From Varroa control to breeding the host

Preface

Varroa destructor on an Apis mellifera pupa inside a capped cell. The mite reproduces in capped brood

Selective breeding of Apis mellifera for resistance to Varroa destructor is one of the most promising ways to make beekeeping more sustainable and gradually reduce reliance on chemical control of the parasite. Resistance results from the interaction of several behavioural, physiological, genetic and social mechanisms. These include general hygienic behaviour (HYG), Varroa sensitive hygiene (VSH), mite non-reproduction (MNR, a more precise term for reduced mite reproduction), recapping and grooming.

This article examines the genetic and behavioural basis of these traits, with particular attention to quantitative trait loci (QTL) and candidate genes, the social structure and patrilines of the colony, heritability, test repeatability and phenotyping methods. These include the pin test, the freeze-killed brood test, the VSH test and the Harbo test. It also compares studies of different sizes, distinguishing research on at least 100 colonies housed in standard Dadant or Langstroth hives from experiments using nucleus colonies or smaller samples.

The available literature shows that breeding for resistance to Varroa can be combined with selection for honey production, spring development, overwintering and gentleness. Pol-line honey bees (a breeding population of Apis mellifera developed in the United States specifically for greater resistance to Varroa destructor while retaining traits useful for commercial beekeeping) provide a prominent example under conditions close to commercial practice. A longitudinal study of 366 colonies in 10-frame Langstroth hives found significantly better survival in the selected population, with no significant reduction in honey production among surviving colonies [1]. Similarly, a Canadian programme following 604 colonies over eight generations found that infestation by Varroa, spring development, honey production, winter food consumption and hygienic behaviour have a genetic component that can respond to selection and can be included in a single selection index [2].

The practical conclusion is to treat VSH as a priority health trait within a multi-trait breeding index. A commercial operation needs a population that combines consistently high resistance to Varroa with productivity, gentleness, a low tendency to swarm, colony development, longevity and adaptation to local rearing conditions.

1. Introduction: from controlling Varroa to breeding the host

Varroa destructor is a major cause of mortality in managed colonies of Apis mellifera worldwide. In addition to feeding on host tissues, particularly the fat body, the mite disrupts host physiology and plays an important role in transmitting and amplifying viruses, impairing the health of both individuals and colonies [3].

Conventional control targets the parasite directly with acaricides, biotechnical methods or both. Selective breeding instead changes the susceptibility of the host over successive generations.

A realistic breeding goal concerns the growth of the population of Varroa. If mite numbers rise slowly enough, the colony may be able to function normally for a substantial part of the year.

The systematic review by Mondet et al. found that survival in the presence of Varroa is associated with a range of phenotypic and genomic mechanisms [4].

This leads to a first principle:

Varroa resistance is a complex colony-level trait that cannot be captured by a single gene or diagnostic test.

2. Individual and social immunity

Honey bee defences against pathogens can be grouped broadly into:

1. individual defences;

2. social defences.

The second group includes behaviours in which workers recognise and remove infected, parasitised or dead individuals. Hygienic behaviour is therefore a form of social immunity.

A worker can:

• detect an abnormality;

• locate the cell;

• uncap it;

• remove the pupa or larva;

• carry the material away.

In the case of Varroa, this mechanism is particularly complex because the worker must recognise cues associated with the mite's presence and reproductive activity.

3. HYG: general hygienic behaviour

Hygienic behaviour (HYG) is the ability of workers to detect and remove dead, diseased or otherwise abnormal brood. The phenotype comprises two main behavioural components: detection and uncapping and removal.

Hygienic behaviour (HYG) is the ability of workers to detect and remove dead, diseased or otherwise abnormal brood

Rothenbuhler originally proposed a relatively simple model with separate genetic factors for uncapping and removal. Subsequent quantitative studies found a much more complex architecture.

Lapidge, Oldroyd and Spivak identified seven suggestive QTL associated with hygienic behaviour, each explaining a limited share of the variation in the population studied [5].

Oxley, Spivak and Oldroyd later identified six QTL associated with the behavioural threshold at which a worker begins hygienic activity [6].

These findings indicate that the phenotype arises from the interaction of multiple genomic regions and, in particular, from how they are expressed in the social context of the colony.

4. VSH: a specialised form of hygienic behaviour

Varroa sensitive hygiene (VSH) is a specific form of hygienic behaviour in which workers preferentially detect brood infested with Varroa and remove it. The distinction from general HYG matters. The conventional HYG test asks:

“How quickly does the colony detect and remove a dead pupa?”

Experimental design used to assess the VSH response to cells containing Varroa and to different cues associated with the mite's presence

The VSH test asks a different question:

“How effectively does the colony detect and remove brood associated with Varroa, particularly when the mite is reproducing?”

Tsuruda et al. identified genomic regions associated with VSH, including a QTL on chromosome 9 [7]. These regions contain candidate genes linked to sensory perception and stimulus processing.

The identified regions explain only part of the variance, and the overall genetic architecture remains polygenic.

