The Potential of Parapanteles hyposidrae and Protapanteles immunis (Hymenoptera: Braconidae) as Biocontrol Agents for the Tea Grey Geometrid Ectropis grisescens (Lepidoptera)

Simple Summary The tea grey geometrid Ectropis grisescens is a significant insect pest of tea plants in China. Parapanteles hyposidrae and Protapanteles immunis are Ectropis grisescens larval parasitoids. Here, we studied the parasitism performance of these two parasitoid species on different host densities under different temperatures as well as the interference effect of parasitoid density. We found that both parasitoid species, Pa. hyposidrae and Pr. immunis, exhibited a type II functional response towards the tea grey geometrid E. grisescens at four tested temperatures. With increasing the density of E. grisescens larvae, the number of parasitized larvae increased until a maximum was reached. Pr. immunis performed better than Pa. hyposidrae under higher temperatures. The parasitism rate by a single female parasitoid decreased with increasing parasitoid density at different temperatures, resulting in a reduction of searching efficiency. The findings of this study showed that Pr. immunis could be a better effective biocontrol agent than Pa. hyposidrae against the tea grey geometrid. Abstract The tea grey geometrid Ectropis grisescens has long been a significant insect pest of tea plants in China. Two parasitoids, Parapanteles hyposidrae and Protapanteles immunis (Hymenoptera: Braconidae: Microgastrinae), are the most important parasitoids in the larval stage of E. grisescens. Yet, the potential of these two parasitoids for controlling the tea grey geometrid is not known. Here, we studied the parasitism performance of these two parasitoid species on different host densities under different temperatures as well as the interference effect of parasitoid density. The results showed that both parasitoid species, Pa. hyposidrae and Pr. immunis, exhibited a Type II functional response towards the tea grey geometrid E. grisescens at four tested temperatures. With increasing the density of E. grisescens larvae, the number of parasitized larvae increased until a maximum was reached. The highest number of hosts parasitized by Pa. hyposidrae or Pr. immunis reached 14.5 or 14.75 hosts d−1 at 22 °C, respectively. The estimated values of instantaneous searching efficiency (a) and handling time (h) for Pa. hyposidrae or Pr. immunis were 1.420 or 3.621 and 0.04 or 0.053 at 22 °C, respectively. Pr. immunis performed better than Pa. hyposidrae under higher temperatures. The parasitism rate by a single female parasitoid decreased with increasing parasitoid density at different temperatures, resulting in a reduction of searching efficiency. The findings of this study showed that Pr. immunis could be a better effective biocontrol agent than Pa. hyposidrae against the tea grey geometrid.


Introduction
The tea plant is an evergreen shrub whose leaves and leaf buds are commonly used to produce the teas we enjoy [1,2]. The tea grey geometrid, Ectropis grisescens Warren, 1894 (Lepidopotera: Geometridae), is one of the most destructive chewing pests widely found in tea plantations in China [3]. Their larvae cause innumerable damage to the leaves of tea plants, resulting in severe yield losses [4][5][6]. Chemical insecticides have been the major method for controlling this pest over the last few decades. Long-term overuse of insecticides, on the other hand, has caused environmental pollution and insecticide resistance [7][8][9]. Furthermore, excessive pesticide residue on tea leaves is a serious problem faced by the tea production industry [10]. In view of these concerns, biological control is thought to be a more appropriate and reliable mean for insect pest control.
Parasitoids are considered the most effective natural enemies used in biological control programs [10][11][12][13]. It is essential to evaluate and test the efficacy of potential biological control agents under laboratory conditions before their release in the field. The functional response is regarded as a valuable measure for evaluating a natural enemy's potential success [14]. It reflects the relationship between the number of preys or hosts attacked by a predator or parasitoid and the population density of the preys or hosts. When a parasitoid meets various densities of hosts, the functional response offers a quantitative explanation of the parasitoids' behavior. Type I linear, Type II rectangular hyperbola, and Type III sigmoidal are the three basic forms of functional response [14][15][16][17]. Parasitoids are frequently linked to the Type II response. Meanwhile, various functional responses of the same species may be influenced by factors such as prey or host type, life stage, and temperature [18][19][20][21]. At high or low temperatures, the longevity and life-time fecundity of parasitoids are decreased [22]. Temperature also affects the functional response of a parasitoid, especially in the range of temperatures where parasitoids search for their hosts [23].
Parapanteles hyposidrae and Protapanteles immunis (Hymenoptera: Braconidae: Microgastrinae) are the most important parasitoids in the larval stage of the tea geometrid E. grisescens in the tea-growing areas of China. They can be easily separated from each other: Pa. hyposidrae has the head and mesosoma black, the metasoma dark brown, the abdominal segments with white areas, and its cocoon smooth and mostly yellowish-green while Pr. immunis has its body black with yellow hind femur, and its cocoon covered with fluffy cotton-like filaments and white [24]. However, it was frequently reported that they had variable parasitism rates and percentage of populations in the different localities and among the different years [24]. Therefore, we hypothesize that the performance of these two parasitoids may be affected by many factors, such as temperature, host population, and density of parasitoids themselves. To test our hypothesis, we studied the parasitism performance of these two parasitoid species on different E. grisescens densities under different temperatures as well as the interference effect of parasitoid density, to determine the potential of Pa. hyposidrae and Pr. immunis as biocontrol agents of E. grisescens. Our research results will provide basic data for biological control of E. grisescens.

