Parasitoids
Infested fruits were collected from underneath and on the trees of pomegranate orchards of Shiraz and vicinity, Iran, and were transferred to the laboratory for dissection and removal of the carob moth larvae. The parasitized larvae were individually placed in plastic containers (10 cm height 20 cm width) until emergence of the adult parasitoids. Parasitoids were reared separately on two hosts, E. kuehniella and E. ceratoniae larvae, in the laboratory (26 ± 1 °C, 50 ± 5% RH and 14 L: 10 D) for three generations before using in the experiments. Honey solution (10%) was placed in the containers as adult food.
Host insects
Original carob moth colonies were obtained from adults emerged from infested pomegranate fruits collected from orchards in Shiraz vicinity, Iran. Emerged adults were transferred into mating cages (50 × 50 × 100 cm) and after 24 h. of exposure, each mated female was removed from the cage and placed separately in a plastic container (1 L volume). The container was inverted on a piece of rough filter paper. A 3-cm-diameter opening was cut on bottom of the plastic container and covered with a fine mesh for ventilation. During oviposition adult females were provided with cotton wool pieces which were soaked in 10% honey water solution for feeding. The hatched larvae were transferred on the artificial diet (wheat bran 300 g, sugar 80 g, yeast 9 g, multivitamin 1.4 g, tetracycline antibiotics 0.6 g, sterile distilled water 120 ml, and glycerin 130 ml) using a fine brush.
Life table analysis
The age-stage, two-sex life table approach was utilized to analyze the raw life-history data for G. legneri (Chi 1988) via the computer program TWOSEX-MSChart (Chi 2020a). The population parameters including age-stage-specific survival rate (sxj), age-stage-specific fecundity (fxj), age-specific fecundity of total population (mx), age-specific survival rate (lx), age-specific maternity (lxmx), intrinsic rate of increase (r), finite rate of increase (λ), net reproductive rate (R0), mean generation time (T), age-stage life expectancy (exj) and reproductive value (vxj) were calculated. In order to estimate the variances and standard errors of population parameters, the bootstrap technique with 100,000 resampling was applied (Wei et al. 2020). Quick paired bootstrapping (paired 1 by 1) function was used for estimating the significant differences between means. Sigma plot v. 12.5 was used to create graphs.
Population projection
Using the computer program TIMING (Chi 2020b), the population growth of G. legneri after 100 days was projected through the method of Chi (1990) and Huang et al. (2018) as follows:
$$N_{t} \mathop{\longrightarrow}\limits^{G, D, F} N_{T + 1}$$
where G, D and F are growth, development and fecundity matrices produced via using TWOSEX-MSChart.
Host stage preference
To determine the host larval stage preferences of G. legneri, a non- and a choice experiment were designed based on the size of larvae of each host, i.e., small (L1 and L2) and large (L4 and L5). Experiments were conducted separately for each host species. In the non-choice experiment, a female and male of G. legneri were confined in a Petri dish (8 cm) and provided with 10% honey solution for food. The insects were 2–3 days old, and the females were naïve. A total of 60 larvae, in batches of 30 for each group (L1 + L2 and L4 + L5), were placed separately in Petri dishes. The larvae were exposed to the parasitoids for 24 h in an incubator at 25° C and a 14:10 (L:D) photoperiod. After exposure, the parasitoids were removed from the Petri dishes, and the parasitized larvae were counted. In choice experiment, the procedures were the same as above, except that in each Petri dish, a mixture of larvae belonging to two groups, i.e., 15 small larvae (L1 + L2) and 15 large larvae (L4 + L5), were provided. Each experiment was replicated 10 times.
Comparisons of host larval stage preferences in non- and a choice experiment were done by independent t test (P < 0.05). Data were normalized using arcsine transformation. All analyses were done with SPSS software version 19 (SPSS Inc., Chicago, USA).
