HAL Id: hal-00890983https://hal.archives-ouvertes.fr/hal-00890983

Submitted on 1 Jan 1992

HAL is a multi-disciplinary open accessarchive for the deposit and dissemination of sci-entific research documents, whether they are pub-lished or not. The documents may come fromteaching and research institutions in France orabroad, or from public or private research centers.

L’archive ouverte pluridisciplinaire HAL, estdestinée au dépôt et à la diffusion de documentsscientifiques de niveau recherche, publiés ou non,émanant des établissements d’enseignement et derecherche français ou étrangers, des laboratoirespublics ou privés.

Active defense against Varroa mites in a Carniolanstrain of honeybee (Apis mellifera carnica Pollmann)

F Ruttner, H Hänel

To cite this version:F Ruttner, H Hänel. Active defense against Varroa mites in a Carniolan strain of honeybee (Apismellifera carnica Pollmann). Apidologie, Springer Verlag, 1992, 23 (2), pp.173-187. �hal-00890983�

Original article

Active defense against Varroa mites in a Carniolanstrain of honeybee (Apis mellifera carnica Pollmann)

F Ruttner H Hänel

1 Universität Frankfurt, Institut für Bienenkunde (Polytechnische Gesellschaft),Fachbereich Biologie, Karl von Frisch-Weg 2, D-W-370 Oberursel;2 Hoechst AG, PO Box 800320, D-W-6330 Frankfurt/M, Germany

(Received 1 December 1992; accepted 25 February 1992)

Summary — Out of 700 hives (A m carnica) owned by a commercial beekeeper 12 colonies wereselected which showed slower Varroa population growth, high mite mortality and good overwinteringwithout treatment. Thirty to 50 percent of the dead mites had damaged legs and (rarely) cuticle ofthe idiosoma. The damage rate increased significantly with mite mortality level. By investigatingfreshly killed mites it was concluded that the damage was caused by the mandibles of worker bees.SEM photos of the mandibles indicate that they act like sharp scissors and that no substantial differ-ences exist between A mellifera and A cerana.

Apis mallifera carnica / Varroa resistance / amputation / mandible / defensive behavior

INTRODUCTION

The Asiatic honeybee Apis cerana F hasa number of defense mechanisms to keepthe population of the ectoparasitic mite

Varroa jacobsoni Oud within restricted lim-its and to ward off damage. One of thesemechanisms as described by Peng et al(1987), is the active removal of adultmites from the bodies of worker bees andbrood cells. This process involves self-

cleaning behavior after showing signs ofirritation, grooming dance, nestmate

cleaning and group cleaning behavior.Each of the different types of cleaning be-havior frequently resulted in the removaland eventual killing of the mite. 73.8% ofthe mites that dropped onto the hive bot-tom showed injuries to the body.

While experimentally introduced miteswere removed from A cerana in a very highpercentage, colonies of European origin (Amellifera of unspecified subspecies)showed only vestiges of this behavior dueto their "limited capability to recognize themite as a pest" (Peng et al, 1987), and totheir mandibles supposedly not suited tobite as strongly as those of the Asian hon-eybee. The removal rate of A mellifera wasobserved to be 0.3% (Peng et al, 1987). InBrazil, however, Moretto et al (1991) founda very similar behavior in Africanized hy-brids of A mellifera with an average remov-al rate of 38.5% (range 10-70%). Killed ormutilated mites were found only rarely. Thesame effect, but to a much reduced de-

gree, was found with simultaneously testedcolonies of A m ligustica with a mean re-moval rate of 5.75%.

The high variance of this behavior

found in these experiments and the experi-ence that other A cerana resistance mech-anisms such as infertility of the Varroa inworker brood also exist in European racesof A mellifera (Ruttner et al, 1984) prompt-ed a search in other Apis mellifera subspe-cies for a higher removal incidence, as al-ready suggested by Peng et al (1987).Proving that A m carnica has the behavior-al, anatomical and physiological capacityof removing and killing living Varroa mitesfrom the colony was the aim of this investi-gation.

