Interspecific hierarchies from aggressiveness and body size among the invasive alien hornet, Vespa velutina nigrithorax, and five native hornets in South Korea

The range of the invasive alien hornet, Vespa velutina nigrithorax, has been expanding since its introduction to Korea in 2003. Here, we compare the aggressive behaviors and body size of V. velutina nigrithorax with five native hornet species to identify the interspecific hierarchies that influence the rate of spread of this species. Aggressive behaviors were classified into 11 categories, and each interaction was scored as a win, loss, or tie. We found that V. velutina was superior to V. simillima in fights that V. velutina won and showed a high incidence of threatening behavior. V. mandarinia outperformed V. velutina in fights that V. mandarinia won and grappling behavior was common. V. analis was superior to V. velutina in fights that V. analis won and showed a high degree of threatening behavior. V. crabro was superior to V. velutina in fights that V. crabro won and showed a high rate of threatening behavior. V. dybowskii was superior to V. velutina in fights that V. dybowskii won and showed a high incidence of threatening and grappling behaviors. The body size of V. velutina was greater than V. simillima (although not statistically significant) and smaller than all other Vespa species. Therefore, according to this study, the low interspecific hierarchies of V. velutina seem to be a major cause of the slower spread in Korea than in Europe. However, over time, its density has gradually increased within the forest, where it seems to be overcoming its disadvantages and expanding its range, possibly because the large colonies and good flying abilities of this species help it secure food.


Introduction
In recent years, invasive alien species (IAS) have spread widely due to rapid climate change and global trade, resulting in a global biodiversity reduction, and economic and ecological impacts [1][2][3]. Indeed, IAS contribute to millions of dollars in economic losses per year [4], with negative impacts from insects and arthropods costing 70 billion US dollars annually [5]. In South Korea, the social and agricultural impacts are also gradually increasing due to Therefore, in this study, we analyzed the interspecific hierarchies of V. velutina among Korean Vespa species by measuring aggressive behavior to secure food sources among native hornet species and V. velutina, in order to understand competitive ability, which is the main factor determining the successful spread of V. velutina. In addition, we analyzed the correlation between fighting ability and size by measuring the body size of each species.

Study species and experiment sites
We planned to test for aggressive behaviors between V. velutina and nine native Korean Vespa species, but because of their different distributions, it was impossible to observe the behavior of all nine species in one place. In particular, Vespa simillima xanthoptera, Vespa binghami, and Vespa crabro crabroniformis have very limited distributions, making them difficult to compare [45]. Therefore, experimental sites at which many Vespa species, including V. velutina, are present were selected by referring to various studies such as those of Choi et al. [45] and Choi and Kwon [46]. As a result, we conducted the experiment in the Piagol Valley of Jirisan National Park, where seven Vespa species (V. velutina, V. simillima, V. crabro, V. dybowskii, V. mandarinia, V. analis, and V. ducalis) occur.
First, we randomly chose three experimental sites in Piagol Valley (Site A: N35˚13'37.96 "E127˚35'48.95", 174m; Site B: N35˚15'17.53 "E127˚35'58.24", 417m; Site C: N35˚15'42.43 "E127˚35'3.57", 398m, Since this experiment was conducted on a private site outside the boundaries of a Jirisan National park, it was conducted with the personal permission of the landlord. Therefore, there is no documented permit). Vespa species were captured using hornet traps in the apiaries near each site, and identified. As a result, four species (V. velutina, V. simillima, V. mandarinia, and V. crabro) were captured in the apiary near site A, six species (V. velutina, V. simillima, V. dybowskii, V. mandarinia, V. analis, and V. crabro) were captured in the apiary near site B, and five species were captured in the apiary near site C (V. velutina, V. simillima, V. mandarinia, V. analis, and V. crabro). Therefore, site B was selected as the experimental site, as it had the highest species diversity. V. ducalis was also seen flying near site B but did not appear in traps or the test site. Therefore, the behavioral experiment was conducted between V. velutina and five native hornet species.

