Review article

Invasion behaviour of Varroa jacobsoni Oud.:from bees into brood cells

Joop Beetsma Willem Jan Boot Johan Calis

Department of Entomology, Agricultural University of Wageningen,PO Box 8031, 6700 EH Wageningen, the Netherlands

(Received 8 September 1998; accepted 11 February 1999)

Abstract - Varroa jacobsoni mites may invade worker or drone brood cells when worker bees bringthem in close contact with these cells. The attractive period of drone brood cells is two to three timeslonger than that of worker brood cells. The attractiveness of brood cells is related to the distancebetween the larva and the cell rim and the age of the larva. The moment of invasion of the mite intoa brood cell is not related to the duration of its stay on adult bees. The fraction of the phoretic mitesthat invade brood cells is determined by the ratio of the number of suitable brood cells and the sizeof the colony. The distribution of mites over worker and drone brood in a colony is determined by thespecific rates of invasion and the numbers of both brood cell types. Knowledge of mite invasionbehaviour has led to effective biotechnical control methods. © Inra/DIB/AGIB/Elsevier, Paris

Apis mellifera / Varroa jacobsoni / invasion / brood cell type / biotechnical control

1. INTRODUCTION

The parasitic mite Varroa jacobsoniOudemans (Acari: Varroidae) is a harmfulpest of the European honeybee races. Up tonow most beekeepers have applied acari-cides to control this mite in their colonies.

However, the use of acaricides has two

major disadvantages: 1) contamination ofhoneybee products [63], and 2) the occur-rence of mite resistance [29, 40, 44]. There-

fore, biotechnical control methods shouldbe preferred. Adult female V. jacobsoni feedon the haemolymph of both adult bees andbrood. While staying on adult bees, theycan survive up to several months; for exam-ple, during the broodless period in coloniesin a temperate climate. This ability allowsthe mites to wait for an opportunity to invadea brood cell. Invasion into a worker or dronebrood cell before cell capping is essentialfor mite reproduction [39, 50]. Studying the

* Correspondence and reprintsE-mail: janboot@users.ento.wav.nl

† After an active life filled with curiousity about bees, beekeeping beekeepers and researchers, we lostJoop Beetsma as he passed away on 26 March, 1999 after a sudden and short illness

process of invasion into brood cells is impor-tant for two reasons. First, invasion isdirectly related to reproduction and there-fore defines population growth. Second, theresults of these studies have led to the

improvement of biotechnical control meth-ods in which trapping combs are used. Inthis paper, current knowledge on the pro-cess of invasion behaviour is reviewed.

2. INVASION BEHAVIOUROF INDIVIDUAL MITES

Boot et al. [10] studied invasionbehaviour of individual mites in small, heav-ily infested colonies. Experiments were car-ried out in a specially designed observationhive using a ’half-comb’ frame [2]. Gener-ally, two frames, both containing a ’half-comb’, one on top of the other, were used tomaintain a brood nest large enough forobservations. The ’half-comb’ consisted ofone layer of cells of which the cell bottomshad been replaced by a transparent sheet.Two movable video cameras were used tomonitor the opening and the bottom of agroup of cells simultaneously. The side ofthe cell openings was illuminated with infra-red light which penetrated the whole cellincluding the larva. When mites invadedbrood cells, they crawled between the cellwall and the larva until they reached the bot-tom of the cell, where they were trapped inthe larval food [8, 24, 38]. The observationssuggested that the movement of a mite fromthe cell opening to the bottom took about1 min. When studying the recordings of theposition of the bees at the cell opening over3 min before the mite reached the cell bot-tom, Boot et al. [10] concluded that it wasnot necessary for the bee to place its headand thorax into the brood cell for mite inva-sion. Apparently, mites left bees whenbrought in close proximity to a brood cell.Because mites were never observed to walkacross the comb surface, Boot et al. [10]suggested that mites went directly from theventral side of the bee into the brood cell.

When using the ’half-comb’ frames, the

bees blocked the view on the cell opening.Therefore, a frame of cells with transparentside walls was constructed. Mites could beobserved through opposite transparent per-spex cell walls of vertical rows of cells [2]into which larvae had been grafted. Whenthe larvae were large enough to attract mites,two video cameras were placed at oppositesides of a few cells to record invasion ofmites. After many attempts, the completemovement and several parts of the move-ment from the bee to the brood cell bottomcould be recorded. The mite walked overthe side of the abdomen of the bee, left thebee and moved onto a cell capping andentered the adjacent cell, walked on the sur-face of the larva and crawled between thelarva and the cell wall to the bottom of thecell. Boot et al. [10] concluded that mitesonly invade brood cells when the distancebetween the mite on the bee and the cellopening is small. Ten minutes after therecorded invasion, eight other mites wereseen on bees passing the recently invadedcell, but they did not invade this cell. Sinceinfested brood cells seem to be just as attrac-tive to mites as non-infested ones [30], thedistance between these mites and the cell

may have been too long for invasion. Inaddition, mites positioned between theabdominal sternites [41, 48] may notrespond to stimuli from the brood cell. Themites were never seen walking on the comb,or entering and leaving the brood cells toselect a cell for invasion. In cells with attrac-tive larvae, the mites have to cover a dis-tance of 4-7 mm from cell opening to thelarva [13, 34]. Therefore, the signal to decidewhether to stay on the bee or to enter thebrood cell is perceived at a distance of atleast 4 mm from the larva and not after directcontact with the larva.

3. ATTRACTIVENESSOF BROOD CELLS

Preference of mites to drone larvae wasfound first in tests outside the colony [42, 45,52], but whether mites could discriminate

between drone and worker brood cells in anatural environment had not yet been shown.Therefore, Boot et al. [8] compared the inva-sion of mites into worker brood cells withthat into drone brood cells in small highlyinfested colonies kept in an observation hive,using half-combs [2]. For each cell, recordswere made of the time that a mite appearedat the transparent cell bottom and the time atwhich the cell had been capped. Invasioninto worker and drone brood cells was stud-ied in separate experiments. Invasion intoworker brood cells could be recorded from15-20 h before cell capping, and in dronebrood cells from 40-50 h before cell cap-ping. Because the ratio between the num-ber of phoretic mites and available broodcells changed gradually within each exper-iment, the rate of invasion of mites intobrood cells must have been affected [11].Therefore, this experiment gave informa-tion only about the duration of the attrac-tive period of each cell, and not on the rateof invasion within the attractive period.When comparing brood cells containing atleast five mites, Boot et al. [8] concludedthat brood cells can be invaded during thewhole invasion period. The number of mitesinvading worker brood cells per hourremained more or less constant until cell

capping, but decreased before capping ofdrone brood cells. Boot et al. [8] attributedthe decreasing rate of invasion into dronebrood cells to a limited number of mites inrelation to the number of drone brood cellsin the small experimental colony. The attrac-tive period of drone brood cells was two tothree times longer than that of worker broodcells (figure 1). This was in agreement withthe results of a similar study by Wieting andFerenz [64] and earlier results based on indi-rect criterions [33, 38].

