Stress Testing Varroa Treatment Regimes

Audio (podcast) version is here or on my YouTube channel

It is difficult to quantify what percentage of honey bee colonies die as a result of varroa mite infestation (and associated viruses) because multiple factors are often involved. From the information I have seen, it is the leading cause of colony losses worldwide. Probably 30 – 50% of all losses are due, in large part, to the mite. This means that choosing a regime for managing the mite is one of the most important decisions a beekeeper can make. In this article, I’ll be doing my best to stress test varroa treatment regimes, to find something reliable.

How bad are varroa mites?

In the USA the Bee Informed Partnership reports annual colony losses of 30 – 50% over the past decade, with varroa consistently coming top of the table as the culprit.

Rosenkranz et al. (2010) highlight that unmanaged varroa infestations can cause complete colony collapse within 1 – 3 years.

Multiple studies have shown the strong link between high varroa loads in the autumn and colony losses over winter, which is not surprising, given that deformed wing virus shortens the lives of bees.

Best way to monitor

As far as I’m concerned, the only monitoring method worth using is the alcohol wash, at least for a one-off assessment of infestation. On the downside, you kill some bees, but it does the job it is intended to do. I have less confidence in the sugar-shake and detergent methods, but each to their own.

Observing mite drop on a board below a mesh floor may be good for long-term monitoring, assuming you have mesh floors. It is certainly a good way to see dead mites after applying a treatment. Having a look at drone brood that has opened up when separating boxes is always worth a go – if I see any mites, I will follow up with an alcohol wash or just treat them.

Spreadsheet Model

The varroa model built by Randy Oliver is my preferred tool of choice when trying to model the effect of miticides on colonies. Obviously, it’s a model, and bees vary considerably, but it is good for comparing different treatment regimes. I used Randy’s spreadsheet model to see how robust different treatment regimes might be.

Mite Treatments

I don’t know if anyone still uses flumethrin or tau-fluvalinate to treat their bees, but I will assume not. These chemicals hang about in bees wax and this leads to resistance by the mites and, just as disturbingly, affect the fertility of queens and drones according to some studies (other studies say not). Regarding amitraz, which is the only synthetic miticide I use, there have been cases of resistance, but I think that happens when it is the only treatment used over many years (no rotation) or strips are left in hives for too long.

The organic acids (oxalic and formic) are effective, don’t contaminate wax, and are unlikely to suffer from mite resistance, at least in the medium term. The essential oil thymol gets in the wax, and can persist for a long time. It definitely causes a reduction in laying eggs by the queen too, so I use thymol much less often than I used to. I’m very pro-organic acids. Evidence is emerging that pads soaked with a solution of oxalic acid in glycerol are effective at helping colonies maintain low mite levels throughout the summer, but that is not an approved treatment in the UK. I have written about Formic Pro previously, and think it’s great in the right circumstances.

Anything else, like dusting with icing sugar, or sprinkling them with snake oil is not worth the effort, in my opinion. The miticides I use are amitraz, formic acid, oxalic acid, and thymol. That gives me plenty of options and allows rotation of treatments to lower the risk of resistance.

Mechanical Methods

Given that varroa mites preferentially seek drone brood in which to reproduce, it makes sense that removal of sealed drone brood has an effect on mite reduction. However, I like drones, and so do my bees. I especially like the drones produced by colonies that have all the traits that I look for in bees, such as low swarming and high honey production.

In parts of Europe, many beekeepers temporarily cage their queens, which stay in the colony but cannot lay eggs, to create a brood break. I have looked at that in the past, and even tried it one year, but for me, it was too much hassle. Queens put on the invisibility cloak just when you need to find them. Used in conjunction with oxalic acid once they are broodless, it’s a great way to kill mites. It’s potentially also a way to clobber the population of bees when a honey flow turns on, or severely weaken them in the run-up to winter if the weather goes against you.

Unreliable Monitoring

Even with a good method of counting mites in a colony, like an alcohol wash, there is plenty of room for error. This is especially the case during spring, when mite numbers are generally quite low and 70% could be inside capped brood. That early in the season, the difference between a mite count of zero (per 300 bees) and two, can have massive ramifications later in the season. For a start, zero in an alcohol wash does not mean there are zero mites in the colony. I don’t believe that doing a wash in April or May is worth it, so I leave it until June. But by then, if I have a colony with lots of mites, my treatment options are … er … limited.

One Treatment Annually In Autumn

Some people rely on zapping the mites once a year, after the supers have been removed at the end of summer. See the chart 1. Based on this model, which does assume that the bees are not especially good a controlling varroa without help, it’s not a great idea.

If you look at the circled cells (showing the mite population at the start and end of the year) you can see the problem. Starting off on 1st Jan (Northern Hemisphere) with just 32 mites – that’s one mite per 500 bees – we still end up with plenty of mites at the year-end (687 in this example). Rolling forward to the following season, the starting mite population would now be 687 rather than 32, with dire consequences – see chart 2.

Any treatment regime that results in more mites at the end of the year than you started with will eventually cause issues, whether it takes six months or three years to show up.

I would also argue that if you use this model with a starting mite population as low as 32 you are probably being a little optimistic. One mite per 500 bees is really low. I tend to use 64 as my best-case scenario (1 mite per 250 bees).

Stress Testing

I have run different treatment regimes through the model and recorded the results below. The important things that I have looked at are:

• starting number of mites
• peak number of mites during the season
• number of mites at beginning of September
• ending number of mites
  

The goal with any regime should be to keep the mite population below 2,000 at all times, as low as possible for brood that becomes the winter bees (September), and make the mite population at the end of the year no higher than the start. The times when these conditions are satisfied are shown in blue in the table below (red if they fail). Furthermore, rotation of treatments is desirable, as well as compliance with guidelines/laws.

The stress testing is when I push up the starting number of mites incrementally to see at what point the colony gets into trouble. The more leeway a treatment regime gives you, the more likely it is that your bees won’t die. I assumed that amitraz, thymol, formic acid, and oxalic (brood free) are 90% effective.

Three Is Best

For the last two years I have followed a “three treatment regime” even though it seems a little over the top to some people. The model showed me that it was by far the best option for keeping mite numbers down. My winter losses were 7.5% in 2024, which I’m happy with. I actually lost more in the summer, due to queen issues and some starving (cringe).

I use amitraz at the beginning of March and remove it before supers go on in April. It works at 10 deg C and is there to clear up the majority of mites that come through winter. This means that the mite levels stay low through the summer. I then hit them with Formic Pro in August and oxalic acid in early December.

There is little evidence (or I can’t find it anyway) that using amitraz results in my summer queen issues (mostly drone laying queens). However, it’s always a possibility, and I would rather be able to use the oxalic acid/glycerol pads through the summer instead, but it’s not allowed.

One more thing: I have seen examples of colonies with very low mite counts in May having loads by August. The idea that having low mite counts early means that they can be left untreated until August carries some risk. Presumably that’s due to the flakiness of mite counts early in the season, but there may be some kind of inward drift at play too.

Brood Break

As you can see from the table, caging the queen in the hive for the month of July (zero brood in early August) does have a positive effect on varroa levels. Just not enough to be worth the effort. And the three treatments per year still seems the way to go, even with the brood break.

Breeding Varroa Resistance

The dream scenario would be for our honey bees to be able to keep mite reproduction levels low enough to enable us to limit treatments to maybe once per year, or even go treatment free. The unhealthy brood odour (UBO) assay is an interesting new tool to help identify bees that can detect and remove infested brood using odour. I am a little skeptical about this dream being fulfilled in my lifetime in my area, but I am certainly keen to raise queens from hygienic stock.