(1991) Behav [,col Sociobiol 29:]61 365Behavioral Ecologyand Sociobiologyf Springcr-Vcrlag 1991The role of the queen in circadian rhythms(Apis mellifera L.)of honeybees Robin F.A. Moritz' and Fritz Sakofski2I Depärtmcnt of Cienctics. University of Natal. P.O. Box 375. Pictcrmaritzburg 3200. South Atiica2 Baycrischc Lzrndesanstalt für Bicncnzucht\" Frieclrich Alcxander Universität, Burgbergstrassc 70, W-8520 ErlangenFcdcral Republic of GermanyReceivcd Novembcr 6. 1990 i Acceptcd Julv 5. 1991Summary. Colonies and smaller social groups of honey-bees (lpl.l mellif aru c'arnicct L.) show distinct free-run-ning circadian rhythms similar to that of individuai ganisms. or-Thc workers of a colony synchronize their indi-vidual rhythms to one overall group rhythm. Caste playsan important role in this synchronization process.Queens were introduced into worker groups which wereentrained to a phase-shifted light/dark cycle. The intro-duction of the queen caused a shift in the free-runningphase under constant dark conditions. Single introducedworkers had no effect on the free-running rhythms. Thisindicates that the queen plays an important role in thesynchronization of circadian rhythms of honeybee colo-nies.IntroductionIt has been known for many years (Forel 1910) thathoneybees and their colonies maintain circadian rhythmsin their activity. The lightldark (LD) cycle (e.g. Beier1968: Moore and Rankin 1983: Frisch and Aschoff19; Moore et al. 19), foraging conditions (Kleber1935) and social parameters (Medugorac and Lindauer1967) strongly influence their phases ofactivity and rest.The circadian activity rhythms of individual workershave been rcpeatedly analyzed (Spangler 1972: Kaiserand Steincr-Kaiscr 1983: Moore and Rankin 1985:Frisch and Koeniger 1992), and free-running periods oflcss than 24 h (rttnge 21.8 to 23.5 h) under (DD) constant darkhavc been revealed. There is considerable variationamong the workers. which is not surprising since evenworkers ol one single colony are not genotypically phcnotypically oridentical. This variation is particularlylarge as a result of rnultiple rnating of the queen (Adamset al. 1978). How can a colony rnaintain an overall circa-di;rn rhythm in light of such large variance in individualOftprint request.t to. R.F.A. Moritz. Institut für Biologie. ischc Universittit Techn-Bcrlin. Franklinstrassc 28i29. W-1000 tserlin Federal Rcpublic 10,of Gcrnranycolony members in a colony of up to 80000 individuals?Intuitively, one presumes some kind of social synchroni-zalion of the individual activity cycles. which producesa properly functioning colony.Social synchronization has been studied mainly invertebrates. For example, species-specific songs can syn-chronize activity rhythms in birds (Gwinner 1966; Men-aker and Eskin 1966). In insects, social synchronizationof free-running activity has not been observed. For ex-ample, the song of male crickets, Theleogn,llus ('ot11tilo-dus, causes females to run, but fails to entrain an activityrhythm (Loher 1979). Social zeitgebers are believed robe most significant for organisms that lack environmen-tal zeitgebers in their nesting site (e.g. cave-nesting or-ganisms such as beavers (Bovet and Oertli 1914) andbats (Marimuthu et al. 1918, 1981)). Honeybees (17;rsmelli/bra) are also cave-nesters, and social zeitgebers maybe equally if not morc important for their colony bio-rhythms. Indeed, a variety of social factors seems tobe involved in the control of colony activity: individualworkers have been shown to adopt the rhythms of socialgroups very rapidly; foragers introduced into a new col-ony foraged at both their own entrained tirne and atthe activity peak of the new colony (Medugorac andLindauer 1967). Frisch and Koeniger (1992) suggesrcdthat within the colony there is mutual synchronizationof the various individual worker rhythms