Treatment-free vs threshold-based varroa control: what the numbers say
Two beekeepers can keep healthy bees twenty miles apart, one who has never put a strip or a vapouriser in a hive and one who treats every colony that crosses a line. Both will tell you the other is doing it wrong, and both have dead-outs in their history they would rather not discuss. We have friends in each camp and have no interest in picking a side for you. What we can do is set out what the published numbers show, what each approach actually asks of the keeper, and the one thing they turn out to have in common.
The baseline: what losses look like with treatment
Start with the whole population, most of which treats. The Apiary Inspectors of America and Bee Informed Partnership survey, the longest-running national series, reports that over April 2023 to April 2024 US beekeepers lost "55.1% [48.1 - 61.3 CI]" of their colonies, the highest annual figure since 2010-11; winter loss alone was 37.3%, against a sixteen-year average of 28.8% (AIA, 2023-2024 survey). The following year was no better: an estimated "55.6% [47.9-61.8 CI]" annual loss, with winter at 40.2% against a seventeen-year average of 29.3% (AIA, 2024-2025 survey).
Those surveys do not split losses by whether the keeper treated, and the respondents are overwhelmingly treating. So the first honest point is this: the system most of us run is losing somewhere between a third and half of its colonies a year, and "I treat" is not, by itself, a plan that works. Whatever the treatment-free argument gets wrong, it did not invent the problem.
What happens when colonies simply stop being treated
Here the evidence is clearer, and it is not kind to the version of treatment-free that means "do nothing and let nature sort it out."
Gary Brook's 2025 review in the Journal of Apicultural Research gathered every published study of colony loss during the first five years after treatment stops. Its summary: "If colonies were left untreated without any management, other than winter feeding, the majority died after 3 years, although up to 11% of colonies survived 5 years." Stopping treatment while carrying on with ordinary management (swarm capture, splits) did better but still cost more than the controls: "Colony losses were higher (27-28%) if treatment was stopped in colonies without any mitigating strategies".
Why does an untreated colony die? The Honey Bee Health Coalition's guide gives the arithmetic: "uncontrolled mite populations can double monthly", faster with lots of drone brood or mites drifting in from neighbours (Tools for Varroa Management). A colony at 1 per 100 bees in May that is at 2 in June, 4 in July and 8 in August has reared its winter bees beside a mite load that no autumn treatment can undo, which is the mechanism behind every "it looked fine in September" dead-out.
What happens when treatment-free is done on purpose
The same review found something the "they all die" camp tends to skip. Two transition strategies produced losses no worse than treated controls: "starting with known Varroa-resistant bees; or starting with non-resistant bees and then selecting bees showing resistant traits, whilst mitigating losses with biotechnical methods in the first summer, followed by requeening in colonies with high Varroa counts." In both, "winter colony loss was 6-17% and was similar to control colonies (8-23%)", and "Once Varroa treatment-free apiaries were established, they have been shown to survive for 20 years or more." Brook concludes that official advice against treatment-free beekeeping "is misplaced".
Read those two strategies again, because they are the whole argument. One begins with stock already proven against the mite. The other begins with ordinary bees and then does three things: it uses non-chemical controls in year one so the colonies live long enough to be selected, it counts mites, and it requeens the colonies whose counts are high. That is not an absence of management. It is a different and more demanding management, and the colonies with high counts are still removed from the gene pool, by the keeper's hand rather than the mite's.
The feral evidence points the same way. Loftus, Smith and Seeley kept two groups of colonies untreated for two years in New York: small 42-litre hives allowed to swarm freely, and large hives managed for honey with swarming suppressed. By October of the first year the small-hive colonies were at about 2.5 mites per 100 bees and the large ones at 7.4; two years on, 8 of 12 small-hive colonies were alive and 2 of 12 large ones, and 7 of the 12 large-hive colonies had shown deformed-wing virus (PLoS ONE, 2016). The authors' point is that wild colonies persist with varroa because they are small and swarm often, and every swarm is a brood break that resets the mite population. A keeper who wants honey from a two-deep colony has removed both of those defences and has to supply something in their place.
What treatment-free actually requires
Put the review and the feral work together and the requirements are not mysterious. They are just expensive.
Stock that is already resistant, or a willingness to lose a lot of bees selecting for it. Penn State lists the known mechanisms: Russian bees that "inhibit mite reproduction", VSH lines that "can recognize and remove mite-infested pupae", and biters that damage the mites' bodies and legs (Penn State Extension). The Coalition names the lines (VSH, Russian, Hilo, Purdue mite-biters, Minnesota Hygienic, Pol-line) and adds a caveat buyers should hear: there is "currently no standardization or labelling for 'varroa-resistant' stock". Randy Oliver adds the second caveat from thirty years of trying: survivor stock that does well in one area may succumb to mites when moved somewhere else (Scientific Beekeeping).
Brood breaks. Penn State explains why they work: during a break "all of the brood emerges, so the mites are forced out of the cells and onto adult bees", where they cannot reproduce and are exposed to grooming and to any treatment you do use. Swarming is the wild version; splits, queen caging and requeening with a cell are the managed ones. The Coalition lists an "Induced Brood Break" among its non-chemical controls.
Isolation, or at least distance. Mites arrive from outside as well as from inside. The Coalition notes they spread "through natural drift of workers and drones, robbing of weak colonies by stronger ones, swarming, absconding", and a collapsing colony sends its mites to every hive within flight range. A treatment-free yard beside a treated commercial yard is importing the neighbours' selection pressure, and the neighbours are importing yours.
