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: Rear declining bumblebees in captivity Bee Conservation

Key messages

Read our guidance on Key messages before continuing

We have captured 22 trials from 13 countries documenting captive rearing of bumblebee colonies by confining mated queens alone (eight trials), with one or more bumblebee workers (seven trials), honey bee workers (one trial), male bumblebee pupae (three trials) or following anaesthetisation with CO2 (four trials). One trial found that over four years of artificial rearing, Bombus terrestris queens gradually decreased in weight.

Three trials have tried to rear North American bumblebees now declining or thought to be declining. Two induced spring queens of the half-black bumblebee B. vagans to rear adults in captivity, one trial induced queen yellow-banded bumblebees B. terricola (attempted in all three trials) and red-belted bumblebees B. rufocinctus (only attempted in one trial) to rear adults in captivity. All three trials tried to rear the yellow bumblebee B. fervidus and in all cases the queens laid eggs but the larvae died before becoming adults. One trial found the same pattern for the rusty-patched bumblebee and the American bumblebee B. pensylvanicus. One study reports rearing the large garden bumblebee B. ruderatus, a Biodiversity Action Plan species in the UK.

Two trials have reported laboratory rearing of a pocket-making bumblebee, the Neotropical B. atratus.

Three replicated trials demonstrated that the pollen diet of captively reared bumblebees influences reproductive success. In one trial, buff-tailed bumblebee B. terrestris colonies fed on freshly frozen pollen produced larger queens that survived better and produced larger colonies themselves than colonies fed on dried, frozen pollen. Two replicated trials demonstrated that B. terrestris workers can produce more offspring when fed types of pollen with a higher protein content.

Two replicated experiments showed that an artificial light regime of eight hours light, 16 hours darkness, can reduce the time taken for artificially reared queen B. terrestris to lay eggs, relative to rearing in constant darkness.

We have captured two replicated trials examining the effect of different artificial hibernation regimes in B. terrestris. One showed that hibernation of queens at 4-5°C for 45 days enhanced egg-laying and colony formation rates, but hibernated queens produced smaller colonies than non-hibernated queens. The second showed that queens should weigh more than 0.6 g (wet weight) and be hibernated for four months or less to have a good chance of surviving.

Supporting evidence from individual studies

Standard rearing conditions now used in commercial bumblebee-rearing facilities are complete darkness, 28 (± 1)°C and 60 (± 5)% relative humidity. The bees are regularly supplied with freshly frozen pollen collected by honey bees and sugar syrup in the ratio 1:1 with water, by volume.

How can queen bumblebees be induced to form colonies in captivity?


Sladen (1912) reared more than eight colonies of the buff-tailed bumblebee Bombus terrestris and one of the red-tailed bumblebee B. lapidarius by confining one or two nest searching queens with between two and seven workers of the same species in wooden boxes supplied with honey and pollen. In the case of the red-tailed bumblebee, the queen was also confined with clusters of cocoons from another nest.


Plath (1923) induced six different species of native North American bumblebee queens to lay eggs and rear colonies of adults, by confining spring queens with one to three bumblebee workers in dark wooden nest boxes supplied with honey bee-collected pollen and diluted honey. These six species included the half-black bumblebee B. vagans, thought to be declining in the USA. Five other species treated the same way laid eggs but did not rear colonies because the larvae died. Four of the species that could not be reared are also declining or thought to be declining in the United States: the rusty-patched bumblebee B. affinis, the yellow-banded bumblebee B. terricola, the American bumblebee B. pensylvanicus and the yellow bumblebee B. fervidus. The latter two species are reported to be pocket-makers (Kearns & Thomson 2001).


Frison (1927) induced nine different species of native North American bumblebee queens to lay eggs, by confining broody spring queens (already secreting wax) singly or in pairs, in wooden boxes in the dark. Fresh honeybee pollen and diluted honey solution were supplied. Colonies were reared to produce adult workers in 11 of the 46 trials between 1917 and 1920, including by the half-black bumblebee B. vagans, thought to be declining in the USA. Two other species reported to be declining: the yellow bumblebee B. fervidus and the yellow-banded bumblebee B. terricola were induced to lay eggs but did not rear colonies. The larvae died. No eggs were laid in two experiments with the American bumblebee B. pensylvanicus.


In Sweden, Hasselrot (1952) induced hibernated spring queens of three bumblebee species (B. terrestris, the tree bumblebee B. hypnorum and B. lapidarius) to form colonies in 26 out of 30 wooden nest boxes. He confined them alone in two linked boxes and provided honey solution and fresh pollen, moss and cellulose nesting material.