Comparison of the rates of invasion ofworker and drone brood cells simultane-

ously in one colony is not practical becauseof the differential time of capping of bothcell types and because of the longer periodof attractiveness of drone brood cells. Ifworker and drone larvae are of the same ageat the start of the experiment, after all workerbrood cells have been capped, invasion intodrone brood cells continues while the densityof mites on the bees has decreased.

A different distribution of mites has beenfound in different types of cells containingthe same type of larva. De Jong and Morse[23] found more mites in worker cells pro-truding above the comb surface than inneighbouring worker cells. De Ruijter and

Calis [27] found more mites in worker broodcells with artificially raised bottoms. Calis etal. [17] and Ramon et al. [46] found moremites in smaller cells, when brood attrac-tivity to mites was tested in cells differing indiameter. Calis et al. [17] and Goetz andKoeniger [34] also found more mites inshortened worker brood cells.

Boot et al. [ 13] measured the period thatbrood cells are attractive to mites, the dis-tribution of mites over different cell types,and the distance between larva and cell rimof different cell types, in relation to the time

preceding cell capping. The attractive periodof brood cells was again measured in half-combs [2, 9]. Two half-combs, clampedtogether with adjacent transparent sheets,were introduced into a heavily infestedcolony. The times of mite invasion and cellcapping were recorded on transparent sheets.Invasion was recorded in normal workerand drone brood cells, shortened worker anddrone brood cells, elongated worker broodcells, drone cells provided with a workerlarva, and worker cells provided with adrone larva. The distance between larva andcell rim was measured with a probe, as usedby Goetz and Koeniger [34]. To comparethe attractiveness for mites between

untreated, shortened or elongated cells orworker cells with a drone larva and vice

versa, Boot et al. [13] used cells with thesame width containing larvae of the sameage in one test colony. Therefore, the rela-tionship between the estimated distancesbetween the cell rim and the nearest surfaceof the coiled larva producing an attractantwould give the same result as estimating thevolume of the unoccupied part of the cells.

Shortening of both worker and dronebrood cells always resulted in a longerattractive period than control brood cells.Elongated worker brood cells were attractiveto the mites for a shorter period than con-trol worker brood cells. Drone cells con-

taining a worker larva seemed to be attrac-tive to mites during a shorter period thancontrol worker brood cells, and drone lar-

vae in worker cells seemed to be attractive

during a longer period than control dronebrood cells.

The cell type strongly affected the num-ber of mites that invaded. In shortenedworker brood cells, two to three times asmany mites were found per cell compared tothe control cells. In elongated worker broodcells and in drone cells containing a workerlarva 1/6 and 1/2 of the number of mites incontrol cells were found, respectively. Inshortened drone brood cells, one and a halfto two times as many mites were found as incontrol drone brood cells. No significantdifference was found in the number of mites

per cell between worker cells containing adrone larva and control drone brood cells.

The distance between the larva and thecell rim decreased linearly in cells contain-ing a worker larva during the 30 h precedingcell capping. Control worker brood cellswere capped when this distance was about5.5 mm. In elongated worker brood cellsthe same relationship between time beforecapping and distance from larva to cell rimwas found, but this distance was about3 mm more than that of control cells at thesame time before cell capping, correspond-ing to the distance by which the cells hadbeen elongated. In drone cells with a workerlarva, the distance from larva to cell rim wasalso much longer (about 2-3 mm) than incontrol worker brood cells. In artificial con-ical worker brood cells (ANP-comb) witha wider cell bottom [64] this distance was0.5-1 mm more than in control worker cells.In drone brood cells, the distance betweenthe larva and the cell rim remained the same,on average 7 mm, during the 35 h precedingcell capping. Before that period, a decreasein this distance was found.

In general, the distance between larvaand cell rim decreased with time. Hence,the critical distance at which mites begin toinvade brood cells may be estimated by tak-ing the distance found at the beginning ofthe attractive period. Similar values werefound for the following critical distances:

between 6.9 and 7.9 mm for control workerbrood cells, between 7.2 and 7.8 mm for con-trol drone brood cells and 6.9 mm for artifi-cial (ANP) worker cells. However, the criti-cal distances for invasion into elongatedworker brood cells and for invasion into dronebrood cells containing a worker larva wereestimated to be longer (from 8.2-9.0 mm).

The mites probably use a signal comingfrom the larva, such as heat production or theproduction of volatile substances, and thedistance between larva and cell rim mayaffect the strength of the signal as it reachesa mite on a bee. To perceive this signal, thedistance between the mite staying on thebee and the larva may have to be within acritical distance. In elongated worker broodcells and drone cells containing a workerlarva, the critical distance at which invasionstarts was estimated to be greater than incontrol worker brood cells. Since the attrac-tive period was shorter in these cases, thelarva was older when invasion of mites

began. Possibly, the critical distance forinvasion is greater when the larva is older,because the strength of a signal coming fromthe larva may increase with age (Calis etal., unpublished data).

Le Conte et al. [42] claimed that odoursof a few aliphatic esters, especially methylpalmitate (MP), are the signals a mite uses toinvade brood cells. Each of these esters,which had been extracted from the larval

cuticle, attracted mites in an olfactometer.The experiment indicated that these esterscan at least be perceived by the mites.Trouiller et al. [60] extracted a maximumof 17 and 320 ng of MP from the cuticle ofa worker and a drone larva, respectively. In drone larvae the aliphatic esters weresecreted over a longer period preceding cellcapping than in worker larvae [61].

Although these data correlate with thedifferential invasion of worker and dronebrood cells, Boot et al. [6] did not find anincreased attractivity of mites to workerbrood cells after application of 2 μL of ace-tone containing 172, 17.2 or 1.72 ng of MP

per larva. Application of 17.2 and 1.72 ng ofMP, or only acetone, did not affect theattractive period of the cells and the numberof mites per cell. Only in one experimentin which 1.72 ng of MP were applied, wasthe number of invaded mites higher than incontrol cells; however, the length of theattractive period was similar. All larvae diedafter application of 172 ng of MP. Treat-ment with 17.2 ng caused some mortality,and treatment with 1.72 ng or acetone alonedid not cause mortality.

