Attack or retreat: Contrasted defensive tactics used by Cyprian honeybee colonies under attack from hornets

Alexandros Papachristoforou^a, , , Agnès Rortais^a, 1, , Jérôme Sueur^b,  and Gérard Arnold^a, c, 

^a Laboratoire Evolution, Génomes, Spéciation, CNRS UPR 9034, Avenue de la Terrasse, 91198 Gif-sur-Yvette, France

^b Muséum National d’Histoire Naturelle, Département Systématique et Évolution, CNRS UMR 7205 OSEB, 45 Rue Buffon, 75005 Paris, France

^c Université Paris-Sud 11, Orsay, France

Received 7 November 2010;  

revised 16 December 2010;  

accepted 20 December 2010.  

Available online 25 December 2010. 

Abstract

This study describes the tactics used by Cyprian honeybees (Apis mellifera cypria) to defend their colonies against hornet (Vespa orientalis orientalis) attacks. We use simulated hornet attacks and a combination of video recordings and image analysis to reveal, for the first time, contrasted intra-subspecies defensive tactics that operate at the colony level during predation. In some colonies, when attacked, the numbers of guards at the hive entrance increases rapidly to attack, engulf, and kill invading hornets. In other colonies, guards avoid conflicts with hornets by retreating gradually and by forming a defensive line of honeybees at the hive entrance. Retreater colonies have propolis walls at the hive entrances with small apertures that are too narrow to allow the hornet to access the hive and that therefore reinforces entrance protection. On the contrary, attacker colonies have propolis walls with large openings through which the hornet can pass; these bees block the hornet's access by intensively guarding the hive entrance. We experimentally destroy propolis walls to test whether colonies consistently rebuild walls with the same intrinsic characteristics and we also monitor the survival rate of each anti-predator tactic after massive natural predation by hornets.

Graphical abstract

 

Research highlights

 Cyprian honeybees present two defensive tactics when attacked by hornets.  Some colonies attack the hornet. Other colonies retreat inside nest cavity.  Defence is correlated with propolis wall architecture at the entrances of hives.  “Attackers” have wide openings on propolis walls that allow hornets access to hives.  “Retreaters” have narrow openings that block hornets from entering hives.

Keywords: Defensive behaviour; Anti-predator tactics; Propolis walls; Apis mellifera cypria; Vespa orientalis

Article Outline

1. 

Introduction

2. 

Materials and methods

2.1. Experimental apiaries

2.2. Behaviour of honeybees under attack

2.3. Propolis walls

2.3.1. Architectural differences in propolis walls

2.3.2. Building the propolis walls

2.3.3. Exchange of propolis walls

2.4. Monitoring colony survival after attacks by hornets

2.5. Statistical analysis

3. 

Results

3.1. Behaviour of honeybees under attack

3.2. Propolis walls

3.2.1. Architectural differences in propolis walls

3.2.2. Building the propolis walls

3.2.3. Exchange of propolis walls

3.3. Monitoring colony survival after hornet attacks

4. 

Discussion

Uncited reference

Acknowledgements

Appendix A. 

Supplementary data

References

1. Introduction

Defence is undoubtedly one of the most important factors in species life history. To prevent capture by predators, prey species – including arthropods (Wilson, 1975) – have evolved many defensive behavioural tactics or morphological structures with behavioural impact on potential predators. These defences include crypsis, mimicry, aposematism, escape tactics, or even retaliation (Alcock, 2001). Most bees of the genus Apis can efficiently defend their nests when confronted by predators such as hornets or wasps ([Butler, 1954], [Ono et al., 1987], [Ono et al., 1995], [Koeniger et al., 1996], [Pirk et al., 2002], [Tan et al., 2007], [Stabentheiner et al., 2007] and [Baracchi et al., 2010]). Ultimately, hornets can seriously damage colonies either by preying on honeybees or by robbing their resources (i.e. honey, pollen, and brood). These resources are then used to sustain their own brood and adults within the colony; a colony that in turn may go on to destroy the entire honeybee colony ([Morse, 1978] and [Abrol, 1994]).

