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: Provide artificial nest sites for solitary bees Bee Conservation

Key messages

Read our guidance on Key messages before continuing

We have captured 30 replicated trials of nest boxes for solitary bees in 10 countries, including Europe, North and South America and Asia. Twenty-nine of these trials showed occupancy by bees. Many species of solitary bee readily nest in the boxes, including some species considered endangered in a study on farmland in Germany, oil-collecting species of the genus Centris in South America and a recently discovered species in lowland tropical forest in Costa Rica. One trial in temperate forest in Canada recorded no bees using nest boxes.

A set of replicated experiments in Germany estimated that four medium to large European species of solitary bee have a foraging range of 150 to 600 m, so nest boxes must be within this distance of foraging resources.

Twenty-three replicated trials have shown nest boxes of cut hollow stems or tubes being occupied by solitary bees. Eleven trials demonstrated occupation of blocks of wood drilled with holes. Two trials in Neotropical secondary forest (one in Brazil, one in Mexico) showed that particular solitary bee species will nest in wooden boxes, without stems or confining walls inside.

Two replicated trials have compared reproductive success in different nest box designs. One showed that reed stem and wooden grooved-board nest boxes produced more bees/nest than four other types. Nest boxes with plastic-lined holes, or plastic or paper tubes were much less productive, due to parasitism or mould. The other, a small trial, found nests of the oil-collecting bee Centris analis in Brazil were more productive in cardboard straws placed in drilled wooden holes than in grooved wooden boards stacked together.

Three trials on agricultural land, one on a carpenter bee in India, one on a range of species in Germany and one on species of Osmia in the USA, have shown that the number of occupied solitary bee nests can double over three years with repeated nest box provision at a given site.

One small replicated trial compared populations of solitary bees in blueberry fields in the USA with and without nest boxes over three years. The estimated number of foraging Osmia bees had increased in fields with nest boxes, compared to fields without nest boxes.

Eleven replicated trials have recorded solitary bees in nest boxes being attacked by parasites or predators. Rates of mortality and parasitism have been measured in 10 studies. Mortality rates range from 13% mortality for cavity-nesting bees and wasps combined in Germany (2% were successfully parasitized), or 2% of bee brood cells attacked in shade coffee and cacao plantations in central Sulawesi, Indonesia, to 36% parasitism and 20% other mortality (56% mortality overall) for the subtropical carpenter bee Xylocopa fenestrata in India.

Two replicated trials of the use of drilled wooden nest boxes by bees in California, USA, showed that introduced European earwigs Forficula auricularia and introduced European leafcutter bee species use the boxes. In one trial, these introduced species more commonly occupied the boxes than native bees.

A small trial tested three soil-filled nest boxes for the mining bee Andrena flavipes in the UK, but they were not occupied.

Supporting evidence from individual studies

Do solitary bees nest in nest boxes?


Red mason bees Osmia rufa readily occupied artificial nest boxes comprising metal food cans filled with drinking straws (straw diameter 5-7 mm; Free & Williams 1970). In the first year of a trial, 349 cans were recovered from 20 sites in southern England; of these 44 (13%) had one or more straws occupied by a red mason bee nest. Over the following two years, there was a tendency by this species to reoccupy cans. Osmia caerulescens and species of Megachile also occupied the cans.


The subtropical carpenter bee Xylocopa fenestrata, a valuable pollinator of cucurbits and other plants, has been shown to nest readily in cut stems of castor Ricinis communis or sarkanda Arundo sp. bundled together (Sihag 1993a). In a trial on agricultural land in Haryanar, India, these bees strongly preferred stems cut to 23-30 cm long, with an internal diameter of 10-12 mm. The number of occupied stems increased from 120 in the first year (1984) to 350 two years later (1986), from a total of 20,000 stems placed out over three years.


In April 1990, in Kraichgau, southwest Germany, 240 bundles of reed stems Phragmites australis in tins were put out, six in each of 40 fields of 10 management types, including various types of set-aside, crop fields and old meadows (Gathmann et al. 1994, also referred to by Tscharntke et al. 1998). Of 43,200 available stems, 292 were occupied by a total of 14 bee species and nine wasp species. Five species of bee considered to be endangered in Germany occupied the reed stem nests: Anthidium lituratum, Heriades crenulatus, Megachile alipcola, Osmia gallarum and Osmia leaiana. The two endangered Osmia species were exclusively found in nests in old meadows (more than 30 years old with several old fruit trees). The other three also nested in stems provided in 2-year-old mown set-aside, and two species (A. lituratum and M. alpicola) used reed stems in a variety of field types, including cereal crops.



