Introduction

The Allee effect is a positive causal relationship between any component of fitness and population density or size (Stephens et al. 1999). Allee effects have been broadly recognized as an important ecological phenomenon in conservation biology (Courchamp et al. 1999; Stephens and Sutherland 1999), biological invasion (Taylor and Hastings 2005; Liebhold and Tobin 2008), and biological control introductions (Hopper and Roush 1993), because they can lead to the extinction of small or sparse populations. Various mechanisms have been suggested to generate Allee effects, including mate-finding difficulty, lack of predator satiation, and reduced foraging efficiency (Courchamp et al. 2008). However, though some species are unlikely to suffer from an Allee effect, other species with similar life history traits are subject to Allee effects (e.g., Myers et al. 1995; Drake 2005; Fauvergue et al. 2007). Thus, predicting which species will be subject to an Allee effect remains a challenge. It is possibly because some of Allee effects are caused by an as yet unknown mechanism.

We propose the novel hypothesis that reproductive interference can generate an Allee effect, owning to its positive frequency dependence. Reproductive interference is a negative interspecific sexual interaction that reduces the fitness of at least one of the species involved (Gröning and Hochkirch 2008; Burdfield-Steel and Shuker 2011). For example, when a male of species A mistakenly recognizes a female of species B as a conspecific and attempts to mate with her, the female might suffer from an energy or time loss, reduced opportunities to mate with conspecifics, gamete wastage, genital damage, or hybrid formation (Gröning and Hochkirch 2008). Notably, in terms of its effect on population dynamics, reproductive interference seems to show frequency dependence (Kuno 1992; Gröning and Hochkirch 2008; Takakura et al. 2009): the higher the frequency of species A, the greater the intensity of the reproductive interference on species B. If for a pair of species, reproductive interference is asymmetric (unidirectional) and the density of interfering species A is constant, the per capita level of interference with the interfered species B should become diluted as the density of species B increases. This dilution of reproductive interference should generate an Allee effect when it is accompanied by increased fitness of species B. Thus, reproductive interference may generate an Allee effect via a dilution effect, just as predation does (Gascoigne and Lipcius 2004).

To test this hypothesis, we examined the relationship between per capita fecundity and conspecific density when heterospecific males were present in a pair of bean weevil species, Callosobruchus chinensis (Linnaeus) and C. maculatus (Fabricius). Nishigaki (1963) measured copulation frequency in C. chinensis and reported that males had difficulty in finding mates. In typical laboratory conditions, however, both C. chinensis and C. maculatus show no positive relationship between population density and lifetime reproductive success in isolation from the other (e.g., Yoshida 1966). Kishi et al. (2009) showed that the fecundity of C. maculatus, but not that of C. chinensis, is asymmetrically reduced when heterospecific males are present; thus, they show asymmetric reproductive interference. We predict specifically that C. maculatus, the species more susceptible to reproductive interference, will suffer an Allee effect only when heterospecific males are present, and that C. chinensis, the non-susceptible species, suffers no Allee effect.

Materials and methods

We used the jC-F strain of C. chinensis (see Harano and Miyatake 2005) and the hQ strain of C. maculatus (see Miyatake and Matsumura 2004), both of which were derived from the stock cultures of Okayama University. We chose these strains, because Kishi et al. (2009) used these same strains to confirm unilateral reproductive interference by C. chinensis on C. maculatus. The stock cultures were maintained under laboratory conditions (25 °C, 50–60 % relative humidity, and 16L8D) and fed on adzuki beans, Vigna angularis (Willd.) cv. Dainagon. Mated females of C. chinensis and C. maculatus lay about 60 eggs on bean surfaces without feeding during their adult lifespan of around 10 days. Males of both species try to copulate with either heterospecific or conspecific females indiscriminately even when conspecific females are present (Kishi et al. 2009). Interspecific copulation is sometimes observed, but no hybrid has been reported (Yamane and Miyatake 2010).

