1. Swarming is biologically normal, not a fault

This chapter first places swarming in biological context: as the natural form of reproduction of the honey bee colony, not as a direct indicator of poor genetics or a deficient queen.

Swarming is the natural form of reproduction of a honey bee colony: part of the workers leaves the original location with the old queen and founds a new colony, while the remaining colony rears a young queen and continues to exist at the original site (Rangel & Seeley, 2012). Studies on colony fissioning show that a large proportion of the workers depart with the swarm, and that the size of this swarm fraction is particularly important for the growth and survival of the swarm colony leaving with the old queen; the observed fraction also agrees well with a model of optimal colony fissioning (Rangel & Seeley, 2012; Rangel et al., 2013). More recent modelling work also classifies colony fission as a reproductive strategy of social insects, whose advantage depends on ecological conditions, swarm size, and establishment probability (Hovestadt et al., 2024).

From the perspective of bee biology, swarming is therefore not "misbehaviour" or an indicator of a deficient queen, but a central reproductive event. From a beekeeping perspective, the same behaviour is judged differently: a natural swarm can lead to bee loss, lower honey yield, extra work, and loss of control over the colony's origin and health status, which is why swarm prevention is an important beekeeping goal. This practical assessment must not, however, be confused with the biological one: what is undesirable for the beekeeper is not automatically "wrong" for the bee colony.

This distinction shapes the entire further discussion: if a colony swarms, this initially only means that it has entered a reproductive phase – not that the queen is genetically unsuitable or that the colony is fundamentally "too prone to swarming". As shown in the following sections, this can only be judged based on the conditions under which a swarm occurred.

2. Genetics plays a role — but not as a simple explanation

This chapter explains what heritability means for swarming behaviour – and why it must not be used to derive a simple genetic test for a single colony.

The fact that swarming is a natural behaviour does not mean that all colonies enter swarming mood equally quickly or strongly. As with many behavioural traits, the threshold at which a colony responds to particular triggers can vary between lines and breeding populations. The question is therefore not whether a colony can swarm in principle, but how easily it enters swarming mood under certain conditions.

Breeding studies show that swarming behaviour can have a genetic component – though with very different estimated values. For Italian and Iranian bee populations, a heritability of 0.34 was calculated for swarming behaviour in each case (Andonov et al., 2019; Tahmasbi et al., 2015). In a large Austrian breeding population with nearly 15,000 colonies, by contrast, the heritability of the selection criterion for swarming behaviour was considerably lower, at 0.08; the estimate was also associated with relatively large uncertainty (Brascamp et al., 2016, corrected 2018).

In other words: depending on the population and model, additive genetic differences in these studies explained roughly 8 to 34% of the observed differences in swarming behaviour. This range is not a fixed biological value and does not mean that a single swarming event would be 8 to 34% genetically caused. Rather, it shows how strongly the estimate depends on the population, environment, survey method, and statistical model.

The Austrian study illustrates particularly well why heritability values must be interpreted with caution. The individual genetic components had not disappeared for swarming behaviour: a heritability of 0.15 was estimated for the worker effect, and as high as 0.33 for the queen effect. At the same time, the two effects were strongly negatively correlated genetically. In the combined selection criterion, this led to the resulting heritability being considerably lower. The same dataset thus shows that the genetic assessment of swarming behaviour depends strongly on whether worker and queen effects are considered separately or jointly (Brascamp et al., 2016, corrected 2018).

Heritability is often misunderstood in practice. It does not describe how "genetic" a particular behaviour is in a given colony, but rather what proportion of the observed differences within a studied population is associated with genetic differences. A heritability estimate therefore cannot be used to conclude that a specific swarm in one's own apiary was primarily genetically caused.

Added to this is the level at which the measurement is made: swarming behaviour is not a trait of a single bee, but a behaviour of the whole colony – dependent on the queen, workers, age structure, brood nest, colony strength, available space, and environment. In this system, genetics is best understood as an influence on the response threshold: some colonies enter swarming mood earlier or more strongly than others under comparable conditions. This view is further supported by genotype-environment experiments (Uzunov et al., 2014).

