Host specificity is a key criterion for the implementation of any biological control program. In parasitoid wasps, host specificity is mediated by their response to chemical cues directly and indirectly associated with their hosts during foraging. Although intraspecific variation in foraging behaviour is documented, it is rarely integrated into risk assessment studies of potential biological control agents, as is the case for Trissolcus japonicus (Ashmead) (Hymenoptera: Scelionidae), an Asian egg parasitoid and biological control agent of the invasive Halyomorpha halys (Stål) (Hemiptera: Pentatomidae). In the present study, we compared the behavioural response of T. japonicus females of an adventive line present in Switzerland and an Asian line (Beijing USDA), recently mass released in Italy, to cues of its target host H. halys and non-target host Pentatoma rufipes (L.) (Hemiptera: Pentatomidae). We observed minimal differences in host acceptance, exploitation and suitability in no-choice tests. In contrast, the behavioural response to host contact kairomones left on leaf substrates significantly differed between the two lines. While both lines preferred chemical footprints of H. halys in comparison to P. rufipes, females of the Beijing USDA line responded more strongly to host kairomones. The observed intraspecific variation in behavioural traits highlights the relevance of considering genetically distinct parasitoid lines and their host foraging behaviour in biological risk assessment studies. However, the implications of intraspecific variation in parasitoid foraging behaviour for biological control of a pest and non-target effects in the field remain challenging to predict.
Notes
Handling Editor: Raul Laumann.
Introduction
Risk assessment studies addressing the host specificity of selected biological control agents are today a prerequisite for the implementation of any biological control program (Barratt et al. 2010). Due to restricted accessibility and high cost involved in host specificity testing, genetically distinct populations of the same parasitoid species originating from different geographical regions are often not considered in risk assessment studies (van Lenteren et al. 2011; Szűcs et al. 2019). However, there is increasing evidence for intraspecific variation in traits in parasitoids relevant to biological control, including host specificity (Goldson et al. 2003; Hopper et al. 2019).
In the early 2000s the invasive Asian brown marmorated stink bug, Halyomorpha halys (Stål) (Heteroptera: Pentatomidae), was first detected in Switzerland (Wermelinger et al. 2008). It quickly spread throughout Europe’s mainland (Claerebout et al. 2019; Gariepy et al. 2021), where it became a serious pest of many agricultural crops (Maistrello et al. 2017; Bosco et al. 2018; Damos et al. 2020). Because effective control of H. halys is proving extremely challenging due to its high mobility and polyphagy, importing its major natural enemy in Asia, the egg parasitoid Trissolcus japonicus (Ashmead) (Hymenoptera: Scelionidae) (Zhang et al. 2017), was considered a promising approach for managing H. halys in the invaded range.
Advertisement
Host range studies in Europe revealed that the fundamental host range of T. japonicus comprises several pentatomids other than H. halys, as well as a few species of scutellerids (Haye et al. 2020; Sabbatini-Peverieri et al. 2021). However, while non-target host range testing in Europe was still underway, adventive populations of T. japonicus were found in Switzerland, Italy and Germany (Sabbatini Peverieri et al. 2018; Stahl et al. 2019; Dieckhoff et al. 2021). Subsequent ‘post arrival’ field studies in Italy and Switzerland demonstrated that the forest bug, Pentatoma rufipes (L.) (Heteroptera: Pentatomidae), was the most parasitized non-target species, while most other species were only sporadically attacked (Haye et al. 2024).
While adventive populations of T. japonicus continue spreading throughout Europe, permission was given to release a T. japonicus line originating from Beijing (China) at more than 700 locations in northern Italy from 2020 onwards (Sabbatini-Peverieri et al. 2020; Falagiarda et al. 2023). A recent study testing the degree of genetic differentiation among the adventive Swiss (Ticino) line and the released Chinese line (Beijing USDA line) revealed considerable genetic variation among these two lines (Abram et al. 2023). However, it remains unknown whether these strains show differences in host specificity, including the response to semiochemicals from non-target hosts.
