Managed Honey Bee Welfare: Problems and Potential Interventions

Ex­ec­u­tive Summary

At any given time there are more than a trillion man­aged[1] honey bees. Globally, the num­ber of man­aged colonies has risen steadily over the last twenty years and this growth will al­most cer­tainly per­sist, at least in the short term. In­creases in de­mand for honey and (es­pe­cially) com­mer­cial pol­li­na­tion ser­vices con­tinue to out­pace the in­crease in sup­ply of man­aged bees. Asia, es­pe­cially China and In­dia, hosts the largest pop­u­la­tions of man­aged bees and ac­counts for much of the re­cent growth in bee stocks. Com­mer­cial bee­keep­ing tech­niques stan­dardly treat man­aged bees as a re­source from which to max­i­mize the ex­trac­tion of value. Bee­keep­ers have a fi­nan­cial in­cen­tive to main­tain the health of their colonies, but they have lit­tle rea­son to look af­ter the welfare of in­di­vi­d­ual bees. Man­aged bees suffer from a va­ri­ety of prob­lems, in­clud­ing pes­ti­cide ex­po­sure, poor nu­tri­tion due to in­ad­e­quate ac­cess to nat­u­ral for­age, in­va­sive hive in­spec­tions and honey har­vest, stress from long-dis­tance trans­port, and par­a­site and pathogen spread ex­ac­er­bated by com­mon man­age­ment tech­niques. The effec­tive an­i­mal ad­vo­cacy move­ment can help a sig­nifi­cant num­ber of man­aged bees by pro­mot­ing welfare-ori­ented man­age­ment tech­niques, sup­port­ing aca­demic re­search that has the po­ten­tial to ad­vance honey bee welfare, en­courag­ing welfare-ori­ented poli­cies and reg­u­la­tions, and re­duc­ing the de­mand for com­mer­cial pol­li­na­tion ser­vices. How­ever, it’s not clear that these in­ter­ven­tions are promis­ing enough to be pur­sued in the near-term. Much un­cer­tainty re­mains, and more re­search would be re­quired to de­ter­mine the ul­ti­mate wis­dom and cost-effec­tive­ness of these in­ter­ven­tions.

In­tro­duc­tion and Context

This re­port is the first in Re­think Pri­ori­ties’ se­ries on farmed in­ver­te­brates. You can find our other work on in­ver­te­brate sen­tience and in­ver­te­brate welfare here. We are fo­cus­ing our cur­rent efforts on farmed in­ver­te­brates be­cause in most cases farmed in­ver­te­brate in­ter­ven­tions are likely to be more tractable than wild in­ver­te­brate in­ter­ven­tions.

Honey bees ap­pear to be a promis­ing in­ter­ven­tion tar­get for three rea­sons. First, ac­cord­ing to my pre­limi­nary es­ti­mates, man­aged honey bees are the most nu­mer­ous farmed in­sect and prob­a­bly the most nu­mer­ous farmed in­ver­te­brate.[2] Se­cond, the case that bees are sen­tient is stronger than that for most other in­ver­te­brates.[3] Honey bees ex­hibit so­cial, cog­ni­tive, and even af­fec­tive com­plex­ity that is com­pa­rable to some ver­te­brates,[4] and they are im­pres­sive com­mu­ni­ca­tors and learn­ers.[5] Third, there ap­pears to ex­ist more pub­lic sym­pa­thy for honey bees than for most other in­ver­te­brates, per­haps be­cause of the es­sen­tial role they play in agri­cul­tural pol­li­na­tion.[6] The plight of honey bees, though not always framed in welfare terms, is a com­mon enough sub­ject of me­dia at­ten­tion that honey bee in­ter­ven­tions are un­likely to gen­er­ate sig­nifi­cant pub­lic out­rage or oth­er­wise dam­age effec­tive al­tru­ism’s long-term image.

Some Caveats

There are two ma­jor caveats worth not­ing at the out­set. The first is that this re­port is heav­ily bi­ased to­ward bee­keep­ing in North Amer­ica and Europe. The rea­son for this bias is that there is much more liter­a­ture available in English on North Amer­i­can and Euro­pean bees than there is for other re­gions. As a re­sult of this bias, the in­for­ma­tion con­tained in this re­port may not always gen­er­al­ize to other parts of the world. This bias is es­pe­cially un­for­tu­nate be­cause much of the marginal growth in bee stocks is in Asia.

The sec­ond caveat is that I am not a bee­keep­ing ex­pert. My for­mal back­ground is in philos­o­phy. Un­doubt­edly, this re­port con­tains mis­takes which, de­spite the sub­stan­tial as­sis­tance I re­ceived from many tal­ented in­di­vi­d­u­als, re­main my own. Any and all cor­rec­tions are wel­come.

Num­ber of Man­aged Honey Bees

I es­ti­mate that at any given time in 2017 there were be­tween 1.4 and 4.8 trillion adult man­aged honey bees.[7] I de­scribe the method­ol­ogy be­hind this es­ti­mate in Ap­pendix 2.

Fu­ture Growth

De­spite much press, some of it quite apoc­a­lyp­tic, con­cern­ing colony col­lapse di­s­or­der (see more be­low), the num­ber of man­aged honey bee colonies is not ac­tu­ally in de­cline. In fact, the num­ber of to­tal man­aged honey bee hives has steadily in­creased over the last 20 years and looks set to con­tinue in­creas­ing, at least in the near fu­ture. (See Figure 1.) Much of this growth is at­tributable to the grad­ual ex­pan­sion and sheer size of Asian bee­keep­ing.[8] Asian pro­duc­tion has grown more than 60% since 2000, while pro­duc­tion in the Amer­i­cas has re­mained es­sen­tially flat since the turn of the mil­le­nium. (See Figures 2 and 3.)

Figure 1: World Bee­hive Pro­duc­tion, 1994-2017 (source: FAOSTAT)

Figure 2: Asian Bee­hive Pro­duc­tion, 1994-2017 (source: FAOSTAT)

Figure 3: North and South Amer­ica Bee­hive Pro­duc­tion, 1994-2017 (source: FAOSTAT)

Although honey bee pop­u­la­tions have been in­creas­ing, the rate of in­crease has not kept pace with de­mand growth (van En­gels­dorp & Meixner 2010), and de­mand for honey and pol­li­na­tor ser­vices will con­tinue to grow, at least in the short term. Ac­cord­ing to this re­port from Zion Mar­ket Re­search, the global honey mar­ket will in­crease at a com­pound an­nual growth rate of roughly 4.8% over the next six years, from a value of ap­prox­i­mately $7.68 billion in 2018 to ap­prox­i­mately $10.34 billion in 2025. Pol­li­na­tion ser­vices will also see an ex­pan­sion in de­mand, as wild pol­li­na­tor pop­u­la­tions con­tinue to de­cline and agri­cul­tural acreage re­quiring pol­li­na­tion con­tinues to in­crease (Potts et al. 2016). Ac­cord­ing to this re­port from the United States Depart­ment of Agri­cul­ture, in the U.S. bee­keeper rev­enue from pol­li­na­tion ser­vices now sur­passes rev­enue from honey.

Pur­pose of Man­aged Honey Bees

Honey bees are man­aged for two main rea­sons: honey pro­duc­tion and pol­li­na­tion ser­vices. See Ap­pendix 3 for more in­for­ma­tion on honey pro­duc­tion, pol­li­na­tion ser­vices, and sev­eral other minor bee prod­ucts.

Where Honey Bees Are Managed

Honey bees are man­aged on ev­ery con­ti­nent ex­cept Antarc­tica. They are man­aged in al­most ev­ery ma­jor coun­try.

Top Regions

Asia leads the world in honey bee man­age­ment with 47% of global bee­hives. The next biggest re­gions are Europe (at 21% of global bee­hives), Africa (at 19% of global bee­hives), and the Amer­i­cas (at 12% of global bee­hives). (See Figure 4.)

Figure 4: Share of Bee­hives by Re­gion, 2017 (source: FAOSTAT)

Top Countries

Ac­cord­ing to FAO data, in 2017 the top ten coun­tries by num­ber of bee­hives were as fol­lows (see Figure 5):

Table 1: Top 10 Bee­hive Pro­duc­ers (source: FAOSTAT)

Figure 5: Bee­hive Top Pro­duc­ers, 2017 (source: FAOSTAT)

Fac­tors That Affect Honey Bee Welfare

Below I dis­cuss six fac­tors that af­fect man­aged honey bee welfare. The sever­ity of these welfare con­cerns varies by re­gion and man­age­ment prac­tice. Bees man­aged in colder cli­mates are more likely to suffer as a re­sult of the poor ther­mal in­su­la­tion of their ar­tifi­cial hives than bees man­aged in warmer cli­mates. Bees man­aged ex­clu­sively to provide pol­li­na­tion ser­vices are more likely to suffer from trans­port and in­ad­e­quate ac­cess to nat­u­ral for­age than bees man­aged ex­clu­sively for honey. Bees man­aged in Europe are less likely to suffer from pes­ti­cide ex­po­sure than bees man­aged el­se­where be­cause sev­eral pes­ti­cides harm­ful to bees have been banned or re­stricted there. Bees man­aged in Aus­tralia are less likely to suffer from par­a­sites and pathogens than bees el­se­where be­cause Var­roa de­struc­tor has not yet be­come es­tab­lished there. Colony Col­lapse Di­sor­der cur­rently only af­fects bees in North Amer­ica and Europe. And so on.[9]

Tak­ing re­gional differ­ences of bee pop­u­la­tion into ac­count, I would very roughly rank the welfare con­cerns from most se­ri­ous to least se­ri­ous as fol­lows:

  1. Par­a­sites and pathogens

  2. Pes­ti­cide exposure

  3. Food access

  4. Transport

  5. Habitat

  6. Colony Col­lapse Disorder


To un­der­stand honey bee welfare and in­ter­ven­tions aimed at im­prov­ing it, one must un­der­stand the habitat in which honey bees live. Man­aged honey bees nest in ar­tifi­cial hives of var­i­ous de­signs. The most pop­u­lar type of hive de­sign is the Langstroth hive, origi­nally patented back in 1852.[10] The Langstroth hive is an up­right mod­u­lar bee­hive of three ba­sic seg­ments: the lower sec­tion, fea­tur­ing a bot­tom board with an en­trance for bee ingress and egress; the mid­dle sec­tion, with boxes for brood cham­bers and honey su­pers, of­ten di­vided by a queen ex­cluder to keep lar­vae sep­a­rate from honey; and the top sec­tion, in­clud­ing both an in­ner cover and a top cover. (See Figure 6). In­side the boxes are 8 to 10 frames, nor­mally with an ar­tifi­cial, wax-coated foun­da­tion upon which bees de­posit honey or grow young.[11] Langstroth hives can be dis­assem­bled rel­a­tively eas­ily to al­low for hive in­spec­tion or har­vest, and the com­pact de­sign makes the hive sim­ple to trans­port (though they can be heavy). Many bee­keep­ers pre­fer Langstroth hives in large part be­cause of their pop­u­lar­ity. The equip­ment is stan­dard­ized and widely available, and there are a large range of ac­ces­sories and re­sources ex­plic­itly tai­lored to the Langstroth. Most api­ary man­u­als are writ­ten un­der the as­sump­tion that the bee­keeper is us­ing a Langstroth hive.

Figure 6: A Typ­i­cal Hive (source: Mtaita 2003)

From a welfare per­spec­tive, the Langstroth hive poses sev­eral challenges. The first is the in­va­sive na­ture of hive in­spec­tion and honey har­vest. There is a lot of var­i­ance among bee­keep­ers, but typ­i­cally hives are in­spected ev­ery two to six weeks, and honey is har­vested once or twice a year. To in­spect the bees or har­vest the honey, a bee­keeper must re­move the com­po­nents one-by-one. That means first re­mov­ing the roof of the struc­ture, lay­ing bare the del­i­cate in­nards of the hive, then re­mov­ing the honey su­pers un­til the brood cham­ber is ex­posed. This pro­cess in­evitably con­fuses, ag­i­tates, and stresses the bees. Th­ese ag­i­tated bees are much more likely to sting in defense of the colony, which they (cor­rectly) per­ceive as un­der at­tack. Bees that sting in defense of the colony nec­es­sar­ily die, and death-by-stinger-ejec­tion prima fa­cie looks to be a par­tic­u­larly painful way to die. (The barbed stinger liter­ally tears a cav­ity right through the mid­dle of the bee’s ab­domen.) To re­duce ag­gres­sive be­hav­ior, stan­dard bee­keep­ing prac­tice is to smoke the bees dur­ing har­vest. (A smoker is just a lit­tle metal con­tainer with bel­lows in which a fire is lit. See Figure 7.) The smoke both masks the scent and hin­ders the pro­duc­tion of the alarm pheromones that would oth­er­wise alert the bees to dan­ger (Gage et al. 2018).[12] Although the smoke is non­toxic, if the tem­per­a­ture of the smoke is too high, it can melt the wings of the bees. But ap­plied ap­pro­pri­ately, smok­ing the bees saves many bee lives and thus is prefer­able to let­ting the bees sting them­selves to death.[13] Fi­nally, when the hive is re­assem­bled (that is, the com­po­nents are stacked on top of each other again) ag­i­tated bees are of­ten crushed be­tween the edges of the boxes, in­jur­ing or kil­ling them.[14] (Figure 7 gives a sense for how difficult it would be to re­assem­ble the hive with­out crush­ing bees.)

Figure 7: Smok­ing the Bees (source: “Why Smok­ing Soothes the Stressed-Out Bee Hive”)

Another prob­lem with the Langstroth hive is that it tends to have worse ther­mal in­su­la­tion than the nat­u­ral hives that feral colonies es­tab­lish. Reg­u­lat­ing the tem­per­a­ture of a hive is an im­por­tant el­e­ment of colony sur­vival. Hives with poor in­su­la­tion will, in cold ar­eas, cause the bees to cluster ear­lier in the sea­son than they oth­er­wise would, in­creas­ing the en­ergy cost to main­tain ther­mal home­osta­sis, and ul­ti­mately in­creas­ing the chance that the colony won’t sur­vive the win­ter. The thin walls of the Langstroth hive also re­duce the hu­midity of the hive com­pared to the hives of feral colonies. Re­cent re­search sug­gests that the lower hu­midity fa­cil­i­tates the breed­ing of harm­ful par­a­sites like Var­roa de­struc­tor (Mitchell 2016). To miti­gate these risks, in cold cli­mates dur­ing win­ter, hives are some­times wrapped in black plas­tic tar pa­per (see Figure 8) or packed in large ware­houses, some­times thou­sands of hives to a build­ing (see Figure 9).

Figure 8: Over­win­ter­ing On­tario Bees (source: Canada Agri­cul­ture and Food Mu­seum)

Figure 9: Over­win­ter­ing Alberta Bees (source: Canada Agri­cul­ture and Food Mu­seum)

Over the last 13 years, U.S. over­win­ter colony mor­tal­ity[15] has av­er­aged 28.8% of all man­aged colonies.[16] That amounts to over 700,000 colonies lost per win­ter. Con­ser­va­tively, that’s about 8 billion pre­ma­ture bee deaths a year in the United States alone.[17] Although these colonies are re­placed in the spring, keep­ing the over­all bee pop­u­la­tion roughly sta­ble year-to-year, the in­di­vi­d­ual bees who suc­cumb to cold ex­po­sure, star­va­tion, or par­a­sitic in­fes­ta­tion suffer need­lessly. An in­ter­ven­tion that re­duced over­win­ter colony mor­tal­ity in the United States by just 5% would pre­vent at least 400 mil­lion pre­ma­ture bee deaths.

Food Access

There are two welfare wor­ries re­gard­ing ac­cess to qual­ity food sources. The first con­cerns the honey har­vest. Com­mer­cial api­aries typ­i­cally har­vest honey once a year, in late sum­mer or early au­tumn, de­pend­ing on the cli­mate. Bees make honey as a win­ter food store, when for­ag­ing is ei­ther im­pos­si­ble or limited. When this es­sen­tial food source is re­moved in its en­tirety, the colony’s sur­vival is put at risk. Without ac­cess to an ad­e­quate food sup­ply, a colony will col­lapse of star­va­tion in a mat­ter of days. For this rea­son, many bee­keep­ers do not har­vest the en­tirety of a colony’s honey sup­ply at one time. How­ever, it is difficult to know in ad­vance how much honey the bees will re­quire to sur­vive the win­ter, as the amount de­pends in part on the weather. Ac­cord­ing to Gar­rido & Nanetti 2019, “star­va­tion is still a com­mon cause of win­ter colony losses” (85). To re­duce the risk of star­va­tion, com­mer­cial bee­keep­ers of­ten provide their colonies with a honey sub­sti­tute dur­ing the win­ter, such as in­ex­pen­sive white cane sugar syrup. Although sugar syrup of­ten se­cures over­win­ter sur­vival of the colony, sugar syrup lacks many of the micronu­tri­ents that honey con­tains.[18] The im­por­tance of these micronu­tri­ents for proper honey bee nu­tri­tion has only re­cently been rec­og­nized. In par­tic­u­lar, it ap­pears that micronu­tri­ent defi­ciency makes a colony more sus­cep­ti­ble to pathogens and par­a­sites and wors­ens the effects of pes­ti­cide ex­po­sure (Mao, Schuler, & Beren­baum 2013).[19] To meet this nu­tri­tional short­fall, some bee­keep­ers have started pro­vid­ing pro­tein sup­ple­ments to their hives. Un­for­tu­nately, these sup­ple­ments do “not ap­pear to rem­edy prob­lems of poor nu­tri­tion or re­duce colony losses” (DeGrandi-Hoff­man et al. 2015: 194). The DeGrandi-Hoff­man et al. study found that colonies with ac­cess to nat­u­ral for­age had lower pathogen loads and higher over­win­ter sur­vival than those fed pro­tein sup­ple­ments.

Re­gret­tably (and this is the sec­ond con­cern), colonies of­ten do not have ac­cess to high-qual­ity and vari­ate nat­u­ral for­age of the kind that would buffer against the worst bee­keep­ing prac­tices. Honey bees re­quire a di­ver­sity of flow­er­ing plants for best nu­tri­tion (Gar­rido & Nanetti 2019).[20] How­ever, many colonies are lo­cated in in­ten­sively-man­aged mono­cul­tural land­scapes, with­out ac­cess to differ­ent types of flow­ers. Many stud­ies sup­port the con­nec­tion be­tween honey bee nu­tri­tion and ac­cess to nat­u­ral or semi-nat­u­ral habitats. For in­stance, Ri­cigli­ano et al. 2019 found that prox­im­ity to US Con­ser­va­tion Re­serve Pro­gram lands en­hanced honey bee health and colony perfor­mance. Without ac­cess to such habitats, honey bees are likely to suffer from malnu­tri­tion, mak­ing them more sus­cep­ti­ble to other stres­sors.


Honey bees are used to fa­cil­i­tate the pol­li­na­tion of a wide va­ri­ety of differ­ent crops. Gen­er­ally speak­ing, it’s not pos­si­ble to sim­ply raise the bees in lo­ca­tions that re­quire pol­li­na­tion ser­vices. Most crops that re­quire com­mer­cial pol­li­na­tion ser­vices are in bloom for only two to four weeks. In the mono­cul­tural land­scapes that are typ­i­cal of fields and or­chards re­quiring pol­li­na­tion ser­vices, there isn’t enough flow­er­ing for­age to sup­port healthy colonies the rest of the year. In the United States, I es­ti­mate that be­tween ½ and ⅔ of colonies are trans­ported at some point dur­ing the year. Con­ser­va­tively, that comes to be­tween 14 and 43 billion in­di­vi­d­ual bees trans­ported in the U.S. alone.[21] As noted above, U.S. bee­keep­ers now earn more rev­enue from pol­li­na­tion ser­vices than they do from honey sales. Given the high de­mand for pol­li­na­tion ser­vices wor­ld­wide, es­pe­cially in East Asia and South Asia, it’s rea­son­able to as­sume that many com­mer­cial bee­keep­ers out­side the U.S. on av­er­age also earn at least half their rev­enue from pol­li­na­tion ser­vices.

