Predator–prey relationships between sympatric species allow the evolution of defense behaviors, such as honeybee colonies defending their nests against predatory wasps. We investigated the predator–prey relationship between the honeybee (Apis mellifera ligustica) and the European wasp (Vespula germanica) by evaluating the effectiveness of attack and defense behaviors, which have coevolved in these sympatric species, as well as the actual damage and disturbance caused to the colonies under attack. Attack and defense behaviors were recorded in front of the hive to observe attacks at the hive entrance (68 attacks in 279 h) and at ground level on isolated and weakened honeybees close to the hive (465 attacks in 32 h). We found that V. germanica attacked the hive entrance infrequently due to the low success rate of this strategy and instead preferred a specialized attack method targeting adult honeybees at ground level, demonstrating opportunistic scavenger behavior. Individual honeybees usually responded effectively to an attack by recruiting an average of two nestmates, causing the wasp to flee, whereas collective balling behavior was only observed on four occasions. V. germanica does not appear to disrupt the foraging activity of the colonies under attack. We found that agonistic events supported by other nestmates were typically the most intense ones, involving physical combat and prolonged attacks at the entrance to the hive. These observations support the hypothesis that A. mellifera ligustica can adapt its behavior to match the severity of the threat and the context of the attack.
Citation: Pusceddu M, Floris I, Buffa F, Salaris E, Satta A (2017) Agonistic interactions between the honeybee (Apis mellifera ligustica) and the European wasp (Vespula germanica) reveal context-dependent defense strategies. PLoS ONE 12(7): e0180278. https://doi.org/10.1371/journal.pone.0180278
Editor: James C. Nieh, University of California San Diego, UNITED STATES
Received: December 19, 2016; Accepted: June 13, 2017; Published: July 5, 2017
Copyright: © 2017 Pusceddu et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: All relevant data are included within the paper and its Supporting Information files.
Funding: This study was financially supported by the Italian Ministry of Education, University and Research (MIUR; Research Project PRIN 2012), ‘Social immunity in honeybee: behavioural, chemical and microbiological aspects’. prot. 2012RCEZWH. The National Institute of Social providence (INPS, Italy) which provide the PhD scholarship granted to Michelina Pusceddu.
Competing interests: The authors have declared that no competing interests exist.
Cooperation among individuals is observed in many phylogenetically diverse taxa and has facilitated the evolution of sociality in the animal kingdom [1,2]. The main advantages of sociality, promoted by natural selection, include more efficient vigilance against predators, a better ability to identify food sources, and the greater survival of developing brood. However, life within a group also presents certain disadvantages, one of the most significant being the ease with which predators can detect prey. For this reason, nest protection to reduce vulnerability is another central aspect in the evolution of sociality [3,4].
Eusocial insects such as the honeybee (Apis mellifera) adopt numerous general and behavioral defense mechanisms against their predators. General mechanisms include nest architecture, site and visibility, as well as species-dependent morphological adaptations such as the size of an individual . In contrast, behavioral defenses are specific to particular enemies and require the prior identification of the predator based on olfactory, visual or tactile cues, recognition of movement, and information from previous encounters [6,7,8]. Behavioral defense can also depend on agonistic behavior by the invader during an encounter, and in eusocial insects, on the caste to which the occupant and/or intruder belong . In the latter case, individuating and blocking specific predators in honeybee societies is the responsibility of guard bees. These bees adopt specialized behaviors that dissuade attacks by invertebrate predators and conspecifics from other colonies, thus preventing the loss of food and brood, and they also recruit “soldiers” to defend the nest against more aggressive predators [10,11]. Defense behavior in the Asiatic honeybee (Apis cerana) was recently shown to vary not only according to the predator, but also based on the context in which the attack takes place, e.g. minimal peril caused by an attack on a single forager contrasting with the substantial threat caused by an attack at the nest entrance .
