a method to feed individual bees 2003

7/31/2019 A Method to Feed Individual Bees 2003

1/6

597Apidologie 34 (2003) 597602 INRA/DIB-AGIB/ EDP Sciences, 2003DOI: 10.1051/apido:2003055

Original article

A method to feed individual bees (Hymenoptera:Apiformes) known amounts of pesticides

Edith LADURNERa, Jordi BOSCHb*, Stefano MAINIa, William P. KEMPc

a Dipartimento di Scienze e Tecnologie Agroambientali, Universit di Bologna, 40126 Bologna, Italyb Biology Department, Utah State University, Logan, UT 84322-5310, USA

c USDA-ARS, Bee Biology & Systematics Laboratory, Logan, UT 84322-5310, USA

(Received 10 October 2002; revised 1 February 2003; accepted 16 May 2003)

Abstract We devised a simple method (flower) to feed bees individually, and compared it with twoother methods commonly used (film canister and glass vial). We tested the three methods on twosolitary species, Osmia lignaria andMegachile rotundata, and one social species,Apis mellifera, under fourdifferent light regimes (natural, artificial, plant growth and darkness). The flower method was the mosteffective for all three bee species: 9095% of the bees fed under natural light, 8095% under artificial light,75100% under plant growth light, and 4570% in darkness. Percent success was 050% with the filmcanister method, and 060% with the glass vial method. The flower method may allow more comprehensivefuture evaluation of pesticide effects on bees.

oral toxicity test / individual feeding /Osmia lignaria/Apis mellifera /Megachile rotundata

1. INTRODUCTION

Bee populations are often exposed to pesti-cide treatments in both natural and agriculturalecosystems (Kevan, 1975; Crane and Walker,1983; Johansen and Mayer, 1990; Peach et al.,1993), and insecticide poisoning is consideredone of the main causes of bee populationdeclines worldwide (Dias and Raw, 1999;Tepedino and Ginsberg, 2000). Given the eco-logical and economic importance of bees aspollinators of wild flowers and cultivatedplants (Kevan, 1991; Southwick andSouthwick, 1992), it is surprising that ourknowledge on bee toxicity is so fragmentaryand mostly restricted to one species, the honeybee,Apis mellifera L. (Apidae) (Johansen andMayer, 1990). Information on pesticide toxic-ity to non-Apis bees is dismally scarce, and

limited to a handful of species managed forcrop pollination (review in Tasi, 2002). A

common first step in bee toxicity studies is theestablishment of LD50 (median lethal dose)values, i.e. the dose, expressed in g of activeingredient per insect, inducing 50% mortalityfollowing application (contact or oral) ofmeasured amounts of active ingredients orcommercial pesticide formulations. The meth-ods used for determining LD

50values on A.

mellifera are defined by official guidelines,both in Europe and in the USA (US EPA,1996; OEPP/EPPO, 2001; Cluzeau, 2002).Methods used to study contact toxicity in A.mellifera can be easily applied to other spe-cies. However, in studies of oral toxicity inA.mellifera a common feeder is provided to agroup of workers assuming that, throughtrophallaxis, all individuals will receive simi-lar doses of test solution (OEPP/EPPO, 2001).Group feeding is not applicable to most other

bee species, which do not perform trophallaxisand thus require individual feeding. In fact,

* Corresponding author: [email protected]


2/6

598 E. Ladurner et al.

oral tests on non-Apis bees are rarely con-ducted because individual feeding is consid-ered time consuming (Felton et al., 1986).

In 2002, we initiated a study on the effect ofpesticide sprays on the solitary bee Osmia lig-naria Say (Megachilidae). We tried to feedindividual O. lignaria known amounts ofsugar solution using some of the methodsavailable from the literature (Tasi, 2002), butour success rate (percent individuals thatwould feed) was very low. Thus, we devised anew method, which we tested in comparison tothe two most commonly used methods availa-ble (Johansen et al., 1984; van der Steen et al.,

1996; Bortolotti et al., 2002; Patetta et al.,2002). All three methods were tested on twosolitary bee species O. lignaria andMegachilerotundata (Fabricius) (Megachilidae), and thesocial bee A. mellifera. All three species areused as crop pollinators (Richards, 1984; Free,1993; Bosch and Kemp, 2001), and thus arefrequently exposed to pesticides.