Insect Cultures
Ectropis grisescens, Parapanteles hyposidrae, and Protapanteles immunis were collected from Yuhang district, located in Zhejiang Province, China (longitude 119.66667 E, latitude 30.15 N). All insect colonies were maintained in an environment-controlled room (21 ± 1 • C, 60 ± 10% RH and 16:8 h L: D photoperiod, designed by Zhejiang University, Model: AGC-1). The tea plant was obtained from the Tea Research Institute, Chinese Academy of Agricultural Sciences (longitude 120.09377 E, latitude 30.18 N) to rear E. grisescens. Second instar larvae of E. grisescens were used as hosts for the subculture of parasitoids. The adults of Pa. hyposidrae and Pr. immunis were maintained in the same environment-controlled room and fed with 10-20% honey water at the same time.

Functional Response at Different Temperatures
Each experimental unit consisted of a plastic box (upper diameter 138 mm, lower diameter 107 mm, height 113 mm) with a circular hole that was a diameter of 6 mm, and the small hole was plugged with absorbent cotton. We tested the functional response of Pa. hyposidrae and Pr. immunis at four temperatures (18 • C, 22 • C, 26 • C, and 30 ± 0.5 • C) under 60 ± 5% relative humidity and 16:8 h light: dark photoperiod in plant growth chambers (Panosonic MLR-352H-PC). Different numbers of 10, 15, 20, 25, and 30 s instar larvae of E. grisescens were offered to one mated female (24 h old) of either Pa. hyposidrae and Pr. immunis at each of four temperatures.
An appropriate number of fresh tea leaves was placed in the experimental containers. The duration of the experiment was 24 h and then the parasitoids were removed and the containers containing larvae were kept in incubators at the same constant temperatures. Under the same conditions, we continued to feed E. grisescens larvae, replaced fresh tea leaves every day, observed, and recorded the number of wasp cocoons and eclosions emerging until the host died from parasitization or pupates. Unhatched pupae were dissected under a stereomicroscope to examine whether they were parasitized. Each treatment with different combination of temperatures and densities was replicated four times.

Parasitoid Performance under Dynamic Parasitoid Density
The experimental environment and apparatus are the same as those in Section 2.2. 60 s instar larvae of E. grisescens were offered to various mated female (24 h old, 1, 2, 3, 4, and 5 adults) of either Pa. hyposidrae or Pr. immunis at two temperatures (22 • C and 26 • C). An appropriate number of fresh tea leaves were placed in the experimental containers. Each treatment with different combination of temperature and density was replicated four times.

