Grooming behavior by Apis mellifera L. in the presence of woodi (Rennie) (: ) Jeffrey S. Pettis, Tanya Pankiw

To cite this version:

Jeffrey S. Pettis, Tanya Pankiw. Grooming behavior by Apis mellifera L. in the presence ofAcarapis woodi (Rennie) (Acari: Tarsonemidae). Apidologie, Springer Verlag, 1998, 29 (3), pp.241-253. ￿hal- 00891491￿

HAL Id: hal-00891491 https://hal.archives-ouvertes.fr/hal-00891491 Submitted on 1 Jan 1998

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

Grooming behavior by Apis mellifera L. in the presence of (Rennie) (Acari: Tarsonemidae)

Jeffrey S. Pettis Tanya Pankiw

Department of Biological Sciences, Simon Fraser University, Burnaby, BC V5A 1S6, Canada

(Received 9 December 1997; accepted 6 January 1998)

Abstract - The role of grooming behavior by the honey , Apis mellifera L., in limiting the infes- tation of, or being elicited by, the parasitic Acarapis woodi was investigated. Grooming behaviors examined included allogrooming and the grooming dance that involves self or auto- grooming. Observation hives monitored over 24 h revealed that dancing increased significantly at night while allogrooming decreased. In 32 mite-infested observation hives the percentage of infested was positively correlated with allogrooming acts and dances observed. In a third expe- riment, young marked bees were introduced into three hives with 0, 50 and 70 % tracheal mite pre- valence and grooming dances increased significantly in the bees 1-3 d of age in the mite-infes- ted colonies. We postulate that mite movement on young bees elicits the grooming dance. Bees from four different single patrilines that had exhibited different propensities to allogroom or dance were marked and placed into eight mite-infested colonies for 5 d. Dissections of marked bees revealed that the allogrooming line was most susceptible and the dancing line least sus- ceptible to mite infestation. We postulate that the dancing line of bees had a lower threshold for detecting on their body resulting in increased dance behavior and autogrooming, which we propose lowered the number of mites that transferred to these bees. This is the first evidence for a mechanism of resistance to the tracheal mite. © Inra/DIB/AGIB/Elsevier, Paris Apis mellifera / Acarapis woodi / grooming / behavior

1. INTRODUCTION obligate parasites [1]. One such obligate parasite of adult bees is the tracheal mite, A colony of honey bees, Apis mellifera L., Acarapis woodi (Rennie) [4, 22]. All mite plays host to a variety of organisms from life stages occur within the tracheal system large vertebrate pests such as mice to small of adult bees and only mated females

* Correspondence and reprints: USDA-ARS Bee Research Laboratory, BLDG. 476 BARC-E, Beltsville, MD 20705, USA E-mail: [email protected] ** Current Address: Department of Entomology, University of California, Davis, CA 95616-8584, USA migrate to new host bees [10]. Mites Dancing bees will autogroom with their appear to be unable to move off the host mesothorasic legs in addition to being allo- onto comb or other structures and then groomed. More recent studies have regain access to a bee; no live mites have demonstrated that allogrooming [5, 12] is been found in close inspection of combs in part genetically determined. Nothing is from mite-infested colonies; ([15]; Pettis known as to the age at which bees per- pers. obs.). Mite movement would thus be form these behaviors. Particular genotypes confined to adult bees. Dispersing females are more likely to allogroom given some preferentially infest adult bees less than 3 d unknown level of stimuli within a colony. of age [7, 13], and locate these bees in part Allogrooming occurs primarily on the tho- by attraction to host cuticular hydrocar- rax and especially around the wing bases bons [21]. Additionally, tracheal mites [2] where tracheal mites are found during disperse more at night, or are more suc- dispersal [23]. Tracheal mite dispersal suc- cessful, compared to daytime dispersal cess increases at night [20] and perhaps [20]. grooming behavior may vary day-night if involved in resistance to this parasite. Bees have been shown to exhibit phe- No factors have been demonstrated and differences in the notypic genotypic which elicit the grooming dance and number of mites they acquire when com- is known about the age at which to other bees in the same nothing pared young bees or dance. no environment Two broad mecha- allogroom Additionally, [6, 17]. benefits from allogrooming have been nisms are for the possible explanations demonstrated, with to the observed differences in mite levels be- except regard exotic parasite Varroa [19]. In the present tween bees in a common environment: 1) studies we examined: behavioral differences between the bees 1) the frequency of allogrooming and or the mites toward the bees (e.g. induced dancing on a day-night basis; chemical cues), and 2) by morphological the between mite levels differences between bees. 2) relationship Physiological and differences between bees would likely grooming behavior; the and of worker bees influence the mite’s success once inside 3) frequency age the host. Mite behavior is difficult to performing allogrooming and dancing in observe without placing the subjects in tracheal mite-infested and uninfested colo- extremely unnatural conditions. However, nies; the host bee is quite easily observed within 4) the success of tracheal mite dispersal a glass-walled observation hive. Morpho- to selected high and low grooming lines logical differences have been proposed of bees to examine the role of grooming and examined in the past with no differ- behavior in limiting mite dispersal. ences demonstrated [10, 13].