The main objective of the study was to investigate whether there are differences in host specificity between the adventive and released T. japonicus lines. Hence, a series of standard no-choice tests comparing the handling time, host acceptance and suitability of H. halys (target) and P. rufipes (non-target) eggs was conducted. Additionally, we compared the response of the two parasitoid lines to contact kairomones (‘chemical footprints’) of both hosts.
Material and methods
Source and rearing of stink bugs and parasitoids
Colonies of H. halys and P. rufipes were established from field collected adults in spring and late summer 2022, respectively. Adults of P. rufipes were collected from maple trees (Acer campestre L.; Sapindaceae) in Delémont (Canton Jura, Switzerland), whereas H. halys adults were collected from common ivy (Hedera helix L.; Araliaceae) in the cities of Bern and Biel (Canton Bern, Switzerland). Adults were maintained in groups of 50 individuals in polyester cages (BugDorm-4090 Insect Rearing Cage 47.5 × 47.5 × 47.5 cm; MegaView Science Co. Ltd., Taichung, Taiwan) at 24 ± 1 °C, 70% RH, and a L:D 16:8 photoperiod. Adults of H. halys were fed with corn, beans, and fruit-bearing branches of common ivy, whereas P. rufipes adults were provided with apple slices and fresh branches of maple (A. campestre), apple (Malus domestica Borkh; Rosaceae) or beech (Fagus sylvatica L.; Fagaceae).
Advertisement
T. japonicus female isolines were maintained on fresh H. halys egg masses taken from the laboratory colonies. The Beijing USDA line was originally collected from wild H. halys egg masses near Beijing (China) in 2009 (Abram et al. 2023). The Swiss line was originally collected from H. halys sentinel eggs exposed in Manno (Canton Ticino, Switzerland) in 2017 (Stahl et al. 2019). Parasitoids were held in 0.2 l plastic containers and fed with a drop of honey placed on a sponge attached to the lid. Parasitoid containers were kept at 24 ± 1 °C, 70% RH in a dark box to reduce activity and thereby extend lifespan. Every two weeks, three to five females of each isoline were transferred into a new container with two cold stored H. halys egg masses (Bittau et al. 2021; Wong et al. 2021). Parasitized egg masses were incubated at 24 ± 1 °C, 70% RH, and a L:D 16:8 photoperiod until offspring emergence. The lifespan of T. japonicus females had been reported to be up to 100 days when kept at constant 26 °C and 16 h light (Sabbatini-Peverieri et al. 2020), but by decreasing temperature and reducing the light phase, female survival and reproductivity was extended up to six months (unpublished data).
No-choice tests
No-choice tests were conducted using 17–19 days old, mated, naïve T. japonicus females of either the Beijing USDA line or the Swiss line. For each T. japonicus line, single females were presented with a fresh host egg mass (maximum three days old) of either P. rufipes or H. halys in a small Petri dish arena (5 cm diameter). White paper rings around each Petri dish served as blinds to prevent visual interference with females of neighbouring Petri dishes. Since average egg mass sizes of H. halys and P. rufipes are 28 and 14 eggs (Hoebeke and Carter 2003; Powell 2020), respectively, the egg mass size was standardized by separating egg masses into smaller clusters (ten eggs per mass). Egg masses were glued with Cementit to 4 cm2 pieces of flat cardboard and kept at 6 °C overnight (maximum 12 h) before being transferred back to room temperature 1.5 h prior to the start of the no-choice tests. T. japonicus females from the respective colonies were isolated in glass pipets one hour before the test, and inspected for injuries that might negatively influence their behaviour. The parasitoid behaviour was videotaped for one hour or until females displayed guarding behaviour (i.e., motionless resting on and patrolling of host egg mass following oviposition, Field 1998) using webcams (papalook® PA150s Web camera HD 1080P) and the videotaping software Aiseesoft Screen Recorder (Aiseesoft Studio, version 2.2.70). After the test, females were removed and exposed egg masses were incubated until emergence of stink bug nymphs or parasitoid offspring. In addition, 31 and 32 unexposed egg masses of P. rufipes and H. halys, respectively, were kept as rearing controls to evaluate baseline host mortality. No-choice tests were performed at 24 ± 1 °C, 70% RH, and a L:D 16:8 photoperiod, and all egg masses were incubated under the same conditions.