Get­ting bees to or­chard and field for the short pe­riod in which crops blos­som and are thus able to be pol­li­nated re­quires trans­port­ing a large num­ber of colonies across vast dis­tances. The bees are shipped con­fined in their hives on the backs of large, open-air semi-trucks. (See Figure 10.) Colonies are of­ten moved mul­ti­ple times a year, cov­er­ing thou­sands of miles. A typ­i­cal North Amer­i­can colony might be­gin the year in North Dakota (home to the largest num­ber of U.S. bee­hives), spend Fe­bru­ary and March pol­li­nat­ing al­monds in North­ern Cal­ifor­nia, then April through June pol­li­nat­ing ap­ples in Cen­tral Wash­ing­ton, fi­nally re­turn­ing to North Dakota in July, travers­ing more than 3,500 miles in to­tal roundtrip. (See Figure 11). This kind of mi­gra­tory bee­keep­ing is also com­mon in In­dia (Tej et al. 2017),[22] and prob­a­bly el­se­where too.

Figure 10: Trans­port­ing Honey Bee Hives (source: “Road Trip: How Hive Trans­porta­tion Puts Stress on Honey Bees”)

Figure 11: Typ­i­cal Trans­port Route for North Amer­i­can Colony (source: Miller Honey Farms)

In gen­eral, “bees from mi­gra­tory colonies have a shorter lifes­pan and higher lev­els of ox­ida­tive stress than work­ers at sta­tion­ary api­aries” (Gar­rido & Nanetti 2019: 75). The ex­act causes of this stress are un­known, but there are sev­eral po­ten­tial welfare con­cerns as­so­ci­ated with long-dis­tance trans­port of honey bee colonies.[23]

Long-dis­tance trans­port has wor­ry­ing effects on honey bee phys­iol­ogy. Ahn et al. 2012 found that “bees ex­pe­rienc­ing trans­porta­tion have trou­ble fully de­vel­op­ing their food glands and this might af­fect their abil­ity to nurse the next gen­er­a­tion of work­ers” (1). Again, the ex­act causes of this ab­nor­mal­ity are un­known, but the sus­tained vibra­tion of trans­port com­bined with changes in air pres­sure, ex­po­sure to diesel ex­haust, and the stress of con­fine­ment may all play a role. A par­tic­u­lar con­cern is tem­per­a­ture reg­u­la­tion. As noted above, main­tain­ing ther­mal home­osta­sis is a ba­sic re­quire­ment for colony sur­vival. Dur­ing trans­port densely packed and poorly ven­tilated colonies of­ten die from over­heat­ing. It’s pos­si­ble to in­crease ven­tila­tion, but do­ing so in­tro­duces cold stress if the out­side air tem­per­a­ture is low, which causes de­vel­op­men­tal ab­nor­mal­ities in the bee lar­vae, of­ten lead­ing to colony col­lapse when the lar­vae emerge as adults a few days af­ter ar­rival (Melicher et al. 2019). The out­side air tem­per­a­ture on a trip from North Dakota to Cal­ifor­nia is go­ing to vary con­sid­er­ably, as the truck has to cross both the Rocky Moun­tains and the Sierra Ne­vada moun­tain range.

Trans­port also likely in­creases the spread of in­fec­tious dis­ease. Bring­ing to­gether bees from many differ­ent re­gions fa­cil­i­tates the trans­mis­sion of par­a­sites and pathogens (Gar­rido & Nanetti 2019: 75). As noted above, dur­ing peak al­mond blos­som sea­son in late Fe­bru­ary, some­where be­tween ⅓ and ½ of all U.S. honey bees are con­cen­trated in Cal­ifor­nia’s Cen­tral Valley. Based on con­ser­va­tive win­ter colony sizes of be­tween 10,800 and 14,000 bees, that comes to be­tween 9.8 and 18.7 billion in­di­vi­d­ual bees. Th­ese bees are trans­ported from as far away as Michi­gan. Due to hom­ing er­rors[24] that are com­mon when hives are crowded to­gether (see be­low), bees from dis­parate ar­eas in­evitably come into close con­tact with one an­other. Be­cause the stress of trans­port in­creases vuln­er­a­bil­ity to dis­ease, the crowd­ing of hives from dis­parate ar­eas can be par­tic­u­larly dev­as­tat­ing.

Par­a­sites and Pathogens

Com­mer­cially man­aged honey bees are threat­ened by a large num­ber of par­a­sites and pathogens: bac­te­rial and fun­gal in­fec­tions, es­pe­cially those lead­ing to Amer­i­can foulbrood, chalk­brood, and stone­brood; par­a­sitic mites, es­pe­cially ec­topar­a­sitic mites like Var­roa de­struc­tor and tra­cheal mites like Acara­pis woodi; microspori­dian pathogens, es­pe­cially Nosema cer­anae and Nosema apis; and the as­so­ci­ated dis­eases that these par­a­sites and pathogens fa­cil­i­tate, es­pe­cially de­formed wing virus and nose­mo­sis. A com­plete sur­vey of honey bee par­a­sites and pathogens is be­yond the scope of this re­port. In what fol­lows I briefly dis­cuss Var­roa mites and de­formed wing virus, the most preva­lent virus in honey bees, which the mites help prop­a­gate (Martin & Bret­tell 2019).

Var­roa mites are small, obli­gate par­a­sites that feed on the hemolymph[25] of both adult and ju­ve­nile bees. (See Figure 12.) They are among the most dev­as­tat­ing par­a­sites of Apis mel­lifera bees. Var­roa mites are rarely fatal to adult hosts, though the mites do shorten the lifes­pan of in­fected bees.[26] Im­por­tantly, Var­roa mites re­quire bee lar­vae or pu­pae to re­pro­duce, and their pres­ence can greatly re­duce colony-wide re­pro­duc­tion.

Figure 12: A Fe­male Var­roa de­struc­tor Mite Feeds on an Adult Worker Bee (source: Univer­sity of Florida En­to­mol­ogy Depart­ment)

Var­roa mites can de­stroy a honey bee colony in at least two ways. The first is by out-pro­duc­ing their hosts and over­run­ning the colony. Bee pop­u­la­tions typ­i­cally peak in early sum­mer, then grad­u­ally de­cline through­out the fall. Var­roa pop­u­la­tions typ­i­cally peak a few weeks af­ter bee pop­u­la­tions. When a mod­est de­cline in bee pop­u­la­tion is com­bined with a mod­est in­crease in Var­roa in­fes­ta­tion, the re­sult can be a pre­cip­i­tous drop in colony strength, ul­ti­mately lead­ing to over­win­ter col­lapse.

The other way Var­roa mites can de­stroy a colony is by spread­ing dis­eases. Var­roa mites are a vec­tor for at least five honey bee dis­eases, the worst of which seems to be de­formed wing virus (DWV). DWV has been de­scribed as “a ma­jor threat to the world’s hon­ey­bees” (Wilfert et al. 2016: 594). DWV is an RNA virus char­ac­ter­ized by mis­shapen, use­less wings and other ab­dom­i­nal de­for­mities. (See Figure 13.) Although DWV can spread di­rectly from one bee to an­other, it is more com­monly trans­mit­ted by the mite Var­roa de­struc­tor. DWV is one of the lead­ing causes of over­win­ter col­lapse in Var­roa-in­fested colonies (Wilfert et al. 2016). More­over, in­fected colonies of­ten pass the dis­ease to wild in­sects, in­clud­ing both feral honey bees and other in­sect pol­li­na­tors, preda­tors, and scav­engers (Villalo­bos 2016: 556), so con­trol­ling the dis­ease would prob­a­bly also im­prove wild an­i­mal welfare.

Figure 13: A Worker Bee Is Stricken by De­formed Wing Virus (source: Univer­sity of Florida En­to­mol­ogy Depart­ment)

Be­cause Var­roa in­fes­ta­tions can be so dev­as­tat­ing, a wide range of Var­roa con­trol meth­ods have been de­vel­oped. (See Table 2.) The welfare im­pli­ca­tions of these con­trol meth­ods, both for the bees and the mites them­selves, are un­clear. Be­cause of the large scale at which these con­trol meth­ods are ap­plied, fu­ture welfare-ori­ented re­search could po­ten­tially un­cover cost-effec­tive in­ter­ven­tions.

Table 2: Var­roa Con­trol Meth­ods (recre­ated from: “Sur­vey Shines Light on Bee­keep­ers’ Efforts to Man­age Var­roa Mites”)

Pes­ti­cide Exposure

Pes­ti­cide ex­po­sure is a ma­jor welfare con­cern for com­mer­cially man­aged honey bees.[27] Gen­eral pur­pose in­sec­ti­cides de­ployed to con­trol crop-dam­ag­ing in­sects can be lethal to bees, as can res­i­den­tial de­ploy­ment of in­sec­ti­cides to con­trol mosquitoes (Long & Krupke 2016).[28] Bees are es­pe­cially at risk for two rea­sons. First, they have a wide for­ag­ing range, up to three miles in any di­rec­tion from the hive. It is thus difficult for a bee­keeper to sta­tion her bees out of range of all pes­ti­cide use. Se­cond, bees are so­cial. If a bee en­coun­ters a pes­ti­cide, she will bring that pes­ti­cide back to the hive. When the pes­ti­cide-ex­posed bee dies, other bees will re­move her body from the hive and be­come con­tam­i­nated by con­tact. In this way, tens of thou­sands of bees can be kil­led in a sin­gle hive in un­der 48 hours.

Sublethal ex­po­sure to pes­ti­cides is also a ma­jor con­cern. Short of kil­ling a bee, pes­ti­cide ex­po­sure can re­duce hom­ing suc­cess,[29] de­grade flight abil­ity, and im­pair bee nav­i­ga­tion (Tosi, Bur­gio, & Nieh 2017). Sublethal pes­ti­cide ex­po­sure can also sup­press a bee’s im­mune sys­tem, mak­ing it more likely to con­tract in­fec­tions and dis­eases (Pet­tis et al. 2012).[30] Fi­nally, pes­ti­cide con­tam­i­na­tion in stored bee­bread and wax comb can diminish a queen’s re­pro­duc­tive rate, low­er­ing colony strength and in­creas­ing the fre­quency of stress-in­duc­ing queen re­place­ments (Traynor et al. 2016).[31] Bees en­gaged in pol­li­na­tion ser­vices are es­pe­cially at risk of sub­lethal pes­ti­cide ex­po­sure. Pes­ti­cides are an in­te­gral part of mod­ern agri­cul­tural pro­duc­tion. When bees are brought to pol­li­nate pes­ti­cide-treated crops, some level of ex­po­sure is in­evitable, though of course the level must be sub­lethal to en­sure pol­li­na­tion suc­cess.

Neon­i­cot­inoids are per­haps the most no­to­ri­ous pes­ti­cide as­so­ci­ated with bees. (See Figure 14.) Neon­i­cot­inoids are among the most widely ap­plied in­sec­ti­cides in the world (Lundin et al. 2015).[32] Though now par­tially banned in Europe (see more be­low), neon­i­cot­inoids have been used ubiquitously the past two decades to sup­press in­sects that re­duce agri­cul­tural yields. Neon­i­cot­inoids are par­tic­u­larly harm­ful for three rea­sons. First, un­like con­tact pes­ti­cides, which re­main on the sur­face of treated plants, sys­temic pes­ti­cides like neon­i­cot­inoids are wa­ter sol­u­ble and are taken up by and trans­ported through­out the plant, which is how they end up in nec­tar and pol­len. In effect, plants that ab­sorb neon­i­cot­inoids be­come toxic to in­sects. Se­cond, neon­i­cot­inoids can per­sist in the soil months or even years af­ter a sin­gle ap­pli­ca­tion. Un­treated plants that grow in the soil dur­ing this time will ab­sorb neon­i­cot­inoid resi­dues and be­come at least some­what toxic (Tosi, Bur­gio, & Nieh 2017). Third, neon­i­cot­inoids are par­tic­u­larly effec­tive at im­muno­sup­pres­sion, mean­ing that even low con­cen­tra­tions of neon­i­cot­inoids can harm bees and other in­sects by ex­ac­er­bat­ing the effect of other stres­sors. Ac­cord­ing to Sánchez-Bayo et al. 2016, “Im­mune sup­pres­sion of the nat­u­ral defences by neon­i­cot­inoid and phenyl-pyra­zole (fipronil) in­sec­ti­cides opens the way to par­a­site in­fec­tions and viral dis­eases, fos­ter­ing their spread among in­di­vi­d­u­als and among bee colonies at higher rates than un­der con­di­tions of no ex­po­sure to such in­sec­ti­cides” (7). Neon­i­con­tinoids are also par­tic­u­larly bad for queens. Willi­ams et al. 2015 demon­strate that “field-re­al­is­tic con­cen­tra­tions of neon­i­cot­inoid pes­ti­cides dur­ing de­vel­op­ment can severely af­fect queens of west­ern honey bees (Apis mel­lifera). In pes­ti­cide-ex­posed queens, re­pro­duc­tive anatomy (ovaries) and phys­iol­ogy (sper­math­e­cal-stored sperm qual­ity and quan­tity), rather than flight be­havi­our, were com­pro­mised and likely cor­re­sponded to re­duced queen suc­cess (al­ive and pro­duc­ing worker offspring)” (1). Im­por­tantly, “many wild bees and other non-tar­get in­ver­te­brates are also ex­posed to the same suite of pes­ti­cides” (Long & Krupke 2016: 10), mak­ing pes­ti­cide ex­po­sure a welfare con­cern for both man­aged and wild an­i­mals.

Figure 14: Neon­i­cot­inoid Pes­ti­cides (source: Com­pound In­ter­est)

Com­mer­cially man­aged honey bees are also ex­posed to pes­ti­cides in­side the hive. As noted above, Var­roa in­fes­ta­tions are a ma­jor cause of colony col­lapse and a ma­jor con­cern for com­mer­cial api­aries. As a con­se­quence, 89% of large-scale U.S. bee­keep­ers[33] re­port ap­ply­ing chem­i­cal var­roacides in­side their hives (Haber, Stein­hauer, & van En­gels­drop 2019). The chem­i­cal treat­ment is effec­tive at con­trol­ling Var­roa pop­u­la­tions, but the long-term effects on the bees are less well un­der­stood.

Colony Col­lapse Disorder

Large-scale and wide­spread colony losses were first re­ported across the United States in 2006 (Villalo­bos 2016). Similar losses were first re­ported in Europe in 2009 (Dainat, van En­gels­dorp, & Neu­mann 2012). Th­ese losses were no­table both for their scale and the pe­cu­liar cir­cum­stances in which the losses oc­curred. Colonies that suc­cumbed to this con­di­tion didn’t leave be­hind a pile of dead bees, as hap­pens when a colony suffers, for in­stance, acute pes­ti­cide poi­son­ing; on the con­trary, the hives of the af­fected colonies were sim­ply empty. Colony Col­lapse Di­sor­der, as the con­di­tion came to be called, is char­ac­ter­ized by the sud­den dis­ap­pear­ance of the ma­jor­ity of worker bees in a colony, de­spite a func­tional queen and ad­e­quate food stores.

Colony Col­lapse Di­sor­der is still a poorly un­der­stood phe­nomenon, but ev­i­dence is con­verg­ing to sug­gest there is no sin­gle un­der­ly­ing cause (Long & Krupke 2016).[34] In­stead, it ap­pears that a num­ber of in­de­pen­dent stres­sors in­crease a colony’s vuln­er­a­bil­ity to col­lapse. When these stres­sors are com­bined at a sin­gle api­ary, colony vuln­er­a­bil­ity reaches a tip­ping point and losses can be catas­trophic, de­stroy­ing 30-80% of hives in a sin­gle sea­son (nor­mally win­ter). As of this writ­ing, Colony Col­lapse Di­sor­der has not been ob­served in sig­nifi­cant num­bers out­side North Amer­ica and Europe, for rea­sons that are not en­tirely clear.

If it’s true that Colony Col­lapse Di­sor­der has no unique fun­da­men­tal cause but is in­stead the re­sult of the con­fluence of other stres­sors, then Colony Col­lapse Di­sor­der is not an in­de­pen­dent welfare con­cern. Rather it is a re­minder that the con­se­quences of other stres­sors can ag­gre­gate in un­pre­dictable, non-ad­di­tive ways. The com­bined stress of two in­de­pen­dent sources may be more than the sum of the in­di­vi­d­ual parts.

For ex­am­ple, ac­cord­ing to Mao, Schuler, & Beren­baum 2013, “The wide­spread api­cul­tural use of honey sub­sti­tutes, in­clud­ing high-fruc­tose corn syrup, may thus com­pro­mise the abil­ity of honey bees to cope with pes­ti­cides and pathogens and con­tribute to colony losses” (8842). Pet­tis et al. 2012 re­ports, “In­ter­ac­tions be­tween pes­ti­cides and pathogens could be a ma­jor con­trib­u­tor to in­creased mor­tal­ity of honey bee colonies, in­clud­ing colony col­lapse di­s­or­der” (153). And Sánchez-Bayo et al. 2016 adds, “The nega­tive im­pact on im­mune bar­ri­ers can be fur­ther ex­ac­er­bated by con­cur­rent stress agents that in­terfere with NF-kB sig­nal­ling (e.g. neon­i­cot­inoids), or ac­ti­vate com­pet­ing stress re­sponses (e.g. poor nu­tri­tion, ex­treme ther­mal con­di­tions), thus favour­ing pathogens and par­a­sites” (10). To sum­ma­rize, the stress of trans­port, honey har­vest, in­ad­e­quate ac­cess to nat­u­ral for­age, and poor ther­mal in­su­la­tion make bees less re­sis­tant to par­a­sites, pathogens, and the worst effects of pes­ti­cide ex­po­sure. Bees are sturdy crea­tures, and bee colonies are re­mark­ably re­silient su­per­or­ganisms, but when pushed to the brink by mul­ti­ple stres­sors, they sim­ply van­ish.

Po­ten­tial Interventions

Below I dis­cuss six broad cat­e­gories of in­ter­ven­tions. It is un­for­tu­nately difficult to as­sess the com­par­a­tive value of these differ­ent in­ter­ven­tions for sev­eral rea­sons. First, as noted above, the sever­ity of a welfare con­cern is con­text-de­pen­dent (e.g., bees in Aus­tralia suffer from differ­ent prob­lems than bees in Europe). Se­cond, not all of the in­ter­ven­tions are aimed at the same time hori­zon. Plant­ing nat­u­ral for­age in some ar­eas could start benefit­ing bees in weeks; build­ing the effec­tive an­i­mal ad­vo­cacy move­ment in Asia might not pay off for decades. Third, where the in­ter­ests of bees al­igns with the in­ter­ests of bee­keep­ers (or farm­ers), fi­nan­cial in­cen­tives will prob­a­bly push adop­tion of the in­ter­ven­tion with­out out­side effort. So, as always, an­ti­ci­pated ne­glect­ed­ness must be fac­tored into the com­par­a­tive anal­y­sis. Fi­nally, many of these in­ter­ven­tions will have knock-on effects, es­pe­cially for wild in­ver­te­brates, that are difficult to calcu­late.

Tak­ing all these fac­tors into ac­count, I would very roughly rank the in­ter­ven­tions from most promis­ing to least promis­ing as fol­lows:

  1. Build EAA in Asia

  2. Re­duce pol­li­na­tion demand

  3. In­crease ac­cess to nat­u­ral forage

  4. Man­age pes­ti­cide risk

  5. Re­duce hom­ing errors

  6. Min­i­mize har­vest disruption

Re­duce Hom­ing Errors

Like other com­mer­cially farmed an­i­mals, honey bees are main­tained at den­si­ties much higher than those found in the wild. Feral colonies of honey bees pre­fer to nest hun­dreds or thou­sands of me­ters apart from each other, whereas man­aged honey bee colonies are typ­i­cally kept less than a me­ter apart on com­mer­cial api­aries (Gar­rido & Nanetti 2019). This level of crowd­ing can lead to a va­ri­ety of welfare prob­lems. One such prob­lem is known as ‘drift­ing.’ Drift­ing is a type of hom­ing er­ror in which a for­ag­ing bee ac­ci­den­tally re­turns to the wrong hive. Drift­ing is a se­ri­ous welfare con­cern be­cause, ac­cord­ing to Seeley & Smith 2015, it is “per­haps the most com­mon mechanism of dis­ease trans­mis­sion be­tween colonies within an api­ary” (717). They add that on a typ­i­cal api­ary, “it is com­mon for 40% or more of all worker bees to drift from their na­tal colony to a neigh­bor­ing colony” (717).

