HAL Id: hal-00891600https://hal.archives-ouvertes.fr/hal-00891600

Submitted on 1 Jan 1999

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.

Economic threshold for Varroa jacobsoni Oud. in thesoutheastern USA

Keith S. Delaplane, W. Michael Hood

To cite this version:Keith S. Delaplane, W. Michael Hood. Economic threshold for Varroa jacobsoni Oud. in the south-eastern USA. Apidologie, Springer Verlag, 1999, 30 (5), pp.383-395. �hal-00891600�

Original article

Economic threshold for Varroa jacobsoni Oud.in the southeastern USA

Keith S. Delaplanea* W. Michael Hoodb

a Department of Entomology, University of Georgia, Athens, GA 30602, USAb Department of Entomology, Clemson University, Clemson, SC 29634, USA

(Received 14 May 1998; revised 20 April 1999; accepted 22 April 1999)

Abstract - This research was designed to determine economic thresholds for Varroa jacobsonimites in mature overwintered colonies under conditions that encourage or discourage mite immi-gration. Congruent data from the present study and our earlier work suggest that a true late-season(August) economic threshold for mites in the southeastern USA lies within a range of mite popula-tions of 3 172-4 261, ether roll mite levels of 15-38, and overnight bottom board insert mite levelsof 59-187 in colonies with bee populations of 24 808-33 699. Overwintering colonies can benefit froman additional early-season (February) treatment. This benefit was realized in colonies which inFebruary had the following average values: mite populations 7-97, ether roll 0.4-2.8, bottom boardinserts 0.6-10.2 and bee populations 12 606-13 500. Continuous acaricide treatment never achievedcolony bee populations or brood number significantly higher than in colonies treated more conser-vatively. There is evidence that minimizing mite immigration has the benefit of delaying the onset ofeconomic thresholds. © Inra/DIB/AGIB/Elsevier, Paris

Apis mellifera / Varroa jacobsoni / integrated pest management / chemical resistance management

1. INTRODUCTION

Synthetic acaricides are the most effectiveand widely used method for controlling Var-roa jacobsoni Oud. But recent evidence ofacaricide-resistant mites in Italy [18], France[6, 22] and USA [4] has underscored the

need to develop management practices thatlimit chemical resistance in mites. One wayto do this is to use economic thresholds [19],that is, to treat a colony only when the mitepopulation reaches a level at which mitesare still tolerable by bees but above whichthere may be serious and possibly irrepara-

* Correspondence and reprintsE-mail: ksd@arches.cc.uga.edu

ble harm to the colony. A corollary to aresearch-based threshold is a reliable and

practical mite sampling method. A programbased on accurate sampling and threshold-based treatments can be expected to reducethe overall number of acaricide treatments,relax the selection pressure for mite resis-

tance, reduce the risk of contaminated hive

products, and reduce production costs forbeekeepers.

Economic threshold recommendations

vary by region [2], owing in part to differ-ences in the length of the brood-rearing sea-son, one of the most important regulatorsof mite population dynamics [14]. Thresh-olds also may vary depending on the riskof mite immigration [17] and on the man-agement history of colonies. This presentpaper is part of a project to develop research-based economic thresholds for the Piedmont

region of the southeastern USA under a vari-ety of beekeeping conditions. In earlier workwe established a threshold for first-yearcolonies set up from mail-order packages[9]. In this paper we present experimentalevidence of an economic threshold formature overwintered colonies under condi-tions that encourage mite immigration(1997) or discourage it (1998). We believethat our design and protocols are applica-ble to workers developing economic thresh-olds elsewhere.

2. MATERIALS AND METHODS

2.1. 1997 study:mite immigration encouraged

The 1997 study was designed to develop eco-nomic thresholds for overwintered colonies underconditions that encourage mite immigration. On17-18 February 1997 we organized 60 over-wintering colonies of honey bees in the Pied-mont region of Georgia and South Carolina, USA(2 states x 2 apiaries per state x 15 colonies perapiary). Within each apiary we equalized coloniesfor initial bee populations, amount of brood, andV. jacobsoni mite populations. We did this byshaking all adult bees from one apiary into acommon cage, distributing the brood equally

