4.4.1.2 Experimental treatments, sample size, and colony arrangements

1. Create a range of varroa colony densities by treating sets of colonies with acaricide at different points of a season, bracketing as widely as possible the months of bee activity particular to one’s region.

At the very least, one treatment should be early in the season when bees are emerging from winter senescence; one treatment should be at the peak of the active season, and another should occur in autumn which is typically the period of highest ratio of mites : bees. More intervals will improve the resolution of the resulting model. It is imperative that the design include a negative control – a set of colonies left untreated, and it is highly recommended that the design include a positive control – a set of colonies treated continuously (Table 3). Including both controls will provide the widest range of varroa densities possible within which the investigator can retrospectively search for mite densities that are damaging or non-damaging.

2. Use an average sample size (initial number of colonies per treatment) of 11.

The literature indicates a range 7-20. Studies using sample sizes within this range never failed to detect treatment effects for at least some dependent variables (Table 3).

3. Stick to one mite control product.

This avoids the risk of experimental confounding error due to variation in acaricide efficacy or unknown sublethal effects on host bees (investigators sometimes included different acaricides in their treatments, apparently with a view to fine-tuning control recommendations for their region).

4. Take into account resistance of mites to acaricides when choosing the treatment.

Whether the population is resistant to a particular product can be tested following the method described in section 3.6.3. ‘Bioassays to quantify the susceptibility of the varroa mite to acaricides’.

5. Control or at least monitor colony spatial arrangement.

 Varroa mites can spread horizontally through infested workers and drones drifting between colonies (Greatti et al., 1992; Frey et al., 2011) and exert a strong influence on results. Depending on one’s objectives, one can set up apiaries to encourage drift (assign treatments within the same apiary) or discourage drift (assign treatments by apiary). The first scenario acknowledges that immigration may confound results, yet this condition is presumed more “real world” because modern beekeeping often encourages mite horizontal transmission with high-density apiaries. This option is however not relevant if treatments are made with persistent active ingredients that are distributed by contact between bees (e.g. fluvalinate). Drifters could contaminate colonies with different treatment regimes (Allsopp, 2006). The second scenario, in contrast, gives a more uncluttered description of the effects of delayed mite treatment. Including both conditions permits a test of the assumption that thresholds occur earlier under conditions of high mite immigration. In either case, it is recommended that the objective be explicit and the spatial arrangement designed accordingly: 1. if immigration is to be minimized then assign all colonies within one apiary the same treatment and space apiaries as widely apart as possible; or 2. if immigration is to be maximized then assign all treatments within the same apiary. Other drift-minimizing practices, such as painting symbols at hive entrances or arranging colonies in circles, will not substitute for wide spatial distances between colonies of different treatments.

Table 3. Experimental treatments, sample size, and colony spatial arrangements recommended and found in the literature on field-derived damage thresholds. a all studies cited here were performed in the Northern Hemisphere, so early season is Feb-May and autumn Sep-Oct; b initial colonies per treatment; c Delaplane and Hood (1997); Delaplane and Hood (1999); e fluvalinate year 1 and thymol year 2; f Delaplane et al. (2010); g Strange and Sheppard (2001); h 10 each for May, 7 for Sep; i Currie and Gatien (2006).

Experimental treatmentsa

nb

Colony spatial arrangement

Reference

1. acaricide X early in the season

2. acaricide X at peak of season

3. acaricide X end of season

4. untreated colonies (negative control)

5. acaricide X continuous treatment

(positive control)

12

according to objective (see point 4)

recommended

1. fluvalinate Jun.

2. fluvalinate Aug.

3. fluvalinate Oct.

4. no treatment

18

divided equally between 2 states (ca. 120 km apart), treatments applied by apiary within state (minimize drift effect)

c

1. fluvalinate Feb.

2. fluvalinate Aug.

3. fluvalinate Feb.+Aug.

4. continuous fluvalinate

5. no treatment

12

divided equally between 2 states, treatments applied within apiary within state (maximize drift effect)

d

1. fluvalinate Feb.

2. fluvalinate Aug.

3. fluvalinate Feb.+Aug.

4. continuous fluvalinate

5. no treatment

8

divided equally between 2 states, treatments applied by apiary within state (minimize drift effect)

d

1. continuous treatmente beginning Jun.

2. treatment Aug.

3. treatment Oct.

20

divided equally between 2 states, treatments applied within apiary within state (maximize drift effect)

f

1. fluvalinate Apr.

2. fluvalinate Aug.

3. fluvalinate Oct.

4. fluvalinate Apr.+Oct.

5. continuous fluvalinate

6. no treatment

7. coumaphos Apr.

8

divided equally among groups (circles) of 8, each circle 15 m apart (minimize drift effect)

g

1. fluvalinate May

2. fluvalinate Sep.

3. 4 formic 4 d apart May

4. 4 formic 4 d apart Sep.

5. 4 formic 10 d apart May

6. 4 formic 10 d apart Sep.

7. coumaphos May

8. coumaphos Sep.

9. no treatment

10 or 7h

divided equally between 2 apiaries 8 km apart, within apiary colonies further subdivided into “low” initial varroa density or “moderate” initial density

i

1. fluvalinate May

2. 5 formic 1 week apart May

3. 1 formic slow release May

4. no treatment

7

treatments applied within 1 apiary of “high” initial varroa density

i