In Cyprus, during the dry hot season, honeybees may be severely impacted by food shortage and native predators such as the Oriental hornet (Vespa orientalis). Cyprian honeybees have developed effective defensive behavioursto reduce the impact of hornets; these defences have been described in previous studies. Cyprian honeybees demonstrate coordinated and massive defence tactics when attacked by hornets such as “shimmering” and “hissing” ([Butler, 1954], [Morse, 1978] and [Papachristoforou et al., 2008]). Furthermore, they carry out an efficient asphyxia-balling behaviour to kill the hornet (Papachristoforou et al., 2007).

After detailed observations, we found that Cyprian honeybees demonstrate two contrasted defensive behaviours, or tactics, among colonies: the “attacking” and the “retreating” behaviour. To reveal these tactics, we tested two assumptions: (1) during a predator attack, the number of honeybees on the flight board decreases in the retreater colonies but increases in the attacker colonies, and (2) at the end of a predator attack, the number of honeybees on the flight board is greater in attacker colonies than in retreater ones. We also looked at differences in the architecture of propolis walls at beehive entrances. We examined the role of propolis walls in defence by testing the hypothesis that propolis walls should provide more protection in retreater colonies than those in the attacker colonies. Overall, this study describes and analyzes the two anti-predator tactics developed by Cypriot colonies when experimentally exposed to hornets.

2. Materials and methods

2.1. Experimental apiaries

All experiments were conducted on A. m. cypria colonies, at two apiaries located in Alassa (N34°46′, E32°57′) between October 2004 and October 2007. A total of 50 colonies were used in the experiments. The behavioural experiments were conducted at an apiary that was maintained by a beekeeper who had never bred queens for his colonies or introduced foreign queens. The second apiary, which had not been maintained by a beekeeper for two years, was observed for colony losses during 2007. Environmental conditions were harsh in 2007; it had been one of the driest years of the last decades. These conditions led to reduced colony strength and caused extensive apiary losses throughout the whole island.

None of the colonies swarmed during any individual experiment and no colony replaced its queen. The strength of all colonies, in terms of the number of frames covered by honeybees, was noted during each experiment.

2.2. Behaviour of honeybees under attack

Ten colonies were observed during simulated hornet attacks in October 2004. This experiment was repeated with ten different colonies in September 2005. To simulate attacks, we collected live hornets in the apiary and exposed them to honeybee colonies, as described by Papachristoforou et al. (2007) with slight modifications. To simulate attacks, hornets were collected in the apiaries and were anaesthetized by CO₂ for 5 s and then tied at the petiole with a 10 cm nylon wire to a stick. When hornets fully recovered, they were brought to the beehive flight board and exposed to colonies guards. Hornets were free to move and flight within the restricted range of 10 cm.

To test and compare the number of honeybees dedicated to defence, we used a digital camera and monitored the behaviour of honeybees in all colonies for a total of three minutes (including 15 s before the simulated attack); this work took place in October 2004 and September 2005. Video recordings were transformed to images with Pinnacle Liquid Edition V6 software. With ImageJ software, pictures were analysed every 15 s and counted the number of honeybees on flight boards during attacks. Colonies were defined as “attacker” or “retreater” depending on whether honeybees, moved forward upon the approach of hornets, tended to get involved in conflict and engulfed the predator or moved backward and retreated inside the beehive, to avoid conflict with the intruder.

2.3. Propolis walls

2.3.1. Architectural differences in propolis walls

During the behavioural experiments in September 2005, we observed that the architecture of propolis entrance walls differed between colonies expressing different anti-predator tactics. To examine the role of propolis walls in defence, we took pictures of five entrances from beehives of colonies expressing the attacker behaviour and five expressing the retreater behaviour. We evaluated the number and cross-sectional area of all openings in each propolis wall. At the same time, we captured 25 hornets attacking honeybee colonies and measured the area of the largest transversal section of their thorax, corresponding to the largest transversal section of their entire body. The openings of the propolis walls were also measured and compared in the ten colonies (i.e. five attacker and five retreater colonies) to estimate the percentage of hornets that would be able to get through the openings.