Plowright & Jay (1966) induced mated queens of seven Canadian bumblebee species to found colonies in captivity, by confining them singly or in pairs in wooden boxes kept at 21° or 29° C, regularly provided with fresh pollen and honey solution. Twenty-eight of the 30 B. terricola tested and four of the nine red-belted bumblebees B. rufocinctus reared adults using this method, but a single yellow bumblebee B. fervidus did not. Some queens confined in waxed paper cartons laid eggs, but none successfully reared adult workers.


Pomeroy & Plowright (1980) described two hive designs in which they had reared several species of bumblebee in the laboratory, including the pocket-making Neotropical species B. atratus and the large garden bumblebee B. ruderatus. Both designs were internally cone-shaped, one made of metal, one moulded from porous concrete. The metal hive had a heating element, and its internal temperature controlled at around 30°C by thermostat.


In a replicated trial in Germany, Röseler (1985) demonstrated that mated queen B. terrestris can be induced to lay eggs by anaesthestising them with CO2 for 30 minutes on two consecutive days. After this treatment, 73% of unhibernated and 81% of hibernated queens began egg-laying within one week of confinement with workers.


Two replicated laboratory trials in France (Tasei & Aupinel 1994, Tasei 1994) showed that an eight hour light, 16 hour dark regime imposed during rearing induces egg-laying more quickly (average 33 and 21 days respectively) in B. terrestris queens than constant dark (average 47 and 39 days), or, in one set of experiments, constant light (average 59 days to egg-laying). Both experiments found that the light:dark regime did not significantly affect the chance of a B. terrestris queen laying eggs (range 61-73% for all treatments). Tasei & Aupinel (1994) used 103 artificially hibernated laboratory reared queens confined alone in standard rearing conditions. Tasei (1994) used 200 non-hibernated laboratory reared queens anaesthetized with CO2 and confined with one B. terrestris worker.


Beekman et al. (1998) tested the effects of different artificial hibernation regimes (temperatures from -5 to 15°C, durations from 1 to 8 months) on 2,210 queen B. terrestris from laboratory-reared colonies in the Netherlands. A queen's initial weight and the duration of hibernation strongly affected survival, but the temperature did not. Queens should weigh more than 0.6 g (wet weight) and be hibernated for four months or less to have a good chance of surviving. Queens weighing less than 0.6 g before hibernation did not survive, but above this threshold, initial weight did not affect survival. Few queens survived hibernation periods of 6 and 8 months (8.5%, compared to 83% of queens hibernated for one, two and four months). Neither temperature, weight nor length of hibernation affected a queen's ability to lay eggs after surviving hibernation.


Beekman et al. (2000) reared B. terrestris in the laboratory over four years, with one to three generations per year, starting with the progeny of 47 wild-caught queens in 1993. A total of 170 colonies were reared altogether. Queens were mated in mating cages, hibernated for two to four months and induced to form colonies by confining with two to four honey bee workers in standard rearing conditions. They found a significant linear decrease in average queen weight over time, from 0.83 g in 1993 to 0.73 g in 1996. Since queens weighing less than 0.6 g do not survive hibernation, this change would be of concern in the context of captive-bred releases. Beekman et al.'s experimental results suggest it is caused by a nutrient deficiency, rather than inbreeding or reallocation of resources within colonies.


Yeninar et al. (2000) reared 96 colonies of the Mediterranean subspecies B. terrestris dalmatinus from nest-searching queens caught in the wild in Turkey. This Mediterranean subspecies aestivates over the dry season from June to October, rather than hibernating over winter. It is the subspecies most commonly reared commercially.

Queens were confined in standard rearing conditions with a single male B. terrestris pupa. They produced an average of 152 workers, 258 males and 31 queens each, but 48% of the colonies produced no queens.


In experiments in South Korea with 132 field-caught, naturally hibernated B. ignitus queens, Yoon et al. (2002) found 27°C and 65% relative humidity produced higher rates of colony foundation and better colony performance than other temperatures and levels of humidity. Queens were confined alone to induce colony formation. At 27°C, 83% of queens founded colonies, 63% produced colonies with more than 50 workers and 46% produced new queens. These percentages were 2.2-5.5 times higher than rates achieved at 23° and 30°C. Colonies reared at 27°C produced more workers and more queens than those at other temperatures, and developed two to five times faster.


Kwon et al. (2003) found that confining 100 B. terrestris queens with a young male pupa 1-2 days old stimulated egg laying and improved overall colony productivity compared to queens that had been confined with older pupae (9-11 days old). Eighty per cent of the 20 queens given a 1-2 day old pupa produced a colony, compared to 30% of 20 queens given a 9-11 day old pupa.


Lopez-Vaamonde et al. (2004) reared colonies from wild-caught queens of the native UK subspecies B. terrestris audax. Rearing methods are not given in detail, but 32 colonies of at least 10 workers were reared from 122 queens.


