Evidence vs Ideology in Honey Bee Breeding

I went to a talk recently, at my local beekeeping association, to hear about ‘sustainable queen production’. The speaker is a genuine chap, who believes in what he is doing, but I found myself disagreeing with nearly everything he said. Well, maybe I disagreed with half, and most of the rest was describing a type of beekeeping which is very different from my own. He is a prominent member of the Bee Improvement and Bee Breeders Association (BIBBA) and also the Bee Farmers’ Association (BFA). That seems like a rarity, but I could be wrong.

I have heard many of the speaker’s beliefs aired by others of a similar persuasion, and, of course, people are free to keep bees however they like, but I decided to express my concerns and comments here. Each to their own, but I don’t think you can go around the country speaking to rooms full of mainly hobby beekeepers without having some accountability for your message. So, where to begin?

Why keep bees?

Beekeeping has changed over the years, along with most other things. Many moons ago, bees were kept in skeps, and the wax was more valuable than the honey. The bees were usually killed so that the beekeeper could get at the wax and honey. Wax was enormously important to churches for candles. From the mid-1800s, demand for wax plummeted. In times of rationing, during and after the second world war, beekeeping popularity surged as sugar, was difficult to come by. I suppose, in recent times, the latest surge in beekeeper numbers came about because of the “save the bees” message that was pushed by the media and environmentally conscious folks. As it happened, wild pollinators needed help but honey bees were doing fine, thank you very much. As an aside, I find it odd that many who began beekeeping to “save the bees” now practise treatment-free beekeeping, which tends to lead to their suffering and death. Ho-hum.

It is worth understanding why people keep bees, as it leads to a more in-depth understanding of the way they keep their bees. During the early 20th Century, because of advances in transport, equipment, and manufacturing, a new breed of beekeeper emerged – the commercial bee farmer. There came about a fork in the road; hobby beekeepers branched one way, and bee farmers the other. There is a world of difference between the two. Farmers are business people, trying to make a profit from their animals and land. Both branches can and do co-exist, and neither side has any moral superiority over the other, in my view.

Dimension	Commercial Outlook	Hobby/”natural” Outlook
Objective	Profit, consistency, risk control	Enjoyment, environmental ethos
Feeding	Prevent starvation, maximise spring build up	“Bees should manage themselves”
Treatment	Cost-benefit: treat if loss risk > treatment cost	Avoid chemicals, Bond (Live & Let Die)
Stock Renewal	Requeening cycles, line selection	Local swarm reliance, minimal intervention
Measurement	Scales, yield records, cost accounting	Anecdotal/ observational
View of Bees	Livestock, managed for outcome	Semi-wild organism to be respected

Neither philosophy is “wrong”; they optimise for different reward functions:

Commercial: economic efficiency and biosecurity

Hobby: personal satisfaction and ecological symbolism

“What the hobbyist sees as interference, the professional sees as management.”

— Eva Crane, 1999

Collecting swarms & cut-outs

The speaker at my recent beekeeping meeting encouraged people to collect swarms to build up their bee numbers. He felt that bees obtained from cut-outs and swarms were more likely to be “survivor stock” – strong enough to thrive without being “molly coddled” by people. At another point in the talk, he encouraged people to raise their own queens, which I applaud. However, with swarms, who knows where they came from? The swarm could be headed by one of those terrible imported queens, or an ancient queen, or they could be riddled with mites or other disease. Ideally, in my opinion, swarms should be kept well away from other bees, assessed after a month, and probably split into nucs with new queens from known lineage. On the plus side, they are great at drawing out comb, and their mites can be almost completely killed using oxalic acid, as they have no brood.

From the reading I have done, and my knowledge of the relatively high beekeeper density across much of the UK, I think that many bees that people call ‘wild bees’ are often from nearby colonies, previously kept in hives. Genuinely wild bees seem to spend a lot of time swarming and dying, as far as I can tell from studies by Tom Seeley and others. They are not the sort of bees that I want to keep, nor, I suspect, do the majority of hobby beekeepers. Small, swarmy colonies that produce little honey – is that really what beekeepers are looking for? Some probably are, but not many, surely.