Wasps are major invertebrate enemies of honeybees, invading hives to steal honey, pollen, larvae and adults to provide sugar and protein for themselves and their offspring [13,14]. The defense mechanisms used by A. mellifera against the Asian predatory hornet (Vespa velutina) has been studied in detail due to the predatory success of the wasp  and the damage caused by its introduction into Europe . However, the study of relationships between sympatric species is also necessary even though the predators are less dangerous  because such relationships may offer insight into the evolution of defense behaviors [18,19]. We therefore investigated the predator–prey relationship between two sympatric species of social Hymenoptera, namely the honeybee (Apis mellifera ligustica) and the European wasp (Vespula germanica) also known as the German wasp or German yellowjacket, in a representative area of the European Mediterranean region (Sardinia, Italy). We evaluated the effectiveness of behavioral displays of attack and defense which have co-evolved in these two species, the defense mechanisms in various peril contexts, and the potential damage and disturbance caused by this predator to the honeybee colony under attack.
Materials and methods
The experimental apiary was set up in the northwest of Sardinia in March 2014, inside the experimental farm of the University of Sassari Department of Agriculture (latitude 40°46'23'', longitude 8°29'34''). The apiary comprised 18 A. mellifera ligustica colonies maintained in new Dadan-Blatt hives containing 10 combs each. They were checked every week to confirm the presence of the queen, as well as pollen and nectar provisions. We also monitored the sanitary status for evidence of microbial infections and varroosis .
Agonistic events between V. germanica and A. mellifera ligustica were examined in two different contexts, one at the hive entrance (which was regularly patrolled by guard bees) and the other on the ground close to the hive, where weakened and dead bees were present. The behavioral observations were based on the “all occurrences sampling” method  in which we recorded the frequencies of a series of behavioral events as set out in the ethogram described below.
Attacks at the nest entrance were recorded in 2014 and 2015 throughout September and October, when the predatory activity of wasps is more intense due to their higher nutritional requirements during reproduction and rearing of offspring . Each colony was recorded for two 15-min sessions per day using a Canon LEGRIA HF R506 video camera placed ~20 cm from the opening of the hives. Recordings were taken during the hottest part of the day (between 9:30 am and 15:30 pm) when the wasps were most active. A total of 279 h of video footage was recorded (63 h in 2014 and 216 h in 2015) and all 18 colonies were observed for the same duration (15.5 h). Subsequently, two operators independently screened the video recordings using a slow motion system (VLC software v2.2.0) and the agonistic behaviors observed were used to establish an ethogram as described below. The ethogram was supplemented with further “attack” and “defense” behaviors not observed by us but reported in the literature for similar species, or in these two species facing different antagonists. This approach allowed us to evaluate the repertoire of agonistic behavior between V. germanica and A. mellifera ligustica in a wider context. The frequency (number of events per unit of time) was reported for all the recorded attack and defense behaviors.
Attacks at ground level (only on individuals still alive and close to the hive) were monitored in 2015 on the same colonies, concurrently with some of the observations at the nest entrance. These observations were conducted by sight, without using the video camera, for a total of 32 h. Two operators simultaneously observed the ground surface under three hives in two sessions per day, each lasting 10 min. The frequency (number of events per unit of time) was reported for all the observed attack and defense behaviors.
Effect of predator attacks on bee foraging activity
The 15-min video clips taken at the nest entrance in each colony were used to evaluate the disturbance caused by wasps on the foraging activity of the honeybees. We compared the frequency of pollen foragers entering the hive 5 min after wasp attack (“attack” context) with the frequency at random times before the attack (“control” context) over a fixed 2-min interval. The comparisons were carried out for 27 agonistic events observed in 2015 to account for any interference that prevented us counting the number of pollen foragers, e.g. continuation of balling, successive attacks, or other bees blocking the view of the video camera.
To determine whether there was a correlation between the degree of agonistic support and the intensity of predator aggression, all attacks at the nest entrance were divided into two behavioral categories described as threats (attacks in which the defender did not make physical contact with the predator) and fights (where physical contact was involved) [23,24]. For each agonistic event, we recorded the duration of the attack, the number of supporters intervening to help a nestmate under attack (in the case of individual support) and any observed cases of balling.
