ARTICLE
Quan i ying he impac o an in asi e ho ne on
Bombus e es is colonies
Thomas A. O’Shea-Whelle 1✉, Robin J. Cu is1, Pe e J. Kennedy 1, Ellen K. J. G oom1, Julie e Poida z1,2,
Da id S. Ra fle1, Sand a V. Rojas-Nossa3, Ca olina Ba olomé 4, Damián Dasil a-Ma ins4, Xulio Maside4,
Salus iano Ma o3& Julie L. Osbo ne 1
The in asi e ho ne Vespa elu ina nig i ho ax is conside ed a p oli e a ing h ea o pollina o s
in Eu ope and Asia. While he impac o his species on managed honey bees is well-
documen ed, e ec s upon o he pollina o popula ions emain poo ly unde s ood. None-
heless, die a y analyses indica e ha he ho ne s consume a di e si y o p ey, uelling
conce ns o a - isk axa. He e, we quan i y he impac o V. elu ina upon s anda dised
comme cially- ea ed colonies o he Eu opean bumblebee, Bombus e es is e es is. Using a
landscape-scale expe imen al design, we deploy colonies ac oss a g adien o local V. elu ina
densi ies, u ilising au oma ed acking o non-in asi ely obse e bee and ho ne beha iou ,
and quan i y subsequen e ec s upon colony ou comes. Ou esul s demons a e ha ho -
ne s equen ly hun a B. e es is colonies, being p e e en ially a ac ed o hose wi h high
o aging a fic, and engaging in epea ed—ye en i ely unsuccess ul—p eda ion a emp s a
nes en ances. No ably howe e , we show ha B. e es is colony weigh s a e nega i ely
associa ed wi h local V. elu ina densi ies, indica ing po en ial indi ec e ec s upon colony
g ow h. Taken oge he , hese findings p o ide he fi s empi ical insigh in o impac s on
bumblebees a he colony le el, and in o m u u e mi iga ion e o s o wild and managed
pollina o s.
h ps://doi.o g/10.1038/s42003-023-05329-5 OPEN
1En i onmen and Sus ainabili y Ins i u e, Uni e si y o Exe e , Pen yn, Co nwall TR109FE, Uni ed Kingdom. 2UMR PVBMT, Peuplemen s Végé aux e
Bioag esseu s en Milieu T opical, CIRAD, 97410 Sain Pie e, La Réunion, F ance. 3Depa men o Ecology and Animal Biology, Facul y o Sciences, Uni e si y
o Vigo, 36310 Vigo, Pon e ed a, Spain. 4G upo de Medicina Xenómica, CIMUS, Uni e sidade de San iago de Compos ela, 15781 San iago de Compos ela, A
Co uña, Spain. ✉email: [email p o ec ed]
COMMUNICATIONS BIOLOGY | (2023) 6:990 | h ps://doi.o g/10.1038/s42003-023-05329-5 | www.na u e.com/commsbio 1
1234567890():,;
The main enance o pollina ion se ices is a global conce n,
as cu en es ima es indica e ha ~87.5% o all e es ial
angiospe ms ely upon animals o ep oduc ion1. When
conside ing ag icul u e, his eliance is u he in ensified o a
ela i ely cons ained g oup o insec axa, wi h bees alone isi -
ing >90% o leading c op ypes2,3. As such, educ ions in polli-
na o heal h cons i u e a h ea o bo h biodi e si y and ood
secu i y a he landscape-scale, wi h consequen e o s being
made o conse e a a ie y o a - isk species4,5. Among hese, wild
bees a e o specific conce n, as hey do no benefi om
an h opogenic managemen 6and compa a i ely li le da a exis s
o moni o na u al popula ion changes7,8. While subs an ial
e o s ha e been made o cha ac e ise he d i e s o pollina o
declines—wi h habi a loss eme ging as a p ima y h ea 2,7,9—
he e emains significan unce ain y ega ding a ple ho a o
addi ional p ospec i e ac o s, indica ing he need o u he
esea ch7,9,10.
In e es ingly, in assessmen s o such ac o s, he impac o
in asi e species is equen ly anked below o he h ea s o pol-
lina o s, os ensibly due o a pauci y o da a2,7. While cau ion is
wa an ed, i is impo an o no e ha a deficiency o da a is no
an amoun o an absence o e ec , and ins ead unde lines he
need o addi ional in es iga ion10,11. This poin is especially
pe inen when conside ing ha o he d i e s o pollina o
decline wi h compa a i ely li le quan i a i e e idence appea o
benefi subs an ially om he ‘p ecau iona y p inciple’, combined
wi h an h opocen ic bias7,10,12. Ne e heless, he e is a g owing
body o e idence ha in asi e o ganisms exe p o ound, hough
o en poo ly- isible, impac s upon wild pollina o s and pollina-
ion se ices13–15. As such, basic empi ical da a is u gen ly nee-
ded o be e iden i y and mi iga e he isks posed by hese
o ganisms.
An eme ging global h ea is ha o he ho ne Vespa elu ina
nig i ho ax—colloquially known as he yellow-legged o Asian
ho ne — he fi s insec o be legisla i ely classified as an in asi e
alien species o conce n in Eu ope16. This social espid is na i e
o sou heas Asia, bu has sp ead apidly h ough pa s o eas
Asia and Eu ope, ollowing acciden al in oduc ions in o be o e
2003 and 2004 espec i ely17–19. The in asion on has expanded
a a a e o up o 78 km pe yea in some egions18,20, acili a ed
by an adap able li e-his o y21, apid colony popula ion g ow h18,
and he abili y o ma ed queens o a el subs an ial dis ances in a
single day22. Con ol e o s a e u he hampe ed by he di fi-
cul y o loca ing nes s20,23, and high po en ial o addi ional
human-media ed dispe sal e en s24, as is he case wi h many
o he in asi e eusocial insec s25–28. Consequen ly, e o s o
con ain he sp ead o hese ho ne s ha e had limi ed success18,21,
while clima ic niche modelling p edic s hei con inued expansion
in o new egions29,30.
The capaci y o V. elu ina o impac insec pollina o s is
e iden h ough hei highly- isible and in ense p eda ion o he
honey bee Apis melli e a21. While hese ho ne s a e na u al p e-
da o s o he congene ic Apis ce ana—which has consequen ly
e ol ed a sui e o de ensi e beha iou s31—A. melli e a cons i u es
an e olu iona ily naï e p ey species, and hus has ew e ec i e
de ence mechanisms, enabling he ho ne s o p eda e upon
colonies wi h nea -impuni y31–33. As such, p eda ion p essu e
can become so se e e ha o aging is ma kedly cu ailed, leading
o he e en ual s a a ion o colonies, wi h epo ed mo ali y
a es exceeding 30% in some egions21.
Al hough managed honey bees ha e occupied much o he
esea ch ocus, p elimina y die a y analyses p o ide s ong e i-
dence ha V. elu ina also p eys upon a di e si y o wild
pollina o s34,35. Indeed, p e ious wo k indica es ha bo h ho ne
p eda ion and o aging ac i i y exe di ec e ec s on plan -
pollina o communi ies36. This is conce ning, as he b oade
impac o V. elu ina upon he su i al and ep oduc ion o wild
pollina o s emains la gely unknown21. Consequen ly, gi en he
scale o damage w ough by o he in asi e Vespidae37, he e is a
p essing need o obus da a o b idge cu en knowledge gaps.