For practice, caution is therefore warranted: swarming tendency can be influenced through breeding, but a single swarm is not a genetic test. A colony that swarms once is not automatically genetically prone to swarming; a colony that does not swarm in a given year is not automatically genetically reluctant to swarm. Observation over several years, locations, or related colonies is more informative. Genetics, then, does not determine the swarm on its own, but influences the threshold for swarming mood – whether this threshold is reached depends substantially on forage, colony development, available space, queen age, and beekeeping practice.

3. Location, forage, and colony management can overlay genetics

This chapter shows why swarming events must always be assessed in connection with location, forage, weather, available space, and the timing of intervention.

How strongly environment and management practice can overlay the genetic threshold is shown by the pan-European COLOSS experiment: Uzunov et al. (2014) studied colonies of different genetic origin at several European locations and found significant effects of both genotype and location on swarming, defensive, and hygienic behaviour. The differences between locations, however, were greater than those between genotypes. For practice, this means: a queen's origin is important, but does not on its own explain how a colony behaves under particular conditions.

Older field observations point in the same direction. Simpson (1957) found in English honey-production apiaries that, under local conditions, in an average year 10 to 40 percent of colonies would swarm if given plenty of room, but would otherwise remain largely unaffected. This variation between years and locations was classified primarily as environmentally caused, and supports the view that swarming should not be understood solely as a property of a queen or line.

An important practical trigger is available space. If a colony grows too quickly during a good nectar flow, the brood nest can become restricted by nectar or pollen while bee density rises – a configuration that can tip development toward swarming mood. Simpson and Moxley (1971) showed in an experiment with small hives that colonies that filled their space and outgrew it built queen cells and swarmed considerably more often than colonies that did not develop to the same extent. This supports the practical experience that timely expansion is a central element of swarm prevention.

Available space is not, however, the only factor: even in continuously expanded, uncongested colonies, queen age remains an independent influencing factor. Hauser and Lensky (1994) found under subtropical conditions that colonies with older queens built considerably more swarm cells than colonies with young queens, even though the experimental colonies were continuously expanded.

For practice, a useful rule of thumb follows from this: if several colonies at the same location enter swarming mood simultaneously, location, forage, weather, space, and management factors should be examined first. If, by contrast, a single colony repeatedly swarms early and strongly under comparable conditions, a genetic component becomes more likely.

4. Swarming as a threshold model: genetics influences the response, practice the trigger

This chapter summarises the findings in a threshold model: genetics shifts the response threshold, while environment and management practice help determine whether it is reached.

The findings discussed so far are best summarised by a threshold model: swarming mood does not arise because a colony has "swarming genes", nor solely because a beekeeper expanded the hive too late. Genetic origin influences the threshold at which a colony responds to particular stimuli with swarm preparations; environment, forage, colony development, and beekeeping management determine whether this threshold is actually reached.

In this model, swarming tendency is not a yes-or-no trait, but a graded property. A colony with a low threshold may build queen cells earlier than a colony with a higher threshold; conversely, even a genetically rather swarm-reluctant colony can enter swarming mood if forage, colony strength, and available space strongly push in that direction. The practical question is therefore not "is this colony genetically prone to swarming?", but "how does this colony respond under the given conditions?"

A biological explanation for this is provided by the model of Fefferman and Starks (2006). They describe swarming as the result of several interacting factors, including colony size, brood nest congestion, worker age structure, and the queen's maximum egg-laying rate. These factors can contribute to a threshold being reached in colony development at which swarm preparations become more likely. The model does not replace a field trial, but it helps explain why strong colonies, under intense forage and unfavourable space management, can enter a critical phase more quickly.

Genetics does not act in isolation in this model: heritability estimates show that the threshold can be partly influenced by heredity, while genotype-environment experiments show that location and environmental conditions strongly co-determine the observed expression (Andonov et al., 2019; Brascamp et al., 2016, corrected 2018; Tahmasbi et al., 2015; Uzunov et al., 2014). Queen age also shifts the threshold independently of available space (Hauser & Lensky, 1994). A colony's genetic disposition thus only becomes visible in interaction with its environment.

