At present, infested colonies are treated withchemical products which give a certain degreeof mite control. However in the long term,the use of miticides can cause a number ofserious problems: (1) mite populations

1. INTRODUCTION

The mite Varroa destructorAndersonand Trueman is the most serious threat thatthe beekeeping industry faces worldwide.

Original article

Relative effect of four characteristics that restrainthe population growth of the mite Varroa destructor

in honey bee (Apis mellifera) colonies

Miguel E. ARECHAVALETA-VELASCOa*, Ernesto GUZMÁN-NOVOAb

a Department of Entomology, Purdue University, West Lafayette, IN, 47907, USAb CENIFMA-INIFAP, Santa Crúz # 29-B, Las Hdas, Metepéc, Méx, Mexico

(Received 25 July 2000; revised 21 December 2000; accepted 6 February 2001)

Abstract – This study was conducted to determine the existence of phenotypic and genotypic vari-ation in the ability of honey bee colonies to restrain the population growth of the mite Varroa destruc-tor Anderson and Trueman, and to asses the relative effect of four characteristics that may confer tol-erance to honey bees toward the mite. Fifty-eight colonies infested with an equal number of mites weresampled monthly during six months to determine their levels of infestation on adult bees and inworker brood. At the end of this period, 16 colonies were selected to study the effect of groomingbehavior, hygienic behavior, brood attractiveness, and host-induced non-reproduction. The infesta-tion-levels in adult bees varied significantly between colonies (range: 6.6–44.7%), but no differ-ences were found in the brood infestation levels. The variation between colonies was partially geneticin origin. Grooming behavior explained most of the variation (r2 = 0.38). Negative correlations werefound between the mite population growth and both the total number of mites and the number of injuredmites collected from the bottom-boards (r = –0.65 and r = –0.76, respectively). Differences were foundfor hygienic behavior but the effect of this mechanism was not clear. No differences were foundamong colonies for brood attractiveness, or for the effect of the brood on the mite’s reproduction.

Apis mellifera / Varroa destructor / tolerance / mechanisms of resistance / Mexico / groomingbehavior / hygienic behavior

Apidologie 32 (2001) 157–174 157© INRA/DIB-AGIB/EDP Sciences, 2001

* Correspondence and reprintsE-mail: miguel@entm.purdue.edu

M.E. Arechavaleta-Velasco, E. Guzmán-Novoa158

are able to develop resistance to chemicalproducts (Lodesani et al., 1995; Hillesheimet al., 1996; Elzen et al., 1999); (2) miti-cides can leave chemical residues in honeyand wax (Faucon and Flamiini, 1990;Slabezki et al., 1991; Lodesani et al., 1992;Wallner, 1999); and (3) treatments withchemical miticides increase the cost ofhoney production.

The complete eradication of V. destruc-tor is impossible. However, the beekeepingindustry needs ways of maintaining pro-ductive colonies with low levels of infesta-tion. Because of the serious disadvantages ofchemical products, it is necessary to developalternative methods of controlling the mite.One option is the development of Varroa-tolerant honey bee (Apis melliferaL.)strains. The development of genotypes thatare able to maintain low levels of mite infes-tation would allow beekeepers to keephealthier and productive colonies and woulddecrease the risks and costs associated withthe use of chemical miticides. This goalwould be feasible if mechanisms of toler-ance to the mite exist among honey bee pop-ulations, if there is variation for these mech-anisms, and if these mechanisms areheritable (Guzmán-Novoa and Correa,1996).

V. destructoris not a serious pest for ApisceranaFabr., its original host. This honeybee species has developed mechanisms oftolerance, as a result of natural selectionthrough a long association with the mite(Peng et al., 1987). In the case of A. mellifera,there are reports of colonies that survived inareas that were seriously damaged byV. destructor, which suggests the existenceof a certain degree of tolerance in some geno-types of honey bees (Engels et al., 1986;Kulincevic and Rinderer, 1986; Kulincevicand Rinderer, 1988; Moosbeckhofer et al.,1988; Wallner, 1990; Morse et al., 1991). InBrazil it has been shown that the climate andthe race of the bees (European or African-ized), has an influence on the colonies’degree of infestation. Warmer climates and

Africanized bees restrain the development ofhigh levels of mite infestation (Morettoet al., 1991). Additionally, different V.destructorgenotypes have been found toparasitize A. mellifera (De Guzman et al.,1997; Anderson, 2000). These genotypesmay represent mite populations with dif-ferent virulence, which could also explainwhy some populations of honey bees sur-vive and maintain low levels of V. destruc-tor infestations.

One defense mechanism against the mitein A. ceranais grooming behavior. A workerbee is able to groom herself with her legsand mandibles to get rid of a mite. If shecannot get rid of it, she performs a dance toattract other workers that may help her toremove the mite from her body (Peng et al.,1987).

Grooming behavior has been measuredwith indirect methods in A. mellifera, butwas observed at a lower frequency thanin A. cerana(Morse et al., 1991; Ruttnerand Hänel, 1992; Boecking et al., 1993;Rosenkranz et al., 1997; Bienefeld et al.,1999). There are reports of mites that appar-ently were injured by worker bees with theirmandibles (Wallner, 1990; Morse et al.,1991; Moosbeckhofer, 1992; Ruttner andHänel, 1992; Boecking et al., 1993; Morettoet al., 1993; Rosenkranz et al., 1997). In thestudy conducted by Moosbeckhofer (1992),a negative correlation between the numberof injured mites and the total level of infes-tation in the colonies was found, suggest-ing that grooming behavior was responsi-ble for lower levels of mite infestation. In thestudy by Moretto et al. (1993) it was foundthat Africanized bees eliminated more mitesfrom their bodies than European bees andthat the heritability of this behavior was0.71, which suggests that breeding for highergrooming behavior may be possible.

Worker bees of A. ceranaalso have theability to detect capped brood that is infestedby V. destructor; the bees open infested cellsto remove mites. This behavior is known ashygienic or removal behavior (Peng et al.,

Honey bee tolerance to Varroa

1994; Harbo and Hoopingarner, 1997;Harbo and Harris, 1999).

In spite of the fact that the above mech-anisms of mite tolerance have been identi-fied in different populations of honey bees,the relative contribution of each of them tothe overall mite tolerance is not clear. Todevelop a breeding program to select honeybee genotypes that restrain the growth ofthe mite population, it is important to deter-mine if there is genetic variation for thistrait, and to determine the relative contri-bution of each of the mechanisms responsi-ble for the tolerance. This information wouldfacilitate the selection of honey bees basedon the mechanisms that have a larger con-tribution to the mite tolerance.

The objectives of this study were (1) todetermine if Mexican honey bees vary intheir ability to restrain the growth ofV. destructorpopulation, (2) to decomposethis variation into genotypic and environ-mental effects, and (3) to determine the rel-ative contribution of four characteristics thatare thought to confer tolerance to honeybees toward V. destructor.

2. MATERIALS AND METHODS

2.1. Establishmentof the experimental colonies

The experiments were conducted in Vallede Bravo, State of Mexico (19°14’ N;100°06’ W), in the central region of Mexico.Fifty-eight colonies were established injumbo-size hives in the same apiary. Eachcolony was established with four frames ofbrood and 2 kg of adult bees. A markedqueen was randomly introduced into eachcolony. The queens were obtained fromseven different breeders (lines) whom donot exchange breeding stock and are geo-graphically isolated from each other.

