Varroa destructor Anderson and Trueman (Acari: Varroidae) is a parasitic mite that has become the most serious health problem of the western honey bee, Apis mellifera Linnaeus (Hymenoptera: Apidae), worldwide. There is evidence that the varroa mite is one of the factors associated with the unprecedented loss of honey bee colonies recently experienced in parts of Europe and North America (Guzman-Novoa et al. Reference Guzman-Novoa, Eccles, Calvete, McGowan, Kelly and Correa-Benítez2010; Le Conte et al. Reference Le Conte, Ellis and Ritter2010).

Honey is the main source of income for beekeepers worldwide. However, despite the damaging effects V. destructor has on honey bee health, remarkably few studies have been conducted to specifically assess its impact on honey production (Arechavaleta-Velasco and Guzman-Novoa Reference Arechavaleta-Velasco and Guzman-Novoa2000; Murilhas Reference Murilhas2002; Medina-Flores et al. Reference Medina-Flores, Guzman-Novoa, Aréchiga-Flores, Aguilera and Gutiérez2011). In Canada in particular, we are aware of only one study conducted in the prairies, in which colonies with different infestation levels of varroa mites were evaluated for honey production and compared with yields of colonies treated with acaricides. Currie and Gatien (Reference Currie and Gatien2006) found that colonies infested with two mites per 100 bees in the spring produced approximately half the amount of honey yielded by colonies treated with acaricides. When infestation levels were high (21 mites per 100 bees in the spring), untreated colonies produced only 1.3 kg of honey, whereas acaricide treated colonies produced 48 kg per hive.

The degree to which honey yields are affected in colonies infested with varroa mites may vary depending on multifactorial interactions in different environments and geographic locations. While mite infestations can be devastating in some regions, they may have less serious effects in others. Therefore, it is important to accumulate more evidence on the relative effect of varroa mite infestations on the honey yield of colonies in different countries and regions of the world. The objective of this study was to compare honey yields between groups of colonies with high and low varroa mite infestation rates in eastern Canada.

Experiments were conducted in two adjacent apiaries located at the Honey Bee Research Centre of the University of Guelph, in Guelph, Ontario, Canada (43°N, 80°W). A total of 152 honey bee colonies housed in Langstroth hives and headed by Buckfast queens that had been mated the previous season in isolation (Thora island, Lake Simcoe, Ontario, Canada), were evaluated twice for adult bee population size and for varroa mite infestation levels. The first evaluations were conducted during the last two weeks of April 2012 (spring), whereas the second evaluations were done 16 weeks later during the third and fourth week of August 2012 (summer). None of the colonies in the experiments had been treated against parasitic mites since the previous fall.

Colony populations were estimated by counting the number of frames covered by bees. A frame was counted as one unit, only if completely covered by bees, or fractions of one unit were used when partial coverage of the frame by the bees was observed; this included frames in the brood chamber as well as those in the supers. The mite population level of the colonies was determined by using screened (4 mm mesh) hive bottom boards containing sticky papers (to trap falling mites) in the surveyed hives. The number of mites that fell onto the sticky papers in a period of three days was counted and divided by three to obtain an average mite fall per 24 hours. This evaluation was repeated on three separate and consecutive occasions over the course of nine days for each colony, both in spring and summer. To determine how much mite populations grew in each colony in 16 weeks, the average mite fall count in the spring was deducted from the average mite fall count in the summer, and the resulting figure was used to select the colonies with the highest (n = 10) and the lowest (n = 10) mite population growth (H and L, respectively; Table 1). The ranges of the varroa population growth in mites fallen per day were 1–10 and 11–226, for the L and H colonies, respectively. In the summer, only the colonies of the two selected groups were additionally tested for infestation in adult bees and honey production.

Table 1 Mean Varroa destructor, bee populations, and honey production (±SE) estimated in honey bee colonies showing low (L, n = 10) and high (H, n = 10) mite population growth in a period of 16 weeks.

*Mite populations were measured as either number of mites fallen per day on sticky boards, or as number of mites in 100 adult bees, whereas adult bee populations were calculated from number of frames covered by bees. Honey yields were measured in kg by weighing the supers harvested from each colony before and after honey extraction.

^†Comparisons between L and H colony groups and P-values are based on t-tests and are only valid within the same row.

Mite infestation levels on adult bees were determined during the summer as follows. For adult bees, a sample of ∼200 workers was collected in a jar containing 70% ethanol from brood nest frames in each hive. The jars with the bees were agitated for 30 minutes on a mechanical shaker (Eberbach, Ann Arbor, Michigan, United States of America) and the mites dislodged from the bees’ bodies strained. The number of bees and mites was counted and the number of mites per 100 bees was calculated for each sample. This test was performed three times per colony every three days during the last two weeks of August 2012 and an average adult bee infestation rate was calculated.

Honey production was determined during the last week of August 2012 and was measured by weighing the supers harvested from each colony before and after honey extraction. Supers were weighed with a floor beam scale (Cardinal Detecto, Atlas, Kitchener, Ontario, Canada). The yield per colony was obtained by subtracting the weight of supers with extracted combs from the weight of supers with combs containing honey.

Paired comparisons of L and H colonies for bee population, mite fall, adult bee infestation, and honey production were done using t-tests. Data on percentage adult bee infestation were square root-arcsine transformed, whereas the data for the rest of the variables were subjected to log transformations because they were not normally distributed. Additionally, honey yield data and mite infestation parameters were analysed for correlations using a Pearson correlation test. All statistical analyses were performed with the R-Statistical Program (R Development Core Team, Auckland, New Zealand).