5. Candidate genes: olfaction, the nervous system and behaviour

One line of VSH research concerns chemical perception. Mondet et al. compared the antennal transcriptomes of bees that displayed VSH with those of bees that did not [8].

The findings suggest that the antennae play a central role in detecting cues from infested brood. The profile of VSH bees was more similar to that of nurse bees than to that of foragers, consistent with the behaviour being expressed mainly by workers performing tasks inside the nest.

The literature discusses several groups of genes and candidates:

• OBPs, odorant-binding proteins;

• olfactory receptors;

• CSPs, chemosensory proteins;

• nanchung;

• dopaminergic pathways;

• Dscam;

• genes involved in neuronal development and connectivity.

Dscam is of particular interest. A genome-wide association study (GWAS) of 155 colonies of A. m. mellifera from France and Switzerland identified two QTL significantly associated with recapping of brood infested by Varroa. One was located in an intron of Dscam, a gene involved in neuronal connectivity [9].

Taken together, the findings suggest that behavioural resistance involves several linked processes:

perception → neural processing → learning and behavioural thresholds → social response.

6. SMR, MNR and VSH: three distinct concepts

The terms SMR, MNR and VSH are a recurring source of confusion in research and technical publications.

SMR – Suppressed Mite Reproduction

The term was used to describe colonies in which a very high proportion of mites in the brood did not reproduce. Harbo and Harris later showed that adult bee behaviour accounted for an important part of this phenomenon [10]. In one experiment, naturally infested combs of freshly capped brood were transferred to SMR and control colonies. After eight days, SMR colonies had a much lower proportion of infested cells than controls, and a much smaller share of the remaining mites were reproductive [10].

MNR – Mite Non-Reproduction

MNR is now the more descriptive term. It refers to the observed mite phenotype, namely failure to reproduce.

VSH

VSH describes the behaviour of the bee.

In summary:

VSH = a behavioural trait of the bee

MNR = an observable outcome in the mite's reproductive cycle

SMR = the historical term for the phenotype and the original selection approach

The distinction matters because non-reproduction can have several causes. Its presence alone does not establish VSH [4,10].

7. The colony as a genetic and social unit

One feature of honey bee genetics sets breeding apart from breeding in many other domesticated animals. The queen generally mates with many drones, so the colony comprises several patrilines. Workers share a mother but may have different fathers. A single colony therefore contains several genotypes.

A colony's VSH phenotype depends on the proportion of workers predisposed to the behaviour and on how they interact. Selection must therefore account for both parents. The queen contributes a substantial share of the colony's genetic potential, and the drones that sire her offspring also help determine the phenotype.

This has a practical consequence:

a queen from a high-VSH line can produce a colony with modest VSH if the paternal lines have not been selected for the trait.

8. The role of nurse bees

Workers that perform tasks inside the nest are chiefly responsible for hygienic behaviour. Mondet et al. found that the profile of VSH bees was closer to that of nurse bees than to that of foragers [8]. This is consistent with the biology of the colony: workers that carry out tasks inside the nest are the most likely to encounter and interact with capped brood.

Selection therefore needs to consider the colony as a social system, with the individual as one part of that system.

9. Measuring HYG: the pin test

The pin test is a simple method for preliminary assessment of hygienic behaviour. A patch of capped brood of uniform age is selected and the pupae are mechanically injured with a needle. It is preferable to use a standard number of cells, such as 50–100. Counts are made at predetermined intervals, typically:

• 12 hours;

• 24 hours;

• 48 hours if needed.

Example of an assessment of hygienic behaviour using artificially damaged brood. The percentage of cells from which brood has been removed is recorded at standard intervals

The final percentage matters, as does the speed of removal. A colony that clears 95% of the cells in 12 hours differs behaviourally from one that reaches the same level after 48 hours. The pin test is therefore particularly useful for:

• preliminary screening;

• rapid selection;

• comparisons within an apiary.

A definitive genetic assessment should draw on other evidence as well.

10. Freeze-killed brood test

The freeze-killed brood test (FKB) is easier to standardise. A patch of brood is frozen and then returned to the colony. Observations determine the percentage of cells from which the colony has removed the brood.

Facchini et al. compared the conventional FKB method with a shortened variant, FKB* [11]. In 2016, the methods were compared in 25 colonies. In the subsequent breeding programme, 56 colonies were phenotyped in 2017 and 95 in 2018, with two records per colony [11].

The correlation between the colony effects for the two methods was: r = 0.93

Estimated heritability for a single record was: h² = 0.23 and rose to the following value when two records were averaged: h² = 0.37.[11]

The practical implication is:

repeating the test improves the reliability of genetic evaluation

Professional breeding programmes should therefore preferably use two HYG assessments.

11. What heritability of 23% or 37% actually means

A common misinterpretation is:

“A trait with h² = 0.37 is transmitted at a rate of 37%.”

In narrow-sense heritability, h² = Vₐ/Vₚ. It represents the proportion of phenotypic variance attributable to additive genetic variance in the population and experimental conditions considered. It describes variation across the population.

The expected response to selection can be approximated by R = h²S, where S is the selection differential.