Each colony was treated with twoplastic strips impregnated with fluvalinate(Apistan®, Novartis laboratories) for nine

1987; Rath and Drescher, 1990; Boecking,1992). Studies conducted by Boecking andDrescher (1991) showed that A. melliferaworkers of the Carniolan race were able todetect, uncap, and remove pupae infestedwith V. destructor. Results of another studyshowed a negative correlation between thedegree of hygienic behavior and the sus-ceptibility to V. destructorof four honeybee lines (Büchler, 1992). In the UnitedStates, genetic programs to develop hygienicbee genotypes have been very successful(Rothenbuhler, 1964; Spivak, 1996; Spivakand Reuter, 1998). Colonies selected forhygienic behavior had lower mite levelsthan non-hygienic ones. However,researchers point out that further experi-ments are necessary to determine to whatextent hygienic behavior reduces the miteload within a colony (Spivak, 1996; Spivakand Reuter, 1998; Boecking and Spivak,1999).

The relative attractiveness of workerbrood and adult bees to the mite may beanother tolerance mechanism (Büchler,1989). Guzmán-Novoa et al. (1996) foundthat the worker brood of Africanized beeswas two times less attractive to V. destruc-tor than the brood of European bees. Workerbrood of hybrid bees (Africanized× Euro-pean) was as attractive to the mite as broodof European bees. In the case of adult bees,European bees were more attractive toV. destructorthan Africanized or hybridworkers.

Other mechanisms that may restrain themite population growth include host factorsthat affect its reproduction. The percent ofV. destructorfemales that reproduce on beeworker brood show variation depending onthe species, race, and gender of the host.V. destructordoes not reproduce success-fully on the worker brood of A. cerana(Koeniger et al., 1981), and there are dif-ferences in the number of mites that repro-duce on the brood of different races andcommercial stocks of A. mellifera (Ritterand De Jong, 1984; Rosenkranz and Engels,

159

M.E. Arechavaleta-Velasco, E. Guzmán-Novoa

weeks to ensure that they were free ofV. destructor, or that they had very low lev-els of infestation before the initiation of theexperiments.

All the colonies were managed in thesame way. The colonies were fed with 2 L of50% sucrose syrup, and were treated withTerramycin® (Pfizer laboratories) every15 days until the beginning of the honeyflow, to provide them with enough food topermit their development, and to prevent afoulbrood outbreak.

2.2. Variation in the abilityof honey bee colonies to restrainthe growth of V. destructorpopulation

Two weeks after the Apistan® strips wereremoved from the colonies, each of themwas artificially infested with 25 mites.Twenty-five workers of each colony wereconfined in two Benton queen cages andone adult female of V. destructorwas placedon the body of each worker with a fine brush(Krause and Page, 1995). The cages withthe infested workers were re-introduced intotheir respective colony, so that the infestedbees could be released by the workers oftheir colony.

The mites used to infest the experimentalcolonies came from one highly infestedcolony located 8 km away from the experi-mental apiary. To collect the mites, a 19 Lplastic container which had a 4 mm wire-mesh screen 20 cm above its bottom wasused. Between the bottom of the containerand the screen, two 40 ml vials half-filledwith ether were placed. Workers of theinfested colony were shaken into the con-tainer that was then closed for 5 min toanaesthetize the bees and to allow the mitesto fall off of the workers. Then the miteswere collected from the bottom of the con-tainer.

The number of frames containing broodand that were covered with adult bees werecounted in each of the experimental colonies

every month, during six months, to deter-mine if there were differences in populationamong the colonies.

To evaluate the relative mite toleranceof the experimental colonies, the growth ofthe mite population was recorded by takingmonthly samples of workers and brood fromthe colonies for a period of six months. Thefirst samples were taken two months afterthe colonies were initially infested.

Three segments of comb each with atleast 50 capped cells of worker brood werecollected from the three central frames ofeach colony and were kept at –5 °C untilthey were examined. Each cell wasuncapped in search for mites, and the levelof infestation of the brood was determinedby dividing the number of cells infested bythe total number of cells in the sample andmultiplying the resulting figure by 100.

The level of infestation of adult bees wasdetermined by collecting and examiningfour samples of 100 workers from eachcolony (400 workers per colony). The sam-ples were taken from the four central framesof each hive, collecting the bees in jars thatwere previously filled with 70% ethanol.Following the technique described byDe Jong et al. (1982), the number of mitesdetected in each sample was divided by thenumber of bees in the sample, and the result-ing figure was multiplied by 100.

Data were transformed by the Box andCox method to provide a normal distribu-tion (Snedecor and Cochran, 1991), andwere subjected to an analysis of varianceunder a complete random design to deter-mine differences in the levels of infestationamong colonies, as evidence of phenotypicvariation in their ability to restrain thegrowth of V. destructorpopulation.

To identify genotypic effects, the datawere subjected to an analysis of varianceunder an unbalanced nested design withrepeated observations. The model was:

Yijkl = µ + Li + Qj(i) + Sk + Eijkl

160

Honey bee tolerance to Varroa

of monitoring, were selected; the eightcolonies with the lowest levels, and the eightwith the highest. The colonies were selectedon the basis of their infestation of adult bees,since no differences in the brood infestationlevels were found among colonies (seeresults). The remaining colonies were treatedwith Apistan® strips.

2.3.1. Grooming behavior

To analyze the effect of the bees’ groom-ing behavior, experiments were conducted inboth the selected colonies and in an incu-bator. Screen (4 mm wire-mesh) bottomboards were placed in each colony. To col-lect the mites that fell from a colony, eachbottom board was assembled with a sheetof white paper smeared with vegetable short-ening beneath the wire-mesh screen. Eachweek during five weeks, the paper sheetswere collected and replaced with new ones.The mites that were stuck on the old sheetswere counted and transferred with a finebrush into a jar containing 70% ethanol.A random sample of 120 mites per colonywas taken from all the mites that were recov-ered. Then, the number of injured mites inthe sample was recorded with the aid of astereoscopic microscope. The injuriesobserved in the mites were similar to thosedescribed by several authors (Wallner, 1990;Moosbeckhofer, 1992; Ruttner and Hänel,1992; Moretto et al., 1993). Only the injuriesof indentations in the idiosoma of the miteswere considered. Injuries in the legs werenot considered because many of them mayhave been caused by the manipulation ofthe samples.

To perform the incubator assay, 30 work-ers were collected from each of the selectedcolonies, and were introduced into a 4 mmwire-mesh cage (10× 10× 20 cm). A piece

where

Yijkl = level of infestationµ = population meanLi = line* effect (i = 1, 2, ..., 7)Qj(i) = colony (queen) effect nested into the

= line (j = 1, 2, ... nj)Sk = sample time effect (k = 1, 2, ..., 5)Eijkl = random error.

The components of the variance were esti-mated by using the restricted maximum like-lihood (REML) (Van Vleck, 1993; Lynchand Walsh, 1998), counting the effect of thelines and the effect of the colony (queen)nested into the line, under the next model:

Yijk = µ + Li + Qj(i) + Eijkwhere

Yijk = level of infestationµ = population meanLi = line* effect (i = 1, 2, ..., 7)Qj(i) = colony (queen) effect nested into the

= line (j = 1, 2, ..., nj)Eijkl = random error.

The genotypic variance was estimated fromthe variation of the line effect and from thecolony (queen) effect nested into the line.The environmental variance was estimatedby the variance of the error (Falconer, 1989;Van Vleck, 1993; Lynch and Walsh, 1998).

