The silent dances of the Himalayan honeybee, Apis laboriosa Wh Kirchner, C Dreller, A Grasser, D Baidya

To cite this version:

Wh Kirchner, C Dreller, A Grasser, D Baidya. The silent dances of the Himalayan honeybee, Apis laboriosa. Apidologie, Springer Verlag, 1996, 27 (5), pp.331-339. ￿hal-00891377￿

HAL Id: hal-00891377 https://hal.archives-ouvertes.fr/hal-00891377 Submitted on 1 Jan 1996

HAL is a multi-disciplinary open access L’archive ouverte pluridisciplinaire HAL, est archive for the deposit and dissemination of sci- destinée au dépôt et à la diffusion de documents entific research documents, whether they are pub- scientifiques de niveau recherche, publiés ou non, lished or not. The documents may come from émanant des établissements d’enseignement et de teaching and research institutions in France or recherche français ou étrangers, des laboratoires abroad, or from public or private research centers. publics ou privés. Original article

The silent dances of the Himalayan honeybee, Apis laboriosa

WH Kirchner C Dreller A Grasser D Baidya3

1 Universität Konstanz, Fakultät für Biologie, 78434 Konstanz, Germany; 2 Department of Entomology, University of California, Davis, CA 95616, USA; 3 Department of Biology, Kathmandu University, Kathmandu,

(Received 25 February 1996; accepted 20 June 1996)

Summary — Comparative studies of the mechanisms of information transfer in the dance language of honeybees have recently led to an evolutionary scenario, assuming that acoustic signals replace visual communication signals in species that are forced to communicate in the dark, either because of nesting in dark cavities like Apis mellifera and or because of nocturnal activity like Apis dor- sata. To test this idea, the dance language of the giant honeybee Apis laboriosa, which is closely related to , but exclusively diurnal, was studied. We observed the dances, determined the flight range and recorded and analyzed acoustic signals emitted by worker bees. The dances of this species do not contain any acoustic signals, indicating that acoustic signalling of the location of food sources is indeed restricted to species which need to dance under low light intensities. The evolu- tionary implications are discussed. communication / sound / Asian honeybee / Apis laboriosa

Honeybee foragers communicate distance Dance communication is found not only and direction of profitable food sources to in A mellifera but in all species of - their nestmates by means of their dance bees studied so far (Lindauer, 1956). Among language (von Frisch, 1965). In Apis mel- the open nesting species of Asian honey- lifera, it has been shown that the information bees, the dwarf honeybee does is transmitted in the darkness of their nests not produce information coding dance through airborne sounds (Kirchner and sounds and presumably communicates the Sommer, 1992; Michelsen et al, 1992; location of food sources by means of visual Dreller and Kirchner, 1993a), which are per- signals (Towne, 1985), whereas the giant ceived by the dance attenders through honeybee Apis dorsata does emit dance Johnston’s organ, an auditory sense organ sounds which contain the information about located in the antennae (Dreller and Kirch- the location of the food source (Kirchner ner, 1993b). and Dreller, 1993). However, the probability of sound production, which is close to one in The purpose of the present study was A mellifera, turned out to be substantially therefore to observe the dances of A labo- lower in A dorsata, indicating that this riosa, compare the dance language of these species might be able to use both visual bees with the dances of other honeybee and acoustic signals. A dorsata forages and species and to clarify whether A laboriosa dances not only during the daytime, but also uses sound to communicate the location of during moon-lit nights (Dyer, 1985). Obser- feeding sites. vations of nocturnal dances of A dorsata revealed that the probability of dance sound production was significantly higher at night METHODS time than during the day (Kirchner and Dreller, 1993), supporting the hypothesis of The study was conducted in March 1995 in north- Towne that acoustic dance commu- (1985) east Nepal, close to the 8 586 m high Mount nication evolved in hon- signals might have Kantchenjunga (87’45"N, 27’30"E) at an eleva- eybees when they were forced to dance at tion of 2 100 m. A large colony of A laboriosa low light intensities. Smith 1871 (width and length of the comb both about 1.2 m) was found under an overhanging The Apis laboriosa, a giant honeybee, cliff about 150 m in a vertical cliff of a close relative of the lowland ledge high giant honey- deep river valley. In this area A laboriosa is geo- bee, A dorsata, can be found in the graphically separated from A dorsata. Further- Himalayan range at altitudes of 1 200 m to more, measurement of the cubital index of the 3 500 m (Sakagami et al, 1980; Roubik et al, wings revealed an extremely high cubital index 1985; Underwood, 1990). The species was of 17.5 ± 6.2 (n = 13), which is typical for A labo- first described by Smith in 1871, but there riosa and different from the cubital index of A dor- sata as et al was some discussion on the species sta- reported by Sakagami (1980). tus of this bee (Maa, 1953; Sakagami et al, Dances were observed through binoculars 1980; McEvoy and Underwood, 1988; Rut- from a distance of about 30 m. The dance tempo was determined by measuring the duration of tner, 1988, 1992; Smith, 1991), which was dance circuits with a stop watch, the directions considered as a mostly being subspecies indicated in the dances were estimated with a of A dorsata until quite recently. Many resolution of 15°. Individually marked bees were aspects of the biology of this bee, which trained to forage at an artificial feeding site, where lives in quite inaccessible places, are still 2 M sucrose was provided. Sound was recorded entirely unknown. This is especially true for using a pressure-sensitive and a velocity-sensitive the dance language of A laboriosa. microphone (previously used to record dance sounds of A mellifera (Michelsen et al, 1987)) on A laboriosa has never been found to be a digital audio tape recorder (Sony DAT recorder active at night and nocturnal activity would TCD-D3). Sounds of flying bees were recorded at seem to be quite unlikely because of low a distance of 1 m from the comb. Sounds emitted temperatures at night time. The dance lan- by bees standing on the comb were recorded at guage of this species is therefore especially a distance of 1-2 cm from the comb. Recorded sound were in the interesting, because it should allow a test signals analysed laboratory a CED 1401 interface and an of the hypothesis that acoustic communi- using laboratory IBM-compatible computer. cation of the location of feeding sites is an adaptation to foraging at low light intensi- ties. on the other acoustic com- If, hand, RESULTS munication of food sources already evolved in a common ancestor of the genus Apis and the silent dances of A florea were a During 8 days of observation of the colony derived trait, one would expect acoustic we observed intense foraging and dancing dance signals to be present in A laboriosa. activity. The main activity time of the bees was between 900 and 1600 hours, depend- in A mellifera could be heard from time to ing on the weather conditions. Returning time. An example of such a signal is shown forager bees advertised a number of differ- in figure 2. This signal was produced by ent feeding sites, mostly nectar sources, at bees standing close to the dancer, pre- different distances and directions. The pat- sumably emitted as a begging signal. tern of feeding sites indicated in a sample of 93 dances recorded on three consecutive days is indicated in figure 1a. The durations DISCUSSION of the dance circuits ranged from 2.0 to 11.1 s with a mean of 4.8 ± 1.8 s. 1b Figure The main result of the is the shows the distribution of circuit durations in present study lack of dance sounds in the dance commu- this of dances. with the sample Compared nication of A laboriosa. As the distribution of circuit durations found for A system recording system used in A laboriosa was dorsata in in a previous study, there the same as in previous studies of the dance was a significant difference in that the dance sounds of other honeybee species, it was circuits of A laboriosa lasted longer on aver- obviously suitable to detect dance sounds, age. As it turned out to be impossible to and as stop and noise train the bees to an artificial feeder at dis- signals background were