MATERIALS AND METHODS

During the summer of 1990, Alois Wallner, acommercial beekeeper in Austria, selected adozen colonies from his apiary of 700 colonieswhich showed a distinctly slower developmentof the Varroa population during the season

(Wallner, 1990). Whether the queens were relat-ed to each other is unknown. They were matedin the open. The selection process, ingeniouslysimple for a field test, was performed by thebeekeeper during the summer of 1990: first, in

spring and early summer, checks of the speedof the mite population increase in all colonieswere made by simply breaking open sealeddrone brood, in order to count infested cells.The majority of colonies was thereby excludedfrom the experiment. In the remaining colonies20-50 capped worker brood cells were investi-gated bi-weekly for the presence of Varroamites in late summer. Finally, less than 2% of allthe colonies remained for further observation.

In contrast to the remainder (which was treat-ed with formic acid) the selected colonies over-wintered excellently without any chemotherapy.Further, the beekeeper noticed a high mite mor-tality and mutilation in a varying percentage ofthe mites. From September 1, 1990, to April 30,1991, all the dead mites were collected dailyfrom the bottom board (covered with a sheet ofwhite paper and protected by a nylon net) ex-cept for 2 periods with very low temperatureswhen the mites were collected weekly. From 5randomly chosen colonies of this group all deadmites (n = 6189) were carefully examined under

a stereomicroscope for injuries and abnormali-ties. The data were processed by an analysis ofvariance, using the SPSS PC package.

Some of the collected and mutilated miteswere examined by using a scanning EM (Leitz1600T, 20 KV). To obtain freshly amputatedmites, infested, freshly capped brood cells (con-taining prepupae or young pupae) of an "effi-cient mite remover colony" were opened andabout a dozen mites set free; then the colonywas closed again. After 10 min, about 5-10mites were found on the bottom board, some ofwhich were still showing movements in spite offresh amputation of one or several legs. Theseamputated mites were easily identified as origi-nating from a capped brood cell by their swollenidiosoma (in contrast to the flat abdomen of a

phoretic female). This method has the advantageof using mites from the same colony and therebyexcluding reactions caused by foreign odors.

To obtain SEM pictures of mites with soft tis-sue, individuals still showing signs of life werefixed in 70% ethanol where they remained for upto 4 weeks. Afterwards the specimens weretransferred for 12 h into 30% amyl acetate inethanol. Passing through 70% amyl acetate inethanol for 12 h, they were then dehydrated inpure amyl acetate for 12 h with 3 changes. Themite specimens impregnated with amyl acetatewere rinsed with liquid CO2 (3 changes every 2h) within a critical-point-dryer (Polaron, Water-ford. UK) and then carefully approached close tothe critical point by slowly rising temperature. Af-ter reaching the critical point (disappearance ofthe liquid phase) the chamber was slowlybrought down to normal pressure level. Thedried specimens were glued to sample plates bymeans of Leit C® (Leitz, Wetzlar, Germany) andinvestigated using the same SEM procedure aswith other dry material.

RESULTS

Mutilated mites

Kind of damage

The mutilation of the mites varied from lossof a single apical segment to the complete

loss of all 8 legs with the gnathosoma(figs 1, 2). Damage to the 2 front legs wasthe most frequent observation. Otherwiseall thinkable combinations of damagedlegs occurred, for instance unilateral lossof several legs or asymmetrical loss of sin-gle legs on both sides, and no patternseemed to exist. Damage to the cuticle ofthe idiosoma was relatively rare (1-2%).The stumps of mutilated legs had the ap-pearance of an empty tube (even when thespecimens had been collected at daily in-tervals), the chitin of the walls looked brit-tle, as if it was broken.