2.3.2. Building the propolis walls

By December 2005, hornet attacks had ceased for the season. We first took pictures of the propolis walls of four colony entrances (i.e. two retreater and two attacker colonies), evaluated the cross-sectional area of their openings and then removed propolis walls from all entrances. We used digital photography to monitor the bees’ repair strategy. The cross-sectional area of all wall openings was measured at monthly intervals until November 2006, to examine the architecture of the new walls.

2.3.3. Exchange of propolis walls

In October 2007, we conducted preliminary trials to observe the defensive behaviour of four additional colonies placed in beehives with different architecture of propolis entrance walls; we exchanged the colonies (population and frames) from two colonies expressing the attacker behaviour and having wide openings in their propolis walls with beehives of two colonies expressing the retreater behaviour and having narrow openings in their propolis walls. The exchange took place during the peak of hornet attacks.

2.4. Monitoring colony survival after attacks by hornets

In 2007, we observed the defensive behaviour of 22 new colonies under intensive natural attacks by hornets and honeybee survival rates following this predation period; the goal was to obtain an initial indication of the effectiveness of each tactic. The apiary had not been maintained by beekeepers for two years. Observations took place in October, at the end of the hornet attack period, after a season of intense hornet predation. Surviving and collapsed colonies were counted and compared according to their anti-predator tactic at the end of November.

2.5. Statistical analysis

For the statistical analysis, non-parametric tests were used in all cases. In particular the Kruskal–Wallis ANOVA and Mann–Whitney for two independent samples tests were used. All statistics were calculated using R Statistical software (R Development Core Team, 2010).

3. Results

3.1. Behaviour of honeybees under attack

In terms of qualitative behaviour, attacker and retreater anti-predator tactics were expressed through the presence or absence of conflict with the predator and through the tendency to form an asphyxia-ball around the hornet (Movies 1 and 2 in Supplementary data). Attacker colonies engulfed the hornets after short conflicts of 30–120 s. In two of the attacker colonies, the hornet was pulled inside the hive and the asphyxia-ball formed just behind the hive entrance. Retreater colonies initially approached the hornet, but honeybees withdrew inside the beehive after the first 20 s. During all experiments, no ball formations were ever noticed around any of the hornets attacking the retreater colonies.

Among the ten colonies examined in 2004, seven expressed the attacker behaviour and three the retreater behaviour. In 2005, five colonies expressed the retreater behaviour and five the attacker behaviour. The total number of honeybees participating in defence on the flight board was higher in 2004 than in 2005 (n₂₀₀₄ = 10, n₂₀₀₅ = 10, Z = 98, p < 0.001, two-tailed Mann–Whitney test) and there was a significant interaction between years and tactics (n_2004,a = 7, n_2004,r = 3, n_2005,a = 5, n_2005,r = 5, K = 16.87, d.f. = 3, p < 0.001, Kruskal–Wallis test). The number of defensive honeybees present on the flight board was higher in the attacker colonies than in the retreater colonies (Fig. 1a) at the end of the attacks in 2004 (n_r = 3, n_a = 7, Z = 0, p < 0.05, two-tailed Mann–Whitney test) and 2005 (n_r = 5, n_a = 5, Z = 0, p < 0.01, two-tailed Mann–Whitney test). In 2004, the average number of honeybees present on the flight board before and at the end of the attack was similar for the attacker colonies (n = 7, Z = 14, p = 1, two-tailed Mann–Whitney test for paired samples) and for the retreater colonies (n = 3, Z = 6, p = 0.25, two-tailed Mann–Whitney test for paired samples) (Fig. 1b).

Full-size image (48K)

Fig. 1. 

(a) Colonies according to anti-predator tactic. Total number of honeybees on the flight board involved in defence during hornet attacks, plotted for both the attacker (white) and retreater (grey) tactics in 2004 and 2005. (b) Honeybees on the flight board defending the colonies before and during hornet attacks during 2004.