The parasitoid behaviour was analysed using Behavioural Observation Research Interactive Software (BORIS, version 7.13.8). In total, 104 observations were recorded. If females failed to make antennal contact with the egg mass (eight out of 104), replicates were discarded and not included in the analysis (valid replicates: Swiss on P. rufipes: n = 24; Swiss on H. halys: n = 21; Beijing USDA on P. rufipes: n = 25; Beijing USDA on H. halys: n = 26). To assess the host acceptance and exploitation, the number of egg masses with at least one parasitized egg (i.e., egg that female drilled and marked), the total handling time of egg masses and the proportion of parasitized eggs within an egg mass were determined. Additionally, the proportion of multiparasitized eggs within an egg mass was recorded. The egg mass handling time was defined as the period from the drilling of the first egg to completing parasitization of all ten eggs (excluding the occasional multiparasitism events observed afterwards), as indicated by egg marking behaviour (Field 1998). If females showed no oviposition behaviour, the egg masses were considered rejected.
To determine the host suitability, the number of emerged nymphs, parasitoids and dead eggs (no emergence) was recorded, as well as the parasitoids’ sex ratio. Additionally, dead eggs from the parasitoid treatment were dissected to determine the presence of undeveloped parasitoids.
Host contact kairomone assay
To compare the behavioural response of T. japonicus to host kairomones of H. halys and P. rufipes, we followed the study by Boyle et al. (2020). As both H. halys and P. rufipes use maple trees as host plant, sycamore maple (Acer pseudoplatanus L.; Sapindaceae) was selected as plant substrate for the assays. Fresh leaves were collected from maples trees in an unmanaged forest in Delémont, Switzerland, and processed immediately for experimental use. Leaves were carefully washed with mild soap water and left to dry prior to assays. To contaminate abaxial leaf surfaces with footprints by either P. rufipes or H. halys females, a 6 cm2 cut-out from the middle of a leaf was placed in a Petri dish (diameter 9 cm) with the abaxial leaf side facing upwards. Single females were confined to the leaf surface for 30 min using a small Petri dish bottom (5 cm diameter). The Petri dish bottom was moved along the midvein in three intervals of 10 min each, forcing the stink bug to walk a ‘uniform’ path. Leaves contaminated with feces or fed on by bugs were discarded. Maple leaves not exposed to stink bugs were used as negative controls.
To test the response of T. japonicus females to contact kairomones of P. rufipes and H. halys, single naïve, mated T. japonicus females (7–90 days old, due to limited availability of wasps of the same age at the time of testing) of the Swiss or Beijing USDA line were placed in Petri dishes containing either a leaf contaminated with stink bug footprints (Swiss: n = 20 each; Beijing USDA: n = 23 for P. rufipes and n = 24 for H. halys) or a clean leaf (control, n = 19 each). At the start of each trial, single females were placed into the Petri dish near the midvein of the leaf using a glass pipette. All trials were conducted within an hour of contamination and in the time window of 10h00 to 16h00 (UTC + 1 h). The wasp’s behaviour was continuously video recorded for ten minutes using the same set-up as described for the no-choice tests.
For each replicate, the wasp’s residence time on the Petri dish, the leaf surface or the underside of the leaf were recorded using BORIS. The total leaf residence time was defined as the duration of time that the female spent on the leaf surface, including short periods of the wasp leaving the leaf surface (maximum 5 s) or moving to the underside of the leaf (maximum 15 s). This was decided based on the observation that T. japonicus would sometimes just briefly leave the leaf surface or check the adaxial side of the leaf while searching for host cues. If females were still on the leaf surface at the end of the recording and had not left the leaf surface after the initial entry, replicates were excluded from statistical analysis. Because pre-experiments demonstrated that both wasp lines positively respond to host kairomones by displaying a typical arrestment behaviour (Peri et al. 2006), replicates in which the wasps did not show any arrestment behaviour were excluded from the analysis. Since no arrestment behaviour was observed on control leaves without kairomones, antennal contact with the midvein of the leaf was used as inclusion criteria of control samples into the analysis.
