Cut your losses

The stats for winter losses in the UK, Europe and USA can make for rather sobering reading.

In the UK, losses over the last 12 years have fluctuated between 9% and 34%. This self-selecting survey includes responses from about 10% of the British Beekeepers Association membership (primarily England and Wales, despite the name). The average number of hives maintained by a BBKA member is about 5, meaning – all other things being equal 1 – that most beekeepers should expect to lose about 1 hive every winter.

BBKA winter losses survey

About 30 countries, mainly Northern hemisphere, contribute to the COLOSS survey which is significantly larger scale. The most recent 2 data published (for the ’16/’17 winter) had data from ~15,000 respondents 3 managing over 400,000 hives. Of these, ~21% were lost for a variety of reasons. COLOSS data is presented as an unwieldy table, rather than graphically. Further details, including recently published results, are linked from their website.

In the USA the Bee Informed Partnership surveys losses – both winter and summer – and claims to have results that cover ~10% of all the colonies in the country (so probably between 250,000 and 275,000 hives). Winter losses in the USA are rarely reported at less than 20% and were as high as 35% in the ’18/’19 winter 4.

Bee Informed Partnership annual colony losses

Are these figures to be trusted?

Who knows?

Each survey is accompanied by a variety of statistics. However, since they all appear to be based upon voluntary reporting by a subset of beekeepers, there are opportunities for all sorts of data to be included (and even more to be missed entirely). 

The problem with surveys

Is the successful beekeeper who managed to get all her colonies through the winter more likely to respond?

A form of ‘bragging rights’.

What about the beekeeper that lost all his colonies?

Does he respond out of a sense of responsibility?

Or does he keep quiet because he doesn’t want to be reminded of those cold, quiet, mouldy boxes opened on the first warm day of spring?

One and two year beekeepers

What about the high level of annual ‘churn’ amongst beekeepers? They buy a nuc in May, filled with enthusiasm about the jars of golden honey they’ll have for family and friends in late summer.

To say nothing of all the “saving the bees” they’ll be doing.

But by late summer the colony is queenless and has an unpleasant temperament

Beekeeping should be enjoyable ...

Beekeeping should be enjoyable …

Psychopathic you might say … if you were feeling uncharitable.

Consequently the Varroa treatment goes on far too late,. Or is quietly forgotten. The winter bees have high viral loads and ‘die like flies’ 5, resulting in the colony succumbing by the year end.

But this colony loss is never recorded on any surveys.

The once enthusiastic beekeeper has moved on and is now passionate about growing prize-winning vegetables or cheesemaking or keeping chickens. 

Beekeeping associations train lots of new beekeepers and – although membership numbers are increasing – it’s well below the rate they’re trained at.

Some may not be ‘joiners’ and go their own way.

Many just quietly stop after a year or two.

How many people have you met that say “Oh yes, I used to keep bees”

Did you ask them whether they ever completed a winter losses survey?

I’m not sure any of the surveys listed above do much ‘groundtruthing’ to establish whether the data they collect is truly representative of the population actually surveyed. With large numbers of respondents spread across a wide geographic and climatic range it’s not an easy thing to do.

So, treat these surveys with a healthy degree of scepticism.

Undoubtedly there are high levels of winter losses – at least sometimes – and the overall level of losses varies from year to year.

Losses and costs

The direct financial cost of these colony losses to beekeepers is very high.

Ignoring time invested and ‘consumables’ like food, miticides and foundation these costs in ’16/’17 for just Austria, the Czech Republic and Macedonia were estimated at €56 million 😯  

These figures simply reflect lost honey production and the value of the lost colonies. They do not include the indirect costs resulting from lost pollination.

But, for the small scale beekeeper, these economic losses are irrelevant.

Most of these beekeepers do not rely on bees for their income.

The real cost is emotional 🙁

It still saddens me when I lose a colony, particularly when I think that the loss was avoidable or due to my incompetence, carelessness or stupidity 6.

Little snow, big snow. Big snow, little snow.