In 2005, the number of honeybees present on the flight board clearly increased over the course of the hornet attack even if the differences before and at the end of the attack were not significant for either the attacker (n = 5, Z = 0,p = 0.063, two-tailed Mann–Whitney test for paired samples) nor the retreater (n = 5, Z = 0, p < 0.10, two-tailed Mann–Whitney test for paired samples) colonies. In 2004, the number of honeybees present on the flight board before the onset of a simulated hornet attack was slightly higher in the retreater than in the attacker colonies (Fig. 1b) though this difference was not significant (n_r = 3, n_a = 7, Z = 4, p = 0.18, two-tailed Mann–Whitney test). In 2005, there was also no difference in the initial numbers of honeybees before the simulated hornet attack, because there were never honeybees outside the beehive entrance in any colony. Like in 2004, honeybees exited the hives at the onset of the simulated hornet attack, but the honeybees of the retreater colonies withdrew inside the nest cavity after the first minute, avoiding conflict with the predator.

3.2. Propolis walls

3.2.1. Architectural differences in propolis walls

The number of openings in the propolis walls of the attacking colonies (10.60 ± 3.36 (mean ± s.d.), n = 5) and of the retreating colonies (12.40 ± 2.51, n = 5) was similar (Z = 17, p = 0.86, one-tailed Mann–Whitney test). However, the size of the openings (Table 1) was greater in the attacker colonies (18.36 ± 6.90 cm², n = 5) than in the retreater colonies (6.99 ± 2.62 cm², n = 5) (Z = 1, p < 0.001, one-tailed Mann–Whitney test) (Fig. 2). Knowing that the transversal area of hornets’ thorax section was 0.55 ± 0.34 cm² (n = 25), the theoretical percentage of hornets that would be able to get through the propolis-secured walls was higher in attacker colonies (58.25 ± 26.82%, n = 5) than in retreater colonies (10.31 ± 11.55%, n = 5). In addition, only one to three honeybees were needed to defend the few, relatively smaller, openings in the retreater colonies (Fig. 3) compared with the numerous honeybees required for the guarding of the wide openings in the attacker colonies.

Table 1. Numbers and area (cm²) of openings on propolis walls (September 2005).

Tactic	Attack					Retreat
Colony	A1	A2	A3	A4	A5	R6	R7	R8	R9	R10
Range	0.31–2.85	0.43–3.55	0.23–2.61	1.85–12.81	0.14–4.52	0.22–7.33	0.11–1.37	0.06–1.02	0.31–0.77	0.11–0.75
Number of openings	11	11	12	4	5	9	13	12	12	16
Area (total)	16.53	20.66	10.43	28.84	15.34	11.19	7.73	4.43	5.74	5.89
Mean	18.33a					6.99b
s.d.	6.90					2.62

Full-size image (63K)

Fig. 2. 

Different architecture of propolis walls. (A) Colony expressing the attacker tactic. Wide openings in propolis walls through which hornets can easily enter the hive. (B) Colony expressing the retreating tactic. Narrow openings in propolis walls through which hornets can rarely enter the hive.

Full-size image (81K)

Fig. 3. 

Honeybee guarding the narrow opening on propolis wall at the entrance of the beehive.

3.2.2. Building the propolis walls

The process of rebuilding the propolis walls proceeded in a manner similar to that seen at the initial construction stage (Figs. 4 and 5 in Supplementary data). The attacker colonies left the hive entrance almost free of propolis. Depending on their size, hornets could potentially get inside from any of the wide propolis openings (100% and 84.6% mean penetrability after and before propolis removal, Table 2 in Supplementary data). On the contrary, the retreater colonies covered the entrances leaving very narrow spaces, making it almost impossible for hornets to enter the hive (37.6% mean penetrability before propolis removal and 12.5% one year later after rebuilding, Table 2 in Supplementary data).

3.2.3. Exchange of propolis walls

The behaviour of the four colonies that were removed from their beehives and placed in beehives having different propolis wall architecture was unaffected. The honeybees of the attacker colonies exited the hive, approached, and tried to engulf intruders. The propolis walls that had narrow openings did not cause any tendency to defend behind this new barrier. They managed to overcome attacks and survived the following wintering. The honeybees of the two retreater colonies moved backwards to the cavity of the beehive and avoided conflict with intruders. However, hornets approached and were able to enter the beehive of retreater colonies without resistance. No balling behaviour was exhibited by the retreater colonies which were invaded by hornets within two weeks of the exchange and which eventually collapsed.

