Functional and numerical responses of the predator Amblyseius swirskii to its prey Tetranychus turkestani in the laboratory

The tetranychid Tetranychus turkestani Ugarov and Nikolskii is a serious pest of many important crops around the world. Management of T. turkestani by augmentative biological control using predators such as the phytoseiid Amblyseius swirskii (Athias-Henriot) is envisioned as an environmentally safe alternative to acaricides. Foundational knowledge on T. turkestani – A. swirskii interactions in the laboratory are necessary to predict the outcome of A. swirskii augmentative releases in the field. In this study, the functional and numerical responses of adult A. swirskii females feeding on immature stages of T. turkestani were determined in the laboratory. Prey densities were 2, 4, 8, 16, 32, 64, or 128 individuals per Petri dish arena. The functional response of A. swirskii to prey showed a Holling's type II response. The attack rate and handling time estimates from the random predator equation were 0.1148/h and 0.3146 h, respectively, indicating that A. swirskii consumed 76.28 individuals per day at the maximum level. The number of eggs laid by the predator, i.e., the numerical response, increased as host density increased up to a maximum of 33.10 eggs per female; then oviposition rate leveled-off. This study suggests that A. swirskii is a suitable candidate for augmentative biological control of T. turkestani but follow-up experiments in greenhouses or open fields are necessary.


Introduction
Tetranychus turkestani Ugarov & Nikolskii (Acari: Tetranychidae) is an important pest of different field and glasshouse crops in the southwestern region of Iran (Mossadegh and Kocheili 2003), Russia (Popov 1981), USA (Jeppson et al. 1975), China (Li et al. 2014), India and some European countries (Zhang 2003). Characterized by frequent outbreaks coupled with acaricide resistance, this mite poses a threat to crop producers (Mossadegh and Kocheili 2003). The hosts of T. turkestani are species of Solanaceae, Cucurbitaceae and Fabaceae in Khuzestan province, southwestern of Iran (Mossadegh andKocheili 2003; Sohrabi andShishehbor 2008). Tetranychus turkestani is often found on field bindweed (Convolvulus arvensis L.), which may serve as a reservoir, and on several other weed species along road margins. This mite has several generations during the growing season, tolerates high temperatures and low humidity and has a short generation time of approximately 10 -12 days (Sohrabi and Shishehbor 2008). Tetranychus turkestani initially feed on the lower leaf surface but can cover an entire plant as population density increases. Typical symptoms of T. turkestani feeding are small, light colored punctures which, in prolonged exposure, develop into irregularly shaped, white or greyishcolored spots. Changes in leaf color from yellowing to bronzing are often characteristics of attack. In some host plant species, leaf curling can be a result of feeding. Necrosis may also occur in young leaves, stems and even growing tips. Sometimes leaf burning and defoliation of the host plant has been observed (Jeppson et al. 1975).
Information on functional and numerical responses of A. swirskii to T. turkestani populations are necessary to predict the outcome of augmentative releases of A. swirskii against T. turkestani in the field. The functional response of a predator is a function of prey density and generally follows one of three mathematical models (Holling 1959a, b, Holling 1961. With a type I functional response model, the number of prey killed increases linearly at constant rate as a function of prey density. With a type II response model, the number of prey killed increases up to a maximum (predator saturation) but the proportion of prey killed declines with prey density. With a type III response model, predation results in a sigmoidal curve. Predators with a type III response can regulate prey populations (Holling 1965). Predators with a type II response have proven to be effective under conditions of low prey density (Koehler 1999; Krebs 1978. The predator's numerical response is an important factor that has been attributed the success of phytoseiids in pest control (Sabelis 1985). The numerical response is a progressive change in the number of offspring in relation to prey density (Solomon 1949), and female predatory mites can vary their progeny production at different prey densities. In phytoseiid predatory mites, the numerical response is linked to the density of prey mites (Cédola et al. 2001; Nomikou et al. 2010; Carrillo and Peña 2012; Fathipour et al. 2020. McMurtry and Rodriguez (1987) showed that the numerical response probably had a more important role in the adjustment of pest mite populations than the functional response because increments in the numerical response, with an increase of prey density, also involved an increment in the reproduction and other important components, such as the survival of progeny and the duration of development time. The analysis of both the functional and numerical responses is helpful and provides information regarding the prey-predator relationship, which is required for the successful usage of a biological control agent.
The functional responses of A. swirskii feeding on an eriophyid Aculops lycopersici (Massee) (Park et al. 2010), astigmatid Suidasia medanensis (Oudemans) (Acari: Suidasiidae) (Midthassel et al. 2014), tenuipalpid Cenopalpus irani Dosse, and the tetranychids Tetranychus urticae Koch (Xiao et al. 2013; Fathipour et al. 2020 and Eotetranychus frosti (McGregor) (Bazgir et al. 2020) have been reported previously. However, no published data are available on the functional response of A. swirskii feeding on T. turkestani (Rahmani Piyani 2021). In addition, only the numerical response of A. swirskii on T. urticae during 24 h is known (Fathipour et al. 2020). Therefore, the aim of this study was to determine the functional and numerical responses of A. swirskii feeding on T. turkestani. This information will provide baseline data necessary to predict the effectiveness of augmentative releases of A. swirskii to control T. turkestani on crop plants in greenhouses, glasshouses, or open fields.