Wasp attack behaviors observed in this study.
Attack—The wasp swoops down to the landing board or to ground level below the entrance plate, grasps the bee from above with its forelegs and starts biting it (usually between the caput and thorax).
Fight—The predator and prey are involved in a physical encounter which may include instances of biting, aggressive gripping, and spinning on a surface or in flight (S1 Photo).
Entering the hive—A wasp may be able to enter the nest if it is overlooked by the bees. Wasps may also enter the nest after contact with bees, i.e. following antennation or following a fight.
Predation—The wasp kills the honeybee. The wasp usually goes on to dismember and consume the honeybee or to carry off parts to its offspring (see below). In some cases, the wasp may also eat the contents of the honey stomach .
Escape—The wasp escapes when the attack has not been successful and one or more honeybees defend themselves or their nest effectively (S1 Movie).
Wasp attack behaviors not observed in this study.
Coalition attack—A coalition of several wasps launches an attack. This behavior has been reported for the Asian giant hornet (Vespa mandarinia), which has developed a strategy of group hunting: certain individuals pillage while others defend the site against conspecifics from other colonies .
Nest defense behaviors observed in this study.
Antennation or antennal boxing—This is often the first physical contact between the occupant and invader, and most likely facilitates the recognition of intruders and conspecifics . It is defined as asymmetric when a dominant and a submissive can clearly be distinguished from the behavioral display of the two opponents, or symmetric when such a distinction is not possible .
Threat—The typical behavior of a honeybee in the presence of a conspecific intruder or other predator, consisting of open mandibles and the adoption of the so-called C posture (gaster flexion with or without extension of the sting) .
Agonistic support—Altruistic behavior in which a honeybee helps another involved in a conflict, thus facing a potential risk. It may involve a single bee or several bees (supporters) that come to the aid of their nestmates.
Balling—The formation of a ball of bees around a wasp until the latter is killed or becomes harmless (S2 Photo). In heat-balling the wasp succumbs to the heat inside the ball because hornets and wasps have a lower thermal tolerance than bees [30,31]. In asphyxia-balling, the heat inside the ball can be lethal to the predator but it dies due to the increased concentration of CO2 in the hemolymph which causes asphyxiation [32,33].
Killing and removal of the predator—Wasps can be killed by a single bee sting or the stings of several bees, or by balling (see above). The dead or dying wasp is then removed from the nest or landing board.
Nest defense behaviors not observed in this study.
Bee carpet—A large proportion of the colony regroups on the landing board and along the sides of the hive, forming a “bee carpet” . This behavior in A. mellifera ligustica has been observed against the European hornet (Vespa crabro) .
Shimmering or shaking signal—When a wasp is seen, bees at the entrance of the hive simultaneously vibrate their abdomens for a few seconds, resulting in the emission of a loud hissing noise caused by the movement of their wings . This behavior has been observed in the giant honeybee (Apis dorsata) [34,35], the dwarf honeybee (Apis florea) , A. cerana , A. cerana nuluensis  and A. mellifera cypria  against V. velutina and the oriental hornet (Vespa orientalis). A. mellifera ligustica also generates a hissing noise in the presence of V. crabro albeit without the shimmering behavior [18,13]. Shimmering is considered a visual signal for the predator and appears to have evolved in order to dissuade the latter from attacking, i.e. it is an honest alert signal that reduces the likelihood of predator success [7,34].
Retreat into the nest—Complete retreat into the nest during an attack has been described in A. cerana  and A. mellifera cypria , as well as the Cape honeybee (A. mellifera capensis), the African honeybee (A. mellifera scutellata) and the Carniolan honey bee (A. mellifera carnica) .
Bees attack—Africanized bees are characterized by a higher propensity than European bees to switch from collective nest defense to attack behavior . The response of Africanized bees is also faster, more aggressive, and involves the recruitment of more nestmates . Attack behavior has also been observed in A. mellifera cypria towards V. orientalis .
