Effect of simulated heat waves on the behaviour of two mirid predators

Extreme climatic events, including heat waves with high‐temperature peaks during the day, are expected to occur more often in many parts of the world due to climate change and may exert negative effects on existing biological control strategies. To assess the effects of high‐temperature peaks on two commonly used, and naturally co‐occurring mirid predators, adults and nymphs of Nesidiocoris tenuis and Macrolophus pygmaeus were exposed either to constant 25°C or daily mean temperatures of 25°C with cycles rising up to 30 or 40°C. Preferred location of the two mirid species on different strata of tomato plants was investigated when they were exposed to the different temperature regimens alone or combined. Activity of mirids under the different temperature regimens was continuously monitored for 48 h using a Drosophila Activity Monitor system. Finally, the efficacy of the mirids preying on Tuta absoluta eggs was measured at the different temperature regimens in a Petri dish assay. Heat waves reaching 30 and 40°C in some instances affected the location of the mirids either when they were on the plant alone or under competition conditions. Locomotory activity of M. pygmaeus in the 40°C treatment was strongly reduced, whereas it remained high in N. tenuis. Effects of temperature peaks on prey consumption were visible in M. pygmaeus nymphs, N. tenuis females and males of both species. We suggest that the different responses of the two species to high‐temperature peaks may reduce their competition and support sustained control when both species are present simultaneously. Moreover, both species were found to be less susceptible to heat peaks when compared to previously reported results for their prey T. absoluta.


| INTRODUC TI ON
Climate change does not only cause higher average temperatures but also leads to a higher frequency and magnitude of extreme climatic events, such as heat waves (Fischer & Schär, 2010;Seneviratne et al., 2014). Because the thermal performance curve (TPC) of insects usually has a distinct shape with an accelerating increase up to an optimum followed by a steep decline, many insects increase in their performance at moderately elevated temperatures (Colinet et al., 2015;Stoks et al., 2017). It is thus predicted that in the non-tropical regions of the world, elevated temperatures lead to an increased number of generations as well as higher activity and metabolism of herbivorous insects and thus an increase in crop losses (Deutsch et al., 2018;Skendžić et al., 2021). However, once the thermal optimum is passed, temperatures quickly reach critical limits for development and survival of insects and therefore maximum temperatures experienced during short-term hot weather events can be more detrimental (Colinet et al., 2015;Stoks et al., 2017;Zhu et al., 2019). Those temperature extremes may lead to decreased energetic efficiency, to altered and potentially costly behaviour, such as shelter seeking and fasting, to decreased and ceased locomotion, and finally to sterility and death due to the denaturation of proteins (Abram et al., 2017;Colinet et al., 2015;Hughes et al., 2010;Ingegno et al., 2021;Sentis et al., 2012).
In an agricultural context, effects on pest control can be negative, when parasitoids and predators are stronger affected by altered climatic conditions than their hosts or prey (Sentis et al., 2013;Skendžić et al., 2021). For biological control, it is therefore important to choose the appropriate agents that can thrive under the thermal conditions that they find in the field or greenhouse environment. Most recently mirid predators have gained increased interest as biological control agents, in particular against the invasive tomato leafminer Tuta absoluta (Lepidoptera: Gelechiidae) (Urbaneja et al., 2012;van Lenteren et al., 2020) despite the fact that they are also known to feed on the plant (Chinchilla-Ramírez et al., 2021;Gillespie & McGregor, 2000;Sanchez, 2008). In Europe, Macrolophus pygmaeus (Rambur) and Nesidiocoris tenuis Reuter (both Heteroptera: Miridae) are commercially available for augmentative release against this pest as well as against whiteflies, aphids, thrips, and spider mites (Ingegno & Tavella, 2019;Pérez-Hedo et al., 2021). Macrolophus pygmaeus is generally used in cooler climates, whereas N. tenuis is used in the Mediterranean region, based on the species' original distribution, and their thermal requirements (Arnó et al., 2010;Messelink et al., 2015;Sanchez et al., 2009Sanchez et al., , 2012, although damage to the crop plants precludes the use of N. tenuis in certain areas, seasons and crop types. Both predator species develop optimally between 27 and 30°C. While above 31°C excessive mortality in M. pygmaeus has been observed, N. tenuis has a high survival at 32°C and can even complete its development at 35°C; the latter is not possible for M. pygmaeus (Ingegno & Tavella, 2019;Martínez-García et al., 2016, 2017. Both species also co-occur in parts of their natural range, for example, in Spain and Greece, on common hosts (i.e. solanaceous plants) and exert control on whiteflies and the tomato leafminer simultaneously (Arnó et al., 2006;Dumont et al., 2021;Ingegno & Tavella, 2019;Lampropoulos et al., 2013). Perdikis et al. (2014) have observed that N. tenuis usually occupies the upper part of the plant, whereas M. pygmaeus preferably forages and oviposits on the middle and lower part. This limits competition and intraguild predation between the two species, which becomes visible at high density and when food is scarce (Dumont et al., 2021;Michaelides et al., 2018;Moreno-Ripoll et al., 2012). It is not clear, however, how this pattern would be influenced by high temperatures that might alter the behaviour of the mirids. We thus assessed how the two species respond to elevated temperatures during the diel cycle and how these temperatures would affect their spatial distribution on the plant, their locomotion activity, and their prey consumption in Petri dishes.
We hypothesized that the preferred locations of the mirids on the plant would be altered by the high-temperature peaks, that the thermophilous N. tenuis would be less affected by the elevated temperatures, and that interactions between the two species would change under high-temperature conditions.