Statistical Analyses
The functional-response data was analyzed with the R (version 4.1.2) package 'Frair' [25], which employed a functional response model based on logistic regressions. Logistic regression analyses of the proportion of parasitized larvae as a function of initial density were used to determine the type of functional response [26]. The coefficient of linear term (p = 0, a linear increase in parasitism rate as host densities rise, until reaching a maximum parasitism rate) characterized Type I, a negative first-order term (p < 0, declining proportional consumption with increasing resource density) characterized Type II, while a positive first order term followed by a negative second-order term characterized Type III (p > 0, initial increase and subsequent decrease in proportional consumption) [26]. The data were fitted to the Type I model described by Holling [25]: The 'frair_test' function was used to assess whether a Type II or III functional response better reflected the relationship between host density and the number of parasitized hosts or each trial combination.
Significant negative linear coefficients from logistic regression suggest Type II functional response model described by Rogers [27]: Significant positive linear coefficients from logistic regression suggest Type III functional response model described by Real [28]: A Type II response is defined by the parameters a (the instantaneous searching efficiency or attack rate), h (the handling time required for parasitic wasps to parasitize or attack a host), and T (experimental duration, 24 h), whereas a Type III response is defined by the parameters h, q (scaling component), b (search coefficient), and T. The scaling component q is a critical determinant of the functional response shape. It shows the extent to which functional response changes from a decelerating hyperbola to a sigmoidal form, with q = 0 in Type II and q > 0 in Type III. Following these analyses and since our data fit a Type II functional response, we used the Rogers Type II equation. The component parameters (a and h) were compared by using the "difference method" outlined in Juliano [24] via "frair compare" z-tests and of optimized coefficients, adopting the significance level of 5% (p < 0.05) for all statistics.
Bootstrapping was used to construct 95% confidence intervals to visualize variability around the fitted curves. Based on the output from bootstrapped fits, a functional response curve could be constructed using the 'drawpoly' function of the 'Frair' package.
To detect mutual interference among the parasitoid wasp, data were analyzed using a one-way ANOVA, and multiple comparisons of means were carried out with Tukey's HSD test.
The equation by Hassell and Varley [29]: log 10 a = log 10 Q − m log 10 P We call the slope m the mutual interference constant, the parameter Q is the quest constant, and log 10 Q is the intercept of the equation. The parameter a is the search rate of the parasitoid wasp and was estimated for each replicate as: The parameter P is the number of parasitoids, T is the duration of the experiment (24 h, also often taken to be (1), A is the total area (often assumed to be 1), N 0 is the number of hosts (60), and N s is the number of hosts not parasitized.

Parasitoid Performance under Dynamic Host and Different Temperatures
The results of the factorial analysis showed that temperature and host density had a significant effect on the proportion parasitized. The number of the E. grisescens larvae parasitized by Pa. hyposidrae or Pr. immunis increased with the increase of temperatures and densities of the host E. grisescens larvae. According to the logistic regression analysis, females of parasitoids exhibited a Type II functional response when attacking second instar larvae of E. grisescens at various temperatures, because the parasitized E. grisescens larvae against the initial density of host larvae yielded significantly negative linear parameters (p < 0) for Pa. hyposidrae (Pr(z) < 0.05; Table 1) and Pr. immunis (Pr(z) < 0.001; Table 1). The functional response curves (Figure 1a,b) indicated that the number of E. grisescens larvae parasitized increased with an increase of the number of host individuals offered until a maximum was reached at all the temperatures tested. Therefore, we concluded that both parasitoid species exhibited Type II functional responses when parasitizing E. grisescens larvae.
In the Pa. hyposidrae-E. grisescens parasitoid-host combination (Figure 1a), as the temperature increased from 18 • C to 30 • C, the maximum number of hosts parasitized increased from 2.  (Table 1). However, in the Pr. immunis-E. grisescens parasitoid-host combination (Figure 1b) The results for the comparison of Type II functional response parameters showed that Pr. immunis exhibited a higher number of parasitized hosts and instantaneous searching efficiency a than Pa. hyposidrae at all temperatures (Figure 1c-f). However, differences were not significant except at 22 • C by using the "difference method" ( Table 2). Unlike the "difference method", bootstrapping results show significant differences at 22 • C and 30 • C (95% CIs clearly overlap) ( Table 3). At 30 • C, the handling time h of Pa. hyposidrae and Pr. immunis differed significantly (95% CIs no overlap) ( Table 3).  In the Pa. hyposidrae-E. grisescens parasitoid-host combination (Figure 1a), as the temperature increased from 18 °C to 30

Mutual Interference at Different Temperatures
We observed that parasitism rate by single female parasitoid decreased with increasing parasitoid density. The resulting mutual interference constant (m) for Pa. hyposidrae and Pr. immunis at 22 • C was estimated to be 0.304 or 0.245, respectively (Table 4), while at 26 • C the constant was estimated to be 0.210 or 0.260. Mutual interference among parasitoids reduces the searching efficiency with increasing parasitoid population density (Figure 2).