The disturbance of foraging activity was measured by comparing the number of pollen foragers in the attack context to the number of pollen foragers in the control context using the Wilcoxon signed rank test (paired comparisons). A chi-squared test was used to measure the proportional difference in support events (individual agonistic support and balling) between the threat and fight categories. To reduce the chance of a type I error, continuity correction was used for the chi-squared tests because the sample size was less than 200 .
The Wilcoxon rank sum test (unpaired comparisons) was used to compare the number of supporters in the threat and fight categories (excluding balling). We also tested for correlation (non-parametric Spearman correlation) between the number of supporters and the duration of attacks. To reduce the chance of a type I error in this analysis, we used Bonferroni correction in the case of multiple testing with significance set at α = 0.05/2 = 0.025. All tests were carried out using R v3.0.2 implemented with library (exactRankTests) and library (coin).
Raw experimental data are available in supporting materials (S1 Data File).
We observed 68 attacks at the hive entrance in 279 h of video footage, specifically 11 attacks in 2014 (63 h) and 57 in 2015 (216 h) representing ~0.24 attacks per hour. The most frequent outcome was wasp escape (55 events, 80.9%) and the least frequent was bee predation (1 event, 1.5%). On three occasions (4.4%) the wasp was observed entering the hive and coming out alive. On another three occasions it was not possible to confirm the fate of the wasp because the observation session terminated while the wasp was still inside the hive. The average attack time was 3.5 ± 0.4 s.
We observed 465 attacks at ground level in 32 h, and these only targeted isolated bees (~14.5 attacks per hour). In this case, the outcome was more balanced. Bee predation was observed 226 times (48.6%) and in 91 of these cases sequestration also occurred. The wasp was chased away 239 times (51.4%), which is a much lower proportion compared to hive entrance attacks. The attack behavioral display data are summarized in Table 1.
We did not observe a collective attack against any of the colonies so our data only represent defense behaviors against individual wasps. Among the 68 agonistic events observed at the hive entrance, 28 cases (41.2%) involved a single bee defending itself or the nest successfully, thus causing the wasp to flee (S2 Movie). In the remaining 40 attacks (58.8%), other bees from the same nest came to the rescue. Agonistic support occurred in 90% of cases, involving an average of 1.9 ± 0.2 supporters per attack (S3 Movie). Balling was observed in 10% of cases. In six cases (8.8%), the wasp was killed and removed from the landing board. Agonistic support of the bees under attack was never observed among the 465 ground level attacks close to the hive. The defense behavioral display data are summarized in Table 2.
Disturbance of foraging
We did not observe any disturbance of foraging activity when the colony was under attack. Indeed, there were no statistically significant differences between the frequency of foraging in the attack context (24.2 ± 4.1) and in the control context (23.0 ± 3.4) in 2015 (U = 154, N1 = N2 = 27, P = 0.6378).
The agonistic events most commonly supported by nestmates either individually or by balling were those involving physical contact (fights) rather than warning behavior (threats). Accordingly, we observed a statistically significant difference between the number of supported threats and the number of supported fights as shown in Fig 1 (chi-squared = 13.07, df = 1, P < 0.001). Moreover, when balling events were excluded, the average number of supporters was significantly higher in fights than threats, as shown in Fig 2 (U = 221, N1 = N2 = 32, P < 0.001). There was also a positive correlation between the number of supporters and the duration of attack (S = 20653, P < 0.001, rho = 0.53). Agonistic support was observed only at the hive entrance, not at ground level.
The difference between the two groups was highly significant (chi-squared test, P < 0.001). N = number of agonistic events observed in 18 colonies.
The 68 agonistic events we observed included four cases of balling which are excluded from the analysis. The difference between the two groups was highly significant (Wilcoxon rank sum test, unpaired comparisons P < 0.001). N = number of agonistic events observed in 18 colonies.