| Experimental set up
All experiments were conducted in climate cabinets (Panasonic MLR 352 H-PE, Labtech Services, Villmergen, Switzerland) to ensure controlled conditions and to comply with Swiss regulations concerning the handling of non-native species.

| Insect rearing
The rearing of T. absoluta was established in the laboratory from individuals obtained by Andermatt Biocontrol AG (Switzerland). The stock colony was maintained on fresh tomato plants inside mesh cages (47.5 × 47.5 × 47.5 cm) (BugDorm-4F4545). Larvae were placed on fresh tomato branches supported by a floral foam on a plate. New branches were added every other day and the floral foams were supplied with water regularly. When the pupal period started, this procedure was stopped until emergence of adults. Adults were collected and transferred to a new cage with fresh plants. The rearing was kept in a walk-in climate cabin at 22 ± 1°C, 70 ± 10% RH and 16:8 L:D. Uniform cohorts of T. absoluta eggs obtained during 24-48 h were used for the experiments.
Nesidiocoris tenuis was obtained from Agrobio S.L. (Spain) and Macrolophus pygmaeus was obtained from Andermatt Biocontrol AG (Switzerland). To rear the predator stock colony, the specimens were kept on tobacco plants inside mesh cages (BugDorm, see above) and feed ad libitum with eggs of Ephestia kuehniella (Zeller) (Lepidoptera: Pyralidae; AGROLINE Bioprotect, Switzerland). To control the age of mirids, they were transferred to a new cage with the same conditions, once the emergence of adults occurred. The rearing of mirids was kept in a climate-controlled cabin at 25 ± 1°C, 70 ± 10% RH, 16:8 h L:D.

| Temperature regimens
To assess the response of the predators to heat waves, we used three temperature regimens as follows: (1) 25°C constant, (2) 25°C average with a peak of 30°C, and (3) 25°C average with a peak of 40°C ( Figure 1). The baseline temperature was chosen to approximate average greenhouse temperatures during the growing season in Switzerland (Janique Koller, Agroscope, personal communication) and to allow for sufficient amplitude of temperature peaks without killing the insects. Light intensity and temperature were increased stepwise. For regimens 2 and 3 with fluctuating temperature, the temperature peak that was held for 2 h coincided with the highest light intensity. For all temperature regimens, the average relative humidity 65 ± 10% and the photoperiod 16:8 L:D. Temperature programs were run within climate cabinets (see above) and actual conditions experienced by the mirids during the assay were recorded using dataloggers (Ecolog TH1, Elpro-Buchs AG, Buchs, Switzerland).

