What Wisconsin and Aberdeen taught us about wintering bees
The November questions
We are into winter, and as I write, the temperature is a chilly 2 °C. The big question, as always, is “when should I treat my colonies with oxalic acid?” The answer is, rather unhelpfully, “When they have no sealed brood.” But when is that, exactly? I have been digging into some work done many years ago, and comparing it to more recent research, to see if there is a satisfactory answer. It’s not just the broodless or not issue, but also what do we know about how, and why honey bees navigate winter successfully? Obviously, if I weren’t such a lazy bugger, I could actually do some field research of my own. Theoretically, I could pull brood frames out of a selection of colonies every week through the winter and record what I see, but I don’t want to. I’m certain my bees don’t want me to.
One thing I do know from experience is that small colonies in big boxes tend to die over winter. I believe that a decision must be made in the autumn, around September time for me, on the small colonies that I have. They either get shaken out or put into a nucleus box so that they are nice and cosy. People talk about combining with another colony, but I tend not to do that. I would rather write them off or take them through winter in a poly nuc, after feeding and treating. It all depends on why the colony is small. If they have a reasonable excuse, they go into a nuc. Otherwise, they are for the chop.
After digging about for a bit, I found a story with three main characters, two of whom almost nobody has heard of. The first is an American government scientist in Wisconsin who told the world to winter big colonies. The other two worked in Aberdeen, a generation apart. Between them, they quietly took the American advice apart, and offered an explanation of the winter cluster that most beekeepers still haven’t caught up with.
None of them ever had to deal with varroa, which is something to keep firmly in mind.
Wisconsin: the case for big colonies
Clayton Farrar was an apiculturist with the US Department of Agriculture, first in Wyoming and then in Madison, Wisconsin. From the 1930s to the 1960s he did as much as anyone to turn colony management into something backed up by good observation and a scientific approach.
His best-known finding was simple. Big colonies make more honey per bee than small ones. In a small colony, a large share of the bees is tied up looking after brood; in a big colony proportionally more are free to forage. A later USDA study at Madison repeated the work and agreed. (You may have seen this as “Farrar’s rule”, the idea that doubling the bees quadruples the honey. That neat square law is a later simplification, not his equation. Big colonies are more efficient, but how much more depends on the season.)
Farrar carried the same logic into winter. His 1944 bulletin for northern states advised going into winter with eight to ten pounds of bees that had emerged after 20 August. That is somewhere around 30,000 to 38,000 bees. By contrast, mine are normally closer to 18,000 bees in November, but Cheshire isn’t Wisconsin.

He also showed that the colonies coming out of winter strongest were the ones with the most pollen stored the previous autumn. That is why I am always happy to see my bees bringing in pollen from balsam and ivy as the season comes to a close. In January 1939 Farrar opened Wisconsin colonies that had been given plenty of autumn pollen and found up to three frames of brood in all stages. Midwinter brood was not news to him.
Two more results from the Madison laboratory matter for this story. A later USDA bulletin reports that colonies whose queens were caged in August, to stop autumn brood rearing, died in droves over winter because they had no young bees to carry them through. And after five winters testing packed, heated and unprotected hives at Madison, he and an engineering colleague concluded that colonies in good condition wintered perfectly well without packing or heating.
So the Wisconsin recipe was: big colonies, young bees, plenty of pollen and honey, and no need to wrap them up. It became the orthodoxy, summed up in an old line that Bernhard Möbus later quoted with some scepticism: the best packing for bees is more bees.
It is probably a good recipe. For Wisconsin.
Aberdeen, part one: Jeffree counts the bees
In the late 1940s and 1950s, at the Bee Research Department of Marischal College, Aberdeen, E.P. Jeffree and his colleague M. Delia Allen did something unglamorous and extremely valuable. They counted bees and brood. Between 1945 and 1953 they made 430 records of how many bees, how much brood and how much pollen were in the college colonies, month by month.
Some of what they found agreed with Farrar’s findings. Pollen and colony size both drove brood rearing, each in its own right. Aberdeen colonies followed much the same annual brood curve as colonies near Washington DC.
Then came the awkward bit. Their 1956 paper on winter losses opens by noting that most authors assumed bigger colonies wintered better, and with good reason in theory. A cluster loses heat from its surface, and surface area grows more slowly than numbers: eight times the bees gives only four times the surface. On paper, the bigger the colony, the easier the winter.