For­tu­nately, a re­cent study sug­gests there may be a low-cost method that greatly re­duces this type of hom­ing er­ror with­out af­fect­ing the den­sity at which colonies can be kept. Dynes et al. 2019 demon­strate that paint­ing hives differ­ent col­ors, mark­ing hives with unique sym­bols, plac­ing hives at differ­ent heights, and ori­ent­ing hives to face differ­ent di­rec­tions can all greatly re­duce drift with­out the need to de­crease den­sity. This re­duc­tion in drift in turn low­ers colony par­a­site load, pro­motes honey pro­duc­tion, and in­creases over­win­ter sur­vival. Pro­mot­ing these changes could thus have a sub­stan­tial welfare im­pact for a sig­nifi­cant num­ber of bees.[35]

The main ob­jec­tion to pro­mot­ing this in­ter­ven­tion is that bee­keep­ers already have a fi­nan­cial in­cen­tive to make these changes if they are in fact effec­tive. How­ever, notwith­stand­ing this in­cen­tive, I found vir­tu­ally no ev­i­dence that com­mer­cial api­aries were im­ple­ment­ing these changes at scale. The most plau­si­ble rea­son for this lack of change is sim­ply that the re­search is so new.[36] If that’s the ex­pla­na­tion, then we should ex­pect these changes to be im­ple­mented over the next few years (again, as­sum­ing that the benefits are as dra­matic as they ap­pear). There may, how­ever, be in­for­ma­tion bar­ri­ers that de­lay the im­ple­men­ta­tion of these re­forms, es­pe­cially in low or mid­dle in­come coun­tries. If such in­for­ma­tion bar­ri­ers ex­ist, then a well-or­ga­nized, low-cost in­for­ma­tion cam­paign might prove to be a cost-effec­tive in­ter­ven­tion.

Re­duce Pol­li­na­tion Demand

Re­duc­ing the de­mand for pol­li­na­tion ser­vices would re­duce the num­ber of bees in com­mer­cial fields and or­chards, where they are more likely to be ex­posed to pes­ti­cides (Traynor et al. 2016).[37] It would also re­duce the num­ber of bees trans­ported long dis­tances, which would im­prove the welfare of those bees both di­rectly, by elimi­nat­ing the stress of trans­port, and in­di­rectly, by re­duc­ing the trans­mis­sion of dis­ease that re­sults from com­min­gling colonies from dis­parate re­gions. There are at least four ways one might re­duce de­mand for honey bee pol­li­na­tion ser­vices: re­duc­ing de­mand for prod­ucts that re­quire com­mer­cial honey bee pol­li­na­tion, re­plac­ing pol­li­na­tor-de­pen­dent crops with self-fer­tile va­ri­eties, pro­mot­ing me­chan­i­cal pol­li­na­tion, and in­creas­ing the pop­u­la­tion of lo­cal wild pol­li­na­tors.

The most ob­vi­ous way to re­duce de­mand for com­mer­cial honey bee pol­li­na­tion is to re­duce de­mand for those prod­ucts that re­quire com­mer­cial honey bee pol­li­na­tion. I sus­pect this ap­proach is the least promis­ing. If di­rectly re­duc­ing con­sumer in­ter­est for cer­tain prod­ucts were pos­si­ble and cost-effec­tive, it would be a more widely used tac­tic in effec­tive an­i­mal ad­vo­cacy. And if it were pos­si­ble and cost-effec­tive to di­rectly re­duce con­sumer in­ter­est in a product, then honey would be a more ob­vi­ous tar­get than prod­ucts that rely in­di­rectly on bees.[38] Nev­er­the­less, there may be limited room for a nar­row cam­paign of this sort. In the United States, al­mond farm­ing is a ma­jor driver of de­mand for pol­li­na­tion ser­vices. Al­monds are a no­to­ri­ously thirsty crop, re­quiring over a gal­lon of wa­ter per nut (Ful­ton, Nor­ton, & Shilling 2019). A cam­paign that em­pha­sized the large en­vi­ron­men­tal im­pact of al­mond agri­cul­tural might suc­ceed in re­duc­ing de­mand for al­monds, thereby re­duc­ing de­mand for com­mer­cial honey bee pol­li­na­tion ser­vices.[39]

As noted above, al­mond farm­ing is a ma­jor driver of de­mand for pol­li­na­tion ser­vices in the United States. Most al­mond va­ri­eties are self-in­com­pat­i­ble and there­fore re­quire both pol­l­eniz­ers[40] and pol­li­na­tors. Re­cently, the surge in the price of pol­li­na­tion ser­vices (see Figure 15) has led to a con­cur­rent surge in in­ter­est in self-fer­tile va­ri­eties of al­mond. The most pop­u­lar self-fer­tile al­mond va­ri­ety is the In­de­pen­dence, de­vel­oped by Zaiger Ge­net­ics and available since 2008. Ac­cord­ing to Dani Ligh­tle, an or­chard sys­tems ad­vi­sor with the Univer­sity of Cal­ifor­nia Co­op­er­a­tive Ex­ten­sion pro­gram, In­de­pen­dence al­monds re­quire 75-80% fewer bees for effec­tive pol­li­na­tion. Ac­cord­ing to the Cal­ifor­nia Depart­ment of Food and Agri­cul­ture Al­mond Acreage Re­port, in 2018 23.4% of all new al­mond acres were planted to the self-fer­tile In­de­pen­dence va­ri­ety. How­ever, not all ex­perts are bullish on the abil­ity of self-fer­tile al­monds to re­duce pol­li­na­tion de­mand. Brit­tney Goodrich, an as­sis­tant pro­fes­sor in the Depart­ment of Agri­cul­tural Eco­nomics and Ru­ral So­ciol­ogy at Auburn Univer­sity and a spe­cial­ist in pol­li­na­tion ser­vice con­tracts, claims that self-fer­tile al­mond va­ri­eties will not have a sig­nifi­cant im­pact on the de­mand for colonies for the fore­see­able fu­ture.

Figure 15: Al­mond Pol­li­na­tion Prices Have More Than Dou­bled Since 2004 (source: USDA Eco­nomic Re­search Ser­vice)

Directly sub­si­diz­ing self-fer­tile va­ri­eties of pol­li­na­tor-de­pen­dent crops does not look to be a promis­ing in­ter­ven­tion be­cause farm­ers already have a fi­nan­cial in­cen­tive to switch to these va­ri­eties if they in fact re­duce pol­li­na­tion costs with­out com­pro­mis­ing qual­ity. How­ever, there may again be in­for­ma­tion bar­ri­ers that slow the adop­tion of such va­ri­eties, es­pe­cially in low or mid­dle in­come coun­tries where in­sti­tu­tional dis­sem­i­na­tion of new knowl­edge may be limited and where farm­ers may be more risk-averse. It might pos­si­ble to mount a rel­a­tively cost-effec­tive out­reach cam­paign that would benefit both poor farm­ers and over­taxed honey bees.

Me­chan­i­cal pol­li­na­tion is an­other al­ter­na­tive to com­mer­cially man­aged honey bee pol­li­na­tion. A typ­i­cal me­chan­i­cal pol­li­na­tion sys­tem uses an elec­tro­static sprayer to ap­ply a sus­pen­sion of pol­len to flow­er­ing fruit trees. (See Figure 16.) It can be used to sup­ple­ment or re­place bee pol­li­na­tion. At around $300 an acre, it com­pares fa­vor­ably to honey bee pol­li­na­tion, which can range from $320 to $450 an acre for al­mond or­chards. The main prob­lem right now is effi­cacy. Ac­cord­ing to Eliz­a­beth Ficht­ner, an­other or­chard sys­tems ad­vi­sor with the Univer­sity of Cal­ifor­nia Co­op­er­a­tive Ex­ten­sion pro­gram, me­chan­i­cal pol­li­na­tion of al­mond or­chards is in­effec­tive be­cause al­mond flow­ers are not all open at the same time. Be­cause me­chan­i­cal pol­li­na­tion is a dis­crete event, if a flower isn’t open when the sprayer drives by, that flower doesn’t get pol­li­nated. Honey bee pol­li­na­tion, by con­trast, is con­tin­u­ous, with for­agers worker through­out daylight hours, weather per­mit­ting.[41] How­ever, the tech­nol­ogy is still rel­a­tively im­ma­ture. Matt Whit­ing, a stone fruit phys­iol­o­gist in Wash­ing­ton State Univer­sity’s Depart­ment of Hor­ti­cul­ture, be­lieves the tech­nique will be widely adopted in 20 to 30 years’ time. Re­search grants that ei­ther en­sure the tech­nol­ogy is de­vel­oped or speeds its progress could thus prove to be cost-effec­tive in­ter­ven­tions for com­mer­cially man­aged honey bees.

Juan Farias drives a spray rig outfitted with test equipment for the application of pollination in a cherry block in the Prosser, Washington on March 3, 2015. <b>(TJ Mullinax/Good Fruit Grower)</b>

Figure 16: A Cherry Or­chard Un­der­goes Me­chan­i­cal Pol­li­na­tion (source: “No bees, but a lot of buzz about ar­tifi­cial pol­li­na­tion”)

Another way to re­duce re­li­ance on com­mer­cial honey bee pol­li­na­tion is to in­crease the pop­u­la­tion of lo­cal wild pol­li­na­tors who will pol­li­nate fields and or­chards for free.[42] There are two ma­jor ad­van­tages to this ap­proach and one se­ri­ous draw­back. The two ad­van­tages are that the means to in­crease wild pol­li­na­tor pop­u­la­tions are well-known and there are sev­eral rep­utable or­ga­ni­za­tions already work­ing to­ward this goal, ready to re­ceive ad­di­tional funds to ex­pand their pro­jects. Ac­cord­ing to Kleijn et al. 2015, “Across all stud­ies, bio­di­ver­sity man­age­ment raised the abun­dance of dom­i­nant crop-vis­it­ing [wild] bees by a fac­tor of 3.2. Or­ganic farm­ing, plant­ing wildflow­ers and es­tab­lish­ing grass mar­gin strips sig­nifi­cantly en­hanced dom­i­nant crop-vis­it­ing bees in arable land­scapes” (3).[43] One sim­ple con­ser­va­tion mea­sure is to plant wild­flow­ers along curb­sides (see Figure 17), which in­creases lo­cal pol­li­na­tor pop­u­la­tions (and sur­pris­ingly de­creases in­sect-ve­hi­cle col­li­sions). Another sim­ple con­ser­va­tion mea­sure is to mow road­side veg­e­ta­tion less fre­quently. Less fre­quent mow­ing im­proves the habitat of lo­cal pol­li­na­tors, while also re­duc­ing main­te­nance costs (Com­mit­tee on Ecol­ogy and Trans­porta­tion Newslet­ter, Jan­uary 2017). Home­own­ers can also im­prove pol­li­na­tor habitat by re­plac­ing lawns with semi-nat­u­ral habitat or mow­ing lawns less fre­quently (Ler­man et al. 2018).[44] The Xerces So­ciety has been ad­vo­cat­ing for sim­ple pol­li­na­tor con­ser­va­tion mea­sures such as these for decades. They main­tain ex­ten­sive re­sources on habitat plan­ning for benefi­cial in­sects, at­tract­ing na­tive pol­li­na­tors, road­side man­age­ment for pol­li­na­tor con­ser­va­tion, and farm­ing with na­tive benefi­cial in­sects. Tar­geted grants to the Xerces So­ciety could re­sult in non-triv­ial in­creases in wild pol­li­na­tor pop­u­la­tions, thus po­ten­tially re­duc­ing the need for com­mer­cial honey bee pol­li­na­tion.[45]

Figure 17: Wild­flow­ers along I-85 in North Carolina (source: “Pol­li­na­tor Con­ser­va­tion at 60 MPH”)

The main draw­back to this ap­proach is that it is es­sen­tially just swap­ping man­aged in­sects for wild in­sects. If the lives of the wild pol­li­na­tors are not any bet­ter than the lives of the com­mer­cially man­aged honey bees, then this ap­proach would not im­prove over­all an­i­mal welfare. If the lives of the wild pol­li­na­tors are worse than the lives of the man­aged bees, then this in­ter­ven­tion would ac­tu­ally be net-nega­tive. This ques­tion de­serves more re­search (see be­low). There may be other rea­sons to pro­mote na­tive pol­li­na­tor pop­u­la­tions, such as con­serv­ing bio­di­ver­sity or im­prov­ing ecosys­tem re­silience, but from a welfare per­spec­tive in­creas­ing the pop­u­la­tion of wild in­sects looks like an un­cer­tain buy.

Man­age Pes­ti­cide Risk

As noted above, pes­ti­cide ex­po­sure is a welfare prob­lem for both com­mer­cially man­aged honey bees, their feral coun­ter­parts, and other non-tar­get wild in­sects.[46] Re­duc­ing non-tar­get pes­ti­cide ex­po­sure could have sev­eral welfare benefits for com­mer­cially man­aged honey bees. The most di­rect effect would be im­proved honey bee health and de­creased chance of colony col­lapse. Re­duc­ing non-tar­get pes­ti­cide ex­po­sure would also benefit wild in­sect pol­li­na­tors, which could in­di­rectly help com­mer­cially man­aged honey bees. If the pop­u­la­tion of wild in­sect pol­li­na­tors were larger, there would be more in­sects available to pol­li­nate agri­cul­tural crops for free, pre­sum­ably de­creas­ing the de­mand for com­mer­cial pol­li­na­tion ser­vices. De­creased pol­li­na­tion de­mand would pre­sum­ably lead to fewer bees trans­ported long dis­tances (see more be­low).

There are three routes to re­duc­ing the pes­ti­cide ex­po­sure of non-tar­get in­sects. The first is out­reach to pro­fes­sional pes­ti­cide ap­pli­ca­tors to en­sure that they fol­low best prac­tices to miti­gate non-tar­get ex­po­sure risk. Train­ing and cer­tifi­ca­tion pro­grams can im­part valuable knowl­edge re­gard­ing proper ap­pli­ca­tion pro­ce­dures. Sea­son, tem­per­a­ture, wind speed, and time of day all con­tribute to a pes­ti­cide’s ex­po­sure risk pro­file, and these de­tails vary by lo­ca­tion and chem­i­cal. En­sur­ing that these pro­grams are fully funded could thus have size­able welfare benefits.[47]

Another route is tech­ni­cal in­no­va­tion in the ap­pli­ca­tion of ex­ist­ing pes­ti­cides. One promis­ing in­no­va­tion is the dropleg noz­zle, de­vel­oped be­tween 2011 and 2015 by a pub­lic-pri­vate part­ner­ship led by the Univer­sity of Ho­hen­heim. Tra­di­tional pes­ti­cide ap­pli­ca­tion tech­niques broad­cast the chem­i­cal over the top of the field. The dropleg noz­zle al­lows ap­pli­ca­tion be­neath the flower canopy, sig­nifi­cantly re­duc­ing spray drift and vir­tu­ally elimi­nat­ing the risk that pol­li­na­tors vis­it­ing flow­ers will be ex­posed to the pes­ti­cide.[48] (See Figure 18.) The dropleg noz­zle was awarded a Euro­pean Bee Award in 2017. It is already com­mer­cially available and al­most all boom sprayers can be equipped with it. Pro­mot­ing the use of dropleg noz­zles, ei­ther through sub­sidies, in­for­ma­tion cam­paigns, or reg­u­la­tory ac­tion, could thus have a sub­stan­tial im­pact.

Figure 18: Dro­pleg noz­zles in use on a rape­seed [canola] field (source: Bayer Bee Care)

Another type of tech­ni­cal in­no­va­tion im­proves co­or­di­na­tion be­tween farm­ers and bee­keep­ers. BeeCon­nected in the United King­dom is an on­line no­tifi­ca­tion tool that alerts reg­istered bee­keep­ers to planned pes­ti­cide ap­pli­ca­tions within 5 kilo­me­ters of their api­aries. In the­ory, the hives can be moved or the bees locked in­side the hives dur­ing the pe­riod of ap­pli­ca­tion to re­duce ex­po­sure. If this pro­gram proves effec­tive, it could be repli­cated in other coun­tries.

Th­ese are promis­ing solu­tions to con­crete prob­lems, and similar sorts of in­no­va­tions ought to be en­couraged and pur­sued. Ul­ti­mately, though, there is only so much that can be ac­com­plished at the level of tech­ni­cal in­no­va­tion or out­reach. A more effec­tive solu­tion is to reg­u­late, re­strict, or ban those pes­ti­cides most detri­men­tal to non-tar­get in­sects. One such edict re­cently took effect in the EU, and the con­se­quences of this de­cree should be closely mon­i­tored. In April 2018 the Euro­pean Food Safety Author­ity voted to pro­hibit the out­door use of three neon­i­cot­inoids: imi­da­clo­prid, cloth­i­ani­din, and thi­amethoxam.[49] If this ban sig­nifi­cantly im­proves the welfare of in­sect pol­li­na­tors, an­i­mal ad­vo­cates could push for similar leg­is­la­tion in other re­gions. The Xerces So­ciety, an Amer­i­can in­ver­te­brate con­ser­va­tion group, main­tains an ex­ten­sive list of recom­men­da­tions to pro­tect pol­li­na­tors from neon­i­cot­inoids, in­clud­ing im­prov­ing U.S. fed­eral and state reg­u­la­tions. Depend­ing on the out­come of the 2020 U.S. pres­i­den­tial elec­tion, a tar­geted part­ner­ship with the Xerces So­ciety or the Na­tional Re­sources Defense Coun­cil could help con­vince the U.S. En­vi­ron­men­tal Pro­tec­tion Agency to en­act a Euro­pean-style ban on neon­i­cot­inoids or at least close the con­di­tional reg­is­tra­tion loop­hole through which many neon­i­cot­inoids first en­tered the mar­ket.

One con­cern about curb­ing the use of neon­i­cot­inoids is what would re­place them. This ques­tion is rele­vant to farm­ers, who want to main­tain the same high yields that neon­i­cot­inoids al­lowed, and an­i­mal ad­vo­cates, who must en­sure that what­ever re­places neon­i­cot­inoids is not, from a welfare per­spec­tive, worse. In some in­stances it seems like neon­i­cot­inoid ap­pli­ca­tion could be halted with­out the need for any re­place­ment. Neon­i­cot­inoids are of­ten used pre­emp­tively as a seed coat, but there is re­cent ev­i­dence that the benefits of this prac­tice are neg­ligible for soy­beans (Mourtz­i­nis 2019) and pos­si­bly so for corn (Del Pozo-Val­divia 2018), the two crops that to­gether con­sti­tute the largest use of neon­i­cot­inoids in the United States. When neon­i­cot­inoids must be re­placed to main­tain high yields, there are many non-chem­i­cal al­ter­na­tives already on the mar­ket. Un­for­tu­nately, many of these non-chem­i­cal al­ter­na­tives in­volve the use of microor­ganisms, such as gran­u­lo­sis virus or the bac­te­ria Bacillus thur­ingien­sis, that, while more tar­geted than neon­i­cot­inoids and thus less likely to im­pact pol­li­na­tors, prob­a­bly in­flict con­sid­er­able suffer­ing on tar­get species. (Both meth­ods dis­rupt lar­vae di­ges­tion, slowly kil­ling tar­get in­sects over about a week.) Hap­pily, there are other tech­niques, such as semio­chem­i­cal meth­ods that dis­rupt mat­ing through the re­lease of sex pheromones, that seem hu­mane and effec­tive and are already com­mer­cially available.