Scott (1994) placed a total of 9,216 wooden nest boxes with small drilled holes of diameters 4.5-11 mm, on the edges of open fields in Upper Michigan USA, in April 1984 and 1985. Three species of the small solitary bee genus Hylaeus used the boxes, with an overall occupancy rate of 4%. These bees only used holes with the smallest diameters (4.5, 5.2 and 6.0 mm). H. ellipticus preferred the smallest 4.5 mm holes. Hylaeus basalis preferred nest boxes at lower heights 0.1 and 0.4 m above the ground and H. verticalis preferred higher boxes set at 1.1 m.


A six-year trial at two experimental farms near Poznan, western Poland demonstrated that the red mason bee Osmia rufa readily nests in bundles of reed stems 7-8 mm in diameter (Wojtowski et al. 1995).


From January 1990 to May 1991, the orchid bee Euglossa atrovenata made 60 nests in 50 wooden boxes placed on a table (1 m above the ground) in secondary forest planted with coffee crops, at Unión Juárez Chiapas, Mexico (Rami­rez-Arriaga et al. 1996). The nests were made on the internal floor and walls of the boxes, constructed with resin.


In a replicated trial in central Germany from 1994-1996, 150 reed stem nest boxes (plastic tubes filled with 150 ´ 20 cm lengths of reed stem) placed at 15 different sites were occupied by 13 species of bee, 19 species of wasp and 17 species of parasite and parasitoid (Gathmann & Tscharntke 1997, also referred to by Tscharntke et al. 1998). In total, 8,303 nests were made.


In a replicated trial in Washington County, Maine, USA, Stubbs et al. (1997) added 50 drilled wooden nest boxes to each of three blueberry fields Vaccinium angustifolium over three years. The nest boxes each had 14 holes and were attached to trees along the field edge, at a height of 1.4 m, with 22-33 m between each box. In the first year, 30 nest boxes were occupied by bees of the genus Osmia, with 120 nests made. The number of nests increased the following year in all three fields. Between 3 and 11.5% of nesting holes were occupied at all three sites, each year.


Thirty to 45 drilled pine wood solitary bee nest boxes were suspended from valley oak Quercus lobata trees on the Cosumnes River Preserve, California, USA in 1989 and 1990 (Barthell et al. 1998). The boxes each had twelve 10 cm deep holes, 0.5, 0.65 or 0.8 cm in diameter. In both years, the European earwig Forficula auricularia was the most common occupant (59-85% of all occupied nests), followed by two introduced leafcutter bee species Megachile rotundata and M. apicalis (19.6% of all occupied nests in 1989, 3.4% in 1990). Four native bee species also occupied the boxes, but in much lower numbers. Megachile angelarum was found in 3.2-3.8% of occupied nests. M. fidelis, M. gentilis and Osmia texana occupied less than 1% of occupied nest boxes in both years.


Frankie et al. (1998) recorded 23 species of bee, mostly from the genera Megachile and Osmia, using nine drilled wooden nest boxes on each of six woodland, shrubland and riparian reserves over three years in northern central California, USA. Three non-native species of Megachile nested in the boxes - M. apicalis, M. rotundata and M. concinna. The former two species were common, but M. concinna was uncommon, recorded less than 12 times overall.


A series of four trials between 1990 and 1996 in Germany documented uptake of reed bundles placed in tins or plastic tubes attached to wooden posts (Tscharntke et al. 1998). Across a variety of agricultural and semi-natural habitats including orchard meadows, old hay meadows, set-aside fields, field margins and chalk grasslands, a total of 33 bee species (not including parasitic bees) used the nests.


Morato & Campos (2000) recorded 14 species of solitary bee using drilled wooden nest boxes in continuous tropical forest and inside and between forest fragments in Amazonas, Brazil. At least 108 nest boxes, each with two holes, were placed at each site. The nest boxes were more frequently occupied in continuous forest (23-29 nests/site) and natural gaps in continuous forest (78 nests/site) than in between forest fragments in pastureland or secondary vegetation (6-23 nests/site).


A trial of 120 reed stem nest boxes at 15 different agricultural sites near Göttingen in Lower Saxony, Germany in 1997 found the boxes occupied by 11 species of bee (Steffan-Dewenter 2002). The red mason bee Osmia rufa and the common yellow face bee Hylaeus communis were the most widespread and common nest box occupants in this study.