2. MATERIALS AND METHODS

Apis mellifera workers were captured in themorning at the entrance of a hive and brought to thelaboratory, where they were chilled for a maximumof 30 min at 4 C prior to being assigned to the dif-ferent feeding methods. Osmia lignaria and M.rotundata wintering females within their cocoonswere incubated until emergence at 25 C and 29 C,respectively. Upon emergence, females were trans-ferred to a holding cage (40 30 30 cm) to allowthem to deposit the meconium. Females werestarved overnight and then assigned to the differentfeeding methods. No chilling was necessary foreither O. lignaria orM. rotundata.

We used the following three individual feedingmethods: (1) Film canister method (after van derSteen et al., 1996; Bortolotti et al., 2002; Patettaet al., 2002). Bees were individually transferred toblack film canisters (3 cm diameter, 5 cm height)with a small hole drilled on the side, near the base.The test solution was pipetted onto a microscopeslide next to the hole. (2) Glass vial method (afterJohansen et al., 1984). Bees were individuallytransferred to 12-mL glass vials with plastic snapcaps. The test solution was injected into a segmentof plastic tubing (length 15 mm, inside diameter

2.9 mm), fitted snugly into the plastic snap caps.(3) Flower method (Fig. 1). A tiny plastic ampoule(inside diameter 2 mm, outside diameter 3 mm,height 5 mm) was inserted into the calyx of aflower, whose reproductive column had been

previously removed with a pair of forceps. Theplastic ampoule was built by approaching one tip ofa 5-mm section of polyethylene tubing to a flame orsome other heat source. The heat caused the poly-ethylene tip to melt, sealing the tubing orifice.Because the polyethylene never actually touchedthe heat source, no charring occurred, and ampouleshad no obvious odor. The test solution was pipettedinto the ampoule. Flowers and bees were individu-ally housed in ice cream cups made of waxed card-board (8 cm diameter, 5 cm height) covered with aplastic Petri dish lid. Flowers were positioned on an

inverted glass vial stopper. A wire mesh screeninsert (mesh size 2 1 mm) in the Petri dish lid pro-vided adequate aeration. To facilitate flower manip-ulation, we used large, actinomorphic flowers withopen corollas. In preliminary trials, we tested two or

Figure 1. Feeding unit for the flower method.Cross-section drawing and photograph ofMegachile rotundata female on test flower. Theplastic ampoule containing the test solution can beseen at the center of the corolla. am: ampoule. ca:flower calyx. co: flower corolla. st: stopper (flowerholder). ts: test solution.


3/6

Method to feed bees in the laboratory 599

more flowers per bee species, including cherry(Prunus avium L.), apple (Malus domestica Borkh),morning glory (Convolvulus arvensis L.) and peri-winkle (Vinca minorL.). All flowers in these pre-liminary trials yielded similar results. In the actualexperiments we used cherry flowers for the early-flying O. lignaria and morning glory flowers for thelater-flying M. rotundata and A. mellifera. Atrained technician could prepare 100 test flowers in40 minutes.

In all three methods, a 10-L-drop of test solu-tion (25% volume of sucrose to 75% volume ofwater) was offered to the bee for one hour. Wetested each method under four different lightregimes: (1) Natural light. Feeding units placed out-doors (on sunny to partially-cloudy days); (2) Arti-ficial light 1. Feeding units placed in an incubatorwith two 15W Cool White Sylvania fluorescenttubes (wavelength peaks at 405, 430, 545 and

575 nm) placed 15 cm above the feeding units; (3)Artificial light 2. Feeding units placed in an incuba-tor with two 20W Gro-Lux/Aquarium StandardSylvania fluorescent tubes (usually used to helpplant growth; wavelength peaks at 460, 545, 630and 660 nm) placed 15 cm above the feeding units;(4) Darkness. Feeding units placed in an incubatorwith no light. We tested darkness because currentguidelines for group feeding toxicology assays areusually conducted in darkness (Arzone and Vidano,1980; OEPP/EPPO, 2001). Temperatures during alltests (including those outdoors) were 22 2 C forO. lignaria and 25 2 C forM. rotundata andA.

mellifera. We used a lower temperature for O. lig-naria because this spring-flying bee is active atlower temperatures thanM. rotundata orA. mellif-era (Burril and Dietz, 1981; Lerer et al., 1982;Bosch and Kemp, 2001).