Grooming behavior is a common 2. MATERIALS AND METHODS means of reducing ectoparasites in many organisms. Honey bees have been shown 2.1. Experiment I to detect and groom another parasitic mite, Varroa jacobsoni Oudemans from Four observation hives were used to exa- their bodies [3, 19]. A grooming dance mine the frequency of grooming behavior in bee colonies on a basis. Four by A. mellifera was first described by Hay- honey day-night observation each of one comb dak and further examined Milum hives, consisting [8, 9] by 20.3 x 42.5 cm and the second comb 11.4 x involves a [14]. Allogrooming normally 42.5 cm, were established at Texas A&M Uni- dancing bee and a bee that responds to the versity, College Station, TX on 15 April 1991 dance by grooming the dancer [8, 14]. and allowed 2 weeks to become established prior to observations. Ten days of observations 2.3. Experiment III were performed beginning on 29 April and every third day thereafter until 27 May 1991. Marked worker bees were monitored from Hives were observed six times over a 24-h per- to 19 d of to examine rela- iod at 0400, 0800, 1200, 1600, 2000 and 2400 emergence age age in with hours for the number of worker bees ted grooming behavior colonies varying solitary tracheal mite Three observation performing the grooming dance and the num- prevalence. ber of allogrooming pairs. Observations consis- hives were established with tracheal mite pre- valences of 0, 50 and 70 % of workers infested. ted of scanning one side of each hive in suc- Two of the hives were established on 15 cession, with five scans made per hive per time Sep- tember 1992 and 1 000 wor- interval (scan time = ca 1 min/hive). Scans on newly emerged kers from a source headed the same hive were separated by more than 5 single (colony by an were marked and min. Adult bee populations were estimated on open-mated queen) paint 26 May 1992 after sundown by averaging the 500 released into each hive on 1-2 October 1992. The third hive with 70 % mite bee counts of 20 randomly selected 5 x 5-cm preva- lence differed in that 2 000 squares, and multiplying that mean by the total newly emerged bees were marked and introduced and the bees number of squares present. Following the from each of four distinct observations a 30-bee sample was collected originated patrilines from each hive, dissected, and tracheal mite (500 bees/patriline). The four patriline colo- prevalence determined. nies were selected during a pilot study in which a pool of colonies was screened for grooming Data on the number of bees or allo- dancing behavior. Colonies were headed by unrelated over 24 h were ANOVA grooming analysed by queens that had been instrumentally insemi- for differences in time of when potential day nated with semen from a single unrelated drone the behaviors and if differences were occurred, [18]. Sealed brood combs were removed from found then means were separated using a each of the four patriline colonies and newly Fisher’s LSD test [25]. emerged bees were marked on the abdomen with a distinct paint color (Aerogloss, Pactra, Medina, OH) by patriline (500/line) and were 2.2. Experiment II 0-16 h old when placed into a single four- frame observation hive on 27 August 1992.