Selecting individuals well above the population mean can therefore improve the average genetic merit of the next generation when the trait has substantial heritability.

12. Heritability of economically important traits

A reference for commercial beekeeping is the Canadian breeding programme of Maucourt et al., which covered 604 colonies across eight generations [2]. The reported estimates were:

Trait	Estimated h²
Infestation by Varroa	0.44 ± 0.56
Spring development	0.30 ± 0.14
Honey production	0.20 ± 0.13
Hygienic behaviour	0.18 ± 0.13
Winter food consumption	0.11 ± 0.09

Some estimates carry considerable uncertainty. In particular, the estimate of 0.44 for infestation by Varroa applies to the population and conditions studied. The main contribution of the study was to show that:

1. the traits can respond to genetic selection;

2. they can be selected simultaneously;

3. a multi-trait selection index can be put into practice [2].

4. 
   

13. Keeping honey yield in the selection index

A VSH breeding programme needs to select for productive colonies as well as for survival. In commercial beekeeping, a key economic outcome is honey yield per colony. The Canadian programme estimated the heritability of honey production at approximately: h² = 0.20 ± 0.13.[2]

This estimate suggests that honey yield can respond to selection, while remaining strongly affected by:

• environmental conditions;

• nectar availability;

• climate;

• colony strength;

• management;

• the apiary;

• geographic location.

Evaluation should therefore cover several years and, ideally, several environments.

14. Evidence from 366 colonies: the Pol-line case

The Pol-line study is particularly relevant because it examined colonies under conditions comparable to commercial beekeeping.

O’Shea-Wheller et al. studied 366 colonies:

• 193 commercial colonies;

• 173 Pol-line colonies.

The colonies were housed in 10-frame Langstroth hives and followed over time [1]. Survival was 59.884% among Pol-line colonies compared with 26.042% among commercial colonies. The hazard of death was significantly lower in the Pol-line colonies. Of particular economic interest, the comparison of honey yields from 141 commercial and 137 Pol-line colonies found no statistically significant difference [1].

For commercial beekeeping, these data offer evidence from a larger and more representative setting than a trial involving a few nucleus colonies. The results support the following conclusion:

a population selected for Varroa resistance can maintain commercially competitive honey yields.

15. Evidence from 268 colonies in Langstroth hives

Another study, by Kefuss et al., began in 1999 with 268 commercial colonies housed in standard deep Langstroth hives and used for honey production and queen rearing [12]. The colonies came from commercial stock that included A. m. ligustica, A. m. carnica, A. m. caucasica and A. m. mellifera. Acaricide treatment was discontinued, and survival became a selection criterion. After more than two-thirds of the original colonies were lost, the survivors supplied breeding stock for a new population. By 2013, the programme had 519 untreated colonies from the two groups, with very low mite levels [12].

The study shows that it is possible to select a resistant population under commercial conditions. Its scope matters when interpreting the results: selection was based heavily on survival without treatment, and the experiment was not designed to isolate VSH. It therefore provides evidence of overall resistance or tolerance, with the role of specific mechanisms still to be established.

16. Evidence from 604 colonies across eight generations

The Canadian programme provides another reference [2]. It included 604 colonies across eight generations. When the population size allowed, each generation comprised about 100 colonies.

Each year, around 20 colonies with the best overall performance were selected. Of those:

• 7–14 supplied maternal breeding stock;

• around 10 supplied drones [2].

This programme illustrates the transition from selecting on individual phenotypes to breeding at population level.

17. Evidence from 649 colonies: a large sample with variation in colony size

Gerdts et al. studied 649 colonies across nine beekeeping operations and 21 apiaries in Australia [13]. The large sample improves the robustness of the estimates. However, colony size ranged from nucleus colonies on about four frames to production colonies with bees covering more than 16 frames. This variation in size needs to be taken into account.

Sixteen per cent of colonies met the threshold for strict hygienic behaviour, defined as removal of at least 95% of dead brood [13]. Colonies in breeding programmes were 5.8 times as likely to be highly hygienic, while colonies headed by queens descended from hygienic mothers were 2.2 times as likely [13]. These associations are consistent with a heritable component that can respond to selection. Environmental conditions can also affect how the trait is expressed.

18. What studies with small samples can tell us

Studies involving a few colonies can be important for:

• discovering a mechanism;

• identifying a candidate gene;

• developing a method;

• demonstrating a causal relationship;

• comparing extreme phenotypes.

The difficulty arises when such studies are used to estimate the genetic response of a commercial population. For example, Harbo and Harris used 17 colonies in their study of the mechanism behind SMR, with nine SMR colonies and eight controls [10]. Their experiment helped explain the mechanism. Estimating the frequency of the trait in a commercial population requires other evidence.

Studies at larger scales include:

• 155 colonies in the recapping GWAS [9];

• 268 colonies in the commercial selection study by Kefuss et al. [12];

• 366 colonies in the Pol-line comparison [1];

• 604 colonies in the Canadian programme [2];

• 649 colonies in the Australian study [13].

Such studies can give breeders a firmer basis for population-level decisions. Study design also matters: a poorly controlled experiment with 150 colonies may be less informative than an experiment with 40 colonies and a rigorous genetic design.