In the same study, separately reported (Gathmann & Tscharntke 2002), nest boxes had a 50% chance of being occupied by two specialised (oligolectic) species of bee - Chelostoma rapunculi and Megachile lapponica - at a distance of 256-260 m from a patch of their required forage plants. There was no colonization of nest boxes by C. rapunculi more than 300 m from a patch of its food plant, bellflowers Campanula spp..


Gathmann & Tshcarntke (2002) used translocation experiments to estimate that female solitary bees of four medium to large European species - Andrena barbilabris, A. flavipes, A. vaga and the red mason bee Osmia rufa - have a maximum foraging range between 150 to 600 m, so nest boxes have to be placed within this distance of forage resources.


From 1998-1999, Steffan-Dewenter & Leschke (2003) recorded 17,278 cells from 13 species of solitary bee using 540 reed stem nest boxes placed in 45 orchard meadows in central Germany.


In 1998, Steffan-Dewenter & Schiele (2004) placed bundles of common reed stems (153 stems per bundle, cut 15-20 cm long) in 10-13 cm diameter plastic tubes, attached to wooden posts, in orchard meadows in Germany. These were used as nest sites by the red mason bee Osmia rufa. Three years later, in autumn 2001, a total of 974 newly developed females were counted in 60 such nests, over five orchard meadow sites.


A study using bamboo stem nest boxes from 1994-1997 at the University of Sao Paulo-Ribeiro Preto, Sao Paulo, Brazil (Augusto & Garofalo 2004) recorded 5% uptake of stems by the euglossine bee Euglossa townsendi. A total of 383 bamboo stems were placed on outdoor shelves on a University campus, in bundles of eight to 11. Those used by female bees were 11.9 to 28.1 cm long, and 1.1-2.2 cm in internal diameter.


Three types of nest box were placed in 20 urban gardens in Sheffield, UK, from 2000‐2002. They were occupied by two bee species – Hylaeus communis (10 gardens) and Osmia rufa (two gardens). The most frequently used were those constructed of 20 cm lengths of bamboo stem in plastic pipe, and 4 mm or 6 mm diameter holes drilled into wooden blocks, with uptake in over 50% of gardens over three years (Gaston et al. 2005). Tin cans filled with paper drinking straws (4‐6 mm diameter) and 8‐10 mm holes drilled in wood were less well‐used.



Six different nesting materials for the red mason bee Osmia rufa were tested at an agricultural experimental station in Poznan County, Poland, in 2000 and 2001 (Wilkaniec & Giejdasz 2003). For each trial, 150 nests of each of the following materials were tested: reed stems, plastic tubes, paper tubes (bundles), wood, cork (grooved boards joined together in blocks), and holes drilled into wood, lined with printer acetate. All materials were used by female bees, but the highest production of bees per nest was from reed stems (3.5 bees/nest in 1999) or wood (7.2 bees/nest in 2000). Nests in paper tubes were all parasitized. Nests in plastic were well occupied (80-100%) but had a low success rate (0.2-1.8 bees/nest), partly due to mould.


Three species of wild megachilid bee (Megachile spp.) nested in boxes made from blocks of pine wood drilled with 14 mm long, 8 mm diameter holes, in a small replicated trial in Arizona USA (Armbrust 2004). Three sites were in the Tucson Mountains, one site on undeveloped land within Tucson city. Four nest boxes, each with 33 holes, were placed at each site. Overall 34% of available nest holes were filled between May and July 2001, but only six nests (4.5% of the available holes) were constructed in boxes at the urban site.


Nest boxes made of 20 cm lengths of common reed Phragmites australis and Japanese knotweed Reynoutria japonica, with internal diameters from 2-20 mm, were readily occupied by the megachilid bee Heriades fulvescens, in a replicated trial on 24 coffee agroforestry systems in Sulawesi, Indonesia (Klein et al. 2004). In total, 671 nests were constructed in 240 nest boxes, over a 14-month period from 2001-2002. Four other species of Megachilidae also used these nest boxes.


A trial in secondary woodland on Santa Catarina Island, Brazil (Zillikens & Steiner 2004) showed that leafcutter bees of the species Megachile pseudanthidioides will nest in wooden boxes with an internal cavity (10 ´ 10 ´ 5 cm with a 10 mm diameter entrance hole), or drilled hardwood blocks (holes 7 cm long, 11-12 mm in diameter) or sections of bamboo stem (15-25 cm long, 5-25 mm in diameter).


Thiele (2002) recorded the recently discovered solitary bee Duckeanthidium thielei nesting in 11 and 13 mm diameter drilled holes in 24 hardwood nesting blocks placed in lowland tropical forest in Costa Rica. The species is known only from Costa Rica and considered to be rare.