to an overallgroup rhythm. Because of the great complexity of socialgroupsl however, the actual mechanisms behind this ex-traordinary social behaviour are still unknown. Tactilecontact was suggested by Southwick and Moritz (1987)to be an important instrument in such synchronization.They entrained workers to phase-shifted lighr-dark (LD)cycles and found that separately entrained workers wereonly able to synchronize their rhythms if physical con-tact and food exchange were possible. Restriction ofcon-tact via a single screen that allowed trophallaxis did notinterfere with synchronization. Howevcr, a doublescreen precluding physical contact prevented the devel-opment of a common group acti\\ ity.A question rarely addressed is the role of the castesystem in colony activity-rhythms. Sex scems to have362somc impact, with drones displaying longer daily activitythan workers (Spangler 1972). However, the special rolequeen poses of the the really intriguing question concern-ing social regulation of colony rhythms. Her part in thedevelopment of a the activity cycle of a colony couldwell be very dilferent than that of any other colonymember. Although many of her activities are controlledvia her worker force, her impact on colony life is muchmore dramatic than that of any worker or drone. Thebehaviour of queenless colonies is completely dilferentfrom that of queenright hives in respect to such diverseactivities as defensive behaviour, foraging and the repro-ductive strategy of the colony. Although queens aresometirnes referred to as egg-laying machines, implyingthey are nothing but a worker bec with better developedovaries, thcre seems to be a strong potential for queensto influence worker rhythms, such as through fluctua-tions in pheromone production or egg laying. This paperreports on a study of the impact of the queen on circa-dian rhythms of honeybee worker groups.phase. Each queen was uscd lbr only one singlc cxperimenl toavoid el-fects of habituation of thc queens to the Ilight-room c1,'clc.The test groups were housed in small cagcs that were suppliedwith sufficient sugar syrup to survivc a 14-clay pcriod ofconllnc-ment in the cagc. The test bccs were kcpt in constaltt darkness(DD) at 20\" C and l0o/o r.h. In control expcrimcnls. instead ofa queen. a worker entraincd to field conditions rvas irrtroducedinto a group of 150 llight-room workers. The leli lbrcwing ol-theintroduced worker was clipped, to determine whethcr shc u'as stillalive at thc cnd of the expcriment. In sevcr:rl cases thc introduccclworkcr was killed by thc foreign workcrs from thc flight roonr;thesc data were not included in the analyses. ancl clrccn-Queenlcss right workers fiom the field and the llight-roorn wcre testcd ascontrols.Metubolic bioassay. Tcn groups could bc tested simultancously inan automalically recording sctup (Fig. 1) which analyzed thc meta-bolic activitl' of the groups by measurin-u thcir oxygen consump-tion. Air was pumped at a ratc of 200 mlirnin through airtightcontainers (250 ml) which houscd the cages. A thin Tvgon tubc(2 mm diam) was connected with a T-piece to thc outlet. and airsamples werc pumped at 30 mllmin through an oxygen analvzcr(Applied Elcctrochemistry S3-A). A computer-controlled manifbldsclccted sample air scquentially 1l'om each of the ten tcst groups.The analyzer read-out was recorded on linc via an AiD intcrfaccto the computer. Every 6 rrin, the solenoid systcrn of the rnaniloldselected thc ncxt test cage: 4 min of llushing thc analyzing systen\"rwith the ncw sample air was followed by a 2-nrin data-rcadinginterval\" alier which the aleragc o\\)gcu conüelttlirrion during thc2-min period was stored on a hard disk. The experinrcnt wits colr-tinucd as long as metabolic activity could bc measurcd (usuallyabout 14 days).The data were analyzed with thc time-series analyscs of