3.3. Monitoring colony survival after hornet attacks

At the apiary that was not maintained by a beekeeper for 2 years, 14 out of the 22 colonies were destroyed during hornet attacks and those beehives were eventually occupied by hornets. Nine of the destroyed colonies were retreater colonies with small propolis openings; five were attacker colonies with large propolis openings. Of the eight surviving colonies, six were attacker colonies with large propolis openings, and two were retreater colonies. These latter two colonies had the smallest openings of all colonies. However, statistical comparisons detected no differences in survival rates between attacker and retreater colonies (Pearson's chi-squared test with Yates’ continuity correction, χ² = 1.78, d.f. = 1, p = 0.18).

4. Discussion

The study of the defensive behaviour of A. m. cypria under attack by hornets revealed that colonies of the same subspecies exhibit two contrasted anti-predator tactics. Though a few examples of the presence of multiple anti-predator tactics in a single species are known in solitary animals ([Bauwens and Thoen, 1981] and [Ohno and Miyatake, 2007]), contrasted collective defensive behaviour among social animals under attack by intruders has not been reported in insects. The attacker anti-predator tactics, which include asphyxia-balling in many cases, is common among honeybees under attack by hornets ([Ono et al., 1987], [Koeniger et al., 1996], [Ken et al., 2005], [Tan et al., 2007], [Papachristoforou et al., 2007] and [Baracchi et al., 2010]). The retreater anti-predator tactics in honeybees under attack by hornets has only been reported in Thailand, among guards of Apis cerana; these guards retreat into the narrow cavity of their nest and engulf the hornet Vespa tropica only if it succeeds in entering the nest (Seeley et al., 1982). However, behavioural variation among colonies of the same species and subspecies has recently been described in detail. This variation was observed during stimulation experiments using mechanical disturbance, alarm pheromones, or the combination of both (Kastberger et al., 2009). During mechanical disturbance, the responses of test colonies ranged from the release of flying defenders to retreat into the nest. After mechanical disturbance combined with alarm pheromones, honeybees of one colony retreated in the nest cavity without releasing flying defenders, while other colonies’ defensive behaviour varied from “aggressive” to “docile”. Our findings support the results of Kastberger et al. (2009), in that the honeybees of the experimental colonies behaved in a similar way (aggressive/attacker or docile/retreater) when attacked by live hornets.

The defensive behaviour of the Cyprian honeybee reveals some similarities and basic differences compared to other Apis mellifera subspecies facing different Vespa species. For example, in a recent study, Baracchi et al. (2010)described in details the behavioural interactions between Apis mellifera ligustica prayed by Vespa crabro. Like in Cyprus, the Italian honeybees demonstrate massive defensive behaviour expressed through coordinated “body shaking” from the “bee-carpet” formed in front of colonies’ entrances. In addition, they perform balling behaviour in many cases. In Cyprus, balling behaviour was present in attacker colonies but it was absent in retreater. The “bee-carpet” was present before the approach of predators in colonies, independently from the anti-predator tactic each colony expressed, as can be seen by the initial numbers of guards at the flight boards, but the behaviourchanged upon the hornet's approach. Attackers recruit more bees and “attack” the hornets while retreaters gradually deform the “bee carpet and retreat behind the propolis walls”. Probably, the differentiations on defensive behaviour of A. m. cypria compared with A. m. ligustica are a result of co-evolution and adaptation to the different predation tactics expressed by V. orientalis compared to V. crabro. V. crabro never attacks and catches individual honeybees from the “bee-carpet” formed in front of the entrance (Baracchi et al., 2010). This is the rule in V. orientalis; they attack directly the honeybees forming the “bee carpet”, trying to catch an individual and escape, avoiding further conflict with the defenders. Furthermore, V. orientalis rarely attacks the few foragers departing or returning from foraging activities (mainly at the colonies expressing the “attacker” behaviour) while V. crabro preys on honeybees engaged in foraging activities.

