Conservation Evidence strives to be as useful to conservationists as possible. Please take our survey to help the team improve our resource.

Providing evidence to improve practice

Action: Plant dedicated floral resources on farmland Bee Conservation

Key messages

Read our guidance on Key messages before continuing

Fourteen trials in Europe and North America have recorded substantial numbers of wild bees foraging on perennial or annual sown flowering plants in the agricultural environment.

Ten trials (eight replicated) have monitored bees foraging on patches sown with a high proportion of phacelia Phacelia tanacetifolia on farmland and all but one found substantial numbers of foraging wild (non-Apis) bees, particularly bumblebees Bombus spp. Six of these trials recorded the number of foraging bee species, which ranged from eight to 35. One replicated trial shows that phacelia is not very attractive to wild bees in Greece.

One replicated controlled trial in the UK showed that planted perennial leguminous herbs, including clovers, were more attractive to bumblebees in landscapes with a greater proportion of arable farming.

Four replicated trials have quantified the wider response of wild bee populations to planted flower patches by measuring reproductive success, numbers of nesting bees or numbers foraging in the surrounding landscape. One trial showed that planted patches of bigleaf lupine Lupinus polyphyllus in commercial apple orchards in Novia Scotia, Canada, significantly enhanced the reproductive success of blue orchard mason bees Osmia lignaria. One trial in the Netherlands showed that bee numbers and species richness are not higher in farmland 50-1,500 m away from planted flower patches. Two trials in Germany found no or relatively few species of solitary bee nesting on set-aside fields sown with phacelia or clover respectively.


Supporting evidence from individual studies


Williams & Christian (1991) planted three 9 m2 plots of phacelia at Rothamsted Research experimental farm, Hertfordshire, England. Seven species of bumblebee, including the long-tongued common carder bee Bombus pascuorum, and one cuckoo bumblebee B. [Psithyrus] vestalis foraged on the phacelia. Of observed worker bumblebee visits, 97% were for nectar, not pollen.


Patten et al. (1993) planted four 1.2 x 1.8 m plots of each of 17 flowering species next to commercial cranberry Vaccinium macrocarpon bogs in Washington State, USA. Five plant species attracted more than 30 bees/plot/count on average: catmint Nepeta mussini, borage Borago officinalis, phacelia, anise hyssop Agastache foeniculum and Korean mint A. rugosa. Short-tongued bumblebee species Bombus mixtus, B. occidentalis and B. sitkensis (cranberry pollinators) strongly preferred three plant species: bird's-foot trefoil Lotus corniculatus, Korean mint and anise hyssop (averages of 17, 23 and 19 bees/plot/count respectively) but did not visit borage or phacelia much (averages of 1 and 5 short-tongued bees/plot/count, respectively). Two long-tongued species, the Californian bumblebee B. californicus and B. caliginosus visited borage and phacelia in large numbers (>70 bees/plot/count).


As part of a larger study with 10 field types, Gathmann et al. (1994) placed bundles of reed Phragmites australis stems for cavity-nesting bees (and wasps) in four set-aside fields newly sown with phacelia in Kraichgau, southwest Germany. The phacelia fields attracted many honey bees Apis mellifera (foraging bees not quantified), but no cavity-nesting solitary bees made nests in reed stems in these fields. By comparison, in the same study, 12 bee species nested in reed stems placed in 2-year-old naturally regenerated set-aside fields mown in late June.


Engels et al. (1994) planted three strips of the commercially available Tubingen nectar and pollen mixture (40% phacelia, 25% buckwheat Fagopyron esculentum) at the edge of an arable field in Baden-Wurttemberg, Germany, over two years. Two strips were sown only in the first year, one strip was sown in both years. They recorded 58 species of wild bee either nesting in grooved board wooden nest boxes or foraging on the plots, including 11 species of true bumblebee Bombus spp. and five species of cuckoo bumblebee Bombus [Psithyrus] spp. Thirty-five bee species foraged on flowers from the Tubingen mixture.


Carreck & Williams (1997) planted two or three plots of two commercial nectar and pollen mixtures Tubingen Mixture (40% phacelia) and Ascot Linde Mixture (25% phacelia) on farmland in Hertfordshire, England. Across two years, the plots attracted 14 species of bee, including all six common UK bumblebee species and three cuckoo bumblebee species Bombus [Psithyrus]. A small number of solitary bees of three species (no more than two individuals on any plot) were recorded. Phacelia attracted 87-99% of all bee visits over the two years of this study. Buckwheat, a nectar source that comprised 20% of both seed mixtures by weight, attracted 1% or less of all bee visits.


Gathmann & Tscharntke (1997) monitored solitary bees and wasps nesting in reed stem nest boxes placed on three set-aside fields sown with a clover grass mix in Germany over three years. Relative to nest boxes placed in semi-natural grasslands, few species occupied these nest boxes (quantitative details are lacking from the report of this trial).


Carreck et al. (1999) recorded 15 species of bee visiting flowers over two summers, in four plots of six annual flowering plant species at Rothamsted Research, Hertfordshire, England. Short-tongued bumblebees buff-tailed and red-tailed (Bombus terrestris/lucorum and B. lapidarius/ruderarius) were the most abundant wild bee visitors, and bees were most numerous on phacelia, borage and (second year only) cornflower Centaurea cyanus.