2.3. Relative contributionof four characteristics to the variationin the growth of V. destructorpopulation

To estimate the relative effect of fourpresumed mechanisms of mite tolerance,the 16 colonies with the most extremelevels of mite infestation after six months

161

* In the two previous models each queen source (breeder) was considered a line, assuming the exis-tence of different allelic and genotypic frequencies for each of the lines. These considerations werebased on the fact that the breeders do not exchange breeding stock, that the honey bee populations fromeach breeder are geographically isolated from each other, and that each population has been exposedto natural and artificial selection in their particular environments.

M.E. Arechavaleta-Velasco, E. Guzmán-Novoa

of white paper smeared with vegetable short-ening and covered with a 4 mm wire-mesh,was placed at the bottom of each cage. Theworkers of each cage were artificiallyinfested with 10 mites with a fine brush.Then, the cages were placed and kept in anincubator (30 °C, 55% RH). Bees wereoffered water ad libitum, and a paste madeof a mixture of honey and powdered sugar.The bees were kept in the incubator for 48 hand the number of falling mites was countedevery 12 h. At the end of this period, thenumber and proportion of mites with injuriesin the idiosoma was determined.

2.3.2. Hygienic behavior

The bees’ hygienic behavior was mea-sured using a modification of the testdescribed by Newton and Ostasiewski(1986). Three frames containing cappedbrood were chosen from each of the selectedcolonies. On each frame, three groups ofseven cells were punctured with a fine pin tokill the brood. The frames were re-intro-duced in the middle of their hive and wereinspected 24 h later to count the number ofcells that had been cleaned by the bees.

2.3.3. Brood attractivenessand its effect on V. destructorreproduction

Brood attractiveness and its effect on theV. destructorreproductive capability weremeasured using a modification of the tech-nique described by Guzmán-Novoa et al.(1996). Each queen of the selected colonieswas confined on a frame that was placedinside a screen cage in its respective colony.The screen of the cages permitted workers topass and take care of the queens, but did notpermit the queens to leave the cages. Queenslaid eggs on these combs within 24 h afterbeing caged. Four days after the queens werecaged, the combs were removed from thecolonies and a square section (approx.20 cm2) containing young larvae was cutout from each comb. Then, the sections were

installed in a jumbo-size frame, alternatingone segment from a low infested colonywith a segment from a high infested colony,to minimize any effect due to their positionin the frame.

The frame containing the comb segmentswas introduced in the center of a colony thatwas highly infested with V. destructor. Theinfested colony was prepared in advance toreceive the frame. Combs containing openbrood cells were removed from the hive toincrease the chances that V. destructormiteswould infest the experimental brood. Thequeen of this colony was confined in a Ben-ton cage to prevent her from laying on theexperimental comb sections.

Fourteen days after being introduced, theexperimental frame was removed from thecolony, and the comb sections were placedin individual plastic bags and kept at –5.0 °Cuntil analyzed. Cells of each comb sectionwere uncapped and examined. The numberof cells containing mites was recorded toinfer brood attractiveness, whereas the num-ber of mature and immature mites found ineach cell were recorded to determine theeffect of the brood on the mite’s reproduc-tive capability.

The data obtained from each of the assayswere tested for normality, and the means ofeach characteristic measured in the eightcolonies with low infestations were com-pared with those of the eight colonies withhigh infestations, using a t-test. In addition,the data of the selected colonies were sub-jected to an analysis of variance under acomplete random design, to identify differ-ences among the selected colonies for thecharacteristics under study.

Finally, for characteristics that showedsignificant differences between groups orcolonies, analysis of regression and corre-lation were performed between the final lev-els of infestation of the colonies, and theirvalues for each trait measured.

The magnitude of thet statistic, the valueof the determination coefficient from theanalysis of regression, and the correlation

162

Honey bee tolerance to Varroa

(F = 0.71; df = 57, 177; P > 0.05). The finallevels of infestation of the adult bees rangedfrom 6.6 to 44.7% and the final levels ofinfestation of the brood range from 25.5 to60.4%.

3.1.2. Genotypic and environmental variation

The levels of infestation on adult beesgrew significantly month to month from thebeginning of the experiment (F = 91.26;df = 4, 201; P < 0.001). There were signif-icant differences among lines (F = 5.32;df = 6, 201; P < 0.001) and among coloniesnested into lines (F = 3.06;df = 51, 201;P < 0.001) for the levels of infestation(Fig. 1).

coefficient, were used to determine the rel-ative contribution of the characteristics tothe ability of the honey bee colonies torestrain the growth of the mite population.

3. RESULTS

3.1. Variation in the abilityof honey bee colonies to restrainthe growth of V. destructorpopulation

3.1.1. Phenotypic variation

Significant differences were found amongcolonies for the levels of infestation of adultbees after a period of six months (F = 1.57;df = 57, 205;P < 0.01), but no differenceswere found for the levels of infestation ofthe brood for the same period of time

163

Figure 1.Levels of Varroa destructorAnderson and Trueman infestation of adult bees of coloniesfrom seven different lines (A, Az, B1, B2, R, V1, and V2) after six months of infestation. Therewere significant differences among colonies nested into the lines (F = 3.06; df = 52, 201; P < 0.001)and among lines (F = 5.32; df = 6, 201; P < 0.001). Different letters indicate differences between themeans of the lines, based on ANOVA and LSM tests. Statistical tests (ANOVA and LSM) wereperformed with transformed data using the Box and Cox function. The values in the figure representactual non-transformed data (n = 58).

M.E. Arechavaleta-Velasco, E. Guzmán-Novoa

The components of the analysis of vari-ance showed that part of the variation wasrelated to the line effect, and part was relatedto the colonies nested into lines. The varia-tion due to the colonies was higher than thevariation due to the lines. The variationrelated to the line effect and to the colonyeffect are components of the genotypic vari-ance. The variation related to the error inthe model is an estimator of the environ-mental variance (Tab. I).

The infestation levels of the broodincreased significantly month to month overthe sampling period (F = 98.49; df = 4, 173;P < 0.001). However, no differences werefound among lines (F = 1.95; df = 6, 173;P > 0.05), and colonies nested into lines(F = 1.34; df = 51, 173; P > 0.05) for thistrait.

3.2. Relative contributionof four characteristics to the variationin the growth of V. destructorpopulation

The infestation levels of the brood andadult bees increased in the 16 selectedcolonies over the six months period afterbeing artificially infested. However, the rateof increase was higher in the high infestedthan in the low infested group (Figs. 2, 3).After the last sampling, the mean levels ofinfestation of adult bees were 10.5 for thelow infested colonies, and 32.3% for thehigh infested colonies. These levels differedsignificantly (t = 8.48; df = 14; P < 0.001).The mean levels of infestation of the broodwere 39.8 and 43.4% for the low, and forthe high infested groups, respectively. Theselevels were not different (t = 0.48; df = 14;P > 0.05). Moreover, the bee populationmeasured as number of combs covered withadult bees, and as number of combs con-taining brood, did not differ among theselected colonies (U = 73.5; U = 71.0 forcombs with bees and for combs with brood,respectively; P > 0.05; Mann-WhitneyU test).

3.2.1. Grooming behavior

Significantly more mites were collectedfrom the low-infested than from the high-infested colonies (t = 2.41;df = 14; P < 0.05;Tab. II). The mean number of mites col-lected per week were 186.3 and 386.3 forthe high and for the low infested groups,respectively. There were significant differ-ences among colonies for the number ofmites collected per week (F = 6.47; df = 15,60; P < 0.001).

164

Table I. Variance sources for the ability of honeybee colonies to restrain the population growthof Varroa destructorAnderson and Truemanestimated by restricted maximum likelihood(REML).