If you are keen on running a breeding programme, with goals of improving your bees, then swarms and cut-outs really do need to be kept far away and possibly re-queened with your own stock. If any of these colonies turn out to be great, after a period of quarantine and assessment, then they are an excellent source of genetic diversity to bring into the fold.

Black bees are best?

This is something that I often hear from BIBBA members. There is a belief that the colour of a bee has some correlation to its behaviour and performance. I have previously written about how incubating queens at lower temperatures produces darker queens. This does not carry across to workers, though, as far as I know.

In my area, and many areas in the UK, if bees are allowed to re-queen themselves, by the third or fourth generation they tend to be rather dark in appearance. They also tend to display less desirable traits – less calm, more defensive, clumpy and drippy on the comb, more likely to follow and sting etc. The speaker explained that, whenever he does a cut-out of bees that have survived for years without human help, they tend to be dark bees. This apparently shows that dark bees are better adapted to that location.

Sorry to be a party pooper, but the colour of bees has no known correlation to behaviour or performance. The genetics of bee colour are basic. In an experiment using CRISPR, changing a single gene led to dark bees becoming light¹. The genes responsible for many of the complex behaviours that we see in our bees are many, and they are dispersed over multiple chromosomes (polygenic). I have never seen evidence linking genes responsible for colour to honey bee behaviour.

Before modern DNA analysis, the appearance of the bee – colour, and certain measurements of wings, for example – were all we really had. We don’t need to rely on these outdated approximations now that we can look at the actual DNA.

• Colour is influenced by a small set of loci affecting melanin and banding pattern.
• Behavioural traits (brood regulation, defensiveness, hygienic response) are polygenic and involve hundreds of loci across the genome.
• There’s no known² linkage between pigmentation genes and key behavioural genes.
  

So colour can drift independently of behaviour during hybridisation. Don’t be too hasty to kill that ‘ginger’ queen!

In the German/European breeding literature (BLUP datasets and modern genomic studies), colour isn’t even analysed as a performance trait. Published correlation tables cover honey yield, gentleness/defence, calmness, swarming, hygiene/VSH, etc. – not colour – and most cross-trait genetic correlations are small. The best evidence-based position is that colour is not a useful predictor of behaviour or productivity. If it were, it would appear in those correlation matrices.

Does Amm even exist?

Back in the day, when the black bee (Apis mellifera mellifera) naturally evolved to be the primary subspecies in Northern Europe, all black bees in the UK were pretty similar, as they were all the same subspecies. Things have changed enormously since then. Across Britain, most managed and feral colonies are admixed (“mongrel”) – a mixture of the native M-lineage (Apis mellifera mellifera, Amm) and imported C-lineage (notably carnica/ligustica) genetics. The historical/native background is Amm, but the current proportion of Amm ancestry varies by place and beekeeper practice.

A FERA-linked PhD, Catherine Eleanor Thompson, sampling managed colonies across England & Wales, reported an average Amm proportion of 42.9% using microsatellite DNA³. Island/very remote and some ‘native-friendly’ programmes were higher (64–66%), and a few individual colonies scored much higher still. Her work also found UK feral honey bees were not more ‘native’ than local managed bees – both were highly introgressed.

More recent DNA work⁴ has shown that there are pockets of nearly pure Amm in the UK and Ireland, but that the majority of our bees are all mixed up (regardless of colour). Given that pure Amm in the UK is a very rare thing, and that our bees are largely a mongrel mix of Amm, Ligustica, and Carnica (with other oddities thrown in), the important thing for beekeepers is no longer the genotype of their bees. It is important to try to conserve the pure subspecies, but for practical beekeeping, that ship has sailed. What we need to do is ignore the colour and start to focus purely on selectively breeding for the traits that we desire.