In a replicated study of foraging bee communities on set-aside fields of different ages and management (four replicates of each) in Germany, Steffan-Dewenter & Tscharntke (2001) found that 1-year-old set-aside fields sown with phacelia had a similar abundance but fewer species of bee (13 species/field on average) than 1- to 5-year-old naturally regenerated set-aside fields (15-29 species/field). Bees found on phacelia were mainly common species of bumblebee and the solitary bee genus Lasioglossum, whereas several endangered and specialised bees were found foraging on naturally regenerated set-aside.


Carreck & Williams (2002) evaluated a sown mix of six annual flowering species: cornflower, common mallow Malva sylvestris (both native), borage, buckwheat, marigold Calendula officinalis and phacelia as forage for insects, in four plots at Rothamsted Research, Hertfordshire, England. The mix attracted 16 bee species, the most numerous insects being honey bee Apis mellifera and red-tailed bumblebee Bombus lapidarius/B. ruderarius (not distinguished in the study). 97% of all bumblebee visits were to phacelia and borage, 67% of all solitary bee visits were to marigold. The common carder bee B. pascuorum and garden bumblebee B. hortorum (both common long-tongued species) were recorded in relatively low numbers.


Fluri & Frick (2002) recorded 0.2 bumblebees/m2 (2,000 bumblebees/ha) foraging on a single 0.3 ha phacelia plot in Switzerland.


Dramstad et al. (2003) recorded numbers of bumblebees visiting a single 2 m x 210 m sown strip of phacelia, in Vestby, Norway, in 1994. They recorded a peak of 237 bumblebees on the strip (0.6/m2) on 17 July, which gradually declined to 93 bumblebees on the strip (0.2/m2) on 28 July.


In a replicated trial of flower-visiting insects foraging on six 0.1 ha sown patches of phacelia on farmland near Thessaloniki, Greece, 95% of all visits were by honey bees (Petanidou 2003). No bumblebees and only small numbers of solitary bees, mostly sweat bees of the family Halictidae (no more than six species in any flowering period, 12 species in total) foraged on the patches.


In a replicated, controlled trial in eastern and central England, Heard et al. (2007) showed that patches sown with a 20% legume seed mix (clovers Trifolium spp. and bird's-foot trefoil Lotus corniculatus) at eight sites attracted significantly higher densities of bumblebees than control patches of non-crop vegetation typical of the site (average 26 bumblebees/200 m2 on forage patches compared to 2 bumblebees/200 m2 on control patches). Honey bees Apis mellifera and cuckoo bumblebees (Bombus [Psithyrus] spp.) were not in greater densities on forage patches. The study also showed that bumblebee densities on sown forage patches were higher in areas with a greater proportion of arable land in a surrounding 1 km radius circle of landscape than in landscapes with less arable and more grassland, woodland and urban habitat. This demonstrates that planted leguminous forage is more valuable to bumblebees in intensive arable landscapes.


Sheffield et al. (2008) demonstrated that planting 3 x 45 m patches of the native bigleaf lupine Lupinus polyphyllus in three apple orchards in Novia Scotia, Canada, significantly enhanced the reproductive success of managed blue orchard mason bees Osmia lignaria. Nests were heavier and contained more new bees/nest in six nest boxes placed next to the lupine patch, relative to six nest boxes placed 600 m away from the lupine patches in the same orchards.


Kohler et al. (2008) planted 100 m2 patches of 17 perennial and annual flowering plant species at five locations on intensive farmland in the central Netherlands. They measured the abundance and diversity of bees during one summer at 10 sampling locations along a 1,500 m transect running away from each patch, and along five 1,500 m control transects. All the transects ran alongside ditches. Bee abundance and diversity were 60-80% higher than on control transects within the flower patches, but not anywhere else along the experimental transects. This suggests that patches of sown forage plants do not enhance numbers of bees in the surrounding landscape, at least in the first year. The lowest values for numbers of bees and bee species were recorded at the sampling point 50 m away from the flower patches.


Tuell et al. (2008) evaluated native perennial plant species in the eastern USA for their attractiveness to wild bees in a replicated experiment (five replicate 1 m2 plots of each species). Nine out of 43 species were highly attractive to bees, having an average of five or more wild bees per m2 plot in vacuum sampling or timed observation. These were Potentilla fruticosa, Scrophularia marilandica, Veronicastrum virginicum, Ratibida pinnata, Agastache nepetoides, Silphium perfoliatum, Lobelia siphilitica, Solidago riddellii and Solidago speciosa. Three other plant species (Zizia aurea, Fragaria virginiana and Coreopsis lanceolata) were identified as attractive to wild bees in the early season (May and June), a crucial time for early-emerging bee species, when flowers are generally less abundant.


A randomised, replicated, controlled trial on four farms in southwest England (Potts et al. 2009) found that 50 x 10 m plots of permanent pasture annually sown with a mix of legumes, or grass and legumes, supported more common bumblebees (individuals and species) than seven grass management options. There were twelve replicates of each management, monitored over four years. No more than 2.2 bumblebees/transect were recorded on average on any grassy plot in any year, compared to over 15 bumblebees/transect in both sown treatments in one year. The legumes sown included white clover Trifolium repens, red clover T. pratense, common vetch Vicia sativa and bird's-foot trefoil Lotus corniculatus.

Referenced papers

Please cite as:

Dicks, L.V., Showler, D.A. & Sutherland, W.J. (2010) Bee conservation: evidence for the effects of interventions. Pelagic Publishing, Exeter, UK