As aware­ness of the harms of neon­i­cot­inoids grows, other con­trol strate­gies will likely be de­vel­oped. Jac­tel et al. 2019 write, “A num­ber of promis­ing meth­ods are still at the re­search and de­vel­op­ment stage and are hardly im­ple­mented in fields at the mo­ment. For ex­am­ple, most plant defense elic­i­tors (Bek­tas and Eul­gem, 2015), mul­ti­ple plant-based semio­chem­i­cals (Mu­rali-Baskaran et al., 2018), gene edit­ing for the de­vel­op­ment of re­sis­tant crop va­ri­eties (Lom­bardo et al., 2016) and con­ser­va­tion biolog­i­cal con­trol meth­ods (Crow­der et al., 2010) are very promis­ing new meth­ods for pest man­age­ment. How­ever, they all re­quire fur­ther de­vel­op­ment, test­ing in the field and fine-tun­ing to farm­ers’ needs be­fore re­lease onto the mar­ket” (428).[50] Some of these novel meth­ods will prob­a­bly have bet­ter welfare im­pli­ca­tions than oth­ers. Selec­tively fund­ing the re­search of those meth­ods that are likely to have the best welfare im­pli­ca­tions could both benefit com­mer­cially man­aged bees and help shift con­trol strate­gies away from the in­dis­crim­i­nate slaugh­ter of vast num­bers of in­sects.

In­crease Ac­cess to Nat­u­ral Forage

One way to im­prove the welfare of com­mer­cially man­aged honey bees is to in­crease their ac­cess to the nat­u­ral and vari­ate for­age that they need for best nu­tri­tion. Numer­ous stud­ies sug­gest colonies are more suc­cess­ful for­ag­ing on non-crop fields that in­clude a wide ar­ray of flow­ers and “nat­u­ral for­age seems to be the only re­source guaran­tee­ing ad­e­quate nu­tri­ent sup­ply and colony health” (Gar­rido & Nanetti 2019: 87). Differ­ent types of flow­ers have differ­ent nu­tri­tional pro­files and with­out ac­cess to an ap­pro­pri­ate di­ver­sity of flow­ers, bees suffer micronu­tri­ent defi­cien­cies. Colonies situ­ated in mono­cul­tural land­scapes (e.g., colonies pol­li­nat­ing large, uniform fields and or­chards) are es­pe­cially at risk of malnu­tri­tion.

There are sev­eral ways to in­crease honey bee ac­cess to nat­u­ral for­age. Bee­keep­ers can con­sider mov­ing their colonies when lo­cal flo­ral re­sources are scarce, as the stress of malnu­tri­tion seems worse than the stress of trans­port (Si­mone-Fin­strom et al. 2016). Govern­ments can set aside more land for con­ser­va­tion and habitat restora­tion in ar­eas that host large com­mer­cially man­aged bee pop­u­la­tions, such as “the North­ern Great Plains, a re­gion that har­bors ap­prox­i­mately 40% of all US bee colonies from the months of May through Oc­to­ber” (Ri­cigli­ano et al. 2019: 1). Pro­grams like the United States Depart­ment of Agri­cul­ture Con­ser­va­tion Re­serve Pro­gram have pro­vided valuable for­age for honey bees. Agri­cul­tural op­er­a­tors that re­quire pol­li­na­tion ser­vices can prac­tice bee friendly farm­ing by plant­ing a con­tin­u­ously bloom­ing plot of differ­ent flow­er­ing plants on 3-6% of their land. The Seeds for Bees pro­gram en­courages grow­ers to plant bee for­age on fal­low or un­used land, in be­tween young, non-bear­ing trees, as a cover crop be­tween rows, and along roads, fence lines, wa­ter­ways, and other or­chard mar­gins or bor­ders. (See Figure 19.) Con­sumers can sup­port these prac­tices by prefer­en­tially pur­chas­ing prod­ucts that have been Bee Bet­ter Cer­tified. Col­lab­o­ra­tion with or­ga­ni­za­tions like Pol­li­na­tor Part­ner­ship, Pro­ject Apis m., and the Xerces So­ciety could help with these efforts.

Figure 19: Bee For­age Between Al­mond Tree Rows (source: “Cover Crops Com­ple­ment, Not Com­pete with Al­mond Bloom”)

In­creas­ing nat­u­ral for­age available to honey bees is re­lated to but im­por­tantly dis­tinct from re­duc­ing pol­li­na­tion de­mand by in­creas­ing wild pol­li­na­tor pop­u­la­tions. The goal of the former in­ter­ven­tion is to main­tain nat­u­ral for­age in the vicinity of large con­cen­tra­tions of com­mer­cially man­aged bees, ei­ther in or near the fields and or­chards they are pol­li­nat­ing, the as­so­ci­ated stag­ing ar­eas, or the api­aries where they are pro­duc­ing honey and over­win­ter­ing. It is likely that na­tive pol­li­na­tors will also be at­tracted to such for­age, but in­so­far as the na­tive pol­li­na­tors will com­pete for re­sources with the man­aged bees, large con­tin­gents of lo­cal pol­li­na­tors is not a de­sired out­come.

In sum, there are two broad ways to in­crease honey bee ac­cess to nat­u­ral for­age: bring the for­age to the bees or bring the bees to the for­age. There are rea­sons to be skep­ti­cal of both meth­ods. Although plant­ing and main­tain­ing nat­u­ral for­age near large con­cen­tra­tions of honey bees would un­doubt­edly im­prove their welfare, pro­mot­ing this in­ter­ven­tion may not be a cost-effec­tive use of money be­cause eco­nomic mo­ti­va­tions can prob­a­bly drive this change on their own. Bee­keep­ers already have ob­vi­ous in­cen­tives to main­tain as much nat­u­ral for­age near their api­aries as pos­si­ble. Grow­ers also already have in­cen­tives to plant for­age on their land. Be­sides aid­ing honey bees, which grow­ers rely on for pol­li­na­tion, plant­ing for­age has nu­mer­ous other benefits, such as im­prov­ing soil qual­ity, giv­ing farm­ers plenty of rea­sons to make this change.[51] And al­though bring­ing bees to nat­u­ral for­age would also im­prove their welfare, it’s not clear that this would be a net-pos­i­tive in­ter­ven­tion. Ev­i­dence is emerg­ing to sug­gest that com­mer­cially man­aged honey bees are bad for wild bees, both be­cause they com­pete for scarce re­sources and be­cause they spread par­a­sites and pathogens. Con­cern­ing par­a­site spread Gray­stock et al. 2016 writes, “One of the best ways to avoid the po­ten­tially dev­as­tat­ing effects of spillover, spillback, and fa­cil­i­ta­tion in wild bees is to pre­vent their mix­ing with man­aged bees” (72). Lo­cat­ing hives on or near pro­tected land might im­prove man­aged bee welfare at the ex­pense of wild bee welfare and thus be over­all welfare neu­tral.

Min­i­mize Har­vest Disruption

As noted above, honey har­vest­ing is an in­va­sive, par­tic­u­larly stress­ful event for man­aged bees. The hive lid is re­moved, the bees are smoked, the in­di­vi­d­ual frames are pul­led out, the bees that re­main on the frames are blown or brushed off, and re­assem­bly in­vari­ably crushes at least a few bees per hive. Bees work ex­tremely hard to keep their hives dark, hu­mid, and ster­ile. When a hive is opened, bees be­come ag­gres­sive, di­s­ori­ented, or both. Bees that sting in defense of the colony nec­es­sar­ily die. Thus, if honey could be ex­tracted with­out open­ing the hive, bees would be spared much stress and death.

In 2014 father-and-son team Cedar and Stu­art An­der­son in­vented a sys­tem called the Flow Hive, which al­lows honey to be ex­tracted from a hive with vir­tu­ally no dis­rup­tion to the bees within.[52] (See Figure 20.) The key in­no­va­tion is the Flow Frame. Flow Frames con­tain ar­tifi­cial hon­ey­comb cells, like other frames with plas­tic foun­da­tions, that the bees fill with honey. The main differ­ence is that the cells in a Flow Frame can split at the turn of a lever, form­ing a chan­nel for the honey to fol­low, pul­led down­ward by the force of grav­ity. (See Figure 21.) Ul­ti­mately the honey ex­its the hive via a small tap, where it can be col­lected in a jar or drum. The whole pro­cess does not seem to af­fect bee ac­tivity in any way, and many am­a­teur bee­keep­ers have been known to har­vest from a Flow Hive in shorts, t-shirts, and san­dals (though offi­cially this is not recom­mended). It thus seems pretty clear that Flow Hives are at least mod­estly bet­ter for bee welfare.

Figure 20: A Flow Hive 2 Arau­caria 7 Frame (source: Flow)

Figure 21: When a Flow Frame Is Oper­ated, the Hexag­o­nal Cells Split, Form­ing a Ver­ti­cal Zig-Zag Chan­nel Through Which the Honey Can Flow Down­ward (source: Flow)

Although Flow Hives can re­duce the la­bor as­so­ci­ated with honey har­vest by as much 90%, there ap­pears to be vir­tu­ally no com­mer­cial up­take of the product.[53] There are sev­eral ob­sta­cles block­ing com­mer­cial adop­tion, but the biggest bar­rier is price.[54] In the United States, a Langstroth hive re­tails for about $110 plus an­other $72 for ten frames with foun­da­tions.[55] The cheap­est Flow Hive (frames in­cluded) costs $629. It’s pos­si­ble to gain most of the welfare benefits of a Flow Hive by putting Flow Frames in a Langstroth Hive, but Flow Frames alone still cost $259.[56]

As an in­ter­ven­tion, Flow Hives would re­quire a hefty sub­sidy to make them cost-com­pet­i­tive with stan­dard Langstroth Hives. More­over, it’s im­por­tant not to over­state the welfare im­prove­ments that such a switch would bring. Flow Hives still need to be opened for in­spec­tion, just like reg­u­lar Langstroth hives, to check for par­a­sites and dis­ease. In­spec­tion fre­quency varies by lo­ca­tion and sea­son, but in gen­eral com­mer­cial pro­duc­tion hives ap­pear to re­ceive at least a quick in­spec­tion ev­ery two to six weeks. Th­ese in­spec­tions can be al­most as in­va­sive as honey har­vest.

Build EAA in Asia

The fo­cus of this re­port is dis­pro­por­tionately cen­tered on honey bees in North Amer­ica and Europe. As noted above, this em­pha­sis is the re­sult of prac­ti­cal limi­ta­tions, namely the fact that there is far more in­for­ma­tion available in English on North Amer­i­can and Euro­pean honey bees than any other re­gion. This re­gional bias is un­for­tu­nate be­cause the ma­jor­ity of man­aged honey bees are lo­cated out­side North Amer­ica and Europe. By a wide mar­gin, the largest pop­u­la­tions of man­aged honey bees are in Asia, es­pe­cially in China and In­dia.[57] In 2017, China and In­dia col­lec­tively man­aged more honey bees (~21.8 mil­lion hives) than all of Europe (~18.8 mil­lion hives) and all of North and South Amer­ica com­bined (~11.1 mil­lion hives).[58] In the long-run, helping the great­est num­ber of honey bees will re­quire im­ple­ment­ing re­forms in coun­tries where the great­est pop­u­la­tions of honey bees are lo­cated. Thus, one of the best long-term in­ter­ven­tions for man­aged honey bees might be build­ing the effec­tive an­i­mal ad­vo­cacy move­ment in Asia, es­pe­cially China and In­dia, so that in the fu­ture the move­ment is bet­ter po­si­tioned to help the largest groups of man­aged honey bees.

In­de­pen­dent of honey bees, there already ex­ist good rea­sons to build effec­tive an­i­mal ad­vo­cacy in Asia. The con­ti­nent hosts the largest pop­u­la­tions of most types of farmed an­i­mals.[59] Con­se­quently, many or­ga­ni­za­tions are already re­search­ing meth­ods to build ca­pac­ity and sup­port in Asian na­tions. The Open Philan­thropy Pro­ject re­cently granted over a mil­lion dol­lars to the Hu­mane So­ciety of In­dia.[60] An­i­mal Char­ity Eval­u­a­tors has pro­duced com­pre­hen­sive re­ports on an­i­mal ad­vo­cacy in both In­dia and China.[61] Other re­search con­sid­ers how effec­tive al­tru­ism more gen­er­ally should be de­vel­oped and ap­proached in Asia.[62] If fea­si­ble, these efforts should be ac­cel­er­ated, though of course one must tread care­fully when deal­ing with im­por­tant cul­tural, poli­ti­cal, and lin­guis­tic differ­ences.[63]

Key Ques­tions and Uncertainties

How likely is it that honey bees are sen­tient? As­sum­ing they’re sen­tient, how does their moral value com­pare to other an­i­mals?

If honey bees do not have sub­jec­tive ex­pe­riences, or if those ex­pe­riences never take on a pos­i­tive or nega­tive af­fect, then they are plau­si­bly not moral pa­tients, and there is thus no in­trin­sic rea­son to care about their wellbe­ing. There may be in­stru­men­tal rea­sons to pro­tect and pro­mote honey bee welfare (e.g., to pre­serve ac­cess to the pol­li­na­tion ser­vices they provide), but if honey bees are not sen­tient, then the jus­tifi­ca­tion for helping them will never bot­tom out in morally valuable states of bee pain or plea­sure.

Re­think Pri­ori­ties has pre­vi­ously in­ves­ti­gated the ques­tion of in­ver­te­brate sen­tience. We con­cluded that there is mod­est ev­i­dence that honey bees are sen­tient. How­ever, we are un­cer­tain what ex­act cre­dence the ev­i­dence de­mands. 2% seems low; 90% seems high. Calcu­lat­ing the cost-effec­tive­ness of honey bee in­ter­ven­tions will prob­a­bly re­quire pre­cisify­ing this num­ber or at least re­fin­ing the range.

Another open ques­tion is what moral weight we should as­sign honey bees rel­a­tive to other an­i­mals. What is one day of honey bee life worth com­pared to one day of cow life? Chicken life? Fish life? Es­ti­mates of moral weight are even more difficult to jus­tify than es­ti­mates of sen­tience. We briefly dis­cuss the is­sue in our in­ver­te­brate welfare cause pro­file, but much more re­search is needed.

Do man­aged honey bees have a bet­ter or worse life than com­pa­rable wild in­sects? Do wild in­sects lead net-nega­tive lives?

Some of the in­ter­ven­tions dis­cussed above would in­crease the pop­u­la­tion of wild in­sects. In some cases that is the ex­plicit goal, for in­stance re­duc­ing com­mer­cial pol­li­na­tion de­mand by in­creas­ing the pop­u­la­tion of lo­cal wild pol­li­na­tors. This sort of in­ter­ven­tion aims to re­place man­aged honey bee pop­u­la­tions with com­pa­rable wild in­sects. Other in­ter­ven­tions would in­di­rectly in­crease the pop­u­la­tion of wild in­sects, for in­stance re­strict­ing the use of in­dis­crim­i­nate in­sec­ti­cides. Although the goal of this sort of in­ter­ven­tion is to im­prove the welfare of man­aged bees, wild in­sect pop­u­la­tions would also rise.

Th­ese and similar in­ter­ven­tions may have un­in­tended welfare con­se­quences. If the lives of wild in­sect pol­li­na­tors are worse than the lives of man­aged honey bees, than re­plac­ing the lat­ter with the former would lead to an over­all de­cline in welfare. There are ar­gu­ments that pull in ei­ther di­rec­tion here. On the one hand, man­aged bees are kept at un­nat­u­ral den­si­ties. Their win­ter food stores are reg­u­larly robbed. They are trans­ported long dis­tances, which is stress­ful, and they have to con­tend with an­thro­pogenic par­a­site and pathogen spread. On the other hand, bee­keep­ers have strong eco­nomic in­cen­tives to keep their colonies healthy. There are no com­pa­rable guardians of wild in­sects. Eu­so­cial in­sects tend to take bet­ter care of their young than nonso­cial in­sects. In honey bees, there is a whole caste (nurse bees) ded­i­cated solely to mak­ing sure lar­vae and pu­pae are well-fed. As a re­sult, ju­ve­nile mor­tal­ity is fairly low, no more than about 30% (Al-Lawati & Bienefeld 2009). In­so­far as ju­ve­nile mor­tal­ity is a good proxy for over­all welfare, this sug­gests man­aged honey bees might lead bet­ter lives than wild nonso­cial in­sects (though of course this point doesn’t ap­ply to feral colonies of honey bees).

More wor­ry­ingly, it might be the case that in­sect lives in gen­eral are net-nega­tive.[64] If that’s the case, then in­ter­ven­tions that in­crease the over­all in­sect pop­u­la­tion are very likely to re­duce over­all welfare. I don’t think we are cur­rently in a po­si­tion to know whether in­sect lives are net-nega­tive. More re­search, mainly em­piri­cal but also philo­soph­i­cal, is needed. For­tu­nately, the ques­tion is at­tract­ing in­creas­ing at­ten­tion. For ex­am­ple, in June 2019 Kim Cud­ding­ton, an­other re­search an­a­lyst at Re­think Pri­ori­ties, re­leased a re­port on the life his­tory of her­bivorous in­sects, in­clud­ing data on lifes­pan, fe­cun­dity, and mor­tal­ity rates by cause. This re­port could serve as both a tem­plate and spring­board for fu­ture in­ves­ti­ga­tions into the over­all qual­ity of in­sect lives.[65]

How will fu­ture tech af­fect honey bees?

As with most as­pects of the world, rapidly de­vel­op­ing in­no­va­tive tech­nol­ogy has the po­ten­tial to greatly al­ter the land­scape of com­mer­cial bee­keep­ing.[66] There are smart hives in China that ac­tively reg­u­late the tem­per­a­ture and hu­midity of a hive and mon­i­tor the num­ber of bees en­ter­ing and leav­ing.[67] Such sys­tems could re­duce the need for in­va­sive on-site in­spec­tions and could po­ten­tially ad­just hive con­di­tions in re­sponse to en­vi­ron­men­tal changes.[68] Jerry Bromen­shenk, a bee re­searcher at the Univer­sity of Mon­tana, is de­vel­op­ing an app that, ac­cord­ing to this writeup in The Economist uses “ar­tifi­cial in­tel­li­gence to analyse the sound that bees are mak­ing in or­der to de­duce whether they are suffer­ing from a num­ber of mal­adies.” A bee­keeper sim­ply holds up her smart­phone for 30 sec­onds near the hive, and the trained al­gorithm re­ports the prob­a­bil­ity of var­i­ous pos­si­ble prob­lems. (See Figure 22.) If the app proves effec­tive and is widely adopted, it could sig­nifi­cantly re­duce the num­ber of in­va­sive hive in­spec­tions a colony must en­dure. MIT’s Me­di­ated Mat­ter Group runs a syn­thetic api­ary, with the goal of fa­cil­i­tat­ing “re­search into biolog­i­cally aug­mented digi­tal fabri­ca­tion with eu­so­cial in­sect com­mu­ni­ties.” This same group has launched bees into space, ap­par­ently with the ul­ti­mate goal of test­ing the fea­si­bil­ity of send­ing bees to pol­li­nate fu­ture Mar­tian crops.

Figure 22: The Bee Health Guru Smart­phone App Uses Ma­chine Learn­ing to Iden­tify Hive Ab­nor­mal­ities by their Au­dial Sig­na­tures (source: Bee Health Guru Kick­starter cam­paign)

Fi­nally, there are ma­te­ri­ally en­g­ineered ar­tifi­cial pol­li­na­tors in de­vel­op­ment and ac­cord­ing to Chechetka et al. 2017 this tech­nol­ogy “could find wide use in var­i­ous ap­pli­ca­tions in differ­ent re­search fields, in­clud­ing agri­cul­ture, biomimetic sci­ence, and robotics. In par­tic­u­lar, it should lead to the de­vel­op­ment of robotic pol­li­na­tors and help counter the prob­lems caused by the de­clin­ing [sic] hon­ey­bee pop­u­la­tions” (233-34). (See Figure 23.) This idea is ex­tremely spec­u­la­tive, and there ap­pears to be lit­tle rea­son to think the tech­nol­ogy is prac­ti­ca­ble. Th­ese ar­tifi­cial pol­li­na­tors would be in­effi­cient, eco­nom­i­cally in­vi­able, and ecolog­i­cally dam­ag­ing (Potts etl a. 2018). Still, if mi­ni­a­ture drone pol­li­na­tors were ever made safe, cheap, and effec­tive, they could rapidly re­place com­mer­cial honey bee pol­li­na­tion.

Figure 23: Con­cept Ren­der­ing of an Ar­tifi­cial Drone Pol­li­na­tor (source: Potts et al. 2018)

How will cli­mate change af­fect honey bees?

