Characterizing thermal tolerance in the invasive yellow-legged hornet (Vespa velutina nigrithorax): The first step toward a green control method

The yellow-legged hornet, Vespa velutina nigrithorax (Hymenoptera: Vespidae, Lepeletier 1836), is native to Southeast Asia and has been unintentionally introduced in France. The species is spreading in many areas of the world. The European Union has classified V. velutina as a species of concern because the hornet significantly affects beekeeping activities, mostly by preying honeybees (Apis mellifera) at beehive entrances. No current control method is simultaneously eco-friendly and effective. Here, we aimed to develop a greener technique for destroying V. velutina nests, inspired by a defense behavior used by the eastern honeybee (Apis cerana), the “heat ball”. In the laboratory, we tested how V. velutina of different sexes, castes, and developmental stages responded to different heat exposure systems employing a range of temperature levels. Overall, the time of death decreased as temperature increased. Hornets died faster when the temperature was gradually increased than when it was instantaneously increased; larvae seemed to be more thermally tolerant. The most promising and potential technique for quickly destroying hornet nests may be steam injection, as the humid airflow system killed all hornets within 13 seconds, and therefore could be a good candidate for a green nest control method.


Introduction
Invasive species are among the leading threats to native wildlife [1][2][3], and their presence requires ever more intensive management approaches. A species is considered to be invasive when it spread over a large area outside their native range, producing negative impacts on biodiversity, human health, and activities [1,4,5]. Human travel and international trade are largely responsible for spreading non-native species, frequently across the entire globe, and often unintentionally [6]. Invasive species pose a direct threat by outcompeting native species for food or other resources [7], preying on native species, and causing or carrying disease [8][9][10][11]. The introduction of a predator into a new ecosystem can have disastrous consequences for economic activities [12,13]. Once invasive species have become established, they are

Collection and rearing of hornet colonies
Hornet colonies were collected from June to November 2019 (n = 66) in western central France (Department of Indre-et-Loire; Fig 1 and S1 Table). As the yellow-legged hornet is officially considered as an invasive species, no permits were required for insect collection, and all studies were carried out in compliance with relevant national guidelines. They were brought back to the laboratory and housed in separate boxes left at room temperature (19-22˚C). To determine the identity of each female (worker vs. reproductive), individuals were weighed, and their wing spacing was measured [49]. All types of individuals (workers, gynes, males, and queens) were given daily a solution of water and honey ad libitum. On the day of colony collection, larvae were removed from the combs and used immediately in the experimental trials. The adults were used as quickly as possible after collection (on average 2.2 days, SD±: 3.2). Before each experimental trial, hornets were anesthetized using carbon dioxide (CO 2 ). As soon as a hornet woke up, the trial began.