(using visual estimates) among the 15 emptyhives, and then distributing equally (by weight)adult bees from that state’s sister apiary into theempty hives. By transferring adult bees to theirstate’s sister apiary, we prevented bees from drift-ing back to their original hive locations and nul-lifying our equalizing efforts. We further dis-couraged bee drift by distinguishing hiveentrances with a variety of colored geometricsymbols [15].We monitored initial colony mite levels

immediately after the equalizing procedure withadhesive bottom board inserts [10]. The numberof mites caught per colony in one overnight sam-pling after the equalizing procedure was 0.7 ± 09(range 0-3) in Georgia and 6.2 ± 14.3 (range0-77) in South Carolina. Colonies were

requeened, treated with Fumidil B (Mid-Con) tocontrol Nosema disease, treated with Terramycinantibiotic (Pfizer) to control brood diseases,treated with vegetable oil/sugar patties to con-trol tracheal mites (Acarapis woodi [Rennie])[8], and managed optimally for honey produc-tion.

Each colony within apiary was randomlyassigned one of the following treatments: 1) treat-ment with Apistan® acaricide in February,2) treatment in August, 3) treatment in Februaryand in August, 4) continuous treatment, and 5) notreatment. Three replicates of each treatmentwere run in each apiary. This design permittedmites from non-treated colonies to emigrate totreated colonies within the same apiary. Apistanstrips inserted in colonies in February wereremoved at day 56. Strips inserted in the contin-uously treated colonies were replaced with newstrips every 56-69 days, and strips inserted inAugust were removed at the termination of theexperiment, days 40-42 (South Carolina) or43-44 (Georgia). We replaced failing queens asnecessary and equalized brood within treatmentgroup and apiary to minimize swarming and vari-ation within treatment groups. Swarming wasminimized in South Carolina with the Demareemethod [1].

On 25-26 February, 20, 27, 30 May, and 21,25 August some colonies were sampled to deter-mine colony bee populations, average bodyweight of bees, number of sealed brood cells,colony mite populations, and mite levels with anether roll test and with an overnight adhesivebottom board insert (20 ± 4 h exposure) usingpublished methods [9]. We did not use acaricideto hasten mite drop on bottom board inserts; how-ever, acaricide unavoidably affected our read-ings for May and August in the continuously

treated colonies. On the 25-26 February dates, wesampled only colonies scheduled for acaricidetreatment that month; however, because colonieshad just recently been set up and no treatmentswere yet in place the February parameters givenin table V accurately estimate the initial conditionof all colonies. From 25 September to 28 Octo-ber all surviving colonies were dismantled(n = 55) to measure the above parameters plus thepercentage of brood cells with visible disease-like symptoms [9] and percentage of colony beesinfested with tracheal mites [8]. The followingJanuary, nearly 1 year after the start of the exper-iment, we assessed the condition of the Georgiacolonies with a blind subjective survivabilityscore; each of four observers independentlyexamined each colony and scored it, to the near-est half unit, from zero (dead or irrecoverablyweak) to three (best condition and greatest prob-ability of surviving winter).

The effects of treatment on colony bee popu-lations, body weight of bees, number of sealedbrood cells, colony mite populations, and per-centage brood cells with visible disease-likesymptoms were tested with a completely ran-domized design analysis of variance [20] blockedon state (Georgia or South Carolina). The effectsof treatment and its interactions with state weretested with the interaction of treatment with api-ary nested within state. The effects of state weretested against the effect of apiary nested withinstate. If the interaction of treatment × state was

significant, then analyses were run separately bystate and the error term was the interaction oftreatment with apiary. Least square means wasused to adjust for non-equal sample size, andmeans were separated with a t-test. Differenceswere accepted at the α ≤ 0.05 level. The rela-

tionship between colony mite populations andmite levels found with ether roll and by bottomboard inserts was tested with regression analyses[20]. With bottom board inserts we were inter-ested in measuring natural mite drop, so weexcluded colonies from the regression analysesthat had Apistan strips in them at the time ofsampling.

Forty-one of 55 surviving colonies were pos-itive for A. woodi, and infestation levels rangedfrom 5 to 95 %. Tracheal mite levels were sig-nificantly higher in South Carolina (40.2 ± 5.3 %,mean ± standard error) than in Georgia(12.5 ± 3.4 %) (F = 47.8; df = 1,2; P = 0.0203),but when tracheal mites were included as acovariate in the main analyses they did notexplain any variation in our parameters of inter-est (P ≥ 0.2025).

2.2. 1998 study:mite immigration minimized

The 1998 study was designed to compare thevarious treatment regimens for overwinteredcolonies under conditions of minimized mite

immigration. The basic protocol was the sameas in 1997 with key differences noted below.