Global warm­ing will plau­si­bly have a nega­tive effect on honey bee wellbe­ing, but the ex­tent of the welfare de­cline is un­cer­tain and the meth­ods api­aries will use to cope with cli­mate change are un­known. More re­search is needed to de­ter­mine where and how honey bees will be most af­fected and how best to miti­gate these harms.

Global warm­ing could dis­rupt syn­chrony with im­por­tant flo­ral re­sources. Bees are sen­si­tive to a va­ri­ety of en­vi­ron­men­tal con­di­tions, in­clud­ing tem­per­a­ture as mea­sured by de­gree days (num­ber of days with a mean tem­per­a­ture above or be­low a cer­tain thresh­old) (Reddy, Vergh­ese, & Ra­jan 2012). If win­ters be­come in­creas­ingly milder, honey bees may ex­tend their sea­sonal for­ag­ing and brood pe­ri­ods. This is prob­le­matic if there are no or fewer flow­ers in bloom dur­ing the ex­tended sea­son. For­ag­ing and brood ac­tivity con­sumes more colony en­ergy than over­win­ter­ing be­hav­ior; win­ter food stores thus may not be suffi­cient to sus­tain the colony through the lean sea­son (Gar­rido & Nanetti 2019).

Cli­mate change may also be chang­ing the phys­iol­ogy of cer­tain flow­ers. There are wor­ry­ing re­ports that “ris­ing at­mo­spheric CO2 is re­duc­ing the pro­tein con­cen­tra­tion of a flo­ral pol­len source es­sen­tial for North Amer­i­can bees” (Ziska et al. 2016). As noted above, pol­len pro­tein is an es­sen­tial nu­tri­ent for honey bee health, and its ab­sence makes colonies more sus­cep­ti­ble to other stres­sors. In­creased fre­quency of se­vere weather like floods or droughts could also de­plete the flo­ral land­scape at crit­i­cal times for the bees.

Global warm­ing will also prob­a­bly ex­ac­er­bate the spread of par­a­sites and pathogens. The range of some honey bee pests, such as hive bee­tles, Asi­atic hor­nets, and Tropilae­laps mites, has been ex­pand­ing as tem­per­ate re­gions have warmed. More­over, many honey bee par­a­sites re­quire bee lar­vae to re­pro­duce. In­creased win­ter brood ac­tivity due to warmer weather thus in­creases the rate and sever­ity of par­a­sitic in­fes­ta­tion (Gar­rido & Nanetti 2019). Taken to­gether, these changes will pose sig­nifi­cant welfare challenges to honey bees and their care­tak­ers.

Will ad­vo­cat­ing for honey bees back­fire?

Long-term, large-scale in­ter­ven­tions to im­prove the lives of honey bees will prob­a­bly re­quire co­or­di­nated ac­tions from a di­verse set of ac­tors. Scien­tists, philan­thropists, con­ser­va­tion­ists, farm­ers, poli­ti­ci­ans, and mem­bers of the gen­eral pub­lic may all have to agree, at least to some de­gree, that honey bee welfare mat­ters. With that goal in mind, it seems nat­u­ral to cam­paign to raise aware­ness about honey bee suffer­ing. Of course, win­ning hearts and minds on this is­sue is not go­ing to be easy. But con­vinc­ing peo­ple honey bee welfare is im­por­tant may be more than just difficult. Directly ad­vo­cat­ing for honey bee welfare at this time might be ac­tively coun­ter­pro­duc­tive, both to the an­i­mal welfare cause area and effec­tive al­tru­ism more gen­er­ally.

Michael Greger (of Nutri­tion Facts fame) ar­gues force­fully that anti-honey ad­vo­cacy hurts the ve­gan move­ment. Many peo­ple ap­par­ently have trou­ble as­cribing morally valuable states to cows and pigs; the idea that bees might suffer (and that we should care about their suffer­ing) strikes these peo­ple as crazy. And if an av­er­age per­son thinks one small part of ve­gan ‘ide­ol­ogy’ is crazy, mo­ti­vated rea­son­ing will eas­ily al­low this por­tion to in­fect her/​his per­cep­tion of the rest of the ve­gan wor­ld­view. Hence, the knowl­edge that ve­g­ans care about bees may lead many peo­ple to show less com­pas­sion to­ward cows and pigs than they oth­er­wise would.

If car­ing about honey bee welfare is per­ceived as an ex­treme, un­pop­u­lar po­si­tion, then ad­vo­cat­ing for honey bee welfare may re­duce re­cep­tive­ness to other po­si­tions the EAA move­ment ad­vo­cates for, even if these other po­si­tions are per­ceived as less ex­treme. This type of phe­nomenon might be re­lated to the so­cial psy­chol­ogy con­cept mes­sage dis­crep­ancy, the ex­tent to which a mes­sage is in­con­sis­tent with ex­ist­ing at­ti­tudes. Mildly dis­crepant at­ti­tudes can in­duce at­ti­tu­di­nal and be­hav­ioral changes, but highly dis­crepant mes­sages can threaten one’s sense of self, un­der­min­ing the effi­cacy of the mes­sage (Choo 1964).[69] It is thus im­por­tant to learn more about pub­lic at­ti­tudes to­ward honey bee welfare in­ter­ven­tions and in par­tic­u­lar how malle­able these at­ti­tudes are.[70]

There is also the worry that rush­ing into an ad­vo­cacy cam­paign may cre­ate hard-to-re­verse lock-in effects. If the ini­tial mes­sage is sub­op­ti­mal, these lock-in effects can im­pose sub­stan­tial costs. Ad­vo­cates for honey bee welfare don’t want to get stuck with one way of fram­ing the is­sue (when an­other fram­ing would be more effec­tive) merely be­cause the ini­tial out­reach cam­paign hap­haz­ardly se­lected its mes­sag­ing.

For these rea­sons, it is im­por­tant to tread care­fully when try­ing to raise aware­ness about honey bee welfare. Any pub­lic cam­paign must be de­liber­ately planned and care­fully ex­e­cuted. Poorly done, such a cam­paign could prove a ma­jor set­back not only to honey bee welfare but to the broader effec­tive al­tru­ism com­mu­nity. If effec­tive al­tru­ism be­comes ir­re­vo­ca­bly as­so­ci­ated with a cause area per­ceived to be not merely ec­cen­tric, but morally per­verse, un­re­lated cause ar­eas are likely to suffer as a re­sult.


In this re­port I’ve dis­cussed the scale of honey bee man­age­ment, its pur­pose and prospects for fu­ture growth, ma­jor re­gions in which it is prac­ticed, and some fac­tors that af­fect honey bee welfare. I’ve also sketched some po­ten­tial in­ter­ven­tions and enu­mer­ated some key ar­eas of un­cer­tainty. Taken to­gether, it is as yet un­clear how much honey bee welfare can be rea­son­ably im­proved and at what cost. How­ever, con­sid­er­ing the vast num­ber of man­aged honey bees, it may be pru­dent to con­sider their con­di­tion in more de­tail at some point, if for no other rea­son than due dili­gence to avert the pos­si­bil­ity of an on­go­ing moral catas­tro­phe. For fun­ders speci­fi­cally in­ter­ested in im­prov­ing honey bee welfare, the next step would be to fi­nance a deep dive into one or more po­ten­tial in­ter­ven­tions or an in­ves­ti­ga­tion into one or more of the key ques­tions. (Ap­pendix 4 lists an as­sort­ment of other open re­search ques­tions.) For fun­ders in­ter­ested in in­ver­te­brate welfare in­ter­ven­tions but un­cer­tain about the cost-effec­tive­ness of helping honey bees, the next step would be to fi­nance more re­views of farmed in­ver­te­brates to es­tab­lish which in­ter­ven­tions can most effec­tively im­prove in­ver­te­brate welfare. For fun­ders in­ter­ested in long-term in­ver­te­brate welfare but un­cer­tain about near-term in­ter­ven­tions, the next step would be to fi­nance work fo­cused on field-build­ing (in­clud­ing aca­demic out­reach and sur­vey­ing pub­lic at­ti­tudes to­wards ne­glected an­i­mals), in­ver­te­brate sen­tience (both em­piri­cal and philo­soph­i­cal), or moral weight (i.e., how to make cross-species welfare com­par­i­sons among tax­o­nom­i­cally dis­tant an­i­mals).

Ap­pendix 1: How This Re­port Was Completed

This pro­ject be­gan in July 2019 as a re­view of farmed in­sects. After ap­prox­i­mately 25 hours of re­search and roughly 16 pages of notes, it be­came clear that farmed in­sects was far too big a sub­ject for a sin­gle re­view, so a de­ci­sion was made to re­strict the fo­cus to a sin­gle group of in­sects. Honey bees were se­lected be­cause they are the most nu­mer­ous farmed in­sect, there is more ev­i­dence that they are sen­tient than there is for other in­sects, there is a wealth of in­for­ma­tion about them, and they seem to at­tract more pub­lic sym­pa­thy than other farmed in­sects. After an­other ~15 hours of re­search ex­clu­sively on honey bees, I be­gan prepar­ing the first draft of the re­port. The first draft was com­pleted be­tween mid-Au­gust 2019 and mid-Septem­ber 2019. It took ap­prox­i­mately 60 hours to write, though many of those hours were spent on ad­di­tional re­search, not ac­tual writ­ing. Six mem­bers of the Re­think Pri­ori­ties re­search team pro­vided ex­ten­sive and very helpful feed­back on the ini­tial draft, prompt­ing an­other ~20 hours of ed­its. When the sec­ond draft was com­plete, I met with Michael Smith, a bee ex­pert at the Univer­sity of Kon­stanz.[71] Around this time, two mem­bers of the effec­tive al­tru­ism com­mu­nity ex­ter­nal to Re­think Pri­ori­ties read and com­mented on the re­port. The third and fi­nal draft of the re­port was com­pleted af­ter an­other ~5 hours of work.

Ap­pendix 2: Es­ti­mat­ing the Num­ber of Man­aged Honey Bees

The guessti­mate model for this es­ti­mate is available here. This es­ti­mate has 7 in­puts: the FAO num­ber of bee­hives, un­cer­tainty of the FAO figure, A. mel­lifera win­ter colony size, A. mel­lifera sum­mer colony size, A. cer­ana colony size, and break­down of colonies by species.

Notably, this es­ti­mate does not in­clude man­aged bees out­side the genus Apis, such as bum­ble bees (genus Bom­bus), al­falfa leaf­cut­ting bees (genus Me­gachile), alkali bees (genus No­mia), ma­son bees (genus Os­mia), and stin­gless bees (genus Melipona). Col­lec­tively, hun­dreds of billions of these non-Apis bees are man­aged wor­ld­wide for their pol­li­na­tion ser­vices or (in the rare case of Melipona bees) their honey.[72]

Ac­cord­ing to FAO data available here (pa­ram­e­ters: Live An­i­mals, World + (To­tal), Bee­hives, Stocks, 2017), there were 90,999,730 man­aged honey bee bee­hives in the world in 2017. The FAO figure is un­likely to be ac­cu­rate, but it is un­clear if it is more likely to be an over-count or un­der-count. The FAO claims that “it is not pos­si­ble to as­sess the over­all ac­cu­racy of the dataset, as the source data is largely col­lected by mem­ber coun­tries.” and that “Es­ti­mates have been made for non-re­port­ing coun­tries as well as for coun­tries re­port­ing in­com­plete data.” To ac­count for this un­cer­tainty, I re­duced the num­ber by 20% for the lower bound and in­creased the num­ber by 20% for the up­per bound.

The FAO figure in­cludes four species: Apis mel­lifera, Apis dor­sata, Apis florea, and Apis cer­ana in­dica.[73] Un­for­tu­nately, the data do not list the num­ber of hives bro­ken down by species, which is prob­le­matic be­cause the av­er­age num­ber of bees per colony varies by species. How­ever, I found no ev­i­dence to in­di­cate any large-scale pro­duc­tion of Apis dor­sata or Apis florea, so for the pur­poses of this es­ti­mate, I ig­nored those species. Ac­cord­ing to nu­mer­ous sources in­clud­ing Potts et al. 2016, A. mel­lifera (col­lo­quially known as the “Western” or “Euro­pean” honey bee) is the most com­monly man­aged bee in the world.[74] A. cer­ana (col­lo­quially known as the “Eastern” or “Ori­en­tal” honey bee) is the sec­ond most com­monly man­aged species, but it ap­pears to be man­aged ex­clu­sively in Asia. Hard data on the num­ber of A. cer­ana colonies is limited, but there are at least two mil­lion colonies in China ac­cord­ing to Wang et al. 2012. As­sum­ing that num­ber was roughly the same in 2017, that amounts to a lit­tle over 2% of global colonies. As best I can tell, A. cer­ana is only man­aged in the fol­low­ing coun­tries: Afghanistan, Bangladesh, Bhutan, China, In­dia, In­done­sia, Ja­pan, Korea, Malaysia, Myan­mar, Nepal, Pak­istan, Pa­pua New Guinea, Thailand, and Viet­nam. Taken to­gether, these coun­tries ac­counted for roughly 27% of global bee­hives in 2017. How­ever, not all of the colonies in these coun­tries are A. cer­ana; most are prob­a­bly A. mel­lifera. All in all, I es­ti­mate that A. cer­ana ac­counts for be­tween 2% and 20% of man­aged colonies wor­ld­wide, with A. mel­lifera ac­count­ing for the rest.

A fur­ther com­pli­ca­tion is that colony sizes vary sea­son­ally, with larger pop­u­la­tions main­tained in sum­mer and lower pop­u­la­tions in win­ter. Based on FAO data I es­ti­mate that be­tween 80% and 90% of man­aged bees are man­aged in the north­ern hemi­sphere. How­ever, be­cause the cold­est and hottest times of the year vary even within a hemi­sphere, not all of the colonies in a given hemi­sphere will peak and trough at the same time. All in all, I es­ti­mate that a max­i­mum of 70-80% of global colonies will peak or trough si­mul­ta­neously. For A. mel­lifera, colony sizes trough at ap­prox­i­mately 10,000 to 15,000 worker bees and peak at ap­prox­i­mately 50,000 to 60,000 worker bees (Sag­ili & Bur­gett 2011). Un­for­tu­nately, I couldn’t find any data on sea­sonal vari­a­tion for A. cer­ana colonies, but in gen­eral, A. cer­ana colonies av­er­age be­tween 20,000 and 40,000 worker bees per colony (Tej et al. 2017).[75] With these in­puts, I es­ti­mate that the global honey bee pop­u­la­tion troughs at be­tween 1.4 and 2.7 trillion adult bees and peaks at be­tween 3 and 4.8 trillion bees.

This es­ti­mate ap­plies only to adult bees and does not in­clude lar­vae or pu­pae. Although data are harder to come by, it seems the num­ber of lar­vae/​pu­pae per colony fluc­tu­ates be­tween 0 (if the weather is cold enough, egg-lay­ing effec­tively stops) and 40,000 (just be­fore peak colony size). The num­ber of lar­vae/​pu­pae does not fluc­tu­ate smoothly; there are a num­ber of lo­cal troughs and peaks in re­sponse to events like spring buildup (when win­ter bees[76] die off), swarm­ing (when a colony re­pro­duces), and fall turnover (when the colony shifts to win­ter bee phys­iol­ogy). It’s worth not­ing, how­ever, that in con­trast to other in­sects, eu­so­cial in­sects like honey bees take ex­cel­lent care of their young. There is a whole caste (nurse bees) ded­i­cated solely to mak­ing sure lar­vae and pu­pae are well-fed. As a re­sult, ju­ve­nile mor­tal­ity is fairly low, no more than about 30% (Al-Lawati & Bienefeld 2009).

A fur­ther note on method­ol­ogy: it’s im­por­tant to es­ti­mate the num­ber of bees based on bee­hives rather than honey pro­duc­tion (as is done here) for at least two rea­sons. First, not all bees are man­aged for honey. Some bees are man­aged ex­clu­sively for pol­li­na­tion ser­vices.[77] The honey these bees pro­duce may be har­vested for lo­cal con­sump­tion or left for the bees to con­sume, but ei­ther way it would not en­ter offi­cial records. Thus, calcu­lat­ing to­tal bees by honey pro­duc­tion is likely to un­der­es­ti­mate the num­ber of bees. Se­cond, the av­er­age amount of honey a bee pro­duces de­pends on the size of the colony. Hives with more bees are more effi­cient at pro­duc­ing honey. Ac­cord­ing to Sag­ili & Bur­gett 2011, “One colony of 30,000 bees pro­duces 1½ times as much honey as the sum of two colonies with 15,000 bees each” and “One colony of 45,000 bees pro­duces 1½ times as much honey as three colonies with 15,000 bees each” (2).[78] Thus, there’s no such thing as av­er­age honey per bee sim­plic­ter. Es­ti­mat­ing the av­er­age amount of honey pro­duced per bee re­quires es­ti­mat­ing the av­er­age size of a bee­hive. But if one has already es­ti­mated the av­er­age size of a bee­hive, then it’s much eas­ier to sim­ply mul­ti­ply that num­ber by the num­ber of wor­ld­wide bee­hives.

Ap­pendix 3: Pol­li­na­tion, Honey, and Other Bee Products

Pol­li­na­tion Services

Bee­keep­ing and agri­cul­ture have be­come thor­oughly in­ter­twined over the last half cen­tury (Lee, Sum­ner, & Cham­petier 2019). Potts et al. 2016 re­port that “the frac­tion of to­tal agri­cul­tural pro­duc­tion that de­pends di­rectly on pol­li­na­tors has in­creased four­fold over the last five decades” (222-223). Wild pol­li­na­tors are not up to the task of pol­li­nat­ing that much acreage. Con­se­quently, many honey bees colonies are trucked hun­dreds or thou­sands of miles ev­ery year and sub­se­quently sta­tioned in or­chards and fields dur­ing peak bloom to pol­li­nate com­mer­cial crops, thereby sig­nifi­cantly in­creas­ing crop yields. Honey bees are the pri­mary pol­li­na­tors for about 73% of the world’s cul­ti­vated crops that re­quire an­i­mal pol­li­na­tion (Reddy, Vergh­ese, & Ra­jan 2012). Honey bees are preferred pol­li­na­tors in large part be­cause of their ver­sa­tility. Un­like many pol­li­na­tors, which visit only a hand­ful of species, honey bees are in­dis­crim­i­nate, col­lect­ing pol­len and nec­tar wher­ever they are available. Bees pol­li­nate 130 crops in the United States alone and more than 400 crops wor­ld­wide (O’Toole 2008).[79] Some com­mon crops pol­li­nated by honey bees in­clude al­monds, av­o­ca­dos, blue­ber­ries, black­ber­ries, canola, co­coa, cran­ber­ries, cher­ries, cu­cum­bers, honey dew mel­ons, kiwis, pears, pump­kins, rasp­ber­ries, straw­ber­ries, and wa­ter­mel­ons.[80] In the Pa­cific North­west colonies ser­vice on av­er­age 2.4 differ­ent crops per year (Bur­gett et al. 2010 cited in Fer­rier et al. 2018: 4).