Thirty-two hives were used to examine Daily observations were conducted between grooming behavior in honey bees with varying 1000 and 1200 hours and consisted of five 3- tracheal mite prevalence. Observation hives min scans per hive over 19 d, with the num- were established in sets of four from overwin- ber of marked or unmarked bees dancing, allo- tered colonies and allowed 7-10 d to become grooming, or being groomed recorded. Scans established. Observations consisted of a scan of were separated by greater than 15 min to avoid each side of the hive (scan time = 3.02 ± 0.32 repeatedly counting individuals. Total number min X ± SD/hive) at 1000 and 1400 hours of marked bees surviving on days 5 and 19 over a 3-d period for the number of bees per- was determined by counting all marked bees in forming the grooming dance or allogrooming. the evening when foraging had ceased. Adult Following the observations a 20-bee sample bee populations were estimated on day 19 by was collected and later dissected to determine counting 20 randomly selected 5 x 5-cm mite prevalence. Thus, observations on groom- squares and multiplying that mean by the total ing behavior were conducted blind with regard number of squares present. to mite hives were prevalence. Thirty-two Tracheal mite levels were confirmed dis- observed from 1992 at Simon by May-August 30 bees from each hive, and external Fraser BC, Canada. secting University, Burnaby, examinations were made for the two external The relationships of mite prevalence were Acarapis species. The observer of grooming tested by correlation with the number of allo- behavior was unaware of the tracheal mite pre- grooming events and grooming dances [25]. valence in each colony and thus observations All counts were log transformed to normalize were made blind with respect to mite preva- the data. To determine if grooming behavior lence. The number of marked bees surviving in varied with time of day or over the 3-d period, the three colonies at days 5 and 19 was com- data were analysed by ANOVA with repeated pared using the LIFEREG procedure in SAS measures [25]. [24, 25 ]. The number of marked bees dancing, allogrooming or being groomed were compa- 3. RESULTS red between the first two hives over 19 d to determine if grooming behaviors differed by 3.1. ExperimentI mite prevalence using a Wilcoxon Rank Sum test The number of unmarked bees [25]. (bees behaviors varied time of of unknown age) dancing and allogrooming Grooming by within the hive. dances were compared between the two hives to eva- day Grooming luate background grooming levels (Wilcoxon increased significantly from midnight to Rank Sum test, [25]). 0800 hours compared to daytime levels In the colony containing four patrilines, data (figure 1, F = 2.70, 5 df, P = 0.02). In on dancing, allogrooming or being groomed contrast, allogrooming decreased signifi- were analysed by a Chi-square goodness of fit cantly at night (figure 1, F = 3.91, 5 df, null test by patriline against the hypothesis that P = 0.002). Adult bee populations were each patriline should perform 25 % of each estimated to be 3 150, 4 500, 2 080 and behavior. To examine potential differences in 2 200, and tracheal mite prevalence was age of task performance, the total number of and 66 % for hives res- occurrences was divided by the 19 d of obser- 80, 07, 05 1-4, vation to yield an expected value for each day. pectively. The observed values was grouped from days 1-5, 6-10, 11-15 and 17-19 (no observations made on day 16) and then compared to expec- 3.2. Experiment II ted values using a Chi-square goodness of fit test [25]. There were significant positive corre- lations between mite prevalence and allo- = 0.27, P < 0.002, n = 2.4. Experiment IV grooming (r 32), and the number of dances observed (r = 0.52, P < 0.001, n = There Workers from four selected lines of bees, 32) (figure 2). was a wide in the number of allo- which had exhibited varying levels of groom- range acts and dances ing behavior in experiment III, were introduced grooming (25-275) into mite-infested colonies to examine the role (10-75) observed. Additionally, mite pre- of grooming behavior in limiting mite disper- valence in the colonies varied from 10-80 %. sal to bees. The four