Seasonal and individual frequency

The total number of dead mites and muti-lated individuals per month from Septem-ber to April is given in table I. There seemsto be a strong seasonal influence, although

this was not statistically significant (F2.40,prob > F 0.0478) probably due to the smallnumber of colonies. The same was true forthe differences between colonies (F 2.68,prob > F 0.0527). There was, however, ahighly significant correlation between thenumber of dead mites and the number of

mutilated mites (F 133.40, prob > F

0.0001 ).As the colonies were part of the private

selection program of the commercial bee-

keeper and were supposed to remain un-treated and unmanipulated, there was nopossibility of assessing their total infesta-tion rate. Excellent control data from the

same race of bees and from similar climat-ic conditions in the same province of Aus-tria and from the same year were pub-lished by Moosbeckhofer (1991). In a

group of 25 experimental unselected colo-nies, he observed an average of 72.8 dead

mites during the period Nov 8, 1990-March 13, 1991. The mean number ofmites per colony, assessed by Apistan®treatment in March-April 1991, was 459.8.From this an average Varroa winter mortal-

ity of 13.6% was calculated. The meannumber of dead mites found in our 5 se-lected colonies for a comparable period(Nov 1-March 15, 1991) was 455.4 (range257-726), ie 6-fold the amount of that ob-served in the aforementioned publication.According to Moosbeckhofer’s figures thiswould correspond to 13.6% of the mitespresent in the colonies, which would implya mean spring infestation of 3 350 mites,which is incompatible with survival of thecolony. Hence a much higher mite mortali-ty has to be concluded for our experimen-tal colonies. Further experiments with se-

lected colonies are needed to confirm this

assumption.

Cleaning and defense behavior

So far, no extensive behavioral observa-tions have been made on the selected col-onies. However, short-term observationson a colony in a glass-walled observationhive or while managing a colony after re-lease of mites revealed the same behavior-al pattern as that described by Peng et al(1987) and Moretto et al (1991), includinggeneral unrest, cleaning movements andgrooming dance of infested bees and nest-mate cleaning. No clear direct observationof the removal of a mite from the body of a

bee by self-cleaning or nestmate cleaninghas been observed.

Freshly amputated mites

From the condition of the mutilated mitesthe cause of the damage cannot be in-ferred with certainty. Some believe that liv-ing mites were killed by the bees; othersregard this as impossible with A melliferaand hold ants or other arthropods respon-sible, or contend that these are old dry andbrittle specimens which broke apart. Sincedirect observations are not yet available(given the relative scarcity of this eventover a restricted period of time) we decid-ed to resort to investigating freshly ampu-tated and still living mites. It was not diffi-cult to obtain specimens of this kind.

Shortly after releasing a number of femaleVarroa from freshly sealed brood cells thetypical unrest and cleaning behavior wasobserved on the respective comb. A fewminutes later some amputated mites withlegs still moving were collected from thebottom board.

The inspection of the amputated miteswas a complete surprise. Instead of findingsigns of gnawing, chewing or tearingapart, the muscular proximal joints of thelegs showed a clear cut, even surface, ofthe chitinuous wall as well as of the mus-cles - as if severed by a sharp knife or apair of scissors (figs 3-5). Sometimes athin slice of muscle tissue was cut off, butstill remaining in place (fig 3). In figure 4,leg 2 was cut off on 2 different planes, leg3 was cut by two-thirds; the rest was brok-en. In figure 5, the whole front part was re-moved, approaching the stage shown in

figure 2. Only if the leg was damaged atone of the end joints, with remaining slen-der muscles, was the surface of the injuryrugged and uneven (fig 6). Seeing theseeffects of the action of honeybee mandi-

bles, a closer study of the mechanics ofthe biting organs seemed desirable.

The cutting instruments:comparative anatomy of the mandiblesand their muscles

We compared bees of the Varroa tolerantstrain belonging to the alpine population ofA m carnica with bees of A cerana ceranafrom northern China. Although belongingto one of the largest subspecies, A c cera-na is distinctly smaller than A m carnica(table II).