Trait selection wins

• Traits such as varroa tolerance, hygienic behaviour, low swarming, temperament, winter hardiness are what determine a colony’s success — not its wing angles or mitochondrial haplotype.
• Heterosis (hybrid vigour) is well-documented in honey bees: crosses between divergent lines often outperform parents for brood area, honey yield, and sometimes disease resistance.
• Several studies have found local adaptation effects stronger than subspecies identity. In mixed environments, selecting the best performers locally improves stock faster than trying to restore a ‘pure’ subspecies that no longer exists in the area.
  

Does selection work in hybrid populations?

Yes, if done systematically. Trait heritability and repeatable improvement have been demonstrated in highly admixed bees.

Trait	Typical Heritability (h²)	Notes
Hygienic behaviour	0.25 – 0.65	Fastest to improve, repeatable in mixed lines
Varroa sensitive hygiene (VSH)	0.3 – 0.4	Heritable in mixed European stocks
Temperament	0.2 – 0.4	Improves with selection; little correlation with genotype lineage
Honey yield	0.3 – 0.5	Clear response to selective breeding even in hybrid lines

Even when bees are hybrids, heritable traits respond strongly to selection. Controlled mating accelerates progress; open mating still works if you saturate the area with your best drones.

Question	Implication
Should we chase pure Amm?	Only if you are in an isolation zone (e.g. Colonsay). For most UK beekeepers functional improvement beats purity
Should we use DNA tests?	Useful for recording background or maintaining a conservation line – not needed for production or resistance breeding
Is hybridisation a problem?	Not if you manage it. It can provide useful genetic variation for selection
Can trait selection change population genetics?	Yes – over time, selection shifts allele frequencies toward desired traits, even in mixed populations

F2 Hybrid aggression

As far as I can tell, many beekeepers believe that if their bees are dark, they are probably nearly pure Amm. Any lighter coloured bees (evil imports) come from Italian stock. When the evil orange drones mate with the perfect black queens, the likely outcome is horrible, aggressive bees (F2 aggression).

The problem with this line of thinking is that their own dark bees are mongrels, and the lighter coloured bees are also mongrels. You are not crossing two hybrid lines made from ‘pure’ parents. Not that this is statistically relevant, but I mix up all sorts of bees when producing queens, and I rarely see unduly aggressive colonies. It does happen – don’t get me wrong – but it is not the norm.

Even with true F₁ and F₂ hybrids, it is the variability that peaks at F₂ (some horrid bees, some lovely), and it settles down thereafter. This pattern is well known in animal breeding — hybrid variability peaks at F₂ due to the random reshuffling of parental alleles. The magnitude of the effect depends on how genetically distinct the parent populations were and which genes are interacting. As already explained, it is not the colour of the bees which necessarily marks one bee as significantly different from another.

Stage	What happens	Behavioural effect
F₁ (first filial)	Controlled cross between two pure subspecies or lines (e.g. Ligustica x Carnica)	Often docile and productive due to heterosis (hybrid vigour) – e.g. Buckfast Bees
F₂ (open mated daughters of F₁ queens)	Recombination splits linked alleles; worker traits can segregate unpredictably	Increased variability – some colonies calm, some more defensive
Later (F₃+)	Selection and local mating stabilise traits again	Temperament normalises once selection pressure resumes

By the way, let’s not forget that the way honey bees reproduce is literally designed to maximise genetic diversity. Virgin queens fly to drone congregation areas and mate with multiple drones from many different locations. It has been found that the higher the number of drones that mate with a queen (different patrilines), the better that queen will be. The colony will tend to do better by many different measures. Nature wants diversity; it likes mongrels. Breeding programmes are continually trying to balance the competition between selection and maintaining diversity.