Calis et al. (unpublished data) measuredthe attractiveness of worker larvae of dif-ferent honey bee races and of different ages.Brood combs with eggs 0-1 day of age [7]were produced in colonies of different hon-eybee races and introduced into strongcolonies for nursing the brood. When thelarvae were 6-7 days of age the combs wereplaced into a strong mite-free colony to pre-vent infestation of these cells. After the firstcells had been capped, the combs were intro-duced into the middle of the brood nest of a

heavily infested colony. In contrast to theprevious experiment [8], brood of differentages was simultaneously exposed to miteinvasion over 3 h in a highly infested colony.It was assumed that the density of the miteson the bees did not change within such ashort period. After 3 h, the combs wereremoved from the infested colony, thecapped cells were marked on transparentsheets, and the combs were returned to themite-free colony. Newly capped brood cellswere marked in 3-h intervals. After six toseven intervals the combs were taken fromthe colony and the number of capped cellsper interval and the number of invaded miteswere recorded. After the data of three exper-iments were weighted to the number ofbrood cells and the number of mites it

appeared that the relative numbers of mitesper cell increased with the age of the larva

(figure 2).

Few differences were found betweenraces (Calis et al., unpublished data). Onthe other hand after Büchler [16] introduced

one frame with nine subunits containingdated (1-2 days) worker brood of differentraces or a Buckfast strain in a highly infestedcolony, he found differences in the rate ofinfestation between races. The average infes-tation per brood cell of pieces of brood of thesame size was lowest in A. m. melliferabrood when compared to that of A. m. car-nica or Buckfast brood.

Queen cells are usually not infested bymites. However, when rearing queens(1 500) under different conditions, Hariza-nis [36] found differential rates of infesta-tion. In queen rearing colonies with openand sealed brood, an average of 2 % of thequeen cells were infested. When only sealedbrood or no brood was present the percent-ages of infested queen cells increased to 4and 9 %, respectively. In colonies withoutbrood, only five mites were found with off-spring in queen cells. The oldest offspringwas a mobile protonymph. Because thecapped stage of queen cells is relatively short(8 days) [49], this offspring never could

become adult. The low attractiveness of

queen cells for mites could be due to aweaker attractive signal. Trouiller et al. [62] explain this weak attractivity of queen cellsas follows: queen larvae produce only halfthe amount of methyl palmitate, methyllinolenate and ethyl palmitate, which areattractants for V. jacobsoni [42], as workerlarvae. In addition queen larvae producemuch more methyl oleate, a substance repel-lent to mites, than worker larvae.

4. DOES INVASIONAND REPRODUCTION DEPENDON THE HISTORYOF THE MITES?

The composition of the V. jacobsoni pop-ulation on adult bees varies constantly.Phoretic mites differ in age and in the dura-tion of their stay on adult bees. These mites

may be callow or have reproduced once orseveral times [26]. In addition these mites

may have a different origin owing to trans-fer by drifting of drones or inexperiencedforager bees [35, 54], or by robber bees [53].Some of the mites may have escaped frombrood cells when the bees removed infestedbrood [3-5]. When invading a brood cellfor the second time, oviposition of thesemites had probably already been initiatedbefore or during the first 2 days [1] or thefirst day [59] of their interrupted stay in acapped brood cell which will affect the startand the rate of egg-laying of the mites [25].

Several authors have assumed that youngmites have to mature and old mites have to

prepare for reproduction in a brood cellwhile staying on adult bees [1, 30, 57]. Inthat case, one would expect that the stay onadult bees would affect the moment of inva-sion into a brood cell, the start of ovipos-tion or the reproductive success of the mite.De Ruijter and Pappas [28] collected youngand old mites from brood cells 10 days aftercell capping. Both categories of mites wereintroduced into recently sealed brood cells,either immediately, or after a stay of 1 weekon adult bees. Ten days later, the offspringof both categories of mites had attained amore advanced stage of development whentheir mothers had been in contact with adult

bees; oviposition of these mites began ear-lier than in mites without previous contactwith adult bees. On the other hand, it

appeared that contacts with adult bees bythe mites are not necessary for the initiationof oviposition [26, 28]. In colonies of hybridsbetween Apis mellifera intermissa and intro-duced European races, Beetsma and Zonn-eveld [1] collected swollen mites (in whichthe dorsal and ventral shields were clearlyseparated) and non-swollen mites from adultbees, and introduced them into recentlycapped worker brood cells. The averagenumber of offspring of swollen mites wassimilar to that of naturally invaded mites,but the non-swollen mites produced signif-icantly fewer offspring. Swollen mites mighthave escaped from capped brood cells whichhad been opened by bees and thereforedemonstrate an increased rate of egg lay-

ing. De Ruijter [25] demonstrated this effectwhen transferring mites 24 or 48 h after cellcapping into another series of recentlycapped worker brood cells. In addition, mitestransferred after a stay of 48 h in a cappedbrood cell did not produce male offspring.When Beetsma and Zonneveld [1 ] collectednon-swollen mites from brood cells, andintroduced them into Eppendorf test tubesprovided with a stretched larva (one in theprocess of spinning a cocoon) every 12 days,these mites did not reproduce over 35 days.However, when the swollen mites wereintroduced into recently sealed brood cells,the number of offspring produced was sim-ilar to that of naturally invading mites.Therefore, Beetsma and Zonneveld [1 ] sug-gested that oviposition could be stimulatedboth by a preceding stay on adult bees or ina brood cell in which the mites did not

reproduce.Boot et al. [9] placed a broodless and

mite-free colony in an isolated place to pre-vent reinfestation by mites. They introducedheavily infested emerging brood for I day toprovide the colony with a large number ofyoung and older mites that started their

phoretic phase on the same day. Boot et al.[9] measured invasion of these mites intobrood cells during a maximum of 20 days. Acomb containing 500 worker larvae 3-4 daysof age was introduced daily and removedafter 3 days. Each day all capped cells weremarked on transparent sheets to register theinvasion time of mites. Finally all mitesremaining on adult bees were killed andcounted. With these data the number of

phoretic mites and the number of mites thatinvaded brood cells could be calculated foreach day. In three replicates with colonies ofdifferent sizes, it appeared that mites beganto invade brood cells on the first day of theirphoretic stage and continued to invade broodcells at a constant rate, although this rateand the number of bees differed betweenthe replicates (figure 3).

Previously assumed differences in inva-sion time between, for example, young andold mites could not be demonstrated. Boot et

al. [ 12] also demonstrated that the time spenton adult bees did not affect the fraction ofmites without offspring, the number of off-spring, the number of viable daughters andthe fraction of mites with only male off-spring. On the other hand, Schmidt-Baileyand Fuchs [55] found that the time spent onadult bees increased the trapping efficiencyof drone brood cells. When groups of 50drone brood cells were introduced, 1, 2, 3and 4 weeks after formation of separateinfested broodless nuclei in Kirchhain mat-

ing boxes, their trapping efficiency increased.