The head capsule of A c cerana wasslightly smaller in both dimensions thanthat of A m carnica (fig 7). The mandibleswere distinctly longer in A m carnica thanin A c cerana (table II, fig 8), but withshovels of about equal width. The 2 majormuscles which move the mandibles wereshorter in A c cerana. but probably withmore fibrils; only the long slender medialmuscle was clearly larger in A m carnica

(fig 7).The working rim of the mandible was a

straight sharp edge in both species, with amean length of 0.51 mm for A c ceranaand 0.59 mm for A m carnica (5 workermandibles were measured in each spe-

cies). At higher magnification (figs 9, 10)the edge compared well with that of a

sharp knife. In figure 8 the diameter of aVarroa hind leg (basal joint) is shown in or-der to demonstrate that this is a powerfultool in both species to work a Varroa sizebody. A characteristic specific to the carni-ca mandible was a double edge in the ba-sal half of the rim (figs 8, 9). Small slices oftissue adhering to the surface of the cutmuscle (fig 3) were probably due to the ef-fect of this double edge. A sparse row ofbristles was positioned at the inner side ofthe edge (fig 9), whereas the outer surfacewas completely bare at a narrow terminalstripe, with rectangularly bent bristles pro-jecting downward over it (fig 10).

DISCUSSION

Quoting an observation of Sakagami(1960), a defense reaction by means of

damaging the Varroa mite is believed tobe a rare event in A mellifera because ofits weak mandibular power. However, inthe original text (p 183) this statementdoes not read as explicitly as it is frequent-ly given: "When C and M are forced to asingle duel, usually M wins the victory be-cause of its stronger muscular power. Butthe powerful mandibles of C compensateto some extent her weakness" (see alsoSakagami, 1959). Our present anatomicalcomparison of 2 northern subspecies of

both species (figs 3, 8) did not support thisstatement. The mandibles were found tobe shorter in A c cerana, and it can be

questioned whether the muscular power issubstantially greater in this species. Thestructure of the mandibular edge did notseem less suited for cutting in A melliferathan in A cerana (figs 8-10).

The strongest argument for the capacityof A mellifera workers to amputate the legsof Varroa mites was the evidence of fresh-

ly damaged mites. The mites were collect-ed from the bottom board, still alive, only10 minutes after they had been releasedfrom capped brood cells. The mutilatedmites collected alive died within a few min-

utes. Then the white tissue visible at the

opening of the cut leg soon darkened andretracted inside the body giving an aspectsimilar to that in figure 1. Although themites were not marked, their identity is un-disputed because of their state of "gravidi-ty" shown by the inflated idiosoma and the

dehiscence of the ventral cuticular plates(in contrast to the flat phoretic individuals).Although not yet observed directly thereseems to be no reasonable other explana-tion for the damaged mites than as the re-sult of an aggressive behavior of the bees.

The inspection of the amputated legsgave a further clue. The surprisingly even,smooth section plane favors the idea ofcooperation between the 2 edges of themandibles like the 2 blades of a pair ofscissors. The enlarged SEM photos of cutlegs are remarkably similar to a neatly cutpiece of ham (figs 3-5). The legs with theirhard, chitinuous surface and the soft tur-gescent muscular content provide an idealsubstratum for cutting even if the materialis tough and hard to tear. The assumptionof a scissor-like action is supported by theobservation of an amputation clearly madeby cuts effected at 2 different angles (fig4), and the presence of tiny slices of mus-cular tissue by the additional mandibularedge (figs 3, 8, 9). It should be kept in

mind that the mandibles of Apoidea arevery well suited for cutting, eg neatly cutcell cups of queens and drones, exactlycut pieces of rose leaves in Megachile.Considering the problem of the capacity ofamputating mite legs, be it by A cerana orby A mellifera, the comparative size ofmites and honey bees is to be kept in mind(fig 8): head width of A mellifera workerbee (northern race) is about 4.1 mm com-pared to 1.2 mm of a female Varroa trans-versal body diameter. The length of the

cutting edge of a worker’s mandible (0.59mm) is more than 5 times the average dia-meter of a Varroa leg (0.06-0.13 mm).Considering this relatively huge tool, smalldifferences in size are not likely to impairefficacy in dealing with a tiny object.