They followed 153 colonies wintered near Aberdeen between 1944 and 1954, without packing, mostly in National hives. They estimated colony strength in late autumn and again in spring, by weighing the bees or comparing combs against calibrated photographs. Some colonies stayed healthy; others caught Nosema (Nosema apis, in those days).
The data disagreed with the theory of “bigger is always better”. Losses in colonies of “supernormal size” were just as severe as in tiny ones. There was an optimum wintering size, and it was well within ordinary apiary practice levels. Their figures, as reported by Randy Oliver, put it at around 13,000 bees in October for Nosema-free colonies and about 16,000 for infected ones. Farrar’s recommendation was two to three times that.
Jeffree also went into colonies in winter, which was not the done thing. According to Möbus, he examined about 360 colonies over ten years, opening them only on days warm enough for the bees to fly. He found patches of brood at any time of winter, getting bigger and more frequent towards spring, and he couldn’t tie them to the weather. In a couple of colonies he opened repeatedly, the brood seemed to stop and start.
That was a hint, but it took the next character in the story to follow it up.
Aberdeen, part two: Möbus opens the winter cluster
Bernhard Möbus held the National Diploma in Beekeeping (NDB) and became the beekeeping adviser at the North of Scotland College of Agriculture in Aberdeen in the 1970s, based at the Craibstone apiary. Scottish beekeepers remember him for championing native bees, for the Maud strain he found in an old beekeeper’s hives in Aberdeenshire, and for a little book on mating in mini-nucs. What interests me here is that he couldn’t leave a puzzle alone.
The puzzle had started years earlier in Lincolnshire, where he found brood in his own hives one winter. Brood in winter costs honey and risks starvation. Why would a colony do it? He later admitted the question became “a veritable compulsion”. In 1998, by then living in the French Pyrenees, he wrote it all up in two articles for the American Bee Journal.
Two winters with the lid off
In 1974/75 and 1975/76 Möbus and his staff opened nearly every colony in the Craibstone apiary every three or four weeks, whenever the weather allowed. They measured every patch of brood on every comb face and recorded eggs, larvae and sealed brood separately.
Here is what they found:
- Some colonies had brood in every winter month, November and December included.
- More colonies had brood, and the patches were bigger, as spring approached.
- Laying was on/off. Queens stopped, then started again, sometimes two or three times in a winter. These “brood stops” became more frequent towards spring, not less.
- Over a winter, colonies reared anything from about 600 to nearly 11,000 bees.
In other words, there was no reliable broodless period. And Möbus made a point that still matters to anyone with a hive monitor: temperature sensors can’t see any of this. A brood nest stays at around 35 °C for as long as there is living brood in it, and a batch of brood takes three weeks or more to emerge. So a queen could stop laying for a while, and start again, without the temperature in the cluster ever showing it. Only looking at the comb revealed the queen switching on and off.

to download two article in the American Bee Journal by Mobus, see Resources
Every baby costs its mother a tooth
Möbus quoted that old saying, and his numbers backed it. Across 33 colonies over the two winters, the amount of brood reared and the amount of food eaten tracked each other almost perfectly (a correlation of 0.92). His regression put the basic cost of wintering a broodless cluster at about 4.2 kg of stores, with brood adding more on top of that. Actual consumption ranged from 3 kg to 9.5 kg. In his words, wintering is “a cluster-specific experience”. That should make us pause before comparing one colony’s winter honey consumption with another’s. The heavy eater may simply be the one rearing the most brood.
Stopping the queen backfired
So why rear brood at all? Möbus tried preventing it. In both winters, he shut some queens in small cages made of queen-excluder material, placed in the middle of the cluster. Workers could squeeze through the slots to feed and tend the queen, but she had nowhere to lay. Most of those colonies developed dysentery within three or four weeks, grew steadily weaker, and hardly any recovered even when the queen was released in spring. In the one colony that did recover, brood rearing restarted and the dysentery stopped. His conclusion was that sometimes not rearing brood is more stressful than rearing it.