One g oup ha may be pa icula ly ulne able a e eusocial
bumblebees o he genus Bombus. This is because hei colony-
based li ecycle has he po en ial o a ac V. elu ina in a simila
manne o ha o A. melli e a; namely by o e ing a dense
agg ega ion o po en ial p ey a colony en ances. Fu he , he
subs an ially smalle colony sizes o Bombus spp38., coupled wi h
ongoing wild popula ion declines in Eu ope39–41, a guably place
bumblebees a e en g ea e isk han managed honey bees. Cu -
en ly, he po en ial impac s o V. elu ina upon bumblebee
colonies emain la gely unknown, howe e , he p esence o
Bombus in he ho ne s’die highligh s he need o u he
inqui y34.
He e, we in es iga ed he e ec s o V. elu ina upon colonies o
he na i e Eu opean bumblebee Bombus e es is e es is. U i-
lising a landscape-scale expe imen al design, we es ablished
s anda dised B. e es is colonies a field si es ac oss Pon e ed a,
Spain—a egion wi h subs an ial he e ogenei y in local V. elu ina
densi ies. Colonies we e fi ed wi h au oma ed came a moni o ing
sys ems (Fig. 1a, b), allowing o he quan ifica ion o ho ne and
bumblebee ac i i y emo ely, along wi h esul an colony ou -
comes. Using a combina ion o de ailed beha iou al obse a ion,
au oma ed acking o o aging ac i i y, and ex ensi e colony
assessmen s, we hen examined fine-scale in e ac ions be ween
he wo species, and quan ified bumblebee colony su i al and
ep oduc i e success ac oss a g adien o V. elu ina densi ies.
Resul s demons a ed ha o aging V. elu ina equen ly
a emp ed o p eda e upon B. e es is wo ke s a colony
en ances, ye in he p esen s udy, all such obse ed a emp s
ended in ailu e. Despi e his, bumblebee colony weigh s showed
a nega i e co ela ion wi h ho ne densi ies a si es, sugges ing
po en ial indi ec e ec s upon colony g ow h. No ably, such
e ec s did no appea o influence ei he colony su i al o queen
p oduc ion, indica ing a deg ee o esilience wi hin he imescale
assessed. Taken oge he , his wo k comp ises he fi s comp e-
hensi e colony-le el analysis o how V. elu ina a ec s a na i e
bumblebee, and p o ides baseline da a wi h which o be e
unde s and he cu en and u u e en i onmen al impac s o he
ho ne .
Resul s
Fac o s influencing ho ne coun s a si es. To assess he impac
o V. elu ina on B. e es is colonies in a ep esen a i e en i -
onmen , we es ablished 12 field si es om an ini ial g oup o 15
loca ions ac oss he p o ince o Pon e ed a, Spain (Fig. S1a). As
we wished o selec si es encompassing a ange o V. elu ina
densi ies and land co e ypes (Figs. S1b, c, S2, and Table S1), we
fi s cha ac e ised ela i e ho ne abundance a a local scale. This
was achie ed using wo modified VespaCa ch (Vé o-pha ma)
aps pe si e, p o iding con inuous coun s o cap u ed ho ne s
e e y wo days o he du a ion o he s udy. We hen u ilised,
hese da a o p o ide a empo ally ele an measu e o ho ne
abundance a each si e, and in es iga e he influence o local
clima ic and en i onmen al a iables on ho ne densi y.
Ho ne coun s showed subs an ial spa io empo al a iabili y
ac oss si es (Fig. 2a), being significan ly g ea e wi h inc eased
minimum empe a u e (GLMM, e ec o minimum empe a u e:
F
1,187
=90.478, P< 0.001; Fig. 2b), dec eased minimum ela i e
humidi y (GLMM, e ec o minimum ela i e humidi y:
F
1,225
=14.170, P< 0.001; Fig. 2c), and a lowe pe cen age o
wa e co e (GLMM, e ec o wa e bodies: F
1,13
=8.988,
P=0.011). Addi ionally, coun s o apped ho ne s peaked in
ARTICLE COMMUNICATIONS BIOLOGY | h ps://doi.o g/10.1038/s42003-023-05329-5
2COMMUNICATIONS BIOLOGY | (2023) 6:990 | h ps://doi.o g/10.1038/s42003-023-05329-5 | www.na u e.com/commsbio
ea ly-Augus , and exhibi ed a dec easing end he ea e
(GLMM, e ec o da e: F
1,19
=495.081, P< 0.001; Fig. 2b). The
pe cen age o discon inuous u ban ab ic a si es had no
disce nible impac (GLMM, e ec o discon inuous u ban ab ic:
F
1,10
=3.356, P=0.099), howe e , si e ID had a significan
andom e ec (si e andom e ec : Z=1.958, P=0.050). No ably,
all o he land co e a iables had li le p edic i e influence, and
hus we e excluded om analyses.
Fac o s Influencing Ho ne Ac i i y a Bumblebee Colonies.
Ac oss he cou se o he s udy, du ing se sampling pe iods,
ho ne ac i i y was eco ded using an ex e nal D agon Touch
Vision 1 came a (D agon Touch), ained on each colony a a
dis ance o 1 m (Fig. 1b). This enabled obse a ion o he colony
ex e io and p oxima e su oundings, including ho ne and bee
in e ac ions in he immedia e icini y (Fig. 1d, ). Semi-
au oma ed quan ifica ion o ho ne p esence and p eda o y
beha iou was hen conduc ed using BORIS ( elease . 8.5), and
combined wi h colony g ow h, su i al, and o aging da a o
in es iga e subsequen e ec s. Addi ionally, 10 bee and 10 ho ne
o age s we e caugh and weighed each day, o assess changes in
he weigh a io o he wo species o e ime, and de e mine
whe he his influenced in e ac ion dynamics.
Ho ne s we e egula ly obse ed ho e ing nea bumblebee
colony en ances, and a emp ing o p eda e upon depa ing o
e u ning o age s. Specifically, esul s demons a ed ha ho ne s
we e mo e likely o be p esen a colonies ha we e hea ie
(GLMM, e ec o colony weigh : F
1,200
=4.285, P=0.040) and
had a highe equency o o aging (GLMM, e ec o o al bee
o aging equency: F
1,200
=6.174, P=0.014; Fig. 3a), indepen-
den o colony ID (colony andom e ec : Z=0.568, P=0.570).
In con as howe e , ho ne p esence a colonies was in e sely
co ela ed wi h bo h empe a u e (GLMM, e ec o si e
empe a u e: F
1,200
=6.631, P=0.011) and ela i e humidi y
(GLMM, e ec o si e ela i e humidi y: F
1,200
=7.649, P=0.006)
a he ime o sampling.
No ably, despi e a o al o 125 eco ded p eda ion a emp s
du ing ideo sampling spanning a pe iod o 27 days, ho ne s had
a 100% ailu e a e when a emp ing o p ey upon bumblebees.