Your hives should be quiet in winter, but it hurts when they are silent in spring.

Anatomy of a death

The COLOSS surveys give a breakdown of winter losses in three categories:

  • natural disasters
  • queen problems
  • dead colonies

Natural disasters are things like bears, honey badgers, flooding or falling trees.

We can probably safely ignore honey badgers in the UK, but climate change is increasing the weather extremes that causes flooding and falling trees.

Moving to higher ground ...

Moving to higher ground …

Don’t assume that poly hives are the answer to potential flooding.

They do float, though not necessarily the right way up 🙁

Queen problems cover a variety of issues ranging from reduced fecundity to poor mating (and consequent drone laying) to very early or late – and failed – supersedure 7.

Beekeepers with a lot more experience than me report that queen problems are increasing.

Drone laying queen ...

Drone laying queen …

Perhaps the issues with fecundity and drone laying are related to toxic levels of miticides in commercial foundation? It’s certainly known that these residues reduce drone sperm fertility significantly. I intend to return to this topic sometime during the approaching winter … perhaps in time to encourage the use of some foundationless frames for (fertile) drone production 😉

In the ’16/’17 COLOSS data, natural disasters accounted for 1.6% of all overwintered colonies (so ~7.5% of losses), queen problems resulted in the loss of 5.1% of colonies (i.e. ~24% of losses) and the remainder (14.1% of colonies, ~68% of losses) just died.

Just died?

We’ll return to natural disasters (but not bears or honey badgers) and queen problems shortly. What about the majority of losses in which the colony ‘just died’?

If you discuss colony post-mortems with beekeepers they sometimes divide the ‘just died’ category (i.e. those not readily attributable to failed queens, marauding grizzlies or tsunamis) into four groups:

  • disease
  • isolation starvation
  • starvation
  • don’t know 

The most important disease associated with overwintering colony losses is high levels of Deformed wing virus (DWV). This results from uncontrolled or inadequately controlled Varroa infestation. For any new readers of this site, please refer back to many of the articles I’ve already written on Varroa management 8.

I strongly suspect that a significant proportion of the reported isolation starvation is actually also due to disease, rather than isolation per se.

A consequence of high levels of DWV is that winter bees die prematurely. Consequently, the colony shrinks faster than it would otherwise do. It starts the size of a basketball but (too) rapidly ends up the size of a grapefruit … or an orange.

Isolation starvation and disease

The small cluster is then unable to remain in contact with stores, and so starves. 

Yes, the colony died from ‘isolation starvation’, but the cause was the high levels of Varroa and the viruses it transmits.

Isolation starvation ...

Isolation starvation …

What about regular starvation?

Not because the cluster became isolated from the stores, but simply because they had insufficient stores to get through the winter.

Whose fault was that?

And the last category, the “don’t knows”?

I bet most of these are due to high levels of Varroa and DWV as well 🙁

Yes, there will be other reasons … but probably not a huge number. 

What’s more … if you don’t know the reason for the colony loss there’s very little you can do to mitigate against it in future seasons.

And, other than wild and increasingly vague speculation, there’s little I can write about if the reason for the loss remains unknown 9.

Avoiding winter losses

So, let’s rationalise those earlier lists into the probable (known) major causes of overwintering colony losses:

  • natural disasters
  • queen problems
  • starvation
  • disease (but probably mainly DWV and Varroa

As the long, hot days of summer gradually shorten and cool as early autumn approaches, you should be thinking about each of these potential causes of overwintering colony loss … and doing what you can to ensure it doesn’t happen to you (or, more correctly, your bees).

Ardnamurchan autumn

Ardnamurchan autumn

Some are easier to deal with than others.

Here’s a whistle-stop tour of some more specific problems and some practical solutions 10. Some, all or none may apply to your bees – it depends upon your location, your climate, your experience and future plans as a beekeeper. 

Natural disasters

These fall into two broad groups:

  • things you can do almost nothing about (but might be able to avoid)
  • things you can relatively easily solve

Flooding, falling trees, lightning, landslides, earthquakes, volcanoes, meteor strikes etc. all fall into the first group.