The super colonies
Then he tested the big-colony advice directly. Having read Jeffree and Allen, he united four strong colonies into two giants. Neither reared any brood. Their bees flew frequently on mild days to collect water. After a cold spell both colonies “suffered massive, disastrous losses”, with chilled bees littering the hive sides and the grass. No acarine, no nosema. The bees he sampled were dehydrated: guts almost empty and low body weights. Two colonies are a tiny sample. But it is a striking result, and it points the same way as Jeffree’s conclusion about an optimal size, rather than bigger is always better.
The hive on the scales
In the winter of 1976/77 he put a single colony on self-recording scales and alternated between encasing it in polystyrene and leaving it “in the cold”. Insulated, it used about 1.5 mg of honey per bee per day. Uninsulated, it was about 3.4 mg. Insulation more than halved consumption.
But whenever the weather allowed a cleansing flight, the insulated colony lost thousands of bees, an estimated 5,500 in all, chilled on the ground. Other colonies flew and came home. Möbus’s reading was that the hive was simply too warm for a mild, damp maritime winter, and that thirsty bees were being driven out to their deaths.
A water theory of the cluster
To explain all this, Möbus leaned on laboratory work by other researchers on caged bees at different temperatures. Bees in the warm centre of a cluster burn little honey and lose more water than they make, so they get thirsty. Bees in the cold outer shell burn more honey to stay alive and build up surplus water. In a colony of the right size the two balance out. Too small, and a larger share of the bees sit in the cold outer shell, building up more water than they can get rid of. Möbus linked this to dysentery. Too big or too warm, and thirst drives bees out of the hive to die.

He thought bees shuffle between core and shell to manage their own water, not out of altruism. It makes sense, but it rests on other people’s lab data, not something he measured himself. I’ve seen it said that he linked winter brood rearing directly to water balance, with brood food acting as a sink for surplus water. That may be the case, but I couldn’t find it in the two articles I have.
Why Wisconsin and Aberdeen disagreed
So who was right? Probably both, for their own areas. I have seen the enormous colony size of Mike Palmer’s Vermont-based bees as he gets them ready for winter, and that seems to work for him. A Wisconsin winter is a long, hard, continental freeze. Colonies are shut in for months, and the bigger the cluster, the less honey each bee needs to stay warm. There is little risk of a big cluster overheating at minus 20.
An Aberdeen winter is cold, wet, and changeable, with mild spells every few weeks. A giant cluster in those conditions can run too warm, get thirsty and fly when it shouldn’t. Randy Oliver, who has written more sensibly about all this than anyone, reconciles the two camps in exactly this way.
There is a third data point from England. In the 1960s, John Free at Rothamsted looked at the same question. According to Randy Oliver, Free found that a bigger share of bees survived the winter in bigger colonies, but that the benefit levelled off at around 18,000 bees. That sits neatly between Jeffree’s 13,000 to 16,000 and Farrar’s 30,000 to 38,000. It also happens to land exactly where I do for my Cheshire bees. Perhaps I’m a beekeeping genius? Anyway, climate is the likeliest explanation of these differences.
As far as I know, nobody has ever run Farrar’s and Möbus’s methods side by side in the two climates. Cheshire is much closer to Aberdeen than to Madison, and what I do seems to work pretty well for me.
What has happened since
As far as I can find, nobody has repeated Möbus’s caged-queen experiment or tested his water theory on real clusters. Research has moved elsewhere. But some of it bears directly on his story.
Winter brood is real, and unpredictable
In 1985, Tom Seeley and Kirk Visscher in New York delayed winter brood rearing by confining queens. By 1 May, colonies that had reared brood in midwinter had more than four times the population of those that hadn’t. Their explanation was that winter brood gave bees a head start for spring. Möbus’s was the water usage theory. The two are not mutually exclusive.
In 2002, the National Bee Unit (NBU) monitored colonies through the winter at sites from Dorset to North Yorkshire because they believed that the exotic pest Tropilaelaps needs brood to survive. They found short broodless spells at unpredictable times, every colony different, and no point at which all of them were broodless. That is basically what Möbus found in Aberdeen, repeated across England, 25 years later.
What about the timing? Recent studies have looked at when colonies restart brood rearing in late winter after a broodless spell. A 2018 German study in climate chambers pointed to rising temperature; a 2025 study of 95 colonies across eight European countries found lengthening days mattered even more. That fits the seasonal trend Möbus saw, with more colonies rearing brood, and bigger patches, as spring approached. It does not explain the rest of what he found: brood in November and December, and queens stopping and starting while the days were still getting shorter and colder. As far as I can tell, nobody has fully explained those yet, though the water regulation idea seems persuasive.
