Functional responses of two predatory bugs (Hemiptera: Anthocoridae) to changes in the abundance of Tetranychus urticae (Acari: Tetranychidae) and Bemisia tabaci (Hemiptera: Aleyrodidae)

Orius spp. (Hemiptera: Anthocoridae) is well-known genus of generalist predators, which feed on numerous pest insects and mites infesting crops. In this study, the functional responses of the predatory bugs, Orius laevigatus (Fieber) and Orius vicinus (Ribaut), to different densities of the eggs of the whitefl y, Bemisia tabaci (Gennadius) (Hemiptera: Aleyrodidae) and red spider mite, Tetranychus urticae Koch (Acari: Tetranychidae), were determined under laboratory conditions. Different numbers of eggs (2, 4, 6, 8, 16, 32, 64 and 128) of both species of prey were offered to females of the above predators for 24-h in a controlled environment of 25 ± 1°C, 60 ± 10% RH and under a 16L : 8D photoperiod. The parameters of the functional responses were assessed using Holling’s Disc Equation. Both predators showed a Type II response to both prey. The attack rates (a) and handling times (Th) of the predators were computed for spider mites eggs: O. laevigatus (a: 0.972, Th: 0.007) and O. vicinus (a: 1.113, Th: 0.005), and whitefl y eggs: O. laevigatus (a: 1.022, Th: 0.002) and O. vicinus (a: 0.772 Th: 0.006). Furthermore, the average number of B. tabaci eggs consumed by O. laevigatus females was greater than by those of O. vicinus. In contrast, O. vicinus was a more effi cient predator of T. urticae eggs than O. laevigatus. Consequently, these results indicate that together these predators might be effective biological control agents in regulating populations of B. tabaci and T. urticae in agricultural ecosystems.


INTRODUCTION
The cotton whitefl y, Bemisia tabaci (Gennadius) (Hemiptera: Aleyrodidae) and the two-spotted spider mite, Tetranychus urticae Koch (Acari: Tetranychidae) are common phytophagous pests that damage many economically important agricultural crops around the world (Jeppson et al., 1975;Helle & Sabelis, 1985;Gerling et al., 2001;Aslan et al., 2004). Bemisia tabaci causes direct and indirect damage by sucking sap, virus transmission and producing honeydew that leads to rapid growth of sooty mould (Breene et al., 1992). Tetranychus urticae feeds on leaves causing damage to chlorophyll and thereafter serious loss of yield (Nachman & Zemek, 2002). For the control of these pests, farmers mostly apply chemical treatments to keep their abundance below economic damage threshold levels (Knowles, 1997;Denholm et al., 1998;Van Leeuwen et al., 2010). However, intensive use of pesticides adversely affects the environment and human health. Moreover, it becomes ineffective due to resistance that is developed by the pest after a while and in restricting naturally occurring biological control agents (Riudavets & Castañé, 1998;Biondi et al., 2012). Therefore, alternative pest control strategies including biological control have been adopted to control Eur. J. Entomol. 117: 49-55, 2020 doi: 10.14411/eje.2020.005

ORIGINAL ARTICLE
days and the old pods (after oviposition) were transferred to new containers to start a new generation. The predators were reared in a climatic chamber (Nüve TK120) at 25 ± 2°C, 65 ± 5 RH% and under a photoperiod of 14L : 10D. Nymphs and adults of B. tabaci reared on cotton plants in a controlled room (at 25 ± 2°C, 65 ± 5 RH%, photoperiod of 14L : 10D) in the Department of Plant Protection were used. This culture of B. tabaci was maintained for about 5 years, and twice a year fi eld collected B. tabaci were added to the culture in order to prevent genetic degeneration.
Adults of the red spider mite, T. urticae were collected from pesticide-free strawberry in a semi greenhouse in the Agricultural Research and Implementation Area of Cukurova University, Adana, Turkey. These mites were reared on potted bean plants in wooden-framed mesh cloth cages (1m × 1m × 1m). Potted bean plants in the cages were replaced when necessary. The cages were kept in a controlled room (25 ± 2°C, 65 ± 5 RH%, 14L : 10D photoperiod). Tetranychus urticae culture has been maintained in the laboratory for two years.