We set up 40 colonies (2 states x 5 apiariesper state × 4 experimental colonies per apiary).Each of the ten apiaries was at least 0.6 km awayfrom other known bee colonies. Each apiarywithin state randomly received one of the fiveexperimental treatments as used in 1997, andexperimental colonies within apiary received thesame treatment. Thus, there was minimized riskof mites immigrating from non-treated coloniesto treated colonies within the same apiary.Colonies were removed from the experiment ifthey died or became queenless and in two caseswhen acaricide strips were found in colonies pastthe prescribed treatment interval.

The number of mites caught per colony perday on adhesive bottom board inserts after initialsetup was 0.6 ± 0.3 (range 0-1.3) in Georgia and0.6 ± 1.1 (range 0-4.5) in South Carolina. Apis-tan strips inserted in colonies in February wereremoved at day 56; strips inserted in the contin-uously treated colonies were replaced with newstrips every 56-67 days, and strips inserted inAugust were removed at day 58 (South Carolina)or at days 80-83, the termination of the experi-ment (Georgia). The experiment was dismantledfrom 26-30 October (n = 36 colonies).

The effects of treatment on colony bee popu-lations, body weight of bees, number of sealedbrood cells, colony mite populations, and per-centage brood cells with visible disease-likesymptoms were tested with a completely ran-domized design analysis of variance blocked onstate. The interaction of month of treatment xstate was the error term. Analyses were run sep-arately by state and the parameters tested againstresidual error when treatment x state interactionwas significant.

3. RESULTS AND DISCUSSION

3.1. General

In the 1997 study under conditions thatencouraged mite immigration, there weretreatment effects in September-October on

colony bee populations and colony mitepopulations (F ≥ 6.5; df = 4,8; P ≤ 0.0125),but not for bee weight nor the percentageof brood cells with disease-like symptoms(table I). There were treatment x state inter-actions for the number of sealed brood cells

(F = 9.63; df = 4,8; P = 0.0038); treatmentsaffected brood in South Carolina (F = 16.4;df = 4,4; P = 0.0096) but not in Georgia(table III).

In the 1998 study under conditions thatminimized mite immigration, there were notreatment effects in October for colony beepopulations, bee weight, number of sealedbrood cells, nor the percentage of broodcells with disease-like symptoms (table II).There was a treatment x state interactionfor colony mite populations (F = 4.87;df = 4,26; P = 0.0046) and significant treat-ment effects in each state (F ≥ 12.2;df = 4,12; 4,14; P ≤ 0.0003) (table IV).

3.2. Colony bee populationsand bee weight

Under conditions encouraging miteimmigration, colony bee populations inSeptember-October were highest in coloniestreated continuously with acaricide. Thepopulations of continuously treated colonieswere not significantly different from those ofcolonies treated once in February and againin August (table I); however, this is based ona conservative interpretation of a P valueof only 0.0501 (t-test) between these twomeans. Nevertheless this supports an argu-ment against treating colonies continuously.Continuous treatment not only risks con-taminated hive products and acaricide-resis-tant mites, but in our study failed to achievebee populations significantly larger than inthe more conservative February + Augustschedule. Moreover, the satisfactory per-formance of the February + August schedulein 1997 was achieved in apiaries in whichmite emigration from non-treated colonieswas an acute threat. In 1998 when apiarieswere managed to minimize mite immigra-

tion, the February + August schedule actu-ally yielded the numerically highest aver-age bee populations (table II).

In both years, treatments did not affect

average bee weight (tables I and II).Although V. jacobsoni can reduce bodyweight in bees parasitized as immatures [7],a remedial effect of acaricide was not appar-ent in this study nor in our earlier work withfirst-year colonies [9].

3.3. Number of sealed brood cellsand percentage brood cells withdisease-like symptoms

In South Carolina in September-Octo-ber under conditions encouraging miteimmigration, the number of sealed broodcells was highest in colonies treated con-tinuously and in colonies treated in February+ August (table III). This contrasts with ourearlier work with first-year colonies in whichbrood number was highest in those colonieswith the most V. jacobsoni-induced dys-functions, a phenomenon we hypothesizedmay indicate efforts by bees to compensatefor high levels of brood parasitism [9]. Our1997 results do not support this hypothesisas brood number was significantly highest inthose colonies in which mite control was

optimized. Instead, our present results sup-port a February + August acaricide treat-ment schedule. Continuous acaricide treat-ment has many inherent risks and in our

study failed to achieve brood productionsignificantly higher than in the more con-servative February + August schedule.Moreover, in 1998 the February + Augustschedule yielded numerically similaramounts of brood to that in continuouslytreated apiaries (table II).