Behavioral observation experiment
The behavioral observation experiment apparatus was set up as follows: a table (1m in height) was installed on a flat plain in a forest where hornets were present, 2-3 sheets of toilet paper were placed on top of the table, and an attractant was poured onto these, following the methods described by Choi et al. [19]. In addition, to increase hornet attraction, a 500-ml nebulizer was filled with attractant, and the liquid was sprayed for approximately 10-20 minutes before observations began. The attractant was composed of 1: 1: 1 brown sugar water, vinegar, and ethanol, which mimics oak sap. It is the most commonly used substance for attracting and capturing Vespinae species in hornet traps in Korea, and it has very little attractiveness bias for a particular species [19,47].
This experiment was conducted only with workers, excluding gynes and males, to get rid of the differences in attack behaviors due to caste differences. Therefore, we conducted our study over a total of four days from August 12-13 and August 17-18, 2017, before the gynes and males came out in mid-September. The experiment was conducted at 8-10 am and 5-7 pm because workers tend to avoid outdoor activities when daytime temperatures exceed 35 degrees. The experiment was conducted eight times in total.

Behavioral description, intensity scores, winning percentages, and aggressive behavior trends
As there were no previous behavioral descriptions of the aggressive behavior of hornets, we prepared the first description based on Jang et al. [48] and Goyens et al. [49], by analyzing images from our experiment. Images of hornet behavior were taken by recording hornets with a camcorder (Digital Camcorder V 2000, Sunwoo Tech. Crop. Goyangsi, Korea), placed 50 cm in front of the experiment table.
Aggressive behavioral interactions were classified as wins, losses, and ties. Losses was given 0 points, and ties given 1 point. For wins, threatening behaviors were given 2 points, grappling behaviors were given 3-4 points, and killing was given 5 points. These intensity scores are summarized in Table 1. Winning percentages were calculated, showing the percentage of wins, losses, and ties. In addition, the tendency towards aggressive behavior between the two species was expressed in the radar chart of the number of each behavior. The score for each behavior is derived from Table 1

Morphological measurements
To measure the body size of the six species, we captured them using an insect net following the behavioral observation experiment and stored them in 95% alcohol. Both winners and losers were included in the sampling. If the test subjects were missed, the remaining opponents were excluded from the measurement. Captured samples were dried at room temperature (20˚C-25˚C) for one week in the laboratory and then pinned. The lengths (mm) of the specimens were obtained by measuring only the head width and thorax length through a stereoscopic microscope (Olympus optical, JP SZ61, Olympus Korea, Seoul, Korea), as the hornet's abdomen was elastic and could affect its body size. Body size of the five native species was measured by selecting 30 individuals from the samples collected during the experiment. In addition, V. velutina was selected from the individuals who fought the five native species, and a total of 30 individuals were measured.

Statistical analysis
We performed an independent t-test on the intensity scores of V. velutina and the five native hornets to verify the significance of any differences in aggression between species. In addition, the size differences among the six hornets, including V. velutina were analyzed using a one-way analysis of variance (ANOVA) with a post hoc Tukey's honest significant difference (HSD) test. All analyses were performed using IBM SPSS Statistics software version 23.0 (IBM, USA).

Description of aggressive behavior
Aggressive behaviors between V. velutina and native hornets were classified into three categories: threatening, grappling, and killing. The full list of all behaviors within these categories are listed in Table 1.
First, threatening occurred when the two individuals maintained a distance from each other and ate the attractant, and as the distance between them became closer, they threatened their opponents without making direct contact. The following behaviors were considered threatening: an individual moving forward, lifting its antennae and front legs, shaking its wings, or opening its mandible (Fig 1B and 1F). In these cases, the individual was considered to have won when the opponent fell back or ran away. An individual flying above its opponent, or chasing the opponent, were also considered threatening behaviors. In these cases, the individual was considered to have won when the opponent ran away. Interactions were considered a tie when opponents threatened and fell behind each other.
Grappling occurred between two individuals if the distance between them became very narrow or touched. They flew into the air, either fighting or chasing each other and throwing their opponent. Pushing their opponent's head, hitting, throwing, climbing on top of, biting, or attacking their opponent with a stinger were all considered grappling behaviors (Fig 1A, 1D and 1E). In these interactions, the individual who fell back, ran away, fell, or was bitten was Finally, killing occurred when an overwhelmingly strong individual fought with an opponent, and killed it (Fig 1C). If predation failed or both individuals died, this was a tie.