Mite colonies
Tetranychus turkestani used in this study originated from a weed (Convolvulus arvensis L.) at Faculty of Agriculture, Shahid Chamran University of Ahvaz, Ahvaz, Iran. A stock colony of T. turkestani was maintained on cowpea (Vigna unguiculata (L.) Walp.) seedlings in compost in plastic pots (20 cm diam). Infested plants were held in wooden-framed rearing cages (120 × 60 × 60 cm) covered with nylon mesh (210 µm aperture). The cages were maintained in a laboratory at 25 ± 2°C, 60 ± 5 % RH and a 16: 8 (L: D) h with illumination (4000 lux) provided by fluorescent lamps. New plants were introduced as required to maintain the colony.
The initial population of A. swirskii was obtained from laboratory colonies at Bu-Ali Sina University, Hamedan, Iran, which had been purchased from Koppert Biological Systems (Berkel en Rodenrijs, The Netherlands) in 2012 and reared on T. urticae. Each rearing unit consisted of a plastic sheet over a foam mat maintained under humid conditions with distilled water in a Plexiglas box (17 cm ×10 cm ×6 cm). To provide food, young cowpea leaves infested with T. turkestani were added to the units twice a week.

Experimental design
Experimental units were prepared according to the method described by Vantornhout et al. (2004) with some modifications. Each experimental unit was comprised of a green plastic sheet (3 cm x 3 cm) located on a thick foam pad, cut to a similar size (3 cm x 3 cm), in a plastic Petri dish (9 cm diam., 2 cm high) arena, half-filled with water. The edges of the plastic sheet were covered with moist tissue paper to prevent A. swirskii females from escaping. In addition, a few cotton threads were placed on each plastic sheet to provide shelter and oviposition substrates for A. swirskii females. Plant substrates were not used in the A. swirskii colony or in experimental units. A. swirskii prefers glabrous leaves (McMurtry et al. 2013), so we used a green plastic sheet as a substrate for experimental arenas. Besides several researchers also used plastic sheet instead of leaf to study the biology predatory behavior and rearing of phytoseiid mites (Vantornhout et al. 2004; Khanamani et al. 2016; Riahi et al. 2017c. To determine the functional response of A. swirskii, a newly emerged female (less than 24 h old) was exposed to seven densities (2,4,8,16,32,64 and 128) of T. turkestani immature stages including larvae, protonymphs, and deutonymphs in similar proportions. Specifically, density 2 included 1 larva and 1 protonymph; density 4 included 1 larva, 1 deutonymph and 2 protonymphs; density 8 included 2 larvae, 2 deutonymphs and 4 protonymphs; density 16 included 5 larvae, 5 deutonymphs and 6 protonymphs; density 32 included 10 larvae, 10 deutonymphs and 12 protonymphs; density 64 included 20 larvae, 20 deutonymphs and 24 protonymphs; and density 128 included 40 larvae, 40 deutonymphs and 48 protonymphs. This procedure was used because our observations indicated that A. swirskii preferred the protonymph stage compared to other developmental stages of T. turkestani. Tetranychus turkestani immatures were transferred from cowpea plant foliage to Petri dish arenas using a fine paintbrush (no. 000). After 24 h, the number of dead larvae and nymphs were counted; all immature stages were combined in the final counts. Each prey density was replicated 10 times. The experiment was conducted in the laboratory at 25 ± 2°C, 60 ± 5 % RH and a 16: 8 (L:D) h photoperiod.
In determining the numerical response, the same methods as described above were used to expose T. turkestani larvae, protonymphs, and deutonymphs to A. swirskii females. Note that A. swirskii mated females were transferred into new Petri dishes containing a fresh supply of T. turkestani prey, each day. Prey replacement was continued for the duration of the lifespan of each A. swirskii female. The number of eggs laid by each female and its longevity was recorded every day from the beginning of the numerical response experiment until the death of each female. Each prey density was replicated 10 times. The experiment was conducted in the laboratory at 25 ± 2°C, 60 ± 5 % RH and a 16: 8 (L:D) h photoperiod.

Statistical analysis
The data on functional response were analyzed in two steps (Juliano 2001). First, the type of functional response was determined by a logistic regression of the proportion of prey consumed (N a /N 0 ) as a function of prey density (N 0 ) (equation 1): ( 0 + 1 0 + 2 2 0 + 3 3 0 ) (1) where N a is the number of prey attacked, N 0 is the initial density of prey, P 0 , P 1 , P 2 and P 3 are the intercept, linear, quadratic, and cubic coefficients, respectively, estimated using the maximum likelihood method. The sign of the P 1 from the above equation can be used to distinguish the type of the functional response curve. If P 1 <0, it describes a type II functional response. If P 1 >0 and P 2 <0, it shows a type III functional response (Juliano 2001). In the next step, handling time (T h ) and attack rate (a) coefficients were estimated using the Roger's random predator equation (Rogers 1972) (equation 2): where N a and N 0 are described above, T is 24 h, the total time available for the predator, a is attack rate and T h is the handling time. Nonlinear regression was used to estimate the attack rate and the handling time parameters (Proc NLIN, SAS Institute 2003), and the maximum predation rate was calculated from 24/T h .
Data on numerical response were analyzed using one-way analysis of variance (one-way ANOVA) to test whether the number of T. turkestani consumed and the number of eggs laid by A. swirskii were different in arenas with different prey density. Statistical analysis was conducted with SAS software (SAS Institute 2003). Differences between treatment means were compared by Tukey's HSD test (α = 0.05).