colonies young represen- Acarapis dorsalis Morgenthaler was pre- ted: dance behavior 1) high grooming (HD), sent in 13 of 32 colonies with mite 2) behavior and pre- high allogrooming (HG), 3) valence within colonies from two colonies that were intermediate for both ranging 10-50 %. There were no differences in traits (I1 and I2). The term ’high’ refers to an increased likelihood of performing either dan- grooming behaviors based on time of cing or allogrooming in a given environment. observation, morning versus afternoon = 1 P = no Groups of 50 bees per patriline were marked (F 0.66, df, 0.42). Additionally, and placed together into each of the four parent significant differences in grooming behav- colonies and four unselected colonies. Brood ior were found over the 3 d of observa- combs from each of the four single patrilines tion (F = 1.69, 2 df, P = 0.15). were held overnight in an incubator at 35 ± 2 °C and the following day bees were paint mar- ked by patriline and were 0-16 h old when pla- III ced into colonies. All eight colonies were infes- 3.3. Experiment ted with tracheal mites. Marked bees were recovered from the colonies after 5 d, frozen, The three observation hives had tra- and later dissected to determine mite levels. cheal mite prevalences of 0, 50 and 70 %, Data were a analysed by Chi-square goodness and had A. dorsalis of 20, 10 of fit test against the null hypothesis that bees prevalences from the four lines should become equally and 20 %, respectively. No A. externus infested and to determine if recovery rates for Morgenthaler were found. No differences each patriline were equal [25]. in marked bee survival were detected be- tween colonies #1 and #2 (P > 0.05) nor dancing in the first 3 d (figure 3). Total between the survival of the four patrilines dancing by marked bees over the 19 d per- within colony #3. The total number of iod did not differ between the two hives unmarked bees (unknown ages) dancing or (figure 3, P > 0.05). The average number allogrooming did not differ between colo- of marked bees allogrooming and the like- nies #1 and #2 (P > 0.05) nor did unmar- lihood of being groomed increased signi- ked bees increase their dancing behavior ficantly in the colony with tracheal mites over days 1-3. The average number of (figure 3, P = 0.04). unmarked workers dancing in the two In #3 colonies was very consistent at 41.8 ± 8.7 colony containing the four workers and 41.4 ± 6.4 (mean ± SD) dances per patrilines, showed age-related day for hives #1 and #2, respectively. In biases for dancing, allogrooming and contrast, the frequency of dancing by mar- being groomed (figure 4, similar to pat- ked bees of known age varied with mite terns in colonies 1 and 2 in figure 3). Dan- levels and age (figure 3), with more dan- cing was most pronounced when workers cing occurring in the first 3 d of a bee’s were 1-5 d of age, with 56 % of all dan- life in the tracheal mite-infested colony. cing occurring during the first 5 d (expec- The marked bees in the colony without ted 26 %). Workers 5-15 d of age were tracheal mites did not exhibit increased most likely to allogroom or be groomed. Greater than 90 % of allogrooming and hood of being groomed did not differ be- 76 % of workers being groomed occurred tween the four patrilines (figure 4). The when workers were between 5 and 15 d patriline named HD (high dancing) of age (expected value of 53 % for days accounted for 48 % of all dances (X2 = 5-15). 25.06, 3 df, P < 0.001) while the HG Dancing and allogrooming varied (high grooming) line accounted for 64 % significantly by patriline but the likeli- of the total allogrooming (X2 = 80.1, 3 df, P < 0.001, figure 4). The likelihood exposure in eight mite-infested colonies of being groomed did not differ signifi- (table I). The null hypothesis was that each cantly between the four patrilines (X2 = of the four lines of bees would acquire 25 % 2.09, 3 df, P > 0.05). of the available mites. The high dancing patriline, HD, acquired significantly fewer than and the allo- 3.4. Experiment IV mites expected high grooming, HG, acquired more mites than Marked bees from the four patrilines expected (figure 5). Recovery rates of mar- exhibited varying mite levels after 5 d of ked bees were not