Variance sources Estimator

Lines 0.0129 Colonies nested into lines 0.0578 Error 0.1228 Total 0.1935

Figure 2. Average levels (± SE) of Varroadestructor Anderson and Trueman infestation ofadult bees from two groups of colonies selectedfor their high (n = 8) and low (n = 8) ability torestrain the mite population growth for fivemonthly sampling periods.

Honey bee tolerance to Varroa

group had a mean of 11.3 mites with injuriesin their idiosoma, which represented 9.4% ofthe total number of mites sampled, whereasthe mean of the low infested group was 13.5injured mites, which represented 11.3% ofthe mites sampled. A negative correlationwas found between the number of injuredmites and the final infestation levels of thecolonies (r = –0.76; n = 16; P < 0.001). Alinear regression analysis between these two

A significant negative correlation wasfound between the number of mitescollected from the bottom boards and thefinal infestation levels of the colonies(r = –0.65; n = 16; P < 0.01). A linearregression analysis between these two vari-ables showed a significant relationship,explaining 38% of the variation (r2 = 0.38;n = 16; P < 0.01; Tab. III; Fig. 4).

Regarding the number of injured mites,significant differences were found betweenthe means of the two groups (t = 2.61;df = 14; P < 0.05; Tab. II). The high infested

165

Table II. Differences for presumed mechanismsof tolerance between two groups of coloniesselected for their Varroa destructor Andersonand Trueman infestation level (high and low).Figures in parentheses are d.f.

Mechanism t P

Grooming behavior 2.41 0.03 No. fell mites from hive tests (14)

Grooming behavior 2.61 0.02No. injured mites from hive tests (14)

Grooming behavior 2.20 0.04No. mites from incubator tests (14)

Grooming behavior 0.92 0.79No. injured mites from incubator tests(14)

Hygienic behavior –0.64 0.53No. cleaned cells (14)

Brood attractiveness –0.11 0.91Proportion infested cells (14)

Varroa reproduction 0.49 0.62

Proportion of mites that reproduced (14)

Figure 3. Average levels (± SE) of VarroadestructorAnderson and Trueman infestation ofthe brood from two groups of colonies selectedfor their high (n = 8) and low (n = 8) ability torestrain the mite population growth for fivemonthly sampling periods.

Table III. Correlation (r) and determination (r2) coefficients between the final Varroa destructorAnderson and Trueman level of infestation of the selected colonies and the number of mites recoveredand injured from the assays used to measure grooming behavior in the hives and in an incubator.

Characteristic n r r2 P

Recovered mites (Hive) 16 –0.65 0.38 0.008Injured mites (Hive) 16 –0.76 0.54 0.0001Recovered mites (Incubator) 16 –0.52 0.22 0.04Injured mites (Incubator) 16 –0.23 0.06 0.45

M.E. Arechavaleta-Velasco, E. Guzmán-Novoa

variables showed a significant relationship,explaining 54% of the variation (r2 = 0.54;n = 16; P < 0.001; Tab. III; Fig. 5).

Significant differences were foundbetween the means of the two groups forthe number of mites recovered from theincubator cages (t = 2.20;df = 14; P < 0.05).The mean of the high infested group was2.3 mites, and the mean of the low infestedgroup was 4.6 mites (Tab. II).

A negative correlation was foundbetween the number of mites recovered inthe incubator tests and the final levels ofinfestation (r = –0.52; n = 16; P < 0.05). Alinear regression analysis between these twovariables showed a significant relationship,explaining 22% of the variation (r2 = 0.22;n = 16; P < 0.05; Tab. III).

No differences were found between thetwo groups for the number of injured mitesin the incubator (t = 0.92;df = 14; P > 0.05;Tab. II), and no correlation was foundbetween this variable and the final infestationlevels of the colonies (r = –0.23; n = 16;P > 0.05; Tab. III).

3.2.2. Hygienic behavior

No differences were found between thetwo groups of colonies for the numberof cells cleaned by the bees (t = –0.637;df = 14; P > 0.05; Tab. II). The colonies ofthe high infested group had a mean of 54.8cells cleaned (87.0%), whereas the lowinfested group had a mean of 51.3 cellscleaned (81.4%). However, an analysis ofvariance showed significant differencesamong colonies (F = 10.71; df = 15, 116;P < 0.001), with a range from 11 to 63 cellscleaned (17.5–100.0%). No correlation wasfound between the final levels of infesta-tion of the colonies and the number of cellscleaned by the bees (r = 0.18; n = 16;P > 0.05).

3.2.3. Brood attractiveness

No differences were found between thetwo groups of colonies for the percentageof cells that were infested with mites(t = 0.11; df = 14; P > 0.05; Tab. II). Themean of the high infested group was 6.0%

166

Figure 4. Relationship between the colonies’final levels of infestation and the mean numberof mites recovered per week from the selectedcolonies. The regression equation is Y = 35.35 –0.043 X (r2 = 0.38; P = 0.008; n = 16).

Figure 5. Relationship between the colonies’final levels of infestation and the numberof injured mites found in a random sample(n = 120) recovered from the hives’ bottomboards of the selected colonies. The regressionequation is Y = 78.105 – 4.582 X (r2 = 0.54;P = 0.0001; n = 16).

Honey bee tolerance to Varroa

1991; Harbo and Hoopingarner, 1997).Additionally, the fact that part of the varia-tion was explained by genotypic effects is inagreement with results reported by Ron andRosenthal (1987), whom showed that part ofthe variation in the V. destructorinfestationlevels of their experimental colonies wasdue to differences in the genetic origin ofthe lines of honey bees that they used.

It is important to notice that the largestpart of the genotypic variation was foundamong the colonies, and not among the lines,which suggest that it is possible to find tol-erant stocks within the different populationsof honey bees that queen breeders keep, andthat this characteristic is not related to a par-ticular line.

Results of this study suggest that the vari-ation in the growth of V. destructorpopu-lation found in the experimental coloniescould be explained mainly by the workers’ability to eliminate mites infesting adultbees, presumably by grooming behavior,since this was the only characteristic mea-sured that showed significant differencesbetween groups and among colonies. Thenumber of mites recovered from the lowinfested colonies in both the incubator andthe hives was significantly higher than thenumber of mites recovered from the highinfested colonies. Moreover, the low infestedcolonies had more injured mites, and sig-nificant negative correlations were foundbetween the final levels of infestation of thecolonies and both the number of recoveredand injured mites. The fact that no differ-ences were found between the groups forthe number of injured mites in the incubatortests may have been a consequence of asmall sample size, or it may be that othermechanisms in addition to grooming behav-ior were also responsible for the differencesobtained in the field tests.

It has been shown that there is a positivecorrelation between mite infestation levels incolonies and mite drop on bottom boards(Fries et al., 1991; Calatayud and Verdu,1993). Our results are not in agreement with

whereas that of the low infested group was5.8%. Also no differences were foundamong colonies for this trait (F = 0.02;df = 15, 84; P > 0.05).

3.2.4. Brood effect on V. destructor reproduction

No differences were found between thetwo groups of colonies for the percentageof infested cells in which the mites wereable to reproduce (t = 0.495; df = 14;P > 0.05; Tab. II). The mites did not repro-duce in 21.0% of the infested cells in thecolonies of the high infested group, whereasin the colonies of the low infested group,mites did not reproduce in 28.5% of theinfested cells. Additionally, no differenceswere found among colonies for this vari-able (F = 0.18;df = 15, 84; P > 0.05).