Sample sizes were 20 individuals of each species(workers inA. mellifera, females in O. lignaria andM. rotundata) per feeding method/light regime. Weanalyzed the proportion of feeding bees as a func-tion of feeding method and light regime. Becausedifferent bee species were tested under differenttemperatures and with different flowers, we ana-lyzed the results of each species separately. Due tothe number of cells with frequencies of < 5 for oneof the species (O. lignaria, Tab. I), we attemptedto fit an exact Logit model using LogXact(Version 5.0). However, the computational size wastoo large for this exact method to work (even on asufficiently equipped computer). In addition, thecomplete separation of the data prevented the com-putation of Likelihood Ratio and Wald test statis-tics. Therefore, because the structure of the O. lig-naria data prevented the use of an exact test, weused an asymptotic Logit model (GENMOD proce-dure, SAS Institute Inc., 1989; Allison, 1999).Because of the pronounced pattern of the results(O. lignaria was the species for which differencesamong feeding methods were most extreme), it isunlikely that the use of this statistical analysiswould compromise our interpretations.

3. RESULTS

The flower method was, by far, the mosteffective for all three bee species (Tab. I):9095% of the bees fed under natural light,8095% under Cool White fluorescent tubes,75100% under Gro-Lux fluorescent tubes,and 4570% in darkness. All bees that fed withthe flower method consumed all 10 L of thetest solution. Percent success was 060% with

Table I. Percent feeding success in Osmia lignaria females, Megachile rotundata females, and Apismellifera workers offered 10L of sugar solution using three individual feeding methods under four lightregimes (n = 20 per species/feeding method/light regime).

Light regime

Bee species Feedingmethod

Naturallight

Cool-whitelight

Gro-Luxlight

Darkness

O. lignaria Film canister 5 0 0 5

Glass vial 5 0 0 10

Flower 90 80 75 70

M. rotundata Film canister 15 25 10 15

Glass vial 20 60 25 40

Flower 95 90 100 70

A. mellifera Film canister 35 50 30 20

Glass vial 45 50 60 15

Flower 95 95 80 45


4/6


5/6

Method to feed bees in the laboratory 601

as many as 100 flowers can be prepared in40 minutes. To further save time, we tried todeposit the droplet of sugar solution directlyon the flower corolla (without the polyethyl-ene ampoule). However, at least when sys-temic pesticides were added to the sugar solu-tion, the droplet was partially absorbed by theflower tissues. In a subsequent study, theflower method (with the ampoule and underartificial light) was used to assess the toxicityof several pesticides to O. lignaria andA. mel-lifera. Success (feeding) rates were 97.7% forthe first species, and 88.2% for the second (n =871 and 755, respectively) (E.L., unpublished

data). In conclusion, the flower method is asimple, highly effective procedure that couldhelp simplify and standardize oral toxicitytests on a diversity of bee species in the labo-ratory. To further standardize procedures,future studies should investigate the possibil-ity of substituting natural flowers by artificialflowers properly colored and scented.

ACKNOWLEDGEMENTS

We are grateful to S. Kalaskar and G. Trostle

(USDA, Logan) for their assistance, to L. Bortolotti(INA, Bologna) and J.-N. Tasi (INRA, Lusignan)for reviewing the manuscript, and to S. Durham andC. Corcoran (Utah State University, Logan) forstatistical advise. This study was partially supportedby a Ph.D. scholarship from the University ofBologna to E. L.

Rsum Mthode pour administrer individuel-lement aux abeilles (Hymenoptera, Apiformes)une quantit dfinie de pesticides. Les tudes enlaboratoire sur laction des pesticides sur lesabeilles cartent souvent les tests de toxicit orale cause de la difficult dadministrer des quantitsdfinies de solutions tester. Les tudes surlAbeille domestique (Apis mellifera L.) reposentsur des mthodes de nourrissement de groupe etconsidrent que des quantits sensiblement galesde solution sont ingres par chaque ouvrire tra-vers les changes de nourriture (trophallaxie).Nanmoins la plupart des abeilles (s.l.) ne prati-quent pas la trophallaxie. Nous avons mis au pointune mthode simple (la mthode de la fleur )pour nourrir des abeilles individuellement etlavons compare deux autres mthodes couram-ment employes (les mthodes de la bote pourpellicule photo et celle du flacon en verre ).

Nous avons test les 3 mthodes sur 2 abeilles soli-taires, Osmia lignaria Say etMegachile rotundata(Fabricius) (Megachilidae), et sur une espcesociale, A. mellifera, sous 4 rgimes de lumirediffrents (lumire naturelle, tubes fluorescents

blanc froid, tubes fluorescents Gro-Lux et obscu-rit). Dans la mthode de la fleur, les abeilles taientconfines individuellement dans une coupe glace.