If you can avoid them, do. 

Your local council will have information on areas at risk from flooding. There are also searchable maps available from SEPA. Do not underestimate the severity of some of the recent flooding. Some parts of Scotland and Northern England had 600 mm of rain in two days in 2015.

You might be surprised (and from an insurance aspect, devastated) at the classification of some areas now ‘at risk’. 

Where did Noah keep his bees? In his Ark hive.

Where did Noah keep his bees? In his Ark hive.

Consider moving hives to higher ground before the winter rains start. One consequence of climate change is that heavy rainfall is now ~20% heavier than it was a few decades ago. This means that floods occur more frequently, are more extensive and the water levels rise faster. You might not have a chance to move the hives if flooding does occur,

More rain and stronger winds (particularly before leaf fall) mean more trees will come down. You might be able to identify trees potentially at risk from falling. It makes sense to remove them (or site your hives elsewhere). 

No risk of this larch tree falling on my hives

Lightning, earthquakes, volcanoes, meteor strikes … all a possibility though I would 11 probably worry about Varroa and woodpeckers first 😉

Solvable natural disasters

The ‘solvable’ natural disasters include preventing your colonies being robbed by other bees or wasps. Or ransacked by mice or woodpeckers after the first hard frosts start. A solution to many of these are ‘reduced size entrances’ which either enable the colony to better defend itself, or physically restricts access to critters.

The L-shaped ‘kewl floors‘ I use prevent mice from accessing the brood box. They are also easier for the colony to defend from bees/wasps, but can also easily be reduced in size with a narrow piece of hardwood. If you don’t use these types of floor you should probably use a mouseguard.

Polyhives and polythene

Polyhives and polythene …

Woodpeckers 12 need to cling onto the outside of the hive to hammer their way through the side. You can either place a wire mesh cage around the hive, or wrap the box in something like damp proof membrane (or polythene) to prevent them gaining purchase on the side walls.

Keep off Woody

Keep off Woody

Doing both is probably overkill though 😉

Strong colonies

Before we move onto queen problems – though it is related – it’s worth emphasising that an even better solution to prevent robbing by bees or wasps is to maintain really strong colonies.

Strong colonies with a well balanced population of bees can almost always defend themselves successfully against wasps and robbing bees.

Nucs, that are both weaker and – at least shortly after being made up – unbalanced, are far less able to defend themselves and need some sort of access restriction.

By ‘balanced’ I mean that the numbers and proportions of bees fulfilling the various roles in the nucleus colony are reflective of a full hive e.g. nurse bees, foragers, guard bees. 

Reduced entrance ...

Reduced entrance …

But the benefits of strong colonies are far greater than just being able to prevent wasps or robbing bees. There is compelling scientific evidence that strong colonies overwinter better

I don’t mean strong summer colonies, I mean colonies that are strong in the late autumn when they are fully populated with the winter bees.

Almost the entire complement of bees in the hive are replaced between late summer and late autumn. Remember that a really strong summer colony may not be strong in the winter if Varroa and virus levels have not been controlled.

How do you ensure your colonies are strong?

  1. Minimise disease by controlling Varroa levels in early autumn to guarantee the all-important winter bees are reared without being exposed to high levels of DWV.
  2. Try and use a miticide treatment that does not reduce the laying rate of the queen.
  3. Avoid blocking the brood nest with stores where the queen should be laying eggs.
  4. Requeen your colonies regularly. Young queens lay more eggs later into the autumn. As a consequence the colonies have increased populations of winter bees.
  5. Unite weak colonies (assuming they are disease-free) with stronger colonies. The former may well not survive anyway, and the latter will have a better chance of surviving if it is even stronger – see below. 
  6. Use local bees. There’s good evidence that local bees (i.e. reared locally, not imported from elsewhere) overwinter better, not least because they produce stronger colonies.