For practice, this model is helpful because it avoids hasty attributions of blame: a swarm is neither automatically proof of poor genetics nor automatically proof of poor colony management. It merely shows that, under the given conditions, the threshold was exceeded – the task is to reconstruct those conditions.

5. Requeening: genetic improvement or age effect?

This chapter separates two frequently conflated questions: requeening can be effective without the original swarm necessarily having to be explained genetically.

After a swarming event, it is often recommended in practice to replace the queen. This recommendation can be sensible, but it is often justified too simply: "requeening can reduce swarming tendency" and "the colony swarmed because of poor swarming genetics" are two different statements.

An important, non-genetic reason for the effect of requeening is the age of the queen. Simpson (1957) showed that colonies with current-year queens built queen cells considerably less often after the end of June than colonies with previous-year queens – an observation he further investigated in a follow-up study on one- and two-year-old queens (Simpson, 1960). Hauser and Lensky (1994) found an even clearer effect under subtropical conditions in Israel: although the experimental colonies were continuously expanded to avoid congestion, and brood area or worker population showed no significant influence on swarm cell formation, colonies with older queens built 2.5 and 3.9 times more swarm cells, respectively, in the two experimental years than colonies with young queens. This corresponds to roughly 150% and 290% more swarm cells. This shows that, even with good space management, queen age can substantially shift the threshold for swarming mood.

This age effect cannot be directly compared with heritability values, because the two measures capture something different. Heritability values describe the proportion of observed differences within a population that is genetically explicable under a given model. The finding by Hauser and Lensky, by contrast, describes a group difference in an experiment. It cannot be concluded from this that queen age generally explains a fixed percentage of swarming behaviour. The finding does, however, show that the age of the queen in a specific colony can be a very strong practical influencing factor – and should therefore be assessed separately from genetic swarming tendency.

Other studies point in the same direction, but are not equally well transferable to Central European practice. Hora et al. (2019) found in a local population of Apis mellifera bandasii that colonies with one-year-old queens on average built hardly any queen cells or swarm cells, while colonies with two-year-old and especially three-year-old queens built considerably more cells. The difference between two- and three-year-old queens corresponded to roughly a fourfold increase. Such values should not be understood as a general rule, but they support the practical statement that queen age can be an independent and sometimes very strong influencing factor.

The underlying mechanism is probably not only egg-laying performance. Hauser and Lensky (1994) discuss age-related changes in the queen's mandibular and tarsal gland secretions as a possible explanation for older queens inhibiting the formation of swarm cells less strongly. Requeening can therefore act via vitality, laying performance, and pheromone effects, without the new queen necessarily having to come from a genetically less swarm-prone line.

Added to this is a practical effect: requeening is rarely an isolated intervention. Anyone replacing a queen usually also checks swarm cells at the same time, alters colony dynamics, or combines the intervention with making increase or correcting available space. If the colony remains calm afterwards, this does not automatically prove that the old queen was genetically unsuitable.

For practice, this means: requeening after a swarm can be the right decision, but should not happen reflexively. First, it should be checked whether management or environmental factors favoured the swarm, and queen age should be explicitly considered. Replacing an old or underperforming queen must be assessed differently, on technical grounds, from prematurely classifying a young queen as genetically unsuitable after a single event.

6. When is requeening technically justified?

This chapter translates the scientific classification into a practical diagnosis: when is requeening a well-founded measure, and when is it merely a reflex?

Requeening is an important tool in beekeeping practice for replacing old, underperforming, or undesirably disposed queens. It becomes problematic when every swarm is automatically taken as proof of a genetically unsuitable queen – a single event is not sufficient for that.

Before deciding on requeening, it is therefore worth running a brief diagnosis using four questions:

1. Has the colony repeatedly and early entered swarming mood – even with sufficient space, timely expansion, and good swarm control?
2. Do related colonies or daughters of the same line show similar behaviour?
3. Does the swarming tendency occur together with other undesirable traits, such as restlessness, aggressiveness, or poor performance?
4. Was the queen already older, declining in laying performance, or did the colony show signs of waning queen influence?

