In 2022 we ran two independent and completely different field trials to test whether installing robber screens would be of benefit for reducing the number of mites that hitchhike their way into a hive.

Mites don’t invade a hive on their own –– they need to be carried in on a bee, either a returning forager or robber, or by a drifted or robbing bee from another colony. In my Mediterranean climate –– where robbing typically doesn’t occur at colony collapse [[1]] –– most mite immigration appears to be due to the normal “drifting” of bees between hives. The net result is a “diffusion” of mites from colonies with high infestation levels to those with fewer mites, with some colonies appearing to be far more “attractive” to drifted bees than others [[2]].

The Question: Could installing “robber screens” reduce the amount of bee and mite drift?

Trial A –– A Crossover Trial, No Mite Donor Hives

Materials and Methods

The Robber Screens Used

There are a number of different styles of anti-robbing screens/entrance guards on the market. For my field trials, I chose to use one manufactured by Beequip in New Zealand (Figures 1 & 2), since I liked several of its features: the simple reversible design (with the option of an offset very small entrance), the fact that they stack neatly for storage in minimal space, and especially that they were made of stainless steel (which would last forever; I have a preference for equipment that I can hand down to my grandkids). This is not a sales pitch for this product –– I haven’t run any comparative tests of the various robber screens available on the market.

Fig. 1 For this trial, I used the stainless steel RobberGards™ from New Zealand. These screens can be flipped to provide a full-width top entrance, or a very small offset top entrance. Due to the high ambient temperatures, for this trial we installed them with the full-width opening, that has a lip below it, which (as you can see) keeps incoming bees from going over it to the entrance.

As with any robber screen, it takes the foragers of a hive a few days to figure out where the entrance is.

Practical application: Since the “home” bees are eventually able to figure out how to enter the hive, the question then is, will drifting bees (carrying mites) “give up” before they find a way in?

Fig. 2 Any investigating robber or drifted bee would be attracted to the odor emanating through the punched screen, but be unable to directly enter (the assumption being that it would eventually give up). And even if it attempted to enter from the top opening, it might be met by unwelcoming guards.

The Trial Yard

We set up this trial of robbing screens in one of our yards near town, which had lots of hobbyists and feral colonies in the neighborhood. In previous years, we suspected that our hives there received considerable mite immigration (a common assumption by beekeepers when their mite wash counts explode late in the season). We also had a number of our own hives, not involved in the experiment, in the same yard, which got moved out partway through the trial.

Preparation for the Trial

To determine the number of mites being carried into a hive from other colonies (“mite immigration”), one must first attempt to eliminate any resident mites in that colony. To do so isn’t easy. In previous immigration studies that I’ve performed, I’ve found a combination of amitraz and fluvalinate to be effective, since we haven’t used any synthetics in our operation for over twenty years (and our mites test to be highly susceptible to amitraz). So we prepared these test hives by treating them with a combination of Apistan and Apivar strips (to hit them with two different modes of action) (Figure 3).

Fig. 3 We got a late start in preparing for this trial (not beginning until June) –– too late to completely eliminate all the mites by our intended start date in early August. We applied a combination of Apivar, Apistan, and for good measure, and extended-release oxalic acid pad.

We applied additional fresh miticide strips from time to time, to ensure than any incoming mites would be quickly killed before they could enter a cell to reproduce. By the time we started taking stickyboard counts in late August, half the hives exhibited daily mite drops of 2 or less, but some were considerably higher. Since there weren’t any rainy days to see whether their mite counts had dropped to zero (which would have validated that they were essentially free of mites), we checked for the locations of the mites on the stickyboards (Figures 4 & 5).

Fig. 4 In a mite-free hive, most of the fallen mites drop in the outer perimeter of the stickyboard, not in the capping debris beneath the broodnest (the dark rows of debris to the right). I took this photo in the field, so while I had my magnifying glasses on, I placed pieces of straw pointing to each mite.

Fig. 5 On the other hand, in suspected mite-producing hives, most mites typically fall in the debris beneath the broodnest.

Based upon our experience from counting mites on hundreds of stickyboards, and “getting to know” each receiver hive in various trials, we find the above observation to be consistent. So unless we get rainy no-flight days for validation, we typically exclude suspect hives from quantification of mite immigration.

Experimental Design

In this trial, we needed to account for the residual mites apparently present in some of the test hives. Since our objective was not to quantify the absolute value of mites entering, but rather only to see whether installing robber screens would make a difference, I decided to use a “crossover design.”

Crossover design trials are widely used in medical research for testing the effect of drugs or other medical procedures, since they help to account for each subject’s innate differences (health, age, lifestyle, environment, etc.), which make it hard to compare one test group to another (similar to trying compare the results between individual bee colonies). On way to get around this problem is to randomly divide the test group in two, and have one half take the med each day, and the other half a placebo. Record any apparent observed effects over a period of time, and then switch their pills to the reverse. Do this back and forth, so that each subject is observed for any effects of being on the drug versus when off it (without them knowing whether they were on or off).

Practical application: From previous research, we already knew that there would likely be a huge degree of hive-to-hive variation in mite immigration. By running a crossover design, each individual hive would toggle back and forth between having a screen on or off, minimizing the variables of hive-to-hive immigration intensity and background residual mite drop.

