Introduction

Predatory mites are economically important because of their potential to control mite pest populations and minimize associated economic injury. Among various groups of predators, phytoseiid mites are known to prey mostly on members of the families Tetranychidae (spider mites), Eriophyidae, Tenuipalpidae and Tarsonemidae. They are thus commonly used as natural control agents of spider mites in agriculture (McMurtry et al. 2013). Although biological control has minimal environmental impact compared with the impact of pesticides, competitive displacement of one predatory species by another might result in poorer control of a pest species (Palevsky et al. 2013). The elimination of a formerly established species from a habitat as a result of direct or indirect competitive interactions with a new candidate species is generally referred to as species displacement (DeBach 1966; Reitz and Trumble 2002). Several factors can contribute to species displacement including resource preemption and degradation, competition from other species (Braks et al. 2004; Reitz and Trumble 2002), reproductive interference (Gröning et al. 2007; Hochkirch et al. 2007; Suzuki et al. 2012), intraguild predation (Gotoh et al. 2014), differences in fecundity and foraging ability (Reitz and Trumble 2002) and differences in insecticide susceptibility (Gao et al. 2012).

Several studies have demonstrated that phytoseiid species have cannibalistic and intraguild predatory behaviors (Schausberger 2003; Gotoh et al. 2014; Zhang et al. 2014; Ji et al. 2015). However, it is often difficult to determine whether strong competition and predation result in species displacement. The population densities of some phytoseiid predators [Phytoseiulus persimilis Athias-Henriot and Neoseiulus californicus (McGregor)] remain stable when they are released alone, but decline when they are released in combination with another phytoseiid predator due to differences of intraguild predation, and differences in cannibalizing severity on conspecifics and heterospecifics (Schausberger and Walzer 2001). In order to predict whether one phytoseiid predator would displace another phytoseiid predator, it is necessary to examine how anthropogenic and other factors affect them differentially.

In Japan, the phytoseiid predatory mite Neoseiulus womersleyi (Schicha) used to be the dominant species among predatory mites in fruit-tree orchards. Populations of the predatory mite N. californicus have increased in central and southwestern Japan since the 1990s, both in size and distribution. Consequently, N. californicus has become the dominant species in conventionally controlled orchards (Wakabayashi 2000; Amano 2001; Kishimoto 2002). Some growers may have purchased a commercial strain of N. californicus (e.g., SPICAL EX®) that had been imported to Japan. The commercial strain was first registered in Japan in 2003 for use in only greenhouses, but in and after 2008 it was used in fruit tree orchards and tea plantations. However, the commercial strain was introduced after the displacement of N. womersleyi in fruit-tree orchards, and thus had no role in the displacement. Recently, N. californicus was found to eliminate N. womersleyi under laboratory conditions, and the cause appeared to be asymmetrical intraguild predation (Gotoh et al. 2014). It was previously found that N. californicus competes very strongly with N. womersleyi, and there might be a possibility of complete species displacement in fruit-tree orchards in Japan (Gotoh et al. 2014).

A combination of various biological mechanisms or anthropogenic factors, including differential reproductive success or susceptibility to pesticides, might influence species displacement (Kishimoto 2002). Moreover, it was evident that the suppression of natural enemies by application of a pesticide resulted in an outbreak of spider mites (Szczepaniec et al. 2011). However, the influence of pesticides on the displacement of natural enemies remains unknown in Japanese fruit-tree orchards. To effectively use native natural enemies for pest control in Japanese fruit-tree orchards through integrated pest management (IPM) programs, it is important to elucidate the role of pesticides in species displacement. Amano et al. (2004) examined the effects of 22 pesticides to see whether they could be responsible for the displacement of N. womersleyi by N. californicus in commercial pear orchards. They found that the two species had very different susceptibilities to milbemectin and pyribaden, which were thus proposed as factors contributing to the displacement, although these two pesticides alone could not account for the displacement. However, their study did not examine the effects of the pesticides on immature stages and did not examine two commonly used pesticides, acetamiprid and imidacloprid. In the present study, we examined the effects of 21 pesticides that were used in Japanese fruit-tree orchards from 1980 to 2000 on both immature and adult stages. Some of these pesticides were not examined by Amano et al. (2004). In addition, in the absence of pesticides, we determined how low the NC:NW ratio had to be to prevent displacement of N. womersleyi by N. californicus.