19. Dadant and Langstroth hives and mini-hives: the question of biological scale

For commercial beekeeping, it helps to distinguish among:

• experimental colonies;

• nucleus colonies;

• mini-hives;

• production colonies.

Colony size influences:

• the amount of brood;

• the number of nurse bees;

• nest temperature;

• the frequency of interactions;

• resource availability;

• pressure from Varroa;

• the expression of social behaviours.

A result obtained in a nucleus colony cannot automatically be extrapolated to a production colony in a Dadant-Blatt or Langstroth hive. Trials intended to guide commercial breeding programmes should therefore be conducted primarily on full-size production colonies. The Pol-line and Kefuss studies are relevant in this respect because they used standard Langstroth hives and large populations [1,12].

20. An important limitation of the freeze-killed brood test

The FKB test measures general hygienic behaviour. A colony that removes a frozen pupa quickly may show a different response to mite-infested brood. Spivak and Danka pointed out that removal of artificially killed brood does not always reliably predict resistance to specific pathogens or parasites [14]. In the Australian study, 16% of the 649 colonies tested were highly hygienic, but hygienic behaviour did not significantly predict the presence or severity of chalkbrood [13].

These findings clarify what HYG measures:

HYG reflects general social immunity, while VSH is a more specific trait.

21. Annual phenotyping schedule

A professional breeding programme should spread testing across the season.

For northern and central Italy, one possible schedule is:

Period	Assessments
February–March	survival, colony strength, food stores, queen status
March–April	spring development, gentleness, swarming
April–May	pin test or FKB test
May–June	second HYG assessment, development, production
June–July	VSH, Harbo test, Varroa in brood
July–August	Varroa population growth, production, colony population
August–September	infestation of adult bees, Varroa in brood, MNR
September–October	final honey yield, winter readiness, health status
Autumn–winter	data analysis and selection of breeding stock

Dates should be adjusted to local climate and phenology. The main point to standardise is the biological stage at which the comparison is made.

22. Pin test protocol

For each colony:

1. select a comb with capped brood of uniform age;

2. mark out 50–100 cells;

3. pierce each pupa with an entomological pin;

4. record the initial number of cells;

5. photograph the area;

6. check after 12 and 24 hours;

7. record the percentage of cells uncapped and the percentage from which brood was removed.

For selective breeding, record the 24-hour result:

HYG₂₄ = (cells cleared by 24 h ÷ cells initially treated) × 100

and the earlier measurement as an indicator of removal speed:

HYG₁₂

so that colonies reaching the same final result at different rates can be distinguished.

23. FKB protocol

Use a standardised patch of brood for the freeze-killed brood test, preferably with at least 100 assessable cells.

After freezing:

1. return the section to the colony;

2. photograph it;

3. check after 24 hours;

4. record the percentage of brood removed;

5. repeat the test at another stage of the season.

Removal of 95% within 24 hours is one possible threshold. Gerdts et al. assessed removal after 48 hours [13].

24. Direct VSH test

Direct measurement of VSH requires brood infested with Varroa.

The general procedure is to:

1. select combs with capped brood;

2. identify cells containing Varroa;

3. record an initial sample;

4. leave the brood in the colony for a defined interval, typically about a week;

5. check which cells have been uncapped or cleared;

6. assess the presence of mites;

7. compare infested and uninfested cells.

The key requirement is sufficient parasite pressure. A colony with negligible infestation cannot be classified as highly VSH solely because no mites are found.

25. Harbo test

Examination of capped brood to assess Varroa destructor reproduction. Classifying the offspring helps distinguish reproductive from non-reproductive mites

The Harbo test assesses the reproduction of Varroa in brood.

Capped cells are opened to assess:

• presence of a foundress mite;

• presence of eggs;

• presence of a male;

• presence of daughters;

• development of offspring;

• failure to reproduce.

Developmental stages of Varroa destructor

The sample should be large enough to contain an adequate number of infested cells. For a professional breeding programme, one proposed target is at least 100 infested cells suitable for assessment, with a larger sample when mites are scarce. A simple count records only: the number of mites found. The reproductive status of the mites also needs to be classified. A colony with few mites, all of which are reproducing, cannot automatically be ranked above a colony with more mites and a very high proportion of non-reproductive mites.

26. Interpreting MNR alongside VSH

For each colony, calculate:

MNR = (non-reproductive mites ÷ total mites that could be assessed) × 100

alongside:

VSH = (infested cells cleared ÷ initially infested cells) × 100

Using both measures helps distinguish:

• colonies that remove a large amount of infested brood;

• colonies in which the remaining mites show low reproductive success;

• colonies with both phenotypes;

• colonies with low infestation because parasite pressure is low.

  
  

27. Varroa on adult bees and in brood

Monitoring resistance requires measures beyond VSH. At intervals, record:

• mites on adult bees;

• mites in brood;

• the seasonal trend;

• the relative growth of the mite population.