On leur donnait une fleur dont le pistil avait t rem-plac par une fine ampoule dans laquelle la solution tester tait pipete (Fig. 1). La mthode de la fleura t la plus efficace chez les 3 espces dabeilles(Tab. I), quel que soit le type dclairage : 9095 %en lumire naturelle, 8095 % avec tubes fluores-cents blanc froid, 75100 % avec tubes fluorescentsGro-Lux et 4570 % en obscurit. Le pourcentagede succs a t compris entre 0 et 50 % avec lamthode de la bote pour pellicule photo et entre 0et 60 % avec la mthode du flacon de verre. La pr-paration des units de nourrissement demandaitplus de travail dans la mthode de la fleur mais avecde la pratique, on peut prparer 100 fleurs en40 min. En raison de sa grande efficacit, lamthode peut en fait conomiser du temps (entermes de rcolte dabeilles et de prparation) et dumatriel animal.

Osmia lignaria / Apis mellifera / Megachilerotundata / test de laboratoire / toxicit orale /nourrissement individuel

Zusammenfassung Eine Methode zur Ftte-rung von Bienen (Hymenoptera: Apiformes) mitdefinierten Mengen von Pestiziden. Bei Labor-

versuchen ber die Wirkung von Pestiziden aufBienen scheiden oft orale Toxizittstests wegen derSchwierigkeit aus, Bienen individuell mit definier-ten Mengen der Testlsungen zu fttern. Versuchebei der Honigbiene, Apis mellifera L. (Apidae),beruhen auf der Ftterung einer Gruppe. Dabei wirdangenommen, dass alle Bienen ber Futteraus-tausch (Trophallaxis) etwa die gleiche Menge derLsung aufnehmen. Die meisten Bienenarten habenjedoch keine Trophallaxis. Wir entwickelten eineeinfache Methode (Blten Methode), um Bienenindividuell zu fttern und verglichen die Effektivi-tt mit zwei anderen allgemein benutzten Methoden(film canister und glass vial Methoden). Wirtesteten die 3 Methoden bei zwei solitren Arten,Osmia lignaria Say und Megachile rotundata(Fabricius) (Megachilidae), und einer sozialen Art,A. mellifera, unter 4 unterschiedlichen Lichtbedin-gungen (natrliches Licht, kalte wei fluoreszieren-de Rhren, Gro-Lux fluoreszierende Rhren undDunkelheit). Bei der Bltenmethode wurden Bie-nen individuell in einem Eiskrembecher gekfigt.Ihnen wurden Blten angeboten, deren Stempel ent-fernt und durch eine winzige Ampulle ersetztworden war, in die die Testlsung pipettiert wurde(Abb. 1). Die Bltenmethode war bei allen 3 Bienen-arten am erfolgreichsten (Tab. I): 9095 % der Bie-

nen, die bei natrlichem Licht gefttert wurden,8095 % bei kalt wei fluoreszierendem Licht,75100 % bei Gro-Lux fluoreszierendem Licht und4570 % in Dunkelheit. Der prozentuale Erfolg beider film canister Methode lag bei 050 % und bei


6/6

602 E. Ladurner et al.

der glass vial Methode bei 060 %. Die Anferti-gung der Futtervorrichtung war bei der Blten-methode am aufwendigsten. Nach einiger bung

kann man jedoch 100 Testblten in 40 Minutenanfertigen und auf Grund der hohen Effektivittspart man mit dieser Methode im Endeffekt sowohlZeit (bei Bercksichtigung der Zeit fr dieSammlung von Bienen und ihrer Prparation) alsauch Tiermaterial.

oraler Toxizittstest / individuelle Ftterung /Osmia lignaria / Apis mellifera / Megachilerotundata

REFERENCES

Allison P.D. (1999) Logistic regression using the SASSystem, SAS Institute Inc., Cary, NC.

Arzone A., Vidano C. (1980) Methods for testingpesticide toxicity to honey bees, Boll. Lab.Entomol. Agr. Filippo Silvestri 37, 161165.

Bortolotti L., Porrini C., Sbrenna G. (2002) Effettidell'imidacloprid nei confronti di Bombusterrestris (L.), Prove di laboratorio, Inf. Fitopatol.3, 6671.

Bosch J., Kemp W.P. (2001) How to manage the blueorchard bee as an orchard pollinator. HandbookSeries, Book 5. Sustainable Agriculture Network,Beltsville, MD.

Burrill R.M., Dietz A. (1981) The response ofhoneybees to variations in solar radiation andtemperature, Apidologie 12, 319328.

Crane E., Walker P. (1983) The impact of pestmanagement on bees and pollination, IBRA,Cardiff, UK.

Cluzeau S. (2002) Risk assessment of plant protectionproducts on honey bees: regulatory aspects, in:Devillers J., Pham-Delgue M.-H. (Eds.), Honeybees: estimating the environmental impact ofchemicals, Taylor & Francis, London, pp. 4255.