Treatment Assignment

Most of the hives in the test yard were in set in pairs near each other, so we simply flipped a coin to assign one hive of each pair to either of two treatment groups (A or B), the treatments being either installing a robber screen or leaving the entrance open. Once we’d collected enough stickyboard counts over time, we then “crossed over” the treatments by swapping the guard to the other hive of the pair (and then later, reversed them again).

Stickyboard Counts

We placed each hive on full-width screened bottoms, with reusable stickyboards made from fiberglass-reinforced plastic (FRP) shower board, on which I draw gridlines with a Marks-a-Lot felt pen [[3]]. This design of stickyboard works very well with a light coating of 1:1 mineral oil: petroleum jelly, and can be used for many years (Figures 6 & 7).

Fig. 6 We set up the bottom boards so that we could pull the stickyboards out from the rear, to avoid disturbing the entrances. Despite inserting wedges to seal the rear opening, the occasional bee would sneak in (not a problem, since any mites that it was carrying would still likely wind up in the count).

Fig. 7 We pulled the stickyboards twice a week to count the mites. Even youthful Rose found that using reading or magnifying glasses helped to differentiate mites from the hive debris. We confirmed the accuracy and consistency of our counts by often recounting the others’ board.

Due to our late start at colony preparation to eliminate their mite, we had to hold off until late August to begin stickyboard counts. At that time, half the 24 hives were dropping fewer than 3 mites per day, with only two showing over 10 per day (one of which we excluded from the trial [[4]]), leaving us 23 test hives. Since the objective of this trial was to determine whether the installation of robbing screens would reduce mite immigration (as opposed to quantification of the absolute amount of immigration), the crossover design would correct for any background drop of residual mites.

We wound up recording 18 semi-weekly stickyboard mite counts from 24 hives over the course of 49 days (a total of 432 stickyboard counts), each hive receiving three alternating treatment regimens.

Results and Interpretation

Based upon past observations and assumptions, we expected considerable mite immigration into the hives to occur in the month of September. That did not occur. That may be partly due to a miscommunication which resulted in my sons moving the extra hives not in the experiment to another yard on the first of September (which appeared to cause a considerable decrease in the mite drop counts). Anyway, the low amount of mite immigration made it difficult interpret the results (Figure 8).

Fig. 8 One of each pair of hives was assigned to Group A, the other to Group B (most pairs sitting side by side). Unfortunately, roughly half the hives exhibited very low mite drop counts over the entire course of the trial (whether or not they had screens on), indicating low mite immigration. That said, the hives with higher mite drop counts generally exhibited a similar pattern of mite count intensity (apparently due to environmental factors), independent of whether they had screens on or not –– since there was no noticeable change in counts due to reversing the treatments (swapping the screens). Note: we did not take counts during the cold rain from September 18-22, so any decrease in mite drop during that time period would not show on these graphs. We also excluded the data from one of the 24 hives, which initially had a very high mite drop count (although its pattern of drops followed the trend).

The data may be clearer if we look at the average semi-weekly counts for each test group (Figure 9).

Fig. 9 Pooling all the counts into group averages, we can look for any overall changes in mite drop counts resulting from a change in robber screens (due to our random treatment assignment, Group A happened to start with higher counts). Pay attention to the expected inflections up or down (circled), following when screens were installed or taken off. For the August 29 and September 30 inflections, there didn’t appear to be an effect. For the September 8 inflections, the slopes were contrary to what one might expect!

Conclusions

As we’d observed in previous experiments, roughly half of the hives exhibited minimal mite drop counts (which made them worthless for looking for a change in immigration due to treatment). We also noted that there was a major drop in mite counts when potential “mite donor” hives were removed from the yard.

We felt that the results of this trial were frustratingly inconclusive.

Practical application: Scientific experimentation is a learning process, and we don’t expect to always “get it right” the first try. So we replicated this trial the next year in another yard, improving the design by starting mite treatment preparation well beforehand (which allowed us to begin the trial earlier in the season), running the trial for a longer duration, moving in collapsing mite-donor hives, and doing an additional crossover swap of screening.

I’ll write about that trial later in this series, but next month I’ll show a different trial that we ran with robbing screens concurrent with the one in this article.

Acknowledgements

My great appreciation to BeeQuip^NZ for supplying the RobberGards, and to Rose Pasetes, for helping me with all the tedious mite drop counting.

Citations and Notes

[1] Oliver, R (2023) A Survey on Robbing at Collapse. ABJ Feb 2023 https://scientificbeekeeping.com/a-survey-on-robbing-at-collapse/

[2] Oliver, R (2023) A Study on Bee Drift and Mite Immigration: Part 4. May ABJ https://scientificbeekeeping.com/a-study-on-bee-drift-and-mite-immigration-part-4/

[3] https://scientificbeekeeping.com/scibeeimages/@Citizen-Science-Mite-Drift-Instructions.pdf

[4] This extreme outlier, which although showing mite wash counts of zero, was still dropping over 30 mites a day in late August.

Category: Varroa IPM Strategies, Varroa Management

ScientificBeekeeping.com

Can Robber Screens Reduce Mite Immigration? Part 2

Contents

Can Robber Screens Reduce Mite Immigration?

Part 2