Intraguild predation between Amblyseius swirskii and two native Chinese predatory mite species and their development on intraguild prey

Amblyseius swirskii, native to the east and southeast Mediterranean region, is a successful biological control agent of whiteflies. In this study, we investigated intraguild predations (IGP) between each stage of A. swirskii and each stage of two Phytoseiid species that occur in China, Amblyseius orientalis and Neoseiulus californicus. When there was no whitefly egg provided as the extraguild prey, IGP between A. swirskii and A. orientalis, and between A. swirskii and N. californicus, was observed in 10 and 20 out of 35 combinations, respectively. When IGP was observed, A. swirskii was the intraguild predator in 70% and 65% cases of A. orientalis and N. californicus predation, respectively. These results suggest that A. swirskii is a more aggressive intraguild predator compared to either A. orientalis or N. californicus. When whitefly eggs were provided as the extraguild prey, IGP between A. swirskii and N. californicus decreased greatly, but no significant decrease of IGP was observed between A. swirskii and A. orientalis. Amblyseius swirskii was able to complete development on both heterospecific predatory mites, and both heterospecific predatory mites completed their development on A. swirskii. Possible impacts that A. swirskii may have on local predatory mite populations in China are discussed.

the negative impact that introduced species have on local competitors 20 , and occurs prevalently in polyphagous phytoseiid mites 21,22 . Therefore, it is necessary to investigate IGP between A. swirskii and native Phytoseiid species to evaluate the environmental risks of introducing A. swirskii to China.
Kuhlmann et al. 23 proposed three principles in selecting native competitors for risk evaluation: (1) ecological similarities, (2) phylogenetic/taxonomic affinities, and (3) safeguard considerations 23 . In this study, Amblyseius orientalis and Neoseiulus californicus were selected mainly based on the first two principles. Amblyseius orientalis (Ehara) (Acari: Phytoseiidae) is a widely distributed native biological control agent of spider mites and whiteflies 24 . Neoseiulus californicus (McGregor) (Acari: Phytoseiidae) is a polyphagous species that is widely distributed worldwide. It co-occurs with A. swirskii in the Mediterranean region, and the first Chinese native population was discovered in Dinghushan, Guangdong Province in 2011 25 .
Lab experiments were conducted in 10 mm diameter arenas to investigate IGP between A. swirskii and A. orientalis on one hand, and between A. swirskii and N. californicus on the other hand. For each pair of species, we estimated IGP levels between each stage of A. swirskii and the native species (eggs, larvae, protonymphs, deutonymphs, female adults, and male adults), either without or with B. tabaci eggs provided as the extraguild prey. We also estimated the developmental duration and age specific survival of each predatory mite when it fed on its competitor as the intraguild prey.

Results
IGP without or with extraguild prey. A. swirskii vs. A. orientalis. In the absence of extraguild prey, IGP was observed in 10 combinations, among which A. swirskii acted as the intraguild predator in 7 combinations. Table 1 summarizes the IGP levels between the two species. When A. swirskii was the predator, adult females preyed on all immature stages of A. orientalis, adult males preyed on A. orientalis protonymphs, and nymphs preyed on A. orientalis eggs. In contrast, when A. orientalis was the predator, only adult females preyed on A. swirskii eggs and larvae, and deutonymphs preyed on A. swirskii eggs.
For the combinations with IGP observed, no significant decrease of IGP level was observed overall (t = 1.087, df = 9, P = 0.305) when extraguild prey was provided. Among all the combinations, significant differences of IGP levels were only detected in 2 combinations (A. swirskii adult female vs. A. orientalis eggs, and A. swirskii adult males vs. A. orientalis protonymph) ( Table 1).
A. swirskii vs. N. californicus. In the absence of extraguild prey, IGP was observed in 20 combinations, among which A. swirskii acted as the intraguild predator in 13 combinations. Table 2 summarizes the IGP levels between the two species. When A. swirskii was the predator, its adults preyed on all immature stages of N. californicus, except predations of N. californicus eggs by A. swirskii males was not observed. Nymphs preyed on all younger stages of N. californicus. When N. californicus was the predator, adult females preyed on all immature stages of A. swirskii. Deutonymphs preyed on A. swirskii eggs and larvae, and protonymphs preyed on A. swirskii eggs. These data suggest that A. swirskii consumes more extensive stages of intraguild prey than N. californicus. However,, the IGP level between the two species decreased significantly overall (t = 9.538, df = 19, P < 0.001) when extraguild prey was provided. One exception was for the combination of A. swirskii adult female and N. californicus eggs ( Table 2). These results suggest that extraguild prey may be an important factor that affects the intraguild predation between A. swirskii and N. californicus.