Functional response experiments
Cups (5 cm × 2 cm) were used as experimental units. Fresh bean (Phaseolus vulgaris L.) leaf discs (2.5 cm diameter) were placed upside down on wet cotton in the bottom of cups to keep the leaf disc fresh. To ensure maximum predation rates the numbers of eggs of the prey provided were determined in a preliminary experiment. With a fi ne camel's hair brush and under a stereomicroscope (Olympus SZ51) (X40), 2, 4, 8, 16, 32, 64 and 128 eggs of both species of prey were transferred onto the leaf discs in separate cups. Within fi ve minutes of transferring the prey, a one-day-old female predator, which had been starved for 24 h was placed on the leaf disc in the each cup. The cups were covered with a perforated (2 cm) lid and sealed with a mesh cloth. The cups were randomly placed in the rearing chamber and kept at 25 ± 2°C, 65 ± 5 RH% and under a photoperiod of 14L : 10D. There were between 10-15 replicates per treatment (prey-predator). In total, there were 4 treatments (prey-predator);

t-test
An independent t-test was used to evaluate differences in the number of eggs consumed by O. laevigatus and O. vicinus at each egg density. The analyses were done using the statistical Package Social Science SPSS (IBM Corp., 2015).

Functional response analyses
A logistic regression equation was used to determine the shape of the functional response recorded in each prey-predator interaction. The logistic regression model is suitable for these analyses since the output variable is dichotomous (consumed or unconsumed). In addition, the distribution of the error terms of this variable is often binomial rather than normal (Trexler et al., 1988). This logistic regression equation (1) determined the proportion of prey consumed (Na/No) as a function of the initial prey density (No) (Juliano, 2001).
The maximum likelihood test was used to estimate the parameters P 0 , P 1 , P 2 , and P 3, which are the intercept, linear, quadratic various species of thrips and especially phytophagous mites (Heitmans et al., 1986). There are several studies on the biological parameters and predation abilities of these predators feeding on different prey (Alvarado et al., 1997;Cocuzza et al., 1997;Wearing & Colhoun, 1999;Pehlivan & Atakan, 2017). However, there is little knowledge of their effi ciency as predators of B. tabaci and T. urticae with the view of using them as biological control agents (Venzon et al., 2002;Arnó et al., 2008).
Their effi ciency in regulating pest populations depends on different biological and behavioural traits. One of the most important methods for evaluating their effectiveness in biological control programs is to determine their response to changes in prey species and densities, namely, their functional responses (Rogers, 1972;Houck & Strauss, 1985). Holling (1965) identifi ed three types of functional response: (i) increasing linear response to increasing prey density, type I response, (ii) initial linear response that reaches a plateau, type II response, and (iii) sigmoidal shaped response with a slow start, type III response. There are functional response studies on anthocorid bugs to some stages of pests indicate that a type II response is the most often recorded for species of Orius (Coll & Ridgway, 1995;Alvarado et al., 1997;Montserrat et al., 2000;Rutledge & O'Neil, 2005). However, there is no data on the functional responses of O. laevigatus and O. vicinus to different egg densities of B. tabaci and T. urticae.
The aim of this study was to determine the functional response of two species of anthocorids to different egg densities (2,4,6,8,16,32,64 and 128) of two species of prey over a 24h period under laboratory conditions. The following topics were addressed: (1) Do the predators in same genus show different types of functional response when feeding on the same prey? (2) How does the functional response of each predator change when they feed on different prey? (3) What is their potential for suppressing pest populations in biological control programs? The results and further investigations might be helpful in estimating the predatory ability of O. laevigatus and O. vicinus when attacking B. tabaci and T. urticae and their value as biological control agents.

Insect rearing
Cultures of the predators O. vicinus and O. laevigatus were established in the Laboratory of Entomology (hereafter referred as laboratory) of Cukurova University, Adana, Turkey. These predators were collected from pepper and eggplant crops grown in open fi elds in the Adana Province, Turkey in 2016. The predators were identifi ed using the identifi cation key of Péricart (1972) and later reared separately in plastic jars (1 l) with a perforated lid (5 cm diameter) and sealed with mesh cloth for ventilation and preventing escape, respectively. Adults and nymphs of these predatory bugs were fed (ad libitum) with frozen eggs of the Mediterranean fl our moth, Ephestia kuehniella Zeller (Lepidoptera: Pyralidae) and pollen of Typha latifolia L. (Typhaceae). The moth culture was obtained from the Biological Control Research Institute in Adana in 2016 and maintained in the laboratory. Bean pods of Phaseolus vulgaris L. were provided every other day as a substrate for oviposition. The bean pods were replaced with fresh ones every two and cubic coeffi cients, respectively. According to Juliano (2001), if the linear term (P 1 ) is not signifi cantly different from zero, it indicates a type I functional response. Also, if P 1 is signifi cant < 0, this implies that the proportion of prey consumed declines monotonically with N 0 indicating a type II functional response. On the other hand, if P 1 is signifi cant > 0, then the proportion of prey consumed is positively density-dependent, hence a type III function response.
Knowing the shapes of the functional response curves of the predators, the next step was to determine the parameters of Holling's disc equation (2). This equation is suitable for estimating the predator attack rate or instantaneous searching rate (a) and the handling time (T h ) since the initial egg densities of prey were depleted without replacement (Rogers, 1972). A nonlinear least square regression (NLIN procedure in SPSS ver. 23) was used to estimate these parameters.
Where N a is the number of eggs consumed, N 0 is the initial density of eggs, a is the predator attack rate or instantaneous searching rate, T is the time of exposure of predator to prey (T = 24 h) and T h is the handling time. It is worth noting that since all functional response curves were type II, no adjustments to equation 2 were necessary in order to estimate the parameters for type III prey-predator functional response curves.