5. EFFECT OF THE BROOD/BEERATIO ON INVASION

Explanations for the differences in therate of invasion between the replicates ofthe experiments of Boot et al. [9] becameclear from the results of a similar experi-ment in which the size of the colony or thenumber of brood cells suitable for invasionwas changed. Boot et al. [11] demonstratedthat the rate of invasion increased with thenumber of suitable brood cells and decreasedwith the number of bees. When the surfacearea of suitable brood cells increases, morebees will come close to a brood cell and the

phoretic mites have more opportunities to

leave the bee and enter a brood cell. Conver-

sely, with a mite population of the same size,the density of mites on bees will decreasewith increasing colony size and thereforetheir opportunities to come close to a broodcell will decrease. The rate of invasion alsodecreased when young brood, not yet attrac-tive to mites, was introduced. The additionof brood probably forced the bees to spreadover more combs and therefore fewer miteswere present in the direct vicinity of theattractive brood cells [11].

After the experiments on invasion intoworker brood cells [9], Boot et al. [ 14] stud-ied the invasion into drone brood cells inrelation to the size of the colony using a sim-ilar design. In these experiments, a combcontaining 50 drone larvae 3-4 days of agewas introduced each day. In six replicates itappeared that the rate of invasion of mitesinto drone brood cells was correlated withthe number of drone brood cells per kg ofbees, but not with the duration of their stayon adult bees, similar to the situation inworker brood cells [11]. However, dronebrood cells were invaded 11.6 times more

frequently than worker brood cells. Note, inthese experiments the invasion into worker ordrone brood cells was tested in separatecolonies (cf. the experiments by Fuchs [31]).

Part of this higher frequency of invasionmay be due to the longer attractive period ofdrone cells [8, 38]. When invasion into abrood cell depends on the frequency withwhich a bee brings a mite close enough to abrood cell to invade, the number of mitesthat invade per cell is expected to be two tothree times higher in drone brood cells, pro-vided that the number of mites on the beesremains the same. In addition, when thefrequency with which a bee brings a miteclose enough to a brood cell for invasion isproportional to the surface of a brood cell,1.7 times more mites are expected per dronebrood cell owing to their 1.7 times largersurface. Combining these two factors wouldresult in drone brood cells being invaded3.4-5.1 times more frequently than workerbrood cells. However, it was found thatdrone brood cells were invaded 11.6 timesmore frequently. Thus, the rate of invasionper cell is increased an additional two tofour times by the presence of a drone larvainstead of a worker larva. Martin [43] addeda third factor to explain the higher number ofmites in drone brood cells when compared toworker brood cells. The weights of droneand worker larvae are 346 and 140 mg,respectively, yielding a proportion of 2.47.Including this factor would lead to a range of8.4-12.6 times more frequent invasions indrone brood cells than in worker brood cells.

However, this factor is probably not relatedto a higher number of bee visits to dronebrood cells as suggested by Martin [43],because Boot et al. [10] demonstrated thatmite invasion was not related to feeding orcell capping. It is more likely that the weightof the larva is related to the strength of thesignal causing mite invasion [42].

6. DISTRIBUTIONOF MITES OVER WORKERAND DRONE BROOD CELLS

Schulz [57] and Fuchs [31] found moremites in drone brood cells than in workerbrood cells. De Jong [22], Rozenkranz et

al. [52] and Otten and Fuchs [45] suggestedthat mites prefer drone brood to workerbrood and Schulz [57], Ifantidis [37] andFuchs and Langenbach [32] suggested thatthis preference is due to the higher repro-ductive success of mites in drone broodcells. The larger number of mites generallyfound in drone brood cells is thought bythese authors to be the result of selection of’drone brood mites’. This preference, how-ever, could not be found in individual mites.When Radtke [47] collected adult mitesfrom worker and drone brood cells, markedeach group of mites distinctly, and intro-duced them into a colony, he found no indi-cation of a selection of ’worker brood mites’or ’drone brood mites’. In fact, he recov-ered about the same numbers of mites inboth categories of marked mites in workerand drone brood cells that had been cappedin the same period.

The differential distribution of mites overdifferent ratios of drone and worker broodcells within one colony can be calculatedaccording to Boot et al. [ 14] using only therelative rates of invasion per drone andworker brood cell per day and the numbersof both brood cell types without making fur-ther assumptions. The relative rates of inva-sion per brood cell per day are 0.56 and 6.49for worker and drone brood, respectively.These values are the result of all possiblefactors that affect brood cell invasion.

Fuchs [31] studied the invasion of workerand drone brood cells simultaneously in thesame colony. He carried out this experimentin 68 colonies containing only one combwith worker brood and one comb with dronebrood. The numbers of worker and dronebrood cells varied between the colonies from

mainly worker brood to mainly drone brood.After all brood cells had been capped, Fuchs[3 1 ] counted the number of mites thatinvaded the two types of brood cells. The

relationship between the percentage of themites in drone brood cells and the percent-age of drone brood cells in the experimentsof Fuchs [31] is presented in figure 4. The

average number of mites per drone broodcell was 8.3 times higher than that perworker brood cell. This distribution (dronebrood cell preference, cf. Fuchs [31]) wasnot affected by the rate of infestation of thecolony nor by the total number of availablebrood cells. However, the distribution wasaffected by the percentage of drone broodcells. The average percentage of mites perdrone brood cell was 12.1 times higher thanthat per worker brood cell when the per-centage of drone brood cells varied between5 and 15 (situation found in untreated colo-nies).

On the basis of the data provided by DrFuchs, nearly the same relationship betweenthe percentage of mites in drone brood cellsto the percentage of drone brood cells couldbe predicted using only the relative rates ofinvasion into worker and drone brood [14],and the numbers of worker and drone brood

cells provided by Dr Fuchs (figure 4). Theobserved distributions were congruent withthe theory on invasion [14] which assumedthat invasions of worker and drone broodcells are independent events. Or, the decisionof the mites to invade a brood cell is deter-mined by the signal they receive from thebrood cell in their direct vicinity.

7. CONCLUSION

Although many aspects of invasionbehaviour have been revealed, it is stillunclear which substances the mites areattracted to when invading a brood cell.These attractive substances could differ in

quantity or even in quality between workerand drone larvae. In addition mites of pop-ulations of different origin (East Russia orJapan) could respond differently to thesesubstances. Invasion behaviour of mites in

A. mellifera and in A. cerana colonies can-not yet be compared because few data areavailable on the behaviour of the Asian mitein colonies of its original host. The results sofar obtained provide possibilities for furtherstudies. The estimation of the relative ratesof invasion per day per worker and dronebrood cell has made it possible to answerthe question as to why it is advantageousfor the mite to invade both worker and dronebrood cells while reproductive success indrone brood cells is higher [15].

The population growth of V. jacobsonidepends entirely on that of the honey beecolony. Since it is known that the rate ofinvasion of the mite depends on the size ofthe colony and the number of worker anddrone brood cells suitable for mite invasion,simulation models of the mite populationcould be improved [21, 43].