Materials and methods

Rearing of predatory and spider mites

An N. womersleyi strain was established from mites collected from an apple orchard in Morioka (39°42′N–140°11′E), Iwate Prefecture, northeastern Japan on 1 October 1999, and an N. californicus strain was collected from a pear orchard in Ichikawa (35°28′N–140°47′E), Chiba Prefecture, central Japan on 17 August 1995. The mite strains were maintained on excised leaves (ca. 16 cm2) of common bean Phaseolus vulgaris L. that were infested with the two-spotted spider mite Tetranychus urticae Koch (green form). Tetranychus urticae was collected from a watermelon Citrullus lanatus (Thumb.) in Takikawa (43°33′N–141°54′E), Hokkaido, northern Japan on 16 July 2001. Each excised leaf was placed on a water-saturated polyurethane mat in a plastic cup (10 cm ø top × 8 cm ø bottom × 4 cm high) with a perforated lid at 25 ± 1 °C, 60–80 % RH and a 16:8 h light:dark (16L:8D) photoperiod. The leaves were replaced whenever they appeared to dry out or be damaged by feeding mites. The lid of each cup had a 50-mm-diameter hole covered with fine nylon mesh to allow ventilation (Gotoh et al. 2005).

Pesticide bioassays

We used 17 insecticides and four fungicides that were sprayed in Japanese fruit-tree orchards from 1980 to 2000 (Table 1). The chemicals were suspended in distilled water just before treatment. Distilled water was used as a control in all trials.

Table 1 Insecticides and fungicides used in this study on eggs and adult females of Neoseiulus womersleyi and N. californicus
Full size table

Suspensions of the chemicals were gently sprayed onto the leaf disc to result in approximately 4 mg chemical/cm2. The leaf discs contained either approximately 40 eggs laid over 24 h or 15 mated adult females (3–5 days old) of predatory mites using a rotary spray tower (Mizuho Scientific, Nagoya, Japan) at 266 hPa with four replicates each. The sprayed samples were allowed to dry in the shade and then were maintained at 25 ± 1 °C with a 16L:8D photoperiod. For the bioassays, the company recommended concentrations for each chemical were used (Table 2). To monitor the impact of pesticides, leaf discs with eggs were gently sprayed and then hatchlings were kept on the sprayed leaf discs until adulthood. Egg-to-adult mortality was then determined based on the number of hatched eggs and dead immature stages from hatchlings. Mortality was evaluated by gentle probing with a fine brush. For the adult bioassay, we counted the N. womersleyi and N californicus prior to and 2 days after application.

Table 2 Mean (±SE, after Abbott’s correction) mortality (%) of egg-to-adult-stage and adult female Neoseiulus womersleyi and N. californicus to 17 insecticides and four fungicides used in Japanese pear orchards from 1980 to 2000
Full size table

For experiments with mixed species, we did not examine chemicals that had similar harmful effects or no effect, or caused nearly 100 % mortality in N. womersleyi or N. californicus for egg-to-adult-stage individuals and adult females.

Single versus mixed species experiment

To determine whether N. californicus excludes N. womersleyi under laboratory conditions, 3- to 5-day-old mated females of each species were placed in each arena (Gotoh et al. 2014). An arena consisted of a 4 × 4 cm section of a bean leaf surrounded by a ‘fence’ of wet cotton (5-mm high), which prevented the mites from leaving. The bean leaf, including its petiole, was placed on wet cotton to keep it fresh. Subsequently, about 700 T. urticae prey individuals of all stages were placed in each arena. The leaf arenas were placed in individual plastic cups (12 cm ø top × 10 cm ø bottom × 6 cm high) and maintained at 25 °C, 60–80 % RH and a 16L:8D photoperiod. Under these conditions, the leaves remained fresh for at least 30 days.