Although there was no major differencein the structure and functioning of the in-struments of biting there certainly exists adifference in the frequency of the actual ef-

fect. Amputated dead mites were found inmany A m carnica colonies, but only at avery low frequency. Higher frequencies(25-50%) were found only in 12 out of 700colonies. The complete behavioral reper-tory (self-grooming, grooming dance,nestmate grooming) as described by Penget al (1987) in A cerana was to be ob-served also in these colonies, but certainlyat a lower frequency. In many cases a

worker bee with a mite on top of the thoraxwas observed for quite a while without thebee in question or the nestmate paying anyattention. The differences between A cera-na and A mellifera as far as the parasiticmite Varroa is concerned are not found intheir morphological and anatomical charac-teristics, but in their behavioral inventory.They are, however, only at a quantitativelevel, that also demonstrates the close rela-tionship between the 2 species in this re-

spect. It is to be expected that the activedefense behavior against Varroa is fre-

quently found in the European subspeciesof A mellifera to a varying degree. It seems

to be especially highly developed in the "Af-ricanized" bees of South America (Morettoet al, 1991).

The active defense reaction againstVarroa mites has evidently nothing to dowith general hygienic behavior, eg, nest

cleaning. A cerana is definitely "disorderly"as far as removing rubbish from the hivebottom is concerned and therefore verysusceptible to wax moths, but it is attentivetowards Varroa; A mellifera carnica, on theother hand, always has a clean bottomboard but it ignores the mites infesting thecolony in most of its colonies. Therefore,these 2 characteristics should be clearlydifferentiated in the terminology.

The impact of this defense behavior onthe population dynamics of Varroa is con-siderable. This is evidenced by our datafrom one colony. Colony No 13 had highmite infestation due to reinfection from

neighboring colonies and also a high am-putation rate (over 40%). The mite mortali-ty from September 1 st to October 10th,1991, assessed by daily counts, amountedto a total of 1 640 individuals. The remain-

ing mite population in the colony, estimat-ed from a sample of 304 bees, was about1 370 mites, that is 45.6% of the total (=dead + remaining mites). Since mite re-

moval continues during autumn and winter,although at a lower rate, while no new

mites can be produced, this colony has agood chance of survival without help. As ahigh degree of heredity has been indicatedby preliminary breeding assays, the char-acter of active Varroa defense will prob-ably play an important role in future pro-grams of selecting a Varroa tolerant

honeybee.

ACKNOWLEDGMENTS

We wish to thank M Keil (Hoechst AG) for takingthe SEM pictures and A Mohr for the morpho-metric data, preparation of mandibles and for fig-ure 3. C Rau and E Hüttinger assisted with pho-totechnical skills and C Boigenzahn withstatistical advice. H Pechhacker provided labor-atory space and apicultural assistance. Themain information, however, we owe to beekeep-er Alois Wallner who was the first to observe theeffects of the active Varroa defense behavior ofA m carnica honeybees and to select a strainwith high manifestation of this characteristic.

Résumé — Défense active contre l’aca-rien Varroa jacobsoni chez une souched’abeilles carnioliennes. Un apiculteurprofessionnel d’Autriche a sélectionné 12colonies parmi ses 700 colonies d’abeillescarnioliennes (Apis mellifera carnica) pourle faible accroissement de leur populationde varroas au cours de la saison apicole.Contrairement à toutes les autres coloniesde la région qui avaient besoin de traite-

ments acaricides réguliers pour survivre,ces colonies ont hiverné sans aucun traite-ment et se sont développées l’année sui-vante normalement en fournissant une ré-colte de miel correspondant à la moyenne.La mortalité des acariens dans les colo-nies sélectionnées, suivie journellement, àquelques exceptions près, du 1er septem-bre 1990 au 30 avril 1991, était très éle-vée. Une partie des acariens morts étaitendommagée (pattes perdues, cuticule

mutilée). Leur taux augmente parallèle-ment à la mortalité de façon significative.Les différences entre les colonies et entreles mois ne sont pas significatives en rai-son, vraisemblablement, du nombre res-

treint d’ob-servations.