The to­tal eco­nomic value of pol­li­na­tion wor­ld­wide has been es­ti­mated at €153 billion ($169 billion), though this figure in­cludes all com­mer­cially man­aged bees, in­clud­ing man­aged bees out­side genus Apis, and also wild non-bee pol­li­na­tors (Gal­lai et al. 2009). A more re­cent es­ti­mate puts the value of com­mer­cially man­aged honey bee pol­li­na­tion at just un­der half[81] the global to­tal value of bee pol­li­na­tion across in­sect-pol­li­nated crops (Kleijn et al. 2015). But bees are not the only in­sect pol­li­na­tors.[82] Another re­cent study es­ti­mates that wild non-bee in­sects ac­count for roughly 38% of to­tal agri­cul­tural pol­li­na­tion (Rader et al. 2016). In the United States, it has been es­ti­mated that honey bee pol­li­na­tion added be­tween $11.7 and $19.2 billion in agri­cul­tural value in 2010 (Calderone 2012). Taken to­gether, I think it’s rea­son­able to es­ti­mate that com­mer­cially man­aged honey bees con­tribute roughly $50-$60 billion in global eco­nomic value via pol­li­na­tion ser­vices.[83]

In the United States, al­monds are a ma­jor and grow­ing driver of pol­li­na­tion ser­vices.[84] In 2016 al­monds ac­counted for 61% of all U.S. pol­li­na­tion ser­vices and, due to higher pol­li­na­tion fees, 82% of all U.S. pol­li­na­tion ex­pen­di­tures. Pol­li­na­tion ser­vices ac­count for ap­prox­i­mately 5% of to­tal farm cost for al­monds com­pared to about 1% for most crops re­quiring pol­li­na­tion. Al­monds, which are al­most en­tirely de­pen­dent on bees for pol­li­na­tion, bloom in Fe­bru­ary, much ear­lier than most other crops. Both the num­ber of colonies and av­er­age bees per colony are at their low­est in the win­ter. As a re­sult, al­mond farm­ers pay be­tween $160 and $180 per colony, com­pared to $50-$70 per colony for ap­ples, blue­ber­ries, black­ber­ries, cher­ries, clover, cran­ber­ries, pears, and rasp­ber­ries (Fer­rier et al. 2018). Al­mond acreage in the U.S. has in­creased from roughly 610,000 bear­ing acres in 2006 to more than a mil­lion bear­ing acres in 2017 (Lee, Sum­ner, & Cham­petier 2019). Cal­ifor­nia dom­i­nates the global al­mond mar­ket. Dur­ing peak al­mond pol­li­na­tion, ap­prox­i­mately ⅓ to ½ of all U.S. honey bees are con­cen­trated in Cal­ifor­nia’s Cen­tral Valley, widely con­sid­ered the largest an­nual man­aged pol­li­na­tion event in the world. (See Figure 24.)

Figure 24: Michi­gan Bees Pol­li­nate Cal­ifor­nia Al­monds (source: “Cal­ifor­nia Al­monds Are Back After Four Years of Bru­tal Drought”)

Pol­li­na­tion ser­vices are in high­est de­mand in East Asia and South Asia (see Figure 25), in part due to the greater abun­dance of pol­li­na­tion-de­pen­dent crops. De­mand for pol­li­na­tion ser­vices thus seems to be an im­por­tant con­trib­u­tor to the growth in Asian bee­keep­ing. In my cur­sory sur­vey, I was not able to find much in­for­ma­tion on the Asian mar­ket for pol­li­na­tion ser­vices. If one were to do a more thor­ough re­view of man­aged honey bees, one would want to un­der­stand the ma­jor fac­tors af­fect­ing de­mand for pol­li­na­tion ser­vices in Asia.

Figure 25: Global Pol­li­na­tion De­mand, 2009 (source: Potts et al. 2016)


Honey is pro­duced from gath­ered flower nec­tar that is pro­cessed in the hive and de­posited into a hon­ey­comb. When most of the wa­ter has evap­o­rated from the hon­ey­comb, worker bees seal each cell by se­cret­ing a liquid from their ab­domen that hard­ens into beeswax. To har­vest the honey, bee­keep­ers first scrape off the wax caps that seal the honey in the hon­ey­comb frames. The frames are then placed in a cen­trifuge that spins the frames, forc­ing the honey out of the comb. The ex­tracted honey is strained and some­times heated, then bot­tled and shipped to mar­ket.

Ac­cord­ing to FAO data available here (pa­ram­e­ters: Live­stock Pri­mary, World + (To­tal), Honey nat­u­ral, Pro­duc­tion Quan­tity, 2017), 1,860,712 met­ric tons of honey were pro­duced in 2017. China dom­i­nates global honey pro­duc­tion. It pro­duced 543,000 met­ric tons in 2017. That’s more than four times as much as the next biggest pro­ducer, Turkey (2017 pro­duc­tion: 114,471 met­ric tons). (See Figure 26.) Cu­ri­ously, In­dia ranks 8th wor­ld­wide in honey pro­duc­tion (2017 pro­duc­tion: 64,981 met­ric tons), de­spite rank­ing 1st in num­ber of bee­hives. A par­tial ex­pla­na­tion is that In­dia dom­i­nates global beeswax pro­duc­tion (see be­low). Con­versely, the United States ranks 10th in to­tal hives but 5th in honey pro­duc­tion.

Figure 26: 2017 Honey Top Pro­duc­ers, Met­ric Tons (source: FAOSTAT)

Other Bee Products

There are a hand­ful of other honey bee prod­ucts sold com­mer­cially, such as beeswax, royal jelly, bee venom, and propo­lis, but the quan­tity and eco­nomic im­por­tance of these prod­ucts pales in com­par­i­son to honey pro­duc­tion and pol­li­na­tion ser­vices. Of these prod­ucts, beeswax seems to be the most im­por­tant. Beeswax is an in­gre­di­ent in many prod­ucts, in­clud­ing cos­met­ics, paints, phar­ma­ceu­ti­cals, resins, glazes, and can­dles. Beeswax is pro­duced by bees to build comb to store honey and house the colony’s young. It is har­vested at the same time as the honey be­cause the beeswax caps must be re­moved in or­der to ac­cess the honey. The force of grav­ity will nat­u­rally sep­a­rate the wax and honey, af­ter which the drained wax can be washed, strained, ren­dered, and fur­ther pro­cessed de­pend­ing on de­sired fi­nal us­age.

Ac­cord­ing to FAO data available here (pa­ram­e­ters: Live­stock Pri­mary, World + (To­tal), Beeswax, Pro­duc­tion Quan­tity, 2016), 66,568 met­ric tons of beeswax were pro­duced in 2017. In­dia dom­i­nates global beeswax pro­duc­tion. It pro­duced 23,500 met­ric tons in 2016. That’s more than four times as much as Ethiopia, the next high­est pro­ducer (2016 pro­duc­tion: 5,529 met­ric tons). (See Figure 27.)

Figure 27: 2016 Beeswax Top Pro­duc­ers, Met­ric Tons (source: FAOSTAT)

Ap­pendix 4: Open Re­search Questions

  1. How many bees are trans­ported wor­ld­wide? What is the av­er­age length of trans­port? What are the trans­port con­di­tions like out­side the U.S.?

  2. How many bees die as a di­rect re­sult of trans­port? How many bee deaths are in­di­rectly at­tributable to trans­port?

  3. What fac­tors af­fect de­mand for pol­li­na­tion ser­vices in Asia?

  4. How do fungi­cides af­fect honey bee welfare?

  5. How eas­ily can home­own­ers be con­vinced to con­vert lawns to semi-nat­u­ral habitat or to mow lawns less fre­quently?

  6. How many bees die as a re­sult of poi­sonous honey sub­sti­tutes?

  7. How effec­tive are the var­i­ous ther­mal in­su­la­tion im­prove­ments (such as wrap­ping hives in black plas­tic tar pa­per) that bee­keep­ers ap­ply dur­ing win­ter in cold ar­eas? What other ways can ther­mal in­su­la­tion be im­proved?

  8. How many bee deaths are avoided due to smok­ing the bees dur­ing in­spec­tions and honey har­vest?

  9. How do ge­net­ics af­fect bee welfare?

  10. What is the dis­tri­bu­tion of views within ma­jor farmed an­i­mal ad­vo­cacy or­ga­ni­za­tions on at­tempt­ing to im­prove bee welfare (or in­ver­te­brate welfare more gen­er­ally)?

  11. How many con­ser­va­tion groups and gov­ern­men­tal or­ga­ni­za­tions are ac­tively work­ing to im­prove honey bee care? How much do they have at their dis­posal? How does the welfare effi­cacy of the var­i­ous pro­grams com­pare?

  12. Which or­ga­ni­za­tions/​re­searchers out­side North Amer­ica and Europe are work­ing on pol­li­na­tor con­ser­va­tion?

  13. What are the best in­di­ca­tors that could be used in or­der to as­sess bee welfare (e.g., cer­tain be­hav­iors, bio­chem­i­cal mark­ers, mor­tal­ity rate)?

  14. Why aren’t api­aries out­side North Amer­ica and Europe af­fected by colony col­lapse di­s­or­der?

  15. Where do the bees go when a colony suc­cumbs to colony col­lapse di­s­or­der?

  16. Globally, which api­aries are the biggest? Does the av­er­age size of an api­ary vary by re­gion?

  17. Which honey com­pa­nies are largest?

  18. How im­por­tant are bee bro­kers for im­prov­ing honey bee welfare?[85] Which bee bro­kers are biggest?

  19. From a welfare per­spec­tive, which Var­roa con­trol meth­ods are best, both for the bees and for the mites?

  20. Could food man­u­fac­tur­ers be con­vinced to sub­sti­tute other in­gre­di­ents for honey or al­monds?

  21. How many man­aged bees get dis­eases due to hom­ing er­rors? How much could this dis­ease bur­den be re­duced by al­ter­ing hive ori­en­ta­tion or color?

  22. Are grow­ers ac­tu­ally act­ing on in­cen­tives to (e.g.) plant more nat­u­ral for­age? Are there coun­ter­vailing in­cen­tives?

  23. Are bee­keep­ers ac­tu­ally act­ing on in­cen­tives to (e.g.) re­duce hom­ing er­rors? Are there coun­ter­vailing in­cen­tives?

  24. For re­forms in which fi­nan­cial in­cen­tives al­ign with bee welfare, where are the in­for­ma­tion bar­ri­ers high­est?

  25. What is the best way to dis­sem­i­nate in­for­ma­tion to bee­keep­ers?

  26. What is the pub­lic per­cep­tion to­ward in­ter­ven­tions tar­get­ing honey bee welfare (or in­ver­te­brate welfare more gen­er­ally)? How malle­able are these at­ti­tudes?

  27. Are there novel pest con­trol strate­gies cur­rently in de­vel­op­ment with pos­i­tive welfare im­pli­ca­tions whose progress to com­mer­cial de­ploy­ment could be coun­ter­fac­tu­ally ac­cel­er­ated through in­creased re­search fund­ing?

  28. Can we coun­ter­fac­tu­ally ac­cel­er­ate the im­ple­men­ta­tion of me­chan­i­cal al­ter­na­tives to honey bee pol­li­na­tion with tar­geted re­search grants?

  29. What is the ideal en­trance height for a hive? Does the ideal height vary by cli­mate?

  30. What per­centage of man­aged honey bees are a species other than Apis mel­lifera?

  31. Do differ­ent sub­species of Apis mel­lifera pose differ­ent welfare con­cerns?

  32. What is the to­tal pop­u­la­tion of feral honey bees? Of wild bees more gen­er­ally? Of in­sect pol­li­na­tors? How have these to­tals changed over time?

  33. What is the best way to build the effec­tive an­i­mal ad­vo­cacy move­ment in Asia, es­pe­cially China and In­dia?


This es­say is a pro­ject of Re­think Pri­ori­ties. It was writ­ten by Ja­son Schukraft. Thanks to Kim Cud­ding­ton, Mar­cus A. Davis, Neil Dul­laghan, Travis Dynes, Per­sis Eskan­der, Kieran Greig, Peter Hur­ford, David Moss, Saulius Šimčikas, Michael Smith, Gavin Tay­lor, and Daniela R. Wald­horn for helpful feed­back. If you like our work, please con­sider sub­scribing to our newslet­ter. You can see all our work to date here.

Works Cited

Ab­bott, K. R., & Dukas, R. (2009). Honey­bees con­sider flower dan­ger in their wag­gle dance. An­i­mal Be­havi­our, 78(3), 633-635.

Ahn, K., Xie, X., Rid­dle, J., Pet­tis, J., & Huang, Z. Y. (2012). Effects of long dis­tance trans­porta­tion on honey bee phys­iol­ogy. Psy­che: A Jour­nal of En­to­mol­ogy, Vol­ume 2012, Ar­ti­cle ID 193029, 9 pages.

Al-Lawati, H., & Bienefeld, K. (2009). Ma­ter­nal age effects on em­bryo mor­tal­ity and ju­ve­nile de­vel­op­ment of offspring in the honey bee (Hy­menoptera: Ap­i­dae). An­nals of the En­to­molog­i­cal So­ciety of Amer­ica, 102(5), 881-888.

Bashir, N. Y., Lock­wood, P., Chas­teen, A. L., Nadolny, D., & Noyes, I. (2013). The ironic im­pact of ac­tivists: Nega­tive stereo­types re­duce so­cial change in­fluence. Euro­pean Jour­nal of So­cial Psy­chol­ogy, 43(7), 614-626.

Bate­son, M., De­sire, S., Gart­side, S. E., & Wright, G. A. (2011). Agi­tated hon­ey­bees ex­hibit pes­simistic cog­ni­tive bi­ases. Cur­rent Biol­ogy, 21(12), 1070-1073.

Büch­ler, R., Costa, C., Hatjina, F., An­donov, S., Meixner, M. D., Conte, Y. L., … & Drazic, M. (2014). The in­fluence of ge­netic ori­gin and its in­ter­ac­tion with en­vi­ron­men­tal effects on the sur­vival of Apis mel­lifera L. colonies in Europe. Jour­nal of Api­cul­tural Re­search, 53(2), 205-214.

Bjerga, A., Co­hen, D., & Hoff­man, C. (2018). Cal­ifor­nia Al­monds Are Back After Four Years of Bru­tal Drought. Bloomberg.

Brod­schnei­der, R., & Crailsheim, K. (2010). Nutri­tion and health in honey bees. Api­dolo­gie, 41(3), 278-294.

Bur­gett, M., Daberkow, S., Rucker, R., & Thur­man, W. (2010). US pol­li­na­tion mar­kets: Re­cent changes and his­tor­i­cal per­spec­tive. Amer­i­can Bee Jour­nal, 150(1):35-41.

Calderone, N. W. (2012). In­sect pol­li­nated crops, in­sect pol­li­na­tors and US agri­cul­ture: trend anal­y­sis of ag­gre­gate data for the pe­riod 1992–2009. PloS One, 7(5), e37235.

Chechetka, S. A., Yu, Y., Tange, M., & Miyako, E. (2017). Ma­te­ri­ally en­g­ineered ar­tifi­cial pol­li­na­tors. Chem, 2(2), 224-239.

Cheng, K., Peña, J., Porter, M. A., & Ir­win, J. D. (2002). Self-con­trol in hon­ey­bees. Psy­cho­nomic Bul­letin & Re­view, 9(2), 259-263.

Chit­tka, L., & Niven, J. (2009). Are big­ger brains bet­ter?. Cur­rent Biol­ogy, 19(21), R995-R1008.

Choo, T. H. (1964). Com­mu­ni­ca­tor cred­i­bil­ity and com­mu­ni­ca­tion dis­crep­ancy as de­ter­mi­nants of opinion change. The Jour­nal of so­cial psy­chol­ogy, 64(1), 65-76.

Court­ney, R. & Mul­li­nax, TJ. (2016). No bees, but a lot of buzz about ar­tifi­cial pol­li­na­tion. Good Fruit Grower.

Dainat, B., vanEn­gels­dorp, D., & Neu­mann, P. (2012). Colony col­lapse di­s­or­der in Europe. En­vi­ron­men­tal Micro­biol­ogy Re­ports, 4(1), 123-125.

DeGrandi-Hoff­man, G., Chen, Y., Rivera, R., Car­roll, M., Cham­bers, M., Hi­dalgo, G., & De Jong, E. W. (2016). Honey bee colonies pro­vided with nat­u­ral for­age have lower pathogen loads and higher over­win­ter sur­vival than those fed pro­tein sup­ple­ments. Api­dolo­gie, 47(2), 186-196.

Del Pozo-Val­divia, A. I., Reisig, D. D., Arel­lano, C., & Heiniger, R. W. (2018). A case for com­pre­hen­sive analy­ses demon­strated by eval­u­at­ing the yield benefits of neon­i­cot­inoid seed treat­ment in maize (Zea mays L.). Crop Pro­tec­tion, 110, 171-182.

Dynes, T. L., Berry, J. A., De­la­plane, K. S., Brosi, B. J., & de Roode, J. C. (2019). Re­duced den­sity and vi­su­ally com­plex api­aries re­duce par­a­site load and pro­mote honey pro­duc­tion and over­win­ter­ing sur­vival in honey bees. PloS One, 14(5), e0216286.

Fer­rier, P. M., Rucker, R. R., Thur­man, W. N., & Bur­gett, M. (2018). Eco­nomic effects and re­sponses to changes in honey bee health (No. 1477-2018-5465), ERR-246, U.S. Depart­ment of Agri­cul­ture, Eco­nomic Re­search Ser­vice, March 2018.

Ful­ton, J., Nor­ton, M., & Shilling, F. (2019). Water-in­dexed benefits and im­pacts of Cal­ifor­nia al­monds. Ecolog­i­cal Indi­ca­tors, 96, 711-717.

Gage, S. L., Ahu­mada, F., Rivera, A., Gra­ham, H., & DeGrandi-Hoff­man, G. (2018). Smoke Con­di­tions Affect the Re­lease of the Venom Dro­plet Ac­com­pa­ny­ing St­ing Ex­ten­sion in Honey Bees (Hy­menoptera: Ap­i­dae). Jour­nal of In­sect Science, 18(4), 7.

Gal­lai, N., Salles, J. M., Set­tele, J., & Vais­sière, B. E. (2009). Eco­nomic val­u­a­tion of the vuln­er­a­bil­ity of world agri­cul­ture con­fronted with pol­li­na­tor de­cline. Ecolog­i­cal Eco­nomics, 68(3), 810-821.

Gar­rido, C., & Nanetti, A. (2019). Welfare of Man­aged Honey Bees. In: Carere C., Mather J. (eds) The Welfare of In­ver­te­brate An­i­mals (pp. 69-104). An­i­mal Welfare, vol 18. Springer, Cham.

Gray­stock, P., Blane, E. J., McFred­er­ick, Q. S., Goul­son, D., & Hughes, W. O. (2016). Do man­aged bees drive par­a­site spread and emer­gence in wild bees?. In­ter­na­tional Jour­nal for Par­a­sitol­ogy: Par­a­sites and Wildlife, 5(1), 64-75.

Green­leaf, S. S., & Kre­men, C. (2006). Wild bees en­hance honey bees’ pol­li­na­tion of hy­brid sun­flower. Pro­ceed­ings of the Na­tional Academy of Sciences, 103(37), 13890-13895.

Haber, A. I., Stein­hauer, N. A., & van En­gels­dorp, D. (2019). Use of Chem­i­cal and Non­chem­i­cal Meth­ods for the Con­trol of Var­roa de­struc­tor (Acari: Var­roidae) and As­so­ci­ated Win­ter Colony Losses in US Bee­keep­ing Oper­a­tions. Jour­nal of Eco­nomic En­to­mol­ogy.

Hop­wood, J. (2017). Pol­li­na­tor con­ser­va­tion at 60 mph. Com­mit­tee on Ecol­ogy and Trans­porta­tion Newslet­ter.

Jac­tel, H., Ver­heggen, F., Thiéry, D., Es­co­bar-Gu­tiér­rez, A. J., Gachet, E., Des­neux, N., & Neon­i­cot­inoids Work­ing Group. (2019). Alter­na­tives to neon­i­cot­inoids. En­vi­ron­ment In­ter­na­tional, 129, 423-429.

Kleijn, D., Win­free, R., Bar­tomeus, I., Car­valheiro, L. G., Henry, M., Isaacs, R., … & Rick­etts, T. H. (2015). De­liv­ery of crop pol­li­na­tion ser­vices is an in­suffi­cient ar­gu­ment for wild pol­li­na­tor con­ser­va­tion. Na­ture Com­mu­ni­ca­tions, 6, 7414.

Lee, H., Sum­ner, D. A., & Cham­petier, A. (2018). Pol­li­na­tion Mar­kets and the Cou­pled Fu­tures of Al­monds and Honey Bees: Si­mu­lat­ing Im­pacts of Shifts in De­mands and Costs. Amer­i­can Jour­nal of Agri­cul­tural Eco­nomics, 101(1), 230-249.

Ler­man, S. B., Con­tosta, A. R., Milam, J., & Bang, C. (2018). To mow or to mow less: Lawn mow­ing fre­quency af­fects bee abun­dance and di­ver­sity in sub­ur­ban yards. Biolog­i­cal Con­ser­va­tion, 221, 160-174.

Long, E. Y., & Krupke, C. H. (2016). Non-cul­ti­vated plants pre­sent a sea­son-long route of pes­ti­cide ex­po­sure for honey bees. Na­ture Com­mu­ni­ca­tions, 7, 11629.