significantly different between patrilines. Mite prevalence in the groomed, thus reducing successful mite eight host colonies ranged from 40 to 100 transfer into the tracheal tubes. % (table I). The number of mites that transferred in the eight colonies ranged Day-night variation in both allogroom- and was observed. The increa- from 99 to 363. ing dancing sed dancing at night can be associated with increased tracheal mite transfers at night as colonies in the current studies were 4. DISCUSSION [20] mite infested. The reduction in allogroo- ming at night may be in association with The performance of allogrooming and the overall decrease in worker activity that the dance worker bees grooming by honey occurs at night [20]. Allogrooming should varied with time of wor- significantly day, increase with increased dancing at night ker and We that age genotype. propose but may not if allogroomers at night are behavior in bees is both grooming honey in a less active state [11]. Monitoring groo- elicited by, and has an impact upon, tra- ming behavior from 32 colonies revealed cheal mites as to bees. they disperse young a positive relationship with mite preva- Our findings, that young bees 1-3 d of lence within these colonies. Both dancing age increased their dancing behavior in and allogrooming increased with increa- the of tracheal mites, is presence strong sing mite levels. No differences were indirect evidence of a factor which induces detected in the time of observation during the dance. grooming Additionally, young the day, 1 000 versus 1 400 hours, nor bet- worker bees that demonstrated dan- high ween the three consecutive days of obser- cing behavior also acquired the lowest vation. number of tracheal mites compared to three other patrilines. We postulate that The age and genotype of a worker these bees had a lower dance threshold influences the timing and likelihood of which was elicited by mite presence on both dancing and allogrooming. Allo- their bodies and that they danced and auto- grooming had previously been shown to be in part genetically determined [5, 12], and rored by the increased dancing behavior is now shown to be age specific. Workers on these days. between 5 and 15 d old are most likely to Workers from different patrilines exhi- serve as allogroomers. The present stu- bited differences in their propensity to dies indicate that workers of a specific acquire tracheal mites in a 5-d bioassay. genotype are more likely to dance than Workers that had a low dance threshold other genotypes. Additionally, dancing acquired the fewest tracheal mites, while occurs more in young bees frequently workers < 5 d old if the hive is infested with tra- exhibiting high allogrooming behavior had acquired the highest num- cheal mites, as mites disperse preferen- ber of mites. Workers that were interme- tially to bees < 5 d of age [7, 13]. The four diate in their grooming behavior were not patrilines were all groomed equally. Thus, significantly different from expected no evidence for preferential allogrooming values. The high allogrooming line was was observed between the four patrilines. most 34 % of the This is in contrast to Frumhoff and Baker susceptible, acquiring available mites. The line [5] who did report preferential allogroo- high dancing the fewest mites which we ming toward full sisters when only two acquired pro- pose was due to their ability to detect the patrilines were present. In the current stu- of mites on their and dies using four patrilines, increased dan- presence body remove these mites with the autogrooming cing by patriline HD did not result in an that occurs the dance. All increased probability of being groomed during dancing bees use their second of to (figure 3). pair legs groom over their thorax as they dance. The raking The hives with and without tracheal action of the legs could remove mites that mites (0 and 50 % prevalence) clearly are attempting to disperse to these bees demonstrated an association between the behavior with level and of dance behavior and Grooming changes timing in tracheal mite levels within tracheal mites. Workers in the tracheal changes honey bee colonies. We found increasing mite-infested danced more in the colony levels of both and as first 3 