A single observation tells us little about the trend. The growth rate is more informative:

r = ln(Nₜ/N₀) ÷ t

where N₀ is the initial infestation level, Nₜ is the level after a defined interval, and t is the duration of that interval. In a breeding programme, a colony that consistently limits the growth of Varroa throughout the season is of greater interest than one with a single very low reading.

28. Environmental effects

Environmental conditions strongly influence hygienic behaviour and VSH. Factors to record include:

• nectar availability;

• protein availability;

• temperature;

• rainfall;

• time of year;

• colony strength;

• amount of brood;

• pressure from Varroa;

• concurrent diseases;

• beekeeping practices.

In the large Australian study, nectar availability influenced the expression of hygienic behaviour [13]. For a fair genetic comparison, assess colonies as far as possible:

• in the same apiary;

• at the same time;

• at comparable strength;

• under comparable feeding conditions.

  
  

29. A selection index for commercial beekeeping

A single-trait index would be: I = VSH.

A multi-trait index could instead take the following form:

I = w₁VSH + w₂M + w₃D + w₄S + w₅H + w₆SW + w₇HYG

where:

• VSH = hygienic response to brood infested by Varroa;

• M = honey yield;

• D = gentleness;

• S = spring development;

• H = overwintering;

• SW = low swarming tendency;

• HYG = general hygienic behaviour.

One possible set of starting weights is:

Trait	Illustrative weight
VSH, MNR and infestation by Varroa	30%
Honey production	25%
Gentleness and calmness	15%
Spring development	10%
Overwintering and longevity	10%
HYG	5%
Low swarming tendency	5%

Adjust the weights to reflect the operation's actual costs and returns.

30. Keeping gentleness as a breeding criterion

Gentleness has economic value in commercial beekeeping. A defensive colony increases:

• handling time;

• use of smoke;

• occupational risk;

• stress for workers;

• management difficulty;

• indirect costs.

A highly defensive colony with exceptional VSH should therefore be assessed against a mandatory minimum standard for gentleness before it is used for breeding.

The same principle applies to swarming tendency.

31. Swarming tendency as an exclusion criterion

A low tendency to swarm is particularly important in commercial beekeeping. A simple record of whether the colony swarmed gives only part of the picture. It is preferable to record:

• number of queen cells;

• date when they first appear;

• number of interventions required;

• number of swarms;

• frequency of swarm cells;

• associated loss of production.

A colony that repeatedly needs swarm control is less economical to manage, even if it has excellent VSH.

32. Measuring honey yield directly

Selection should use kilograms of honey produced per colony

alongside indirect measures such as:

• number of occupied frames;

• brood area;

• apparent colony strength.

Colony strength is an intermediate trait. Honey yield is the economic outcome. Production should therefore be measured over at least the main nectar flows.

A commercial breeding programme needs to distribute selection pressure across a range of biological and economic traits. These include general hygienic behaviour and VSH, reduced mite reproduction, colony development, honey yield, gentleness, swarming tendency and overwintering ability. The aim is a stable genetic balance between parasite resistance and overall value to beekeepers.

Multi-trait selection in Apis mellifera

Breeding for Varroa resistance should include the principal traits of economic interest. VSH, general hygienic behaviour, reduced mite reproduction and survival are the main resistance-related traits. Honey yield, colony development, gentleness, low swarming tendency and overwintering ability determine the overall value of the line to beekeepers. Selection aims to improve the combination of traits over successive generations.

33. A five-year breeding plan for Varroa-resistant bees suited to commercial beekeeping

Year 1 – Baseline assessment

Suggested population: 100–200 colonies.

Preferred conditions:

• standard Dadant-Blatt or Langstroth hives;

• individually identified queens;

• initial pedigree records;

• comparable apiaries;

• uniform management.

The first year is for collecting a phenotypic baseline before applying strong selection pressure. Record the following for each colony:

• overwintering;

• colony strength;

• brood;

• honey yield;

• gentleness;

• swarming tendency;

• HYG;

• Varroa infestation;

• VSH or Harbo test results when parasite pressure permits.

  
  

Year 2 – Initial selection

The breeding pool can be narrowed to 20–30 candidate maternal colonies and 20–40 drone-producing colonies. Selection should proceed in two stages.

First screen

Exclude colonies that are:

• highly defensive;

• highly prone to swarming;

• low in productivity;

• diseased;

• weak;

• poor at overwintering.

Second screen

Among the remaining colonies, select for:

• high HYG scores;

• high VSH scores;

• high MNR;

• slow population growth of Varroa.

This prevents a colony with poor commercial performance from entering the programme on the strength of a single test.

Years 2–3 – Instrumental insemination

Instrumental insemination of an Apis mellifera queen. Controlling the paternal contribution makes planned crosses possible in a breeding programme

Instrumental insemination allows control of the paternal contribution. There are two main approaches.

Single-drone insemination

Semen comes from one drone.

Advantages:

• precise pedigree records;

• close genetic control;

• useful for studying segregation;

• useful for fixing combinations of traits.

Disadvantages:

• risk of inbreeding;

• strong dependence on how genetically representative the chosen drone is.

• 
  

Multiple-drone insemination

Semen comes from several selected drones.