Lundin, O., Rundlöf, M., Smith, H. G., Fries, I., & Bom­marco, R. (2015). Neon­i­cot­inoid in­sec­ti­cides and their im­pacts on bees: a sys­tem­atic re­view of re­search ap­proaches and iden­ti­fi­ca­tion of knowl­edge gaps. PLoS One, 10(8), e0136928.

Lundin, O., Ward, K. L., Artz, D. R., Boyle, N. K., Pitts-Singer, T. L., & Willi­ams, N. M. (2017). Wild­flower plant­ings do not com­pete with neigh­bor­ing al­mond or­chards for pol­li­na­tor vis­its. En­vi­ron­men­tal En­to­mol­ogy, 46(3), 559-564.

Mao, W., Schuler, M. A., & Beren­baum, M. R. (2013). Honey con­stituents up-reg­u­late detox­ifi­ca­tion and im­mu­nity genes in the west­ern honey bee Apis mel­lifera. Pro­ceed­ings of the Na­tional Academy of Sciences, 110(22), 8842-8846.

Martin, S. J., & Bret­tell, L. E. (2019). De­formed wing virus in Honey­bees and Other In­sects. An­nual Re­view of Virol­ogy, 6.

Melicher, D. (2019). Road Trip: How Hive Trans­porta­tion Puts Stress on Honey Bees. En­to­mol­ogy To­day.

Melicher, D., Wil­son, E. S., Bow­sher, J. H., Peter­son, S. S., Yocum, G. D., & Rine­hart, J. P. (2019). Long-Dis­tance Trans­porta­tion Causes Tem­per­a­ture Stress in the Honey Bee, Apis mel­lifera (Hy­menoptera: Ap­i­dae). En­vi­ron­men­tal En­to­mol­ogy, 48(3), 691-701.

Mitchell, D. (2016). Ra­tios of colony mass to ther­mal con­duc­tance of tree and man-made nest en­clo­sures of Apis mel­lifera: im­pli­ca­tions for sur­vival, clus­ter­ing, hu­midity reg­u­la­tion and Var­roa de­struc­tor. In­ter­na­tional jour­nal of biome­te­o­rol­ogy, 60(5), 629-638.

Mota, T., Giurfa, M., & San­doz, J. C. (2011). Color mod­u­lates ol­fac­tory learn­ing in hon­ey­bees by an oc­ca­sion-set­ting mechanism. Learn­ing & Me­mory, 18(3), 144-155.

Mourtz­i­nis, S., Krupke, C. H., Esker, P. D., Varen­horst, A., Ar­ne­son, N. J., Bradley, C. A., … & Kan­del, Y. R. (2019). Neon­i­cot­inoid seed treat­ments of soy­bean provide neg­ligible benefits to US farm­ers. Scien­tific Re­ports, 9(1), 1-7.

Mtaita, T. A. (2003). Food. In Field Guide to Ap­pro­pri­ate Tech­nol­ogy (pp. 277-480). Aca­demic Press.

Omholt, S. W., & Am­dam, G. V. (2004). Epi­ge­netic reg­u­la­tion of ag­ing in hon­ey­bee work­ers. Science of ag­ing knowl­edge en­vi­ron­ment: SAGE KE, 2004(26), pe28-pe28.

O’Toole, C., (2008). For­ward. In: James, R.R., Pitts-Singer, T.L. (Eds.), Bee Pol­li­na­tion in Agri­cul­tural Ecosys­tems, Oxford Univer­sity Press, New York, pp. V–vii.

Perry, C. J., & Bar­ron, A. B. (2013). Honey bees se­lec­tively avoid difficult choices. Pro­ceed­ings of the Na­tional Academy of Sciences, 110(47), 19155-19159.

Pet­tis, J. S., John­son, J., & Dively, G. (2012). Pes­ti­cide ex­po­sure in honey bees re­sults in in­creased lev­els of the gut pathogen Nosema. Natur­wissen­schaften, 99(2), 153-158.

Potts, S. G., Im­per­a­triz-Fon­seca, V., Ngo, H. T., Aizen, M. A., Bies­meijer, J. C., Breeze, T. D., … & Van­ber­gen, A. J. (2016). Safe­guard­ing pol­li­na­tors and their val­ues to hu­man well-be­ing. Na­ture, 540 (7632), 220.

Potts, S. G., Neu­mann, P., Vais­sière, B., & Vereecken, N. J. (2018). Robotic bees for crop pol­li­na­tion: Why drones can­not re­place bio­di­ver­sity. Science of the To­tal En­vi­ron­ment, 642, 665-667.

Rader, R., Bar­tomeus, I., Garibaldi, L. A., Gar­ratt, M. P., Howlett, B. G., Win­free, R., … & Bom­marco, R. (2016). Non-bee in­sects are im­por­tant con­trib­u­tors to global crop pol­li­na­tion. Pro­ceed­ings of the Na­tional Academy of Sciences, 113(1), 146-151.

Reddy, P. R., Vergh­ese, A., & Ra­jan, V. V. (2013). Po­ten­tial im­pact of cli­mate change on hon­ey­bees (Apis spp.) and their pol­li­na­tion ser­vices. Pest Man­age­ment in Hor­ti­cul­tural Ecosys­tems, 18(2), 121-127.

Ri­cigli­ano, V. A., Mott, B. M., Maes, P. W., Floyd, A. S., Fitz, W., Copeland, D. C., … & An­der­son, K. E. (2019). Honey bee colony perfor­mance and health are en­hanced by api­ary prox­im­ity to US Con­ser­va­tion Re­serve Pro­gram (CRP) lands. Scien­tific Re­ports, 9(1), 4894.

Sag­ili, R. R., & Bur­gett, D. M. (2011). Eval­u­at­ing honey bee colonies for pol­li­na­tion: a guide for com­mer­cial grow­ers and bee­keep­ers. PNW. 623: 1–8.

Sánchez-Bayo, F., Goul­son, D., Pen­nac­chio, F., Nazzi, F., Goka, K., & Des­neux, N. (2016). Are bee dis­eases linked to pes­ti­cides?—A brief re­view. En­vi­ron­ment In­ter­na­tional, 89, 7-11.

Seeley, T. D., & Smith, M. L. (2015). Crowd­ing hon­ey­bee colonies in api­aries can in­crease their vuln­er­a­bil­ity to the deadly ec­topar­a­site Var­roa de­struc­tor. Api­dolo­gie, 46(6), 716-727.

Si­mone-Fin­strom, M., Li-Byarlay, H., Huang, M. H., Strand, M. K., Ruep­pell, O., & Tarpy, D. R. (2016). Mi­gra­tory man­age­ment and en­vi­ron­men­tal con­di­tions af­fect lifes­pan and ox­ida­tive stress in honey bees. Scien­tific Re­ports, 6, 32023.

Tej, M. K., Aruna, R., Mishra, G., & Srini­vasan, M. R. (2017). Bee­keep­ing in In­dia. In In­dus­trial En­to­mol­ogy (pp. 35-66). Springer, Sin­ga­pore.

Tosi, S., Bur­gio, G., & Nieh, J. C. (2017). A com­mon neon­i­cot­inoid pes­ti­cide, thi­amethoxam, im­pairs honey bee flight abil­ity. Scien­tific Re­ports, 7(1), 1201.

Traynor, K. S., Pet­tis, J. S., Tarpy, D. R., Mul­lin, C. A., Fra­zier, J. L., & Fra­zier, M. (2016). In-hive Pes­ti­cide Ex­po­some: Assess­ing risks to mi­gra­tory honey bees from in-hive pes­ti­cide con­tam­i­na­tion in the Eastern United States. Scien­tific Re­ports, 6, 33207.

van En­gels­dorp, D. & Meixner, M. D. (2010). A his­tor­i­cal re­view of man­aged honey bee pop­u­la­tions in Europe and the United States and the fac­tors that may af­fect them. Jour­nal of in­ver­te­brate pathol­ogy, 103, S80-S95.

Villalo­bos, E. M. (2016). The mite that jumped, the bee that trav­eled, the dis­ease that fol­lowed. Science, 351(6273), 554-556.

Walker, M.S. (2018). Why Smok­ing Soothes the Stressed-Out Bee Hive. En­to­mol­ogy To­day.

Wang, Z. L., Liu, T. T., Huang, Z. Y., Wu, X. B., Yan, W. Y., & Zeng, Z. J. (2012). Tran­scrip­tome anal­y­sis of the Asian honey bee Apis cer­ana cer­ana. Plos One, 7(10), e47954.

Wilfert, L., Long, G., Leggett, H. C., Sch­mid-Hem­pel, P., But­lin, R., Martin, S. J. M., & Boots, M. (2016). De­formed wing virus is a re­cent global epi­demic in hon­ey­bees driven by Var­roa mites. Science, 351(6273), 594-597.

Willi­ams, G. R., Trox­ler, A., Retschnig, G., Roth, K., Yañez, O., Shut­ler, D., … & Gau­thier, L. (2015). Neon­i­cot­inoid pes­ti­cides severely af­fect honey bee queens. Scien­tific Re­ports, 5, 14621.

Yoon, Y., Gürhan‐Canli, Z., & Sch­warz, N. (2006). The effect of cor­po­rate so­cial re­spon­si­bil­ity (CSR) ac­tivi­ties on com­pa­nies with bad rep­u­ta­tions. Jour­nal of Con­sumer Psy­chol­ogy, 16(4), 377-390.

Ziska, L. H., Pet­tis, J. S., Ed­wards, J., Han­cock, J. E., Tome­cek, M. B., Clark, A., … & Pol­ley, H. W. (2016). Ris­ing at­mo­spheric CO2 is re­duc­ing the pro­tein con­cen­tra­tion of a flo­ral pol­len source es­sen­tial for North Amer­i­can bees. Pro­ceed­ings of the Royal So­ciety B: Biolog­i­cal Sciences, 283(1828), 20160414.

[1] The ter­minolog­i­cal con­sen­sus in the rele­vant liter­a­ture is that honey bees are man­aged not farmed. I have adopted this ter­minol­ogy through­out.

[2] Silk­worms are the sec­ond most nu­mer­ous type of farmed in­sects. Over­all, in­sects ap­pear to con­sti­tute the largest group of farmed in­ver­te­brates, though it’s pos­si­ble that ne­ma­todes beat them out for that honor. (Reli­able figures are hard to come by for ne­ma­todes. Ne­ma­todes are raised com­mer­cially for many rea­sons, in­clud­ing pest con­trol and soil amend­ment, and be­cause they are so small, a sin­gle box can con­tain half a billion of the crea­tures.)

[3] The ev­i­dence of sen­tience is strongest for coleoid cephalopods (oc­to­puses, cut­tlefish, squid). Group-liv­ing in­sects, of which honey bees are a part, prob­a­bly have the next best case, though one could rea­son­ably ar­gue that the ev­i­dence is slightly bet­ter for de­ca­pod crus­taceans. See our sum­mary of find­ings by taxa or the case stud­ies sec­tion in our in­ver­te­brate welfare cause pro­file.

[4] One liter­a­ture re­view doc­u­mented 59 dis­tinct be­hav­ior types in honey bees; that num­ber com­pares fa­vor­ably against many mam­malian species, such as the North Amer­i­can moose (at 22), De Brazza mon­keys (at 44), and bot­tlenose dolphins (at 123) (Chit­tka & Niven 2009). Honey bees ex­hibit more self-con­trol (defined as the ten­dency to choose large de­layed re­wards over small im­me­di­ate re­wards) than rats and pi­geons (Cheng et al. 2002). And ag­i­tated honey bees ex­hibit the same sorts of pes­simistic cog­ni­tive bi­ases that anx­ious mam­mals do (Bate­son et al. 2011).

[5] Honey bees are able to com­mu­ni­cate the dis­tance, di­rec­tion, and rel­a­tive re­ward-to-dan­ger ra­tio of nearby flower patches to hive­mates us­ing their fa­mous wag­gle dance (Ab­bott & Dukas 2009). Honey bees can come to un­der­stand that when a cer­tain color is pre­sented to them, odor A pre­dicts a su­crose re­ward and odor B does not, but when a differ­ent color is pre­sented, the re­la­tion­ship be­tween the odors is re­versed (Mota, Giurfa, & San­doz 2011). There is even ev­i­dence of metacog­ni­tive abil­ities in honey bees. Us­ing the “un­cer­tain re­sponse” paradigm, one study tasked honey bees with dis­crim­i­nat­ing be­tween two stim­uli. A cor­rect an­swer earned a re­ward (su­crose) and an in­cor­rect an­swer earned a pun­ish­ment (qui­nine). When given the choice to opt out, honey bees were found to opt out more of­ten when the trial was difficult (when the honey bee was pro­por­tionately more likely to re­ceive a pun­ish­ment than a re­ward). The honey bees on av­er­age im­proved their suc­cess-to-failure ra­tio when given the op­tion to opt out of tri­als (Perry & Bar­ron 2013).

[6] Stijn Bruers makes many of these same points in his ex­cel­lent blog post “About In­sect Welfare and How to Im­prove It.”

[7] The num­ber of honey bees varies sea­son­ally, with lower num­bers in win­ter than sum­mer, but I es­ti­mate with 90% con­fi­dence that it stays within this range.

[8] As the Asian mar­ket ma­tures, we should ex­pect a slow­down in growth rate. How­ever, it is un­clear to me how far the Asian mar­ket is from ma­tu­rity. Growth in Asia ap­pears to be slow­ing, which is some ev­i­dence that the mar­ket is near­ing ma­tu­rity.

[9] Fur­ther com­pli­cat­ing any gen­eral sum­mary, there are big differ­ences in the availa­bil­ity of in­for­ma­tion (in English) among re­gions. Asia hosts the largest pop­u­la­tions of man­aged bees, but the con­di­tions in which many of these bees are man­aged are un­clear.

[10] Modern in­car­na­tions have differ­ent di­men­sions than the 1852 ver­sion, but the same ba­sic lay­out. The age of the de­sign sug­gests honey bee hives might be ripe for an up­date.

[11] So-called “nat­u­ral bee­keep­ers” of­ten opt for foun­da­tion­less frames, which al­low bees to build their own foun­da­tion. Found­less frames are allegedly cleaner and more hy­gienic, but com­mer­cial api­aries don’t use them be­cause (1) the bees have to be closely mon­i­tored to make sure they are con­struct­ing their foun­da­tions straight and (2) the en­ergy bees use to con­struct their own foun­da­tions could in­stead be used gath­er­ing nec­tar and pol­len.

[12] It also ap­par­ently con­vinces the bees that the for­est is on fire, to which their in­stinc­tual re­sponse is to ig­nore in­trud­ers and eat as much of the colony’s honey as pos­si­ble be­fore the hive is con­sumed in flames. (The effect wears off quickly, so the to­tal amount of honey con­sumed is neg­ligible.)

[13] It would be even bet­ter if the hive did not need to be opened at all. See dis­cus­sions be­low about Flow Hives, which al­low for honey har­vest with­out open­ing the hive, and the Bee Health Guru app, which could plau­si­bly re­duce the fre­quency of hive in­spec­tions.

[14] Ac­cord­ing to Michael Smith, a bee ex­pert at the Univer­sity of Kon­stanz, 10-20 bees might be crushed in this man­ner if the bee­keeper is rush­ing; 3-5 bees might be crushed in this man­ner if the bee­keeper is be­ing care­ful.

[15] This refers to the num­ber of colonies that do not sur­vive the win­ter.

[16] In Europe, the over­win­ter mor­tal­ity rate varies from sin­gle digits in some Med­it­ter­anean coun­tries to nearly 30% in colder climes.

[17] The lifes­pan of honey bees varies by sea­son. In the sum­mer, when bees are most ac­tive, av­er­age lifes­pan can be as short as 3-4 weeks. In the win­ter, when bees are mostly dor­mant, av­er­age lifes­pan can stretch as long as 6 months (Omholt and Am­dam 2004). (Th­ese figures are for worker bees. Queens typ­i­cally live 3-5 years.)

[18] More­over, some of these sub­sti­tutes may be di­rectly poi­sonous to bees. Brod­schnei­der & Crailsheim 2010 write, “About 40% of the sug­ars found in soy­beans, which are used as pol­len sub­sti­tutes, are toxic to bees… Another sub­stance toxic to bees is hy­drox­ymethylfur­fural (HMF), formed from the acid-catal­ized de­hy­dra­tion of hex­ose sug­ars, es­pe­cially fruc­tose, and formed in honey as a re­sult of heat treat­ment or stor­age… Su­gar solu­tions con­tain­ing 150 ppm HMF re­sult in 58.7% mor­tal­ity within 20 days in caged bees, whereas solu­tions con­tain­ing 30 ppm cause a mor­tal­ity of only 15.0%, which was not sig­nifi­cantly differ­ent from the con­trols (12.5%). HMF lev­els of 30 ppm can there­fore be re­garded as safe for bees. Re­cently, the HMF con­tent of com­mer­cially available HFCS was de­ter­mined to be be­tween 3.1 and 28.7 ppm by LeBlanc et al 2009” (281).

[19] “the prac­tice of us­ing honey sub­sti­tutes is wide­spread in com­mer­cial bee­keep­ing op­er­a­tions as a cost-sav­ing mea­sure. This long­stand­ing prac­tice was adopted af­ter lab­o­ra­tory stud­ies demon­strated the ac­cept­abil­ity and nu­tri­tional equiv­alence of sub­sti­tutes. Th­ese stud­ies, how­ever, were con­ducted be­fore the in­tro­duc­tion of var­roa mites in the mid-1980s; since that in­tro­duc­tion, the pathogen load of US bees has been sub­stan­tially in­creased be­cause of the abil­ity of var­roa mites to act as vec­tors, and pes­ti­cide ex­po­sure have in­creased due to the use of in-hive aca­ri­cides and non­tar­get en­coun­ters with pes­ti­cides in agri­cul­tural fields. In view of cur­rent knowl­edge of con­tem­po­rary lev­els of honey bee ex­po­sure to pes­ti­cides and of in­creased pathogen loads caused by global­iza­tion of trade, ex­am­in­ing the abil­ity of honey and honey sub­sti­tutes to reg­u­late ex­pres­sion of detox­ifi­ca­tion and im­mu­nity genes would seem to be a high pri­or­ity” (Mao, Schuler, & Beren­baum 2013: 8845).

[20] “the nu­tri­tional qual­ity of differ­ent flow­er­ing plants differs, and honey bees tend to for­age on a large va­ri­ety of differ­ent plants if pos­si­ble (Re­quier et al. 2015b). More­over, it was shown that the qual­ity of food, not the caloric in­take, in­fluenced age­ing and lifes­pan of honey bee work­ers (Paoli et al. 2014). Re­quier et al. (2015a, b) sup­port this: in spring, oilseed rape pol­len was un­der-rep­re­sented in honey bee colonies in in­tensely man­aged agri­cul­tural land­scape. Pol­len for­agers preferred plants from nearby sem­i­nat­u­ral habitats offer­ing a big­ger di­ver­sity of plants. The api­cul­tural prac­tice of trans­port­ing honey bee colonies to mass-flow­er­ing crops for honey or pol­li­na­tion pur­poses could there­fore nega­tively im­pact the nu­tri­tional stage of the colonies if ar­eas with big­ger plant di­ver­sity are ab­sent. This may lead to malnu­tri­tion and could fur­ther af­fect colony welfare with­out be­ing no­ticed by the bee­keeper.” (Gar­rido & Nanetti 2019: 87)

[21] The lower figure as­sumes 10,800 bees (a grade B or­chard colony) per colony; the higher num­ber as­sumes 24,000 bees (a grade A field colony) per colony. It’s doubt­ful that colonies be­low ~10,000 work­ers would be trans­ported long dis­tances be­cause a colony that small isn’t par­tic­u­larly use­ful for pol­li­na­tion. It’s pos­si­ble colonies as large as 30,000 to 40,000 are trans­ported long dis­tances, but my im­pres­sion is that colonies aren’t nor­mally moved at peak ca­pac­ity. Worker bees are rel­a­tively short-lived, so gen­er­ally in­di­vi­d­ual bees won’t be trans­ported more than once in their lives.