Virgin males and females of C. chinensis and C. maculatus were prepared for our experiments as follows. About 40 g of adzuki beans were placed in each of two plastic Petri dishes (ϕ90 mm × 20 mm). Then, about 10 females of each species were randomly chosen from the stock cultures (200–400 individuals in each generation), and allowed to lay eggs on the beans in one of the dishes. Beans with one, two, or three eggs on them were selected and placed individually in wells of a 24-well plastic plate (each well, ϕ15.5 mm × 17.3 mm). Beans with four or more eggs were discarded to avoid potential overcrowding effects. Each well was checked every day for newly emerged virgin individuals (i.e., individuals that emerged singly or in a unisexual group).

The effect of C. chinensis males on per capita fecundity (i.e., average number of hatched eggs per female) of C. maculatus was investigated under different C. maculatus density regimes as follows. From one to five C. maculatus pairs, each consisting of one male and one female individual, were housed with three C. chinensis males throughout their lives in a plastic Petri dish (ϕ70 mm × 15 mm) containing 9.0 g of adzuki beans. Before the two species were put together, we allowed each C. maculatus female to copulate with a conspecific male to exclude the possibility that the presence of heterospecific males would lead to a potential mate-finding difficulty. We introduced only male heterospecifics, rather than heterospecific pairs, because we wished to investigate how reproductive interference by heterospecific males affected the fecundity of the C. maculatus females. As the control, we also examined the per capita fecundity of one to five pairs of C. maculatus housed without any C. chinensis males. After 20 days, by which time all females had died, we counted the number of hatched eggs in each experimental replication and control treatment (we performed 16 or 17 replications of each) in each density regime to determine the per capita fecundity of C. maculatus and the effect of reproductive interference by C. chinensis males. We similarly examined the effect of C. maculatus males on the relationship between fecundity and density in C. chinensis. All experiments were conducted at 25 °C, 50–60 % relative humidity, and 16L8D. We used analysis of covariance (ANCOVA) to examine the effects of the presence of heterospecific males, conspecific density, and their interactions on per capita fecundity. All statistical analyses were conducted with R software version 2.11.1 (R Development Core Team 2010). The significance level for the statistical tests was set at P = 0.05.

Results

The presence of C. chinensis males significantly reduced the per capita fecundity of C. maculatus, and the impact was greater when the density of conspecifics was lower (ANCOVA: C. chinensis males, F 1,165 = 130.05, P < 0.001; density, F 1,165 = 18.71, P < 0.001; interaction, F 1,165 = 16.59, P < 0.001; Fig. 1a). In contrast, in the absence of C. chinensis males, density and fecundity of C. maculatus were not significantly related (linear regression analysis: F 1,83 = 0.028, P = 0.87; Fig. 1a). The presence of C. maculatus males had no significant effect on per capita fecundity of C. chinensis, but C. chinensis density had a significant negative effect on the fecundity of that species (ANCOVA: C. maculatus males, F 1,164 = 0.56, P = 0.45; density, F 1,164 = 8.81, P = 0.003; interaction, F 1,164 = 2.79, P = 0.10; Fig. 1b).

Fig. 1
figure 1

Per capita fecundity of C. maculatus (a) and C. chinensis (b) in the absence (open circles) or presence (closed circles) of heterospecific males, plotted against population density. In a, the broken line and the solid line are regression lines for the data in the absence or presence of C. chinensis males, respectively. In b, the regression was calculated using all data, because the effect of C. maculatus males on C. chinensis fecundity was not significant

