Introduction

Pesticide hazard and risk assessment for bees in the EU, USA or Australia have so far focused on managed western honey bees (Apis mellifera). However, honey bees may have different intrinsic susceptibility to pesticides than other bees. They may also be exposed in a different manner due to variations in behaviour and life history. Finally, bee populations may respond in various ways to pesticides because of distinct population dynamics. Consequently, the pesticide risk assessment procedures currently applied for managed honey bees are not necessarily directly applicable to other bees. Only recently have pesticide risk assessment methods for bees other than honey bees received more attention, but no clear consensus on risk assessment procedures has yet been established.

Extrapolation of honeybees to non-Apis bees

Susceptibility

The large majority of toxicity data has been generated for the western honeybee; relatively few toxicity data are available for non-Apis bees, such as bumblebees, stingless bees and solitary bees. As  a result, there is uncertainty to what extent honeybees can serve as a surrogate for non-Apis species.

A few recent reviews compared acute toxicity among bee species (e.g. Mommaerts & Smagghe, 2011; Arena & Sgolastra, 2014). They find that the relative toxicity of pesticides between honeybees and other bee species is not consistent, and depends on the pesticide and the type of bee. Bumblebees are often less sensitive to pesticides than honeybees, while stingless bees tend to be more sensitive, however variation is large.

Both Arena & Sgolastra (2014) and EFSA (2013) suggest that, if honeybee toxicity data are used as a surrogate for bumble bees and solitary bees, an extrapolation or safety factor of 10 may be used. This extrapolation factor would be protective of non-Apis bees in about 95% of the cases.

Exposure

The probability and degree of exposure to a pesticide may be very different from one bee species to another. Exposure will be determined by when and where bees forage, live and reproduce. Bee phenology (their seasonal development, adult emergence and activity patterns) and also their behaviour will determine exposure to pesticides when they are flying and visiting flowers, or nesting in a wide variety of localities.

For example, the exposure of bees that nest in the fields where pesticides are being applied is likely to be much higher than those nesting farther away. Bees that have large foraging ranges are likely to diminish their exposure by visiting a larger diversity of crops and flowers, some differing in pesticide load. Some bees may emerge and complete their life cycle almost entirely within the blooming period of a crop, making them completely exposed to all pesticide applications. Other bees may have more prolonged life cycles with reproduction taking place before crop bloom, and therefore experience less risk of pesticide exposure.

Many wild bees collect other materials than pollen and nectar, such as soil, wood, mud, resin, leaves or leaf hairs used to build nests, and they may also collect water, sap from wounded plants or sap feeding insects, or floral oils, used to feed larvae. When any such materials are contaminated with pesticide there is an additional risk factor.

Overall, it cannot be concluded that non-Apis bees are consistently more or less exposed to pesticides than honeybees. This will depend on the bee species and the local agronomic and ecological situation. More guidance on this topic is provided by Roubik and colleagues (2014), who review the life histories of various groups of non-Apis bees and the implication for exposure to pesticides.

Population dynamics

The population dynamics of a bee species will also affect the long-term survival of the population after exposure to a pesticide.

Honeybees and other highly social bees have colonies with thousands of individuals, whereas the large majority of wild bees are solitary. A solitary female bee mates, provisions nests, forages for resources and lays eggs. She will not be able to produce further offspring, if she succumbs to pesticide exposure. On the other hand, a honeybee colony is capable of continuing to produce more bees, despite the loss of individual workers, or even the queen.

Fecundity of wild bees tends to be far more limited than that of honeybees. A honeybee queen that has mated with 10-30 males has a lifetime supply of sperm to fertilize eggs during her life of at least a year. The queen of a stingless bee, however, mates with a single male. The range for less social or solitary bees may be seen, for example, among the small colonies or groups of African carpenter bees (Xylocopa spp.) which only produce 1 to 5 offspring per generation, and European mason bees (Osmia rufa) that produce up to 20 offspring. This contrasts with up to several tens of thousands of offspring per queen for Apis mellifera. Fewer offspring mean that after bee mortality caused by a pesticide, population recovery will be slower.

Again, no consistent difference exists between the population dynamics of honey bees and non-Apis bees and its resulting influence on pesticide effects.

Risk factors

Different factors will characterize the risk of a pesticide to non-Apis bees. A table of key risk factors determining the risk of pesticides to bees is provided, which can be used by a registrar to compare the locally relevant non-Apis bee(s) with the honeybee (for which more detailed risk assessments are generally available).

A more detailed set of risk factors is discussed by Van der Valk et al (2012).

Risk assessment

Two main approaches are presently taken for risk assessment of pesticides to non-Apis bees.

While it is recognized that the use of honeybees as a surrogate for other insect pollinators has limitations, it is assumed that data on individual honeybees as well as colony level data will provide some relevant information on the potential effects of a pesticide on both solitary bees as well as social bees.

If additional data are available on other bee species, these data can be included in the tiered risk assessment process as an additional line of evidence.

This approach is taken by regulators in North America (US-EPA et al., 2014)

In spite of the fact that pesticide risk assessment for bumblebees and solitary bees has not been routinely carried out, and uncertainties exist about inter-species sensitivity and exposure, a risk assessment methodology is proposed. This methodology is based on the best presently available information on bumble bees and solitary bees, and further applies processes that have been described previously for the honeybee. 

To allow for the fact that bumble bees and solitary bees are potentially more vulnerable to pesticides than honey bees, extra safety factors are incorporated in the risk assessment.

This approach is taken in Europe (EFSA, 2013)

Arena M & Sgolastra F (2014) A meta-analysis comparing the sensitivity of bees to pesticides. Ecotoxicology 23: 324–334

EFSA (2013) EFSA Guidance Document on the risk assessment of plant protection products on bees (Apis mellifera, Bombus spp. and solitary bees). European Food Safety Authority. Available here

Mommaerts V & Smagghe G (2011) Side-effects of pesticides on the pollinator Bombus: An overview. In: Pesticides in the Modern World – Pests Control and Pesticides Exposure and Toxicity Assessment, Stoytcheva M (ed.). InTech. Available here

Roubik DW (ed.) (2014) Pollinator safety in agriculture. Food and Agriculture Organization of the United Nations. Available here

US-EPA et al. (2104) Guidance for Assessing Pesticide Risks to Bees. United States Environmental Protection Agency, Health Canada & California Department of Pesticide Regulation. Available here

Van der Valk H, Koomen I, Nocelli RCF, Ribeiro M de F, Freitas BM, Carvalho S, Kasina JM, Martins D, Mutiso M, Odhiambo C, Kinuthia W, Gikungu M, Ngaruiya P, Maina G, Kipyab P, Blacquiere T, Van der Steen J, Roessink I, Wassenberg J & Gemmill-Herren B (2012) Aspects determining the risk of pesticides to wild bees: risk profiles for focal crops on three continents. Julius-Kuhn-Archiv, 437. Available here

[Toolkit Reference A09-03-01-Y]