Statistical analysis
Logistic regressions were used with the glm function of the stats package in R (R Core Team 2022) to analyse the effect of parasitoid line (Swiss and Beijing USDA), host species (P. rufipes and H. halys), and their interaction on (1) host acceptance (proportion of egg masses with at least one parasitized egg), (2) exploitation (proportion of parasitized eggs within an accepted egg mass), (3) multiparasitism rate (proportion of eggs that were parasitized more than once by the same female), (4) host suitability (proportion of parasitoid emergence from parasitized eggs) and (5) sex ratio (proportion of female offspring). Similarly, (6) egg mortality (proportion of dead eggs with undefined content) was analysed using a logistic regression with the same variables as described above, except that the variable parasitoid line contained a control (i.e., rearing control), for which egg mortality without parasitism was determined. In all models, the data was modelled using a binomial error distribution with a logit link function. For models 2–6, overdispersion, as determined by Pearson’s χ2 test, was corrected by including a dispersion parameter estimated from the data. Post-hoc multiple comparisons of means were conducted with Tukey’s HSD adjustments.
Since data were not normally distributed and the variance of the residuals was not constant, a non-parametric Kruskal–Wallis rank sum test was used to analyse the effect of hosts and parasitoid lines on egg mass handling time. If egg masses were not fully parasitized, replicates were excluded from the handling time analysis.
A linear mixed effect model using the lmer function of the lme4 package (Bates et al. 2015) was applied to analyse the effect of parasitoid line and kairomone treatment (P. rufipes, H. halys, and control), and their interaction, on the total residence time of the parasitoid on the leaf. The elapsed time from the start of the experiment to the initial contact of parasitoids with either host kairomones (as indicated by arrestment behaviour in stink bug treatments) or the leaf midvein (control treatment) was discretized into ten intervals of 60 s each. This discretized time variable was then treated as a random effect in the analysis. Since parasitoid age can influence patch time allocation (Wajnberg 2006; Wajnberg et al. 2006; and references therein; Zhang et al. 2022), and the age of females used for the experiment was variable, we assumed that older females spend more time searching for host eggs on leaf surfaces upon encountering host footprints than younger females. To test whether the effect of female wasp age differs between wasp line and kairomone treatment, the interactions between female age and wasp line, and female age and kairomone treatment, were included in the model. The residence time was square-root-transformed to meet the assumptions of normality and homoscedasticity of residuals. Post-hoc multiple comparisons of means were conducted with Tukey’s HSD adjustments. All statistical analyses were conducted in R (version 4.2.2) using RStudio (version 2022.07.2) (R Core Team 2022).
Results
No-choice tests
The no-choice tests revealed minimal differences in host acceptance and exploitation between the two wasp lines and host species (Table 1). The egg mass handling time did not differ between the two wasp lines and host treatments. In contrast, the number of accepted (i.e., parasitized) host egg masses differed between host species. While both T. japonicus lines parasitized all H. halys egg masses offered, both lines rejected two P. rufipes egg masses each. Further, host exploitation (i.e., the proportion of parasitized eggs within an egg mass) significantly depended on both the host species and the wasp line. Of the accepted host egg masses, females of the Beijing USDA line parasitized the entire egg masses of both hosts without exception, whereas females of the Swiss line only parasitized H. halys egg masses completely. However, the latter still exploited on average 96% (SE = 4%) of P. rufipes eggs per egg mass. The proportion of multiparasitized eggs was significantly higher for wasps of the Beijing USDA line than for the Swiss line.
