Findings: Long-term population trends for the honey bee, the most important managed pollinator, are demonstrably downward. Similar data are not available for other managed pollinators, such as alfalfa leafcutting bees and bumble bees.
Among the various pollinator groups, evidence for decline in North America is most compelling for the honey bee, Apis mellifera. Honey bees enable the production of no fewer than 90 commercially grown crops, and beekeeping is a large commercial industry that leases honey bee colonies for pollination services across the continent.
Since 1947, the U.S. Department of Agriculture’s (USDA’s) National Agricultural Statistics Service (NASS) has tracked honey bee colonies managed by beekeepers in the United States. Statistics demonstrate declines in 1947–1972 and 1989–1996, and a recent drop in 2005. Reports from industry journals suggest higher rates of winter kill in honey bee colonies since the advent of the parasitic mite Varroa destructor in the 1980s, causing temporary shortages of healthy honey bee colonies (for early season almond pollination) that are not captured by the NASS data. However, putting those declines into context is complicated by the peculiarities of NASS data collection. Because its annual survey focuses on honey production and pollinating colonies are not monitored unless they also produce honey, there are limits on the extent to which those data can be extrapolated to inform population estimates. NASS methods result in undercounting because the annual survey group consists of beekeepers with five or more hives; there is no mechanism to count hobbyist beekeepers who might nevertheless contribute to the supply of honey-producing or pollinating colonies. Moreover, because surveys do not consider that some honey-producing colonies travel—they are leased in different regions of the country for different seasons—these colonies can be counted more than once.
NASS also conducts a 5-year census of agriculture survey that counts all honey bee colonies just once, but definitional differences make the data incompatible with data from the annual honey survey. Yet another complicating factor is that no surveys account for colony health or for intrayear volatility in colony numbers. (Colonies that die early in the year, when they are critically needed for pollination, can be replaced by purchasing packages of bees or splitting surviving colonies later in the year.) Finally, there is an additional complication for assessing the supply of honey bee pollinators in North America. U.S. data collection does not match what is done in Canada or Mexico. Canadian data are collected on all honey bee colonies, whether they are kept for pollination, for honey production, or both. Mexico has