RESULTS
The results obtained from the logistic regression analysis of the functional response experiments had signifi cantly negative P 1 values indicating that the functional responses of both species of Orius to different densities of the eggs of T. urticae and B. tabaci were of type II (Table 1). The proportion of eggs of T. urticae and B. tabaci consumed by each predator declined and plateaued above a density of 64 eggs (Fig. 1).
Following the determination of the Type II functional responses, attack rate (a) and handling time (T h ) of Orius spp. were estimated using Holling's Disc Equation (Holling, 1959 Table 2).
The mean number of eggs of T. urticae and B. tabaci consumed at different densities by the Orius spp. is given on Table 3. The number eggs of T. urticae eaten by predators increased and reached a plateau, and there were statistical differences in numbers eaten by the two species of Orius spp. at egg densities 2 and 32. For B. tabaci eggs, each predator consumed more eggs at higher densities and generally O. laevigatus and O. vicinus consumed similar numbers of eggs at each of the densities (except 4 eggs).
Our fi ndings indicate that predators are not able to consume an infi nite number of prey as prey density increases. Consequently, beyond a certain level of prey abundance more predators will be required to keep prey numbers below the economic threshold level.

DISCUSSION
Many predators that have been successfully used as biocontrol agents for important pests in greenhouses exhibit a type II response to their prey (Pervez & Omkar, 2006;Xiao & Fadamiro, 2010 , 2015;Salehi et al., 2016). Predators with high attack rates (a) and low handling times (T h ) are considered to be more effi cient biological control agents of pests (Fathi & Nouri-Ganbalani, 2010;Salehi et al., 2016). Our results indicate that O. laevigatus is more effective than O. vicinus in terms of the number of eggs of B. tabaci eaten, whereas O. vicinus had a higher predation rate when fed eggs of T. urticae. According to Heitmans et al. (1986) O. vicinus is considered to be a potential biological control agent of phytophagous mites. In addition, some researchers report that O. laevigatus prefers mixed stages of T. urticae and B. tabaci to species of thrips (Venzon et al., 2002;Arnó et al., 2008). This could be due the mites and whitefl ies having stages (e.g. eggs, larvae or nymphs) that are more vulnerable to attack by predators. When prey are immobile, such as eggs or pupae, the predation rates can be high under natural fi eld conditions (Andow, 1990;Cook et al., 1996) (De Clercq et al., 2000;Cedola et al., 2001;Madadi et al., 2007;Jalalizand et al., 2012;Banihashemi et al., 2017).
In conclusion, this is the fi rst report of the functional responses of O. laevigatus and O. vicinus to the abundance of two important species of pests, B. tabaci and T. urticae. In Turkey, the parasitoid Eretmocerus mundus Mercet (Hymenoptera: Aphelinidae) and the predators Macrolophus melanotoma (Wagner) and M. pygmaeus (Rambur) (Hemiptera: Miridae), are considered to be important natural enemies of B. tabaci (Karut et al., 2016). In addition, there are many predatory mites, especially those belonging to family Phytoseiidae (Şekeroglu & Kazak, 1993;Attia et al., 2013), which have been released against T. urticae. With respect to our results, the expectation is that both of these predators can be used as biological control agents of whitefl ies and two spotted mites as they are likely to mainly attack their eggs. The reproduction and foraging abilities of these predators when attacking the above mentioned pests have not been extensively investigated. In addition, these experiments were done in the laboratory in small arenas that are very different from natural conditions. Thus, further studies regarding the biological parameters and behavioural responses of these predators when attacking these preys are needed in order to clearly understand their potential capacity in terms of biological control.