For the initial release, each experimental arena received one or two mated adult females of N. californicus alone, and there were 10 replicates. For N. womersleyi, eight or nine mated adult females were released alone, and there were five replicates. In other arenas, we released both predators at NC:NW ratios of either 2:8 (2 N. californicus and 8 N. womersleyi adult females/leaf disc) or 1:9 (1 N. californicus and 9 N. womersleyi adult females/leaf disc), with five replicates per ratio. When the number of prey individuals in an arena fell below about 500, additional prey individuals from the stock culture were brushed onto the arena with a fine brush. The mites were observed every 3 days for 30 days. At each observation time, the adult females of each species were counted under a stereomicroscope. Adult females of the two species could be easily distinguished by the lengths of their dorsal setae.

Neonicotinoids application on mixed species

Chemicals were applied to combinations of N. californicus and N. womersleyi at an NC:NW ratio of 1:9 in an arena, with five replicates. The mites were observed at 3-day intervals for 30 days. Both pesticides were applied on day 9. At each observation time, the adult females of each species were counted under a stereomicroscope.

Data analysis

The mean numbers of adult females in mixed species arenas were compared with paired t tests. The percentage of dead mites was corrected using Abbott’s (1925) formula. The mortality of both species using different chemicals was compared using Fisher’s exact test or Chi square test. All tests were performed with IBM SPSS version 22 (IBM, Armonk, NY, USA).

Results

Pesticide bioassays

The bioassays indicated that N. womersleyi and N. californicus had significantly different responses to most of the chemicals in the egg-to-adult stage but had significantly different responses to only a few of the chemicals in the adult stage (Table 2). Four of the chemicals (acetamiprid, imidacloprid, milbemectin and silafluofen) had more harmful effects on N. womersleyi than N. californicus. However, silafluofen had similar negative effects on adult females of both species and milbemectin caused more than 99 and 98 % mortality of N. womersleyi in the egg-to-adult and adult stages, respectively. Thus, only acetamiprid and imidacloprid had substantially different effects on N. womersleyi and N. californicus. Both of them were more harmful to N. womersleyi than to N. californicus in egg-to-adult stage (Table 2). After acetamiprid application, egg mortality was 50.5 % for N. womersleyi and 0.8 % for N. californicus (data not shown), whereas after imidacloprid application, egg mortality was 37.8 % for N. womersleyi and 38.6 % for N. californicus (data not shown). Acetamiprid caused much higher mortality of eggs and egg-to-adult stage individuals in N. womersleyi than in N. californicus, while imidacloprid caused slightly higher mortality in egg-to-adult stage individuals in N. womersleyi (Table 2). However, alanycarb and thiometon had more severe effects on N. californicus than on N. womersleyi (Table 2).

Single and mixed species experiments in the absence of pesticides

Changes in the populations of N. californicus and N. womersleyi females when released alone or in combination are shown in Fig. 1. Initially, gravid females of a single species of predator mite were released into an arena containing T. urticae. For each predator species, the number of females (males were not counted) increased to about 100 in the 30-day experimental period, even though the starting numbers of N. californicus were lower (Fig. 1a–d).

Fig. 1
figure 1

Changes in the number of adult females per arena after the following releases of Neoseiulus californicus (NC) and N. womersleyi (NW). a 2 NC; b 1 NC; c 8 NW; d 9 NW; e 2 NC and 8 NW; f 1NC and 9 NW. The conditions were 25 °C and a 16L:8D photoperiod. Paired t test; ns: P > 0.05; *P < 0.05; **P < 0.01; ***P < 0.001. Vertical lines indicate standard deviation

Full size image

When the two predator species were introduced together in an NC:NW ratio of 2:8, N. californicus began to displace N. womersleyi after about 15 days (Fig. 1e), although the number of N. californicus females was less than when N. californicus was released alone (Fig. 1a, b). On the other hand, when the initial NC:NW ratio was 1:9, the N. californicus population grew more slowly than, and did not displace, the N. womersleyi population (Fig. 1f).