The equency o hese a emp s was no influenced by he weigh
a io o indi idual ho ne s o bumblebees (GLMM, e ec o
wo ke weigh a io: F
1,204
=0.003, P=0.959; Fig. 3c), o ho ne
coun s a si es (GLMM, e ec o ho ne coun : F
1,83
=1.237,
P=0.269). Ins ead, p eda ion a emp equency co ela ed wi h
he amoun o ime ha ho ne s spen o aging nea colonies
(GLMM, e ec o ho ne du a ion a colonies: F
1,204
=12.410,
P< 0.001; Fig. 3b), independen o colony ID (colony andom
Fig. 1 Au oma ed moni o ing se up. a Expe imen al colony se up. A ows indica e (1) angefinde ma kings; (2) ame calib a ion ma kings; (3)
obse a ion en ance; and (4) en ance came a. bEx e nal came a se up 1 m om colony. cTypical iew om he en ance came a demons a ing
au oma ed acking o B. e es is wo ke s. A ows indica e (1) digi al en ance and exi coun e s; (2) AI-assis ed acking locks (boxes) and ajec o ies
(lines) o indi idual bumblebees. dView om he ex e nal came a showing he colony ex e io and su oundings. eHea map de ailing he ime spen by
bumblebees a e sing he obse a ion en ance (sho e , blue; longe , ed), confi ming he op imum egion o acking. A ypical in e ac ion be ween he
wo species p eceding a p eda ion a emp . A ows deno e (1) a B. e es is o age exi ing he colony; (2) a V. elu ina wo ke in es iga ing he colony
en ance.
COMMUNICATIONS BIOLOGY | h ps://doi.o g/10.1038/s42003-023-05329-5 ARTICLE
COMMUNICATIONS BIOLOGY | (2023) 6:990 | h ps://doi.o g/10.1038/s42003-023-05329-5 | www.na u e.com/commsbio 3
e ec : Z=1.107, P=0.268), and was again in e sely ela ed o
empe a u e (GLMM, e ec o si e empe a u e: F
1,204
=15.121,
P< 0.001) and ela i e humidi y (GLMM, e ec o si e ela i e
humidi y: F
1,204
=19.724, P< 0.001) a he ime o sampling.
Fac o s influencing bumblebee colony o aging ac i i y. Fo a-
ging ac i i y was eco ded using a calib a ed D agon Touch
Vision 1 came a (D agon Touch) posi ioned di ec ly abo e he
en ance o each colony (Fig. 1a). Du ing sampling, his allowed
o he filming o all bumblebees passing h ough he anspa en
en ance ube agains a s anda dised high-con as backg ound
(Fig. 1c), along wi h confi ma ion o o aging h ough he p e-
sence o pollen. Au oma ed acking o bumblebees was hen
achie ed ia a modified e sion o Camly ics ( elease .2.2.5) AI-
assis ed acking so wa e, yielding indi idual ajec o ies and
imes amps o all wo ke ansi s, along wi h hei di ec ionali y
(Fig. 1c, e, S3, and Supplemen a y Video 1).
The equency o wo ke depa u es om colonies inc eased
wi h he pe cen age o he e ogeneous ag icul u e a si es (GLMM,
e ec o he e ogeneous ag icul u e: F
1,8
=6.182, P=0.036),
peaking in mid-Augus (Fig. 4a), and was posi i ely associa ed
wi h he p obabili y o subsequen ho ne p esence (GLMM,
e ec o ho ne p esence a colonies: F
1,130
=15.097, P< 0.001;
Fig. 4b). Addi ionally, depa u e equency dec eased wi h highe
ela i e humidi y a he ime o sampling (GLMM, e ec o si e
ela i e humidi y: F
1,60
=133.816, P< 0.001; Fig. 4c), while he
pe cen age o discon inuous u ban ab ic had no disce nible
influence (GLMM, e ec o discon inuous u ban ab ic:
F
1,8
=1.720, P=0.226), and colony ID had a significan andom
e ec (colony andom e ec : Z=2.023, P=0.043).
In a simila ashion, he equency o wo ke e u ns o
colonies dec eased wi h ela i e humidi y a he ime o sampling
(GLMM, e ec o si e ela i e humidi y: F
1,51
=104.780,
P< 0.001; Fig. 4c), peaked in mid-Augus (Fig. 4a), and was
again posi i ely associa ed wi h subsequen ho ne p esence
(GLMM, e ec o ho ne p esence a colonies: F
1,140
=16.916,
P< 0.001; Fig. 4b). This sugges s ha ho ne s we e a ac ed o
colonies wi h highe o aging a fic (Fig. S4), bu ha hei
p esence did no lead o subs an ial o aging inhibi ion, as is he
case in honey bees42. Colony ID again had a significan andom
e ec (colony andom e ec : Z=2.190, P=0.029).
Fac o s influencing bumblebee colony g ow h and su i al.To
assess he influence o ac o s upon colony g ow h and su i al,
we eco ded colony weigh change, su i al s a us, and isible
symp oms o disease ac oss he sampling pe iod. Then, upon
s udy conclusion, we des uc i ely sampled colonies o quan-
i y he o al adul popula ion, mean weigh o adul bumblebees,
numbe o open and closed pupal cells, and he p esence o
absence o pa asi es, new queens, and males (Fig. S5). These
measu es we e used o e alua e he si e-le el e ec s o ho ne
densi y, land co e , and clima ic condi ions upon colony g ow h
and su i al. Fu he , om a subse o 10 colonies, he p e alence
o common pa hogens was assessed, allowing us o cha ac e ise
en i onmen al pa hogen exposu e in isibly symp oma ic and
asymp oma ic colonies (Supplemen a y Me hods, Tables S2–S5).
11 Sep embe 29 Augus 17 Augus 5 Augus 22 July
nuoC en oH
150
100
50
0
Minimum Rela i e
Humidi y (%)
100
80
60
40
20
0
Da e
17
Sep embe
5
Sep embe
23
Augus
11
Augus
Da e
17
Sep embe
5
Sep embe
23
Augus
11
Augus
nuoC en oH
60
50
40
30
20
10
0
Minimum Tempe a u e
(°C)
20
15
10
5
0
a
bc
15
14
13
12
11
10
9
8
7
6
5
4
3
2
1
Humidi y
Tempe a u e
Vespa elu ina
Fig. 2 Ho ne coun s a si es. a Ho ne coun s a si es o e ime, de e mined by he numbe o indi iduals caugh in wo VespaCa ch aps e e y wo days
(N=5302). Line colou s indica e si e IDs, he dashed e ical line deno es he iming o colony es ablishmen , and poin s p io o his indica e hose used
o si e cha ac e isa ion. bMean ho ne coun s and minimum daily empe a u e o all si es ac oss he du a ion o he s udy (N=3798). Each poin
ep esen s a single measu e o ho ne coun o minimum daily empe a u e a a si e om he co esponding da e. Lines indica e mo ing a e ages, and
colou s deno e he measu ed a iable (ho ne coun , g ey; minimum daily empe a u e, blue). cMean ho ne coun s and minimum daily ela i e humidi y
o all si es ac oss he du a ion o he s udy (N=3798). Each poin ep esen s a single measu e o ho ne coun o minimum daily ela i e humidi y a a
si e om he co esponding da e. Lines indica e mo ing a e ages, and colou s deno e he measu ed a iable (ho ne coun , g ey; minimum daily ela i e
humidi y, g een). E o ba s ep esen 95% confidence in e als.