Over both years, incidence of brood withdisease-like symptoms occurred in the non-treated, February, August, and February +August treatment schedules, but valuesnever differed significantly from zero (tablesI and II. V. jacobsoni mites vector or acti-vate several bee pathogens [3, 16], and

brood parasitized by mites often display vis-ible disorders [21]. In our study, visiblebrood pathology was eliminated only incolonies treated continuously with acari-cide. We do not believe that this is a com-

pelling defense for continuous treatmentsconsidering the low incidence of visiblebrood pathology in the February + Augustgroups (0 % [1997] and 0.3 % [1998]).

3.4. Colony mite populations

In September-October under conditionsencouraging mite immigration, colony mitepopulations were highest in colonies treatedin February and in non-treated colonies(table I). Mite populations did not differamong the August, February + August, norcontinuously treated colonies, all of whichhad Apistan acaricide strips in them at thetime of sampling. Our data suggest that one

February acaricide treatment is not satis-factory under conditions of mite immigrationin the southeastern USA. Apistan acaricidestrips were removed from February-treatedcolonies by 23 April, and by Septem-ber-October mite populations in thesecolonies had rebounded to the same levelas non-treated colonies. This rapid growthoccurred from reproducing survivors andalso from mites emigrating from non-treatedcolonies in the same apiary.A rapid rebound of mites in February-

treated colonies also occurred in 1998 underconditions of minimized mite immigration.Acaricide was removed from February-treated colonies by 23 April, and by Octobermite populations in these colonies hadrebounded to a level approaching or exceed-ing that of non-treated colonies (table IV).

These results imply that if we are todevelop an optimally conservative single-

treatment acaricide schedule for the south-eastern USA, it will occur late in the sea-son rather than early. A single treatmentearly in the season is unsatisfactory becausemany months of brood rearing remain tosupport reproduction by surviving mites;reproduction rate of mites is high when theratio of mites per brood cell is low [11]. Asingle late-season treatment in August gaveoptimum results in first-year colonies underconditions of minimized mite immigration[9]. In the present study with maturecolonies under varying conditions of mitedrift, single August treatments generallygave intermediate results (tables I-IV).

3.5. Implications foran economic threshold

In the 1997 study under conditionsencouraging mite immigration, satisfactorymite control with minimized use of acari-cide was achieved in mature, overwinteredcolonies treated once in February and againin August. Continuous treatment, a non-sus-tainable practice included in this study toprovide a positive check, failed to achievebee populations, brood number, and mitepopulations significantly different from themore conservative February + August sched-ule. A single February treatment permittedan unacceptably high rebound of mites bySeptember-October which was associatedwith reduced brood number (in South Car-olina) compared to February + August-treated colonies (tables I and III). In 1998under conditions of apiary isolation, averageOctober bee populations were numericallyhighest in February + August-treatedcolonies (table II).

However, there is reason to hypothesizethat a more conservative single late-seasontreatment may work, especially if the bee-keeper is able to minimize mite immigra-tion by isolating apiaries from knownsources of mite contamination and by treat-ing all colonies in an apiary simultaneouslywith acaricide. Our 1998 study with isolated

apiaries failed to demonstrate significantdifferences in October among treatmentsfor all variables except colony mite popu-lations (tables II and IV). But it is worth not-ing that October bee populations and inci-dence of diseased brood in colonies treated

only in August compared favorably withother treatments; bee populations evenexceeded numerically the bee populationsin continuously treated colonies (table II).

It is insightful to compare the condition ofAugust- and February + August-treatedcolonies in the Augusts of both years (tablesV and VI). Comparing the February + August-treated colonies, the mite populations, etherroll, and bottom board insert levels were

predictably lower in 1998 under conditionsof isolation, and this was associated withgood colony condition in October (table II).It is possible that the February + Augustcolonies in 1998 were below economicthreshold when they were treated in August.However, it is possible that the August-treated colonies in 1998 were in fact at eco-nomic threshold when they were treated.Note the similarity between August mitepopulations in August-treated colonies in1998 (4 057, table VI) and in February +August-treated colonies in 1997 (4 261,table V).