recorded and of similar amplitude as in tances further than 50 m, we could not deter- other of the was mine the distance dialect of A laboriosa. species honeybees, system also the There- Assuming the distance dialect is similar to operating during recordings. fore we can exclude the possibility that the the distance dialect of A dorsata, we can lack of dance sounds is due to calculate the distances inferred by the inappropriate, insensitive or dancers. The mean distance would then be malfunctioning equipment. Dance sounds have been described inde- about 700 m, the longest distance indicated in A mellifera Esch and by the dancers about 2 km. pendently by (1961) Wenner (1962). These sounds are produced In all species of honeybees found to pro- by dorsoventral vibrations of the wings in duce dance sounds studied so the fre- far, round and wagging dances (Kirchner et al, of the dance sounds is similar to quency 1988) and are exclusively airborne sounds the of beat in There- frequency wing flight. (Michelsen et al, 1986), perceived by the fore, we determined acoustically the wing dance follower bees through Johnston’s beat of workers of A labo- frequency flying organ, located in the antennae (Kirchner et riosa. We found a mean beat fre- wing al, 1991; Dreller and Kirchner, 1993b). The of 133 ± 18 Hz This is quency (n = 20). sig- dance sounds provide the dance attenders lower than in the nificantly (P < 0.01) wing with the information about the location of beat frequency of A mellifera, but not sig- the food source (Michelsen et al, 1992; different from the beat fre- nificantly wing Kirchner and Sommer, 1992; Dreller and of A dorsata. quency Kirchner, 1993a). Similar dance sounds A sample of 40 vigorous and long-last- have also been found in the Asian bees A ing dances was observed from a short dis- cerana (Towne, 1985) and in A dorsata tance and all sounds produced by the (Kirchner and Dreller, 1993), whereas A flo- dancer and the dance attendance recorded rea does not produce dance sounds on digital audio tape. (Towne, 1985). Both open nesting species, None of the dancers produced any dance A florea and A dorsata, however, can per- sounds. The dances were entirely silent. ceive those sounds like A mellifera (Dreller Just short squeaking sounds similar to those and Kirchner, 1994). Therefore, we assume which have been described as stop signals that A laboriosa might also be able to per- that A laboriosa might also be able to per- location of food visually during the daytime, ceive sound. a species that is dancing exclusively diur- does not need additional acoustic Dance attenders in A dorsata are able nally sig- nals. Our result that the dances of the to pick up the information about the loca- A laboriosa are tion of the food source acoustically. How- Himalayan honeybee silent thus this ever, only a fraction of the dances of A dor- entirely supports hypothe- sis. sata for natural food sources contains dance sounds. This finding supported the hypoth- The squeaking sound (fig 2) is basically esis of Towne (1985) that acoustic signalling similar to the stop signal of A mellifera (Esch, of the location of food sources evolved when 1964; Nieh, 1993). The frequency is sub- ancestors of recent honeybees began to stantially lower, corresponding to the lower forage at low light intensities. In this sce- wing beat frequency of the much larger nario A dorsata reflects an evolutionary A laboriosa. A stop signal has also been stage in which acoustic signals can be used, described in A dorsata by Towne (1985). It and are used during nocturnal foraging, but is similar in signal duration. Sound frequency are probably not essential for successful was not analysed by Towne (1985). dance communication during the day. The distribution of dance circuit durations This idea was tested in our study of the indicates that the observed colony foraged dance language of A laboriosa. The rough in a wider area compared to a colony of climate at the high altitudes where this bee A dorsata observed in a previous study in lives excludes nocturnal foraging. If giant India (Kirchner and Dreller, 1993). The lack honeybees are able to communicate the of a calibration of circuit durations by train- ing bees to artificial feeders at different dis- A laboriosa, A mellifera and A cerana