Effects of neonicotinoids on mixed species

Because the two neonicotinoid insecticides (acetamiprid and imidacloprid) were more harmful to N. womersleyi than N. californicus, their influence on species displacement was examined in mixed species arenas. Nine days after N. californicus and N. womersleyi were released in an NC:NW ratio of 1:9 (1 N. californicus and 9 N. womersleyi adult females/leaf disc), acetamiprid and imidacloprid were individually applied (Fig. 2a, b, respectively). Just before addition of the insecticides, the number of N. womersleyi adult females was higher than the number of N. californicus adult females, in agreement with the results shown in Fig. 1f. After application of the insecticides, the numbers of N. womersleyi adult females declined sharply, while the numbers of N. californicus adult females immediately increased (Fig. 2). These results are consistent with the lower mortalities the pesticides caused in immature N. californicus than in immature N. womersleyi (Table 2). Finally, N. womersleyi was nearly completely displaced by N. californicus.

Fig. 2
figure 2

Changes in the number of adult females per arena after application of acetamiprid (a) and imidacloprid (b) when Neoseiulus californicus and N. womersleyi were released in a combination of 1:9 ratio at 25 °C and a 16L: 8D photoperiod. Arrows indicate acetamiprid and imidacloprid applications. Paired t test; ns: P > 0.05; **P < 0.01; ***P < 0.001. Vertical lines indicate standard deviation

Full size image

Discussion

The present results clearly show that pesticides could accelerate the displacement of N. womersleyi by N. californicus even though, in the absence of pesticides, no displacement was observed when the two species were released at an NC:NW ratio of 1:9. This finding, based on laboratory experiments, suggests that the use of some pesticides, such as acetamiprid and imidacloprid, can promote species displacement of N. womersleyi by N. californicus in Japanese fruit-tree orchards.

Differential susceptibility to pesticides has previously been associated with changes in the demographics of arthropod pest complexes. For example, the spirea aphid Aphis spireacola Patch displaced Aphis pomi De Geer, the predominant aphid pest of apple (Malus domestica Borkh.) in eastern and northwestern USA (Hogmire et al. 1992; Brown et al. 1995; Lowery et al. 2006). Aphis spireacola was substantially less susceptible than A. pomi to a variety of commonly used insecticides, which influenced species displacement.

Similarly, displacement of the B biotype by the Q biotype of the whitefly Bemisia tabaci (Gennadius) was attributed to higher resistance of the Q biotype to insecticides, such as pyriproxyfen and some neonicotinoids (Horowitz et al. 2003, 2005; Chu et al. 2010; Dennehy et al. 2010). Similar observations have been made on the B and Q biotypes of B. tabaci in Japan (Tsueda and Tsuchida 2011) and in China (Rao et al. 2012). Differences in resistance to neonicotinoids was a major factor that affected species displacement of the B biotype by the Q biotype whitefly (Luo et al. 2010; Wang et al. 2010; Rao et al. 2012; Yuan et al. 2012; Sun et al. 2013). These observations indicate that insecticides shifted species competitive interactions effects in favor of the Q biotype across China (Sun et al. 2013). Likewise, the leafmining fly Liriomyza sativae (Blanchard) was displaced by Liriomyza trifolii (Burgess) in California (Parrella et al. 1984; Reitz and Trumble 2002) and in China (Gao et al. 2012) due to differential responses to pesticides.