Table 1
Host acceptance and exploitation of the Swiss and the Beijing USDA Trissolcus japonicus lines, when exposed to a host egg mass of Pentatoma rufipes (non-target host) or Halyomorpha halys (target host), and their comparative host suitability
Host species
Statistical significance
Variable
Wasp line
H. halys
P. rufipes
Wasp
Host
Wasp × Host
Number of accepted egg masses
Swiss
21 (21)
22 (24)
χ2 = 0.00
χ2 = 5.53
χ2 = 0.00
df = 1
df = 1
df = 1
Beijing USDA
26 (26)
23 (25)
p = 0.966
p = 0.019
p = 1.000
% Egg mass exploitationa
Swiss
100.0 ± 0.0 (21)
95.9 ± 4.1 (22)
χ2 = 5.84
χ2 = 5.47
χ2 = 0.00
df = 1
df = 1
df = 1
Beijing USDA
100.0 ± 0.0 (26)
100.0 ± 0.0 (23)
p = 0.016
p = 0.019
p = 1.000
Handling time (min)b
Swiss
31.8 ± 0.4 (21)
34.7 ± 1.7 (21)
Kruskal–Wallis χ2 = 6.69
df = 3
Beijing USDA
32.4 ± 1.3 (26)
30.8 ± 0.8 (23)
p = 0.082
% Multiparasitismc
Swiss
1.0 ± 0.7 (21)
2.7 ± 1.3 (22)
χ2 = 6.27
χ2 = 1.38
χ2 = 1.88
df = 1
df = 1
df = 1
Beijing USDA
12.7 ± 4.8 (26)
4.8 ± 3.5 (23)
p = 0.012
p = 0.241
p = 0.170
% Parasitoid emerged
Swiss
80.5 ± 7.4 (21)
93.2 ± 3.9 (22)
χ2 = 1.95
χ2 = 9.32
χ2 = 1.41
df = 1
df = 1
df = 1
Beijing USDA
86.2 ± 5.0 (26)
99.1 ± 0.6 (23)
p = 0.163
p = 0.002
p = 0.235
Female sex ratio
Swiss
89.2 ± 0.9 (21)
82.6 ± 5.7 (22)
χ2 = 51.18
χ2 = 2.52
χ2 = 0.60
df = 1
df = 1
df = 1
Beijing USDA
51.8 ± 6.2 (26)
62.9 ± 5.4 (23)
p < 0.001
p = 0.113
p = 0.439
% Undeveloped parasitoids (dissection)
Swiss
3.3 ± 1.3 (21)
0.0 ± 0.0 (22)
–
–
–
–
–
–
Beijing USDA
2.7 ± 0.9 (26)
0.4 ± 0.4 (23)
–
–
–
% Nymphs emerged
Swiss
0.0 ± 0.2 (21)
3.6 ± 3.2 (22)
–
–
–
–
–
–
Beijing USDA
0.0 ± 0.0 (26)
0.0 ± 0.1 (23)
–
–
–
% Dead eggs (undefined content) (dissection)d
Swiss
15.7 ± 7.0 (21)A
7.3 ± 3.8 (22)A
χ2 = 18.66
χ2 = 3.56
χ2 = 0.36
df = 2
df = 1
df = 2
Beijing USDA
10.8 ± 4.8 (26)A
0.4 ± 0.4 (23)A
p < 0.001
p = 0.059
p = 0.835
Control
33.8 ± 5.4 (32)B
24.2 ± 4.6 (31)B
Single naïve, mated females of both T. japonicus lines were exposed to a host egg mass (ten eggs) of either host species in no-choice assays
For each variable, the total number of cases (number accepted egg masses) or the mean ± SE (all others), the sample size (in brackets), and the results of statistical tests are indicated
aUnparasitized egg masses were excluded. Eggs were defined as parasitized, when drilled and marked
bWasps that did not parasitize all ten eggs were excluded. Multiparasitism events after initial parasitism of all ten eggs were not considered for handling time calculation
cMultiparasitized eggs were defined as eggs that were drilled and marked at least twice
d“Control” indicates the percent egg mortality per egg mass in rearing controls of either host species. Capital letters indicate significant differences between treatments
The comparison of host suitability for development for both wasp lines is displayed in Table 1. Adult emergence was higher when developing on P. rufipes eggs for both wasp lines. The proportion of female offspring was significantly higher for the Swiss line than for the Beijing USDA line, independent of the host species. Further, the unattributed mortality (i.e., dead eggs containing undefined content) for each wasp line and host treatment combination was significantly lower than the observed egg mortality in the two respective rearing controls.