d introduction dancing allogrooming following (similar tracheal mite levels increased. Most stri- results were obtained in colony #3 with king was the increased dancing observed four In contrast, when workers patrilines). in bees 1-3 d of age in mite-infested colo- from the same source were intro- colony nies that corresponds with the movement duced into a tracheal mite-free colony they of mites onto young bees. No increased danced at a lower level over the first 3 d dancing was found in young bees in a than their sisters in a with tracheal colony free of tracheal mites. We mites. The in both colony propose background dancing that dancing behavior is elicited by mites colonies by bees of unknown age was very on the body of workers and that the auto- consistent over the 19 41 d, averaging associated with the dance dances/d in both hives. the increased grooming Thus, reduces the number of mites that succeed 1-3 can be linked to dancing during days in transferring. Thus, the grooming dance and mite levels and not to to age day day worker bees is as a within the hive. The increased by honey proposed changes resistance mechanism to the bee in the tracheal mite-infested honey dancing tracheal mite. colony on days 1-3 is strong indirect evi- dence that mite movement onto these wor- Since submitting this manuscript the kers is eliciting the dance behavior. Tra- authors have become aware of research in cheal mites disperse to young bees and Baton Rouge, LA (USDA-ARS) with their preference for 1-3-d-old bees is mir- Buckfast bees that demonstrates that auto- grooming is linked to tracheal mite resis- comparée à la colonie non infestée, les tance, which corroborates our findings. danses de toilettage ont augmenté signi- ficativement chez les abeilles âgées de un à cinq jours puis ont diminué (figure 3), ACKNOWLEDGEMENTS indiquant que le déplacement de l’acarien sur les jeunes abeilles peut déclencher la We wish to thank Pete Teel of Texas A&M danse de toilettage. Dans la colonie infes- University and W.T. Wilson of USDA-ARS tée à 70 %, des groupes d’abeilles récem- for advice and support, Yvonne Higo, Heather ment écloses de quatre lignées paternelles and Marainna for technical assis- Higo Lightly uniques ont été marquées et leur compor- tance, and M.L. Winston, H. and three Higo tement de a été observé durant reviewers for comments toilettage anonymous helpful 19 Les abeilles des dif- on the manuscript. Research was supported by jours. ont présenté en the USDA-ARS and a National Science and férences significatives, fonction de leur Engineering Research Council Grant to MLW. lignée paternelle, dans leur propension à danser ou à s’allotoiletter. Une lignée paternelle a rendu compte de 64 % de tout Résumé - Comportement de toilettage l’allotoilettage, une autre de 48 % de toutes par Apis mellifera L. en présence d’Aca- les danses, tandis que les deux autres occu- rapis woodi (Rennie) (Acari, Tarsone- paient une position intermédiaire pour ces midae). On a étudié le rôle du comporte- deux caractères (figure 4). L’allotoilettage ment de toilettage chez l’abeille Apis est lié à l’âge, la majeure partie (90 %) mellifera dans la limitation de l’infesta- ayant lieu quand les abeilles ont entre 5 tion par l’acarien des trachées Acarapis et 15 jours. Dans une quatrième expé- woodi ainsi que le déclenchement de ce rience, des groupes d’abeilles récemment comportement par la présence du para- écloses issues de colonies de lignée pater- site. On a relevé le nombre d’abeilles qui nelle unique ont été marqués et placés soit exécutaient une danse de toilettage durant cinq jours dans des colonies infes- (= autotoilettage), soit étaient engagées tées, réparties en quatre colonies paren- dans une activité de toilettage mutuel tales et quatre colonies d’observation (= allotoilettage), qui consiste à toiletter (tableau I). Sur les quatre colonies de un congénère ou à être toiletté par lui. La lignée paternelle simple, l’une correspon- première expérience a porté sur quatre dait à des abeilles ayant un comportement ruches d’observation suivies 24 h sur 24 d’allotoilettage élevé, une seconde à des durant dix jours. Le comportement des abeilles ayant un comportement de danse abeilles a été mis en relation avec le