did tances prevented us from calculating a flight use acoustic signals in their dances. On the range of the observed colony. However, the other hand, we can ask whether the use of topography of the area where these bees visual signals for the transfer of information were studied, which is characterized by high in the dances of A laboriosa is a primitive mountains and deep and steep valleys, or a derived trait. In A dorsata dance sounds would limit the value of a distance dialect are found in only a fraction of the dances curve so long as we do not know exactly performed during the daytime and it seems how bees calculate distances. Whereas in likely that A dorsata can communicate the most places slope can be ignored in calcu- location of food sources without the aid of lations of distances, they may become sig- sound signals. Therefore, the common nificant in mountainous areas. In A mellif- ancestor of both giant honeybee species era indication of distance seems to be should have been able to use both chan- affected by the topography of the landscape nels of information transfer, sound as well as (von Frisch, 1965). visual cues. Our result that A laboriosa dances However, the question remains whether silently, whereas A dorsata produces dance this is also true for the common ancestor of sounds, supports, as discussed above, the the giant honeybees and the A mellifera hypothesis that acoustic signalling of the group. The dwarf honeybee A florea nests in location of food sources is an adaptation to the open, dances silently and uses visual dancing at low light intensities. In an evo- cues to transfer dance information. If we lutionary perspective the silent dancing of assume, as did Lindauer (1956), that this A laboriosa could either reflect a primitive represents the ancestral mode of information trait or a derived trait. The phylogeny within transfer in the dance language of honey- the genus Apis is not yet fully clarified (see bees, the next step must have been the evo- Koeniger, 1976; Smith, 1991). However, lution of the acoustic channel of communi- one can assume that A laboriosa is quite cation in the common ancestor of the closely related to A dorsata, the lowland A mellifera group and the A dorsata group. giant honeybee. According to Sakagami et However, it is also possible, as Koeniger al (1980), it is most likely that the origin of (1976) pointed out, that the common ances- A laboriosa is interglacial or postglacial. The tor of the recent species in the genus Apis first giant honeybees appeared already in was a cavity-nesting species. In this case the Miocene and lived under a warm Ter- we have to assume that the use of visual tiary climate. If the silent dancing in A labo- signals in the dance communication system riosa was a primitive trait, we would have has evolved twice, in the ancestor of A flo- to assume that A dorsata and the A mellifera rea and independently in the ancestor of group have independently developed acous- the giant honeybees, whereas the A mellif- tic signals in the dances. The facts that era group would have conserved the prim- A mellifera and A cerana use sound signals itive trait of an exclusively acoustical trans- that are quite similar to those of A dorsata fer of information. This hypothesis can be (Towne, 1985; Kirchner and Dreller, 1993) tested: A mellifera has been observed to and that A mellifera- and A dorsata-like bees dance in the open under experimental con- are known since Tertiary times, whereas ditions (von Frisch, 1965). If A mellifera A laboriosa is much younger, make it could communicate the location of food unlikely that the silent dancing of A labo- sources without the aid of sound signals riosa is a primitive trait. We assume instead, when they are dancing in the open, it would that the common ancestors of A dorsata, be possible to conclude that visual transfer of information must be a primitive trait and Tanzrichtung bestimmt. Im Mittel betrug die that the latest common ancestor of all recent Umlaufdauer 4,8 ± 1,8 s, die längste beob- honeybees was nesting in the open. achtete Umlaufzeit war 11.1 s. Unter der Voraussetzung, daß der Entfernungscode dem von Apis dorsata gleicht, entspricht ACKNOWLEDGMENTS das einer mittleren Entfernung der Futter- quellen von etwa 700 m, als größte beob- achtete