Regarding bee populations, no differences were detected between colonies (P > 0.05). Additionally, no differences were found between L and H colonies for the number of mites dropped on sticky boards in the spring (P > 0.05), but significant differences were detected between the two groups of colonies for this variable in the summer (P < 0.001). Mite fall counts were 11 times lower in the L colonies compared with the H colonies. This result was corroborated in adult bees; workers of H colonies were six times more infested with varroa mites than workers of L colonies and this difference was significant (P < 0.0001) (Table 1).

H and L colonies differed significantly for honey production (P < 0.001). L colonies produced 28.91 ± 2.34 kg of honey per hive versus 18.49 ± 0.77 kg for the group of H colonies (Table 1). Additionally, the mite infestation level of colonies measured as mite fall or as number of mites per 100 bees was negatively correlated with honey production (r = −0.62, P < 0.05, and r = −0.76, P < 0.01, respectively).

H colonies produced 36% less honey than L colonies, which agrees with results of previous studies conducted in Mexico (Arechavaleta-Velasco and Guzman-Novoa Reference Arechavaleta-Velasco and Guzman-Novoa2000; Medina-Flores et al. Reference Medina-Flores, Guzman-Novoa, Aréchiga-Flores, Aguilera and Gutiérez2011) and Portugal (Murilhas Reference Murilhas2002) in which honey yields of colonies highly infested with varroa mites were 45–66% lower than those of treated colonies. Honey loses were also found in the study conducted in the Canadian prairies, in which groups of colonies with different levels of infestation by varroa mites were tested (Currie and Gatien Reference Currie and Gatien2006). The average yield per untreated colony of the group with low infestation levels (2 mites/100 bees) was 66 kg, whereas that of the highly infested group (21 mites/100 bees) was only 1.3 kg (98% difference). The larger difference in honey yields between the high and low groups of the study by Currie and Gatien (Reference Currie and Gatien2006) relative to the difference found between H and L colonies of this study, could be due to the fact that mite loads between the high and low groups of the two studies, were higher in the study conducted in the prairies (about 10 and 6 times, respectively). Additionally, the colonies of the high group in the study by Currie and Gatien (Reference Currie and Gatien2006) had about three times more mites on adult bees compared with the H colonies of our study. Another factor that could have accounted for larger differences in honey production between the colony groups in the prairies relative to eastern Canada, are environmental effects.

The Canadian prairies comprise a region with very favourable conditions for high honey production. However, nothing was known about the damaging potential of varroa mite infestations in regions with less favourable conditions for honey production in temperate climates; this is the first time that evidence is generated on the effect that varroa has on colonies in a region of relatively low honey production in a temperate climate. Theoretically, varroa infestations of colonies would be expected to be more damaging in the prairies than in Ontario for several reasons, including higher honey yields, lower economic threshold for mite infestations, and shorter time for colony development in spring. Colonies in the prairies produce more honey than colonies in Ontario because the prairies have longer day lengths (based upon latitude) and have many more flowering plants, which provides more foraging hours and resources for the bees to produce honey. Due to these differences, colonies typically produce about twice as much honey in the prairies, relative to colonies in Ontario. Therefore, the amount of honey loss as a result of similar varroa infestation levels would be expected to be higher in the prairies. Additionally, and because less honey per hive is produced in Ontario, the economic threshold would be higher in this region (more mites could be tolerated without a reduction in honey yield). Moreover, the season is shorter in the prairies because it normally does not warm up enough before late April for the bees to develop their colonies. Thus, it would be expected that lower varroa infestation levels would cause more damage for the development of colonies in the prairies. However, and despite the less favourable conditions for mite damage in Ontario relative to the prairies, we found in this study, that varroa mite infestations significantly reduced honey yields in this geographical area of Canada. This result demonstrates that the varroa mite can be very harmful to honey bee colonies in different regions with different potentials for honey yields. Furthermore, regardless of the magnitude of the loss in honey yields of infested colonies, it is clear that V. destructor is highly detrimental for honey production. This study confirms that negative effect on honey yields in a different geographical region of Canada.

One intriguing result of this study was that the high varroa infestation levels in August of the H colonies did not seem to have an effect on the size of the colonies. We do not have a clear explanation for this result, but it is possible that the queens in H colonies had laid more eggs than queens in L colonies to compensate for the possible differential loss in population by early deaths of infested bees. If more bees died earlier in H colonies relative to L colonies, but queens in H colonies compensated the loss by laying more eggs, differences in colony sizes would not be apparent, but there would be fewer bees of foraging age in H colonies, which would explain the differences in productivity. However, the egg-laying rate of queens was not measured in this study and therefore that and other alternative hypothesis that could shed light on this result remain to be investigated.

The reasons as to why there were differences in mite infestation levels in the summer between the H and L groups of colonies are the subject of another ongoing study, and point towards differences in seasonal mite reproduction. A higher proportion of mites reproduced during the spring in the colonies of the H group than in the colonies of the L group (B.E., personal observation), although the underlying mechanisms of this distinct level of mite reproduction are unknown.

Further studies will be needed to more precisely establish the infestation levels at which V. destructor reduces honey yields at a cost higher than that of its control in honey bee colonies in different geographical regions, as well as to investigate about the underlying reasons as to why honey yields are reduced in infested colonies.