Robbing: understanding what happens between honey bee colonies
Brian R. Johnson & James C. Nieh in Journal of Insect Behavior (2010) 23:459–471. DOI: 10.1007/s10905-010-9229-5. “Modeling the Adaptive Role of Negative Signaling in Honey Bee Intraspecific Competition”, Brian R. Johnson & James C. Nieh provides an illuminating model of robbing, rather than a complete explanation of every aspect.
To understand it, we need to take a step back. Foragers that find a rich food source tell their nestmates through the waggle dance (Seeley 1989, 1995; Passino and Seeley 2006). This signal rapidly increases the number of new bees visiting the advertised food source. Positive feedback and the resulting recruitment decline when the source is exhausted (von Frisch 1967; Seeley 1995).
Bees also have a STOP signal that causes waggle dancers to stop dancing (Kirchner 1993; Nieh 1993; Pastor and Seeley 2005; Lau and Nieh 2010; Nieh 2010). Its role was clarified by Nieh (2010). The stop signal provides negative feedback in certain contexts, including fights over resources. Bees that lose fights, provided they survive, produce many stop signals directed at other foragers working at the same site (Nieh 2010). Robbing is a clear example of this behavioural context.
At colony level, bees must respond to the defensive strength of colonies they might attack. Individual bees have only their own experience of the foreign nest: whether they can steal honey without fighting, or have to fight. For example, if 10 bees obtain honey, with or without fighting, while 50 are driven off by guards and some robbers are killed, this may indicate that the target colony is strong enough to defend itself and the attack should cease. Robbing can cost the attacking colony dead bees, making it unprofitable. The STOP signal may therefore provide a selective benefit by shutting down waggle dances when costs are high and benefits low.
Nectar work is divided between foragers, which collect it, and receiver bees inside the hive, which accept and process it. In the model, receivers do not retain food-source information, whereas foragers do. Receivers move randomly around the dance area until they encounter a forager needing to unload. They accept nectar and are unavailable for further unloading until they have dealt with that load. Essentially, a forager returning from a food source may produce four kinds of signal: waggle dance, tremble dance, shaking signal or stop signal, depending on its experiences. If an inactive forager encounters an available waggle dancer, it begins to follow. The model assumes one follower at a time and an average following time of 5 seconds, based on Judd (1995). If the inactive bee has previously visited the advertised site, it is reactivated and returns there (Seeley 1995). If the site is unfamiliar, it tries to locate it. The tremble dance instead recruits additional nectar-receiving bees (Seeley 1992; Seeley et al. 1996; Thenius et al. 2008).
The shaking signal activates bees (Schneider et al. 1986; Nieh 1998; Biesmeijer 2003). The stop signal causes waggle dancers to stop dancing (Lau and Nieh 2010; Nieh 2010). In the model’s parameterisation, nectar receivers take 20 minutes to complete a trip to the storage combs and back (Seeley 1989). A receiver that cannot find a forager to unload within 120 minutes leaves that task, reducing the total receiver population by one. These are model settings, not fixed timings for every colony.
Robbing
In an apiary experiencing forage scarcity, some foragers will soon discover neighbouring colonies and begin assessing
whether to attack them. The study used simulations to examine attack dynamics and the effect of stop signals in two settings: attacks on colonies capable of defending themselves but requiring different response times, and attacks on colonies with different defensive abilities.
When a target colony eventually recruits enough guards to stop at least 95% of attackers, but does so at different speeds, two situations arise in the simulations. With rapid guard recruitment, most robbers encounter guards, lose fights and neither persist nor recruit others. As guard recruitment slows, the stop signal increasingly influences the number of attacks and their abandonment. A few bees can slip unnoticed into the nest when guards are too few relative to the entrance size and leave with honey.
Back at their own hive, these successful bees perform waggle dances to recruit a few more nestmates to rob. Meanwhile, other bees that were repelled by guards during the same attack may also return. Their stop signals counter the recruitment dances of the few successful intruders, and the attacking colony abandons the raid. The scenario changes when the target colony’s defensive ability varies or is absent.
The proportion of robbers that guards stop at the entrance is crucial. In the simulations, if the target blocks only a small proportion of intruders, returning robbers recruit increasing numbers and the attack grows. If it blocks an intermediate proportion, the attack continues but attacker numbers increase little or stabilise. When it repels most attackers—more than half in the scenario described—the attacking colony gives up; this is not a universal biological threshold. In every case, stop signals from defeated bees reduce the number of attackers, and thus the number the target must repel.
Essentially, attacks are more likely to be called off when the attacking colony uses stop signals than when it does not. In the model, this makes colonies more conservative: they mainly attack weak neighbours and quickly abandon attacks on colonies that can defend themselves. For the defenders, strongly repelling all intruders, or at least a substantial proportion, encourages returning robbers to inhibit further recruitment. The numerical outcome depends on the model conditions.
Robbing is neither an easy nor a cost-free option for bees: its costs can exceed its rewards, and misleading initial success can draw a colony into a poor choice. Later research also found an association between infection and risky robbing: robbers were more heavily infected with Nosema and shorter-lived. Calling them “kamikaze” is a metaphor; this does not prove that viral infection generally causes suicidal robbing. First, a bee that gets inside another hive returns laden with honey and may perform vigorous waggle dances to recruit others. Second, when those recruits reach the target, it is likely to have mobilised its defences, so fighting becomes necessary and bees may die (Butler and Free 1952; Winston 1987).
A forager may work intensively for several weeks (Schmid-Hempel and Wolf 1988; Visscher and Dukas 1997). Losing one in a fight can therefore cost much more than the benefit of a small stolen load. Colonies should consequently be selective about attacking another hive and quick to stop when costs approach benefits.
For defence, Downs and Ratnieks (2000) and Couvillon et al. (2008) showed that colonies vary in the number of guards stationed at the entrance, which can increase after an attack. Couvillon et al. (2008) found that the response can take several hours even in colonies strong enough to repel attackers. This highlights the importance of doing everything possible as a beekeeper to prevent attacks, because the defensive capacity immediately displayed by each colony influences intruders. It makes little sense, in evolutionary cost terms, to lose many bees before discovering that a raid was unprofitable; yet direct encounters provide that information. The paper discusses roughly 50 guards as sufficient for the small entrance considered, not as a universal requirement for every hive.
Therefore, a useful measure is to narrow entrances promptly in late summer as forage ends, concentrating guards over a smaller opening while maintaining adequate ventilation. This should help prevent unfamiliar bees entering and repel intruders that return home with negative experience. If some intrusions occur, effective defence may reduce robbing by encouraging the attackers to stop. Bees can slip unnoticed in and out of foreign hives, increasing in number each day until an obvious outbreak. How often do we hear, “There was nothing yesterday, and today it erupted”? More accurately, it may already have been happening yesterday without being visible, perhaps for some time.
Guard determination and selectivity are not constant, and queen pheromone signalling can influence colony behaviour. For small or weak colonies, prompt entrance restriction is one preventive measure. Synthetic queen pheromone products such as Bee Boost® have also been studied for colony responses, but the evidence does not establish guaranteed prevention of robbing or improved growth in every situation. Any use should follow the product instructions and cannot replace adequate colony strength, a healthy queen and appropriate management.
Comments