Discussion
In the present study, intraguild predations were observed between A. swirskii and A. orientalis/N. californicus. In each combination, both species could serve as the intraguild predator, and was able to complete its development on the intraguild prey. In most cases where IGP was observed, intraguild predator adults or nymphs preyed on younger stages of the intraguild prey, but adult female of A. swirskii also preyed on adult male of N. californicus. Amblyseius swirskii was the intraguild predator in 70% and 65% IGP cases, when it co-existed with A. orientalis and N. californicus, respectively. In addition, A. swirskii was capable of preying on nymphs of A. orientalis 22 reviewed studies on cannibalism between phytoseiid mites and concluded that the degrees of diet specialization of predatory mites are reflected in their performances in cannibalism 22 . In addition, generalist predators have higher preferences for heterospecific immatures than conspecific immatures, while often not for specialist predators 28 . There is an increased positive correlation between the competitiveness of predatory mites and their prey range. McMurtry and Croft 31 categorized Phytoseiid mites into 4 categories based on their prey. According to their classification, A. swirskii belongs to type III (generalist predators), and N. californicus belongs to type II (selective predators of tetranychid mites) although it was also questioned whether this species should be classified as a member of type III 31,32 . Amblyseius orientalis is a native biological control agent in China. Similar to N. californicus, it used to be categorized as type II and was widely used in spider mite control in China 33,34 , but recent data indicate that this species is more appropriately categorized as type III 24,35 . Our interpretation is that type II and type III include phytoseiid species with prey ranges increasing continuously, and the border between the two categories is actually vague. Both N. californicus and A. orientalis are critical species ranging between type II and type III, and their prey ranges are narrower than A. swirskii, which is consistent with our results suggesting A. swirskii is a stronger intraguild predator. McMurtry et al. 32 revised the types of lifestyles of phytoseiid mites. In their study, type II remained almost the same, while type III was subdivided according to their habitat, instead of according to the prey range 32 .These data suggest that it will be valuable to further evaluate the prey range of type II and type III species quantitatively, and analyze the correlation between intraguild competitiveness and prey range.
Higher immature survival was observed for A. swirskii than A. orientalis and N. californicus, when they fed on intraguild prey. The developmental duration of A. swirskii was 6.38d and 6.02d when fed on A. orientalis and N. californicus, respectively, both shorter than its developmental duration when fed on B. tabaci (6.96d) 36 . The developmental duration of N. californicus fed on A. swirskii (4.79d) also is shorter than that when they fed on Tetranychus urticae (6.46d) 37 . In contrast, the developmental duration of A. orientalis on A. swirskii (5.38d) is longer than the same on a mixed prey of Panonychus ulmi and Tetranychus viennensis (4.20d) 38 .