Functional response
The one-way ANOVA revealed a significant effect of prey density on predation rate (F = 14.87; df = 6, 63; P < 0.001). The consumption rate of A. swirskii females increased as T. turkestani density increased (Figure 1). The functional response was density-dependent and corresponded to a Holling's type II response curve (Table 1); the response increased on average from two individuals consumed a day when prey density was two T. turkestani immatures to 58 individuals when prey density was 128 T. turkestani immatures per arena.
The attack rate and handling time estimates from the random predator equation (equation 2) were 0.1148 h -1 and 0.3146 h, respectively ( Table 2). The maximum number of T. turkestani immatures that one A. swirskii female could attack was 76.28 individuals in a day.

Discussion
The functional response displayed by A. swirskii was well-defined by the Roger's model. The functional response curve reached a plateau at approximately 64 prey individuals during 24 h, suggesting a type II functional response where predation rate decreased as prey density increased. This type of response has been reported for A. swirskii on tetranychid (Xiao et al. 2013; Fathipour et al. 2017; Fathipour et al. 2020, tenuipalpid (Bazgir et al. 2020), and eriophyid (Park et al. 2010) mites. However, Fathipour et al. (2020) reported that the functional response of A. swirskii feeding on T. urticae eggs and nymphs changed to a type III response when maize pollen was provided as an additional food source. Attack rate and handling time are two parameters that describe the magnitude of functional responses (Pervez and Omkar 2005). Handling time is an important indicator of predator effectiveness and reflects the cumulative time required for identifying, chasing, capturing, killing, and eating prey (Holling 1959a, b, Holling 1965, Holling 1966. Different factors can influence handling time such as the speed of predator, movement of prey, and the time needed for predators to overcome individual prey (Hassell 1978), which could be related to defensive behavior of prey (Ali et al. 2011). The handling time found in this study (0.3146 h) was lower than the values reported for A. swirskii feeding on T. urticae eggs (0.5184 h, Xiao et al. 2013), T. urticae deutonymphs (0.9959 h, Fathipour et al. 2020) and E. frosti deutonymphs (0.8052 h, Bazgir et al. 2020); suggesting that A. swirskii had a greater predation response when feeding on a mixture of T. turkestani immature stages. The shorter handling time in this study may be due to the presence of T. turkestani larvae, which are easier to capture and consume in comparison to other immature prey stages. Similarly, Ganjisafar and Perring (2015) reported that the predatory mite Galendromus flumentis (Chant) had the shortest handling time when consuming larvae of the bank grass mite, Oligonychus pratensis (Banks).
In this study, A. swirskii females produced an average maximum number of 33.10 eggs at a prey density of 64 T. turkestani immatures prey arena. Other laboratory studies with adequate densities of mature and immature stages of prey have reported a variety of fecundity (number of eggs) values for A. swirskii. For example, at 25°C, Riahi et al. (2017 b) reported that A. swirskii produced an average of 40.85 eggs on T. urticae; Asgari et al. (2020) documented that A. swirskii produced an average of 82.17 eggs on Carpoglyphus lactis L. (Acari: Carpoglyphidae); Bazgir et al. (2018) found the average number of eggs produced by A. swirskii was 34.69 eggs and 25.96 eggs on E. frosti and Cenopalpaus irani Dosse (Acari: Tenuipalpidae), respectively. Hosseininia et al. (2020) reported an average of 25.21, 21.46, and 14.87 eggs produced by A. swirskii feeding on T. urticae, C. lactis and Trialeurodes vaporariorum Westwood (Hemiptera: In conclusion, this study provides baseline knowledge on the association between A. swirskii and T. turkestani. The existing evidence proposes that A. swirskii can play an important role in reducing T. turkestani populations. However, results of the functional response experiment indicate that A. swirskii could be more effective at low T. turkestani population density. Nevertheless, it is necessary to consider the results with caution because these experiments have been conducted under simplified conditions using small arenas and no plant leaf substrates. In a whole plant system, A. swirskii adults can spread from one plant to the next and interact with other prey and predator populations, which could significantly affect its efficiency. Future research on prey stage preferences of A. swirskii and intraguild predation should add to the understanding of the capacity of A. swirskii as an effective biological control agent of T. turkestani. Moreover, glasshouse (greenhouse) and field studies are necessary to determine the effectiveness of A. swirskii against T. turkestani under more realistic conditions.