Location on plants
We assessed the location of predators on whole plants under the three temperature regimens. One tomato plant with 10 fully developed composite leaves and 25-30 cm height was introduced into a cage (BugDorm, see above) with about 100-200 adult T. absoluta of mixed sexes for 24 h prior to the experiment, to allow for oviposition.
As young females lay 25 or more eggs per day (Silva et al., 2015) at least 1250 eggs were assumed to be present on each plant as food source for the mirids. One plant with eggs was used per experimental unit.
Adults were not older than 7 days and of mixed sex with either a 1:1 or slightly female biased sex ratio. As nymphs of the two species could not be reliably distinguished in the chosen setup on the plant, no treatment with mixed nymphs was considered. The plants were divided in upper, middle and lower strata, each having 8-10 cm of height depending on the original plant size. After 48 h, the location of the mirids present per plant strata was recorded, as well as the number of dead individuals. This was done visually through the transparent plastic front of the bugdorm and without manipulating F I G U R E 1 Temperature regimens and humidity as measured by dataloggers during the experiment in which constant (25°C) and fluctuating temperatures (with either a peak at 30 or 40°C) were employed to generate heat waves. Light conditions during the experiment were as follows: (1) dark period from 21:00 to 05:00 h, (2) intermediate light period from 06:00 h to 11:00 h and 18:00 h to 20:00 h, and (3) intense light period from 12:00 h to 17:00 h. the plant during the morning in a period with baseline temperature.
Ten replicates per treatment were assessed.

Locomotory activity
To test the hypothesis that the two mirid species differ in their response to temperature and in the threshold temperatures for heat damage, the activity of individual predators was measured using a Drosophila Activity Monitor (DAM) system (TriKinetics, Waltham, MA). This system registers how many times each individual crosses an infrared light beam within a glass tube. Each crossing is counted, i.e., a higher activity of the insect becomes visible as a higher count.
Male and female adults less than 7 days old and nymphs from

| Temperature modulating biocontrol efficacy
In the efficacy experiments, females, males, and nymphs of N. tenuis and M. pygmaeus were individually kept in 5-mL plastic tubes to starve for 1 h before release into the experimental arena. Up to 7-day-old adults of both predators were used in the experiments.
A Petri dish (7 cm diameter) served as experimental arena containing tomato leaflets placed on a solid water agar solution. 100 fresh (<24 h old) eggs of T. absoluta were placed on the tomato leaflet using a fine brush. A single predator was released into each dish in front of the prey and the Petri dishes were covered with a film of PVC paper.
Several small holes were pinched into the PVC paper using a fine needle to ensure air circulation. After 24 h, the predation rate was estimated by counting the number of intact eggs under a stereomicroscope (Leica MZ 125 10X). The quantity of the eggs was chosen according to Michaelides et al. (2018), who reported that N. tenuis and M. pygmaeus are able to prey between 70 and 90 eggs at maximum within 24 h at 25°C. To determine how varying temperature modulates their biocontrol efficacy we selected a density with excess eggs for both mirid species.
The following experimental treatments were set up: (1) female Mp, (2) female Nt, (3) male Mp, (4) male Nt, (5) nymph Mp, (6) nymph Nt. Adults were not older than 7 days and nymphs were in the 4th or 5th instar. For each treatment, 20 replicates were considered, no mortality occurred during the 24 h experimental period.