Host contact kairomone assay
The total leaf residence time was significantly dependent on the host kairomone treatment (χ2 = 156.06, df = 2, p < 0.001). Naïve T. japonicus females of both lines spent significantly more time searching on leaves with H. halys kairomones than on leaves with kairomones of P. rufipes (Swiss: p = 0.027, Beijing USDA: p = 0.001, Fig. 1). Furthermore, we observed a significant interaction between the wasp line and the kairomone treatment (χ2 = 13.67, df = 2, p = 0.001). Beijing USDA females spent significantly less time on control leaves than on leaves treated with kairomones from either stink bug species (both p < 0.001). In contrast, Swiss females spent significantly less time on control leaves in comparison to leaves with H. halys kairomones (p < 0.001), but not to leaves with P. rufipes kairomones (p = 0.056). In addition, independent of the stink bug species, females of the Beijing USDA line spent on average significantly more time on leaves with stink bug kairomones (343.02 ± 23.13 s, mean ± SE) than females of the Swiss line (213.40 ± 21.53 s; wasp line: χ2 = 22.97, df = 1, p < 0.001). Furthermore, as expected, older females stayed significantly longer on leaf surfaces than younger females (female age: χ2 = 18.01, df = 1, p < 0.001). However, the increased leaf residence time with increasing wasp age was only observed for leaves with host footprints but not for control leaves (kairomone treatment × female age: χ2 = 7.86, df = 2, p = 0.020, Fig. 2). There were no age-specific differences in the behavioural response of the Swiss and the Beijing USDA line (wasp line × female age: χ2 = 0.02, df = 1, p = 0.899).
Fig. 1
Residence time of Trissolcus japonicus females of the Beijing USDA line (grey) and the Swiss line (white) on maple leaves with either no host kairomones (control; nBeijing USDA = 19, nSwiss = 19), or contact kairomones of its target host Halyomorpha halys (nBeijing USDA = 24, nSwiss = 20) or non-target host Pentatoma rufipes (nBeijing USDA = 23, nSwiss = 20). Boxplots visually summarize the distribution of the data. Each box represents the interquartile range (IQR), which spans from the first quartile (Q1) to the third quartile (Q3), with the median indicated by a line inside the box. Whiskers extend to the smallest and largest observations within 1.5 times the IQR from the first and third quartiles, respectively, to show the range of the data. Any data points beyond this range are considered outliers and are depicted as individual points. The black dots show model predicted means of the different treatments with error bars representing SE. Different letters indicate significant differences among all treatment means with Tukey’s HSD adjusted p-values and α = 0.05
Fig. 2
Effect of female age on residence time of Trissolcus japonicus females of the Beijing USDA line (grey dots) and the Swiss line (black triangles) on maple leaves with either no host kairomones (control; nBeijing USDA = 19, nSwiss = 19), or contact kairomones of its target host Halyomorpha halys (nBeijing USDA = 24, nSwiss = 20) or non-target host Pentatoma rufipes (nBeijing USDA = 23, nSwiss = 20). Solid and dotted lines denote fitted regression lines with 95% confidence intervals for the Beijing USDA and the Swiss line, respectively, relating leaf residence time to female age
×
×
Discussion
With the mass releases of the T. japonicus Beijing USDA line in Italy starting in 2020 (Sabbatini-Peverieri et al. 2020; Falagiarda et al. 2023), two genetically distinct lines of T. japonicus currently occur in the same geographic area in Europe (Stahl et al. 2019; Abram et al. 2023). To date, little is known about the effects of intraspecific variation between the two lines with regard to biological control of H. halys, including host specificity. The present study represents the first direct comparison between the released and adventive line regarding host acceptance, host suitability and response to host contact kairomones of H. halys and the most frequently attacked non-target host of T. japonicus in Europe, P. rufipes (Haye et al. 2020, 2024; Falagiarda et al. 2023).