von We thank the administration of the Kathmandu Entfernung Futterquellen ergibt sich etwa 2 km. Da die Tanztöne anderer University for support of our applications for a research permission for the Kantchenjunga area, Honigbienenarten in ihrer Frequenz der Flü- Ming-Ma Sherpa, a skilled mountain climber and gelschlagfrequenz im Flug entsprechen, honey-hunter, for his assistance at the research wurde diese in 1 m Entfernung von der spot which would have been inaccessible for us Wabe an vorbeifliegenden Bienen bestimmt. without of his and the German Research help, Sie beträgt im Mittel 133 Hz ± 18 Hz. Bei Council DFG for financial support of our research. einer Stichprobe von 40 lebhaften und lan- ganhaltenden Tänzen wurden alle Geräu- sche aus einer Entfernung von 1-2 cm von Die stummen Zusammenfassung — der Wabe aufgezeichnet und analysiert. In Tänze der aus dem Hima- Honigbiene keinem der Tänze traten Tanzlaute auf. laboriosa. In der laya, Apis Tanzsprache Während die Tänzerinnen von Apis labo- teilen ihren Honigbienen Nestgenossinnen riosa völlig stumm sind, konnten gleichzei- Richtung und Entfernung lohnender Futter- tig kurze piepende Signale, die den bei Apis quellen mit. In der Dunkelheit ihrer Nester mellifera und Apis dorsata auftretenden benutzt die westliche Honigbiene Apis mel- Stop-Signalen ähneln, registriert werden lifera akustische zur Signale Weitergabe (Abb 2). Die Tatsache, daß die Tänze von dieser Informationen. Unter den im Freien Apis laboriosa stumm sind, unterstützt die nistenden Arten, die auch mit Hilfe optischer Vorstellung, daß sich akustische Kommu- Signale kommunizieren können, benutzt nikationssignale bei Honigbienen als Anpas- florea keine für die Tanz- Apis Schallsignale sung an die Notwendigkeit von Tanzkom- kommunikation, während Apis dorsata Tanz- munikation bei niedrigen Lichtintensitäten töne produziert, die Information über die entwickelt haben. Lage der Futterquelle enthalten. Diese Töne treten allerdings nicht wie bei Apis mellifera Tanzkommunikation / Schallproduktion / in allen Tänzen auf. Bei nächtliche Tänzen asiatische Honigbiene / Apis laboriosa ist die Wahrscheinlichkeit der Schallpro- duktion erheblich höher als am Tage. Diese Beobachtung führte zu der Hypothese, daß Résumé — Les danses silencieuses de die Evolution akustischer Signale in der l’abeille de l’Himalaya, Apis laboriosa. Tanzsprache eine Anpassung an die Not- Par le langage dansé les abeilles mellifères wendigkeit ist, bei geringen Lichtintensitäten (genre Apis) communiquent aux congénères zu kommunizieren. Diese Hypothese wurde du nid la direction et la distance des sources an der Riesenhonigbiene des Himalaya, de nourriture à exploiter. Dans l’obscurité Apis laboriosa, die eng mit Apis dorsata ver- de son nid l’abeille mellifère occidentale, wandt, aber nicht nachtaktiv ist, getestet. Apis mellifera, utilise des signaux acous- Die Tänze von Apis laboriosa wurden an tiques pour transmettre ces informations. einer Kolonie in Ost-Nepal in 2100m Höhe Les espèces qui nidifient en plein air peu- beobachtet. In einer Stichprobe von 93 Tän- vent aussi communiquer par signaux zen (Abb 1) wurde die Umlaufdauer und die visuels. Parmi celles-ci, Apis florea n’utilise pas de signaux sonores alors qu’Apis dor- les danses dans les cas de faible intensité sata produit au cours des danses des sons lumineuse. qui contiennent l’information sur la situation de la source de nourriture. En fait ces sons Apis laboriosa / communication sonore / ne sont pas présents, comme chez A mel- danse lifera, dans toutes les danses. Leur proba- bilité est considérablement plus élevée la nuit que le jour. Cette observation a conduit REFERENCES à émettre l’hypothèse selon laquelle l’évo- lution des dans le lan- signaux acoustiques Dreller C, Kirchner WH (1993a) How honeybees per- gage dansé serait une adaptation à la ceive the information of the dance language. Natur- nécessité de communiquer dans les cas de wissenschaften 80, 319-321 faible intensité lumineuse. Cette hypothèse Dreller C, Kirchner WH (1993b) Hearing in honeybees: localization of the auditory sense organ. J Comp a été testée sur l’abeille de l’Hima- géante Physiol A 173, 275-279 laboriosa, est étroitement laya, Apis qui Dreller C, Kirchner WH (1994) Hearing in the Asian hon- apparentée à A dorsata