All about the queen

Many readers will know that I love raising queens and believe that great queens are the cornerstone of productive, healthy, and good-to-work-with colonies. But I do not ever want anybody to forget the importance of drones. I have recently heard several speakers state that ‘varroa resistance’ is passed on by queens, and that drones are not very relevant. This ignorance baffles me. Of course drones matter! When you run a selective breeding programme the drones that your queens mate with carry the genetics of their mum. It is the combination of resistance traits carried by the queen and drones that matters. If the drones come from resistant queens, then the outcome is more likely to be positive than if the drones are from normal non-resistant stock. I discussed Randy Oliver’s work recently⁵, which shows how this works.

The other ‘chestnut’ is the story of how changing the queen of defensive bees makes them almost immediately become docile – well before the new queen’s brood has even been capped, let alone emerged. This proves that it is “all about the queen” and her lovely pheromones. Well, sometimes that is the case, but other times not. It rather depends on what is causing the defensive behaviour in the first place. If the old queen is failing, then replacing her with a younger model will probably induce the desired effect very quickly (assuming they don’t kill her – bees don’t always take the sensible option). However, if the defensive behaviour is due to starvation, robbing, wasp attack, or bad weather, it will be a different story. And, of course, if the defensive behaviour is genetic in nature, it won’t change until the genetics of the colony become those of the new queen – probably about six weeks.

Local is best?

The evidence shows that for somebody trying to breed varroa resistance bees, local certainly is best. Whatever magic happens to make a colony varroa-resistant, it seems to weigh heavily on environmental factors as well as genetics. You can take a resistant colony from one area, move it to another area, and the resistance can be lost.

The oft quoted study⁶: COLOSS pan-European “Genotype × Environment (GEI)” experiment led by Costa, Büchler, Meixner, Hatjina and colleagues, is the one generally used in the context of local bees being best. In this rather impressive experiment, the bees were not treated for varroa, and so most died. However, the local bees survived over 80 days longer than bees from far-off places, showing that local bees do better than outsiders. That is all well and good, but it can’t really be applied to bees that are kept alive by beekeepers. In that scenario – one where competent beekeepers treat for varroa, feed where necessary, and so on – I can only assume that most of the bees would not have died, local or otherwise.

I will concede that the ‘local is best’ approach makes intuitive sense, is backed up by the aforementioned authoritative study, and carries much lower risks of spreading hideous beasties like hive beetles and tropilaelaps mites. When I buy breeder queens, they come from UK-based breeders who control the genetics by using instrumental insemination. I’m really happy with the quality, and don’t need to bring in breeders from anywhere else. I’m also delighted to breed from any of my queens that I consider worthy.

However, in my opinion, the real issue is bringing in queens⁷ from areas far to the south, with entirely different climates to ours. I believe that queens coming in from Denmark or Germany are pretty close to ‘local’ from an ecological perspective. These are the areas, after all, which were originally populated by the native black bee – it was spread across northern and Western Europe, not just little old England. So the climate and forage can’t be that different. Bees from southern Italy, Greece, and Romania will not be so well adapted to the UK. I think it is important to differentiate between northern and southern bees.

Let them die

Apologies for going on a bit, but we are in the home straight. The final generalisation made by my local speaker recently was that feeding and treating bees merely propagates weak bees. He does not feed his bees, apparently, and if they die, they were rubbish anyway. Ditto for varroa treatments. I think most people know what I think of that!

• Yes, natural selection can produce varroa-tolerant or locally adapted lines under certain conditions.
• But: uncontrolled die-offs cause massive genetic bottlenecks, disease spillover (robbing and drifting), and loss of productivity traits.
• Environmental stress (e.g. wet summers) hits all bees — even adapted ones — so feeding is husbandry, not genetic interference.
• Modern breeding programmes that combine selection and management produce far better, more sustainable outcomes than “let them die” philosophies.
  