[22] “In north­ern In­dia, com­mer­cial bee­keep­ers shift the colonies be­tween plains and hills for mi­gra­tory bee­keep­ing. Dur­ing Oc­to­ber–Novem­ber, colonies are mi­grated to the plains of Ut­taran­chal, Ut­tar Pradesh, Haryana, Pun­jab and Ra­jasthan to ex­ploit rape­seed and mus­tard. Dur­ing De­cem­ber–Jan­uary, colonies are mi­grated to eu­ca­lyp­tus plan­ta­tion of Ut­tar Pradesh and Haryana. Bee colonies will also be mi­grated to litchi or­chards at Ram­na­gar and Dehradun from Fe­bru­ary to March. Some bee­keep­ers will also mi­grate to sun­flower fields of Pun­jab and Haryana. Bee­keep­ers will also mi­grate to for­est plan­ta­tions of Ut­tar Pradesh for shisham till May. In south­ern In­dia, mi­gra­tion of bee colonies from south­ern Tamil Nadu (mainly Marthandam of Kanyaku­mari District) to Ker­ala dur­ing Jan­uary–March is a renowned prac­tice. The com­mer­cial bee­keep­ers mi­grate the colonies to rub­ber plan­ta­tions which are spread over about 0.40 mil­lion hectares. Dur­ing that pe­riod, bee­keep­ers will har­vest tons of honey and store to sell when they get bet­ter price. Rub­ber is con­sid­ered as the third ma­jor source of honey next to rape­seed/​mus­tard and sun­flower in In­dia. Bee­keep­ers from Ker­ala and Tamil Nadu mi­grate their colonies mainly to Quilon, Kot­tayam, Changanacherry, Trichur, Palghat, Kozhikode and Can­nanore dis­tricts for rub­ber–honey flow. In Tamil Nadu, dur­ing May–June, bee­keep­ers mi­grate the colonies for har­vest­ing nec­tar from tamarind flow­ers. Colonies are also mi­grated to high ranges of De­viku­lam, Peer­medu, Idukki and other dis­tricts to car­damom es­tates.” (Tej et al. 2017: 57-58).

[23] Un­re­lat­edly, queen bees are of­ten mailed in small boxes from breed­ers to bee­keep­ers. Pre­sum­ably, these bees ex­pe­rience con­sid­er­able stress in tran­sit, though ob­vi­ously this is­sue only af­fects a small num­ber of bees. More im­por­tantly, queens shipped long dis­tances may pro­duce weaker colonies. Re­cent re­search sug­gests colonies with queens of lo­cal ori­gin sur­vive longer than colonies with queens of non-lo­cal ori­gin (Büch­ler et al 2014).

[24] A hom­ing er­ror is when a for­ag­ing bee re­turns to the wrong hive. Hom­ing er­rors in­crease par­a­site and pathogen spread among colonies.

[25] He­molymph is the in­sect equiv­a­lent of blood.

[26] Although it’s only pos­si­ble to spec­u­late about the pain and suffer­ing mite at­tacks in­flict on in­di­vi­d­ual bees, the welfare im­pli­ca­tions of a large (rel­a­tive to the size of a bee) par­a­site slowly suck­ing one’s blood and fat away are surely not good. It might be com­pa­rable to a tick the size of a base­ball latch­ing on to a hu­man.

[27] Other ar­tifi­cial chem­i­cals, es­pe­cially fungi­cides, may also have a detri­men­tal effect on bees. Thanks to Michael Smith for bring­ing this to my at­ten­tion.

[28] “Although a va­ri­ety of agri­cul­tural pes­ti­cides were found at all sites, the con­tam­i­nants likely to provide the great­est haz­ard to honey bees in our study were non-agri­cul­tural pyrethroid in­sec­ti­cides tar­get­ing nui­sance pests such as mosquitoes.” (Long & Krupke 2016: 2)

[29] This refers to a bee’s abil­ity to re­turn suc­cess­fully to its home hive.

[30] “Nosema in­fec­tions in­creased sig­nifi­cantly in the bees from pes­ti­cide-treated hives when com­pared to bees from con­trol hives demon­strat­ing an in­di­rect effect of pes­ti­cides on pathogen growth in honey bees. We clearly demon­strate an in­crease in pathogen growth within in­di­vi­d­ual bees reared in colonies ex­posed to one of the most widely used pes­ti­cides wor­ld­wide, imi­da­clo­prid, at be­low lev­els con­sid­ered harm­ful to bees.” (Pet­tis et al. 2012: 153)

[31] “pes­ti­cide con­tam­i­na­tion in stored bee­bread and wax comb was as­so­ci­ated with colony mor­tal­ity and in­creased queen re­place­ment” (Traynor et al. 2016: 13)

[32] “Neon­i­cot­inoid com­pounds are used in more than 120 coun­tries with at least 140 differ­ent crop uses (e.g. soil and fo­liar ap­pli­ca­tions of the same com­pound in the same crop are defined as two differ­ent crop uses). Since their com­mer­cial in­tro­duc­tion in the early 1990s, neon­i­cot­inoids have quickly be­come the most com­monly used class of in­sec­ti­cides in the world” (Lundin et al. 2015: 2).

[33] Defined as those man­ag­ing 50 colonies or more

[34] “Ac­cu­mu­lat­ing ev­i­dence sug­gests that no sin­gle stres­sor alone is re­spon­si­ble for de­clines. Rather, it is prob­a­bly a com­bi­na­tion of abiotic and biotic fac­tors act­ing in syn­chrony, to have a nega­tive im­pact on pol­li­na­tor pop­u­la­tions” (Long & Krupke 2016: 2)

[35] The ex­act ex­tent to which this in­ter­ven­tion would re­duce par­a­site trans­mis­sion is un­known, but will hope­fully be stud­ied in more de­tail soon. From page 12 of the Dynes et al. 2019 pa­per: “To­gether these find­ings sug­gest rel­a­tively lower amounts of dis­ease trans­mis­sion, com­pared to the al­ter­na­tives of bees drift­ing back and forth be­tween na­tal and non-na­tal colonies, or bees ran­domly drift­ing to any colony which causes mites to be spread greater dis­tances from the origi­nal in­oc­u­la­tion. Although our ex­per­i­ment was not able to di­rectly quan­tify be­tween-colony trans­mis­sion of mites through drift, it is highly likely that in­creased rates of drift in­crease dis­ease trans­mis­sion.”

[36] Alter­na­tively, the con­clu­sion is false for rea­sons that are ob­vi­ous to com­mer­cial bee­keep­ers but not to me (or the re­view­ers at PLOS One).

[37] “Colonies perform­ing pol­li­na­tion ser­vices were sub­ject to in­creased pes­ti­cide ex­po­sure com­pared to honey-pro­duc­tion and hold­ing yards” (Traynor et al. 2016: 1) and “Pol­li­na­tion en­vi­ron­ments al­most in­vari­ably con­tribute more pes­ti­cide resi­dues” (Traynor et al. 2016: 14).

[38] Rough plant-based honey sub­sti­tutes already ex­ist (e.g., agave nec­tar and maple syrup). One way to re­duce de­mand for honey would be to con­vince food com­pa­nies to swap honey for a plant-based sub­sti­tute in their recipes. ProVeg has had some suc­cess con­vinc­ing com­pa­nies to adopt other ve­gan-friendly recipes. I’m un­sure about the cost-effec­tive­ness of this ap­proach.

[39] For al­tru­ists who re­frain from honey con­sump­tion on welfare grounds, it seems similar rea­son­ing ap­plies to al­monds and other crops heav­ily de­pen­dent on com­mer­cial bee pol­li­na­tion ser­vices.

[40] A pol­l­enizer is a plant that pro­vides pol­len. In com­mer­cial agri­cul­ture, the trees that are planted to pro­duce crops to be sold are of­ten dis­tinct va­ri­eties from the trees planted to serve as pol­l­eniz­ers. The pol­l­enizer va­ri­eties have been cul­ti­vated to pro­duce abun­dant, com­pat­i­ble, and vi­able pol­len, whereas the mar­ket va­ri­eties have been cul­ti­vated to max­i­mize the qual­ity of fruit/​nut and are of­ten ster­ile. For ex­am­ple, in al­mond or­chard plant­ing, the stan­dard is to al­ter­nate the Non­pareil va­ri­ety (the kind of al­mond one buys in a store) with ei­ther the Ne Plus Ul­tra or Peer­less va­ri­ety, which serve as pol­l­eniz­ers for Non­pareil.

[41] Honey bee pol­li­na­tion is also more tar­geted than the ran­dom blow­ing of me­chan­i­cal pol­li­na­tion.

[42] The pres­ence of wild pol­li­na­tors prob­a­bly also benefits other wild species. Ac­cord­ing to this re­port from the NRDC, “North Amer­ica’s ap­prox­i­mately 4,000 wild bee species help pol­li­nate… the seeds, nuts, and fruits that are con­sumed by an­i­mals from song­birds to griz­zly bears.”

[43] They add, “Mea­sures to miti­gate loss of pol­li­na­tion ser­vices are most cost effec­tive in rel­a­tively in­ten­sively farmed land­scapes be­cause here mea­sures have the high­est im­pact, ecosys­tem ser­vice de­liv­ery is likely to be re­duced ow­ing to the in­ten­sive farm­ing prac­tices, and re­turns on in­vest­ments are greater ow­ing to higher yields in in­ten­sively farmed ar­eas” (Kleijn et al. 2015: 4).

[44] Im­por­tantly, these mea­sures will only aid agri­cul­tural pol­li­na­tion if the ar­eas are lo­cated near fields or or­chards.

[45] In ad­di­tion, wild bees in­crease the effi­ciency of honey bee pol­li­na­tion for at least one crop (hy­brid sun­flower), and the au­thors re­port that “it is likely that wild bees en­hance honey bee pol­li­na­tion effi­ciency for other crops” (Green­leaf & Kre­men 2006: 13892).

[46] Of course, the an­i­mals most af­fected by pes­ti­cide use are the tar­get pop­u­la­tions kil­led by the chem­i­cals. Th­ese an­i­mals are be­yond the scope of this re­port, but for their sake we should pro­mote pes­ti­cides that are as hu­mane as pos­si­ble.

[47] Ad­vo­cat­ing for ap­pro­pri­ate pes­ti­cide la­bel­ling also falls into this cat­e­gory.

[48] Note that this in­no­va­tion is only rele­vant for con­tact pes­ti­cides. Sys­temic pes­ti­cides, which are ab­sorbed by the plant, will af­fect pol­li­na­tors no mat­ter how they are ap­plied.

[49] Ac­cord­ing to the Amer­i­can Bee Jour­nal, “Sev­eral Cana­dian provinces have also re­cently placed par­tial re­stric­tions on neon­i­cot­inoids, and some U.S. states have par­tially re­stricted the use of neon­i­cot­inoids as part of their Pol­li­na­tor Pro­tec­tion Plans.”

[50] They add, “Strength­en­ing sci­en­tific re­search and shift­ing it to­ward the needs of farm­ers will be a key challenge in the new neon­i­cot­inoid-free area that will re­quire not only biotech­nol­ogy, but also in­put from the so­cial sci­ences to ad­dress tech­nol­ogy trans­fer, as par­ti­ci­pa­tory and cit­i­zen sci­ences meth­ods are of par­tic­u­lar rele­vance for im­prov­ing the adop­tion of new bio­con­trol tech­niques through joint de­vel­op­ment with farm­ers (Wy­ck­huys et al., 2018)” (Jac­tel et al. 2019: 428).

[51] Grow­ers used to fear that nat­u­ral for­age would dis­tract bees from the al­mond blos­soms that they were rented to pol­li­nate, but re­cent re­search shows that al­though wild­flower plant­ings are highly at­trac­tive, gar­ner­ing much at­ten­tion from the bees, they do not re­duce al­mond blos­som vis­its (Lundin et al. 2017). It seems the wild­flow­ers strengthen the honey bee colonies, in­creas­ing to­tal for­ag­ing and re­duc­ing the risk of colony col­lapse.

[52] In Fe­bru­ary 2015 the An­der­sons launched a crowd­fund­ing cam­paign hop­ing to raise $70,000. They fa­mously raised over $12 mil­lion, break­ing sev­eral crowd­fund­ing records in the pro­cess.

[53] Flow Hives are mainly mar­keted to­ward am­a­teur “back­yard” bee­keep­ers, but there is a bit of effort to sell to small, bou­tique com­mer­cial api­aries.

[54] Another prob­lem is that beeswax can­not be har­vested from a Flow Frame. Ad­di­tion­ally, there seems to be a fair bit of re­sent­ment among pro­fes­sional bee­keep­ers to­wards the Flow Hive. The origi­nal ad­ver­tis­ing for the Flow Hive was slightly mis­lead­ing, giv­ing the im­pres­sion that bee­keep­ing had been rev­olu­tionized into a ba­si­cally zero-effort ac­tivity, which lead to a bunch of (lazy? ig­no­rant?) am­a­teurs get­ting into bee­keep­ing.

[55] Th­ese are con­sumer-fac­ing prices. Pre­sum­ably com­mer­cial api­aries buy in bulk and re­ceive a dis­count for it.

[56] Flow Frames also elimi­nate the need for a cen­trifuge (a type of honey ex­trac­tor), which can cost a few hun­dred dol­lars, but a cen­trifuge can be used re­peat­edly. For a com­mer­cial api­ary pro­duc­ing at scale, this seems to be a min­i­mal ex­pense rel­a­tive to the cost of the hive, which scales lin­early with num­ber of colonies.

[57] Fried bees are eaten as a del­i­cacy in parts of China. Although the num­ber of af­fected bees looks to be quite small, this prac­tice may af­fect how Chi­nese view bee welfare. (Thanks to Gavin Tay­lor for bring­ing this prac­tice to my at­ten­tion.)

[58] 2017 In­dia: 12,763,684 hives; 2017 China: 9,031,457 hives; 2017 Europe: 18,764,349 hives; 2017 Amer­i­cas: 11,139,203 hives (source: FAOSTAT)

[59] See this ACE re­port on trends in the pro­duc­tion of farmed an­i­mals.

[60] Open Phil also re­cently granted $362,000 to The Pol­li­na­tion Pro­ject to sup­port move­ment-build­ing and re-grant­ing to farm an­i­mal groups in Brazil, In­dia, Mex­ico, Thailand, and Viet­nam. (Note that the name is a metaphor: The Pol­li­na­tion Pro­ject has noth­ing to do with ac­tual plant pol­li­na­tion or pol­li­na­tor welfare.)

[61] The China re­port is only available by re­quest be­cause it con­tains con­tent that could be per­ceived as for­eign in­terfer­ence by the Chi­nese gov­ern­ment and hence have nega­tive con­se­quences for an­i­mal ad­vo­cates cur­rently op­er­at­ing in China.

[62] The au­thor of this re­port is cur­rently look­ing for fund­ing, ad­vi­sors, and col­lab­o­ra­tors.

[63] Char­ity En­trepreneur­ship has con­ducted some pre­limi­nary re­search into the rel­a­tive value of start­ing an­i­mal char­i­ties in differ­ent coun­tries which dis­cusses some of these ob­sta­cles, though I don’t en­dorse their method­ol­ogy or their con­clu­sions.

[64] See Brian To­masik’s work for the canon­i­cal rep­re­sen­ta­tion of this view.

[65] Kim’s im­pres­sion from that re­search is that at least some in­sect lives are not net-nega­tive.

[66] The pro­jects de­scribed in this sec­tion are ex­tremely spec­u­la­tive and are meant only to illus­trate the gen­eral promise of tech­nolog­i­cal in­no­va­tion.

[67] See this IoT For All ar­ti­cle for other smarthive tech­nolo­gies in de­vel­op­ment.

[68] Di­ag­nos­tic Ra­dioen­to­mol­ogy is an­other ap­proach in de­vel­op­ment that promises to re­duce the need for hive in­spec­tions. Us­ing com­put­er­ised to­mog­ra­phy (CT) scan­ning, re­searchers are able to mon­i­tor in-hive com­po­nent vol­umes and pop­u­la­tion dy­nam­ics non-in­va­sively.

[69] Here are some loose ex­am­ples of mes­sage dis­crep­ancy at work. The proso­cial ac­tivi­ties of com­pa­nies whose mo­tives are per­ceived to be in­sincere or am­bigu­ous hurts the com­pany’s image (Yoon, Gürhan‐Canli, & Sch­warz 2006). Nega­tive stereo­types of so­cial ac­tivists make peo­ple more re­sis­tant to the adop­tion of so­cial changes that the ac­tivists pro­mote (Bashir et al. 2013). Other po­ten­tial ex­am­ples in­clude pro­po­nents of mar­riage equal­ity gal­va­niz­ing op­po­si­tion, ad­ver­tis­ing for Cal­ifor­nia’s Propo­si­tion 12 re­duc­ing its ap­peal to con­ser­va­tives, and anti-to­bacco cam­paigns failing to per­suade smok­ers to give up the habit.

[70] Re­think Pri­ori­ties is cur­rently in­ves­ti­gat­ing at­ti­tudes to­ward wild an­i­mal welfare in­ter­ven­tions, and this work might be rele­vant to at­ti­tudes about in­ver­te­brate welfare.

[71] The meet­ing with Smith should not be con­strued as Smith’s en­dorse­ment of any part of this re­port. Notwith­stand­ing his gra­cious as­sis­tance, any er­rors and non-sci­en­tific value judg­ments re­main my own. I thank Gavin Tay­lor for helping to ar­range this meet­ing.

[72] The an­nual pro­duc­tion of Melipona honey is triv­ial com­pared to Apis honey. Fol­low­ing stan­dard con­ven­tion, when I re­fer to “honey bees,” I re­fer ex­clu­sively to Apis bees.

[73] It is un­clear to me what con­di­tions the use of one species rather than an­other. A more thor­ough re­view would ac­count for pos­si­ble welfare differ­ences among the species.

[74] There are sev­eral sub­species within Apis mel­lifera, each with slightly differ­ent char­ac­ter­is­tics. For ex­am­ple, in many parts of Africa the most com­monly man­aged bee is Apis mel­lifera scutel­lata (the so-called “Afri­can bee”), which is more ag­gres­sive than sub­species man­aged el­se­where.

[75] For con­ve­nience I’ve ex­cluded drones (male bees) from these figures; drones never num­ber more than a few hun­dred per colony, gen­er­ally only ap­pear­ing in late spring or early sum­mer.

[76] The lifes­pan of a worker bee varies by sea­son. Win­ter bees are longer-lived than sum­mer bees.

[77] Ac­cord­ing to Fer­rier et al. 2018, in 2016 of the slightly more than 3 mil­lion colonies in the United States, roughly 400,000 were used ex­clu­sively for pol­li­na­tion ser­vices (6).

[78] Although the au­thors are not ex­plicit, these figures re­fer to mel­lifera bees.

[79] “Now we are mostly not hunter-gath­er­ers and we have pro­duced new habitats, the agroe­cosys­tems, where in­ten­sive agri­cul­ture has re­sulted in crop yields un­dreamed of a cou­ple of gen­er­a­tions ago. In so do­ing, we have not eman­ci­pated our­selves from de­pen­dency on bees: we rely on them to pol­li­nate 63 of the 82 (77%) most valuable crops. Wor­ld­wide, bees pol­li­nate more than 400 crop species and in the United States more than 130 crop species” (“For­ward,” vi).

[80] The ex­tent to which these crops de­pend on bees for pol­li­na­tion varies. Some, like al­monds, are al­most en­tirely de­pen­dent on bees for pol­li­na­tion. For oth­ers, like straw­ber­ries, bee pol­li­na­tors merely in­crease yields.

[81] $2,913 per hectare for man­aged honey bees vs $3,251 per hectare for wild bees

[82] Some com­mon non-bee in­sect pol­li­na­tors in­clude blow flies, but­terflies, hover flies, moths, wasps, and in rare cases ants (man­goes) and bee­tles (pomegranate).

[83] Note that this is not an es­ti­mate of to­tal rev­enue re­ceived for pol­li­na­tion ser­vices, which, as far as I can tell, is about an or­der of mag­ni­tude smaller. This sug­gests that pol­li­na­tion ser­vices may be cur­rently un­der-priced.

[84] Al­monds are also a ma­jor driver of de­mand for pol­li­na­tion ser­vices in Aus­tralia.

[85] Bee bro­kers ar­range for pol­li­na­tion ser­vices.