Mutual Interference at Different Temperatures
We observed that parasitism rate by single female parasitoid decreased with increasing parasitoid density. The resulting mutual interference constant (m) for Pa. hyposidrae and Pr. immunis at 22 °C was estimated to be 0.304 or 0.245, respectively (Table 4), while at 26 °C the constant was estimated to be 0.210 or 0.260. Mutual interference among parasitoids reduces the searching efficiency with increasing parasitoid population density ( Figure 2).

Discussion
The parasitoid functional response is regarded as crucial to host-parasitoid dynamics [30]. Our study is the first research to define the parasitoid functional response types of Pa. hyposidrae and Pr. immunis on E. grisescens. The significantly negative linear parameters (p < 0) obtained in the present study confirm that both parasitoids display Type II functional responses towards E. grisescens. Interestingly, Pr. immunis parasitized a higher number of hosts and exhibited higher instantaneous searching efficiency than Pa. hyposidrae at all temperatures. The handling time estimated for Pr. immunis was generally shorter than those for Pa. hyposidrae at higher temperatures, especially at 30 • C, which suggests that Pr. immunis could parasitize more E. grisescens at higher temperatures than Pa. hyposidrae. Therefore, Pr. immunis may be more effective for biological control of E. grisescens. The results are consistent with the data observed in the field that Pr. immunis was the predominant parasitoid of E. grisescens by investigating the species and number of parasitoids of E. grisescens (data unpublished).
In the functional response of Type II, the proportion of host consumed declines as host density increases [15]. The Type II functional response is theoretically less capable of suppressing host density when compared to the Type III functional response. Although Type II functional responses are common in host-parasitoid systems [30][31][32][33][34][35][36][37] and Type III is not prevalent in parasitic insects [34,[38][39][40][41], the form of the functional response on its own does not determine the success or failure of parasitoids in biological control [30]. Other factors that influence the efficiency of natural enemies in pest control include prey growth rates, behaviors, and distribution [42][43][44], as well as temperature [16] and host plant [45].
Mutual interference amongst parasitoids can lead to a reduced rate of parasitism of host populations as parasitoid density increases [46,47]. We investigated the mutual interference of Pa. hyposidrae and Pr. immunis by observing and analyzing the parasitic behavior of single and multiple parasitoids, respectively. The results showed that the increasing number of parasitoids did not result in a proportional increase in the number of hosts parasitized, due to the negative effects of mutual interference. This is likely due to the limit space in the experimental arena that generated high conspecific encounter rates. As the density of conspecifics increases, each individual parasitoid spends less time searching for a host and more time interacting with other conspecifics [48], and this explains why one single female parasitoid parasitizes less hosts with the increase of parasitoid density. By fitting the data to the Hassell and Varley (1969) equation, the mutual interference constant (m) for Pr. immunis has no difference at 22 • C and 26 • C, while for Pa. hyposidrae the mutual interference constant estimated at 22 • C is higher than at 26 • C. Handling time or egg limitation of parasitoids influence the parasitism rate of parasitoids with a Type II functional response. The equation used to estimate the search rate does not take into account this influence, thus the mutual interference constant was consistently underestimated [49]. Mutual interference for a shared resource may have a consequence for the behavior of parasitoids, such as reducing the proportion of female offspring, which may change the population dynamics of the system, including higher host equilibrium densities and decreased stability [10,50,51].
In general, our study suggests Pa. hyposidrae and Pr. immunis are active throughout the entire temperature range tested, therefore they may contribute markedly to E. grisescens suppression and could be incorporated into IPM systems that rely on natural enemies, as well as considered for augmentative releases.