ARTICLE COMMUNICATIONS BIOLOGY | h ps://doi.o g/10.1038/s42003-023-05329-5
4COMMUNICATIONS BIOLOGY | (2023) 6:990 | h ps://doi.o g/10.1038/s42003-023-05329-5 | www.na u e.com/commsbio
Ho ne coun s a si es had no significan influence on
bumblebee colony su i al (GLMM, e ec o ho ne coun :
F
1,676
=0.474, P=0.491; Fig. 4), howe e a highe pe cen age o
he e ogeneous ag icul u e (GLMM, e ec o he e ogeneous
ag icul u e: F
1,10
=6.206, P=0.032), highe minimum ela i e
humidi y (GLMM, e ec o minimum ela i e humidi y:
F
1,676
=14.602, P< 0.001), and highe minimum empe a u e
(GLMM, e ec o minimum empe a u e: F
1,676
=10.225,
P=0.001), inc eased he likelihood ha colonies would su i e.
Colony ID also had a significan andom e ec when
conside ing su i al ou comes (colony andom e ec :
Z=2.696, P=0.007).
In con as , colony weigh o e ime was nega i ely co ela ed
wi h ho ne coun s (GLMM, e ec o ho ne coun :
F
1,653
=83.175, P< 0.001; Fig. 5), and highe maximum em-
pe a u es a si es (GLMM, e ec o maximum empe a u e:
F
1,640
=9.548, P=0.002), while being posi i ely influenced by
highe minimum empe a u es (GLMM, e ec o minimum
empe a u e: F
1,634
=9.436, P=0.002). When conside ing hese
in e ac ions, colony ID had a significan andom e ec (colony
andom e ec : Z=3.952, P< 0.001), indica ing a deg ee o
unexplained in e -colony a iabili y, as would be expec ed.
Ho ne coun s had no significan e ec on he p oduc ion o
new queens by colonies (GLMM, e ec o ho ne coun :
F
1,716
=0.051, P=0.821), while a highe pe cen age o he e o-
geneous ag icul u e (GLMM, e ec o he e ogeneous ag icul u e:
F
1,10
=11.250, P=0.007) and wa e bodies a si es (GLMM,
e ec o wa e bodies: F
1,9
=6.121, P=0.035) inc eased he
likelihood o new queens being p esen , independen o si e ID
(si e andom e ec : Z=1.677, P=0.094).
Colony popula ion size was posi i ely co ela ed wi h colony
weigh ac oss he s udy (GLMM, e ec o colony weigh :
F
1,698
=4.154, P=0.042). This ela ionship was independen o
ho ne coun s (GLMM, e ec o ho ne coun : F
1,698
=0.428,
P=0.513), and he pe cen age o he e ogeneous ag icul u e a
si es (GLMM, e ec o he e ogeneous ag icul u e: F
1,10
=2.054,
P=0.182), wi h si e ID ha ing a significan andom e ec (si e
andom e ec : Z=2.159, P=0.031). Simila ly, he a e age
weigh o indi idual adul bumblebees in colonies was posi i ely
co ela ed wi h colony weigh o e ime (GLMM, e ec o colony
weigh : F
1,698
=7.663, P=0.006), ye una ec ed by ho ne coun s
(GLMM, e ec o ho ne coun : F
1,689
=0.995, P=0.319), and
he pe cen age o e e g een b oadlea ed and coni e ous o es s a
si es (GLMM, e ec o e e g een b oadlea ed and coni e ous
o es s: F
1,10
=3.362, P=0.097). When conside ing a e age adul
bee weigh , si e ID again had a significan andom e ec (si e
andom e ec : Z=2.167, P=0.030).
Discussion
Ou esul s demons a e ha V. elu ina wo ke s equen ly
a emp o p eda e upon B. e es is o age s a colony en ances,
ye in he p esen s udy, we obse ed such a emp s o be en i ely
unsuccess ul. Ho ne s we e p e e en ially a ac ed o colonies
wi h a highe equency o o aging a fic, and he du a ion o
his a ac ion was significan ly co ela ed wi h he p obabili y o
p eda ion, sugges ing ha con inuous p ey flux is a equisi e
15 Sep embe 6 Sep embe 27 Augus 18 Augus 11 Augus
ycneu
q
e
Fgni
ga oFeeBla oT
200
150
100
50
0
Da e
12
Sep embe
31
Augus
21
Augus
11
Augus
hgieW ek o
W(
g)
0.8
0.6
0.4
0.2
0.0
P eda ion A emp s
>543210
anoi a uD en oH
sein
o
l
o
CeeB(s)
1500
1000
500
0
Vespa elu ina
Bombus e es is
bc
Ho ne Absen
Ho ne P esen
a
(h-1)
Fig. 3 Ho ne beha iou a bumblebee colonies. a To al bee o aging equencies o colonies o e ime (N=36). Each poin ep esen s a single sample o
bee o aging equency pe hou o a colony om he co esponding da e. Poin s a e colou ed based on whe he ho ne s we e p esen a he colonies
(absen , blue; p esen , ed), and he solid line indica es a LOESS (locally weigh ed sca e plo smoo hing) mo ing a e age o all colonies o e ime. bTime
spen by ho ne s o aging a colonies (N=28), g ouped by he numbe o esul an p eda ion a emp s obse ed. Ou lie s (g ea e han 1.5 imes he
in e qua ile ange om he median) a e indica ed wi h ci cles. cWeigh s o indi idual bee and ho ne wo ke s sampled ac oss he cou se o he s udy (N
b.
e es is
=350, N
V. elu ina
=350). Boxplo s a e colou ed by species (B. e es is, blue; V. elu ina, g ey) and g ouped by sampling da e. Ou lie s (g ea e han
1.5 imes he in e qua ile ange om he median) a e again indica ed wi h ci cles.