Size still matters, up to a point
Later work broadly supports “optimum, not maximum”. In Louisiana in the 1980s, John Harbo found small colonies used noticeably more honey per bee. A 2019 Pennsylvania study found that pre-winter colony size and weight predicted survival far better than where the bees came from. Germany’s huge bee monitoring project found that weak colonies in autumn were one of four main risk factors for winter loss, alongside varroa, viruses and old queens. Nobody, though, seems to have tested whether there is an upper limit, as Möbus claimed.
Warm autumns wear bees out
There is a newer worry Möbus would have recognised. When pollen stops coming in during autumn, colonies wind down brood rearing and the long-lived winter bees take over. That still happens after a long, warm autumn; forage runs out eventually. The problem comes next. Bees will fly whenever the weather allows, whether or not there is anything worth collecting, and flying shortens a bee’s life. A 2024 modelling study in the Pacific Northwest suggested that extended warm spells in autumn and winter use up the lifespan winter bees need, leaving an older, more worn-out cluster that fails in spring. It is a model, not field data.
Varroa changed everything
And then along came the varroa mite. Winter brood means winter mites: Möbus himself pointed this out in 1998. Dutch research in 2012 showed that the bees reared in late summer and autumn become the long-lived winter bees, and that controlling varroa before and during that transition gave longer-lived bees and better colony survival. In their trial, July treatment worked best.
Every colony Farrar, Jeffree and Möbus studied was varroa-free. Their figures on optimum size and winter consumption come from a different universe. The principles may hold, but the numbers may not.
The insulation argument: physics versus colonies
If you want a fight at a beekeeping association meeting, mention poly hives. Actually, if you really want fisticuffs, saying that you use imported queens and treat for varroa three times per year would probably also do the trick! Both camps (pro and anti poly) might quote Möbus, and both would be only half right.
The most influential voice on the pro-insulation side is Derek Mitchell, a physicist and engineer who did a PhD at Leeds on the thermofluid engineering of the honey bee nest. His work shows, convincingly, that standard thin-walled hives lose heat several times faster than the tree cavities bees may choose for themselves. He goes further, arguing from a computer model that the outer layer of a winter cluster is not insulation at all but a heat sink, that clustering is a stressful last resort, and that forcing bees to do it may be a welfare issue. Möbus would have agreed the outer shell isn’t insulation. Where I part company is the leap from a heat-flow model to a welfare verdict.
Mitchell’s key hive measurements used empty hives with electric heaters and foam models of trees. The cluster research paper is a model built on other people’s data. The one study I found with live colonies measured the gaps around combs in imitation tree nests, not survival or honey use. None of it tests whether bees in insulated hives actually survive better.
So what happens when people put real colonies in insulated and uninsulated hives and compare?
- Honey use: fairly consistently lower with insulation. Möbus’s scale hive halved its consumption. Norwegian work in the 1970s is reported to have found that thick walls cut winter store use from about 12 kg to 8.5 kg. A 2022 randomised trial in Illinois found covered hives used less food.
- Survival: much less clear. In that Illinois trial, 1 of 22 covered colonies died against 6 of 21 uncovered ones; impressive, but a small, single-winter study that didn’t record why colonies died. A Massachusetts trial found 20% losses whether hives were wrapped or not. Germany’s monitoring of over 1,000 colonies found no difference between wooden and polystyrene hives, which is also what I have found (I have plenty of both types). And two Turkish studies, in a hard continental climate, found colonies survived better in wood than in polystyrene.
- The cautionary tale: Möbus’s insulated colony saved honey and lost thousands of bees on cleansing flights.
Where does that leave us? Insulation saves stores. Whether it saves colonies, in a British winter, I’m not so sure. However, when we are talking about small colonies, in my experience, poly wins for nucs. A poly hive is also not simply a wooden hive with better walls; ventilation, damp, floors and fit all differ, and any of those could matter more than the R-value.
Lessons for beekeepers
None of this adds up to a recipe. A colony is a system: bees, queen, brood, stores, pollen, mites, viruses, hive, weather and beekeeper, all working for or against colony survival. Change one and the others may change. With that warning, here is what I’m taking from ninety years of research spread across Wisconsin, Aberdeen and beyond.