The more of these questions answered "yes", the more justified requeening is on technical grounds – particularly when several daughters or sister lines show the same tendency, since the comparison can then no longer be explained solely by location or management. Genetic correlations between swarming behaviour and other traits such as honey yield or defensive behaviour have indeed been described, but they vary depending on population and model and should therefore be interpreted with caution (Andonov et al., 2019).

The situation is different for a single swarming event following delayed expansion, a blocked brood nest, or an exceptionally strong nectar flow: here, requeening as a genetic measure is not very convincing, since location and forage effects can be larger than genetic differences (Simpson, 1957; Uzunov et al., 2014). In such cases, management practice should be analysed first, before regarding the queen as the cause.

The central rule is: requeening is a targeted measure when a diagnosis is in place – without a diagnosis, it is merely a reflex.

7. What breeding can achieve — and what it cannot

This chapter shows why low swarming tendency is a sensible breeding goal, but does not allow complete control over swarming events.

Low swarming tendency is a legitimate breeding goal: colonies with low swarming tendency make management easier, reduce workload and swarm losses, and improve the predictability of the honey harvest. The fact that this tendency can be partly influenced by heredity is shown by the heritability estimates from various breeding populations (Andonov et al., 2019; Brascamp et al., 2016, corrected 2018; Tahmasbi et al., 2015). Breeding can therefore indeed shift the threshold for swarming mood.

This possibility must not, however, be confused with complete control. The reported heritability values vary not only between populations, but also depending on whether queen and worker effects are considered separately or as a joint selection criterion. Genotype-environment experiments also show that location effects can be larger than genetic differences (Uzunov et al., 2014). For selection, it follows that a queen or line should not be assessed based on an isolated result at a single location, but on how stably a trait appears across several years, locations, or daughter groups.

Low swarming tendency should also not be assessed in isolation from other traits. Andonov et al. (2019) report genetic correlations between swarming behaviour and traits such as honey yield and defensive behaviour; such correlations are population-dependent and should be interpreted with caution. They do, however, show that breeding decisions must take several traits into account simultaneously. A colony that swarms little but is restless, aggressive, or underperforming is not automatically valuable for breeding.

It is also important to distinguish between perspectives: swarming is, for the bee colony, a natural reproductive behaviour. When breeding programmes select for low swarming tendency, this happens from a beekeeping perspective – for the sake of manageability, yield security, and labour economy. This is legitimate, but should not be presented as though swarming itself were a biological defect. Breeding against swarming tendency is artificial selection in favour of particular management goals, not a "correction" of bee biology.

Finally, the same applies here too: a successful requeening is not automatically a genetic breeding success, since queen age and pheromone effects can explain the same outcome (Hauser & Lensky, 1994). Breeding is therefore an important tool, but not a substitute for good colony management: a swarm-reluctant line can swarm under unfavourable management, while a more swarm-prone line can remain manageable under good management.

8. Practical conclusion: from reflex to diagnosis

This chapter summarises the consequences for training, advisory work, and the apiary: the best swarm prevention comes from diagnosis, not reflexes.

A swarm should not automatically be interpreted as a genetic defect of the queen. Swarming is a natural reproductive behaviour whose occurrence depends on several levels: genetics influences the threshold for swarming mood; environment, forage, and available space help determine whether this threshold is reached; and queen age can shift it independently of these. The blanket rule "colony swarmed = replace the queen" is therefore too crude – it may be correct in an individual case, but it does not replace a causal analysis.

In training and advisory work, the focus should therefore lie less on attributing blame and more on diagnosis. After a swarm, the conditions should be reconstructed: How strong was the colony compared with the rest of the apiary? How did the forage situation develop? Was expansion carried out in time? Was the brood nest blocked? How old was the queen? Did several colonies at the same location show swarming mood simultaneously? Such questions help distinguish between genetic disposition, queen age, environmental influence, and management errors.

As a rule of thumb: the more colonies at the same location enter swarming mood simultaneously, the more strongly forage, weather, available space, and the timing of intervention should be examined first. Conversely, the more strongly a single colony or a related line repeatedly swarms early under good management, the more plausible a genetic component becomes. What is always decisive is the comparison, not the single event.