Parasitoid’s preference for the host stage was evaluated by calculating a preference index according to Manly (1974). This index is as follows:
$$\beta_{i} = \frac{{\log \left[ {\frac{{e_{i} }}{{A_{i} }}} \right]}}{{\mathop \sum \nolimits_{s = 1}^{k} \log \left[ {\frac{{e_{s} }}{{A_{s} }}} \right]}}$$
where βi is the preference for prey group i, ei are the numbers of hosts remaining after the experiment; Ai and As are the number of prey groups i and s offered, respectively. This index provides a value between 0 and 1. With two-prey choices, as in our experiment, the value 0.5 for βi shows that the predator has no preference for any of prey groups, whereas values greater and lower than 0.5 indicate a preference for prey group i and prey groups, respectively (Meyling et al. 2004). To assess if the estimates deviated from 0.5, a two-tailed t test (P < 0.05) SPSS 19 software was utilized.
Parasitoid’s preference for the host stage was also assessed by calculating a preference index according to Jervis and Kidd (1996) using the following equation:
$$\frac{{E}_{1}}{{E}_{2}}=c\left(\frac{{N}_{1}}{{N}_{2}}\right)$$
where the N1 and N2 are the number of small and large larvae, and E1 and E2 are the number of small and large parasitized larvae, respectively. The value c < 1 indicates preference for prey 2 (group 2), whereas c > 1 depicts preference for prey 1.
Functional response
Since in nature, only one larva in each pomegranate fruit is exposed to G. legneri females for being parasitized, so the functional response with E. ceratonia larvae seemed to be meaningless. Therefore, only the flour moth larvae were included in the experiment. The 4th and 5th instar larvae of E. kuehniella were placed within the Petri dishes in different densities of 4, 6, 10, 20, 30, 50 and 60. One mated 10-day-old female parasitoid (fed on drop of 20% honey/water) was introduced into each Petri dish. Ten replicates for each host density were utilized. The parasitoids were removed after 24 h, and the parasitized larvae were counted. In order to determine the shape (type) of functional response, the logistic regression of the proportion of parasitized hosts (Na/N0) as a function of host density (N0) is used. For doing this, a polynomial function (Eq. 1) is fitted:
$$\frac{{N}_{a}}{{N}_{t}}=\frac{[\mathrm{exp}\left({P}_{0}+{P}_{1}{N}_{t}+{P}_{2}{N}_{t}^{2}+{P}_{3}{N}_{t}^{3}\right)]}{1+\mathrm{exp}\left({P}_{0}+{P}_{1}{N}_{t}+{P}_{2}{N}_{t}^{2}+{P}_{3}{N}_{t}^{3}\right)}$$
(1)
where Na/Nt is the proportion of parasitized hosts, Nt is the initial host density, P0, P1, P2 and P3 are the intercept, linear, quadratic, and cubic coefficients estimated through the CATMOD procedure in SAS, respectively. If the signs of the linear parameter (P1) and the quadratic parameter (P2) are both negative, the proportion of parasitized host decreases monotonically with host density (De Clercq et al. 2000) implying type II functional response. If P1 and P2 are positive and negative, respectively, the proportion of parasitized host is positively density-dependent depicting type III functional response (Juliano 2001).
After determination of the type of functional response, to estimate the parameters associated with functional response models, a nonlinear least square regression (SAS Institute 2014) was utilized. As our data fit a type III functional response, the type III equation for parasitoids was utilized as follows (Eq. 2):
$${N}_{a}={N}_{t}\left[1-\mathrm{exp}\left(-\frac{bT{N}_{t}{p}_{t}}{1+c{N}_{t}+b{T}_{h}{N}_{t}^{2}}\right)\right]$$
(2)
where Na is the number of parasitized hosts, Nt is the initial number of hosts, Pt is the number of the parasitoid, T is the total time of the experiment (24 h in our study), Th is the handling time, b and c are constants, (24 h). The coefficient of determination (R2) was calculated using the following equation: R2 = 1 − (residual sum of squares/corrected total sum of squares).