COMMUNICATIONS BIOLOGY | h ps://doi.o g/10.1038/s42003-023-05329-5 ARTICLE
COMMUNICATIONS BIOLOGY | (2023) 6:990 | h ps://doi.o g/10.1038/s42003-023-05329-5 | www.na u e.com/commsbio 5
15 Sep embe 6 Sep embe 27 Augus 18 Augus 11 Augus
y
cneu
qe Fgniga oFe
eB
200
150
100
50
0
Rela i e Humidi y (%)
100806040200
200
150
100
50
0
Ho ne P esence
Absen P esen
gn
iga o
F eeB
y
cneu
qe
F
250
200
150
100
50
0
a
c
Bee Depa u es
Bee Re u ns
b*
Bee Depa u es
Bee Re u ns
Bee Depa u es
Bee Re u ns
(h-1)
(h-1)
Fig. 4 Bumblebee colony o aging ac i i y. a F equency o bee depa u es and e u ns pe hou o all colonies o e ime (N=36). Each poin ep esen s a
single measu e o depa u e o e u n equency pe hou o a colony om he co esponding da e. Lines indica e LOESS (locally weigh ed sca e plo
smoo hing) mo ing a e ages, and colou s deno e wo ke mo emen di ec ion ( e u ns, ligh blue; depa u es, da k blue). bF equency di e ences in bee
depa u es and e u ns pe hou when ho ne s we e p esen o absen (N
P esen
=28, N
Absen
=36). Boxplo s a e colou ed by wo ke mo emen di ec ion
( e u ns, ligh blue; depa u es, da k blue) and g ouped by ho ne p esence o absence a colonies. As e isk indica es significan di e ences be ween
g oups (en ances, F
1,130
=15.097, P< 0.001; exi s, F
1,140
=16.916, P< 0.001). Ou lie s (g ea e han 1.5 imes he in e qua ile ange om he median) a e
deno ed wi h ci cles. cRela ionship be ween si e ela i e humidi y a he ime o sampling and he equency o bee e u ns and depa u es pe hou ac oss
all colonies (N=36). Each poin ep esen s a single measu e o bee depa u e o e u n equency pe hou and he co esponding ela i e humidi y a he
ime o sampling. Poin colou deno es wo ke mo emen di ec ion ( e u ns, ligh blue; depa u es, da k blue), he solid line is a linea endline fi ed
agains he da a, and he dashed lines ep esen he 95% confidence in e al.
Ho ne Coun
100806040200
hgieWynoloC(kg)
2.4
2.2
2.0
1.8
1.6
1.4
1.2
Su i ed
Died
50
100
150
200
Fig. 5 Bumblebee colony g ow h and su i al. Rela ionship be ween colony weigh , o al bee o aging equency, and ho ne coun s a si es (N=36). Each
poin ep esen s a single measu e o colony weigh and he co esponding ho ne coun om wo VespaCa ch aps a he ime o sampling (N=3798).
Poin colou deno es he final su i al ou come o he colony (blue, su i ed; g ey died), while poin size indica es he o al bee o aging equency a he
ime o colony weigh sampling (la ge , highe equency). The solid line is a linea endline fi ed agains he da a, and he dashed lines ep esen he 95%
confidence in e al.
ARTICLE COMMUNICATIONS BIOLOGY | h ps://doi.o g/10.1038/s42003-023-05329-5
6COMMUNICATIONS BIOLOGY | (2023) 6:990 | h ps://doi.o g/10.1038/s42003-023-05329-5 | www.na u e.com/commsbio
s imulus o he main enance o p eda o y beha iou . No ably,
bumblebee colony weigh s we e nega i ely associa ed wi h ho ne
coun s a si es, indica ing possible colla e al e ec s upon colony
g ow h, os ensibly h ough o aging dis up ion ia epea ed
p eda ion a emp s a o away om colonies, o compe i ion o
local esou ces. Howe e , while colony g ow h was educed, his
did no appea o impac ei he su i al o queen p oduc ion,
sugges ing a deg ee o esilience. In conce , ou esul s p o ide
he fi s colony-le el da a quan i ying he impac s o V. elu ina
on B. e es is, and u he elucida e he p eda o y beha iou o
hese in asi e ho ne s, hus in o ming u u e mi iga ion e o s
o wild and managed pollina o s.
The epea ed ailu e o ho ne s o success ully cap u e and
subdue bumblebees is no able. P eda ion a emp s gene ally ol-
lowed a p edic able sequence: a ho ne would fi s pu sue and
g ab a bee in fligh (Fig. S6a–c); he bee would subsequen ly d op
o he g ound in esponse, causing he ho ne o all wi h i
(Fig. S6d– ); and hen, upon impac wi h he g ound, he ho ne
would lose i s g ip, hus allowing he bee o escape (Fig. S6g–i and
Supplemen a y Video 2). In some cases, a ho ne was able o
main ain i s hold upon a bee a e hi ing he g ound, howe e he
bee would hen assume a ypical de ensi e pos u e, alling on o i s
back wi h legs and s ing aised38,43. Wi hou excep ion, his
beha iou e en ually o ced he ho ne o abandon he p eda ion
a emp , and e u n o o aging a he colony en ance. In e es -
ingly, i is unlikely ha such ailu e can be a ibu ed solely o he
compa a i e size o B. e es is adul s, as we obse ed p eda ion
a emp s di ec ed owa ds indi iduals spanning he ange o
wo ke polymo phism p esen in his species (Fig. 3c), all o
which we e unsuccess ul.
Such in e ac ions appea in s a k con as o p eda ion a -
ge ing A. melli e a, wi h p e ious wo k finding ha in ~30% o
o aging isi s o colonies, V. elu ina we e able o success ully
cap u e, subdue, and kill wo ke s44. Indeed, confi med cases o
he ho ne s p eda ing upon Bombus species in hei in aded
ange a e a e o da e, wi h limi ed isual epo s45,46, and a
comp ehensi e die a y analysis finding only a single eco d o B.
lapida ius among a sample o 2515 p ey i ems34. Fu he , a ecen
s udy o ho ne s hun ing a Hede a hibe nica flowe s obse ed
agonis ic in e ac ions, bu ne e a success ul p eda ion a emp
owa ds wild B. e es is queens, indica ing ha ailed p eda ion
a emp s a e p esen in na u e47. While he pauci y o de ec ions
in o he die a y analyses may be due o he limi ed pe iod o
sampling, and p ime biases ha supp ess he amplifica ion o
bumblebee DNA35, addi ional e idence o success ul p eda ion
upon Bombus spp. is sca ce, e en wi hin he na i e ange o V.
elu ina48,49. Taken wi h ou p esen findings, his sugges s ha
B. e es is is no a majo p ey species o V. elu ina; howe e as
no ed in ea lie wo k36,48,49, his ails o de e he ho ne s om
making epea ed p eda ion a emp s.
While he impac s o p eda ion pe se appea minimal, ou
esul s do sugges ha he p esence o ho ne s—especially a e y
high densi ies—may impa a nega i e impac on bumblebee
colony g ow h. The exac mechanisms unde lying his emain
unclea , hus i is impossible o comple ely ule ou addi ional
con ounding ac o s. No ably, al hough ou s udy accoun ed o
he influence o land co e a a b oad scale, i is easible ha he
a ailabili y o specific esou ces wi hin si es—such as flo al nec a
o o he insec p ey—ga e ise o he obse ed ends. The local
densi y o managed honey bee colonies may also ha e been a
ac o , as apia ies can supp ess bumblebee colony g ow h;50
howe e we we e unable o collec his da a wi hin he confines o
ou s udy. While such e ec s a e plausible, he e a e also se e al
di ec mechanisms by which ho ne densi y may ha e impac ed
colony weigh . Specifically, p e ious wo k has shown ha bum-
blebees o aging a flowe s al e hei beha iou in esponse o
ho ne a acks36, concomi an ly educing he e ficacy o esou ce
collec ion. Fu he mo e, inc eased exploi a ion o flo al nec a in
a eas wi h high ho ne densi ies, combined wi h ha assmen o
o aging bumblebees, would se e o educe esou ce a ailabili y,
while imposing ene ge ic cos s associa ed wi h epelling a acks.