nyc- élevé, les deux autres colonies ayant une thémère : la danse augmentait significa- position intermédiaire concernant ces deux tivement la nuit alors que l’allotoilettage caractères. La dissection des abeilles mar- diminuait (figure 1). Dans une seconde quées a montré des différences cohérentes expérience, 32 ruches d’observation ont entre les quatre lignées parentales dans le été suivies durant l’été 1992 à Burnaby, niveau d’infestation (tableau I). La lignée Bristish Columbia, Canada. Une corréla- la plus sujette à l’allotoilettage était la tion positive a été trouvée entre le pour- plus sensible à l’infestation par l’acarien centage d’abeilles infestées et le nombre (figure 5) ; la lignée qui dansait le plus d’auto- ou d’allotoilettages (figure 2). étant la moins sensible. Ceci peut être dû Dans une troisième expérience de jeunes à la capacité de ces abeilles à détecter la abeilles marquées ont été introduites dans présence de l’acarien sur leur corps et à trois ruches ayant un taux d’infestation de l’éliminer par l’autotoilettage qui a lieu 0, 50 et 70 %. Leur comportement de toi- au cours de la danse. Ces expériences lettage a été observé chaque jour durant montrent que les abeilles qui ont une 19 jours. Dans la colonie infestée à 50 % prédisposition génétique à effectuer des danses de toilettage sont susceptibles de tet. Die Arbeiterinnen der unterschiedli- détecter les acariens présents sur leurs chen Vaterlinien zeigten eine unter- corps et à s’en débarrasser. C’est la pre- schiedliche Neigung zu Putztänzen oder mière fois que l’on propose un mécanisme zum Fremdputzen. 64 % des Fremdput- de résistance de l’abeille mellifère à l’aca- zens entfiel auf eine der Patrilinien, 48 % rien des trachées. © Inra/DIB/AGIB/ aller Putztänze auf eine andere, zwei der Elsevier, Paris Linien nahmen in beiden Verhaltenswei- sen eine mittlere Position ein (Abb. 4). Apis mellifera / Acarapis woodi / com- Das Fremdputzen war vom Alter abhän- 90 % durch portement toilettage / résistance au gig, erfolgte 5-15 Tage alte parasite Bienen. In einem vierten Experiment wur- den Gruppen frischgeschlüpfter Bienen aus vier jeweils nur eine Vaterlinie enthal- Zusammenfassung - Über das Putzver- tenden Völkern markiert und fünf Tage halten bei Apis mellifera L. in Gegen- lang in vier milbenbefallene Völker ein- Die vier wart von Acarapis woodi (Rennie). Die gesetzt (Tabelle I). Spendervöl- Auslösung des Putzverhaltens der Honig- ker standen für Bienen mit 1) ausgepräg- biene Apis mellifera L durch den Bie- tem Fremdputzen, 2) ausgeprägtem nenparasiten Acarapis woodi und des- Putztanzen und 3) mittelmäßiger, Aus- beider Verhaltensweisen Völ- sen Rolle in der Begrenzung des Befalls prägung (2 wurde untersucht. An zehn Tagen wurde ker). Die Präparation der markierten Arbei- jeweils während 24 Stunden die Anzahl terinnen zeigte hiermit übereinstimmende Unterschiede des Befallsniveaus von Arbeiterinnen in Beobachtungsstök- (Tabelle I), ken erfasst, die entweder einen Putztanz wobei die Vaterlinie mit hoher Neigung ausführten oder andere putzten (Fremd- zum Fremdputzen den höchsten Befall putzen). In der Nacht stieg die Tanzhäu- aufwies (Abb. 5). Die Linie mit ausge- figkeit signifikant an, während die prägtem Putztanzverhalten war gegenü- Putzhäufigkeit im Vergleich zur Aktivität ber einem Befall durch die Milben am während des Tages abnahm (Abb. 1). In wenigsten empfänglich. Dies könnte einem zweiten Experiment wurden wäh- möglicherweise auf eine geringere rend des Sommers 1992 32 Beobachtungs- Wahrnehmungsschwelle gegenüber der völker in Burnaby, Britisch Columbien, Anwesenheit von Milben auf dem Körper Canada untersucht. Hierbei zeigte sich hindeuten, mit der Folge eines verstärk- ein positiver Zusammenhang (Abb. 2) zwi- ten Tanzverhaltens und damit verbunden schen dem Anteil infizierter Arbeiterin- einem erhöhten Selbstputzen. Wir neh- nen in diesen Völkern und der Anzahl von men an, daß dies die Anzahl von Milben Putztänzen und von Fremdputzen. In verminderte, die auf diese Bienen über- einem dritten Experiment wurden junge tragen wurden. Die Experimente zeigen, markierte Bienen in drei Völker mit 0, 50 daß Bienen mit einer genetischen Neigung und 70 % Befallshäufigkeit eingesetzt und zur Ausführung von Putztänzen