Dias B.S.F., Raw A., Imperatriz-Fonseca V.L. (1999)International pollinator initiative, the So Paulodeclaration on pollinators, Ministrio do Meio

Ambiente, Braslia DF, http://www.bdt.fat.org.br/polinizadores/declaration (consulted on16 September 2003).

Felton J.C., Oomen P.A., Stevenson J.H. (1986)Toxicity and hazard of pesticides to honeybees:harmonization of test methods, Bee World 67,114124.

Free J.B. (1993) Insect pollination of crops, 2nd ed.,Academic Press, London.

Johansen C.A., Rincker C.M., George D.A., MayerD.F., Kious C.W. (1984) Effects of Aldicarb andits biologically active metabolites on bees,Environ. Entomol. 13, 13861398.

Johansen C.A., Mayer D.F. (1990) Pollinator

Protection: A bee and pesticide handbook,WicWas, Cheshire, CT.Kevan P.G. (1975) Forest application of the

insecticide Fenitrothion and its effect on wild beepollinators (Hymenoptera: Apoidea) of lowbush

blueberries (Vaccinium spp.) in southern NewBrunswick, Canada, Biol. Conserv. 7, 301309.

Kevan P.G. (1991) Pollination: keystone process in

sustainable global productivity, Acta Hortic. 288,103110.Lerer H., Bailey W.G., Mills P.F., Pankiw P. (1982)

Pollination activity of Megachile rotundata(Hymenoptera: Apoidea), Environ. Entomol. 11,9971000.

Naumann K., Currie R.W., Isman M.B. (1994)Evaluation of repellent effects of a Neeminsecticide on foraging honey bees and otherpollinators, Can. Entomol. 126, 225230.

OEPP/EPPO (2001) Revised draft of EPPOGuidelines PP 1/170: Guidelines for the efficacyevaluation of plant protection products: side-effects on honeybees. Appendix I, in: Belzunces

L.P., Plissier C., Lewis G.B. (Eds.), Hazards ofpesticides to bees, Les Colloques de lINRA,n 98, Paris, pp. 279288.

Patetta A., Marletto F., Manino A. (2002) Pesticidesand bumblebees: poisoning tests in the labora-tory, Insect Soc. Life 4, 101106.

Peach M.L., Tepedino V.J., Alston D.G., GriswoldT.L. (1993) Insecticide treatments for rangelandgrasshoppers: potential effects on the reproduc-tion of Pediocactus sileri (Englem.) Benson(Cactaceae), pp. 309319, in: Sivinski R.,Lightfoot K. (Eds.), Proc. Southwestern Rare andEndangered Plant Conf., Santa Fe, NM.

Richards K.W. (1984) Alfalfa leafcutter bee

management in Western Canada, AgricultureCanada Publication 1495/E, Ottawa, Canada.SAS Institute Inc. (1989) SAS/STAT Users guide,

Version 6. Cary, NC.Solomon M.G., Hooker K.J.M. (1989) Chemical

repellents for reducing hazard to honeybees inapple orchards, J. Apic. Res. 28, 223227.

Southwick E.E., Southwick L. (1992) Estimating theeconomic value of honey bees (Hymenoptera:Apidae) as agricultural pollinators in the UnitedStates, J. Econ. Entomol. 85, 621633.

Tasi J.N. (2002) Impact of agrochemicals on non-Apis bees, in: Devillers J., Pham-Delgue M.-H.(Eds.), Honey bees: estimating the environmental

impact of chemicals, Taylor & Francis, London,pp. 101131.Tepedino V.J., Ginsberg H.S. (2000) Report of the US

Department of Agriculture and US Department ofthe Interior joint workshop on declining pollina-tors. Information and Technology Report USGS/BRD/ITR 2000-0007, Fort Belvoir, VA.

Thompson H.M. (2003) Behavioural effects ofpesticides in bees their potential for use in riskassessment, Ecotoxicol. 12, 317330.

US EPA (1996) Honey bee acute contact toxicity test(OPPTS 850.3020). Ecological effects testguidelines, EPA 712-C-96-147, Washington, DC.

van der Steen J.J.M., Gretenkord C., Schaeffer H.

(1996) Methods to determine acute oral andcontact LD50 of pesticides for bumblebees(Bombus terrestris L.), Appendix 28, in: Lewis G.(Ed.), Proc. VIth Int. Symp. on Hazards ofpesticides to bees, Braunschweig.

a method to feed individual bees 2003

Documents