Indeed, da a om honey bees indica es ha he p esence o
ho ne s can impac colony heal h h ough educed queen
ecundi y33,51, and he inc eased exp ession o oxida i e s ess-
ela ed genes, e en in bees ha a e no di ec ly a ge ed52. Despi e
his, i is impo an o no e ha he obse ed educ ions in
bumblebee colony weigh did no esul in educed colony su -
i al o queen p oduc ion o e he cou se o he s udy. While his
is p omising, u he wo k would be needed o ule ou
popula ion-le el isks o e longe ime scales, because such
g ow h e ec s ha e he po en ial o accumula e ac oss mul iple
gene a ions.
Local land co e and clima e appea ed o be o a iable
impo ance be ween species. While ho ne coun s a si es we e
la gely independen o land co e , bumblebee colony su i al,
queen p oduc ion, and o aging ac i i y we e all posi i ely co -
ela ed wi h he pe cen age o he e ogeneous ag icul u e in he
su ounding a ea. This is pe haps unsu p ising, as V. elu ina is a
highly adap able gene alis o age 34, while B. e es is is neces-
sa ily sensi i e o he local a ailabili y o sui able pollen and
nec a sou ces53. The nega i e co ela ion be ween ho ne coun s
and wa e bodies a si es is less clea , and may be indica i e o
co a iance wi h addi ional unmeasu ed pa ame e s, howe e , as
all wa e bodies we e ≥0.1 km2in a ea, i may also sugges
po en ial exclusion e ec s. Bo h species esponded posi i ely o
highe daily minimum empe a u es, while ha ing a nega i e
esponse o highe daily maximum empe a u es, sugges ing ha
hese we e limi ing ac o s. Addi ionally, ela i e humidi y a he
ime o sampling nega i ely influenced bo h ho ne ac i i y and
bumblebee o aging equency, as would be expec ed due o i s
igh associa ion wi h p ecipi a ion.
Se e al colonies showed signs o disease du ing he s udy,
namely dysen e y, and esul s om he pa hogen analyses we e
able o p o ide some u he insigh in o he po en ial causa i e
agen . Specifically, e idence o dysen e y was co ela ed wi h he
p e alence o he ungal pa hogen Vai imo pha bombi— o me ly
Nosema bombi54—which is consis en wi h he obse ed symp-
omology (Fig. S7). I should be s essed ha he limi ed sample
size used he e p ecluded any causa i e analyses, howe e as
colonies we e sc eened p io o es ablishmen (Table S2), i is
easonable o deduce ha V. bombi in ec ions we e acqui ed
na u ally, especially gi en i s high p e alence in he wild55.
The ela ionship obse ed be ween bee o aging equency,
ho ne isi s o colonies, and he equency o p eda ion a emp s,
p o ides some insigh in o he po en ial s imuli equi ed o
p omo e p eda o y beha iou in V. elu ina. Specifically, ho ne s
appea o equi e cons an o aging a fic o elici con inued
in e es in a colony, and disengage once his equency alls below
a ce ain h eshold (Fig. S4). The o me poin is salien , as i
sugges s ha o age ac i i y— a he han any o he unique
a ibu e o he expe imen al colonies—is a key d i e o ho ne
a ac ion. This end is consis en wi h he ho ne s’p e e ence
o o he dense p ey agg ega ions, including pollina o s a flowe
pa ches;36,48 flies a ound li es ock and e use;34,45 and colonies o
honey bees42,44,56. An impo an dis inc ion when conside ing B.
e es is, howe e , is ha colony o aging ac i i y appea s o be a
he edge o he h eshold equi ed o main ain he ho ne s’
in e es , as e idenced by he equen cessa ion o p eda o y
beha iou a lowe o aging equencies (Fig. S4). Indeed, o a-
ging ac i i y showed li le disce nible esponse o ho ne p esence,
suppo ing a causal influence o o age flux upon ho ne
a ac ion, a he han any de ensi e eac ion by he colony.
COMMUNICATIONS BIOLOGY | h ps://doi.o g/10.1038/s42003-023-05329-5 ARTICLE
COMMUNICATIONS BIOLOGY | (2023) 6:990 | h ps://doi.o g/10.1038/s42003-023-05329-5 | www.na u e.com/commsbio 7
Due o he need o weigh s anda disa ion, de ailed colony
assessmen s, and acking o su i al ou comes, we we e con-
s ained o he use o comme cially- ea ed colonies in he p esen
s udy. Howe e , a no able ea u e o hese colonies is hei
inc eased isibili y when compa ed o na u al B. e es is nes s. I
is hus possible ha wild colonies o his species a e less a ac i e
o o aging ho ne s, because hey a e o en loca ed
unde g ound38, and may no always each he o age densi ies
obse ed in he p esen s udy. Despi e his, as he flow o o aging
a fic alone appea ed o be he key de e mining ac o in he
ho ne s’a ac ion o colonies— ega dless o he inc eased
en ance isibili y o ou expe imen al se up—i can be eason-
ably su mised ha na u al colonies wi h su ficien o aging
ac i i y would be simila ly a ge ed. Indeed, da a om wild B.
e es is colonies indica es compa able wo ke popula ions o
hose eco ded in he p esen s udy57, wi h subsequen o aging
ac i i y being used by humans and p eda o s o success ully loca e
nes s58. Addi ionally, while ou use o anspa en en ance ubes
inc eased he isibili y o a e sing bumblebees, i should be
no ed ha ho ne p eda ion a emp s gene ally occu ed away
om hese, implying ha hei p esence did no subs an ially al e
ho ne beha iou . As such, i can be pos ula ed ha ou esul s
a e easonably ans e able o wild colonies, al hough u he
esea ch would be needed o de e mine hei impo ance in he
con ex o na u al popula ions; and o di e en bumblebee
species.
The appa en p e e ence o ho ne s o concen a ions o p ey
may be pe inen o he p o ec ion o pollina ing insec s in
in aded a eas, as i p o ides impe us o limi ing he clus e
densi y o conse a ion measu es ha a ificially agg ega e soli-
a y species59. In e es ingly, he con e se is indica ed o managed
honey bee colonies, because high densi ies a e hough o ‘dilu e’
he pe capi a p eda ion bu den60. The la e should be unde -
s ood in he con ex o ‘ o aging pa alysis’howe e , as he p in-
ciple h ea o honey bee colonies is hei eac ion o he in ensi y
o p eda ion, a he han di ec mo ali y om p eda ion
i sel 42,51. Consequen ly, o species ha do no unde ake col-
lec i e o e win e p o isioning, including bumblebees, such
e ec s a e o limi ed applicabili y. This sugges s ha managed
honey bee colonies should be conside ed as a special case, and o
o he less nume ous agg ega ions o insec s, s a egies ha di use
local densi y a e likely o be e ec i e in educing p eda ion
in ensi y.
