothers lay their eggs on galls of certain aphid species (Tuberocephalus sasakii, which forms galls on cherry tree leaves, or Ceratovacuna nekoashi, which forms galls on fruits of Japanese snowbell, Styrax japonica), and the hatched larvae forage exclusively on the aphids in the gall. In contrast, a sympatric congener, Pseudoscymnus pilicrepus, utilizes the eusocial aphid Colophina arma on the subshrub Clematis stans [24]. At present, the drivers of this niche partitioning in regionally coexisting predatory ladybirds are not well understood. Here, we focus on Adalia and Coccinella species pairs, because each of these genera contains generalist and specialist species pairs distributed sympatrically in Britain and other parts of Europe and considerable information is available about their ecology.

Adalia bipunctata is one of the most common ladybird species in Europe and it occupies a great variety of habitats, including woodlands, scrub, and grassland. By contrast, its sibling species, Adalia decempunctata, is more habitat-specific; it is found mainly on deciduous trees and in hedgerows [7678]. Because in A. decempunctata larvae can develop and forage on various aphid species [79] and mothers produce more eggs when fed on an aphid that is not regularly used in the field [80], this habitat restriction cannot be explained solely by food suitability. Moreover, Sloggett and Majerus [78] have shown that the prevalence of parasitoids is similar between A. bipunctata and A. decempunctata, suggesting that the habitat specialization of the latter is not driven by a need for a habitat with low predation risk (i.e., an enemy-free space [81]).

Under laboratory conditions, interspecific copulation occurs between A. bipunctata and A. decempunctata, and the resulting sterile hybrids occasionally develop into adults [82]. Thus, interspecific mating interactions should reduce their reproductive success when the two species occupy the same local patch. During fieldwork in the Netherlands, Brakefield [76] importantly observed that the principal spring mating habitats of A. bipunctata were shrubs, whereas he found almost all mating A. decempunctata on trees. This use of different mating habitats suggests that past negative mating interactions might have caused small-spatial-scale divergence of reproductive sites in these species. It would be interesting to explore this possibility empirically in the future.

Coccinella septempunctata is also very common in gardens and agroecosystems, but it is also found in heathland, scrub, and coniferous and deciduous woodlands. In contrast, each of the four other Coccinella species in Britain specializes in a different niche (Table 2). In particular, Coccinella magnifica is a myrmecophile that lives only near ant nests and it preys on aphids that are attended by the ants [83]. Obviously, foraging on ant-attended aphids is costly for C. magnifica adults and larvae, because the ants often attack the foraging ladybirds [84, 85], suggesting that the ant attendance itself cannot explain the niche differentiation between C. septempunctata and C. magnifica. Importantly, C. magnifica can be successfully reared in the laboratory from adult to adult by feeding nonattended aphid species such as Acyrthosiphon pisum and Aphis fabae (i.e., alternative foods). Thus, its potential dietary breadth appears to be similar to that of its generalist congener C. septempunctata [83]. Even though interspecific mating behaviors have not been thoroughly studied in these two ladybird species, the adult body size and color pattern are very similar between them. It might be illuminating, therefore, to consider interspecific mating interactions in future investigations of niche partitioning among Coccinella species. For example, to test the hypothesis that reproductive interference prevents a species from occupying its fundamental niche and restricts it to a narrower realized niche, experiments should be performed to examine whether reproductive interference occurs in these species and whether the cost of the interaction is incurred more by C. magnifica than by C. septempunctata.

4.2. Conifer Specialists

It is well known that some species of predatory ladybirds utilize specialized habitats on pine trees and other conifers (Table 3) and that this specific specialization has independently evolved several times in the ladybird lineage [86]. Not all conifer-specialist species coexist in the same region with a congeneric relative (Table 3), which suggests that, assuming that reproductive interference is mainly likely to occur among phylogenetically close species, these specializations might have occurred through mechanisms other than reproductive interference. Some conifer specialists, however, do have sympatrically distributed generalist congeners (Table 3). In Britain, the habitats of Scymnus nigrinus and Scymnus suturalis are restricted to conifer trees, whereas a congener, Scymnus schmidti, utilizes a wide range of habitat types [35]. A similar pattern has been observed in Japan in specialist and generalist species belonging to genus Oenopia [24]. Therefore, we can hypothesize that, as in Harmonia species in Japan, this specialization to conifers might have been driven by reproductive interference from sympatric congeners. To test this hypothesis, it will be necessary to quantify the cost of interspecific mating interactions as well as the quality of the conifer habitat with respect to larval and adult development. It would be interesting to know if reproductive interference has caused some species to specialize to conifers more than once in the ladybird phylogeny.

5. Aphidophagous Guilds

In this section we consider intraguild predation as an alternative mechanism of the niche partitioning of aphidophagous insects. Multiple species of aphidophagous insect predators, as well as parasitoids that attack aphids, often coexist despite the occurrence of intraguild predation. Osawa [87], for example, studied aphidophagous ladybirds and hoverflies preying on eight species of aphids infecting seven species of deciduous trees in the Botanical Garden of Kyoto University, central Japan. He collected five aphidophagous ladybird species and eight hoverfly species (see also [88]). Notably, both field [89] and laboratory studies [90, 91] have shown that four of the collected ladybird species (except Scymnus posticalis) engage in intraguild predation. Harmonia axyridis, in particular, is an intraguild predator that preys predominantly on other ladybird species [89, 92, 93]. Nevertheless, the abundance of H. axyridis is not negatively correlated with the abundance of the other four ladybird species in this area [88]. Thus, some mechanism must exist that allows these ladybird species to coexist in the same patch.

