Native pollinator contributions to crop yield for two cultivars of highbush blueberry

Background/Question/Methods

Honeybees, commonly managed for crop pollination, are declining, and crop yield may be reduced as a result. Native pollinators may reduce pollination deficits, however, and so we assessed the contribution of native pollinators and honeybees to yield in 26 fields of highbush blueberry (Vaccinium corymbosum; 14 ‘Duke’ and 12 ‘Bluecrop’) in the Greater Vancouver area of British Columbia, Canada. To estimate pollination deficits, we measured proportion fruit set of pairs of branches on 10 stratified random plants at each of three distances from the most natural (‘bee friendly’) edge of fields (0m, 50m, and 100m). We added saturating amounts of pollen by hand to 10-20 flowers on one branch while the other served as an open-pollinated control; if pollen-supplemented branches had higher fruit set (assessed using ANOVA with field included as a random factor) we interpreted this as a pollination deficit. We also analyzed effects of cultivar and distance from natural edge on fruit set, using ANOVA.  We assessed the composition of the pollinator community by counting pollinator visits to blueberry flowers, identifying insects to the lowest taxonomic level possible on the wing, and determined the effect of visit rate by different pollinators (honeybees, native bees or total bees) on fruit set using multiple linear regressions.

Results/Conclusions

We found a pollination deficit in blueberry in our study region in 2011; fruit set was significantly higher for the supplemented inflorescences in both cultivars (‘Duke’ 2% deficit, F₃₅₈=3.95, P=0.048; ‘Bluecrop’ 13% deficit, F₃₁₁=54.48, P<0.001). Distance from the natural edge had no effect on proportion fruit set or pollinator visit rate. For both cultivars, proportional fruit set tended to increase with total number of observed visits (‘Duke’ R²=0.09, P=0.08; ‘Bluecrop’ R²=0.15, P=0.02).  However, a comparison between multiple linear regression models showed that proportion fruit set in ‘Duke’ was best predicted by both native and honeybees (adjusted R² = 0.17, P=0.01) while ‘Bluecrop’ was best predicted by native bees alone (adjusted R²=0.09, P=0.05). This difference may be due to morphological and phenological differences in flowering between the cultivars. Other unmeasured variables (degree of cross-pollination, management practices, etc.) could explain the low proportion of variance explained by pollinators in our models. Because they appear to contribute to fruit set, pollination by native pollinators should be encouraged. Future research will integrate information on landscape composition and on-field practices (pesticides, honeybee stocking rates, etc.) to understand the relative contributions of each to native pollinator abundance and fruit set.