Discussion and conclusions
Our data revealed that V. germanica attacks on nest of A. mellifera ligustica occur infrequently, particularly attacks on the landing board of the hive, so there is a low risk of predation. Only in extremely rare cases did the predator manage to overcome the barrier of guard bees, enter the hive, pillage it and escape. Instead, V. germanica predatory activity is clearly directed at ground level, targeting weak or isolated bees that have fallen from the entrance plate. This is a specialized form of attack, as opposed to a direct attack on the hive, which achieves high predation efficiency. It can be explained by the optimal foraging theory, which postulates a trade-off between energy returns and mortality due to predation [26,42,43,44]. Our observations revealed a compromise between the reward obtained and the risk taken by the wasp, with a direct attack on the hive entrance attracting a greater risk than attacks on isolated bees. Similar observations have been reported in other species: for example, the isolation of foragers is essential to improve the hunting efficiency of hornets in the vicinity of A. dorsata nests because this species is extremely effective in repelling hornets by shimmering .
Our observations also indicated that V. germanica is predominantly a solitary predator, because we found no evidence of coordinated attacks involving other conspecifics. In contrast, competition for food and pillaging among wasps were observed during predation. This probably reflects the individual and independent foraging typology of this species : individuals from different colonies can find themselves at the same foraging site, explaining why each individual defends its own prey . In contrast to predators such as V. velutina, V. crabro , and V. tropica , V. germanica has never been observed attacking forager bees in flight and returning to the hive, only bees on the ground or on the landing board.
The relatively weak predation practiced by V. germanica is confirmed by the ability of 1–3 bees to repel an attack without recourse to truly collective defense strategies such as balling. Indeed, balling by A. mellifera ligustica against V. germanica has never been reported before, and we observed this behavior only four times throughout our observation period. In contrast, balling is deployed much more frequently against Asian wasps [18,38]. This is important because balling often kills some of the participating bees in addition to the predator, suggesting that A. mellifera ligustica regulates its defense behavior depending on the intensity of the threat in order to prevent unnecessary sacrifices [7, 46]. We also saw no evidence of alternative collective defense strategies such as a bee carpet or shimmering, which are often deployed against V. crabro . Again this suggests that A. mellifera ligustica adjusts its defense strategy in response to different predators, which can likewise be interpreted as a trade-off between the involvement of the colony in collective defense (with the associated risks discussed above) and the peril represented by the predator.
Our study also revealed that more intense agonistic events (i.e. fights rather than threats) attract stronger support from nestmates, and showed a correlation between the degree of support and the duration of attack. This suggests that A. mellifera ligustica can adapt its defense behavior according to the context of the peril . The hypothesis is further supported by the different defense responses observed at the hive entrance and on the ground. Hive entrance attacks usually attracted supporters in the events because such attacks are obviously recognized as a potential form of peril for the entire colony. On the other hand, we observed that attacks against those individuals which had fallen from the hive plate were never supported. Possibly, these individuals were no longer recognized as nestmates.
The distinct nestmate-recruiting capacity observed in different behavioral peril categories probably reflects the emission of warning signals such as alarm pheromones by bees that are under attack [6,12]. In contrast to recent observations in colonies of A. mellifera ligustica attacked by V. velutina in Liguria, Italy (Cervo, personal communication), we never observed the interruption of foraging or complete retreat into the nest. We can therefore exclude the possibility that vibration stop signals are used to recruit supporters . Indeed, we found no evidence that V. germanica disrupts A. mellifera ligustica foraging activity, providing more support for the hypothesis that the prey–predator relationship between these two sympatric species has reached a state of balance and that V. germanica need not be considered a threat to apiculture. However, to exclude the threat to bee foraging activity completely, further observations are required in areas with a greater density of wasp colonies.