Pesticides have different impacts on different species of spider mites and their predators. In the present study, the number of N. womersleyi adult females was higher before acetamiprid and imidacloprid application, whereas the proportion of N. californicus immediately increased after application (Fig. 2). Thus, the species displacement of N. womersleyi by N. californicus in Japanese fruit-tree orchards (Wakabayashi 2000; Amano et al. 2004; Gotoh et al. 2014) might have resulted from their differential susceptibilities to these neonicotinoids. An N. womersleyi population in several Japanese pear orchards was affected negatively by different pesticides including imidacloprid (Izawa et al. 2000). In the present study, the egg-to-adult stage of N. womersleyi was strongly affected by acetamiprid and imidacloprid. Although Amano et al. (2004) did not examine acetamiprid and imidacloprid, they reported that milbemectin and pyribaden were more lethal to adults of N. womersleyi than to adults of N. californicus. However, they were unable to determine to what degree these differences were responsible for the species displacement. In our study, milbemectin caused more than 90 % mortality in the egg-to-adult stages of both N. womersleyi and N. californicus. Because the mortalities were very high for both species, these pesticides probably did not affect the species displacement. In apple orchards in Washington State, the predatory mite Galendromus occidentalis (Nesbitt) outnumbered the predatory mite Zetzellia mali (Ewing; Beers et al. 1993), but after several treatments of acetamiprid, Z. mali outnumbered G. occidentalis (Beers et al. 2005). In the phytoseiids Amblydromella caudiglans (Schuster) and G. occidentalis, imidacloprid caused adult female mortalities of 85 and 78 %, respectively (Schmidt-Jeffris and Beers 2015). In G. occidentalis, acetamiprid and imidacloprid were moderately toxic to larvae and had moderate to severe effects on fecundity (Beers and Schmidt 2014). Thus, these chemicals are not only detrimental to adults but also to the fecundity and immature stages of phytoseiids. In this study, we used only one strain of each species, although strains of N. womersleyi and N. californicus were found to vary in susceptibility to pesticides (Amano et al. 2004; Mochizuki 1990). Therefore, in future studies, we plan to test susceptibilities to chemicals in different strains of these two species.

Neonicotinoids have been used for insect and mite pest management for approximately 25 years in Japan. The population density of N. californicus has been increasing since the 1990s (Amano 2001). Acetamiprid and imidacloprid were first registered in Japan in 1989 and 1992, respectively. According to the spray calendar maintained by Chiba Prefecture, farmers started using acetamiprid and imidacloprid in fruit-tree orchards in Chiba Prefecture, in 1993 and 1997, respectively. The spraying period roughly corresponds with the time of displacement of N. womersleyi by N. californicus. Furthermore, the laboratory bioassays showed that the two pesticides are less harmful to N. californicus than to N. womersleyi. The failure of N. californicus to displace N. womersleyi even when the NC:NW ratio at the time of release was 1:9, together with the observation that N. womersleyi was displaced after the application of acetamiprid and imidacloprid, suggest that these two pesticides caused the displacement of N. womersleyi. These results indicate that the different modes of action of pesticides at different life stages have a role in phytoseiid mite species displacement.

Because of the increasing use of natural enemies in IPM programs, growers must understand how the natural enemies are affected by pesticides. It is likely that spider mite pest management programs with neonicotinoids need to be adjusted, as N. californicus exhibits dominance over N. womersleyi in Japanese fruit-tree orchards and displacement might occur. Continued use of either acetamiprid or imidacloprid at the recommended dose might lead to complete displacement of N. womersleyi by N. californicus. On the other hand, alanycarb and thiometon were more lethal to N. californicus than to N. womersleyi (Table 2), so that continued use of these pesticides might cause N. californicus to completely disappear from fruit-tree orchards. The N. womersleyi populations in Japanese fruit-tree orchards might rebound if changes are made in the pesticides that are used. For example, diamides suppress the growth of N. californicus more than they suppress the growth of N. womersleyi (Hirotsuna Hoshi, pers. comm.). Finally, this research provides new insights into the side effects of acetamiprid and imidacloprid (i.e., causing predatory species displacement of N. womersleyi by N. californicus) and highlights the importance of considering these side effects in Japanese fruit-tree orchards.