4. DISCUSSION

Our results showed high phenotypic andgenotypic variation among colonies for thelevels of infestation of V. destructoron adultbees, but no significant variation was foundfor the levels of infestation in the brood.These results apparently suggest that theability of the colonies to restrain theV. destructorpopulation growth found inthis study was related to the colonies capac-ity to control the mite population when theywere phoretic on adult bees. The lack of sig-nificant variation in the infestation levels inthe brood could have resulted from the lackof differences between groups for hygienicbehavior, brood attractiveness to the mite,and brood effect on the mite reproduction,which are mechanisms that affect the mitepopulation growth at the brood stage of thebee’s life.

The variation in the mite populationgrowth of the experimental colonies agreeswith results reported by several authors(Engels et al., 1986; Kulincevic andRinderer, 1988; Kulincevic et al., 1988;Moosbeckhofer et al., 1988; Moretto et al.,

167

M.E. Arechavaleta-Velasco, E. Guzmán-Novoa

these findings, since more mites were recov-ered from the bottom boards of the coloniesbelonging to the low infested group thanfrom those of the colonies of the highinfested group. Results of our study supportthose of Moosbeckhofer (1992), who alsofound a negative correlation between thenumber of injured mites and the V. destruc-tor infestation levels of the colonies of hisstudy. Our findings could be due to the factthat we selected colonies with extreme lev-els of infestation as a result of their differ-ential ability to restrain the mite populationgrowth. Thus, our colony groups were notaverage colonies, but colonies with differ-ences in susceptibility to mite infestationon adult bees. Differences in the suscepti-bility of adult bees to become infested withV. destructormites have been foundbetween European and Africanized beecolonies. Africanized and hybrid bees weretwo times less susceptible to becomeinfested than European workers (Guzmán-Novoa et al., 1996). In another study,Moretto and Mello (2000) measuredV. destructormortality rates by recoveringmites from bottom board trays installed inAfricanized colonies. They found that thecolonies with the highest V. destructor mor-tality rates had the lowest levels of infesta-tion on adult bees (r = –0.83). The abovestudies support our findings, and suggestthat the bees of some colonies eliminatemore mites than the bees of others.

An alternative hypothesis to explain whywe found more falling mites in the lowinfested than in the high infested coloniesis that there may have been differences inthe amount of brood and bees between ourgroups of colonies. Positive correlationsbetween number of mites on bottom boardswith amount and emergence of brood havebeen reported (Lobb and Martin, 1997;Beetsma et al., 1999). However, there wereno differences in the number of combs withbees and with brood between our two groupsof colonies. Therefore, it seems that theworkers of colonies from the low infestedgroup had a higher grooming ability than

the workers of colonies from the highinfested group. It could also be that a com-bined effect of grooming and hygienicbehavior was responsible for the differencesbetween our two groups of colonies. Dam-aged mites found on bottom boards can notbe attributed exclusively to the groomingbehavior of worker bees. Hygienic bees andwax moths may cause damage to mites seenon bottom boards too (Szabo and Walker,1995; Boecking and Spivak, 1999). Thus,the methods that we used could not assesand separate what proportion of mites felland what proportion of mites were injured asa result of the bees’ grooming behavior,hygienic behavior, or other causes. How-ever, it was clear that adult bees from thetwo groups differed significantly in theirlevels of mite infestation, but did not differin their levels of brood infestation. There-fore, it is likely that mechanisms associatedwith adult bees were responsible for thedifferences in the degree of infestation ofour two groups of colonies.

The results reported here, as well as thosefrom studies conducted in other countries(Morse et al., 1991; Moosbeckhofer, 1992;Ruttner and Hänel, 1992; Boecking et al.,1993; Moretto et al., 1993; Rosenkranzet al., 1997; Bienefeld et al., 1999), suggestthat grooming behavior may be an importantmechanism conferring tolerance to honeybee colonies toward V. destructor. How-ever, the effects of grooming behavior mustbe interpreted with caution, because thereliability of the indirect methods used toquantify this behavior is controversial(Rosenkranz et al., 1997; Bienefeld et al.,1999; Boecking and Spivak, 1999).

The importance of hygienic behavioron conferring tolerance to honey bees toV. destructoris not clear. In the presentstudy, the differences found among thecolonies suggest that this mechanism couldcontribute to the bees’ tolerance, but itseffect was apparently smaller than that ofgrooming behavior, since no differenceswere found between the two groups ofexperimental colonies. The pin-killed

168

Honey bee tolerance to Varroa

The fact that no significant differenceswere found for the above two mechanisms inthis study could be a consequence of a rela-tively small sample size. However, if theirrelative contribution had been high, differ-ences should have been found even with thesample size that was used in the experi-ments. The honey bee colonies used in thisstudy were obtained from commercial queenbreeders, and thus, the experimental colonieswere not highly Africanized, which couldbe another reason for the lack of agreementwith the above studies.

Tolerance toward V. destructorhas beendocumented in South America and Mexico,but the degree of tolerance and the mecha-nisms involved in this tolerance vary in dif-ferent places (Guzmán-Novoa et al., 1999;Rosenkranz, 1999). Studies conducted inthe Americas have shown that the factorswhich contribute to mite tolerance in honeybees are not the same in all bee populations.The low fertility of the mite is the main tol-erance factor in honey bees in Brazil. How-ever in other countries, characteristics suchas lower attractivity of the bee brood,increased grooming, and hygienic behav-ior, have shown to be associated with lowinfestation levels of V. destructorin honeybee colonies (Moretto et al., 1993; Guzmán-Novoa et al., 1996; Vandame et al., 1997;Guerra et al., 2000). Thus, it may be thatsome characteristics that have a strong influ-ence on conferring tolerance toward the mitein some bee populations may not have ahigh influence in others.

It is critical to determine which and towhat extent different characteristics andmechanisms confer tolerance to honey beecolonies toward V. destructorto facilitatethe development of successful breeding pro-grams for mite tolerant honey bees. How-ever to accomplish this goal, studies areneeded to develop direct and reliable tech-niques for measuring presumed tolerancecharacteristics. It is also important to takeinto account that tolerance to V. destructorin honey bees is influenced by multiple

method that we used to determine hygienicbehavior might be questioned because itsreliability has not been tested, and perhapswe were not able to accurately measure theactual hygienic ability of the experimentalcolonies. In other studies, Boecking andDrescher (1991) and Büchler (1992),reported that Carniolan bees that showedhigh hygienic behavior, also showed lowlevels of mite infestation. Spivak (1996)found that for two lines of honey bees, onehygienic and one non-hygienic, there weresignificant differences in the bees’ abilityto detect and remove pupae from cells thatwere previously infested with V. destructormites in her first period of tests. But for asecond period of tests, no differences werefound between the two groups of colonies.Bär and Rosenkranz (1992) concluded thathygienic behavior does not confer signifi-cant resistance to the colonies against themite.

No differences were found for the broodattractiveness to V. destructorbetweengroups or among colonies. These resultsagree with those reported by Camazine(1986), who did not find differences betweenAfricanized and European bees for themite’s preference for a particular type ofbrood. However these results differ fromthose of Guzmán-Novoa et al. (1996) whomfound differences among colonies of Euro-pean, Africanized, and hybrid genotypes.Additionally, no differences betweencolonies or between groups were found forthe brood effect on the reproductive capa-bility of the mite, which again suggests thatthis mechanism did not have a significantcontribution on restraining the growth ofthe mite population in the experimentalcolonies of this study. These results alsodiffer from those reported by several authors(Rosenkranz and Engels, 1994; Guzmán-Novoa et al.,1996; Harbo and Hoopingarner,1997), whom reported differences in thereproductive capability of the mite as a resultof the effect of the infested brood’s geno-type.