thepackage. STATGRAPHICS The pcriods o1- the group actirirl-rhythms were determincd by regression ünalysis using thc pcakactivities in a periodogran.r plot (Enright 1981). The phase-shifiwas calculated by lotting the pcak activities of oxygcn lcvcl againsttimc and smoothing thc data with tlrc moving averagc tcchnique(meanover*5h).MethodsBee hrecding and phuse entrainment. All experimcnls were con-ducted with Apis ntellifera turnica L. from commercial stock com-monly used fbr honey production. A full-size honeybcc colonywas kept in an indoor flight-room similar to that of van Pr.aagh(1972). The bees were allowcd to forage for pollen and sugar syrupad lib. Thc tempcrature was kcpt constant at 25'C*1\" C andrclative humidity (r.h.) ranged from 60 65%. The indirect lightingof thc flight-room \\vas switched on 8 h after sunrise and switchcdofT 8 h after sunsct. In this artiflcial LD regime, thc flight-roomcolony w:rs cntraincd to a phasc shift corrcsponding to the activitycyclcs ol'colonies kept in the field.Twenty sister qucens (thcy were also sislers of thc queen inthc f'light-room colony) wcrc reared by standard techniques(Ruttner 191t0) and kept in small mating nuclei (approximately2000 3000 bees) in the fleld. They were allowed to mate freelyand werc used for the experiments only after- they had startedegg-laying. The quccns were introduced to groups of 1 50 workersliom the flight-room colony whcn both fleld and flight-room colo-nies wcrc expcricncing light conditions and werc in their activcResultsA total of 47 groups of Api,s ntelli/bra carnit'a weregroups tested: 16 flight-room bee with a field quecn.14 queenless flight-room groups. 10 flight-room beeIHaFig. 1. Diagram of cxpcrimcntal sctup. Ten testgroups of Api.:; tnellifera trtrnitu were vcntilatcd atratc of 200 mlimin. A compuler-controllcd manifbldsequentially sclcctcd sarrplcs ol- air fronr sach ol thcten test groups (30 mlrmin). Orygen concerrtlrtionwas analyzed in a zirconia cell and the dutu storcd lbr'further processing363groups with one introduced field worker_bee. 3 queenlessfield-bee groups, and 4 queenright field-bee groups.Fig. 2 is a typical plot of oxygen level as a functibnof time, for a 7-day period. In rare cases (r:2) thebees suffered from confinement and displaye; 1.rs_pro_nounced activity peaks after a few days. We did notinclude data from groups that showed circadian activitvof 7 days in our analyses; the period was estirnated bvregression analysis of the periodograms (Fig. 3). Theräwere no significant differences among the lengths of thevarious activity cycles (Table 1). Also, the activity phasesof queenright and queenless groups from the field didnot differ.o/o oxygen in airAlthough the light cycle in the flight room was olflerby 8 h, the phase-shift in the activitycycle of the indoorbees in relation to that of the fleld beös was onlv 6.1+0.4 h (92 angle). The introduction of sinsle fielcl worker_bees into groups of 150 flight room riorkers dicl notcause any_significant phase-shift (+0.++0.6 h; r:10).The introduction of queens from field colonies into theflight-room groups did have significant cffects. On anaverage, the activity phase was shifted towards thequeen's rhythm by -1.38+0.26 h (Student's l_tesr:2.93, p<0.01), i.e., a 22.5o/o of the totul shift betweenqueen and workers. This indicates that the queen hada significant effect on the endogenous group rhythnt.