Aggressive intensity, winning percentages, and aggressive behavior trends
In V. velutina vs. V. simillima, V. velutina was superior to V. simillima in 153 fights (t (304) = 9.89, P < 0.001 , Fig 2A), and won 71% of the encounters (Fig 3). V. velutina had a high rate of threatening behavior, such as rushing the opponent, and lifting the antennae and front legs and shaking the wings. V. simillima, however, rushed the opponent more often than V. velutina (Fig 4A).
In contests between V. velutina and V. crabro, V. crabro was superior to V. velutina in 93 fights (t (184) = -9.62, P < 0.001, Fig 2D) and won 73% of the encounters (Fig 3). Lifting antennae and front legs and shaking wings were the most common threatening behaviors in V. crabro, and banging or pushing with the head were the most common grappling behaviors. In contrast, V. velutina displayed minor aggressive behavior such as threateningly flying over the opponent (Fig 4D).
Finally, in encounters between V. velutina and V. dybowskii, V. dybowskii was superior to V. velutina in a total of 132 fights (t (262) = -24.22, P < 0.001, Fig 2E) and won 91% of the contests (Fig 3). The most common threatening behavior employed by V. dybowskii was chasing the opponent, and the second most common action was rushing the opponent; chasing and grabbing were the most common grappling behaviors for this species. V. velutina's counterattack action was to fly threateningly over the opponent but not vigorously or often (Fig 4E).

Spread effect of V. velutina on interspecific competition in Korea
The invasion of IAS generally results in interspecific competition for resources with similar native species [50]. If the native species are strong, invasive species tend to avoid competition temporally or spatially, but if the native species are weak or similar, exploitation or competition may interfere [29,51], resulting in interspecific displacement [50].
In Europe, where V. velutina was introduced in 2004, there were already two native hornet species, V. crabro and V. orientalis, so that the two species were in close competition because of their similar ecological niche, due overlaps in food preference, nesting sites, and life history [34,35]. However, in Korea, where there are nine native hornets, V. velutina clearly differ in their ecological niche. V. mandarinia is the most aggressive among hornets, with the highest hierarchy, nesting in the ground, and mainly foraging on honeybees and large beetles [37]. V. crabro nests in the ground, under trees, and V. analis nests under grasses and leaves, hunting small and medium-sized insects such as bees, flies, and moths [37]. V. dybowskii nests in tree trunks and often display social parasitism [37]. V. velutina, however, nests at the tops of trees, with a lifespan of about one month longer and a higher rate of honeybee foraging [28,52]. V. simillima are phylogenetically closest to V. velutina [53] and have the most ecological overlap because of similarities in life history, nesting, and food preference. Therefore, V. simillima may be most severely affected by V. velutina [19]. According to the results of this study, V. velutina is less aggressive than the other four species except V. simillima, indicating that the ecological niche is low. Therefore, V. velutina avoids competition from four species with strong ecological niche when nesting, as it nests in a place with low competition. Indeed, after the invasion of V. velutina in Busan, Korea, there was a decrease in V. similima [19]. Therefore, after V. velutina's invasion of Korea, its distribution and spread seem to be affected by its ecological niche. From Busan, the point of first invasion of V. velutina, it spread northwards into cities such as Ulsan, Gimhae, and Changwon, where the road network is directly connected, and the east and south coasts [19,45,54]. Since there are nine native hornets within Korean forest habitats, V. velutina would have had considerable difficulty invading deep forests and crossing large forests and mountains [54].
This ecological niche is often determined by differences in fighting ability, which is largely influenced by body size. In many animals, large body size is likely to corelate with physical fighting ability, which can lead to intraspecific and interspecific resource competition. This is termed resource-holding potential (RHP) [55,56]. For example, in various insects, larger individuals in the intraspecific have higher aggressiveness and higher RHP than small ones, indicating that they are competitive [48,49,57,58]. Large invasive species have been shown to spread more easily in nature with competitive advantages at higher RHP than smaller native species [59]. Therefore, in this study, as V. velutina is a similar size to or larger than V. simillima, but smaller in size than the other four native hornets, its RHP seems to be relatively low, as shown in the aggressiveness results of this study.
Therefore, V. velutina appears to avoid direct competition with native hornets, inhabiting urban centers with relatively low-density hornet populations, or on mountain edges adjacent to cities. According to Choi et al. [19], in Busan, the occurrence rate of V. velutina was overwhelmingly higher in cities than in forests.
Meanwhile, the low interspecific hierarchy of V. velutina may have influenced their initial spread. In general, the initial incubation period is determined according to the presence or absence of competing species after the invasion of alien species, and after a certain period of time, when the alien species have a firm ecological niche, the population expands and spreads rapidly [60]. Thus, in Europe where the competition was V. crabro at the beginning of the invasion [34, 35], the annual spread rate was 60-80 km/year [61,62]. In Korea, however, the rate of spread is slower than that of Europe at 10-20 km/year, because 63% of the land is forested, and nine Vespa species already inhabit forests [19,63,64]. The forest landscape of Korea and presence of many competing hornet species had the effect of prolonging the incubation period. Thus, for V. velutina, one of the main reasons why it could survive without natural extinction, despite its low ecological niche, seems to be its choice of habitat in cities with low competition among hornets during incubation period.
V. velutina has ended its initial incubation period, and the rate of spread has increased rapidly from 2008-2010 [64], culminating in its distribution throughout South Korea as of 2018 [20]. Of course, this spread may have been affected by human factors, such as human activity or transport [61], but competition among hornets seems to be the main reason.
However, despite its low aggression and small size, V. velutina population density tended to increase not only in urban areas but also in forested areas over time, and this trend is present in most parts of southern-central Korea [46].
Many invasive alien social insects, such as ants, hornets, yellowjackets, and paper wasps, have large colonies, high polygyny rates, or invasive generalist insect predators, which often lead to dominant distribution and species displacement with successful invasion. This is because alien species with larger colonies than native species are much easier to feed [16,17,[65][66][67][68][69]. Although V. velutina has a lower ecological niche than native hornets, the colonies are two to four times larger [19,70], which is an overwhelming number of individuals that can obtain food resources in nature. In addition, V. velutina has a good flight ability compared to other hornets [71,72] and hunts for food quickly and safely through hawking [46,73], making it very efficient. Altering the concentration and composition of alarm pheromones in areas it has invaded also creates an effective defense strategy against potential predators [74,75]. Therefore, in the early stages of the invasion, it spread to urban areas, competing with native hornets, but due to its large colonies and high foraging ability, they seemed to overcome their disadvantages and gradually increase their abundance even in the forest.