Uniting – take your losses in the autumn

My regular colony inspections every 7-10 days during May and June are pretty much abandoned by July. The risk of swarming is very much reduced after the ‘June gap’ in my experience. 

I still check the colonies periodically and I’m usually still rearing queens. However, the rigour with which I check for queen cells is much reduced. By July my colonies are usually committed to single-mindedly filling the supers with summer nectar.

They are already making their own preparations for the long winter ahead.

Although the inspections are less rigorous, I do keep a careful watch on the strength of each colony. Often this is directly related to the number of supers I’ve had to pile on top.

Colonies that are underperforming, and – more specifically – understrength are almost always united with a stronger colony.

An Abelo/Swienty hybrid hive ...

An Abelo/Swienty hybrid hive …

Experience has taught me that an understrength colony is usually more trouble than it’s worth. If it’s disease-free it may well overwinter reasonably well. However, it’s likely to start brood rearing more slowly and build up less well. It may also need more mollycoddling 13 in the autumn e.g. protection from wasps or robbing bees.

However, a colony that is not flourishing in the summer is much more likely to struggle and fail during the winter. Perhaps the queen is not quite ‘firing on all cylinders’ and laying at a really good rate, or she might be poorly mated.

Far better that the workforce contributes to strengthening another hive, rather than collect an underwhelming amount of honey before entering the winter and eventually becoming a statistic.

My winter losses are low and, over the last decade, reducing.

That’s partly because my Varroa management is reasonably thorough.

However, it’s probably mainly due to ensuring only strong colonies go into the winter in the first place.


I’ve dealt with uniting in several previous posts.

It’s a two minute job. 

You remove the queen from the weak colony, stack one brood box over the other separated by a sheet or two of newspaper with a very small (~3mm) hole in the middle. Add the roof and leave them to get on with things.

I don’t think it makes any difference whether the strong colony goes on the top or the bottom.

I place the colony I’m moving above the box I’m uniting it with. My – wildly unscientific – rationale being that the bees in the top box will have to negotiate the route to the hive entrance and, in doing so, will help them orientate to the new location faster 14.

If you unite colonies early or late in the day most foragers will be ‘at home’ so not too many bees will return to find their hive missing.

If there are supers on one or both hives you can separate them with newspaper as well. Alternatively, use a clearer the day before to empty the supers prior to uniting the colonies. You can then add back the supers you want and redistribute the remainder to other hives in the apiary.

Successful uniting ...

Successful uniting …

Don’t be in too much of a hurry to check for successful uniting.

Leave them a week. The last thing you want is for the queen to get killed in an unseemly melee caused by you disturbing them before they have properly settled.

Done properly, uniting is almost foolproof. I reckon over 95% of colonies I unite are successful.

That’s all folks … more on ‘Cutting your losses’ next week 🙂


At just over 3000 words this post got a bit out of control … I’ll deal with more significant queen problems, feeding colonies, the weather and some miscellaneous ‘odds and sods’ next week.


  1. Which they’re not
  2. At least that I could find.
  3. Actually, far more than this but they do a lot of ‘sanity checking’ on the submissions and chuck out those that are nonsensical e.g. those from people ending the winter with more hives than they started.
  4. Note that the total annual losses are not simply the addition of the summer and winter figures.
  5. If you’ll excuse the entomological misclassification.
  6. I have an abundance of all three.
  7. Perhaps that should be attempted supersedure?
  8. Regular readers are probably sick and tired of the subject. I make no apologies for this as it’s probably the most important part of good beekeeping practice.
  9. Not that that’s ever stopped me in the past.
  10. I’ll deal with feeding colonies, the weather and a few other miscellaneous things next week.
  11. And that’s even though I live quite close to the Ardnamurchan Ring Complex, an ancient volcano (but not active for about 60 million years).
  12. Interestingly it’s only some green woodpeckers that cause problems. It’s a learned activity and many have yet to learn how to do it.
  13. Interesting etymology … moll meaning a girl or prostitute, and ‘coddle’ derived from caldus/m, the Latin for warm/warm drink via caudle (administer invalid’s gruel) to coddle in the mid-16th Century meaning to gently boil fruit.
  14. Though I’ve done it – deliberately or by accident – using every possible permutation of box and location.