Some previous studies suggest that the quality of intraguild prey is often lower than the quality of extraguild prey 39,40 . However, Buitenhuis et al. 41 indicated that intraguild prey might be an equally good or better food source than the extraguild prey (thrips) for both A. swirskii and Neoseiulus cucumeris (Oudemans) 41 . Similarly, in this study, the intraguild prey appeared to be more suitable food than extraguild prey for A. swirskii and N. californicus. We observed that all three predatory mite species are able to reproduce when fed on intraguild prey. Further investigation on fecundity of predatory mites on intraguild prey will allow better estimates of how the introduction of A. swirskii will impact population dynamics of local species.
Based on our results, A. swirskii is a stronger competitor compared either to A. orientalis or N. californicus. In addition, A. swirskii could have greater negative impact on A. orientalis than on N. californicus. Therefore it is not possible to deduce whether A. swirskii and A. orientalis populations are able to coexist without negative consequences, based on the coexistence of A. swirskii and N. californicus in the Mediterranean region. Possible negative impact of A. swirskii on A. orientalis populations is herein attributed to 3 main concerns: 1) A. swirskii is a stronger competitor than A. orientalis, 2) extraguild prey does not lead to significant decrease of IGP levels between the two species, and 3) A. orientalis is an equally good or better food source for the development of A. swirskii than B. tabaci. In contrast, A. swirskii appears to be a less appropriate prey for A. orientalis.
The current study provided evidences of intraguild predations between A. swirskii and A. orientalis/N. californicus under laboratory conditions, which is an early step in investigating interactions between A. swirskii and other predatory mite species in China. IGP between predatory mites will be more complicated in real agroecosystems. For example, IGP may decrease with increasing habitat complexity 42 , and intraguild prey could avoid or escape areas with the intraguild predator 43,44 . Therefore, the impact of these factors, such as habitat complexity and dispersal capability of prey and/or predators on intraguild predations can only be estimated. Furthermore, the impact of intraguild predation on extraguild prey, intraguild predator, and intraguild prey should also be linked with other interactions among these species, which also are influenced by the complexity of real agroecosystems. For example, intraguild predations do not always result in reduced biological control efficiency, while predators that do not prey on each other do not always perform better. Previous studies on biological control of whiteflies provided examples on both sides 45,46 . The complexity of the agroecosystem will increase when other organisms sharing the same agroecosystems, including secondary prey, parasitoids, and neutral insects, etc., also are involved 45,47-49 . Chailleux et al. 48 stated the necessity to enhance the link between community ecology theory and biological control to develop better pest management strategies 49 . In our case, it is necessary to further evaluate potential interactions between B. tabaci, A. swirskii, native predatory mite species, and the environment comprehensively, and to find the equilibrium between the values and risks of introducing A. swirskii to control B. tabaci and other whitefly pests.