Full size image

Discussion

We suggest that reproductive interference by C. chinensis generated an Allee effect in C. maculatus. The presence of C. chinensis males reduced the per capita fecundity of C. maculatus (Fig. 1a), a result in accordance with the findings of a previous study that documented asymmetric reproductive interference by C. chinensis males on C. maculatus females (Kishi et al. 2009). A higher density of C. maculatus, however, mitigated the reduction in per capita fecundity, generating an Allee effect (Fig. 1a). Although the presence of C. chinensis males reduced the per capita fecundity of C. maculatus, in the absence of C. chinensis, the fecundity of C. maculatus did not depend on density (Fig. 1a). Thus, we can attribute the observed fecundity reduction in C. maculatus exclusively to the presence of C. chinensis males. Callosobruchus chinensis males attempt to mate with females of either species indiscriminately (Kishi et al. 2009). In this study, the relative abundance of C. chinensis males to C. maculatus decreased as C. maculatus density increased. A single C. maculatus female was, therefore, expected to experience fewer mating attempts by C. chinensis males as the density of C. maculatus increased, as a result of the dilution effect. If this held true, the hypothesized reduction in the number of mating attempts by C. chinensis males can explain the enhanced per capita fecundity of C. maculatus with increasing C. maculatus density (Fig. 1a). We also observed no Allee effect in C. chinensis, which shows little susceptibility to reproductive interference by C. maculatus, a result consistent with our hypothesis (Fig. 1b).

The effects of the presence of heterospecific males on per capita fecundity were asymmetric; although C. chinensis males reduced the fecundity of C. maculatus (Fig. 1a), C. maculatus males had no significant effect on the fecundity of C. chinensis (Fig. 1b). Interspecific copulation sometimes occurs between C. chinensis males and C. maculatus females, but it seldom occurs between C. maculatus males and C. chinensis females (Yamane and Miyatake 2010). Thus, it is likely that the interspecific copulation between C. chinensis males and C. maculatus females plays an important role in the asymmetric reduction in fecundity. Yamane and Miyatake (2010), however, suggested that neither sperm nor seminal fluid was transferred during interspecific copulation between C. chinensis males and C. maculatus females. In Callosobruchus, male genitalia have spines, which injure the genital tracts of the females during intraspecific copulation (e.g., Crudgington and Siva-Jothy 2000; Rönn et al. 2007). The genital spines of C. chinensis males may thus injure the genital tracts of C. maculatus females, and the putative injury might cause the fecundity reduction in C. maculatus. Another possibility is interspecific sexual harassment. Though males of both species attempt to mate with heterospecific females (Kishi et al. 2009), some behavioral differences between these species may cause asymmetric fecundity reduction. Further experiments are necessary to find out whether structural damage or behavioral interference is actually responsible for the asymmetric reproductive interference between C. chinensis and C. maculatus.

In our experiments, difficulty in finding mates was eliminated by allowing females to copulate with a conspecific male prior to experiments, and reproductive interference generated an Allee effect in C. maculatus even with the presence of conspecific sperm (Fig. 1a). However, reproductive interference may cause lack of conspecific sperm. For example, heterospecific matting attempts may interfere with conspecific copulations (Liu et al. 2007). Thus, without copulation with a conspecific male prior to interaction with heterospecifics, the Allee effect might be enhanced. The interaction between reproductive interference and conventional mechanisms of Allee effects is an important topic for further research.

It is not clear to what extent reproductive interference causes Allee effects in nature. The importance of reproductive interference as a mechanism generating an Allee effect depends on the generality of reproductive interference itself. The role of reproductive interference in population dynamics, however, has not been appreciated until recently (Gröning and Hochkirch 2008). Reproductive interference is especially likely when a biological invasion occurs (Burdfield-Steel and Shuker 2011), that is, when originally allopatric species begin to interact with each other. Biological invasions, therefore, may be one of the most promising settings in nature for finding an Allee effect generated by reproductive interference. On the other hand, several studies have documented reproductive interference between originally sympatric species (e.g., McLain and Pratt 1999; Hettyey and Pearman 2003; Van Gossum et al. 2007), suggesting that it has the potential to produce Allee effects in sympatric species as well.

Allee effects are increasingly appreciated for their importance in applied ecology problems, such as conservation and biological invasion (Taylor and Hastings 2005; Courchamp et al. 2008; Tobin et al. 2011). Here we showed that reproductive interference can indeed generate an Allee effect. If reproductive interference can generate an Allee effect in nature, then programs dealing with small or sparse populations need to examine interactions between closely related species, even those that do not share a limited resource.