In a separate report of the same study (Thiele 2005), 16 species of solitary bee were recorded nesting in 24 hardwood nesting blocks, each with 80 drilled holes, in Costa Rican lowland tropical forest. Most nests were made in boxes placed in the canopy of dead trees, 21-37 m high (69% of all nests in the first year). The author stresses the importance of retaining dead standing emergent trees for bee conservation in this habitat.


Tylianakis et al. (2005) placed 432 reed stem nest boxes across 48 plots in agricultural areas of Manabi Province, southwest Ecuador, on posts or hanging from trees 1.5 m above the ground. Traps were monitored from June 2003 to May 2004. In total, 31 species of bee and wasp used the traps, with averages between 8 and 12 species per plot over the entire year for different land use types. The number of bee species is not specified.


Taki et al. (2008) put six cardboard tube/milk carton nest boxes in each of eight forest sites in Norfolk County, southern Ontario, Canada from summer 2003 to November 2004. They recorded 12 wasp species and two species of parasitic wasp, but no bee species nesting in the boxes.


In a replicated trial in two fragments of semi‐deciduous tropical forest in the State of Sao Paulo, Brazil, Gazola & Garofalo (2009) reported 16 species of solitary bee using nest boxes comprising bamboo stem sections or cardboard tubes. Overall, 2,708 cardboard tubes inserted in drilled wooden blocks, and 1,619 sections of bamboo cane were placed out for two years from 2000 to 2002. A total of 528 bee nests were recovered.


Oliveira & Schlindwein (2009) reported the use of nest boxes made with cardboard straws inserted into drilled wooden blocks or grooved wooden boards by the oil-collecting bee Centris analis in orchards in Pernambuco, Brazil. Seventeen nests were made in five wooden blocks with 40 cardboard-lined drilled holes in each (8.5% occupancy). Forty-eight nests were made in 12 nest boxes made of grooved wooden boards (10 groove holes per box) stacked together (40% occupancy). Cardboard straw nests had more brood cells (average 3.8 cells/nest) than nests in grooved boards (2.3 cells/nest). The species showed a preference for cavities with internal diameters between 6 and 10 mm.


Roubik & Villanueva-Gutiérrez (2009) monitored solitary bees using drilled wooden nesting blocks in the Sian Ka'an Biosphere Reserve, Quintana Roo, Mexico, during two 4-year stretches within a 17 year time period (1988 to 2005). Twenty 5 ´ 10 ´ 15 cm blocks, each with 12 holes 7 cm long, were placed at each of five sites. The boxes were used by solitary bees from at least five genera, with the most common occupants being Megachile zaptlana and the oil-collecting bee Centris analis.


Sobek et al. (2009) document three species of host bee (Hylaeus communis, Hylaeus confusus and Megachile ligniseca) and one parasitic bee (Coelioxys alata) using twelve reed stem nest boxes placed at each of 12 broadleaf woodland sites in the Hainich National Park, Thuringia, Germany. Bees made up only 9% of host cells (347 cells), with the majority of nest occupants being wasps. The abundance of nesting insects was higher in nest boxes in the canopy than in nest boxes mounted on wooden posts at chest height.

Do solitary bee nest boxes enhance local populations?


Numbers of the subtropical carpenter bee Xylocopa fenestrata nesting in cut stems of castor Ricinis communis or sarkanda Arundo sp. bundled together increased from 120 in the first year (1984) to 350 two years later (1986), in a trial on agricultural land in Haryanar, India (Sihag 1993a).


In a replicated trial on field margins, set aside fields and extensively managed meadows in central Germany from 1994-1996, 150 reed stem nest boxes (plastic tubes filled with 150 ´ 20 cm lengths of reed stem) were placed at 15 sites (Gathmann & Tscharntke 1997). The number of occupied stems almost doubled over three years from 1,761 in 1994 to 3,326 in 1996.


In a small replicated trial in Washington County, Maine, USA, Stubbs et al. (1997) added 50 drilled wooden nest boxes to three experimental blueberry fields Vaccinium angustifolium over three years from 1993 to 1995. The percentage of holes occupied rose from around 3% in the first year to over 7% in the third year in two fields, but did not rise substantially in the third field, remaining between 5 and 7%. Numbers of bees of the genus Osmia foraging on blueberry flowers in the experimental fields and in three control fields without nest boxes were monitored, using quadrat counts and sweep net sampling. In the first year, estimated numbers of Osmia ranged from 0 to 879 bees/ha in both control and experimental fields. In the third year, control fields had between 0 and 440 bees/ha, while experimental fields with nest boxes had from 219 to 1328 bees/ha. The numbers of foraging bees had increased in two of the three fields with nest boxes.