Advantages:

• greater diversity;

• lower risk of inbreeding;

• greater colony robustness.

For commercial populations, retaining some paternal diversity is generally preferable.

Selection of drone-producing colonies

Breeding programmes also need to select the paternal lines. Drones should come from colonies with:

• high VSH scores;

• high HYG scores;

• good productivity;

• gentleness;

• good health;

• good survival;

• low infestation.

The Canadian programme provides a precedent: around 20 colonies were selected each year for breeding, some to produce queens and others to produce drones [2].

Year 3 – Progeny testing

The focus shifts from the mother to the family. Produce enough daughter queens from each mother to assess the family. The question moves from

“How strong is this queen’s VSH?”

to

“How strong is VSH among her offspring on average, and how stable is the trait?”

Expression can decline in subsequent generations. An exceptional mother alone gives limited evidence that the line will remain stable.

Years 3–4 – Family selection

For each family, calculate:

• mean VSH score;

• standard deviation;

• mean HYG score;

• mean infestation level;

• mean yield;

• mean gentleness score;

• swarming tendency;

• overwintering performance.

A family with less extreme but more consistent results may have greater breeding value than a family with one exceptional colony. The principle is:

family breeding value ≠ best individual phenotype

Year 4 – Testing under production conditions

Test the best lines in:

• Dadant-Blatt hives;

• Langstroth hives;

• production apiaries;

• actual nectar flows;

• migratory beekeeping conditions where relevant.

At a minimum, compare:

• honey yield;

• colony strength;

• swarming;

• gentleness;

• Varroa infestation;

• overwintering;

• mortality;

• queen longevity.

This phase establishes whether the lines have enough commercial value to justify their use in production.

Year 5 – Consolidation

In the fifth year, the programme should be able to identify:

• 10–15 principal maternal families;

• 20–40 paternal families;

• numerous tested daughter queens.

The aim is to build a genetically resistant population that retains sufficient diversity. Genetic diversity is especially important in honey bees because of complementary sex determination, which involves the csd gene. Intense selection among a few lines can raise the risk of inbreeding and reduce overall fitness.

34. The role of natural selection

A commercial breeding programme can combine artificial selection with controlled natural selection. Kefuss et al. describe an approach in which survival in the presence of Varroa was used as an initial screen. Surviving colonies were then selected for economic traits as well [12]. This approach can exert strong selection pressure, at the cost of high mortality. Commercial operations may need to limit the initial colony losses.

A more conservative strategy uses:

• Varroa monitoring;

• VSH testing;

• MNR;

• HYG;

• survival;

• honey yield;

as combined criteria.

35. Selective breeding and acaricide treatment

A professional breeding programme should define how to manage heavily infested colonies and distinguish the breeding apiary from the production apiary.

Breeding apiaries can use predefined experimental thresholds and authorised protocols. In production apiaries, colony protection and health management take priority. If a candidate colony shows uncontrolled population growth of Varroa, its infestation calls for intervention under the applicable protocol.

The purpose of selection is to improve population fitness while limiting avoidable colony losses.

36. Exclusion criteria

Exclude a candidate breeding colony if it shows:

• defensive behaviour incompatible with commercial management;

• recurrent swarming;

• low honey yield;

• poor overwintering;

• high susceptibility to disease;

• a markedly irregular brood pattern;

• rapid population growth of Varroa;

• abnormal queen behaviour.

A high VSH score should not outweigh a serious production or management problem.

37. Suggested selection method

A practical approach combines minimum thresholds with a selection index.

First screen

The colony must meet the minimum resistance criteria:

• low infestation;

• slow population growth of Varroa;

• HYG above a defined threshold;

• VSH or MNR above a defined threshold.

Second stage

Among colonies that pass the screen, select for:

• higher honey yield;

• greater gentleness;

• lower swarming tendency;

• better colony development;

• better overwintering.

In shorthand: breeding stock = minimum resistance standard + highest economic value

This approach aligns breeding decisions with the practical needs of commercial beekeeping.

38. Which studies should carry most weight in breeding decisions?

Pol-line colonies developed through US breeding programmes for Varroa resistance. Pol-line illustrates selection under commercial beekeeping conditions

For a commercial breeding programme, the evidence can be ranked as follows:

First tier

Studies with:

• ≥100 colonies;

• production colonies;

• standard hives;

• pedigree records;

• multiple generations;

• repeated measurements.

Examples include:

604 colonies across eight generations – Maucourt et al.[2]

366 colonies in Langstroth hives – Pol-line.[1]

268 colonies in Langstroth hives – Kefuss et al.[12]

649 colonies – Australian study.[13]

155 colonies – recapping GWAS.[9]

Second tier

Genetic and mechanistic studies with smaller samples and tightly controlled experimental designs:

• Tsuruda et al.;

• Harbo & Harris;

• Mondet et al.;

• Oxley et al.

Third tier

Studies involving:

• nucleus colonies;

• mini-hives;

• <30 colonies;

• a single season.

These studies are mainly exploratory.