Culling or requeening, not waiting. Oliver's line is the one both camps should pin up: "simply allowing colonies to die is not part of The Solution, you need to propagate the queen lines that survive" (Scientific Beekeeping). A colony left to collapse rewards the most virulent mites in the yard. The treatment-free keepers in Brook's successful studies did not let that happen; they requeened the high-count colonies.
And the cost in the first years is real. Oliver reported a California treatment-free club whose membership survey found "some six out of ten colonies perished" in a year (Scientific Beekeeping, 2012). That is consistent with Brook's "majority died after 3 years" for the unmanaged version, and with the losses the managed version avoided. The difference between the two outcomes was management, not philosophy.
What threshold-based control actually requires
It is tempting to describe the other camp as "treat on a schedule", and some do, but that is not what the guidance says. Every extension service publishes a threshold, a count above which you act, and the point of a threshold is that below it you do not.
The published lines, in mites per 100 bees from a wash or roll: the University of Minnesota treats above 2 and aims to stay under 1 (UMN Bee Lab); the Honey Bee Health Coalition's ninth edition says 1 during the dormant and build-up phases and 2 at peak and through the autumn decline; Penn State Extension's IPM page works "below or around a mean abundance of 2 mites per 100 bees"; and Penn State's López-Uribe lab, at the same university, puts "a reasonable threshold" at "about 5 mites per 100 bees" (López-Uribe lab). A spread from 1 to 5 among careful institutions is not carelessness. It is different climates, different stock and different appetites for risk, and our piece on reading a varroa count goes through where each figure comes from.
Threshold control has its own failure modes. Treat the same way every year and you select for mites that shrug it off: the Coalition warns about "miticide-resistant varroa populations" and caps the number of times a season most products may be used. Treat and never re-count and you do not know whether it worked; the Coalition expects a post-treatment check and says to "apply another suitable control method without delay" if the count is still over the line. And treat on the calendar instead of on a count and you are back to the surveys at the top of this piece.
The one habit both camps need
Look at what the successful version of each approach has in common. The treatment-free keeper in Brook's review counted mites and requeened the colonies that ran high. The threshold keeper counted mites and treated the colonies that ran high. The failed versions of both, the yard that stopped treating and did nothing, and the yard that treated on a date and never checked, share one thing too: nobody counted.
A wash count every month, from the same brood-nest comb, in the same unit, is the one practice that works in either philosophy, and it is cheap. UMN says monthly; Penn State says monthly; the Coalition says monthly from the spring build-up, "or at least 4 times from Population Peak to Population Decline"; López-Uribe says it is "extremely important to monitor mites on a monthly basis". That is four institutions with four different thresholds agreeing on the cadence, which tells you the cadence is the part that is not in dispute.
Two details carry across. First, the count has to be a rate: mites per 100 bees from a wash, roll or CO2 shake. A sticky board gives mites per day off an unknown population, and López-Uribe says outright "there is not a reliable threshold for mites per sticky boards". Second, the count belongs to one colony. A yard average hides the one hive at 9 that is about to seed its neighbours, and that hive is the one both camps want to find first: one to requeen, the other to treat.
HiveMind AI records a mite check per hive with the method named, keeps board counts in their own unit so they never get compared to a wash, lets you set the threshold yourself (the default is 3, labelled as a choice), and flags a reading over a month old as stale rather than quoting it as current. Nothing in it tells you whether to treat. It tells you which colony crossed the line you drew, which is the question both camps are actually asking.
What to do with this
If you treat: keep treating on counts, not dates; count after you treat; rotate what you use; and consider that stock selection and brood breaks are not the other side's property.
If you are treatment-free or want to be: start with proven stock or accept heavy early losses as the price of selection; build brood breaks into the year; keep your distance from treated yards where you can; count monthly; and requeen a colony that runs high before it collapses, because a collapse is not selection, it is a mite bomb.
If you are new: the surveys say the median experienced keeper loses a third of their bees a year with treatment, and the review says the unmanaged treatment-free beginner loses most of them in three. Learn to do a wash before you choose a camp. It is the one skill you will need in either.
Sources
- Apiary Inspectors of America and Bee Informed Partnership, US beekeeping survey results, 2023-2024 and 2024-2025: https://apiaryinspectors.org/US-beekeeping-survey-23-24 and https://apiaryinspectors.org/US-beekeeping-survey-24-25
- Brook, G. (2025). Varroa treatment-free colony losses in the European honey bee (Apis mellifera): a review of published literature. Journal of Apicultural Research: https://www.tandfonline.com/doi/full/10.1080/00218839.2025.2514936
- Loftus, J.C., Smith, M.L., Seeley, T.D. (2016). How honey bee colonies survive in the wild: testing the importance of small nests and frequent swarming. PLoS ONE: https://pmc.ncbi.nlm.nih.gov/articles/PMC4788434/
- Honey Bee Health Coalition, Tools for Varroa Management, ninth edition (June 2026): https://honeybeehealthcoalition.org/tools-for-varroa-management-guide-9th-edition/
- Penn State Extension, Methods to control varroa mites: an integrated pest management approach: https://extension.psu.edu/methods-to-control-varroa-mites-an-integrated-pest-management-approach
- University of Minnesota Bee Lab, Varroa mite testing: https://beelab.umn.edu/varroa-mite-testing
- Penn State López-Uribe lab, Varroa mite monitoring: https://lopezuribelab.com/varroa-mite-monitoring/
- Scientific Beekeeping (Randy Oliver), The "rules" for successful beekeeping (2012) and The varroa problem, part 17b (2018): https://scientificbeekeeping.com/the-rules-for-successful-beekeeping/ and https://scientificbeekeeping.com/the-varroa-problem-part-17b/