3.6.1. Overview

Participants reported the number of sighted bees with crippled/deformed wings in four categories. There was a significantly higher loss rate in the category "often" 36.3% (95% CI: 28.0–45.5%) than in "seldom" 16.8% (95% CI: 15.4–18.3%). The category "seldom" was also significantly higher than "none" with 13.5% (95% CI: 12.5–14.7%) (Figure 9B).

**Figure 9.** (**A**) Categories for if monitoring for varroa mites during the period April 2018–April 2019 was practiced and the corresponding winter honey bee colony loss rates (and 95% CI). (**B**) If and how often the participants noticed bees with crippled/deformed wings in their colonies (during summer season) in four categories and their loss rate (and 95% CI). (**C**) Distribution of how many different treatment methods (including drone brood removal) per operation were used, i.e., formic acid—short term and thymol would be two. (**D**) Histogram of which months monitoring of varroa infestation level (e.g., counting mite fall) was done and by how many beekeepers (*n* = 1360) in the survey. Spring, summer, and winter are color coded; see Figure 10.

Beekeepers who monitored varroa infestation level had a significantly lower loss rate at 14.7% (95% CI: 13.8–15.7%) than those that did not at 21.7% (95% CI: 18.3–25.5%) (Figure 9A). The monitoring of varroa infestation level was primarily performed in the months between July and October (Figure 9D).

Varroa control information was provided by 1455 participants. Most beekeepers (87.4%) used between two and four different treatment methods (Figure 9C). The most common treatment in spring (April–May) was drone brood removal (Figure 10A). In summer (June–October) the prevalent treatment was formic acid, with more participants using a long-term than short-term method (Figure 10C,D). In the autumn/winter period (November–January), oxalic acid treatment in its various forms was the most dominant (Figure 10B,E,F). Lactic acid, synthetic methods (including Amitraz, Coumaphos, and other synthetic methods), and "another method" (without synthetic methods) were the least common used treatments in the survey (Figure 10J–L).

**Figure 10.** Varroa control methods and their month of usage as histograms starting with April 2018 and sorted by total frequencies. (**L**) "Synthetic methods" are combined and include Amitraz, Coumaphos, and "other synthetic methods"; (**G**) "other biotechnical methods" do not include drone brood removal or hypothermia; and (**K**) "another method" excludes synthetic methods. We defined April–May as spring (blue), June–October as summer (green), and November–January as autumn/winter (black). The months February, March, and April 2019 are excluded. The Y-axis has separate scales for each treatment. *n* = 1455 participants.

#### 3.6.2. Treatment as Single Factor

The colony loss rates of operations with different treatment methods were compared. Methods were grouped into spring, summer, and winter according to the month of usage (Figure 10). Within the groups, we did not differentiate if the same treatment methods were used multiple times by any given beekeeper, only if they were performed.

In spring (April–May), four control methods were applied by at least 20 participants (Figure A2). Results of drone brood removal as a single factor is discussed in the next section. The application of the other three spring control methods (hyperthermia, other biotechnical method, and oxalic acid trickling) had no significant effect on winter loss rates.

For summer (June–October), eleven treatment methods (yes = *n* > 19) were compared (Figure A3). Significantly lower loss rates were found for participants using other biotechnical methods (13.1%, 95% CI: 11.7–14.8%) than for those participants not using them (16.2%, 95% CI: 15.2–17.4%) (Figure A3C). Participants who performed oxalic acid trickling (including oxalic acid mixtures) in summer had a significantly higher loss rate (17.4%, 95% CI: 15.7–19.4%) compared to those who did not use this treatment method (14.6%, 95% CI: 13.7–15.7%) (Figure A3H).

Only oxalic acid methods (sublimation and trickling) were applied in winter (November–January) with a minimum of 20 reports. No significant differences between the application or non-application of these methods were found (Figure A4).

#### 3.6.3. Drone Brood Removal Combination

Drone brood removal performed only in spring, only in summer, and in both seasons did not have an effect on winter losses compared to no drone brood removal at all. Removing drone brood only in summer was significantly worse (17.6%, 95% CI: 15.3–20.3%) than drone brood removal in spring and summer (13.3%, 95% CI: 11.8–15.0%) (Figure 11).

Drone brood removal as a single factor in spring (which includes also the abovementioned beekeepers that apply this treatment in both spring and summer) resulted in a just significant effect compared to those not removing drone brood in spring: confidence intervals are overlapping (yes: 14.1% (95% CI: 12.8–15.4%), no: 16.3% (95% CI: 15.1–17.5%)) (Figure A2A), but *χ*2 = 29.7 (*p* < 0.05). In summer, no significant difference was found between whether beekeepers applied this control method.

#### 3.6.4. Treatment Combinations

To gain a more thorough understanding of the used methods, their combinations, and their success, all possible combinations of different treatment methods were calculated. How the combinations

were generated and selected is described in the Materials and Methods section. The used letters and abbreviations for the generated treatment combinations as seen in Figure 12 are explained in Table 2.

The most used method combination in the survey was (A) formic acid—long term in summer—and oxalic acid—trickling in winter. It showed the smallest confidence interval and significantly lower loss rate of 10.8% (95% CI: 9.2–12.6%) than the second most frequent combination (B) formic acid—short term in summer—and oxalic acid—trickling in winter—with 16.1% (95% CI: 13.0–19.6%), combination (D) formic acid—short term in summer—and oxalic acid—sublimation in winter—with 16.6% (95% CI: 12.6–21.5%), and the combination (H) formic acid—long term in summer—and oxalic acid—sublimation in winter and summer with—20.1% (95% CI: 15.1–26.1%). The highest loss rates were observed for (L) formic acid—short term in summer—and oxalic acid—trickling in summer and winter—with a loss rate of 26.9% (95% CI: 18.6–37.3%), combination (N) formic acid—long term—and biotechnical method in summer followed by oxalic acid—trickling in winter—with 23.8% (95% CI: 15.0–35.6%) and the single treatment (O) formic acid—short term in summer—with 22.4% (95% CI: 15.2–31.9%). The widest confidence interval was observed for the combination (P) oxalic acid—trickling in summer and winter—with a loss rate of 24.3% (95% CI: 11.5–44.2%).

Thymol was only found in one of the generated combinations: (T) thymol in summer and oxalic acid—trickling in summer and winter—resulting in a loss rate of 20.1% (95% CI: 9.8–36.9%).

**Figure 12.** Winter honey bee colony loss rates (and 95% CI) resulting from combinations of various varroa control methods ordered by number of participants using the given combination: The point fill colour represents the number of different methods (single method = white, two methods = gray, and three methods = black). Please refer to Table 2 for explanation of the combination letter.