39. High heritability estimates from small samples

A high estimate of h² can have a large standard error when it comes from a small sample. An estimate of h² = 0.65 from a small experiment may be less reliable than an estimate of h² = 0.20 from hundreds of colonies across several generations. Report the estimate alongside:

• sample size;

• standard error;

• pedigree structure;

• environment;

• statistical model.

  
  

40. Following performance over time

An isolated result gives limited information:

VSH in 2026 = 80%.

A more useful record would be:

VSH in 2026 = 80%; VSH in 2027 = 76%; VSH in daughter colonies = 74%; low autumn Varroa infestation; above-average honey yield; high gentleness.

Breeding value is better assessed by examining the consistency of the phenotype over time. The five-year programme should therefore be designed as a longitudinal study.

41. Suggested colony breeding record

Assign each colony a permanent identity.

Identification

• colony ID;

• queen ID;

• queen’s year of birth;

• mother;

• paternal source;

• apiary;

• hive.

Health

• Varroa infestation in May;

• Varroa infestation in July;

• Varroa infestation in August or September;

• MNR;

• VSH;

• HYG.

Production

• spring honey yield;

• summer honey yield;

• total honey yield.

Colony biology

• adult bee population;

• brood;

• colony development;

• overwintering.

Behaviour

• gentleness;

• calmness;

• swarming tendency;

• tendency to rob other colonies.

Over time, these records become more informative than individual tests.

42. Final decision model

A candidate colony could receive a standardised score from 0 to 100 for each trait. For example:

Trait	Score
VSH	90
MNR	85
HYG	88
Honey yield	92
Gentleness	95
Colony development	90
Overwintering	87
Low swarming tendency	91

The overall index could then be calculated as follows:

I = 0.30(VSH) + 0.25(honey yield) + 0.15(gentleness) + 0.10(development) + 0.10(overwintering) + 0.05(HYG) + 0.05(low swarming tendency)

It is particularly important to set minimum thresholds.

For example:

• VSH below threshold → exclude;

• gentleness below threshold → exclude;

• honey yield below threshold → exclude.

This prevents an exceptional result on one trait from masking a serious shortcoming for commercial production.

43. Conclusions

The available research supports the conclusion that resistance of Apis mellifera to Varroa destructor is a complex, heritable trait that can respond to selection.

VSH is one of several mechanisms involved in overall resistance. The broader biological picture includes:

general hygienic behaviour, VSH, MNR or SMR, recapping, grooming, brood dynamics, the colony response and survival.

The genetic component is polygenic. QTL identified by Tsuruda et al., Lapidge et al. and Oxley et al. show that behavioural variation is associated with multiple genomic regions [5–7]. Molecular studies of the antennae also suggest that olfaction and the peripheral nervous system play an important role in VSH [8].

Colony social structure also matters. A colony consists of cooperating workers from different patrilines. Breeding programmes therefore need to select both maternal and paternal lines and test the offspring.

Heritability estimates need to be interpreted in context. In the Italian study by Facchini et al., estimated heritability was 0.23 for a single HYG measurement and 0.37 when two measurements were averaged [11]. The Canadian programme involving 604 colonies found that infestation by Varroa, spring development, honey yield, winter food consumption and HYG all have a genetic component that can respond to selection [2].

Large-scale studies are especially relevant. In the Pol-line study, which involved 366 colonies in 10-frame Langstroth hives, survival was about 60% compared with 26% in the commercial stock, with no significant difference in honey yield between the surviving groups assessed [1]. Kefuss and colleagues started with 268 commercial colonies in Langstroth hives, and the programme later reached 519 untreated colonies. Their results show that selection under commercial conditions can yield populations with low levels of Varroa [12].

These results are especially relevant to commercial breeding because they extend beyond trials in nucleus colonies or in a few dozen colonies. The main implication concerns which colonies to breed from.

A high VSH score is one criterion among several. Select colonies with:

• adequate VSH;

• slow population growth of Varroa;

• high HYG;

• a high proportion of non-reproductive mites;

• high honey yield;

• gentleness;

• calmness on the comb;

• low swarming tendency;

• good spring development;

• good overwintering;

• good queen longevity;

• good overall health.

The breeding objective combines:

Varroa resistance + production + ease of management + fitness

while taking into account:

VSH as one component of resistance

In commercial beekeeping, VSH is most useful as part of multi-trait selection.

A five-year programme starting with 100–200 colonies provides a firmer basis for selection when it includes repeated phenotyping, pedigree records, selection of maternal and paternal families, instrumental insemination, testing of daughter colonies and evaluation under standard production conditions.

The ultimate aim is to build a stable, productive, gentle commercial population with progressively greater resistance to Varroa destructor.

Final methodological note

The heritability estimates reported here describe variation within the populations studied. They depend on the population, pedigree, environment, measurement method, season and statistical model. Operational thresholds for HYG and VSH serve as selection tools and can vary with the protocol. An Italian breeding programme could use estimates from the literature as initial prior information and gradually replace them after three to five generations with genetic parameters estimated from its own population. Keep individual records for each colony and queen, including maternal and paternal lines. The performance of a VSH line is best evaluated over time and across its offspring. Look for slow population growth of Varroa and adequate survival alongside sustained honey yield, gentleness, colony development and consistent productivity.