V. jacobsoni can effectively be trappedwhen using large numbers of worker broodcells. The finding that mites invade dronebrood cells in larger numbers than workerbrood cells [14, 31, 57] inspired severalauthors [19, 51, 58] to develop biotechni-cal control methods in which mites are

trapped in drone brood combs that are sub-sequently removed from the colony. Fromthe experiments on the process of invasion,it follows that in a colony of given size thenumber of phoretic mites that can be trappeddepends mainly on the number of cells usedfor trapping. The methods developed byCalis et al. [19] are already effective withrelatively small amounts of drone broodcells. Boot et al. [14] calculated that in abroodless colony of 1 kg of bees only 462drone brood cells are needed to trap 95 % ofthe mites. To obtain the same result, how-ever, 5 357 worker brood cells would beneeded. The principle of trapping mites inbroodless colonies with drone brood has ledto the development of several biotechnicalcontrol methods [18, 56]. Without tests inthe field, the effectiveness of biotechnicalcontrol methods can now be predicted usingthe simulation model developed by Calis etal. [20].

ACKNOWLEDGEMENT

We thank Dr S. Fuchs for kindly providingthe data he collected in 1987.

Résumé - Le comportement d’invasionde Varroa jacobsoni Oud. : des abeillesaux cellules de couvain. L’acarien Varroa

jacobsoni (Acari : Varroidae) peut envahirles cellules de couvain d’ouvrières (cco) oude mâles (ccm) d’abeilles (Apis mellifera L.)lorsque les ouvrières le mettent en contactétroit avec ces cellules. Les acariens pas-sent du flanc de l’abdomen de l’abeille surle rayon et pénètrent immédiatement dansla cellule de couvain adjacente, grimpent àla surface de la larve et se glissent entre lalarve et la paroi de la cellule jusqu’au fondde celle-ci. On n’a jamais vu d’acariens sedéplacer sur le rayon ni entrer dans des cel-lules de couvain et en ressortir pour choisirune cellule à envahir [10].À l’aide de « demi-rayons », Boot et al. [8]ont pu enregistrer le moment où un acarienapparaissait au fond de la cellule transpa-rente et le moment où la cellule était oper-culée. L’invasion des acariens dans les ccos’est produite entre 15 et 20 h avant l’oper-culation de la cellule, alors que dans les ccmelle a commencé 45 à 50 h avant l’opercu-lation. La période d’attractivité des ccm estdonc 2 à 3 fois plus longue que celle des cco.Boot et al. [13] ont mesuré i) la périoded’attractivité des cellules de couvain, ii) larépartition des acariens sur les différentstypes de cellules, et iii) la distance entre lalarve et le bord de la cellule de divers typesde cellules, en relation avec la durée qui aprécédé l’operculation. La période d’attrac-tivité des cco et des ccm raccourcies a été

plus longue et on y a trouvé 1,5 à 3 fois plusd’acariens que dans les cellules témoins. La

période d’attractivité des cco allongées aété plus courte et elles ne renfermaient qu’unsixième des acariens présents dans les cel-lules témoins. Les cellules de mâles avecune larve d’ouvrière semblent avoir eu une

période d’attractivité plus courte et ne ren-

fermaient que la moitié des acariens trouvésdans les cco témoins. Les cellules d’ouvrièresavec une larve de mâle semblent avoir euune période d’attractivité plus longue, maisle nombre d’acariens ne différait pas de celuitrouvé dans les ccm témoins.

L’attractivité des cellules de couvain sembleêtre en rapport avec la distance entre la larveet le bord de la cellule. On a estimé la dis-tance critique à laquelle les acariens enva-hissaient les cellules de couvain, en mesurantcette distance au début de la période d’attrac-tivité. Dans les cco allongées et dans lesccm enfermant une larve d’ouvrière, la dis-tance critique était plus grande que dans lescco témoins. Puisque la période d’attractivitéétait plus courte dans ces deux cas là, leslarves étaient plus âgées lorsque les acariensont commencé à envahir les cellules.À partir de colonies indemnes d’acariens,Calis et al. [20] ont obtenu des rayons decouvain daté. Lorsque les premières cellulesde couvain ont commencé à être operculées,les rayons ont été placés durant 3 h dansune colonie fortement infestée, puis remisdans les colonies indemnes. L’operculationdes cellules de couvain a été observée toutesles 3 h. Le nombre relatif d’acariens qui enva-hissaient les cellules a augmenté avec l’âgede la larve.

Boot et al. [9] ont introduit un grand nombred’acariens jeunes et vieux dans une colonieindemne et les ont laissé une journée. Lesacariens ont commencé à envahir les cel-lules de couvain au premier jour du stadephorétique et ont continué à le faire à untaux constant, bien que ce taux et le nombred’abeilles aient varié d’une répétition àl’autre. Quand on introduisait chaque jour ungrand nombre de cellules de couvain, lamême proportion d’acariens phorétiquesenvahissait chaque jour une cellule de cou-vain. L’invasion d’une cellule de couvainn’était pas liée à la durée de leur séjour surles abeilles adultes. Le temps passé sur lesabeilles adultes n’influence donc pas lesuccès reproductif [12].Le taux d’invasion est déterminé par le rap-port du nombre de cellules de couvain appro-

priées à la taille de la colonie. Plus le nombrede cellules de couvain attractives est grand,plus grand est le taux d’invasion et plus lataille de la colonie est grande, plus faible estla densité des acariens sur les abeilles et,donc, plus faible est le taux d’invasion. Bootet al. [ 14] ont établi ces rapports en utilisantdes colonies sans couvain dans lesquelles ilsintroduisaient chaque jour un nombre connude cellules de couvain de mâles daté. Lesccm ont été envahies 1 1,6 fois plus souventque les cco.Boot et al. [14] ont calculé la répartition desacariens dans une colonie sur le couvaind’ouvrières et le couvain de mâles. La répar-tition est déterminée par les taux spécifiquesd’invasion par jour et les nombre de cel-lules des deux types. Les répartitions obser-vées par Fuchs [34] concordent avec la théo-rie sur l’invasion [14]. L’étude du comportement d’invasion deV. jacobsoni a permis de mettre au point desméthodes de lutte biotechniques efficaces,qui utilisent le piégeage du couvain de mâlesdans des colonies sans couvain [21, 58].Calis et al. [18] ont mis au point un modèlede simulation qui permet de prédire l’effi-cacité des méthodes biotechniques. Avec cemodèle il n’est plus nécessaire de tester lesvariations de ces méthodes au champ.© Inra/DIB/AGIB/Elsevier, Paris

Apis mellifera / Varroa jacobsoni / com-portement d’invasion / lutte biotechnique

Zusammenfassung - Von den Bienen indie Brutzellen: Ein Überblick über dasEindringverhalten von Varroa jacobsoni.Sobald Arbeiterinnen die Varroamilben innahen Kontakt zu Arbeiterinnen - oderDrohnenbrutzellen brachten, konnten diesein sie eindringen. Die Milben bewegten sichhierbei von der Seite des Abdomens derBiene auf die Wabe, wobei sie sofort in einenahegelegene Brutzelle eindrangen. Dannliefen sie auf der Oberfläche der Larveherum und krochen zwischen der Larve undder Zellwand hindurch bis zum Boden der