Claim	Kernel of truth	Why it’s flawed
“Treating for varroa props up weak genes”	True that treatments relax natural selection pressure for resistance traits	But they prevent total population collapse and buy time for managed resistance breeding. Bond approach needs large isolated populations
“Feeding stops natural selection for thriftiness”	True that overfeeding can select for poor storage behaviour	But feeding during abnormal weather is a welfare action, not artificial selection. Even locally adapted bees can starve in extreme years
“If they die without help, they are unfit”	Natural selection does not aim for productive, gentle bees – just keeps survivors alive	Survivors might be tiny, swarmy, aggressive, or unproductive. Fit does not equal useful
“Local adaptation only comes from letting bees die”	Environment plays a role in adaption	But selective breeding can accelerate adaptation while keeping diversity and desirable traits

Studies and NBU reports show large winter losses occurred after prolonged rain in some years. Even established local stocks required feeding. This is ecological stress, not a genetic defect. “Locally adapted” means tuned to average conditions, not invulnerable to extremes.

Feeding is environmental buffering, not a genetic crutch.

• Analogy: Farmers irrigate during droughts; that doesn’t mean their wheat is weak.
• A colony can be genetically thrifty yet starve if forage disappears for six weeks.
• Proper husbandry protects valuable genetics from random environmental extinction.
  

Proper breeding

I just want to finally shine a light on the incredible work done by elite bee breeders in Europe. There is a reason why their bees are purchased by many UK bee farmers, and it’s not just convenience. Several large UK bee breeders use genetic material from these programmes. They really are excellent bees that do exceptionally well for UK bee farmers. Here is a quick list of some things elite breeders do:

1) Control the mating, using isolated island mating stations and instrumental insemination

2) Test performance under standard protocols: multi trait, recorded in national and international databases

3) Calculate breeding values (EBVs): not just ‘nice colonies’. EBVs and tools (mate planning, sibling comparisons, line overviews) are visible on BeeBreed⁸ to guide who mates with whom next season.

4) Use selection indices: EBVs for several traits are combined into a selection index (with economic or programme-specific weights) so you don’t accidentally ‘win’ on yield while losing on temper or varroa traits.

5) Manage inbreeding and line structure: Registries/studbooks track pedigrees and inbreeding coefficients; mating plans avoid close matings and help keep effective population size up while still making genetic gain.

6) Adopt genomic selection where available: using DNA as well as traits and pedigrees.

7) Quality assure queen production and distribution: age windows, banking rules etc.

Now, compared to that lot, I am indeed whistling in the dark. Nevertheless, I enjoy raising my own queens and I know that my bees are good. I see it every day and hear good feedback from nuc customers.

Now that I have got that lot off my chest, I can have a nice cup of tea and watch a K-drama on Netflix. It’s not all about bees, even for me!

References

1. Hong-Yi Nie, Li-Qiang Liang, Qiu-Fang Li, Zheng-Han-Qing Li, Ya-Nan Zhu, Yong-Kang Guo, Qiu-Lan Zheng, Yan Lin, Dong-Lin Yang, Zhi-Guo Li, Song-Kun Su, CRISPR/Cas9 mediated knockout of Amyellow-y gene results in melanization defect of the cuticle in adult Apis mellifera, Journal of Insect Physiology, Volume 132, 2021, 104264, ISSN 0022-1910, https://doi.org/10.1016/j.jinsphys.2021.104264.
2. A bold statement, but I did search. If I’m wrong about this, please let me know in the comments.
3. The health and status of the feral honeybee (Apis mellifera sp) and Apis mellifera mellifera population of the UK https://etheses.whiterose.ac.uk/id/eprint/5211/1/CorrectedThesis3.pdf.
4. Buswell, V., Huml, J., Ellis, J., Brown, A., & Knight, M. (2024) ‘Whole genome analyses of introgression in British and Irish Apis mellifera mellifera’, Journal of Apicultural Research, . Available at: 10.1080/00218839.2024.2411483.
5. https://thewalrusandthehoneybee.com/non-treatment-sanity-check/
6. https://scispace.com/pdf/the-influence-of-genetic-origin-and-its-interaction-with-1wbvxo3ptt.pdf
7. https://thewalrusandthehoneybee.com/local-bees/
8. https://www2.hu-berlin.de/beebreed/ZWS/