It is arguable that a single August treat-ment gave satisfactory results even in 1997under conditions of mite immigration.August-treated and February + August-treated colonies did not differ significantlyin September-October in bee populations,bee weight, and incidence of diseased brood,and August-treated colonies were subjec-tively scored the following January asequally likely to survive winter (table I).These similarities are paralleled by similarcolony mite populations in August (4 144and 4 261, table V).

If late-season mite populations of ∼4 000represent a tolerable level, we must alsoidentify levels that are damaging. Our earlierwork with first-year colonies [9] establishedthat late-season mite populations of 6 662

exceed economic threshold, and in the pre-sent study mite populations approaching6 000 were associated with decreasing beepopulations and increasing brood disease inSeptember-October and poor colony con-dition the following January (table I).

In summary, we believe that in this studyonly the August-, February + August- (1997)and August-treated (1998) colonies were infact at economic threshold when they weretreated in August. Moreover, the Augustparameters of these mature colonies (tablesV and VI) and the August threshold param-eters we developed for first-year colonies[9] are noticeably congruent. Thus, we sug-gest that a true late-season economic thresh-old for varroa mites in the southeastern USAlies within a range of mite populations of3 172-4 261, ether roll mite levels of 15-38,and bottom board insert mite levels of59-187 in colonies with bee populations of24 808-33 699. Our present results suggestthat overwintering colonies can benefit froman additional early-season (February) treat-ment. This benefit was realized in colonieswhich in February had the following averagevalues: mite populations 7-97, ether roll0.4-2.8, bottom board inserts 0.6-10.2, andbee populations 12 606-13 500 (tables Vand VI). And finally, our present results sug-gest that apiary isolation has the benefit ofdelaying the onset of economic thresholds;August colony mite levels were consider-ably lower in February + August-treatedcolonies in 1998 under conditions of isola-tion than they were in 1997 under condi-tions of mite immigration.

Economic thresholds vary by region [2],owing to such differences as brood rearingseason and possibly unknown genetic dif-ferences in bee and mite populations. Thus,we believe that this type of research shouldbe replicated in other regions. Our proto-cols described here and earlier [9] were ableto identify mite levels that were damagingand levels that were tolerable, and webelieve that these protocols can give practi-cal results elsewhere.

3.6. Sensitivity of mite samplingmethods, 1997 data

The relationship of ether roll mite levelswith colony mite populations was describedby a model (r = 0.74) with a simple posi-tive linear term (coefficient = 0.013 ± 0.001). The same relationship with bottom boardinserts was described by a model (r = 0.76)with linear (0.05 ± 0.01), quadratic(-5.2*10-6 ± 1.8*10-6), and cubic (1.6*10-10± 0) terms. The simple positive linear rela-tionship in the first model means that onecan expect increasing ether roll levels toreliably predict increasing colony mite pop-ulations. The complicated cubic relation-ship with bottom board inserts means thathigh colony mite population may yield highor low insert levels.

These results diametrically contradict ourearlier work in which bottom board inserts

yielded the more straightforward linearmodel [9]. Bottom board inserts is gener-ally regarded as the more reliable samplingmethod [12, 13] because inserts potentiallysample the entire adult bee population. Wedid not use acaricide to hasten mite drop,so our values reflect natural mite drop offadult bees during the sampling interval(20 ± 4 h). Excluding the continuouslytreated colonies which contained Apistanstrips at the time of sampling, the number ofmites retrieved on inserts expressed as a per-centage of colony mites on adult bees was2.8-23.9 % (derived from table V) whichis similar to the daily range of 0.6-17.8 %published by others [5].

The ether roll test is easy to do, and ourpresent results suggest that it can reliablypredict colony mite populations. Collec-tively, these results suggest that threshold-based treatment decisions should be basedon an average of several samplings. Themethod of sampling, whether by ether roll orbottom board inserts, may be of secondaryimportance.

ACKNOWLEDGMENTS

We thank Carl Webb, Clarkesville, Georgiaand Berry Wright, Gainesville, Georgia who gra-ciously donated colonies, time, and labor for thisexperiment. Technical assistance was provided byLloyd Allison, Priscilla Bennett, Jennifer Berry,Jamie Ellis, Hosafy Eshbah, Art Limehouse,Clyde McCall, Denny Smith, and Troy Usher.Funding was provided by the Georgia Beekeep-ers Association, Wellmark International, and theAgricultural Experiment Stations of the Univer-sity of Georgia and Clemson University.