There was no evidence for relationship between a colony's anti-predator tactics and its strength; all colonies were of equal strength in 2005 (number of occupied bee ways) and the difference in colony strength was not significant in 2004. The initial number of honeybees on the flight boards before a hornet approach did not affect the behavioural tactic of the colonies. Differences in the number of honeybees participating in defence between years might be linked to predation pressure from bee-eater birds (Merops apiaster) that prey on both hornets and honeybees and therefore impacted activity levels of both. Such differences could also be linked to the harder environmental conditions during September resulting in lower colonies’ strength compared to October.

It is not clear yet if either of the two contrasted anti-predator tactics is more effective. Our observations at the apiary that faced intensive hornet predation showed no differences in survival rates between attacker and retreater colonies.

From the exchange of propolis walls, we obtained some preliminary indications that an alteration to the structure of the propolis walls might affect the defensive capability of retreating colonies. On the one hand, the honeybees of the two retreater colonies that were shifted to hives with attacker-style openings in the propolis walls were not able to confront the hornets by attacking them outside the beehive. Therefore, hornets were able to enter the beehive, leading to collapse of the colonies. No balling-behaviour was exhibited from the retreater colonies against hornets. On the other hand, attacker colonies faced no defence difficulties as a result of being switched to a hive with retreater-style openings in the propolis walls.

Colonies with the same defensive strategies rebuilt the same structure of propolis walls that was present before removal. Though our sample size was small (two attacker and two retreater colonies), results after the exchange provide a clear indication of what might happen if the defence pattern is disturbed artificially, especially for retreater colonies. These data, though preliminary, provide a first indication that genetic factors might control the expression of these tactics and provide an avenue for future experimentation. Genetics have indeed been proven to influence the defensive behaviour of honeybees and Quantitative Trait Loci (QTL) are implicated in defence at both the individual and the colony levels ([Breed et al., 2004] and [Hunt et al., 2007]).

Honeybee defence is believed to have evolved in response to predation pressure and environmental factors; this evolutionary response has led to various behavioural displays (DeGrandi-Hoffman et al., 1998). However, these same mechanisms known to drive the evolution of defence cannot explain the differences found in Cyprian honeybee colonies. These colonies have coevolved with V. orientalis under the same climatic factors within an “isolated” environment (i.e. the Cyprus Island) and it would be expected that they should display a common and successful defensive behaviour. The peculiar phenomenon of two contrasted defensive tactics within the same population under the same predation and environmental pressure could be a focus of future research.

It is also not yet known how these alternate defence strategies are maintained in this subspecies, or under what circumstances each strategy confers a fitness advantage to the colony. Since the environmental conditions in Cyprus are often extremely harsh during summer and autumn, with limited food for colonies, there are predicted trade-offs between growth and defence (Rivera-Marchand et al., 2008). The retreater tactic would limit the risk of honeybee loss due to conflict. For these colonies, it seems that resources are also better protected because hornets can rarely enter the colony. However, their foraging activity is totally stops during the hornet attack period, contrary to the attacker colonies where foraging activity is also reduced but never ceases completely (data not shown). Retreater colonies are, therefore, liable to colony weakness and starvation. We assume that if the duration of hornet attacks is short, the retreater tactic would be optimal, but when it is extended, the attacker tactic may be preferable because colonies would still be able to collect food. Moreover, the timing of attacks, in combination with weather conditions, might be crucial. If colonies had successfully stored food, then retreating could be advantageous. If sufficient supplies had not been stored, then the attacker behaviour may be the best tactic. The observation in our preliminary data that more retreater colonies collapsed during the harsh conditions of 2007 supports these hypotheses, but further work is required. The coexistence of these two defensive tactics raises crucial questions regarding the importance of the environment in the evolution of honeybee social behaviour.

Uncited reference

Nolan (1928).

Acknowledgments

We thank Max Watkins for reviewing the manuscript and Vita-Europe Ltd. for supporting our research. We are grateful to Sharilynn Wardrop who revised the English and provided useful comments and corrections of the manuscript. We are indebted to George Kellenos and Giannakis Varnava for providing the honeybee colonies. We dedicate this research to the memory of Jérôme Trouiller who continuously supported our research on Cypriot honeybees.