Are solitary bees in nest boxes attacked by predators and brood parasites?


From 1988-1991, Bosch (1992) recorded rates of parasitism for the orchard bee Osmia cornuta, nesting in nest boxes comprising paper straws in milk cartons, drilled holes in wooden blocks, grooved wooden boards, or bundles of reed stem. Overall parasitism rates ranged from 5-18% of cells in wild populations, and 0.1-13% of cells in managed populations.


Sihag (1993b) measured rates of mortality and parasitism for X. fenestrata nesting in stems of castor or sarkanda provided on agricultural land in Haryana, India, from 1985-1987. Up to seven generations a year were reared in stems 23-27 cm long, 10-12 mm in diameter. Cell numbers peaked in early summer (April to mid-May) and early autumn (late September to October). Mortality rates were highest in the late summer (mid-May to late-July), when parasitism was 33-36% and mortality 6-20%. Only two brood parasite species were seen, and the wasp Monodontomerus obscurus was responsible for 90% of the parasitism. Other larval mortality was caused mainly by ants and other predators. Birds, lizards and rodents destroyed some nests in domiciles not protected by wire cages (not quantified). No mortality caused by fungal or bacterial agents was observed.


In April 1990, in Kraichgau, southwest Germany, 240 bundles of reed stems in tins were put out, six in each of 40 fields of 10 management types, including various types of set-aside, crops fields and old meadows (Gathmann et al. 1994). The proportion of larvae in the nests that died from disease or failed parasitism was 13%; 2% were successfully parasitized.


In January 1990, 22% of nests of the orchid bee Euglossa atrovenata in wooden boxes were parasitized by the Megachilid brood parasite Coelioxys costaricensis, in secondary forest planted with coffee crops, at Union Juarez Chiapas, Mexico (Ramirez-Arriaga et al. 1996).


Two studies between 1990 and 1996 in Germany documented predation and parasitism rates in reed bundles in tins or plastic tubes attached to wooden posts in various semi-natural and agricultural habitats (reported in Tscharntke et al. 1998). The percentage of bees and wasps killed by predators or parasites was 21 or 28% on average.


In a trial of 120 reed stem next boxes in lower Saxony, Germany in 1997 (Steffan-Dewenter 2002), 14% of bee brood cells were attacked by natural enemies (brood parasites, parasitoids or predators).


Armbrust (2004) placed four nest boxes made from blocks of pine wood drilled with 14 mm long, 8 mm diameter holes at each of four sites in or near Tucson, Arizona, USA for one summer. Overall, 34% of nest holes were filled, by three species of Megachilidae. Of the filled nests examined, 27% were subsequently occupied by parasites or predators.


Klein et al. (2006) found 25 species of natural enemy attacking 14 bee and wasp species nesting in nest boxes in shade coffee and cacao plantations in central Sulawesi, Indonesia. 2.1% of bee brood cells were attacked in this study, although no predators or parasitoids were recorded for the most commonly found bee, the megachilid Heriades fulvescens, which made up 20% of all brood cells.


Gazola & Garofalo (2009) report five and 13 species of parasite attacking bees nesting in bamboo stem and cardboard tube nest boxes respectively, in two different fragments of semi-deciduous tropical forest in the State of Sao Paulo, Brazil.


Sobek et al. (2009) record parasitism rates of 16% of solitary bee and wasp cells made in canopy nest boxes and 13% of cells in understorey nest boxes, during a five-month study in Hainich National Park, a semi-natural broadleaf forest in Thuringia, Germany.

Do artificial nest sites for ground-nesting bees work?

The alkali bee Nomia melanderi, endemic to arid and semi-arid regions of the western USA, nests in dense aggregations in patches of salty soil known as alkali flats. Artificial nesting sites, or bee beds have been successfully maintained for over 35 years in the United States, for the purpose of alfalfa pollination (Torchio 1987). Brief instructions for creating a bee bed are contained in Torchio's review: the bed is lined with a 40 cm deep gravel layer to hold water, covered with 1 m of soil mixed with salt and seeded with soil cores from active nesting areas. We have not sought evidence on the take-up rates, or effects of management regimes on these artificial nest sites, because the alkali bee is a managed pollinator with very specific requirements.


A small-scale study with three replicates tested soil filled nest boxes for the mining bee Andrena fulvipes, a host of the Nationally Scarce dotted bee-fly Bombylius discolor in the UK. These nest boxes were not occupied, despite being placed alongside active colonies of nesting bees (Gibbs 2004).

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