Les acariens morts de 5 colonies ontété observés au microscope (tableau I). Lamortalité moyenne par colonie durant la

période expérimentale a été de 1 235,8acariens. Même réduite aux mois d’hiver

(novembre-mars), la mortalité est 6 fois su-périeure à celle trouvée par Moosbeckho-fer (1991) chez des colonies non sélection-nées. La mortalité naturelle et l’infestationrésiduelle ont été calculées pour une colo-nie : entre le 1 er septembre et le 10 octo-bre 1 640 acariens ont été trouvés sur le

plancher de la ruche; 40% étaient endom-magés. À partir de l’examen d’un échan-tillon de 304 abeilles, on a estimé à envi-ron 1 370 le nombre d’acariens encore

présents dans la colonie, soit 45,6% del’effectif total (morts + survivants).

La lésion la plus fréquente est la pertetotale ou partielle d’une ou de plusieurspattes (fig 1), mais elle peut aller jusqu’àl’ablation totale de tous les organes ven-traux (fig 2). Des mutilations de ce typesont connues comme faisant partie du

comportement de défense d’A cerana,mais les mandibules d’A mellifera étaientconsidérées comme trop faibles pour agirainsi. Des acariens fraîchement amputésmais encore vivants ont été obtenus en li-

bérant 10-15 individus de cellules de cou-vain operculé de la même colonie sélec-tionnée. Dix minutes plus tard, on a retrou-vé régulièrement un certain nombred’acariens amputés remuant encore sur leplancher de la ruche. Les photos au MEBmontrent clairement les faisceaux de mus-cles et la carapace chitineuse sectionnés

par un instrument tranchant (figs 3-6). Laseule cause plausible est le comportementde défense semblable à celui d’A cerana.L’anatomie comparée des mandibules etde leurs muscles (figs 7-10) ne fournit au-cune preuve du manque de force des man-dibules d’A mellifera. Il faut supposer quela différence dans la défense active contre

Varroa, qui existe entre A cerana et A mel-lifera et visiblement aussi entre les diver-ses races d’A mellifera, est due au fait queles ouvrières d’A mellifera ne reconnais-sent pas Varroa comme un ennemi. Ellesreconnaissent par contre très bien d’autres

intrus, comme les larves de Galleria mello-nella. On a pu montrer qu’en l’absence detraitement chimique, le comportement dedéfense active contre Varroa accroît les

chances de survie des colonies d’A cerana

sélectionnées. Cela fournit une base pro-metteuse pour sélectionner une abeillecarniolienne plus résistante à Varroa.

Apis mellifera carnica / résistance àVarroa / comportement de défense / am-putation / mandibule

Zusammenfassung — Aktive Varroa-Abwehr bei selektierten Völkern einesCarnica-Stammes. Ein Berufsimker in

Österreich selektierte aus seinem Bestandvon 700 Carnica-Völkern 12 Völker miteinem nur langsamen Anstieg der Varroa-Population während der Saison. Im Ge-

gensatz zu allen anderen Völkern der

Region, die zum Überleben regelmäßigermedikamentöser Behandlung bedürfen,überwinterten diese Völker ohne jede Be-

handlung und sie entwickelten sich in derfolgenden Saison zu normaler Stärke miteinem dem Standmittel entsprechendenErtrag. Der Milbenabfall der selektierten

Völker, der vom 1 September bis zum

30 April mit wenigen Ausnahmen täglichprotokolliert wurde, war sehr hoch. Dietoten Milben zeigten teilweise Schädenverschiedenen Ausmaßes (Verlust von

Beinen, Beschädigung des Chitinpanzers;Abb 1, 2). Der Anteil beschädigter Tierestieg mit der Milbenmortalität signifikantan. Die Unterschiede zwischen den Völ-kern und Monaten waren statistisch nicht

signifikant, wahrscheinlich wegen zu gerin-ger Zahl von Beobachtungen.

Die toten Milben aus fünf Völkern

wurden unter dem Mikroskop untersucht(Tabelle I). Der mittlere natürliche Milben-abfall je Volk betrug in der Untersuchungs-periode 1 235,8 Tiere. Selbst wenn manden Abfall nur für die eigentlichen Winter-monate (November - März) heranzieht, soentspricht die Zahl dem Sechsfachen

dessen, was Moosbeckhofer (1991) beiunselektionierten Völkern festgestellthatte. Bei einem Volk wurden natürlicherAbfall und Restbefall bestimmt: zwischen

1,9 und 10,10 wurden insgesamt 1 640

Milben auf dem Bodenbrett gefunden,40% von ihnen beschädigt; aus der Hoch-rechnung des Befalls von 304 untersuch-ten Bienen wurde geschätzt, daß im Volknur etwa 1 340 Milben verblieben waren,das sind 45,6% der ursprünglichen Ge-samtzahl.