References

Abrol, 1994 D.P. Abrol, Ecology, behaviour and management of social wasp, Vespa velutina Smith (Hym., Vespidae), attacking honeybee colonies, Korean J. Apic. 9 (1994), pp. 5–10.

Alcock, 2001 J. Alcock, Animal Behavior: An Evolutionary Approach (7th ed.), Sinauer Associates Inc., Sunderland, MA (2001).

Baracchi et al., 2010 D. Baracchi, G. Cusseau, D. Pradella and S. Turillazzi, Defence reactions of Apis mellifera ligustica against attacks from the European hornet Vespa crabro, Ethol. Ecol. Evol. 22 (2010), pp. 1–14.

Bauwens and Thoen, 1981 D. Bauwens and C. Thoen, Escape tactics and vulnerability to predation associated with reproduction in the lizard Lacerta vivipara, J. Anim. Ecol. 50 (1981), pp. 733–743. Full Text via CrossRef

Breed et al., 2004 M.D. Breed, E. Guzmán-Novoa and G.J. Hunt, Defensive behaviour of honeybees: organization, genetics, and comparisons with other bees, Annu. Rev. Entomol. 49 (2004), pp. 271–298. Full Text via CrossRef |View Record in Scopus | Cited By in Scopus (59)

Butler, 1954 C.G. Butler, The World of the Honeybee, Collins, London (1954) 226 pp.

DeGrandi-Hoffman et al., 1998 G. DeGrandi-Hoffman, A.M. Collins, J.H. Martin, J.O. Schmidt and H.G. Spangler, Nest defence behavior in colonies from crosses between Africanized and European honey bees (Apis mellifera L.) (Hymenoptera: Apidae), J. Insect Behav. 11 (1998), pp. 37–45. Full Text via CrossRef

Hunt et al., 2007 J.G. Hunt, V.G. Amdam, D. Schlipalius, C. Emore, N. Sardesai, E.C. Williams, O. Rueppell, E. Guzmán-Novoa, M. Arechavaleta-Velasco, S. Chandra, K.M. Fondrk, M. Beye and E.R. Page, Behavioral genomics of honeybee foraging and nest defence, Naturwissenschaften 94 (2007), pp. 247–267.

Kastberger et al., 2009 G. Kastberger, R. Thenius, A. Stabentheiner and R. Hepburn, Aggressive and docile colony defence patterns in Apis mellifera. A retreater–releaser concept, J. Insect Behav. 22 (2009), pp. 65–85. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (2)

Ken et al., 2005 T. Ken, H.R. Hepburn, S.E. Randloff, Y. Yusheng, L. Yiqiu, Z. Danyin and P. Neumann, Heat-balling wasps by honeybees, Naturwissenschaften 92 (2005), pp. 492–495. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (17)

Koeniger et al., 1996 N. Koeniger, G. Koeniger, M. Gries, S. Tingek and A. Kelitu, Observations on colony defence of Apis nuluensis Tingek, and predatory behavior of the hornet, Vespa multimaculata Perez, 1910, Apidologie 27(1996), pp. 341–352. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (7)

Morse, 1978 R.A. Morse, Honey Bee Pests, Predators, and Diseases, Cornell University Press, Ithaca, NY, USA (1978) 430 pp..

Nolan, 1928 W.J. Nolan, Seasonal brood-rearing activity of the Cyprian honeybee, J. Econ. Entomol. 21 (1928), pp. 392–403.

Ohno and Miyatake, 2007 T. Ohno and T. Miyatake, Drop or fly? Negative genetic correlation between death-feigning intensity and flying ability as alternative anti-predator strategies, Proc. R. Soc. B 274 (2007), pp. 555–560. Full Textvia CrossRef | View Record in Scopus | Cited By in Scopus (11)

Ono et al., 1987 M. Ono, I. Okada and M. Sasaki, Heat production by balling in the Japanese honeybee, Apis cerana japonica as a defensive behavior against the hornet, Vespa simillima xanthoptera (Hymenoptera: Vespidae),Experientia 43 (1987), pp. 1031–1032.