Docile characteristics are preferred when selecting genetic lines of A. mellifera and several methods have therefore been developed to evaluate the aggression of reared honeybees. It follows that an understanding of agonistic behavioral displays in A. mellifera against natural enemies could be used to develop more effective tests to replace the current evaluation methods, which have been called into question . Indeed, the method recommended by Apimondia (International Federation of Beekeepers’ Associations) is based on subjective evaluation by the operator on a four-point scale, where one point is assigned to the most aggressive bees and four points to the most docile, thus indicating the protective equipment the beekeeper must use . This method does not account for climatic, chemical, visual, social and environmental variables that can play a role in the aggressive behavior of a colony, and is therefore too subjective and difficult to reproduce. A much more selective method has been developed to distinguish between aggressive and docile states in A. mellifera colonies representing the subspecies carnica, scutellata and capensis . Our data could also facilitate the selection of genetic lines of honeybees that are less hostile to humans while maintaining aggressive behavior towards their natural enemies.
In conclusion, our study provides insight into the mechanisms of attack and defense deployed by V. germanica and A. mellifera ligustica both in terms of predator–prey coevolution  and in terms of potential defense strategies that can be used by native bees against alien species such as V. velutina.
S1 Photo. Predator and prey involved in a physical encounter.
S2 Photo. Formation of a ball of bees around a wasp until the latter is killed or becomes harmless (balling).
S1 Movie. Guard bees stop the intruder wasp at the entrance of the hive.
S2 Movie. Individual honey bee defense under wasp attack.
S3 Movie. Several individual bees help a nest mate involved in a conflict.
We thank Rita Cervo and Alice Siguret for their valuable comments on the early versions of this manuscript.
- 1. Szathmary E, Maynard Smith J. The major evolutionary transitions. Nature. 1995;374: 227–232. pmid:7885442
- 2. Wilson EO. Sociobiology: the new syntesis. Cambridge MA, twenty-fifth anniversary edition; 1975.
- 3. Hermann HR. Defensive mechanisms: general considerations. In: Defensive mechanisms in Social Insects. Hermann HR. Ed., Praeger, New York; 1984. pp. 1–31.
- 4. Shorter JR, Rueppell O. A review on self-destructive defense behaviours in social insects. Insect Soc. 2012; 59: 1–10.
- 5. Seeley TD, Seeley RH, Akratanakul P. Colony defense strategies of the honey bees in Thailand. Ecol Monogr. 1982; 52: 43–63.
- 6. Breed MD, Guzman-Novoa E, Hunt GJ. Defensive behavior of honey bees: Organization, genetics, and comparisons with other bees. Annu Rev Entomol. 2004;49: 271–298. pmid:14651465
- 7. Tan K, Wang Z, Li H, Yang S, Hu Z, Kastberger G, Oldroyd BP. An ‘I see you’ prey–predator signal between the Asian honeybee, Apis cerana, and the hornet, Vespa velutina. Anim Behav. 2012a;83: 879–882.
- 8. Wood MJ, Ratnieks FLW. Olfactory cues and Vespula wasp recognition by honey bee guards. Apidologie. 2004;35: 461–468.
- 9. Breed MD, Silverman JM, Bell WJ. Agonistic behavior, social interactions, and behavioral specialization in a primitively eusocial bee. Insectes Soc. Paris. 1978;25(4): 351–364.
- 10. Breed MD, Robinson GE, Page RE. Division of labor during honey bee colony defense. Behav Ecol Sociobiol. 1990;27: 395–401.
- 11. Breed MD, Smith TA, Torres A. Role of guard honeybees (Hymenoptera:Apidae) in nestmate discrimination and replacement of removed guards. Ann Entomol Soc Am. 1992;85: 633–637.
- 12. Tan K, Dong S, Li X, Liu X, Wang C, Li J, Nieh JC. Honey bee inhibitory signaling is tuned to threat severity and can act as a colony alarm signal. PLoS Biol. 2016;14 (3): e1002423. pmid:27014876
- 13. Baracchi D, Cusseau G, Pradella D, Turillazzi S. Defence reactions of Apis mellifera ligustica against attacks from the European hornet Vespa crabro. Ethol. Ecol. Evol. 2010; 22: 1–14.