Similarly, in hoverfly communities generalist predators often use many of the same prey species [94, 95]. Moreover, aphidophagous ladybirds and hoverflies often coexist where their distributions overlap [88, 89, 96], even though predatory ladybirds often attack hoverflies with which they compete for the shared aphid resource (reviewed by [97]). Thus, in aphidophagous communities, predation and resource depletion do not necessarily result in strict niche partitioning.

If parasitoids that attack aphids are assumed to belong to the same guild as aphidophagous insects [13], then aphidophagous communities become even more species-rich and complex. In this case, intraguild predation includes predators such as ladybirds that feed on parasitized aphids (e.g., aphids that have been mummified), as well as different parasitoid species that, when parasitizing the same aphid, attack each other. Despite this complex food web structure, multiple parasitoid species are known to utilize the same aphid colonies simultaneously. For example, Müller et al. [98] reported that, in Silwood Park, southern England, some species of aphids were infected with multiple parasitoid species. Moreover, Osawa [87] identified various primary and secondary parasitoids on aphids on which both predatory ladybirds and hoverflies intensively foraged. These observations suggest that some factor must exist that permits the coexistence of multiple species in aphidophagous communities with multiple and complex prey-predator links.

The conditions that allow the stable coexistence of a predator and its prey, together with a shared prey species, have been extensively analyzed by mathematical models, as well as by some empirical studies. One classic model [12] predicts that stable coexistence should be possible if a trade-off exists between resource consumption and predation ability, but only under a limited range of environmental conditions. In addition, mechanisms such as a temporary refuges for prey [99], optimal predator foraging behavior [100, 101], and intense intraspecific interactions [102] have been proposed to allow species coexistence despite the occurrence of intraguild predation. These mechanisms likely explain the seemingly paradoxical robust persistence of many natural intraguild predation systems.

The local coexistence of phylogenetically closely related species, however, may still be inhibited. In Japan, two sibling aphidophagous ladybirds, Propylea japonica and Propylea quatuordecimpunctata, are distributed parapatrically; the former lives at low elevation and the latter at high elevation, but within a narrow zone they coexist [24]. In Britain, Coccinella septempunctata utilizes a wide range of prey types and habitats, whereas Coccinella hieroglyphica is mainly found on heather, where it forages on the heather aphid, Aphis callunae, and on larvae of the heather leaf beetle, Lochmaea suturalis [35]. It has often been demonstrated that in both animal and plant communities local assemblages tend to be composed of phylogenetically distant species (e.g., [103105]), probably owing to exclusive interactions among closely related species. Therefore, to understand the observed spatial distributions of aphidophagous insects, in addition to intraguild predation, some factor that is tightly connected to phylogenetic closeness should be considered.

Interspecific mating behavior is more likely to occur among phylogenetically close species because they often share similar mating signals [56, 106]. The resulting, often costly mating interactions can lead to local species extinction, which in turn can cause diversification of niche use and resource-use traits [49]. In contrast, phylogenetically distant species, which generally have distinctive, easily discriminated mating signals, focus their mating efforts on high-quality conspecifics even when their ecological niches largely overlap. As a result, distantly related species can stably coexist within a patch [49]. These observations suggest that reproductive interference might be a mechanism of niche diversification in closely related species.

6. Conclusion

We have shown that the effect of reproductive interference on local population dynamics has probably been underestimated, in contrast to the intraguild predation mechanism, which is specific to the predator community. Therefore, to evaluate the validity and generality of the reproductive interference mechanism, more empirical studies, and extensions of published studies, of various predator species are needed. We have suggested some possible approaches earlier in this paper. Moreover, the relative importance and synergistic effects of predation, exploitative competition, and reproductive interference are largely unknown. Because these mechanisms of species exclusion are not mutually exclusive, the impact of each individually, and of their interactions, on niche partitioning needs to be quantified. A theoretical research priority is the incorporation of reproductive interference into the intraguild predation model to examine how it changes the population dynamics. Together, such empirical and theoretical efforts will enable us to evaluate whether niche partitioning in predators is explicable mainly by reproductive interference, a mechanism that occurs in several functional groups, including herbivores [42, 50] and plants [107], or by a specific mechanism that is unique to carnivores such as intraguild predation.

At present, it cannot be conclusively decided whether the observed niche partitioning between phylogenetically close species has been caused by sympatric speciation or reproductive interference after allopatric speciation, because it is difficult to distinguish between these two evolutionary scenarios by examining current phenotypic traits and fitness. Thus, a fruitful approach might be to identify the locality of the origin of the species by appropriate techniques such as phylogeography and genomics. Sympatric speciation has been frequently assumed to explain niche partitioning among phylogenetically similar relatives such as cryptic species complexes [108], but compelling evidence that two sympatrically distributed species (races) have diverged is still lacking [109]. A theoretical basis now exists for ecological character displacement via reproductive interference [49]. Incorporation of reproductive interference into existing models should make it possible to design more rigorous tests and to further refine the models and thus improve their skills in reconstructing the natural world.

Competing Interests

The authors declare that there is no conflict of interests regarding the publication of this paper.


The authors thank Yasuko Kawakami for information on Japanese ladybirds and Jun-Ya Ide for useful comments on the manuscript. This study was funded by support from the Japan Society for the Promotion of Science to Suzuki Noriyuki (no. 24-3112, no. 26840137, and no. 26650154) and to Naoya Osawa (no. 20405047 and no. 24255013).