Zelle. Es wurde niemals beobachtet, daß dieMilben auf der Wabenoberfläche herumlie-fen oder daß sie Zellen betraten und wie-der verließen um eine Zelle für den Befallauszusuchen [10].Durch die Verwendung von ’Halbwaben’konnten Boot et al. [8] feststellen, zu wel-chem Zeitpunkt die Milben auf dem durch-sichtigen Zellboden erschienen und wanndiese Zellen verdeckelt wurden. In ver-schiedenen Völkern begann der Befall derArbeiterinnenbrutzellen 15-20 Stunden, derder Drohnenbrutzellen 40-50 Stunden vorder Verdeckelung der Zellen. Der Zeitraum,in dem die Zellen attraktiv wirken, ist daherbei den Drohnenbrutzellen 2-3 mal längerals bei den Arbeiterinnenbrutzellen.

Boot et al. [13 ] bestimmten: 1) die attraktiveZeit der Brutzellen, 2) die Verteilung derMilben zwischen verschiedenen Brutzellen,und 3) den Abstand zwischen der Larve unddem Zellrand bei den verschiedenen Zell-

typen in Beziehung zu der Zeit bis zur Zell-verdeckelung. Die attraktive Zeit gekürzterArbeiterinnen - und Drohnenbrutzellen war

verlängert, dies führte zu im Vergleich zuKontrollzellen 1,5 bis 2 mal höheren Anzah-len von Milben in diesen Zellen. Demge-genüber war die attraktive Zeit von verlän-gerten Zellen kürzer, in diesen Zellen fandensich im Vergleich zu Kontrollzellen nur 1/6der Milbenzahlen. Drohnenzellen, die eineArbeiterinnenlarve enthielten, wiesen einekürzere attraktive Zeit auf und enthieltenhalb so viele Milben wie die Kontrollgruppevon Arbeiterinnenbrutzellen. Arbeiterin-nenbrutzellen mit einer Drohnenlarve wareneine längere Zeit attraktiv, der Befall mitMilben unterschied sich aber nicht von den

Kontrollgruppe mit Drohnenzellen.Die Attraktivität der Brutzellen schien mitdem Abstand zwischen der Larve und demZellrand zusammenzuhängen. Der kritischeAbstand, ab dem die Milben in die Brut ein-zudringen beginnen, kann zu Beginn derattraktiven Periode bestimmt werden. In ver-

längerten Arbeiterinnenbrutzellen und inDrohnenzellen, die eine Arbeiterinnenlarveenthielten, war dieser gegenüber den Kon-

trollzellen verlängert. Da die attraktive Zeitin diesen Fällen verkürzt war, waren dieLarven zum Zeitpunkt des Beginnes desZellbefalls bereits älter.

Calis et al. [20] entnahmen Waben mitdatierter Brut aus milbenfreien Völkern.Sobald die Verdeckelung der ersten Zellenbegann, wurden diese drei Stunden lang inhochbefallene Völker eingebracht unddanach wieder in die milbenfreien Völker

zurückgebracht. Danach wurde in Abstän-den von drei Stunden die Verdeckelung derZellen verfolgt. Offensichtlich nahmen mitsteigendem Alter der Larve die relativenAnzahlen von Milben zu, die die Zellenbefallen hatten.

Boot et al. [9] setzten eine große Anzahljüngerer und älterer Milben in ein milben-freies Volk ein. Die Milben begannen bereitsam ersten Tag ihres phoretischen Stadiumsdie Brutzellen zu befallen, danach drangensie mit konstanter Befallsrate in die Brut-zellen ein. Diese Befallsrate war zwischenden Versuchen allerdings unterschiedlich.Sobald jeden Tag die gleiche Anzahl vonBrutzellen eingebracht wurde, befiel jedenTag ein gleichbleibender Anteil der Milbendie Brut. Der Befall der Zellen stand nichtmit dem Alter der Milben im Zusammen-

hang. Die Zeit, die diese auf Bienenarbei-terinnen verbrachten, stand in keinemZusammenhang mit ihren Reproduktions-erfolg [12].Die Befallsrate wird von dem Verhältniszwischen geeigneten Brutzellen und derGröße des Volkes bestimmt. Je größer dieAnzahl attraktiver Brutzellen ist, desto höherist auch die Befallsrate. Je größer das Volkist, desto geringer ist die Dichte der Varro-amilben auf den Bienen und daher auch dieBefallsrate. Boot et al. [14] ermittelten die-sen Zusammenhang, indem sie täglich einebestimmte Menge von Drohnenbrutzellenin ansonsten brutfreie Völker einbrachten.Hieraus bestimmten sie, daß Drohnenbrut-zellen 11,6 mal mehr befallen werden alsArbeiterinnenbrutzellen.

Boot et al. [14] berechneten die Verteilungvon Milben zwischen Arbeiterinnenbrut und

Drohnenbrut in einem Volk. Diese Vertei-

lung wird von den spezifischen Befallsra-ten beider Bruttypen bedingt. Die von Fuchs[34] beobachten Verteilungen stimmten mitdieser Theorie des Befallsverhaltens überein

[14].Die Untersuchung des Befallsverhaltens vonBrut durch Varroa jacobsoni ermöglichtedie Entwicklung wirksamer biotechnischerBekämpfungsmethoden unter Verwendungvon Fangwaben in brutlosen Völkern [29,58]. Calis et al. [18] entwickelten ein Simu-lationsmodell, das Voraussagen der Wirk-samkeit dieser Methoden ermöglicht. Hier-durch ist es nicht länger nötig, Variantendieser Methoden in Feldversuchen zu testen.© Inra/DIB/AGIB/Elsevier, Paris

Apis mellifera / Varroa jacobsoni /

Eindringverhalten / Brutzellentyp /Biotechnische Bekämpfung

REFERENCES

[1] Beetsma J., Zonneveld K., Observations on theinitiation and stimulation of oviposition of theVarroa mite, Exp. Appl. Acarol. 16 (1992)303-312.

[2] Beetsma J., Boot W.J., Calis J.N.M., Specialtechniques to make detailed observations of Var-roa mites staying on bees and invading honeybeebrood cells, in: Connor L.J., Rinderer T.,Sylvester H.A., Wongsiri S. (Eds.), Asian Api-culture, Wicwas Press, Cheshire CT, USA, 1993,pp. 510-520.

[3] Boecking O., Removal behavior of Apis mellif-era towards sealed brood cells infested with Var-

roa jacobsoni: techniques, extent and efficacy,Apidologie 23 (1992) 371-373.