- 14. Matsuura M, Sakagami SF. A bionomic sketch of the giant hornet, Vespa mandarinia, a serious pest for Japanese apiculture. HUSCAP J. 1973;19: 125–162.
- 15. Tan K, Radloff SE, Li JJ, Hepburn HR, Ang MX, Zhang LJ, et al. Beehawking by the wasp, Vespa velutina, on the honeybees Apis cerana and A. mellifera. Naturwissenschaften. 2007;94: 469–472. pmid:17235596
- 16. Monceau K, Bonnard O, Thiéry D. Vespa velutina: a new invasive predator of honeybees in Europe. J pest sci. 2014;87: 1–16.
- 17. Markwell T.J, Kelly D., Duncan KW. Competition between honey bees (Apis mellifera) and wasp (Vespula spp.) in honeydew beech (Nothofagus solandri var. solandri) forest. N Z J Ecol. 1993;17(2): 85–93.
- 18. Arca M, Papachristoforou A, Mougel F, Rortais A, Monceau K, Bonnard O, et al. Defensive behaviour of Apis mellifera against Vespa velutina in France: Testing whether European honeybees can develop an effective collective defence against a new predator. Behav Process. 2014;106: 122–129.
- 19. De Grandi-Hoffman G, Collins AM, Martin JH, Schmidt JO, Spangler HG. Nest defense behavior in colonies from crosses between Africanized and European honey bees (Apis mellifera L.) (Hymenoptera: Apidae). J. Insect Behav. 1998;11: 37–45.
- 20. Pappas N, Thrasyvoulou A. Searching for an accurate method to evaluate the degree of Varroa infestation in honeybee colonies, in: Cavalloro R. (Ed.) 1988. European research on Varroatosis control, Proceedings of a meeting of the EC-Experts´Group, Bad Homburg, Germany, 15–17 October 1986, Commission of the European Communities, A.A. Balkema, Rotterdam; 1986. pp. 85–91.
- 21. Altmann J. Observational study of behaviour: sampling methods. Behaviour. 1974;49: 227–265. pmid:4597405
- 22. Turillazzi S. Le società delle Vespe. Alberto Perdisa Editore, Bologna; 2003. pp. 55.
- 23. Johnson LK, Hubbell SP. Aggression and competition among stingless bees: field studies’.Ecology. 1974; 55: 120–127.
- 24. Nieh JC, Kruizinga K, Barreto LS, Contrera FAL, Imperatriz-Fonseca VL. Effect of group size on the aggression strategy of an extirpating stingless bee, Trigona spinipes. Insect. Soc. 2005;52: 147–154.
- 25. Coelho JR, Hoagland J. Load-lifting capacities of three species of yellowjackets (Vespula) foraging on honey-bee corpses. Funct Ecol. 1995;9: 171–174.
- 26. Free JB. The behavior of wasp (Vespula germanica L. and V. vulgaris L.) when foraging. Insectes Soc. Paris. 1970;12(1): 11–20.
- 27. Matsuura M. The social biology of wasp. In: Ross KG, Matthews RW, editors. Ithaca: Cornell University Press; 1991. pp. 232–262.
- 28. De Wroey C, Pasteels JM. Agonistic Behavior of Myrmica rubra L. Insectes Soc. Paris. 1978;25(3): 247–265.
- 29. Denis D, Chameron S, Costille L, Pocheville A, Chaline N, Fresneau D. Workers agonistic interactions in queenright and queenless nests of a polydomous ant society. Anim Behav. 2008;75: 791–800.
- 30. Ono M, Igarashi T, Ohno E, Sasaki M. Unusual thermal defense by a honeybee against mass attack by hornets. Nature. 1995;377: 334–336.
- 31. Tan K, Yang MX, Wang ZW, Li H, Zhang ZY, Radloff SE, et al. Cooperative wasp-killing by mixed-species colonies of honeybees, Apis cerana and Apis mellifera. Apidologie. 2012b;43: 195–200.