References

[1] O’Shea-Wheller, T. A., Rinkevich, F. D., Danka, R. G., Simone-Finstrom, M., et al. (2022). A derived honey bee stock confers resistance to Varroa destructor and associated viral transmission. Scientific Reports, 12, 4852. DOI: 10.1038/s41598-022-08643-w. (PubMed Central (PMC))

[2] Maucourt, S., Fortin, F., Robert, C., & Giovenazzo, P. (2020). Genetic Parameters of Honey Bee Colonies Traits in a Canadian Selection Program. Insects, 11(9), 587. DOI: 10.3390/insects11090587. (PubMed Central (PMC))

[3] Noël, A., Le Conte, Y., & Mondet, F. (2020). Varroa destructor: how does it harm Apis mellifera honey bees and what can be done about it? Emerging Topics in Life Sciences, 4, 45–57. DOI: 10.1042/ETLS20190125. (PubMed Central (PMC))

[4] Mondet, F., Beaurepaire, A., McAfee, A., Locke, B., Alaux, C., Blanchard, S., Danka, B., & Le Conte, Y. (2020). Honey bee survival mechanisms against the parasite Varroa destructor: a systematic review of phenotypic and genomic research efforts. International Journal for Parasitology, 50, 433–447. DOI: 10.1016/j.ijpara.2020.03.005. (PubMed Central (PMC))

[5] Lapidge, K. L., Oldroyd, B. P., & Spivak, M. (2002). Seven suggestive quantitative trait loci influence hygienic behavior of honey bees. Naturwissenschaften, 89, 565–568. DOI: 10.1007/s00114-002-0371-6.

[6] Oxley, P. R., Spivak, M., & Oldroyd, B. P. (2010). Six quantitative trait loci influence task thresholds for hygienic behaviour in honeybees (Apis mellifera). Molecular Ecology, 19, 1452–1461. DOI: 10.1111/j.1365-294X.2010.04569.x.

[7] Tsuruda, J. M., Harris, J. W., Bourgeois, L., Danka, R. G., & Hunt, G. J. (2012). High-Resolution Linkage Analyses to Identify Genes That Influence Varroa Sensitive Hygiene Behavior in Honey Bees. PLoS ONE, 7(11), e48276. DOI: 10.1371/journal.pone.0048276. (PubMed Central (PMC))

[8] Mondet, F., Alaux, C., Severac, D., Rohmer, M., Mercer, A. R., & Le Conte, Y. (2015). Antennae hold a key to Varroa-sensitive hygiene behaviour in honey bees. Scientific Reports, 5, 10454. DOI: 10.1038/srep10454. (PubMed)

[9] Guichard, M., et al. (2022). Two quantitative trait loci are associated with recapping of Varroa destructor-infested brood cells in Apis mellifera mellifera. Animal Genetics, 53, 156–160. DOI: 10.1111/age.13150. Genome-wide association study of 155 colonies. (PubMed)

[10] Harbo, J. R., & Harris, J. W. (2005). Suppressed mite reproduction explained by the behaviour of adult bees. Journal of Apicultural Research, 44(1), 21–23. DOI: 10.1080/00218839.2005.11101141. (ARS USDA)

[11] Facchini, E., Bijma, P., Pagnacco, G., Rizzi, R., & Brascamp, E. W. (2019). Hygienic behaviour in honeybees: a comparison of two recording methods and estimation of genetic parameters. Apidologie, 50, 163–172. DOI: 10.1007/s13592-018-0627-6. (Springer Nature)

[12] Kefuss, J., Vanpoucke, J., Bolt, M., & Kefuss, C. (2016). Selection for resistance to Varroa destructor under commercial beekeeping conditions. Journal of Apicultural Research, 54(5). DOI: 10.1080/00218839.2016.1160709. (Taylor & Francis Online)

[13] Gerdts, J., Dewar, R. L., Simone-Finstrom, M., Edwards, T., & Angove, M. (2018). Hygienic behaviour selection via freeze-killed honey bee brood not associated with chalkbrood resistance in eastern Australia. PLoS ONE. Study of 649 colonies. (PubMed Central (PMC))

[14] Spivak, M., & Danka, R. G. (2021). Perspectives on hygienic behavior in Apis mellifera and other social insects. Apidologie, 52, 1–16. (Springer Nature)

[15] Mondet, F., Parejo, M., Meixner, M. D., et al. (2020). Evaluation of Suppressed Mite Reproduction (SMR) Reveals Potential for Varroa Resistance in European Honey Bees (Apis mellifera L.). Insects, 11(9), 595. (PubMed Central (PMC))

[16] Bigio, G., Al Toufailia, H., & Ratnieks, F. L. W. (2014). Honey bee hygienic behaviour does not incur a cost via removal of healthy brood. Journal of Evolutionary Biology, 27, 226–230. DOI: 10.1111/jeb.12288.

Author: Davide Simeone

Published: 20 August 2026