-21 20t'\'l18'*\\o['\"1,t1 \\'''ri'\\ül4n\"r,,r]Table 1. Apis mcllifera curnica. Chatactcristics of free-runtrtng ac_tivity rhythms of bees entrained to flighr-room (in t h.It seems unlikely that the queen moves around andphysically entrains each individual group member to thcnew group rhythm. Queen control over workers is gener'-ally achieved via pheromones; this leads to the intriguingquestion as to whether pheromonal cues n-rodily the cir-cadian activity cycles of thc workers. Apparcntly thefactor that controls the synchronization mechanisrn isqueen-specific, since the introduction of out-of-phaseworkers had no detectable effect on the group. Sinceit is mainly her pheromonal bouquet that distinguishesa queen from a worker honeybee (Crewe 1982), phero-monal fluctuations would seem to be prime candidatesfor the synchronization between the activity rhythms ofthe queen and her colony. Furthermore. workers in-crease their metabolic activity when they are exposedto various fatty acids of the queen's mandibular glandsecretions (Moritz and Crewe 1988);they do not incrcascactivity when exposed to worker-specific compounds.The phenomenon of circadian pheromone rclcasc hasbeen demonstrated for various insect spccics. Scx attrac-tants, especially, have been found to be released by \"call-ing\" females only during an active period of potentialmating. A typical example is the dermestid beetle Trogo-clernu glabrum (Roelofs 1978).Although the honeybee queen has partral control ovcrcolony activity. this does not mean that her workcr fbrceis unimportant for the determination of phase and peri-od of the activity cycle. Certainly, the workers also all'cctthe activity of the queen. There is an initial phase-shifton the first day after the introduction of the field-queen.but thereafter, the free-running period is identical to thatof queenless flight-room groups. If the queen had rnain-ttrined her individual rhythm. phase shifts should haveoccurred on the following days, also affecting the periodestimates. If the workers had no effect on the queen,the groups should eventually have been entrained to thefree-running cycle of the queen. Effccts of workers onqueen activity are not new. One example is the preparer-tion of the virgin queen lbr the mating flight, when theworkers push, pull and harass her through the nest untilshe, almost \"reluctantly\leaves the hive for hcr matingflight (Hammann 1954). Furthermore, freshly emergedworker bees seem to be arrhythmic r,rntil they are en-trained by older workers of the colony. In a naturalcolony the sarne may be true for the freshly cmergedqueen, and she will adopt the synchronized colonyrhythm. Therefore, under natural conditions. there isno phase angle between the activity cycles of the queenand the workers because both are synchronized anyway.Our artificial experimental conditions have revealed therelative importance of the queen and the workers in es-tablishing the overall colony rhythn and havc shownclearly that the queen plays a very special role in thecolony, controlling the workers in ways not previouslyconsidered.How the synchronization ol queen and workerrhythms is achieved remains to be resolved. Various test-able mechanisms are possible. Close contact seems tobe important for workers. since groups separated by adouble wire mesh are unable to synchronize theirrhythms (Southwick and Moritz '1987). Thc queen'spheromoncs arc also dispersed throughout the colonyby tactile worker-interactions (Seeley 1980), which is inagrcement with thc hypothesis of pheromonal controlof the group activity-rhythm. However, the final pieccof evidence in support ol this hypothesis, the circadiandynamics of pheromone relcase. still requires documen-tation. Irrespective of the actual synchronization mecha-nism, our data clearly show that social zeitgebers arenot confined to vertebratcs and that the oueen ls animportant gear wheel in the clock controlling the activityrhythms of social groups of honeybees.Atkrunrledscmcl?