- 32. Papachristoforou A, Rortais A, Zafeiridou G, Theophilidis G, Garnery L, Thrasyvoulou A, et al. Smothered to death: hornets asphyxiated by honeybees. Curr Biol. 2007;17: 795–796.
- 33. Sugahara M, Sakamoto F. Heat and carbon dioxide generated by honeybees jointly act to kill hornets. Naturwissenschaften. 2009;96: 1133–1136. pmid:19551367
- 34. Kastberger G, Schmelzer E, Kranner I. Social waves in giant honeybees repel hornets. PLoS ONE. 2008; 3(9): e3141. pmid:18781205
- 35. Kastberger G,Weihmann F, Hoetzl T, Weiss ES, Maurer M, Kranner I. How to join a wave: decision-making processes in shimmering behavior of giant honeybees (Apis dorsata). PLoS ONE. 2012; 7(5): e36736. pmid:22662123
- 36. Abrol DP. Defensive behaviour of Apis cerana F. against predatory wasps. J Api Sci. 2006;50: 39–46.
- 37. Papachristoforou A, Rortais A, Sueur J, Arnold G. Attack or retreat: contrasted defensive tactics used by Cyprian honeybee colonies under attack from hornets. Behav Process. 2011;86: 236–241.
- 38. Tan K., Hepburn HR, Radloff SE, Yusheng Y, Yiqiu L, Danyin Z. Heat-balling wasps by honeybees. Naturwissenschaften. 2005;92: 492–495. pmid:16151794
- 39. Kastberger G, Thenius R, Stabentheiner A, Hepburn R. Aggressive and docile colony defence patterns in Apis mellifera. A retreater-releaser concept. J Insect Behav. 2009;22: 65–85.
- 40. Collins A, Rinderer T, Harbo J, Bolten A. Colony defence by Africanized and European honey bees. Science. 1982;218: 72–74. pmid:17776713
- 41. Sokal RR, Rohlf FJ. Biometry, 2nd ed. W. H. freeman and Co., New York; 1981.
- 42. Kasper ML, Reeson AF, Mackay DA, Austin AD. Environmental factors influencing daily foraging activity of Vespula germanica (Hymenoptera, Vespidae) in Mediterranean Australia. Insect. Soc. 2008;55: 288–295.
- 43. Whitford WG and Bryant M. Behavior of a predator and its prey: the horned lizard (Phrynosoma cornutum) and harvester ants (Pogonomyrmex spp.). Ecology. 1979;60: 686–694.
- 44. D’Adamo P, Lozada M. The importance of location and visual cues during foraging in the German wasp (Vespula germanica F.) (Hymenoptera: Vespidae). N Z J Zool. 2003;30: 171–174.
- 45. Monceau K, Arca M, Leprêtre L, Mougel F, Bonnard O, Silvain F, et al. Native prey and invasive predator patterns of foraging activity: the case of the yellow-legged hornet predation at European honeybee hives. PLoS ONE. 2013;8(6): e66492. pmid:23823754
- 46. Helfman GS. 1989. Threat-sensitive predator avoidance in damselfish-trumpet fish interactions. Behav Ecol Sociobiol. 1989;24: 47–58.
- 47. Zakour M. K, Bienefeld K. Subjective evaluation of defensive behavior in the Syrian honeybee (Apis mellifera syriaca). J Apic Sci. 2013; 57(2): 137–145.
- 48. Ruttner, H. Technische Empfehlungen zur Methodik der Leistungsprüfung von Bienenvöikern. In: Proceedings of Paarungskontrolle und Selektion bei der Honigbiene. Internationales Symposium. Lunz am See, Austria; 1972. pp. 112.
- 49. Futuyama D.J. (1986) The Evolution of Interactions Among Species, in: Evolutionary Biology, 2nd ed., Sinauer Associates Inc., Sunderland, Massachusetts; 1986. pp. 482–504.