In conclusion, we p o ide he fi s empi ical da a comp e-
hensi ely quan i ying he impac s o V. elu ina upon B. e es is
colonies. Despi e he finding ha bumblebees we e equen ly
a ge ed as p ey, all such p eda ion a emp s ha we obse ed
we e wholly ine ec ual, highligh ing he impo ance o de ailed
beha iou al moni o ing o asce ain ou comes. Fu he , he
nega i e co ela ion be ween bumblebee colony weigh s and
ho ne coun s a si es indica es ha sub-le hal e ec s should be
conside ed when e alua ing colony-le el isks, as his end pe -
sis s despi e an appa en lack o p eda ion-media ed mo ali y.
Taken oge he , ou findings sugges ha V. elu ina in e ac s
qui e dis inc ly wi h di e en bee gene a, e en wi hin he eusocial
Apidae alone. As such, he e is s ong p eceden o conside
h ea ened species in a bespoke ashion, i we a e o e ec i ely
unde s and and mi iga e in asion impac s a he ecosys em-le el.
Me hods
Colonies. Colony p epa a ion occu ed p io o he ini ia ion o
he s udy du ing Augus 2021, nea Tomiño in Pon e ed a, Spain
a 41.97903, −8.76852 (DD). To ensu e s anda disa ion, and
acili a e con olled assessmen s, 36 coe aneous Bombus e es is
e es is colonies we e ob ained om a local supplie in Alme ia,
Spain (Biobes G oup), each consis ing o a queen and ~80
wo ke s, and housed wi hin wa e p oo polys y ene ou e coa -
ings (Biobes G oup). Along wi h hese, an addi ional h ee
colonies we e pu chased o equipmen es ing and op imisa ion.
All colonies we e sc eened by he supplie o an ex ensi e panel
o pa hogens and pa asi es using RT-qPCR, and he absence o
in ec ion was confi med (Table S2). Upon deli e y, colonies we e
weighed, assigned a andom ID numbe , and p o ided wi h
supplemen al suc ose solu ion o assu e adequa e p o isioning.
Each was hen alloca ed a si e ID numbe , o alling h ee colonies
pe si e, ensu ing an e en s a ing weigh dis ibu ion ac oss si es
(mean weigh <±50 g be ween si es, Fig. S8a). Following ans-
po a ion o si es, supplemen al suc ose was emo ed o s imula e
o aging, and colony en ances we e opened. Colonies we e hen
allowed o acclima ise and amilia ise hemsel es wi h hei su -
oundings o 24–48 h p io o sampling.
Expe imen al design. P eceding colony se up, 15 p ospec i e
field si es we e iden ified h oughou he p o ince o Pon e ed a,
Spain. These si es we e selec ed o include a ange o local Vespa
elu ina nig i ho ax densi ies and land co e ypes, allowing us o
e alua e colony ou comes in a ep esen a i e en i onmen . To
cha ac e ise ela i e ho ne abundance and ac i i y, we deployed
wo VespaCa ch aps wi h VespaCa ch a ac an (Vé o-pha ma)
a each si e61, a a heigh o 1.5 m om he g ound, and ou o
di ec sunligh . Foam inse s we e placed in o all aps o abso b
excess liquid a ac an and p e en cap u ed insec s om
d owning, hus limi ing mo ali y and unwan ed e ec s upon
bo h na i e species and ho ne s62. Cap u ed ho ne s we e coun-
ed and eleased e e y second day o he wo weeks p eceding
s udy ini ia ion in la e July and ea ly Augus 2021. The esul an
measu e consis ed o he o al numbe o ho ne s caugh in bo h
aps a a si e o e wo days—hence o h e med ‘ho ne coun ’.
This p o ided a p ofile o ho ne ac i i y o e ime, allowing us o
selec a final g oup o 12 field si es ha bes -encompassed he
ange o local ho ne densi ies. Ac oss he cou se o he s udy, we
con inued moni o ing ho ne coun s a si es o ensu e ha ou
da a emained empo ally ele an , as ac i i y can fluc ua e sub-
s an ially o e ime.
Each si e was p o isioned wi h a wa e sou ce and suc ose
eede , allowing he colonies o o age ad libi um, while ensu ing
ha wa e and ca bohyd a e a ailabili y we e no limi ing ac o s
—al hough colonies we e a ely obse ed u ilising hese. Colonies
we e si ua ed in he shade amongs ege a ion o a oid
o e hea ing, and in cases whe e su ficien na u al shade was no
p esen , we e p o ided wi h shade in he o m o pi ched oo es
o whi e Co ex shee ing (Co plex UK). Expe imen al sampling
was ini ia ed on he 9 h o Augus 2021, and con inued o a
pe iod o 40 days, concluding on he 18 h o Sep embe .
Sampling. Colony sampling occu ed e e y wo days o he
du a ion o he s udy, p oducing 20 ime poin s in o al. Speci-
fically, colonies we e weighed using a PS 2100.R2.M.H p ecision
(linea i y ±0.02 g) balance (RADWAG), checked o isible
symp oms o disease, and hei su i al s a us was no ed. To
quan i y indi idual beha iou , wo D agon Touch Vision 1
came as (D agon Touch) we e ained on each colony. The fi s
o hese was an ‘en ance came a’si ua ed 110 mm abo e a cus-
omised anspa en ac ylic en ance ube, yielding a ame a ea
o ~0.011 m2, and hus eco ding all bumblebees en e ing and
exi ing he colony du ing sampling (Fig. 1a). The second was an
‘ex e nal came a’posi ioned 1m om he colony, wi h i s field o
iew encompassing he ex e io and p oxima e su oundings
wi hin ~1.5 m, enabling obse a ion o all ho ne s in he
immedia e icini y, and hus allowing p eda ion beha iou o be
ARTICLE COMMUNICATIONS BIOLOGY | h ps://doi.o g/10.1038/s42003-023-05329-5
8COMMUNICATIONS BIOLOGY | (2023) 6:990 | h ps://doi.o g/10.1038/s42003-023-05329-5 | www.na u e.com/commsbio
eco ded (Fig. 1b). Video eco dings occu ed be ween 09:00 and
17:00 h o bes align wi h colony o aging ac i i y63, and a e aged
3–4 h in leng h, dependen on ba e y li e (Fig. S8b). Spen
came as we e collec ed and eplaced a each ime poin , allowing
da a e ie al and ini ia ion o eco ding o occu simul aneously.
This o a ion employed a o al o 72 came as, enabling comple e
co e age o all 36 colonies.
In addi ion o colony assessmen s, con inuous si e-le el
empe a u e (°C) and ela i e humidi y (%) measu emen s we e
made using Tiny ag Plus 2 - TGP-4500 da a logge s (Gemini Da a
Logge s L d). These eco ded alues a 15 min in e als, enabling
he calcula ion o daily maxima and minima, along wi h ambien
condi ions a he ime o ideo eco dings. To cha ac e ise land
co e , sa elli e da a was ex ac ed om a 1.5 km adius a ound
each si e, u ilising he 2018 CORINE64 land co e da ase in
A cGIS P o ( elease . 3.0) (Fig. S2). This p o ided land co e
da a encompassing he maximum o aging ange o colonies65 o
a esolu ion o 100m2, which was hen alida ed agains ecen
sa elli e image y o p oduce modified class defini ions ha mo e
accu a ely ep esen ed he local ege a ion (Fig. S3 and Table S1).