Ono et al., 1995 M. Ono, T. Igarashi, E. Ohno and M. Sasaki, Unusual thermal defence by a honeybee against mass attack by hornets, Nature 377 (1995), pp. 334–336. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (49)

Papachristoforou et al., 2007 A. Papachristoforou, A. Rortais, G. Zafeiridou, G. Theophilidis, L. Garnery, A. Thrasyvoulou and G. Arnold, Smothered to death: hornets asphyxiated by honeybees, Curr. Biol. 17 (2007), pp. R795–796.

Papachristoforou et al., 2008 A. Papachristoforou, J. Sueur, A. Rortais, S. Angelopoulos, A. Thrasyvoulou and G. Arnold, High frequency sounds produced by Cyprian honeybees Apis mellifera cypria when confronting their predator, the Oriental hornet Vespa orientalis, Apidologie 39 (2008), pp. 468–474. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (2)

Pirk et al., 2002 C.W.W. Pirk, R. Hepburn, S.E. Radloff and J. Erlandsson, Defence posture in the dwarf honeybee, Apis florae, Apidologie 33 (2002), pp. 289–294. View Record in Scopus | Cited By in Scopus (3)

R Development Core Team, 2010 R Development Core Team, R: A Language and Environment for Statistical Computing, R Foundation for Statistical Computing, Vienna, Austria (2010) ISBN: 3-900051-07-0, http://www.R-project.org.

Rivera-Marchand et al., 2008 B. Rivera-Marchand, T. Giray and E. Guzmán-Novoa, The cost of defence in social insects: insights from the honey bee, Entomol. Exp. Appl. 129 (2008), pp. 1–10. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (1)

Seeley et al., 1982 T.D. Seeley, R.H. Seeley and P. Akratanakul, Colony defence strategies of the honey bees in Thailand, Ecol. Monogr. 52 (1982), pp. 43–63. Full Text via CrossRef

Stabentheiner et al., 2007 A. Stabentheiner, H. Kovac and S. Schmaranzer, Thermal behaviour of honeybees during aggressive interactions, Ethology 113 (2007), pp. 995–1006. Full Text via CrossRef | View Record in Scopus |Cited By in Scopus (4)

Tan et al., 2007 K. Tan, S.E. Radloff, J.J. Li, H.R. Hepburn, M.X. Yang, L.J. Zhang and P. Neumann, Bee-hawking by the wasp, Vespa velutina, on the honeybees Apis cerana and A. mellifera, Naturwissenschaften 94 (2007), pp. 469–472. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (6)

Wilson, 1975 E.O. Wilson, Sociobiology, Harvard University Press, Cambridge, MA, USA (1975).

Appendix A. Supplementary data

Download this Video (4636 K)

Download this Video (4047 K)

Download this File (31 K)

Download this File (200 K)

Corresponding author. Tel.: +33 1 69 82 37 17; fax: +33 1 69 82 37 36.
¹ Present address: Emerging Risks Unit, European Food Safety Authority, Largo N Palli 5/A, 43100 Parma, Italy.

Note to users: The section "Articles in Press" contains peer reviewed accepted articles to be published in this journal. When the final article is assigned to an issue of the journal, the "Article in Press" version will be removed from this section and will appear in the associated published journal issue. The date it was first made available online will be carried over. Please be aware that although "Articles in Press" do not have all bibliographic details available yet, they can already be cited using the year of online publication and the DOI as follows: Author(s), Article Title, Journal (Year), DOI. Please consult the journal's reference style for the exact appearance of these elements, abbreviation of journal names and the use of punctuation.

There are three types of "Articles in Press":

Accepted manuscripts: these are articles that have been peer reviewed and accepted for publication by the Editorial Board. The articles have not yet been copy edited and/or formatted in the journal house style. Uncorrected proofs: these are copy edited and formatted articles that are not yet finalized and that will be corrected by the authors. Therefore the text could change before final publication. Corrected proofs: these are articles containing the authors' corrections and may, or may not yet have specific issue and page numbers assigned.