Swarm prevention therefore does not begin only with breaking down swarm cells, but with timely expansion, sufficient space, observation of colony development, and a realistic assessment of queen age. Breeding for swarm-reluctant colonies can ease this work, but cannot replace it. Anyone who merely replaces the queen after a swarm may be treating only the symptom. Anyone who considers genetics, queen age, forage, space, and management together makes better decisions – for breeding, colony management, and training alike.

 

---

Learn more:

• The mechanisms of natural swarming
• Does selection in beekeeping allow for heritability?
• Queen rearing and honey bee genetics
• Swarm prevention
• Why does a bee colony replace its queen?

 

References

Indonov, S., Costa, C., Uzunov, A., Bergomi, P., Lourenco, D., & Misztal, I. (2019). Modeling honey yield, defensive and swarming behaviors of Italian honey bees (Apis mellifera ligustica) using linear-threshold approaches. BMC Genetics, 20.

Brascamp, E. W., Willam, A., Boigenzahn, C., Bijma, P., & Veerkamp, R. F. (2016). Heritabilities and genetic correlations for honey yield, gentleness, calmness and swarming behaviour in Austrian honey bees. Apidologie, 47, 739–748. https://doi.org/10.1007/s13592-016-0427-9

Brascamp, E. W., Willam, A., Boigenzahn, C., Bijma, P., & Veerkamp, R. F. (2018). Correction to: Heritabilities and genetic correlations for honey yield, gentleness, calmness and swarming behaviour in Austrian honey bees. Apidologie, 49, 462–463. https://doi.org/10.1007/s13592-018-0573-3

Fefferman, N. H., & Starks, P. T. (2006). A modeling approach to swarming in honey bees (Apis mellifera). Insectes Sociaux, 53, 37–45.

Hauser, H., & Lensky, Y. (1994). The effect of the age of the honey bee (Apis mellifera L.) queen on worker population, swarming and honey yields in a subtropical climate. Apidologie, 25(6), 566–578.

Hora, Z., Yadeta, G., & Hundie, T. A. (2019). Young honeybee (Apis mellifera bandasii) queens perform better than old ones. Journal of Biology, Agriculture and Healthcare.

Hovestadt, T., Kohl, P., & Mitesser, O. (2024). Optimal fissioning strategies of social insects with respect to colony dynamics and nest founding probability. Insectes Sociaux, 71, 171–183.

Rangel, J., & Seeley, T. D. (2012). Colony fissioning in honey bees: Size and significance of the swarm fraction. Insectes Sociaux, 59, 453–462.

Rangel, J., Reeve, H. K., & Seeley, T. D. (2013). Optimal colony fissioning in social insects: Testing an inclusive fitness model with honey bees. Insectes Sociaux, 60, 445–452.

Simpson, J. (1957). The incidence of swarming among colonies of honey-bees in England. The Journal of Agricultural Science, 49, 387–393.

Simpson, J. (1960). The age of queen honeybees and the tendency of their colonies to swarm. The Journal of Agricultural Science, 54, 195.

Simpson, J., & Moxley, E. (1971). The swarming behaviour of honeybee colonies kept in small hives and allowed to outgrow them. Journal of Apicultural Research, 10, 109–113.

Tahmasbi, G., Kamali, M., Ebadi, R., Javaremi, A., Babaei, M., Gharadaghi, A., & Bahraini, R. (2015). Genetic trends and parameters of honey production, swarming and defense behavior in Iranian honeybee (Apis mellifera meda) colonies. Journal of Agricultural Science and Technology, 17, 1735–1742.

Uzunov, A., Costa, C., Panasiuk, B., Meixner, M., Kryger, P., Hatjina, F., Bouga, M., Andonov, S., Bieńkowska, M., Le Conte, Y., Wilde, J., Gerula, D., Kiprijanovska, H., Filipi, J., Petrov, P., Ruottinen, L., Pechhacker, H., Berg, S., Dyrba, W., Ivanova, E., & Büchler, R. (2014). Swarming, defensive and hygienic behaviour in honey bee colonies of different genetic origin in a pan-European experiment. Journal of Apicultural Research, 53, 248–260.