- Don’t assume your colonies are broodless in winter. Aberdeen in the 1970s and England in 2002 both found no reliable broodless window, and every colony was different. If you rely on an oxalic trickle when there is “no brood”, look first: cappings on the varroa tray are a good clue that brood is emerging. I can’t do that, as I don’t have trays and many of my floors are solid. My winter oxalic acid trickle treatment will be less effective on colonies with sealed brood, which is why a spring treatment has become part of my anti-varroa armoury of late. I suppose one could theoretically do three rounds of oxalic treatments by sublimation, spaced maybe five days apart. Repeated sublimation is used abroad, but the UK label allows one treatment a year, so that’s not an option for me. It would bring costs down if I didn’t need to use strips in April.
- Big enough beats as big as possible. A healthy, decent-sized colony is the goal. Uniting a genuinely weak colony in autumn might make sense, but see my points earlier on (I tend to nuc or shake out). Making a monster may not be wise, especially in a mild, damp winter.
- Autumn decides winter. Young bees, raised on good pollen, protected from varroa while they are being reared, are what carry a colony through. Farrar knew it in the 1940s; the Dutch varroa work confirms it. In practice that means getting mites down in late summer, before the winter bees are reared, not after. Many people still leave it far too late.
- Don’t try to stop the queen. Caging queens in autumn killed Wisconsin colonies; caging them in winter gave Aberdeen colonies dysentery. Colonies seem to manage their own brood for good reasons, even if we don’t fully understand them.
- A hungry colony may be a busy one. Winter consumption tracks brood rearing. A colony that is getting light fast in a mild winter may simply be rearing a lot of brood. Heft hives through late winter and early spring, when brood rearing is taking off and stores fall fastest. The light ones need fondant, and probably lots of it (a 2.5 kg block is likely to be a better move than a silly little piece).
- Insulation saves honey; it is not proven to save colonies. It may help production colonies, and I believe it definitely helps nucs. An insulated box can’t fix a colony that is too small, too old or too full of mites.
- Watch the warm autumns. Bees flying late into a mild autumn are wearing themselves out. I like the winters to be cold and the bees to be clustered for much of the time, as warmer winters seem to cause problems.
- Remember the mite. Winter brood means winter mites. The old figures on optimum size and stores come from a varroa-free world. Use the principles; treat the numbers with care.
What we still don’t know
Quite a lot, as it turns out.
- Why colonies switch brood rearing on and off in winter, and whether Möbus was right that water is part of the answer.
- Whether there really is an upper limit to useful colony size in a British winter, and where it lies today, with varroa.
- Whether insulated or poly hives improve survival here, rather than saving stores.
The frustrating thing is that Möbus’s questions could be tested by any bee farmer with a few hive scales, a notebook, and the patience to open colonies on mild winter days. I suspect that they are all either having a well-earned break (I’m off to Lanzarote after Christmas) or are busy with equipment maintenance and building in preparation for the season ahead. What strikes me most is how much careful work was done by a handful of people and then largely forgotten. Jeffree and Möbus deserve to be better known. So, for that matter, does the habit they shared: going and looking, rather than repeating what the books say.

Sources
A note on sources: I have read the Möbus articles, the first page of Jeffree and Allen’s paper, and the abstract of most of the modern studies. A few older papers I know only through other people’s accounts; they are listed separately, with whose account I relied on. I must confess to being helped by an AI model (Claude) with my research, but I’ve checked key claims against the originals where I could get them. Hopefully, that means this article is more useful than if I’d done it all myself, which would have meant fewer research papers looked at.
Primary sources consulted
- Allen, M.D. & Jeffree, E.P. (1956) The influence of stored pollen and of colony size on the brood rearing of honeybees. Annals of Applied Biology 44: 649-656. (Abstract.)
- Dodoloğlu, A., Dülger, C. & Genç, F. (2004) Colony condition and bee behaviour in honey bees housed in wooden or polystyrene hives. Journal of Apicultural Research 43: 3-8. (Abstract.)
- Döke, M.A. et al. (2019) Colony size, rather than geographic origin of stocks, predicts overwintering success. Journal of Economic Entomology 112: 525-533. (Abstract.)