169

M.E. Arechavaleta-Velasco, E. Guzmán-Novoa

factors of the host, by differences in themite’s virulence, and by different environ-mental conditions, which could result in dif-ferent levels of tolerance in bee colonies.Therefore, additional studies are necessaryto confirm the actual contribution and impor-tance of different bee characteristics in dif-ferent environments, with different popula-tions of honey bees, and with differentstrains of V. destructormites.

ACKNOWLEDGEMENTS

This work was funded with grants fromCONACYT and from the National AutonomousUniversity of Mexico (UNAM). We thank GregJ. Hunt for critical review of this manuscript.We are grateful to Ruben Mendoza, EnriqueCoronado, Gabriela Ortiz, Modesto Bautista,Enrique Estrada, Alberto Barrera, SalvadorCajero, and Antonio Zozaya whom providedassistance in various ways.

Résumé – Influence relative de quatrecaractéristiques qui limitent la croissancede la population de l’acarien Varroa des-tructor dans les colonies d’abeilles domes-tiques (Apis mellifera). L’acarien VarroadestructorAnderson et Trueman représentela menace la plus grave pour l’industrie api-cole mondiale. Actuellement les coloniesinfestées sont traitées avec des produits chi-miques qui permettent dans une certainemesure de contrôler le parasite. Pourtant surle long terme, les acaricides peuvent être àl’origine d’un certain nombre de problèmes.Une solution pour lutter contre V. destructorest de développer des lignées d’abeilles tolé-rantes à l’acarien. L’étude a été faite pourdéterminer s’il existe une variation phéno-typique et génotypique de l’aptitude descolonies d’abeilles à limiter la croissancede la population de V. destructoret pourétablir l’influence relative de quatre carac-téristiques supposées conférer aux abeillesune tolérance à l’acarien. Cinquante huitcolonies ont été installées dans des ruches detaille jumbo. Les reines à la tête de ces

colonies provenaient de sept lignées diffé-rentes fournies par des sélectionneurs. Lescolonies ont été infestées artificiellementavec 25 acariens. Pour évaluer la tolérancedes colonies expérimentales vis-à-vis del’acarien, on a enregistré la croissance de lacolonie en prélevant chaque mois des échan-tillons d’ouvrières et de couvain pendantune période de six mois. Les données ontsubi une analyse de variance afin de déter-miner la variation phénotypique et génoty-pique. Les composantes de la variance ontété également estimées. Pour déterminerl’influence relative des quatre mécanismesde tolérance, on a choisi 16 colonies : leshuit colonies ayant le taux d’infestation leplus bas et les huit qui avaient le taux d’infes-tation le plus élevé. On a étudié les méca-nismes suivants : comportement de toilet-tage, comportement hygiénique, attractivitédu couvain et non reproduction induite parl’hôte. Dans les colonies sélectionnées,chaque mécanisme a été mesuré à l’aided’un test biologique. La valeur moyennedes 8 colonies les moins infestées a été com-parée avec celle des huit les plus infestées àl’aide d’un test t. Les données ont subi enoutre une analyse de variance pour identifierles différences entre colonies sélectionnées.Pour les mécanismes qui ont montré des dif-férences significatives entre groupes ou entrecolonies, on a réalisé une analyse de régres-sion et une analyse de corrélation entre leniveau final d’infestation des colonies et lesvaleurs moyennes de chaque mécanismemesuré. Une variation phénotypique a ététrouvée entre les colonies pour les niveauxd’infestation des abeilles adultes, mais paspour l’infestation du couvain (Figs. 2, 3).On a trouvé des différences significativesentre lignées et entre colonies au sein deslignées (Fig. 1), ce qui prouve une influencedu génotype et du milieu (Tab. I). Concer-nant les mécanismes étudiés, des différencessignificatives ont été trouvées entre groupeset entre colonies pour le comportement detoilettage. Des corrélations significative-ment négatives ont été trouvées entre leniveau final d’infestation et le nombre

170

Honey bee tolerance to Varroa

gegenüber den Milben bewirken. Achtund-fünfzig Völker wurden in Beuten der Gröβe„Jumbo“ etabliert. Die Königinnen in diesenVölkern entstammten der Linienzucht vonsieben verschiedenen Bienenzüchtern. DieVölker wurden mit jeweils 25 Milben infi-ziert. Um den Toleranzgrad der Völkergegenüber V. destructorzu bestimmen,wurde das Populationswachstum der Mil-ben anhand von monatlichen Probennah-men von Arbeiterinnen und Brut über einenZeitraum von 6 Monaten ermittelt. ZurBestimmung phenotypischer und geneti-scher Variabilität wurden die Daten einerVarianzanalyse unterworfen. Zur Bestim-mung des relativen Einflusses von vier Tole-ranzmechanismen wurden 16 Völker aus-gewählt, dieses waren die acht Völker mitdem niedrigsten und die acht Völker mitdem höchsten Befallsgrad. Die untersuch-ten Mechanismen waren das Putzverhalten,das hygienische Verhalten, die Brutattrak-tivität und das vom Wirt verursachte Nicht-reproduzieren der Milben. Jeder dieserMechanismen wurde in einem Biotest erfasstund die Mittelwerte der acht hochbefalle-nen Völker mit denen der acht niedrigbe-fallenen anhand von t-Tests verglichen.Zusätzlich wurden die Daten mittels einerVarianzanalyse auf Unterschiede zwischenden ausgewählten Völkern hin untersucht.Für die Mechanismen, die zwischen denGruppen oder den Völkern signifikanteUnterschiede aufwiesen, wurden Regres-sionen und Korrelationen zwischen demEndbefall der Völker und den Mittelwertender erfassten Mechanismen berechnet. Zwi-schen den Völkern gab es signifikanteUnterschiede im Befallsgrad der Bienen,aber es wurden keine signifikanten Unter-schiede im Befallsgrad der Arbeiterinnen-brut gefunden (Abb. 2, 3). Zwischen Bie-nenlinien und zwischen Völkern innerhalbder Bienenlinien bestanden signifikanteUnterschiede im Befall der Bienenarbeite-rinnen (Abb. 1), die genetische und phäno-typische Einflüsse belegen (Tab. I). Bei denuntersuchten Mechanismen bestanden signi-fikante Unterschiede zwischen Bienenlinien

d’acariens récupérés d’une part (Tab. II ;Fig. 4) et le nombre d’acariens blessésd’autre part (Tab. III ; Fig. 5). Pour le com-portement hygiénique, des différences ontété trouvées entre colonies, mais pas entregroupes. Par contre aucune différence n’aété trouvée ni entre groupes ni entre coloniespour l’attractivité du couvain et pour l’effetdu couvain sur la reproduction de l’acarien.Ces résultats suggèrent que, dans cette étude,la tolérance des colonies vis-à-vis de V. des-tructor était liée à l’aptitude des abeilles àlimiter la population d’acariens lorsqu’ilssont dans la phase phorétique sur les abeillesadultes, vraisemblablement par le compor-tement de toilettage. Les problèmes et lesimplications liées aux méthodes utiliséespour mesurer les caractéristiques de tolé-rance, ainsi que les recherches à venir, sontdiscutées.