/,t. We thank the Br.rndesministcrium für For-schung und Technik lbr llnancial support. Wc arc grateful to E.1...Southwick fbr hclpful comments on thc manuscript.ReferencesBcicr W (1968) I3ccinllulJung der innercn Uhr der Bienen durchPhaserrvcrschiebung dcs Licht-Dunkel-Zeilgcbcrs. Z Bienen-forsch 9:356 378Bovet J, Ocrtli EF (1974) Frcc running circadian activity rhythmsin f rec living bcaver ((?,rtol tctrtutlen.si.s). J Comp Physiol 92: I10Crewc RM (19ti2) Corripositional variability: the key to thc socialsignals produccd by honcy b!'c mandibular glar-rds. In: BrceclMD. Michcncr CD. Evans HE (eds) Thc biology ol social in-sccts. Westvicw Press. Bouldcr. CO, pp 3llJ 321Enright T (1981) Data analysis. In: AschofT.l (cd) Handbook ofbchavioral ncurobiology. vol ,1. Biological rhythnrs. PlcnumPress. New York. pp 21 39Forel AII (1910) Das Sinnesleben der Insekten. Ernst RcinhardtVerlag. MunichFrisch B (19134) Circadian rhythms of horrcy bces. Apidologic1 5:278 280Frisch B. AscholT.l ( 1 987) Circadiirn activity in honcybees : entrain-rnent by I'eeding cycles. Phlsiol Entornol 12:41 49Frisch B. Kocui-rlcr N (1992) Mr.rtual s1'nchronization of the activityrhythms of honcybccs within a colony. Bchav Ecol Sociobiol(rn press)Grvinncr Ir (1966) []ntrajnmcut ol- circadiau rhl'thrn in birds byspecics specilic strng cycles (Avcs. Fringillidac'. Curduelis spinu.s.Serintrs sarinus). Erpcrientia 22 : 765llamnrann E (1957) Wer hat dic lnitiatiie bci den Ausflügen dcrKirnigin. dic Königin odcr die Arbcitsbicnen'l lnsectes .l:91 Soc106Karser W. Steincr-Kaiser .l (1983) Neurotral corrclates of sl.365wakefulness and arousal in a diurnal inscct. Nature (London)301:701 709Klcbcr I (1935) Hat das Zeitgedächtnis der Bier.ren biologischc Bc-deutung? ZYcrgl Physiol 22:221 262Loher W (1979) The el-fect of n-rale calling on ltmalc locornotoractivity of Tltelcogrrllus comtnotlus. Behav L.col Sociobiol5:383 390Marimuthu G, Subbaraj R, Chandrasckaran MK (1978) Socialsyncronizatiou of the activity rh.vthnt in a cavc-dwclling insccti-vorous bat. Naturwisscnschaften 65 : 600Marimuthu G. Rajan S. Chandrasckirran MK (1981) Social cn-lrainmcnt of the circadian rhythm in thc l-light activity ol- thcmicrochiroptcran bat Hippo.sidcros .speori,s. Bchav Ecol Socio-biol 8: 147 1 50Mcdugorac l\" Lindauer M (1967) Das Zcitächtnis dcr Bicncnuntcr dem Einfluß von Narkosc und sozialcn Zcit-rrcbct'n. ZVcrgl Physiol 55:450 474Mcnakcr M. Eskin A (1966) L,ntrainnrcnt of circadian r-hythnrsby sound rn Passcr dontc.stit u,y. Scicncc I 5.1 : I 579 I 58 1Moole D. Rankin MA (1983) Diurnal changes in the accuracyof thc honeybee fbraging rhythrn. Biol Bull Mar Uiol l-abWoods Hole 1:471 482Moore D. Rankin MA (1985) Circadian locomotor rhythnrs inindividual honcybccs. Physiol E,nton.rol 10:191 197Moore D. Siegliied D, Wilson R. Rankin MA (19) The rnlluenccof timc of day on the lbraging behalior ol- thc honcybcc. .,17;r.rnellileru. J Biol Rhythms 4:305 325Moritz RFA. Crcwc RM (1988) Reaction o1- honel,bee u'orkers(Api.s mclli./eru) to latty acids in quccn signals. Apidolo_uic19:333 342Praagh JP van (1972) Towards a controllcd cnvironmcnt roon.rsuitablc lor normal colonl' lif'e in honcybccs. 1. Dcscriptionand gencral observations. J Apic Res 11:71 8lRoelofi WL (1978) Chemical control of insects by phcronroncs.In: Rockstein M (cd) Biochernistry' of insects. Acadenric [)ress.Ncr.v York, pp 419 1Ruttner F (1980) Königinnenzucht. Apimodia. BucharcstSccley T (1979) substancc dispersal by messenger workcrsin honeybee Quccn colonics. Bchav Ecol Sociobiol 5:391 41 5Southwick Eh,. Moritz RFA (19ti7) Social sl,nchronizarion o1'circa-dian rhythms of melabolism in honevbe e s (;17rl.r nu,llileru).Physiol Entomol 12:2.09 212Spangler HC (1912) Daily activity rhythrns of indir,idual n'orkerand dronc honeybees. Ann Entomol Soc Arr 65: 1 073 1 075Visvanathan N. Chandrasekaran MK (1984) Mothcr mousc sctsthe circadian clock of pups. [)roc Indian Acad Sci (Anim Sci)93:235 241Visvanathan N, Chandrasckaran MK (1988) Limits o1- malcrnalcntrainment of the activity rhythm in thc fielcl mouse Mir,r ärlo-rftrga. .l Conrp Ph1'siol (A) 163:231 242h-