Résumé - Seuil de dégâts économiquespour Varroa jacobsoni dans le sud-est desÉtats-Unis. Dans ce travail nous apportonsla preuve d’un seuil économique pour Var-roa jacobsoni dans des colonies ayanthiverné dans des conditions qui soit facili-tent, soit entravent la réinfestation par l’aca-rien.En février 1997 nous avons installé dans la

région du Piedmont de la Géorgie et de laCaroline du Sud, États-Unis, 60 coloniesd’abeilles ayant hiverné. Les ruchers étaientinfestés par V. jacobsoni. Chaque colonieau sein d’un même rucher a reçu l’un destraitements suivants : 1) traitement à l’Apis-tan® en février, 2) en août, 3) en février et enaoût, 4) en continu , 5) pas de traitement.Ce dispositif permettait aux acariens descolonies non traitées de réinfester les colo-nies traitées d’un même rucher. En février,mai et août, on a déterminé dans certainescolonies les paramètres suivants : taille de lapopulation d’abeilles, poids corporel moyendes abeilles, nombre de cellules de couvainoperculé, taille de la population d’acariens etniveau d’infestation à l’aide du « test durouleau d’éther » [9] et d’un lange adhésifplacé sur le plateau de fond. En septembre eten octobre, toutes les colonies survivantesont été démantelées afin de mesurer ces

paramètres ainsi que le pourcentage de cel-lules de couvain présentant des symptômesvisibles de maladie.Le protocole de base a été répété en 1998,mais cette fois-ci la réinfestation a été réduiteau minimum. Nous avons installé 40 colo-

nies (2 états x 5 ruchers/état x 4 colonies/rucher). Chaque rucher était au moins distantde 0,6 km des autres colonies connues. Dansun même état chaque rucher a reçu l’un descinq traitements et les colonies d’un rucheront toutes reçu le même traitement. Ainsile risque de réinfestation des colonies traitéespar les acariens des colonies non traitéesétait réduit au minimum. L’expérience s’estterminée en octobre (n = 36 colonies).Nos résultats suggèrent que seules les colo-nies traitées en août 1997, en février + août1997 et en août 1998 étaient au seuil éco-

nomique lorsqu’elles ont été traitées en août.En outre les paramètres pour le mois d’aoûtde ces colonies (tableaux V et VI) concor-dent nettement avec les paramètres du seuilen août que nous avions développés pourdes nuclei. Nous estimons donc qu’il existeun véritable seuil économique d’arrière sai-son pour l’acarien V. jacobsoni dans le sud-est des États-Unis, situé dans une fourchettede population d’acariens comprise entre3 172 et 4 261. Cela correspond à un nombred’acariens entre 15 et 38 par rouleau d’étheret entre 59 et 87 par lange sur le plateaupour une colonie de 24 808 à 33 699abeilles. La concordance entre cette étudeet la précédente [9] suggère qu’en arrièresaison des populations d’acariens ∼ 6 000sont au-dessus du seuil économique. Lescolonies qui hivernent peuvent bénéficierd’un traitement supplémentaire d’avant-sai-son (février). C’était le cas des colonies quien février avaient les valeurs moyennes sui-vantes : population d’acariens 7-97, rou-leau d’éther 0,4-2,8, langes 0,6-10,2, popu-lation d’abeilles 12 606-13 500 (tableaux Vet VI). Un traitement acaricide continu n’a jamais réussi à augmenter significativementla population d’abeilles ni la quantité decouvain par rapport au traitement de février+ août, plus traditionnel. Enfin nos résul-tats montrent que l’isolement d’un rucherretarde le dépassement des seuils écono-miques ; les niveaux d’acariens dans lescolonies d’août étaient considérablement

plus faibles dans les colonies traitées enfévrier + août dans les conditions d’isole-

ment de 1998 que dans celles de 1997 enconditions de réinfestation. © Inra/DIB/

AGIB/Elsevier, Paris

Apis mellifera / Varroa jacobsoni / lutteintégrée / gestion de la résistance

chimique

Zusammenfassung - Ökonomische Scha-densschwelle durch Varroa jacobsoni imSüdosten der USA. In dieser Arbeit belegenwir die ökonomische Schadensschwelle fürSchäden durch Varroa jacobsoni an über-winternden Wirschaftsvölkern unter Mil-beninfektion fördernden oder behindernden