mais n’est pas eybees, Apis dorsata and Apis florea. Insectes Soc active la nuit. Les danses d’A laboriosa ont 41, 291-299 été observées sur une colonie située à Dyer FC (1985) Nocturnal orientation by the Asian , dorsata. Anim 769-774 2 100 m d’altitude dans le nord-est du Népal. Apis Behav 33, La durée et la direction de la danse ont été Esch H (1961) Über die Schallerzeugung beim Werbe- tanz der Honigbiene. Zvergl Physiol 45, 1-11 déterminées sur un échantillon aléatoire de Esch H (1964) Beiträge zum Problem der Entfer- 93 danses durée a été (fig 1 ). La moyenne nungsweisung in den Schwänzeltänzen der Honig- de 4,8 ± 1,8 s (maximum observé 11,1 s). À biene. Z vergl Physiol 48, 534-546 condition que le code de distance soit le Frisch K von (1965) Tanzsprache und Orientierung der même que celui d’A dorsata, la distance Bienen. Springer, Berlin moyenne des sources de nourriture a été Kirchner WH, Dreller C (1993) Acoustical signals in the dance of the dor- évaluée à 700 m environ language giant honeybee, Apis (maximum sata. Behav Ecol Sociobiol 33, 67-72 observé : 2 km environ). Puisque chez les Kirchner WH, Dreller C, Towne WF (1991) Hearing in fré- autres espèces d’abeilles mellifères la honeybees: operant conditioning and spontaneous quence des sons des danses correspond reactions to airborne sound. J Comp Physiol A 168, à la fréquence des battements d’ailes en 85-89 vol, on a déterminé celle-ci sur des abeilles Kirchner WH, Lindauer M, Michelsen A (1988) Honeybee dance communication: acoustical indication of direc- volaient à 1 m du Elle était en qui rayon. tion in round dances. Naturwissenschaften 75, 629- moyenne de 133 ± 18 Hz. Sur un échan- 630 tillon aléatoire de 40 danses vigoureuses Kirchner WH, Sommer K (1992) The dance language et de longue durée on a enregistré et ana- of the honeybee mutant diminutive wings. Behav Ecol Sociobiol 30, 181-184 lysé tous les bruits à une distance de 1 à N Neue der inner- 2 cm du rayon. Aucune de ces danses n’a Koeniger (1976) Aspekte Phylogenie halb der Gattung Apis. Apidologie 7. 357-366 été accompagnée de sons. Les danseuses Lindauer M (1956) Über die Verständigung bei den indi- sont d’A laboriosa restées totalement silen- schen Bienen. Z vergl Physiol 38, 521-557 mais on a de cieuses, pu enregistré temps Maa TC (1953) An inquiry into the systematics of the en temps des petits cris aigus qui ressem- tribus Apidini or honeybees (). Treubia blaient aux signaux d’arrêt déjà décrits chez 21, 525-640 A mellifera et A dorsata. Le fait que les McEvoy MV, Underwood BA (1988) The drone and danses d’A laboriosa soient silencieuses species status of the Himalayan honey bee, Apis laboriosa (Hymenoptera: ). J Kans Entomol conforte l’idée que les signaux acoustiques Soc 61, 246-249 chez les abeilles mellifères sont le résultat Michelsen A, Andersen BB, Storm J. Kirchner WH, Lin- d’une adaptation de la communication par dauer M (1992) How honeybees perceive commu- nication dances, studied by means of a mechanical Ruttner F (1992) Naturgeschichte der Honigbienen. model. Behav Ecol Sociobiol 30, 143-150 Ehrenwirth, Munich Michelsen A, Kirchner WH, Lindauer M (1986) Sound and Sakagami SF, Matsumura T, Ito K (1980) Apis laboriosa vibration signals in the dance language of the honey- in Himalaya, the little known world largest honey- bee, Apis mellifera. Behav Ecol Sociobiol 18, 207-212 bee (Hymenoptera, Apidae). Insecta Matsumurana Michelsen, A, Towne WF, Kirchner WH, Kryger P (1987) (NS) 19, 47-77 The acoustic near field of a dancing honeybee. Smith DR (1991) Diversity in the Genus Apis. Westview J Comp Physiol A 161, 633-643 Press, Boulder, CO Nieh JC (1993) The stop signal of the honey bee. Behav Towne W F (1985) Acoustical and visual cues in the Ecol Sociobiol 33, 51-56 dances of four honey bee species. Behav Ecol Socio- biol 16, 185-187 Roubik DW, Sakagami SF, Kudo I (1985) A note on the distribution and nesting of the Himalayan honey bee Underwood BA (1990) Seasonal nesting cycle and migra- Apis laboriosa Smith (Hymenoptera: Apidae). J Kans tion patterns of the Himalayan honey bee Apis labo- Entomol Soc 58, 746-749 riosa. Natl Geogr Res 6, 276-290 Ruttner F (1988) Biogeography and of Hon- Wenner AM (1962) Sound production during the waggle eybees. Springer, Berlin dance of the honey bee. Anim Behav 10, 79-85