Als häufigste Beschädigung der Milbenwurde der Verlust eines oder mehrerer

Beine, ganz oder teilweise, festgestellt(Abb 1), bis zur beinahe vollständigen Ent-fernung sämtlicher ventraler Organe (Abb2). Verletzungen dieser Art sind als Teil

des Verteidigungsverhaltens von Apiscerana bekannt, aber bei A melliferawurden die Mandibeln für derartige Effektefür zu schwach gehalten.

Frisch amputierte, noch lebende Milbenwurden durch Freisetzung von 10-15 Indivi-duen aus frisch verdeckelten Brutzellen

desselben, selektierten Volkes gewonnen.Regelmäßig wurde etwa 10 min später eineAnzahl amputierter, noch bewegungsfähi-ger Milben am Bodenbrett gefunden. Ra-sterelektronenmikroskopische Aufnahmen

zeigen sehr deutlich frische Muskelbündelund das Chitin der Körperdecke, alles glattdurchgeschnitten von einem offensichtlichscharfen Instrument (Abb 3-6). Die einzigeplausible Erklärung für diesen Effekt ist die-selbe aktive Varroaverteidigung durch Ar-beitsbienen, wie sie bei A cerana beobach-tet wurde. Ein anatomischer Vergleich derMandibeln und ihrer Muskeln (Abb 7-10)ergab kein Argument für eine wesentlich ge-ringere Mandibelkraft bei A mellifera.

Es ist anzunehmen, daß die Unterschie-de in der aktiven Varroaabwehr, die zwi-schen A cerana und A mellifera und offen-sichtlich auch zwischen verschiedenenRassen von A mellifera bestehen, aufeinem Mangel der Fähigkeit der Bienendieser Art beruhen, die Varroa-Milbe als

Feind zu erkennen. Andererseits erkennensie aber andere Feinde, wie zB Wachs-

mottenlarven, sehr gut als solche. Eskonnte gezeigt werden, daß die aktive Var-roaabwehr die Chance von selektiertenCarnica-Völkern erhöht, auch ohne chemi-sche Behandlung zu überleben. Das ergibteine vielversprechende Grundlage zur

Auslese einer besser varroatoleranten Car-nica-Biene.

Apis mellifera carnica / Varroa-Resistenz / Verteidigungsverhalten /

Amputation / Mandibel

REFERENCES

Moosbeckhofer R (1991) Varroaverluste wäh-rend der Überwinterung. Bienenvater 112,300-303

Moretto G, Gonçalves LS, De Jong D (1991) Af-ricanized bees are more efficient at removingVarroa jacobsoni - preliminary report. AmBee J 131 (7) 434

Peng YS, Fang Y, Xu S, Ge L (1987) The resis-tance mechanism of the Asian honeybee,Apis cerana Fabr, to an ectoparasitic miteVarroa jacobsoni Oud. J Invertebr Pathol 40,54-60

Ruttner F (1988) Biogeography and Taxonomyof Honeybees. Springer, Heidelberg

Ruttner F (1991) Auf dem Wege zu einer varroa-toleranten Carnica. Allg Dtsch Imkeztg 25(11) 10-15

Ruttner F, Marx H, Marx G (1984) Beobachtun-gen über eine mögliche Anpassung von Var-roa jacobsoni an Apis mellifera in Uruguay.Apidologie 15, 43-62

Sakagami SF (1959) Some interspecific rela-tions between Japanese and Europeanhoneybees. J Anim Ecol 28, 51-68

Sakagami SF (1960) Preliminary report on thespecific difference of behaviour and other ec-ological characters between European andJapanese honeybees. Acta Hymenopterol 1,171-198

Wallner A (1990) Imkern heute. SelbstverlagA-263 Randegg, Austria