Apis mellifera / Varroa destructor/tolérance / mécanisme de résistance /comportement toilettage / comportementhygiénique / Mexique

Zusammenfassung – Der relative Einflussvon vier das Populationswachstum derMilbe Varroa destructorin Honigbienen-völkern (Apis mellifera) begrenzendenEigenschaften.Die Milbe Varroa destruc-tor Anderson und Trueman ist weltweit dieernsthafteste Bedrohung der Bienenindu-strie. Gegenwärtig wird der Befall der Völ-ker mit chemischen Behandlungsmitteln biszu einem gewissen Grad unter Kontrollegehalten. Langfristig verursachen Mitizideallerdings eine Reihe von Problemen. EineMöglichkeit zur Bekämpfung der Milben istdie Entwicklung von gegenüber V. destruc-tor toleranten Bienenlinien. Diese Untersu-chung sollte die Existenz phenotypischerund genetischer Variation der Fähigkeit vonBienenvölkern belegen, das Populations-wachstum von V. destructorzu begrenzenund weiterhin den relativen Effekt von vierFaktoren einschätzen, von denen angenom-men wird, dass sie eine Toleranz der Bienen

171

M.E. Arechavaleta-Velasco, E. Guzmán-Novoa

und Völkern bezüglich des Putzverhaltens(Tab. II). Signifikante negative Korrelatio-nen bestanden zwischen dem Endbefall derVölker und der Anzahl wiedergefundenerMilben (Tab. III, Abb. 4), und zwischendem Endbefall der Völker und der Anzahlverletzter Milben (Tab. III, Abb. 5). Unter-schiede des hygienischen Verhaltens bestan-den zwischen Völkern, aber nicht zwischenden Bienenlinien. Weder Brutattraktivitätnoch der Einfluss der Brut auf die Repro-duktion der Milben unterschied sich zwi-schen Völkern oder Bienenlinien. DieseErgebnisse legen nahe, dass die in dieserStudie gefundene Toleranz der Völkergegenüber V. destructorzu der Fähigkeitder Bienen in Beziehung stand, die auf denArbeiterinnen phoretische Milbenpopula-tion zu begrenzen, dies geschah offensicht-lich durch das Putzverhalten. Die mit derMessung von Toleranzmechanismen zusam-menhängenden Probleme und Folgerungensowie zukünftige Untersuchungen werdendiskutiert.

Apis mellifera / Varroa destructor /Toleranz / Resistenzmechanismen /Mexico

REFERENCES

Anderson D.L., Trueman J.W.H. (2000) Varroa jacob-soni (Acari: Varroidae) is more than one species,Exp. Appl. Acarol. 24, 165–189.

Bär E., Rosenkranz P. (1992) Spezifisches putzver-halten von honigbienen (Apis mellifera) unter-schiedlicher rassen gegenüber Varroa-infiziertenbrützellen. (Specific grooming behaviour towardsbrood cells infested with Varroa jacobsoniof dif-ferent honey bee races.), Ann. Univ. M. Curie-Skl.47, 1–6.

Beetsma J., Boot W.J., Calis J. (1999) Invasionbehaviour of Varroa jacobsoni Oud.: from bees intobrood cells, Apidologie 30, 125–140.

Bienefeld K., Zautke F., Pronin D., Mazedd A. (1999)Recording the proportion of damaged Varroa jacob-soni Oud. in the debris of honey bee colonies (Apismellifera), Apidologie 30, 249–256.

Boecking O. (1992) Removal behavior of Apis mellif-era towards sealed brood cells infested with Varroajacobsoni: Techniques, extent and efficacity,Apidologie 23, 371–373.

Boecking O., Drescher W. (1991) Response of Apismellifera L. colonies to brood infested with Var-roa jacobsoniO., Apidologie 22, 237–241.

Boecking O., Rath W., Drescher W. (1993) Groom-ing and removal behavior: strategies of Apis mel-lifera and Apis ceranabees against Varroa jacob-soni, Am. Bee J. 133, 117–119.

Boecking O., Spivak M. (1999) Behavioral defenses ofhoney bees against Varroa jacobsoniOud., Api-dologie 30, 141–158.

Büchler R. (1992) Test auf Varroatoleranz in Rahmenvon Leistungsprüfungen (Test for Varroa-toler-ance performance), Neue Bienen Zeitung 3,162–167.

Büchler R. (1989) Attractivity and reproductive suit-ability for the Varroa-mite of bee brood from dif-ferent origins, in: Proceedings of the Meeting ofthe Economic Community Experts Group. Udine,Italy, pp. 139–145.

Calatayud F., Verdu M.J. (1993) Hive debris countsin honeybee colonies – A method to estimate thesize of small populations and rate of growth of themite Varroa jacobsoniOud. (Mesostigmata: Var-roidae), Exp. Appl. Acarol. 17, 889–894.

Camazine S. (1986) Differential reproduction of themite Varroa jacobsoni(Mesostigmata: Varroidae)on Africanized and European honey bees(Hymenoptera: Apidae), Ann. Entomol. Soc. Am.79, 801–803.

De Guzman L., Rinderer T.E., Stelzer J.A. (1997) Theevidence of the origin of Varroa jacobsoni Oud. inthe Americas, Biochem. Gen. 35, 327–335.

De Jong D., Roma D.A., Gonçalves L.S. (1982) Acomparative analysis of shaking solutions for thedetection of Varroa jacobsonion adult honey bees,Apidologie 13, 297–306.

Elzen P.J., Eischen F.A., Baxter J.R., Elzen G.W.,Wilson W.T. (1999) Detection of resistance in USVarroa jacobsoniOud. (Mesostigmata: Varroidae)to the acaricide fluvalinate, Apidologie 30, 13–17.

Engels W., Gonçalves L.S., Steiner J., Buriolla A.H.,Cavichio-Issa M.R. (1986) Varroa-Befall voncarnica-Volkern in Tropenklima (Varroa infestationgrowth in carnica bees in tropical climate),Apidologie 17, 203–216.

Falconer D.S. (1989) Introduction to quantitative genet-ics, Longman Scientific and Technical.

Faucon J.P., Flamiini C. (1990) Résidus de fluvalinatedans la cire et dans le miel (Residues of fluvalinatein honey and wax), Santé Abeille 118, 182–184.

Fries I., Aarhus A., Hansen H., Korpela S. (1991) Com-parison of diagnostic methods for detection of lowinfestation levels of Varroa jacobsoniin honey bee(Apis mellifera) colonies, Exp. Appl. Acarol. 10,279–287.

Guerra J.C., Gonçalves L.S., De Jong D. (2000)Africanized honey bees (Apis melliferaL.) are moreefficient at removing worker brood artificiallyinfested with the parasitic mite Varroa jacobsoniOudemans than are Italian bees or Italian/African-ized hybrids, Genet. Mol. Biol. 23, 89–92.

172

Honey bee tolerance to Varroa

Lodesani M., Pellacani A., Bergomi S., Carpana E.,Rabitti T., Lasagni P. (1992) Residue determina-tion for some products used against Varroa infes-tation in bees, Apidologie 23, 257–272.

Lynch M., Walsh B. (1998) Genetics and analysis ofquantitative traits, Sinauer Associates, Inc., Mas-sachusetts, pp. 779–800.

Moosbeckhofer R. (1992) Beobachtungen zumauftreten beschädigter varroamilben im natürlichentotenfall bei völkern von Apis mellifera carnica,Apidologie 23, 523–531.

Moosbeckhofer R., Fabsicz M., Kohlich A. (1988)Untersuchungen über die abhängigkeit dernachkommensrate von Varroa jacobsoni vonbefallsgrad der bienenvölkern, Apidologie 19,181–207.

Moretto G., Mello L.J. (2000) Resistance of Africanizedbees (Apis mellifera L.) as a cause of mortality ofthe mite Varroa jacobsoniOud. in Brazil, Am. BeeJ. 140, 895–897.

Moretto G., Gonçalves L.S., De Jong D., BishuetteM.Z. (1991) The effects of climate and bee raceon Varroa jacobsoniOud. infestations in Brazil,Apidologie 22, 197–203.