[4] Boecking O., Drescher W., The reaction ofworker bees in differentApis mellifera coloniesto Varroa infested brood cells, in: Proceedings ofthe International Symposium on Recent Researchon Bee Pathology, 1990, Gent, Belgium,pp. 41-42.

[5] Boecking O., Drescher W., Response of Apismellifera colonies to brood cells infested withVarroa jacobsoni Oud., Apidologie 22 (1991)237-241.

[6] Boot W.J., Methyl palmitate does not elicit inva-sion of honeybee brood cells by Varroa mites,Exp. Appl. Acarol. 18 (1994) 587-592.

[7] Boot W.J., Calis J.N.M., A method to obtaindated brood in honeybee colonies, Bee World72 (1991) 19-21.

[8] Boot W.J., Calis J.N.M., Beetsma J., Differentialperiods of Varroa mite invasion into worker anddrone cells of honey bees, Exp. Appl. Acarol.16 (1992) 295-301.

[9] Boot W.J., Calis J.N.M., Beetsma J., Invasionof Varroa jacobsoni into honey bee brood cells:a matter of chance or choice?, J. Apic. Res. 32(1993) 167-174.

[10] Boot W.J., Beetsma J., Calis J.N.M., Behaviourof Varroa mites invading honey bee brood cells,Exp. Appl. Acarol. 18 (1994) 371-379.

[11] Boot W.J., Sisselaar D.J.A., Calis J.N.M.,Beetsma J., Factors affecting invasion of Var-roajacobsoni (Acari: Varroidae) into honeybee,Apis mellifera (Hymenoptera: Apidae), broodcells, Bull. Entomol. Res. 84 (1994) 3-10.

[12] Boot W.J., Calis J.N.M., Beetsma J., Does thetime spent on adult bees affect reproductive suc-cess of Varroa mites?, Entomol. Exp. Appl. 75(1995) 1-7.

[13] Boot W.J., Driessen R.G., Calis J.N.M., Beetsma J.,Further observations on the correlation betweenattractiveness of honey bee brood cells to Varroajacobsoni and the distance from larva to cell rim,Entomol. Exp. Appl. 76 (1995) 223-232.

[14] Boot W.J., Schoenmaker J., Calis J.N.M.,Beetsma J., Invasion of Varroa jacobsoni intodrone brood cells of the honey bee, Apis mellif-era, Apidologie 26 (1995) 109-118.

[15] Boot W.J., Van Baalen M., Sabelis M.W., Whydo Varroa mites invade worker brood cells ofthe honey bee despite lower reproductive suc-cess?, Behav. Ecol. Sociobiol. 36 (1995)283-289.

[16] Büchler R., Attractivity and reproductive suit-ability for the Varroa-mite of bee brood fromdifferent origin, in: Cavalloro R. (Ed.), PresentStatus of Varroatosis in Europe and Progress inthe Varroa Mite Control, Office for Official Pub-lications of the European Communities, Lux-embourg, 1989, pp. 139-145.

[17] Calis J.N.M., Boot W.J., Beetsma J., Behavior ofVarroa mites invading honey bee brood cells ofdifferent sizes, in: Connor L.J., Rinderer T.,Sylvester H.A., Wongsiri S. (Eds.), Asian Api-culture, Wicwas Press, Cheshire CT, USA(1993), pp. 499-509.

[18] Calis J.N.M., Boot W.J., Beetsma J., Van denEijnde J.H.P.M., De Ruijter A., Van der SteenJ.J.M., Trapping Varroa jacobsoni in dronebrood combs of Apis mellifera in broodlesscolonies, Apidologie 27 (1996) 295.

[19] Calis J.N.M., Schmidt-Bailey J., Beetsma J.,Boot W.J., Van den Eijnde J.H.P.M., Fuchs S.,De Ruijter A., Van der Steen J.J.M., Successfultrapping of Varroa jacobsoni with drone broodin broodless Apis mellifera colonies, Apiacta 32(1997) 65-71.

[20] Calis J.N.M., Boot W.J., Beetsma J., Model eval-uation of methods for Varroa jacobsoni mitecontrol based on trapping in honey bee brood,Apidologie 30 (1999) 197-207.

[21] Calis J.N.M., Fries I., Ryrie S.C., Populationmodelling of Varroa jacobsoni Oudemans, Api-dologie 30 (1999) 111-124.

[22] De Jong D., Current knowledge and open ques-tions concerning reproduction in the honey beemite Varroa jacobsoni, in: Engels W., ClarkW.H. Jr, Fisher A., Olive P.J.W., Went D.F.(Eds.), Adv. Invertebr. Reprod. 3 (1984)547-552.

[23] De Jong D., Morse R.A., Utilisation of raisedbrood cells of the honey bee, Apis mellifera(Hymenoptera: Apidae), by the mite, Varroajacobsoni (Acarina: Varroidae), Entomol. Gen.14 (1988) 103-106.

[24] De Jong D., Morse R.A., Eickwort G.C., Mitepests of honey bees, Annu. Rev. Entomol. 27(1982) 229-252.

[25] De Ruijter A., Preference of Varroa jacobsoniOudemans for different cell types and some fac-tors affecting reproduction, in: ProceedingsXXXth International Apicultural Congress,Nagoya, Japan, Apimondia Publ. House,Bucharest 1985, pp. 165-167.

[26] De Ruijter A., Reproduction of Varroa jacob-soni during successive brood cycles of the hon-eybee, Apidologie 18 (1987) 321-326.

[27] De Ruijter A., Calis J.N.M., Distribution of Var-roa jacobsoni female mites in honey bee workerbrood cells of normal and manipulated depth(Acarina: Varroidae), Entomol. Gen. 14 (1988)107-109.

[28] De Ruijter A., Pappas N., Karyotype and sexdetermination of Varroa jacobsoni Oud., in:Cavalloro R. (Ed.), Varroa jacobsoni Oud.Affecting Honey Bees: Present Status and Needs,A.A. Balkema, Rotterdam, 1983, pp. 41-44.

[29] Elzen P.J., Eischen F.A., Baxter J.B., Pettis J.,Elzen G.W., Wilson W.T., Fluvalinate resistancein Varroa jacobsoni from several geographiclocations, Am. Bee J. 138 (1998) 674-676.

[30] Fuchs S., Untersuchungen zur QuantitativenAbschätzung des Befalls von Bienenvölkern mitVarroa jacobsoni Oudemans und zur Verteilungdes Parasiten im Bienenvolk, Apidologie 16(1985) 343-368.

[31] Fuchs S., Preference for drone brood cells byVarroa jacobsoni Oud. in colonies of Apis mel-lifera carnica, Apidologie 21 (1990) 193-199.

[32] Fuchs S., Langenbach K., Multiple infestationof Apis mellifera L. brood cells and reproduc-tion in Varroa jacobsoni Oud., Apidologie 20(1989) 257-266.

