Dependent variables and sampling protocols

A variety of parameters is measured to relate to damage thresholds of the colonies hosting different parasite population sizes (Table 4). For some of the most invasive measures, i.e., measures of bees or brood (Table 4), it is common during the mid-experiment sampling intervals to sample only a subset of colonies in each treatment. These numbers have ranged from 3 (Delaplane and Hood, 1999) to 4 (Delaplane and Hood, 1997) to 2 (Strange and Sheppard, 2001). It is important to remember that these mid-experiment measures are made not in pursuit of statistic differences, per se, but to provide a retrospective snapshot of colony condition at time of treatment. They will also slightly increase within group variability of the parameters measured. All colonies are dismantled and fully measured at the end of the study when statistical rigor is desirable to identify the treatment regimens that optimized colony condition. 

1. Place bottom board sheets under a wire mesh and positioned under the colony (see section 4.2.4. ‘Natural mite fall’).

2. Remove the sheet after 24-48 h.

3. Count number of mites and the adjust number to mites recovered per 24 h.

4. Collect samples of ca. 300 adult bees.

5. Count the number of mites and number of bees (see section ‘Infestation rates of adult bees’).

6. Report the data are as mites / 100 bees.

7. Derive fresh bee weight (mg) by sampling live bees off the comb into pre-weighed, or tarred, containers, weighing the container, subtracting container weight to get net bee weight, counting bees, and dividing by net bee weight to get mg per bee.

This sampling can be combined with sampling for mites per adult bee as described above.

8. Obtain ending colony bee population according to the methods described under section ‘Measuring colony strength at end of experiment: objective mode’ of the BEEBOOK paper on estimating colony strength (Delaplane et al., 2013), or from a variation from this method using net colony bee weight (kg) and average fresh bee weight (mg):

8.1. Obtain net colony bee weight by first closing the colony entrance with a ventilated screen in the evening or early morning before sampling to trap all bees inside.

8.2. Weigh the whole hive in the field (Fig. 19).

8.3. Open the hive.

8.4. Brush all bees off every comb and surface (usually into a temporary holding hive).

8.5. Re-weigh the hive without bees.

8.6. Calculate the difference in weight, which is the net weight of bees.

8.7. Divide this number by fresh bee weight to derive colony bee population.

9. Count the number of sealed brood cells as described in section, steps 17 and 18.

10. Colony mite population is derived from the methods described in section

Once the investigator knows ending bee population, phoretic mites per bee, number of sealed brood cells, and mites per sealed brood cell, then one can extrapolate to (phoretic mite population + mite population in brood) = total colony mite population.

11. Quantify visible brood disorders.

Brood disorders are sometimes associated with varroa parasitism (Shimanuki et al., 1994) and can contribute to colony damage.

11.1. Select two relatively contiguous patches of brood in the late larval – capped stages (stages more likely to express visible symptoms)

11.2. Overlay on each patch a 10-cm horizontal transect and a 10-cm vertical transect intersecting at the centre (Fig. 20).

11.3. Examine along each transect every cell of brood under strong light and magnification for visible disorders, i.e., symptoms typical of American foulbrood (see the BEEBOOK paper by de Graaf et al., 2013), European foulbrood (see the BEEBOOK paper by Forsgren et al., 2013), sacbrood (see the BEEBOOK paper on honey bee viruses by de Miranda et al., 2013), or chalkbrood (see the BEEBOOK paper on fungal diseases by Jensen et al., 2013).

11.4. Report the parameter as percentage of brood expressing visible disorders.

12. When honey yield needs be considered as a parameter of the economic threshold, calculate colony yield (kg) by weighing honey supers before and after they are placed on colonies during a nectar flow. 

13. Measure the infestation rate by Acarapis woodi in the colonies.

This rate could be affected by the presence of varroa and its measure might be needed as economically relevant. For this sample 50 workers per colony and place in alcohol. See the BEEBOOK paper on tracheal mites (Sammataro et al., 2013) for the method to detect the presence of the mites.

Table 4. Dependent variables and sampling protocols recommended and employed in the literature on field-derived damage thresholds. Months have to be considered according to the season at the location of the study. a Delaplane and Hood (1997); b Delaplane and Hood (1999); c fluvalinate year 1 and thymol year 2; d derived from regression predictions based on sticky board mite counts; e Delaplane et al. (2010); f Strange and Sheppard (2001); g Currie and Gatien (2006).

Experimental treatments

Dependent variables measured in each treatment group

Sampling intervals


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)

1. mites / 24 h on bottom sheets

2. mites / 300 bees

3. fresh bee weight (mg)

4. colony bee populations

5. number sealed brood cells

6. colony mite populations

7. visible brood disorders and other diseases

8. colony honey yield (kg)

9. subjective “survivability” score at beginning of the season of following year

at regular intervals in the year

at mid-season


1. fluvalinate Jun.

2. fluvalinate Aug.

3. fluvalinate Oct.

4. no treatment

1. mites / 18 h on sticky sheets

2. mites / 300 bees

3. fresh bee weight (mg)

4. colony bee populations

5. number sealed brood cells

6. colony mite populations

7. visible brood disorders (Dec only)

Jun., Aug., Oct., Dec.


1. fluvalinate Feb.

2. fluvalinate Aug.

3. fluvalinate Feb.+Aug.

4. continuous fluvalinate

5. no treatment

1. mites / 20 h on sticky sheets

2. mites / 300 bees

3. fresh bee weight (mg)

4. colony bee populations

5. number sealed brood cells

6. colony mite populations

7. visible brood disorders (Sep.-Oct. only)

8. subjective “survivability” score (following Jan only)

Feb., May., Aug., Sep.-Oct.


1. continuous treatmentc beginning Jun.

2. treatment Aug.

3. treatment Oct.

1. mites / 100 bees

2. fresh bee weight (mg) (Dec only)

3. colony bee populations

4. colony weight (kg) (Dec only)

5. cm2 brood (all stages)

6. colony mite populations d

7. % bees infected with Acarapis woodi (Dec. only)

Aug, Oct, Dec


1. fluvalinate Apr.

2. fluvalinate Aug.

3. fluvalinate Oct.

4. fluvalinate Apr.+Oct.

5. continuous fluvalinate

6. no treatment

7. coumaphos Apr.

1. mites / 20 hr on sticky sheets

2. mites / 300 bees

3. fresh bee weight (mg)

4. colony bee populations

5. number sealed brood cells

6. colony mite populations

Jun., Aug., Oct., Nov., Apr.


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

1. mites per bee

2. colony honey yield (kg) (Aug. only)

At pre-treatment (May and Sep.), then post-treatment weekly for 3 weeks (spring) then biweekly thereafter


Fig. 19.
The whole hive is weighed in the field to obtain final colony bee population. Photo: Keith Delaplane.

Figure 19


Fig. 20. Selecting and inspecting brood cells for diseases. Photo: Keith Delaplane.

Figure 20