| Statistical analyses
Mortality of mirids during the on-plant experiment on predator location was modelled separately for each species and separately for nymphs and adults, since nymphs were only kept alone but not mixed. Generalized linear models (GLM) with binomial distribution were employed to assess the effects of the fixed factors temperature and -only for adult mirids alone/mixed as well as their interaction on the binary outcome of survival (dead/alive). Non-significant interaction terms (p > 0.05) were excluded from the final models.
An analysis of deviance was conducted for all models followed by Dunn post-hoc tests. To assess the preferred location of mirids on the plants, the number of surviving individuals per stratum was counted at the end of the 48 h experimental period. The location of individuals at the different plant strata was analysed separately for N. tenuis and M. pygmaeus adults and nymphs alone and compared across temperature regimens using contingency tables. Also with contingency tables, the location of adult individuals of the two species and at different temperature regimens was analysed separately to assess the effect of alone/mixed on their preferred location. To account for the twofold use of data, the significance threshold was corrected to p = .025.
Mortality of mirids in the experiment on locomotory activity was modelled separately for each species using generalized linear models (GLM) with binomial distribution to assess the effects of the fixed factors temperature and -only for adult mirids sex/stage as well as their interaction on the binary outcome of survival (dead/ alive). An analysis of deviance was conducted for all models followed by Dunn post-hoc tests. Movement of mirids was analysed as total count during the 48 h experimental period. To account for unequal variances, GLMs with negative binomial distribution and log link function were used to analyse the effect of the fixed factors sex/ stage and temperature as well as their interaction on the counts of movement for each species separately. Differences between day-and nighttime activity were analysed by comparing the mean number of movement-counts during the 8 h of total darkness (21:00 to 5:00) and the 5 h of highest light intensity (12:00 to 17:00) with Wilcoxon signed rank tests for each species, stage/sex and temperature regimen.
To assess predation, the number of predated eggs was counted over the full experimental period of 24 h. To account for unequal variances, GLMs with negative binomial distribution and log link function were used to analyse the effect of the fixed factors sex/ stage and temperature as well as their interaction on the number of predated eggs for each species separately. Due to significant interaction effects separate models were run for females, males and nymphs, followed by Tukey's post hoc tests adjusted for a family of 3 estimates to analyse the effect of temperature.
All statistical analyses using generalized linear models were performed in R version 4.0.2 (R Core Team, 2020) contingency analyses and Wilcoxon signed rank tests were conducted with SPSS version 26 (IBM, 2019). All raw data are stored in a public repository (Duarte .

| Location on plants
Across all temperature treatments and sex/stages, 28 ± 19% (mean ± SE) of M. pygmaeus and 26 ± 19% of N. tenuis died during the experimental period of 48 h (Table 1). In M. pygmaeus adults kept alone and in nymphs, temperature regimen had a significant influence on mortality. In both cases, lower mortality was observed in individuals exposed to peaks of 30°C compared to those at constant 25°C. Furthermore, significantly higher mortality of adults was observed in the mixed-treatment as compared to the treatment in which adults were placed on the plant alone (χ 2 = 6.3515; df = 2, p = 0.012).
In adult N. tenuis no effect of temperature regimen on mortality was visible. However, nymphs of N. tenuis were significantly affected by

| Locomotory activity
The mortality during the locomotory assay was lower in M. pygmaeus (12 ± 6%; mean ± SE) across all temperature treatments and sex/ stages compared to N. tenuis (29 ± 4%) with the highest mortality recorded in male N. tenuis at 30°C (50%) ( Table 2). In M. pygmaeus the mortality was significantly affected by temperature regimen (χ 2 = 12.1377; df = 2, p < 0.001) but not by sex/stage or the interaction of sex/stage with temperature. No significant differences were detected during multiple comparisons of the temperature regimens (Dunn test). In N. tenuis mortality was significantly affected by sex/ stage (χ 2 = 31.2940; df = 2, p < 0.001) as well as by the interaction of sex/stage and temperature regimen (χ 2 = 7.8481; df = 4, p = 0.020) but not by temperature regimen. Again, no significant differences were detected during multiple comparisons of males, females and nymphs (Dunn test).
In the overall models for M. pygmaeus, the locomotory activity differed significantly for the factors sex/stage (χ 2 = 87.174, df = 2, p < 0.001) and temperature regimen (χ 2 = 94.901, df = 2, p < 0.001) and a significant interaction between both factors was observed (χ 2 = 28.222, df = 4, p < 0.001) (Figure 3). When separated by sex/  (Figure 4), whereas no such clear pattern was visible for N. tenuis ( Figure 5). Significantly higher nighttime activity was detected for M. pygmaeus nymphs at 30°C, for females at 25 and 30°C and for males at all temperature regimens (Table 3). In N. tenuis higher nighttime activity was only visible in nymphs at 30°C, and males at 25 and 30°C. Females displayed higher daytime activity at 40°C (Table 3).