Fu he , om a single si e pe day, 10 bee and 10 ho ne
o age s we e caugh and weighed using HSW07 p ecision
(linea i y ±0.001 g) scales (Hoosiwee), hus compa ing he weigh
a io be ween species o e ime, o de e mine whe he his
influenced p eda ion dynamics. Ra ios we e calcula ed using he
ollowing o mula:
Wo ke weigh a io ¼mean ho ne weigh ðgÞ=mean bee weigh ðgÞ
ð1Þ
Upon conclusion o he s udy, colonies we e des uc i ely
sampled wi h e hyl ace a e o quan i y he o al adul popula ion,
mean weigh o adul bumblebees (g), numbe o open and closed
pupal cells, colony weigh (g) and dimensions (mm), and he
p esence o absence o gynes, males, nec a , and pollen. Males
we e iden ified ia ex usion o he endophallus, while gynes we e
di e en ia ed by mapping hei posi ion ela i e o known queens
in he o e all weigh (g) equency dis ibu ion (Fig. S5).
Addi ionally, he p esence o absence o bumblebee wax mo h
la ae (Aphomia sociella) was eco ded, along wi h hose o he
black soldie fly(He me ia illucens).
To in es iga e en i onmen al pa hogen exposu e, a subse o
10 colonies was sc eened o common ungal, ypanosomal,
and neog ega ine pa hogens. This subse consis ed o fi e
colonies showing isible symp oms o disease—specifically
dysen e y—and fi e showing no isible symp oms. Subsequen
analyses we e desc ip i e a he han in e en ial, and aimed o
cha ac e ise he pa hogen p ofiles o symp oma ic and asymp-
oma ic colonies. A panel o se en a ge s was chosen, consis ing
o Vai imo pha ce anae,Vai imo pha apis,Vai imo pha bombi,
C i hidia bombi,C i hidia mellificae,Lo ma ia passim,and
Apicys is bombi (Table S3). Analyses we e conduc ed using a
combina ion o mul iplex and monoplex PCR, and DNA
ex ac ion p o ocols ollowed p e iously es ablished me hods60.
Fo ull de ails o he PCR and DNA ex ac ion me hodologies,
p ime sequences, and he mal p o ocols, see (Supplemen a y
Me hods, Tables S4, S5).
Video analyses. Video analyses u ilised wo complemen a y
p ocessing echniques. Videos om he en ance came as we e
analysed ia au oma ed acking so wa e, while hose om he
ex e nal came as we e p ocessed semi-au oma ically using e en -
logging so wa e (Fig. 1c, d). Analyses we e dispe sed e enly
ac oss he 40 day sampling pe iod, wi h he equency o sam-
pling being cons ained by came a o a ion, as e e y hi d day all
72 came as we e deployed simul aneously. This yielded 13 ime
poin s o da a comp ising 135,983 indi idual ajec o ies om he
en ance came as, and 6 ime poin s o da a comp ising 648 h o
beha iou al da a om he ex e nal came as. P io o p ocessing,
all ideos we e no malised o a s anda d leng h o 3 h, and
imes amps combined wi h an audio cue we e used o synch o-
nise imings be ween came as.
Au oma ed acking. To assess whe he ho ne beha iou had an
impac on o aging ac i i y, we used a modified e sion o
Camly ics ( elease .2.2.5) AI-assis ed acking so wa e o ack
indi idual bumblebees en e ing and lea ing colonies h ough he
obse a ion en ances (Fig. 1c, S3, and Supplemen a y Video 1).
The cus om design o hese en ances ensu ed ha bumblebees
emained on a fixed plane o mo emen , while a whi e plas ic
inse below he en ance simplified he backg ound o op imise
de ec ion (Fig. 1e). Fu he , in compa ison o s anda d com-
me cial nes boxes, he ex ended en ance leng h se ed o be e
app oxima e he unnel en ances p esen in na u e57. Colonies
we e ma ked wi h calib a ion and ange-finding indica o s abo e
he en ances, ensu ing ha all came as we e moun ed a a
consis en posi ion and dis ance, hus allowing se ial p ocessing
o ideos (Fig. 1a). When a bee a e sed an en ance, i was
indi idually iden ified as an objec and i s ajec o y was acked,
enabling mo emen di ec ion o be de e mined. This was hen
de ec ed by a digi al ‘en ance coun e ’ egion, allowing he exac
iming and numbe o en ances and exi s o be eco ded (Fig. 1c,
S3, and Supplemen a y Video 1). F om he esul an da a, h ee
key me ics we e ex ac ed, equency o bee depa u es, defined
as he numbe o bee exi s pe hou ; equency o bee e u ns,
defined as he numbe o bee en ances pe hou ; and o al bee
o aging equency, his being calcula ed as he sum o depa u es
and e u ns pe hou , and confi med ia he p esence o pollen
ca ied by e u ning bumblebees. To alida e he sys em, a subse
o ideos we e hen analysed manually, ensu ing conco dance
be ween esul s.
Semi-au oma ed obse a ion. To assess whe he ho ne s a emp-
ed o p eda e upon bumblebees a colony en ances, ideos om
he ex e nal came as we e analysed using BORIS ( elease . 8.5)66,
allowing obse e s o eco d a sui e o p especified e en s and
beha iou s (Fig. 1 ). Specifically, hese included ho ne p esence
a colonies, defined as ho ne s o aging in p oximi y o a colony;
ho ne du a ion a colonies, defined as he o al ime spen (s) by
ho ne s in p oximi y o a colony; p eda ion a emp s, defined as
ho ne s pu suing and subsequen ly making physical con ac wi h
bumblebees; p eda ion successes, defined as ho ne s subsequen ly
ca ching bumblebees; and p eda ion ailu es, defined as ho ne s
ailing o ca ch a bee a e making an a emp . Obse e s eco ded
hese e en s manually and he so wa e subsequen ly ex ac ed
beha iou ype, iming, equency, and du a ion da a (Fig. 1d).
P io o and du ing hese analyses, in e obse e consis ency was
alida ed using he BORIS in e - a e eliabili y unc ion.
Range o assessmen ac o s. Measu ed a iables we e b oadly
g ouped in o hose o ho ne abundance and beha iou , colony
g ow h, su i al, and o aging ac i i y, si e clima e, and land
co e . Wi hin hese g oups, we examined a combina ion o col-
ony-le el, indi idual-le el, and si e-le el pa ame e s.
Ho ne abundance and beha iou we e assessed using ho ne
coun s a si es o e ime, ho ne p esence a colonies, ho ne
du a ion a colonies, p eda ion a emp s, p eda ion successes,
p eda ion ailu es, and wo ke weigh a io.
Bumblebee colony g ow h and su i al we e e alua ed using
colony weigh , colony su i al s a us, o al adul popula ion,
mean weigh o adul bumblebees, numbe o open and closed
COMMUNICATIONS BIOLOGY | h ps://doi.o g/10.1038/s42003-023-05329-5 ARTICLE
COMMUNICATIONS BIOLOGY | (2023) 6:990 | h ps://doi.o g/10.1038/s42003-023-05329-5 | www.na u e.com/commsbio 9