P. rufipes was readily accepted by both T. japonicus lines and highly suitable for parasitoid development with a mean parasitoid emergence rate of 96% (SE = 2%). In addition, P. rufipes was more suitable for parasitoid development than H. halys for both wasp lines, confirming previous results of laboratory host range studies (Haye et al. 2020). Despite the statistically significant effects, differences in the observed number of accepted host egg masses and eggs parasitized within an egg mass between wasp lines and host species were small (Table 1) and likely of little or no biological relevance. The recorded statistical significance is most likely attributed to a lack of variance in the data (i.e., number of accepted host egg masses and proportion of parasitized eggs) and few strong outliers (i.e., proportion of multiparasitized eggs).
Overall, there is little indication for biologically relevant intraspecific variation in host acceptance and suitability, and thus we expect that both T. japonicus lines successfully parasitize H. halys and P. rufipes egg masses upon encounter in the field. This conclusion is supported by field observations of parasitized sentinel or natural host egg masses of both host species and by both lines (Falagiarda et al. 2023; Haye et al. 2024). A more comprehensive comparison of the wasp’s fundamental host range using additional non-target hosts or other genetic T. japonicus lines that are currently not present in Europe may reveal different results.
Because the ecological host range of scelionid egg parasitoids such as T. japonicus is largely determined by the wasps’ ability to successfully locate host resources that are suitable for parasitoid development (Conti et al. 2004; Salerno et al. 2006), host foraging behaviour should also be considered in risk assessment studies. In the present study, kairomonal traces of both target and non-target host elicited an arrestment response in females of both T. japonicus lines present in Europe. In the presence of host kairomones, T. japonicus resided about twice as long or more on leaf substrates while searching for a potential host egg mass than on leaves without. However, kairomone traces of H. halys did elicit a stronger arrestment response in comparison to P. rufipes kairomones for both parasitoid lines, resulting in a considerably longer total leaf residence time (Fig. 1). The preference of T. japonicus and other scelionid wasps for chemical footprints of associated hosts over non-associated hosts, and non-associated hosts over control treatments, has been demonstrated in several studies (Colazza et al. 1999; Salerno et al. 2006; Scala et al. 2022). For instance, females of the Beijing USDA line and an adventive North American line of T. japonicus reduced their walking velocity and increased their residence time on substrates more strongly upon encountering contact kairomones of H. halys in comparison to kairomones of a non-associated host or control treatments (Boyle et al. 2020; Malek et al. 2021).
The present study bears further evidence for the ability of foraging Trissolcus spp. to discriminate between different host cues to locate suitable oviposition substrates (Scala et al. 2022). The differences in their behavioural response to host-related semiochemical cues are likely the result of varying chemical composition or concentration of chemical compounds of the cues being associated with different host stages (Tognon et al. 2016; Zhong et al. 2017) or different host species (Conti et al. 2004; Salerno et al. 2006). For example, Malek et al. (2021) found that the reduced preference of T. japonicus for Podisus maculiventris (Say) (Hemiptera: Pentatomidae) traces correlated with a considerably lower concentration of n-tridecane and (E)-2-decenal in comparison to H. halys traces. Whether the lower response rate to P. rufipes footprints observed in this study is similarly caused by differences in the chemical composition of host footprints remains unknown. Another explanation for the stronger arrestment response elicited by H. halys contact kairomones could be an innate preference for chemical cues of the parental host species due to preimaginal conditioning or early adult learning (Turlings et al. 1993), since host preference shifts upon changes of rearing host were observed previously (Botch and Delfosse 2018; Chierici et al. 2023).