Moretto G., Gonçalves L.S., De Jong D. (1993) Hered-ability of africanized and european honey bee defen-sive behavior against the mite Varroa jacobsoni,Braz. J. Gen. 16, 71–77.

Morse R.A., Miska D., Robinson G.E. (1991) Varroaresistance in US honeybees, Am. Bee J. 131,433–434.

Newton D.C., Ostasiewski N.J. (1986) A simplifiedbioassay for behavioral resistance to american foul-brood in honey bees (Apis mellifera L.), Am. Bee J.126, 278–281.

Peng Y.S., Fang Y., Xu S., Ge L. (1987a) The resis-tance mechanisms of the Asian honey bee, ApisceranaFabr., to an ectoparasitic mite, Varroa jacob-soniOudemans, J. Invert. Pathol. 49, 54–60.

Peng Y.S., Fang Y., Xu S., Ge L., Nasr M.E. (1987b)Response of foster Asian honey bee (Apis ceranaFabr.) colonies to the brood of European honey bee(Apis mellifera L.) infested with the parasitic mite,Varroa jacobsoni Oudemans, J. Invertebr. Pathol.49, 259–264.

Rath W., Drescher W. (1990) Response of Apis cer-ana Fabr. colonies towards brood infested withVarroa jacobsoniOud. and infestation rate ofcolonies in Thailand, Apidologie 21, 311–321.

Ritter W., De Jong D. (1984) Reproduction of VarroajacobsoniO. in Europe, the Middle East and trop-ical South America, Z. Angew. Entomol. 98, 55–57.

Ron M., Rosenthal C. (1987) Genetic differences inthe resistance to varroasis of bees in Israel, in: Pro-ceedings of the XXXI Int. APIMONDIA Congress,Warsaw, Poland, pp. 263–265.

Rosenkranz P. (1999) Honey bee (Apis mellifera) tol-erance to Varroa jacobsoniOud. in South America,Apidologie 30, 159–172.

Guzmán-Novoa E., Correa A. (1996) Selección de abe-jas melíferas (Apis melliferaL.) resistentes al ácaroVarroa jacobsoniOud. (Selection of honey beesresistant to the mite Varroa jacobsoniOud.), Vet-erinaria Méx. 27, 149–158.

Guzmán-Novoa E., Sanchez A.A., Page R.E. Jr.,Garcia P.T. (1996) Susceptibility of European andAfricanized honeybees (Apis melliferaL.) and theirhybrids to Varroa jacobsoni Oud., Apidologie 27,93–103.

Guzmán-Novoa E., Vandame R., Arechavaleta M.E.(1999) Susceptibility of European and Africanizedhoney bees (Apis melliferaL.) to Varroa jacob-soniOud. in Mexico, Apidologie 30, 173–182.

Harbo J.R., Hoopingarner R. (1997) Honey bees(Hymenoptera: Apidae) in the United States thatexpress resistance to Varroa jacobsoni(Mesostig-mata: Varroidae), J. Econ. Entomol. 90, 893–898.

Harbo J.R., Harris J.W. (1999) Selecting honey bees forresistance to Varroa jacobsoni, Apidologie 30,183–196.

Hillesheim E., Ritter W., Bassand D. (1996) First dataon resistance mechanisms of Varroa jacobsoni(Oud.) against tau-fluvalinate, Exp. Appl. Acarol.20, 283–296.

Koeniger N., Koeniger G., Wijayagunasekar N.H.P.(1981) Anpassung von Varroa Jacobsonian ihrennatürlichen wirt Apis cerana in Sri Lanka (Obser-vations about the adaptation of Varroa Jacobsonionits natural host Apis cerana in Sri Lanka), Api-dologie 12, 37–40.

Krause B., Page R.E. (1995) Effect of Varroa jacobsonion feral Apis melliferain California, Environ. Ento-mol. 24, 1473–1480.

Kulincevic J.M., Rinderer T.E. (1986) Differential sur-vival of honey bee colonies infested by Varroajacobsoniand breeding for resistance, in: Pro-ceedings of the XXX Int. APIMONDIA Congress,Nagoya, Japan. pp. 175–177.

Kulincevic J.M., Rinderer T.E. (1988) Breeding honeybees for resistance to Varroa jacobsoni: Analysis ofpopulation dynamics and structure of mites progeny,in: Needham G.R., Page R.E. Jr., Delfinado-BakerM., Bowman C.E. (Eds.), Africanized Honey Beesand Bee Mites, Ellis Horwood, Chichester, UK,pp. 360–369.

Kulincevic J.M., Rinderer T.E., Urosevic D.J. (1988)Seasonality and colony variation of reproducingand non-reproducing Varroa jacobsonifemales inwestern honey bee (Apis mellifera) worker brood,Apidologie 19, 173–179.

Lobb N., Martin S. (1997) Mortality of Varroa jacob-soniOudemans during or soon after the emergenceof worker and drone honeybees Apis melliferaL.,Apidologie 28, 367–374.

Lodesani M., Colombo M., Spreafico M. (1995) Inef-fectiveness of Apistan® treatment against the miteVarroa jacobsoniOud. in several districts of Lom-bardy (Italy), Apidologie 26, 67–72.

173

M.E. Arechavaleta-Velasco, E. Guzmán-Novoa

Rosenkranz P., Engels W. (1994) Infertility of Varroajacobsonifemales after invasion into Apis melliferaworker brood as a tolerance factor against varroa-sis, Apidologie 25, 402–411.

Rosenkranz P., Fries I., Boecking O., Stürmer M.(1997) Damaged Varroa mites in the debris ofhoney bee (Apis mellifera L.) colonies with andwithout hatching brood, Apidologie 28, 427–437.

Rothenbuhler W.C. (1964) Behaviour genetics of nestcleaning in honey bees: I. Responses of four inbredlines to disease-killed brood, Anim. Behav. 12,578–583.

Ruttner F., Hänel H. (1992) Active defense againstVarroa mites in a carniolan strain of honey bee(Apis mellifera carnica Pollmann), Apidologie 23,173–187.

Slabezki Y., Gal H., Lensky Y. (1991) The effect of flu-valinate application in bee colonies on populationlevels of Varroa jacobsoni and honey bees (Apismellifera L.) and on residues in honey and wax,Bee Sci. 1, 189–195.

Snedecor G., Cochran W. (1991) Statistical methods,8th ed., Iowa State University, Iowa.

Spivak M. (1996) Honey bee hygienic behavior anddefense against Varroa jacobsoni, Apidologie 27,245–260.

Spivak M., Reuter G.S. (1998) Performance of hygieniccolonies in a commercial apiary, Apidologie 29,285–296.

Szabo T.I., Walker C.R.T. (1995) Damages to deadVarroa jacobsonicaused by the larvae of Galleriamellonella, Am. Bee J. 135, 421–422.

Van Vleck L.D. (1993) Selection index and introduc-tion to mixed model methods, CRC Press.

Vandame R., Colin M.E., Otero-Colina G. (1997)Africanized honey bees tolerance to Varroa in Mex-ico: mite infertility is not the main tolerance fac-tor, XXXVth Int. APIMONDIA Congress,Antwerp, Belgium.

Wallner A. (1990) Beobachtungen naturlicher Varroa-abwehrreaktionen in meinen bienñenvolkern(Observations of honey bee natural Varroa-defen-sive reactions), Imkerfreund 9, 4–5.

Wallner K. (1999) Varroacides and their residues inbee products, Apidologie 30, 235–248.

174

To access this journal online:www.edpsciences.org