PLOS ONE
Thermal tolerance in the invasive yellow-legged hornet

Experimental trials
To test the individual thermal tolerance of yellow-legged hornets on the time of death, four heat exposure systems were assessed. Instantaneous & gradual heat. First, we explored the use of instantaneously applied radiant heat (hereafter, the flash system; n = 31 colonies) and gradually applied radiant heat (hereafter, the gradual system; n = 50 colonies); four temperature levels were used (50˚C, 60˚C, 70˚C, and 80˚C). Both systems were applied to the colony as a whole: all sexes (males and females), castes (workers and reproductives), and developmental stages (larvae and adults). For both systems, we performed 30 replicates using gynes, males, workers, and larvae; across the replicates, on average 6 different colonies (SD± 2.07) were represented (n = 120 at each temperature level for each system). Due to the scarcity of queens, they were only used in these first two systems (at each temperature level, n = 5 for the flash system and n = 3 for the gradual system). Second, two direct airflow systems were used to evaluate the influence of humidity on the time of death of adult hornet: one system employed dry heat (hereafter, the dry airflow system), and the other system employed moist heat (hereafter, the humid airflow system). Three temperature levels were used (100˚C, 120˚C, and 140˚C). The dry airflow system was tested using just workers (n = 90), while the humid airflow system was tested using gynes, males, and workers (n = 90). In all the experimental trials, the temperature was measured with a digital thermometer (HI 935005, ± 0.2˚C). All the trials were filmed using a Panasonic camera (4K, HC-VX980) to better quantify the time of death.
For both the flash and gradual systems, a water bath was used (Fisherbrand™ Polystat™ Immersion Circulator [±0.05˚C], Bioblock Scientific). In the flash system, hornets were individually placed in an octagon glass jar (95 x 87 mm) closed with a lid. The jar had been previously placed in a water bath of the target temperature (50˚C, 60˚C, 70˚C, or 80˚C). A probe was inserted into the jar to measure its internal temperature. In the gradual system, hornets were individually placed in a glass tube (90 x 24 mm) at room temperature that was closed with a foam lid. Then, when the water bath reached the target temperature (50˚C, 60˚C, 70˚C, or 80˚C), the glass tube was put into the water. In both systems, we noted the exact time and temperature at which the hornets died. For the gradual system, calibration curves were established (n = 3 for each temperature), allowing us to determine the internal temperature of the glass tubes at the time of death (S1 Fig). The experimental trial ended when the hornet could no longer move. In all experiments, dead hornets were then observed outside the experimental system tested for 15 minutes to confirm their death.
Dry and humid airflow. High temperatures are needed to quickly immobilize the hornets and thus prevent them from attacking or escaping. In the dry airflow and humid airflow systems, high temperatures were combined with airflow. For the dry airflow system, we carried out preliminary tests on hornet nest in the laboratory to assess the potential fire hazard associated with different temperature levels (120˚C, 130˚C, 140˚C, and 150˚C, n = 15 each). Hot air at the target temperature was applied to a fragment of nest comb for 4 minutes using a heat gun (Tacklife HGP72AC, 2000 W/230 V, 500 l/min; n = 15 for each temperature level). The presence or absence of combustion, defined as the appearance of smoke and/or a color change in the comb, was noted. Using this information, we were able to select experimental temperatures that did not pose a fire risk: 100˚C, 120˚C, and 140˚C. These temperatures were tester on hornets from 11 colonies. In the dry airflow trials, a hornet was placed in a cylindrical metal cage (55 x 85 mm) that was located 10 cm away from the heat gun. The trial ended when the hornet could no longer move and died. For the humid airflow system, hornets, from 11 colonies, were placed in a rapid bottle warmer (NUK Thermo Rapid, 500 W, 220V); they were surrounded by wire mesh (50 x 70 mm) and kept 0.9 cm away from the water (also using wire mesh). To create steam (mean temperature: 92.2˚C), 6 ml of water was added to the bottle warmer before the machine was switched on. The interior temperature was recorded using a digital thermometer, as mentioned above.

Statistical analysis
All the statistical analyses were performed using R (v. 3.1.2, R Development Core Team, 2014). The figures were generated using R or SigmaPlot (v. 12.5). First, data normality was assessed using a Shapiro-Wilk test. Then, to analyze how each system affected the time of death of hornet, Kruskal-Wallis one-way analysis of variance (ANOVA) on ranks using the time-to-death data was used; posthoc pairwise comparisons were performed using the Tukey test and the Dunn test. To determine whether there was a relationship between hornet mass and time to death, the Pearson correlation (with normal distribution) and the Spearman correlation (without normal distribution) were used. To assess how well the logarithmic calibration curves obtained using the gradual system reflected internal temperature in the glass tube, we calculated the coefficient of determination (R 2 ).