| Temperature modulating biocontrol efficacy
Overall, N. tenuis consumed more eggs within the 24-h experimen-

| DISCUSS ION
Our results revealed that daily heat peaks compared to comparable (i.e. same daily mean) constant temperature regimens affected the on-plant location, locomotion and predation activity of some of the stages of the mirid predators M. pygmaeus and N. tenuis.
When ten individuals of one species were present on a plant, temperature significantly affected mortality of M. pygmaeus adults and nymphs as well as of N. tenuis nymphs. In all cases, mortality at 25°C was higher than at a regimen with a temperature peak at 30°C, indicating that the mirids actually benefitted from the fluctuating intermediate temperature, while no clear negative effect of the 40°C regimen was found. It is known that fluctuating temperatures within the permissive range generally improve insect performance (Colinet et al., 2015). This is based on the unequal form of the thermal performance curve as well as other factors such as adjustments in gene expression (Liefting et al., 2017 (Gols et al., 2021).
Thus, the availability of microclimatic refuges is of vital importance (Thakur et al., 2020). For example, Tetranychus urticae Koch (Trombidiformes: Tetranychidae) seek out places with optimal temperature on apple leaves (Caillon et al., 2014). Similarly, aphids are more likely to drop from plants under heat stress, which was suggested to be a form of behavioural thermoregulation to reach cooler microclimates (Ma & Ma, 2012).
Within the temperate range, activity in ectothermic insects generally increases with temperature (Abram et al., 2017). Accordingly, mirids faced with higher temperatures respond with higher activity (i.e. faster walking speed) compared to lower activity at cold temperatures (Hughes et al., 2010;Ingegno et al., 2021). However, once a critical threshold is passed, locomotion becomes impaired rapidly, up to a status of heat coma, where insects can no longer escape from harmful situations (González-Tokman et al., 2020;Gunderson & Stillman, 2015;Ingegno et al., 2021). In the experiment measuring locomotory activity, temperature peaks had no significant effect on N. tenuis adults and little effect on nymphs, whereas all stages of M. pygmaeus were significantly impaired in their locomotion in the regimen with a peak at 40°C. Therefore, it can be assumed that a certain extent by changing their location, whereas they were fully exposed in the tubes and dishes for the other experiments.
Biological processes ranging from growth and development to feeding and dispersal are determined by metabolic processes directly influenced by temperature. Heat peaks thus influence physiology as well as behavioural interactions between species and trophic interactions (Gillespie et al., 2012;Sentis et al., 2013). Sometimes behavioural changes may explain observed ecological changes related to increased temperatures even better than the physiological and metabolic processes alone (Abram et al., 2017). In the case of predators, extreme heat events have been reported to have a strong effect on aspects such as locomotion, biology, niche preference and efficacy (Damien & Tougeron, 2019;González-Tokman et al., 2020;Halsey, 2016;Laws, 2017). In our study, we found strong impacts of the heat peaks on the mirids evident in altered locomotion, diurnal rhythm, preferred location and altered efficacy. It is likely, that some of these effects are buffered under natural conditions when mirids are able to seek out their preferred habitat or have access to various food sources. Yet the observed effects point to abilities to deal with temperature peaks in the two species. Both predatory mirids act against tomato pests in greenhouses and compete for prey when co-occurring in the same habitat (Duarte Martínez et al., 2022). It is possible that in the case of N. tenuis and M. pygmaeus the competition would be reduced, due to the different preferred location on the plant, different response to temperature peaks and different daytime activity. Indeed, enhanced control of T. absoluta due to complementarity of the two predators was suggested (Konan et al., 2021;Lampropoulos et al., 2013) and sustained biocontrol efficiency over 10 weeks was observed when both predators were applied together in greenhouse compartments (Yao et al., 2022

ACK N O WLE D G E M ENTS
Leticia Duarte Martinez was funded by the Swiss Government Excellence Scholarship Program. Mario Waldburger assisted with plant and insect rearing. Antonio Biondi helped to establish the initial mirid rearing. Open access funding provided by Agroscope.

CO N FLI C T O F I NTE R E S T S TATE M E NT
The authors declare that they have no conflicts of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
All raw data are stored in a public repository: https://doi.