wahr- täglich 19 Stunden lang ihr Putzverhalten scheinlich die Milben auf ihrem Körper beobachtet. Die Putztänze waren bei 1-5 wahrnehmen und abputzen. Dies ist der Tage alten Bienen signifikant häufiger als erste vorgeschlagene Mechanismus für die in dem nicht infizierten Volk (Abb. 3). Resistenz von Honigbienen gegenüber Tra- Dies zeigt an, daß das Überwechseln der cheenmilben. © Inra/ DIB/AGIB/Elsevier, Milben auf junge Bienen den Putztanz Paris auslösen kann. In dem zu 70 % befallenen Volk wurden Gruppen frischgeschlüpfter Apis mellifera / Acarapis woodi / Bienen aus vier Patrilinien markiert und Putzverhalten / Verhalten / ihr Putzverhalten 19 Tage lang beobach- Arbeitsteilung REFERENCES [14] Milum V.G., Honey bee communication, Am. Bee J. 95 (1955) 97-104. [1] Bailey L., Ball B.V., Honey Bee Pathology, [15] Morgenthaler O., An acarine disease experi- Second edition, Academic Press, London, 1991. mental apiary in the Bernese Lake-District and some results obtained Bee World 12 [2] Bozic J., Valentincic T., Quantitative analysis there, of social grooming behavior of the honey bee (1931) 8-10. Apis mellifera carnica, Apidologie 26 (1995) [16] Morse R.A., Honey Bee Pests, Predators and 141-147. diseases, Cornell Univ. Press, Ithaca, NY, 1978, [3] Büchler R., Drescher W., Tornier I., Groo- 419 p. ming behavior of Apis cerana, Apis mellifera [17] Page R.E., Gary N.E., Genotypic variation in and Apis dorsata and its effect on the parasitic susceptibility of honey bees (Apis mellifera) to mites Varroa jacobsoni and Tropilaelaps cla- infestation by tracheal mites (Acarapis woodi), reae, Exp. Appl. Acarol. 16 (1992) 313-319. Exp. Appl. Acarol. 8 (1989) 275-283. [4] Delfinado-Baker M., Baker E.W., Notes on [18] Pankiw T., Winston M.L., Slessor K.N., Varia- honey bee mites of the genus Acarapis Hirst. tion in worker to honey bee (Apis mellifera L.) (Acari: Tarsonemidae), Int. J. Acarol. 8 (4) queen mandibular pheromone (Hymenoptera: (1982) 211-226. Apidae), J. Insect. Behav. 7 (1994) 1-15. [5] Frumhoff P.C., Baker J., A genetic component [19] Peng Y., The resistance mechanism of the to division of labor within honey bee colonies, Asian honey bee (Apis cerana) to the mite Var- Nature 333 (1988) 358-361. roa jacobsoni, in: Needham et al. (Eds.), Afri- [6] Gary N.E., Page R.E., Phenotypic variation in canized Honey Bees and Bee Mites, Ellis Hor- susceptibility of honey bees, Apis mellifera, to wood Ltd., Chichester, 1988, pp. 426-429. infestation by tracheal mites Acarapis woodi, [20] Pettis J.S., Wilson W.T., Eischen F.A., Noc- Exp. Appl. Acarol. 3 (1987) 291-305. turnal dispersal by female Acarapis woodi in [7] Gary N.E., Page R.E., Lorenzen K., Effects of honey bee (Apis mellifera) colonies, Exp. Appl. age of worker honey bees (Apis mellifera L.) on Acarol. 15 (1992) 99-108. tracheal mite (Acarapis woodi Rennie) infes- [21] Phelan L.P., Smith A.W., Needham G.R., tation, Exp. Appl. Acarol. 7 (1989) 153-160. Mediation of host selection by cuticular hydro- [8] Haydak M.H., Some new observations of the carbons in the honeybee tracheal mite Acarapis bee life, Cesky Vcelar 63 (1929) 133-135 (in woodi (Rennie), J. Chem. Ecol. 17 (2) (1991) Czech). 463-473. [9] Haydak M.H., The language of the honeybees, [22] Rennie J., White P.B., Harvey E.J., Isle of Am. Bee J. 85 (1945) 316-317. Wight Disease in hive bees, Trans. R. Soc. 52 737-779. [10] Hirschfelder H., Sachs H., Recent research on Edingburg (1921) acarine disease, Bee World 33(12) (1952) [23] Royce L.A., Krantz G.W., Ibay L.A., Burgett 201-209. D.M., Some observations on the biology of [11] Kaiser W., Busy bees need rest, too, J. Comp. Acarapis woodi and Acarapis dorsalis in Ore- Physiol. 163 (1988) 565-584. gon, in: Needham et al. (Eds.), Africanized Honey Bees and Bee Mites, Ellis Horwood Kolmes wor- [12] S.A., Grooming specialist among Ltd., Chichester, 1988, 498-505. ker honey bees in Apis mellifera, Anim. Behav. pp. 37 (1989) 1048-1049. [24] SAS, SAS Stat User’s Guide version 6.03, SAS Institute, 1988. [13] Lee D.C., The susceptibility of honey bees of Cary, NC, different ages to infestation by Acarapis woodi [25] Zar J.H., Biostatistical Analysis, Prentice Hall, (Rennie), J. Insect. Pathol. 5 (1963) 11-15. Inc., Englewood Cliffs, NJ, 1984.