16 thoughts on “Cut your losses

  1. Damian Ford

    Hi David! Thanks for another interesting article. What is your yardstick for an acceptably strong colony for overwintering? I personally aim for about seven frames of brood by the end of august. The terms
    ”weak ”and ”strong ”colonies are probably the most common in bee literature. However, it seems that trying to pin any beekeeper down to actually define what a strong or weak colony looks like, and what one should be aiming for, is often as difficult as getting blood from a stone. So I just wondered if you would like a stab at it? 😄

    1. David Post author

      Hi Damian

      Yes, 7-8 frames is probably about right. However, I also try and guesstimate whether the queen is still laying well or if the colony is rapidly shrinking (in terms of brood production). I prefer the former. It’s also worth noting that it’s going to depend upon the strain of bee and the location. Here on the west coast, all my dark, native bees are in single brood boxes now. On the east coast my ‘Heinz 57 varieties’ local mongrels are much stronger, with several going into winter this season on double brood and 12+ frames of brood.

      This type of question is the type that emphasises just how ‘local’ beekeeping is, and reinforces how useful it is to have two or more colonies to make comparisons between.


  2. Paul Evans

    Hi David,
    Interesting your mention of mouse guards. I’ve been tempted to give up on them since watching Autumn Watch last year when they talked about mice coming into your house as the days get colder.
    They experimented with, admittedly only one mouse, getting under a door to get to food. At 11 mm it failed to get under the door. All my entrance blocks have 8 mm high apertures.
    Has any beekeeper lost a strong colony to a mouse over winter? Stored supers are a more likely home for an over wintering mouse.
    Paul Evans from a very temperate Cardiff

    1. David Post author

      Hi Paul

      Not sure I’m qualified to answer the “Has any beekeeper …” question. I’ve lost nucs and I’ve found mice in full colonies in very cold weather. They make a real mess of the frames. Narrow entrances are the solution as you say … mine are 8mm (with the exception of the cavernous gaping maw on Thorne’s Everynuc’s which I seal off with wire mesh or Correx).


  3. Frazer

    Ah yes, the Kewl floor gets another mention. Im in the process of making up another batch of 6 of the GT model. All have rear doors and this time I am using decking ripped up the middle with the grooves facing each other so I can slide a correx varroa monitoring board in opposing grooves….Coffee is for your IP payment 😉

    1. David Post author

      Hello Frazer

      Decking planks are a good idea. I’ve added little strips of softwood to hold my Varroa trays in place. It’s worth making sure you can form a relatively tight seal between the board and the underside of the hive if you vaporise OA. Before I did this it used to billow out in clouds. Now I keep the stuff where it’s going to do some good 😉

      Thanks for the coffee 🙂 Writing 3000 words late last Thursday required a lot of caffeine.


  4. Joe Mc Donald

    I am a new beekeeper and living on the east coast of Ireland, in lowland county Wicklow.I don’t know when to use apivar or what week in Sept or October to help
    Keep down the mites in the winter bees.It’s still reasonably mild here and can you say definitely when it is too late to use the mite treatment. I know I should do it soo and lm afraid I will be too late using it.
    Regards Joe
    County Wicklow

    1. David Post author

      Hello Joe

      Read the post on Rational Varroa control … the timing depends upon your climate and nectar flows. But, as I make clear in that post, it will vary year to year. You can’t make decisions based on the calendar alone. Assuming you don’t have a heather crop, or produce ivy or balsam honey, then I’d treat as soon as the summer supers are removed. October is almost certainly too late.


  5. Archie McLellan

    Thanks for this 3000-words-short post, David. You wrote:
    ‘How do you ensure your colonies are strong?
    1. Minimise disease by controlling Varroa levels in early autumn
    3. Avoid blocking the brood nest with stores where the queen should be laying eggs.’