- Farrar, C.L. (1937) The influence of colony populations on honey production. Journal of Agricultural Research 54: 945-954.
- Genersch, E. et al. (2010) The German bee monitoring project. Apidologie 41: 332-352.
- Harbo, J.R. (1986) Effect of population size on brood production, worker survival and honey gain in colonies of honeybees. Journal of Apicultural Research 25: 22-29. (Abstract.)
- Jeffree, E.P. & Allen, M.D. (1956) The influence of colony size and of Nosema disease on the rate of population loss in honey bee colonies in winter. Journal of Economic Entomology 49: 831-834. (First page.)
- Mitchell, D. (2016) Ratios of colony mass to thermal conductance of tree and man-made nest enclosures of Apis mellifera. International Journal of Biometeorology 60: 629-638.
- Mitchell, D. (2023) Honeybee cluster: not insulation but stressful heat sink. Journal of the Royal Society Interface 20: 20230488.
- Möbus, B. (1998) Brood rearing in the winter cluster. American Bee Journal 138(7): 511-514.
- Möbus, B. (1998) Rethinking our ideas about the winter cluster, Part II. American Bee Journal 138(8): 587-591.
- Nürnberger, F., Härtel, S. & Steffan-Dewenter, I. (2018) The influence of temperature and photoperiod on the timing of brood onset. PeerJ 6: e4801.
- Owens, C.D. & Farrar, C.L. (1967) Electric heating of honey bee hives. USDA Technical Bulletin 1377.
- Parrott, J. (2006) Effects of hive wrapping strategies on honey bee survival in western Massachusetts. SARE report FNE05-560.
- Rajagopalan, K. et al. (2024) Warmer autumns and winters could reduce honey bee overwintering survival. Scientific Reports 14: 5410.
- St Clair, A.L., Beach, N.J. & Dolezal, A.G. (2022) Honey bee hive covers reduce food consumption and colony mortality during overwintering. PLOS ONE 17: e0266219.
- Ulgezen, Z.N., van Dooremalen, C. & van Langevelde, F. (2025) Photoperiod and temperature as seasonal cues for the initiation of brood rearing in honeybees. Ecology and Evolution 15: e72066.
- van Dooremalen, C. et al. (2012) Winter survival of individual honey bees and honey bee colonies depends on level of Varroa destructor infestation. PLOS ONE 7: e36285.
Known only through secondary accounts
- Brown, M.A., Thompson, H.M. & Bew, M.H. (2002) Risks to UK beekeeping from the parasitic mite Tropilaelaps clareae and the small hive beetle. Bee World 83(4). Via the Irish Beekeeping summary of winter brood studies.
- Farrar, C.L. (1936) Influence of pollen reserves on the surviving populations of over-wintered colonies. American Bee Journal 76: 452-454. Via Haydak & Tanquary (1943) and contemporary Wisconsin reports.
- Farrar, C.L. (1944) Productive management of honeybee colonies in the northern states. USDA Circular 702. Via Randy Oliver.
- Free, J.B. & Racey, P.A. (1968) The effect of the size of honeybee colonies on food consumption, brood rearing and the longevity of the bees during winter. Entomologia Experimentalis et Applicata 11: 241-249. Via Randy Oliver.
- Jeffree, E.P. (1956) Winter brood and pollen in honey bee colonies. Insectes Sociaux 3: 417-422. Via Möbus (1998).
- Jeffree & Allen (1956), optimum colony sizes of 13,000 and 16,000 bees. Via Randy Oliver.
- Seeley, T.D. & Visscher, P.K. (1985) Survival of honeybees in cold climates: the critical timing of colony growth and reproduction. Ecological Entomology 10: 81-88. Via Randy Oliver.
- USDA bulletin on overwintering honey bee colonies (author and date unconfirmed), source of the August queen-caging results.
- Villumstad, E. (1974) Importance of hive insulation for wintering, development and honey yield in Norway. Apiacta 9: 277-281. Via a secondary summary.
Secondary sources
- Oliver, R. Understanding Colony Buildup and Decline, Part 13b, and Part 3. Scientific Beekeeping.
The book Healthy Bees, Heavy Hives by Steve Donohoe and Paul Horton covers wintering honey bees, among many other things
Discover more from The Walrus and the Honey Bee
Subscribe to get the latest posts sent to your email.