Bedingungen.Ab Februar 1997 betreuten wir 60 über-winterte Bienenvölker in der Piedmont

Region von Georgia und Süd Carolina,USA. Die Bienenstände waren mit Varroa

jacobsoni infiziert. Jedes Volk pro Standwurde einer der folgenden Behandlungenunterzogen: 1) Behandlung mit Apistan imFebruar, 2) Behandlung im August, 3)Behandlung im Februar und August, 4) Dau-erbehandlung und 5) keine Behandlung. Beidieser Anordnung konnten die Milben vonunbehandelten Völkern in behandelte inner-halb des Bienenstandes eindringen. ImFebruar, Mai und August wurden einigeVölker untersucht, um die Volksstärke, dasmittlere Körpergewicht der Bienen, die Zahlder verdeckelten Brutzellen, die Milbenpo-pulation und den Milbenbefall durch einen’Ätherrolltest’ und mit klebriger Bodenein-lage zu bestimmen. Im September und Okto-ber wurden alle überlebenden Völker (n =

55) aufgelöst, um die oben aufgeführtenParameter und zusätzlich den Prozentsatzder Brutzellen mit sichtbaren Schadens-

symptomen zu messen.Dieser Versuch wurde 1998 wiederholt, nurwurde diesmal die Milbeneinwanderungminimiert. Dazu stellten wir 40 Völker auf

(2 Staaten x 5 Bienenstände pro Staat x 4Versuchsvölker pro Stand). Jeder Stand warmindestens 0,6 km von anderen bekanntenVölkern entfernt, auf jedem Stand pro Staatwurde eine der 5 Behandlungen angewendet,

und alle Versuchsvölker innerhalb einesStandes wurden gleich behandelt. Dadurchwurde das Risiko einer Rückinfizierungdurch Milben aus nicht behandelten Völ-kern innerhalb eines Standes minimiert. DerVersuch wurde im Oktober aufgelöst (n = 36Völker).Nach unseren Daten scheinen nur die 1997im August bzw. im August + Februar unddie 1998 im August behandelten Völkerunter der ökonomischen Schadensschwellezu bleiben. Zusätzlich stimmen die AugustParameter dieser Wirtschaftsvölker (Tabel-len V und VI) und die August Schwellen-parameter, die wir für Ableger entwickel-ten, deutlich überein. Deshalb nehmen wir

an, daß eine reale spätsaisonale Schadens-schwelle durch Varroamilben im Südostenvon USA bei einer Milbenpopulation von3 172-4 261 besteht. Das entspricht einerÄtherroll-Milbenzahl von 15-38 und einerMilbenzahl auf der Bodeneinlage von59 - 187 bei einer Volksstärke von 24 808bis 33 699 Bienen. Übereinstimmende Datendieser Untersuchung und unserer früherenArbeit [9] legen nahe, daß eine spätsaiso-nale Milbenzahl von über 6 000 über derökonomischen Schadensschwelle liegt.Überwinterte Völker können von einerzusätzlichen Frühjahrsbehandlung (Februar)profitieren. Dieser Vorteil wurde bei Völ-kern mit Volksstärken von 12 606-13 500Bienen erreicht, bei denen im Durchschnittfolgende Werte im Februar vorlagen: Mil-benpopulation 7-97, Ätherrolltest 0,4-2.8;Bodeneinlage 0,6-10,2 (Tabellen V und VI).Eine Dauerbehandlung erreichte nie einesignifikant höhere Bienenzahl oder Brut-menge als die herkömmliche Behandlungim Februar + August. Schließlich zeigenunsere Ergebnisse auch, daß die Isolationvon Bienenständen das Überschreiten derSchadensschwelle behindert. Im August1998 war der Milbenbefall unter isolierten

Bedingungen bei im Februar + Augustbehandelten Völkern verhältnismäßig nied-riger als es 1997 unter Bedingungen mitRückinfizierung durch Milben der Fall war.© Inra/DIB/AGIB/Elsevier, Paris

Apis mellifera / Varroa jacobsoni /

integrierte Behandlung / Managementder chemischen Resistenz

REFERENCES

[1] Ambrose J.T., Management for honey produc-tion, in: Graham J.M. (Ed.), The Hive and theHoney Bee, Dadant and Sons, Hamilton, Illi-nois, USA, 1992, pp. 601-655.