Methods
Laboratory experiments were conducted in small arenas to measure IGP levels between A. swirskii and A. orientalis, and between A. swirskii and N. californicus, in the presence or absence of an extraguild prey (eggs of B. tabaci), with different combinations of predator stages.

Predatory mites and whiteflies colonies.
Amblyseius swirskii used in the present study were obtained from a commercial producer (Koppert Biological Systems, The Netherlands) in 2012, and have been reared on Carpoglyphus lactis (Linnaeus) (Acari: Carpoglyphidae). Amblyseius orientalis and N. californicus were obtained from colonies maintained in the Laboratory of Predatory Mites, Institute of Plant Protection, Chinese Academy of Agricultural Sciences. Amblyseius orientalis were reared on C. lactis and N. californicus were reared on Tetranychus cinnabarinus (Boisduval) (Acari: Tetranychidae). The colonies were maintained in a climate chamber at 25 ± 1 °C, 80% ± 5%RH and 16L: 8D photoperiod. Bemisia tabaci were obtained from Dry-land Farming Institute of Hebei Academy of Agricultural and Forestry Sciences, and were cultured on tobacco plants in climate chambers (22 ± 1 °C, 60% ± 5%RH and 16L:8D photoperiod).

Experimental arena. Each experimental arena was built with a transparent acrylic board (30 × 20 × 3 mm 3 )
with a 10 mm diameter hole in the center. The top and bottom sides were sealed with a rectangular piece of glass (30 × 20 × 1 mm 3 ), and a black plastic sheet, respectively. The three layers were tightly clipped on both ends to avoid predatory mites escaping. When estimating IGP levels with extraguild prey, the bottom side was replaced by a bean leaf disc that contained B. tabaci eggs, a piece of filter paper soaked with water to keep the leaf disc moist, and another piece of rectangular glass (30 × 20 × 1 mm 3 ). Experimental arenas were maintained in a climate chamber at 25 ± 1 °C, 80% ± 5%RH and 16L:8D photoperiod.
IGP without or with extraguild prey. IGP between A. swirskii and A. orientalis or N. californicus was investigated without or with extraguild prey. For each pair of predators, IGP levels were estimated at 35 combinations, representing full combinations of all stages (eggs, larvae, protonymphs, deutonymphs, female adults, and male adults) of each species, except the combination of eggs of both species. For each combination, 1 individual of each active stage or 5 eggs of each predator were placed in the experimental arena. To avoid hatching or molting of predators during the experimental duration, eggs < 12 h old, larvae < 1 h old, protonymphs and deutonymphs < 3 h old, and newly emerged adults were used. Twenty replicates were prepared for each combination.
Each combination was examined 12 h later, except for the combinations with the larval stage. Due to the short developmental duration of the larval stage (often < 12 h), these combinations were observed 6h later. For each treatment, the number of replicates with 1 active stage killed or consumed, or at least 1 egg consumed were recorded as the number of IGP that occurred.
For all the combinations with IGP observed, IGP levels also were estimated with the extraguild prey provided. For each of these combinations, another 20 replicates were prepared, each provided with 20 B. tabaci eggs. Based on our observations, twenty whitefly eggs exceed the maximum consumption rates of both predators. For the predator combinations where no IGP was observed, no replicate in the presence of extraguild prey was conducted. A. swirskii, A. orientalis or N. californicus that fed on intraguild prey. For each of the three predatory mite species, eggs laid within 6 hours (considered as 3 hours old on average when the experiment started) were collected and reared individually to estimate development on intraguild prey. Based on the results of the IGP experiment, only adult and nymph stages of the three species performed as intraguild predators (Tables 1-2). Larvae of the three species do not feed on intraguild prey and are able to develop to protonymph stage without feeding. Therefore, intraguild prey were provided and renewed daily after each predatory mite entered protonymph stages.

Development of
When A. swirskii and A. orientalis coexisted, eggs of both species were intraguild prey in IGP cases (Table 1). In contrast, when A. swirskii and N. californicus coexisted, both eggs and larvae of intraguild prey often were consumed by elder stages of the intraguild predator (Table 2). Therefore, four sets of predatory mite development were estimated: 1) A. swirskii on A. orientalis eggs (20 eggs per d), 2) A. orientalis on A. swirskii eggs (20 eggs per d), 3) A. swirskii on N. californicus eggs and larvae (10 eggs and 10 larvae per d), and 4) N. californicus on A. swirskii eggs and larvae (10 eggs and 10 larvae per d). Based on preliminary observations, the amount of intraguild prey provided per d exceeded the consumption rate of the intraguild predators.
For each replicate, the experimental arena was checked once every 12 h, with the stage of the intraguild predator recorded, and its stage specific survival estimated. A minimum of 28 replicates completed their development for each treatment. The mean developmental duration of egg to larva, protonymph, and deutonymph stages were estimated based on the individuals that successfully developed to adults. Statistical Analyses. In the intraguild predation experiment, the IGP level of each treatment, either with or without B. tabaci eggs provided as the extraguild prey, was estimated as the proportion of replicates with IGP observed. For each combination with IGP, IGP levels without and with extraguild prey were compared using a Chi-squared test. The overall IGP level between the two predators without or with extraguild prey was compared with a paired t-test. ANOVAs were conducted to compare the developmental durations of each immature stage among the predatory mite species. Tukey's HSD tests were used for multiple comparisons. In all analyses, comparisons with p < 0.05 were considered to have statistical significant differences. Statistical analyses were performed using SPSS 18.0.