[33] Fuchs S., Müller K., Invasion by honeybee broodcells by Varroa jacobsoni in relation to the ageof the larva, in: Cavalloro R. (Ed.), EuropeanResearch on Varroatosis Control, A.A. Balkema,Rotterdam, 1988, pp. 77-79.

[34] Goetz B., Koeniger N., The distance betweenlarva and cell opening triggers broodcell invasionby Varroa jacobsoni, Apidologie 24 (1993)67-72.

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

[36] Harizanis P.C., Infestation of queen cells by themite Varroa jacobsoni, Apidologie 22 ( 1991 )533-538.

[37] Ifantidis M.D., Parameters of the populationdynamics of the Varroa mite on honeybees,J. Apic. Res. 22 (1984) 200-206.

[38] Ifantidis M.D., Some aspects of the process ofVarroa jacobsoni mite entrance into honey bee(Apis mellifera) brood cells, Apidologie 19(1988) 387-396.

[39] Ifantidis M.D., Rosenkranz P., Reproduktion derBienenmilbe Varroa jacobsoni (Acarina: Var-roidae), Entomol. Gen. 14 (1988) 111-122.

[40] Jacobs F., Bruneau E., Trouiller J., Resultatenvan de campagne voor de opsporing van resisten-tie van Varroa jacobsoni t.o.v. pyrethrinoïdenin België, Maandblad Vlaamse Imker Bond 83(1997) 369-372.

[41] Kraus B., Koeniger N., Fuchs S., Unterschei-dung zwischen Bienen verschiedenen Altersdurch Varroa jacobsoni Oud. und Bevorzugungvon Ammenbienen im Sommerbienenvolk,Apidologie 17 (1986) 257-266.

[42] Le Conte Y., Arnold G., Trouiller J., Masson C.,Chappe B., Ourisson G., Attraction of the para-sitic mite Varroa to the drone larvae of honeybees by simple aliphatic esters, Science 245(1989) 638-639.

[43] Martin S., A population model for the ectopara-site Varroa jacobsoni in honey bee Apis mellif-era) colonies, Ecol. Model. 109 (1998) 267-281.

[44] Milani N., The resistance of Varroa jacobsoniOudemans to acaricides, Apidologie 30 (1999)229-234.

[45] Otten C., Fuchs S., Individual differences in Var-roa jacobsoni of preference for drone larvae toworker bee larvae, in: Cavalloro R. (Ed.), Euro-pean Research on Varroatosis Control,A.A. Balkema, Rotterdam, 1988, pp. 69-71.

[46] Ramon H., Van Laere O., Ifantidis M., Size ofcomb cell and reproduction of Varroa jacob-soni, in: Connor L.J., Rinderer T., Sylvester H.A.,Wongsiri S. (Eds.), Asian Apiculture, WicwasPress, Cheshire CT, USA, 1993, pp. 521-529.

[47] Radtke J., Zur Frage individueller Bevorzugungvon Drohnen- bzw. Arbeiterbrut durch Varroa

jacobsoni Oud., Apidologie 28 (1997) 188-190.[48] Rath W., Aspects of preadaptation in V. jacob-

soni while shifting from its original host ceranato A. mellifera, in: Connor L.J., Rinderer T.,Sylvester H.A., Wongsiri S. (Eds.), Asian Api-culture, Wicwas Press, Cheshire CT, USA, 1993,pp. 417-426.

[49] Rehm S.M., Ritter W., Sequence of the sexes inthe offspring of Varroa jacobsoni and the result-ing consequences for the calculation of the devel-opment period, Apidologie 20 (1989) 339-343.

[50] Ritter W., Varroa disease of the honeybee Apismellifera, Bee World 62 (1981) 41-153.

[51] Rosenkranz P., Engels W., KonsequenteDrohnenbrutentnahme, eine wirksame biotech-nische Massnahme zur Minderung van Var-roatose-Schäden an Bienenvölkern, Allg. Dtsch.Imkerztg. 19 (1985) 265-271.

[52] Rosenkranz P., Tewarson N.C., Engels W., Opti-mal host selection by reproductive female Var-roa jacobsoni, in: Engels W., Clark W.H. Jr,Fisher A., Olive P.J.W., Went D.F. (Eds.), Adv.Invertebr. Reprod. 3 (1984) 628.

[53] Sakofsky F., Transfer of Varroa jacobsoni byrobbing, in: Cavalloro R. (Ed.), Present statusof Varroatosis in Europe and progress in the var-roa mite control, Office for Official Publicationsof the European Communities, Luxembourg,1989, pp. 177-181.

[54] Sakofsky F., Koeniger N., Natural transfer ofVarroa jacobsoni among honeybee colonies inautumn, in: Cavalloro R. (Ed.), EuropeanResearch on Varroatosis Control, A.A. Balkema,Rotterdam, Brookfield, 1988, pp. 81-83.

[55] Schmidt-Bailey J., Fuchs S., Versuche zur Fang-effizienz von Varroa jacobsoni mit Drohnen-brut-Fangwaben, Apidologie 28 (1997) 184-186.

[56] Schmidt-Bailey J., Fuchs S.. Büchler R., Effec-tiveness of drone brood trapping combs in brood-less honeybee colonies, Apidologie 27 (1996)294.

[57] Schulz A.E., Reproduktion und Populationsen-twicklung der parasitischen Milbe Varroa jacob-

soni Oud. in Abhängigkeit vom Brutzyclus ihresWirtes, Apis mellifera L., Apidologie 15 (1984)401-420.

[58] Schulz A.E., Koeniger N., Ruttner F., Drohnen-brut als Varroafalle, Allg. Dtsch. Imkerztg. 17(1983) 52-54.

[59] Steiner J., Dittmann F., Rosenkranz P., Engels W.,The first gonocycle of the parasitic mite Varroajacobsoni) in relation to preimaginal develop-ment of its host, the honey bee (Apis melliferarcarnica), Invertebr. Reprod. Dev. 25 (1994)175-183.

[60] Trouiller J., Arnold G., Le Conte Y., Masson C.,Temporal pheromonal and kairomonal secretionin the brood of honeybees, Naturwissenschaften78 (1991) 368-370.

[61] Trouiller J., Arnold G., Chappe B., Le Conte Y.,Masson C., Semichemical basis of infestationof honey bee brood by Varroa jacobsoni,J. Chem. Ecol. 18 (1992) 2041-2053.

[62] Trouiller J., Arnold G., Chappe B., Le Conte Y.,Billion A., Masson C., The kairomonal estersattractive to the Varroa jacobsoni mite in thequeen brood, Apidologie 25 (1994) 314-321.

[63] Wallner K., Varroacides and their residues inbee products, Apidologie 30 (1999) 235-248.

[64] Wieting J., Ferenz H.J., Behavioral study on theinvasion of honeybee brood by the mite Varroajacobsoni on wax combs and ANP combs, Am.Bee J. 131 (1991) 117-118.