Ings et al. (2006) reared wild-caught queens of the native UK subspecies B. terrestris audax in southern England. Queens were confined in dual compartment nest boxes at 25-28°C, 60% relative humidity, with two or three male pupae or artificial pupae made of clay. From 79 nest searching queens caught in March, 20 colonies were reared to the second brood of workers.


Almanza (2007) reported rearing four colonies of the Neotropical pocket-making species B. atratus from wild-caught queens in Colombia. Kept alone under standard rearing conditions, the queens began egg-laying within one week. The colonies lasted between 97 and 183 days, and produced between 40 and 145 workers, but none produced any new queens or males.


Gurel & Gosterit (2008) found B. terrestris queens were more likely to lay eggs and found colonies, and laid eggs more quickly, when they had been confined with a single B. terrestris worker, compared to queens confined alone, with a honey bee worker or with a B. terrestris pupa. These laboratory experiments were carried out at the University of Akdeniz, Turkey, following hibernation and CO2 anaesthetic treatment of laboratory-reared queens.


Li et al. (2008) reared colonies of two bumblebee species native to China - the white-tailed bumblebee B. lucorum and B. ignitus - from queens caught in the field in May (rearing methods not described in detail); 84% of 150 B. ignitus queens, and 89% of 200 B. lucorum queens laid eggs. Colonies produced 105-107 workers/colony on average, with no difference between species, but B. lucorum colonies produced significantly more queens (average 121 queens/colony) than B. ignitus (average 55 queens/colony).


Gurel & Gosterit (2009) reared 50 wild-caught B. terrestris dalmatinus over two generations in the laboratory. Queens were anaesthetized with CO2, then confined with B. terrestris workers to induce colony formation. They found second-generation queens produced around 60% more workers (average 121 workers/colony, compared to 72 workers/colony in the first generation), significantly more males (average 71 males/colony, compared to 30 for first-generation colonies) and completed the colony cycle significantly more quickly than first-generation colonies.


In another replicated controlled laboratory experiment, Gosterit & Gurel (2009) found that hibernating B. terrestris queens at 4-5°C for 45 days followed by anaesthetizing with CO2 for 30 minutes produced the highest egg-laying and colony formation rates, compared to non-hibernated queens, or those hibernated for 75 or 105 days. However, non-hibernated queens (also anaesthetized) ultimately produced larger colonies, with more workers and more new queens and males. These experiments used 148 mated, laboratory reared queens, with 30-40 queens in each treatment group.


Chiang et al. (2009) documented rearing of two montane oriental species, B. eximius and B. sonani in Taiwan. Queens were induced to form colonies by confining them alone in wooden boxes at 26°C and 65% relative humidity, under red light. Of 53 B. eximius queens, 40 (76%) laid eggs, and 31 produced mature colonies. Of 37 B. sonani queens, 27 (73%) laid eggs and 22 produced mature colonies. B. eximius produced significantly larger colonies with on average 120 workers, 210 males and 25 queens, compared to 53 workers, 102 males and nine queens on average for B. sonani.


Whitehorn et al. (2009) reared colonies of B. terrestris from 210 commercially-reared queens by confining queens alone, under standard rearing conditions, following artificial hibernation for 47 days at 6°C. Ninety-three queens (44%) survived artificial hibernation and 47 of them (51% of those surviving hibernation) founded colonies with at least one offspring.

Does pollen diet affect reproductive success?


A replicated laboratory experiment by Regali & Rasmont (1995) showed that four groups of captive B. terrestris workers fed on pollen mainly from oilseed rape Brassica napus ssp. oleifera (22% protein) reared more, larger, longer-lived males than four groups fed on pollen mainly from sunflower Helianthus annuus (13% protein).


Ribeiro et al. (1996) showed that eight laboratory-reared B. terrestris colonies fed on freshly frozen honey bee Apis mellifera pollen produced larger queens, which survived better and produced larger colonies after hibernation than queens from seven colonies fed on dried, commercially available honey bee pollen. There was no difference in the number or weight of workers or males from colonies fed on these two types of pollen.


In a replicated, controlled trial, Génissel et al. (2002) demonstrated that the pollen content of the diet significantly affects the fecundity of B. terrestris in captivity. Twenty groups of three workers fed pollen from fruit trees Prunus spp. or a mix of pollen including fruit tree pollen, produced more offspring (average 14-19 adult males produced/group in 95 days) than 20 groups fed pollen from dandelion Taraxacum sp.or willow Salix sp. (average 0-8 adult males produced/group). The protein content of Prunus pollen was shown to be higher (average 27.5%) than other pollens in the trial.

Additional reference:

Kearns C.A. & Thomson J.D. (2001) The natural history of bumblebees: a sourcebook for investigations. University Press of Colorado, USA.

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