Discussion
Due to its negative impacts worldwide, the yellow-legged hornet (V. v. nigrithorax) is the target of various control efforts [33]. As there are no signs that the species' expansion is slowing, it is essential to bolster detection and control strategies. In the future, the global economic costs of invasive species will increase rapidly [15,50] and the effects of climate change on the invasive expansion will grow [50]. At present, current methods for controlling the yellow-legged hornet are limited: they are sometimes effective but not eco-friendly. Thus, new techniques must be developed.
Due to their small size and ectothermic physiology, insects are particularly vulnerable to temperature, which can have a direct effect on their behaviors (flight. . .), and fitness (lifespan. . .) [51]. For these reasons, the heat was previously tested in control methods against insect pests in different conditions, as in the museum [52] and food storage [53]. Classically, a true core temperature of above 45˚C sustained for several hours will kill insect adults [53].
In this study, we found that humid airflow was the most efficient system for killing hornets (100% mortality within 13 seconds at 92.2˚C), and it entailed no risk of nest combustion. In contrast, when dry airflow was used, workers survived for up to a minute at 100˚C. Therefore, heat exposure alone does not suffice as a control method because it could give hornets time to flee. In natural condition, various species seem to use three complementary parameters for killing hornets: heat, CO 2 , and/or humidity. Cyprian honeybees (A. m. cypria) utilize heatballs, which combine heat and CO 2 to effectively smother Oriental hornets (V. orientalis) [41]. The same combined approach is used by Japanese honey bees (A. c. japonica) in their heatballs [45] as they defend themselves against Japanese giant hornets (V. mandarinia japonica). Furthermore, Japanese and European honey bees can kill hornets (V. mandarinia, V. analis, V. simillima, and V. crabro) by forming heat-balls in which there is a simultaneous rapid increase in temperature, CO 2 , and humidity [54]. In the same study, it was shown that the giant hornet had a lower median lethal temperature at 90% relative humidity (under artificial laboratory conditions) than at 40-50% relative humidity (mean under average natural conditions). Many social insects can tolerate high temperatures (30˚C to 60˚C) thanks to evaporative cooling [55,56]. Classical social insect behaviors for regulating nest temperature include directing warm air away from the nest (i.e., wing fanning) and evaporatively cooling the nest (i.e., workers collect water and place droplets on the surface of the brood comb, [57]). However, evaporative cooling, is no longer viable when humidity is high; consequently, temperature regulation is disrupted, and lower temperatures become more lethal [55,56].
Here, according to our data obtained in the laboratory, the idea will be to quickly inject steam water into a nest, which should immobilize and kill hornets in few seconds. However, in the field, insects could escape the lethal thermal shock by taking refuge or fleeing. For this reason, it is crucial that any heat diffusion method be rapidly applied. Further studies are necessary to test the efficacy of our proposed control technique in the field on entire yellow-legged hornet colonies.
The data collected in these experiments also demonstrated that a gradual increase in temperature was more deadly to hornets than an instantaneous increase in temperature. Workers and males died significantly faster when the air temperature rose from room temperature to 50˚C, 60˚C, or 70˚C. Indeed, more than 50% of the hornets had died before the target temperature was reached. For example, it took about 170 seconds for the temperature inside the glass tube in the 50˚C water bath to climb from 21˚C to 50˚C (±0.1); however, most of the workers had died before then (93.3%). In contrast, when the temperature was instantaneously set at 50˚C, only 70.6% of workers died. The median lethal temperature (44.6˚C) was reached at around 134 seconds in the flash system versus at around 63 seconds in the gradual system. In an experiment using an incubator, where the temperature rose 1˚C every 20 minutes from an initial setting of 42˚C, it was observed that the yellow-legged hornet's lethal thermal limit was 45±0.48˚C [44], the hornets died after an average of 60 minutes in the incubator. In our study, in the gradual system, the temperature rose 1˚C every 5 seconds until it reached 50˚C. Previous research has shown that initial temperature and the rate of temperature change have a highly significant effect on critical thermal limits [58]. Here, it is possible that quickly and permanently increasing the temperature during extreme thermal stress wore out the hornet's metabolism, preventing the organism from thermally acclimating or regulating its temperature. In contrast to what found with the tsetse fly (Glossina pallidipes) [58], we discovered that slower rates of temperature change (i.e., longer experimental trials) resulted in greater survival in the yellow-legged hornet. At 50˚C, workers died significantly faster than did gynes, larvae, and males. This heat exposure system mimics the functioning of a heat-ball in honey bees (where a maximum temperature of 45.9 ± 1.0˚C [45] is reached within 5-10 minutes [44]). Consequently, in nature, we can assume that it would take more time and effort for honey bees to kill a male or gyne yellow-legged hornet with a heat-ball. Of all the groups considered in this study, the larvae seemed to be the most thermally tolerant. However, because they cannot survive without the presence of adults, the larvae would die of starvation or become prey if all the workers died. Therefore, the temperature needed to kill the entire colony is the maximum temperature needed to kill the workers and the queen early in the season (in Europe: May to late July) and all the adults later on.
Thermal tolerance could be influenced by several natural factors that could not be controlled in this experiment. For example, hornets differ in their phenology and/or health status, and it might be that older workers are less resistant than younger workers. However, age does not appear to influence the critical thermal limits of insects [58]. In this study, there was variability in time to death at lower temperatures (50˚C and 60˚C) for workers and males. This variability was less pronounced at higher temperatures (70˚C and 80˚C). Overall, mass and time to death were not correlated, a result that contrasts with those of certain previous studies in social insects, in which a relationship between mass and thermal tolerance was observed. In the ant Cataglyphis cursor, large workers survived longer than small workers at 48˚C and 52˚C (application of radiant heat; [59]). It is possible that, in the yellow-legged hornet, the difference in mass between early-season hornets (around 350 mg) and end-of-season hornets (around 550 mg) is not large enough to evaluate the above relationship. Indeed, we did observe a trend of heavier workers surviving longer at 50˚C, although the results were not significant. Another possibility is that the lowest temperature tested (50˚C) was already quite high, given that the thermal tolerance of the hornet V. velutina is 45.9 ± 1.0˚C [44]. Such a situation would leave no room for variation in phenology or health status to manifest itself.
In conclusion, we have shown that a method based on steam injection may hold the most promise for controlling the yellow-legged hornet given that humid airflow was capable of killing individual hornets within a few seconds. Such a technique would be eco-friendly, have limited collateral impacts, and pose no risk of combustion. The method's speed will be crucial to its successful use in nature on entire nests. Future research must take place in the field to explore hornet behavior in response to steam injection, ascertain whether hornets have enough time to escape, and determine whether the steam acts upon the entire nest.