    Re 1. You’ve told us umpteen times before, but could you briefly tell me again, please, the rationale to treating at this time of year if you have very low varroa counts by sugar roll.

    Re 3. I might read this as a suggestion to feed little and often at this time of year – but I don’t think you mean, or do, that. So how should we try to ‘Avoid blocking the brood nest with stores…’?
    Many thanks!

    1. David Post author

      Hi Archie

      I really ought to write an article on phoretic mite counts and reality … ! I’ll add it to the list.

      1. I guess there are two ways to justify this … mite levels tend to rise very fast at the end of the season. Quite when this occurs depends upon your particular climate and your bees. If you get a low counts but the colony actually continues rearing brood for quite a bit longer then mite levels will inevitably continue to rise. How sure are you that you’ve not sampled too early?
      2. The winter oxalic acid treatment is, let’s say, 90% effective. If you do not treat in late summer you will inevitably be starting the winter treatment with more mites … therefore the 90% reduction will leave more mites. This means the early brood cycles will be made in the presence of more mites. If you start the season with more mites you may have to intervene before summer’s end if mite levels get out of hand.

      95% of my colonies only get treated late summer and ‘midwinter’. By doing this mite levels are always under control and I never need to intervene during the season. If I don’t treat in late summer I don’t think this would be the case. Furthermore, there’s an assumption that treatment is detrimental to the colony (in the absence of the impact on mites). It may be, but I think it’s probably very minor.

      I’m going to write about feeding next week.


  6. Paul Lindstrom

    I have also wondered how statistically solid the results of the Winter losses are, especially since there seems to be no cooperation between what the people at BeeBase do and BBKA. But it’s better than nothing when trying to compare your own losses with the average in your area. What I find to be a good complement is the report on the total number of managed colonies in the UK, by BeeBase. Logically with more and more beekeepers starting out we should see a slow but steady increase in the number of managed colonies (if the new beekeepers are well trained and sign up on BeeBase. But that statistic is relatively new, so we will have to wait a while before we can draw any definite conclusion. And the numbers are based on some kind of algorithm that BeeBase applies to extrapolate the total based on answers given to what the total might be if all the beekeepers had reported the number of colonies they have at the end of the season (it would be interesting to see how that algorithm is setup).

    1. David Post author

      Hi Paul

      For those wondering about these NBU/BeeBase surveys, you can find them here, though the information is rather limited.

      I’ve talked with the NBU about how representative BeeBase is and they clearly know something about how many are registered vs. not registered. Whether this varies across the country is a different matter altogether. The colony density (from the ‘apiaries within 10 km’ statistic for any of your own apiaries) is one of the bits of information I most regularly use. It’s striking how this varies in different locations … and (in my experience) the broad correlation with problems encountered managing Varroa.

      Don’t hold your breath waiting for any publication of the algorithm for hive numbers or anything else from the NBU. All that information is kept very close to their chest.


      1. Paul Lindstrom

        Have have reported on those surveys in our own magazine, and will continue to do so. As with any statistic you have to be aware of the limited accuracy depending on the reply rate, but it’s better than nothing. It would be good if the NBU noted what areas starts to be overcrowded, but that mean we need to agree on how many colonies per square Km or square Mile counts as overcrowded.

        1. David Post author

          Again … don’t hold your breath 😉

          Whether an area is overcrowded would also depend upon the forage available and the density of other pollinators. Not as simple as five hives per square kilometre is too many.

          For reference … when I was in the Midlands I was in an area with ~220 apiaries within 10 km. Let’s assume there were 5 colonies per apiary. The hive density was therefore something like 3.5/km2. My two Fife apiaries are 20 km apart, but both have ~50 apiaries within 10 km, so 0.8 hives/km2. Honey yields in the Midlands were consistently about twice as good as Fife, and one of the Fife apiaries consistently outperforms the other.

          I’ve no idea about pollinator density other than honey bees …



Leave a Reply

Your email address will not be published. Required fields are marked *