Apart from distinct responses to kairomonal traces of the target and non-target host, we documented evidence for intraspecific variation in host foraging behaviour in T. japonicus. Females of the Beijing USDA line displayed a stronger arrestment response and resided on contaminated leaf substrates at least 1.8 times longer on average than females of the Swiss line, independent of the kairomone treatment (Fig. 1). Furthermore, although not explicitly measured in the present study, females of the Beijing USDA line displayed a notably higher walking velocity in comparison to females of the Swiss line in the absence of host kairomones. As expected, within each T. japonicus line, older females stayed significantly longer on leaf surfaces contaminated with host kairomones than younger ones (Wajnberg 2006; Wajnberg et al. 2006; Zhang et al. 2022). However, the behavioural difference between the two lines described above was not related to the varying ages of tested females, as no age-specific variation in behavioural response was observed between the two wasp lines (Fig. 2). Because all females used in this study were naïve and had no previous host foraging experience influencing their response (e.g., Peri et al. 2016), the observed behavioural differences are likely genetic. Yet, this would have to be verified in a comparison of different female isolines of both T. japonicus lines and their hybrids. For example, genetically determined, intraspecific variation in the patch-leaving tendency was observed between female isolines of Trissolcus brochymenae (Ashmead) and Telenomus busseolae (Gahan) (both Hymenoptera: Scelionidae) (Wajnberg et al. 1999; Sevarika et al. 2021). Differences in patch-leaving tendencies might originate from either a lower initial responsiveness to host kairomone cues (Waage 1979), or a more rapid habituation to chemical cues experienced during foraging on a host patch (Peri et al. 2016; Abram et al. 2017). Other possible explanations for the observed differences include different amounts of time in laboratory rearing (Visser et al. 1992; Thiel et al. 2006) or differences in other traits (e.g., fecundity) that correlate with foraging activity (Minkenberg et al. 1992).
A difficult yet relevant question is how the intraspecific variation in foraging behaviour of the two T. japonicus lines in Europe may affect their capacity for biological control of H. halys and the risk of non-target effects, as well as their coexistence in the field. Bigler et al. (1988) observed a positive relationship between the walking speed of an egg parasitoid in the laboratory and parasitism rates in the field and concluded that a strain’s activity level is positively correlated with its host location efficiency. Similarly, Colazza and Rosi (2001) argued that a more responsive strain of T. busseolae that scans plants faster for host cues and invests more time in host search upon locating highly reliable host cues might be a more efficient biological control agent in the field in comparison to a less responsive strain. Following the same line of argument, the higher activity and stronger arrestment response of the Beijing USDA line of T. japonicus may lead to higher parasitism of H. halys in the field, along with a higher risk of non-target effects on P. rufipes. Contradictory to this, parasitism rates of field collected H. halys egg masses in Northern Italy were higher at sites with previous records of the adventive Swiss line as opposed to sites of mass releases of the Beijing USDA line (Falagiarda et al. 2023). However, whether the observed differences were due to differences in T. japonicus population densities, adaptation to local conditions or biological control efficiency of the two lines remains unknown.
Which T. japonicus line is better adapted to efficiently control H. halys in Europe while minimizing the risk for non-target effects remains difficult to predict, but ongoing post-release studies on the population genetics of T. japonicus in Northern Italy, where both genetic lines occur in sympatry (Sabbatini-Peverieri et al. 2020; Falagiarda et al. 2023), will potentially reveal valuable insights. Since hybridisation between the two lines is a plausible scenario, the effects of hybridisation of T. japonicus lines should be investigated in the future to assess their biological control efficiency and potential non-target effects in comparison to their parental lines.
Acknowledgements
We would like to thank Sandra Younie and Nathan Heuver for technical assistance in the laboratory and for their help in bug collection. We would also like to thank Lukas Seehausen for his help with the statistical analysis and Kim Hoelmer for the USDA laboratory line of T. japonicus, which was provided by Paul Abram. We thank Paul Abram for comments on an earlier version of the manuscript.
Declarations
Competing interests
All authors declare that there is no conflict of interest.
Ethical approval
This article does not contain any studies with human participants or animals (vertebrates) performed by any of the authors.
Informed consent
Informed consent for participation and approval for submission was obtained from all authors.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.