Interspecific hierarchies of Korean Vespa species
In this study, the interspecific hierarchies of Korean Vespa species was determined in part through the aggression behavior of six Vespa species, but the remaining four Korean species were not tested [45]. However, of these four species, V. crabro crabroniformis is a subspecies of V. crabro (the V. crabro used in this experiment is V. crabro flavofasciata), and V. simillima xanthoptera is a subspecies of V. simillima (the V. simillima used in this experiment was V. simillima simillima). These are very similar to the two species we observed, so we expect there to be little difference in ecological characteristics between them. In particular, V. simillima xanthoptera only inhabits Jeju Island in Korea, and has no geographical overlap with the mainland species. V. binghami was first reported in Korea in 2011 [76], and there is a lack of ecological information on this species, which requires further study. Finally, V. ducalis was not addressed in this study, but Yoshimoto and Nishida [39] showed it was less aggressive than V. crabro and V. analis, and Matsuura [77] showed it was less aggressive than V. simillima. It is therefore likely to be the least aggressive among the Vespa species.
As a result, according to Matsuura and Yamane [37], the order of hierarchies Vespa species in Japan is V. mandarinia > V. crabro > V. analis > V. simillima.
However, in this study, our results suggested the order for Korean Vespa species should be V. mandarinia > V. dybowskii > V. analis > V. crabro > V. velutina > V. simillima (Fig 6). Although the order of hierarchies V. analis and V. crabro changed in comparison with the order in Matsuura and Yamane [37], the sequence is similar. If V. ducalis were to be added here, we would expect the order to be V. mandarinia > V. dybowskii > V. analis > V. crabro flavofasciata (or V. v. crabroniformis) > V. velutina > V. simillima (or V. s. xanthoptera) > V. ducalis.
Supporting information S1 Table. Distribution of aggressiveness scores between an invasive alien hornet, V. velutina, and five native Korean hornet species. (DOCX)