[2] Bach J.C., Danka R.G., Ellis M.D., Mussen E.C.,Pettis J.S., Sanford M.T., Protecting Honey Beesfrom Varroa jacobsoni, American Association ofProfessional Apiculturists, Lincoln, Nebraska,USA, 1998.

[3] Ball B.V., Allen M.F., The prevalence ofpathogens in honey bee (Apis mellifera) coloniesinfested with the parasitic mite Varroa jacob-soni, Ann. Appl. Biol. 113 (1988) 237-244.

[4] Baxter J.. Eischen F., Pettis J., Wilson W.T.,Shimanuki H., Detection of fluvalinate-resistantvarroa mites in US honey bees, Am. Bee J. 138(1998) 291.

[5] Boot W.J., Calis J.N.M., Beetsma J., Does timespent on adult bees affect reproductive success ofVarroa mites? Entomol. Exp. Appl. 75 (1995)1-7.

[6] Colin M.E., Vandame R., Jourdan P., DiPasquale S., Fluvalinate resistance of Varroajacobsoni Oudemans (Acari: Varroidae) inMediterranean apiaries of France, Apidologie28 (1997) 375-384.

[7] De Jong D., De Jong P.H., Gonçalves, L.S.,Weight loss and other damage to developingworker honeybees from infestation with Varroajacobsoni, J. Apic. Res. 21 (1982) 165-167.

[8] Delaplane K.S., Controlling tracheal mites(Acari: Tarsonemidae) in colonies of honey bees(Hymenoptera: Apidae) with vegetable oil andmenthol, J. Econ. Entomol. 85 (1992)2118-2124.

[9] Delaplane K.S., Hood W.M., Effects of delayedacaricide treatment in honey bee colonies para-sitized by Varroa jacobsoni and a late-seasontreatment threshold for the southeastern USA,J. Apic. Res. 36 (1997) 125-132.

[10] Dietz A., Hermann H.R., Biology, Detection andControl of Varroa jacobsoni: a Parasitic Miteon Honey Bees, Lei-Act Publishers, Commerce,Georgia, USA, 1988.

[11] Eguaras M., Marcangeli J., Fernandez N.A.,Influence of ’parasitic intensity’ on Varroajacobsoni Oud. reproduction, J. Apic. Res. 33(1994) 155-159.

[12] Ellis M.D., Baxendale F.P., Comparison offormic acid sampling with other methods todetect varroa mites (Varroa jacobsoni Oud.) andmite distribution within colonies in Nebraska,BeeScience 3 (1994) 139-144.

[13] Ellis M., Nelson R., Simonds C., A comparisonof the fluvalinate and ether roll methods of sam-

pling for varroa mites in honey bee colonies,Am. Bee J. 128 (1988) 262-263.

[14] Fries I., Camazine S., Sneyd J., Populationdynamics of Varroa jacobsoni: a model and areview, Bee World 75 (1994) 5-28.

[15] Frisch von K., Bees: Their Vision, ChemicalSenses, and Language, Jonathan Cape, London,1984.

[16] Gli&jadnr;ski Z., Jarosz J., Varroa jacobsoni as a car-rier of bacterial infections to a recipient bee host,Apidologie 23 (1992) 25-31.

[17] Greatti M., Milani N., Nazzi F., Reinfestationof an acaricide-treated apiary by Varroa jacob-soni Oud, Exp. Appl. Acarol. 16 (1992) 279-286.

[18] Lodesani M., Colombo M., Spreafico M., Inef-fectiveness of Apistan® treatment against themite Varroa jacobsoni Oud. in several districtsof Lombardy (Italy), Apidologie 26 (1995)67-72.

[19] Luckmann W.H., Metcalf R.L., The pest-man-agement concept, in: Metcalf R.L., LuckmannW.H. (Eds.), Introduction to Insect Pest Man-agement, 2nd ed., John Wiley and Sons, NewYork, 1982, pp. 1-31.

[20] SAS Institute, SAS/STAT User’s Guide, ver-sion 6, 4th ed., SAS Institute, Cary, North Car-olina, USA, 1989.

[21] Shimanuki H., Calderone N.W., Knox D.A., Par-asitic mite syndrome: the symptoms, Am. BeeJ. 134 (1994) 827-828.

[22] Trouiller, J., Résistance du varroa au fluvalinate:résultats des campagnes de détection de 1996,Rev. Fr. Apic. 571 (1997) 115-117.