A Non-lethal Water-based Removal-reapplication Techniquefor Behavioral Analysis of Cuticular Compounds of Ants

Olivier Roux & Jean-Michel Martin &

Nathan Tene Ghomsi & Alain Dejean

Received: 3 March 2009 /Revised: 6 July 2009 /Accepted: 23 July 2009 /Published online: 4 August 2009# Springer Science + Business Media, LLC 2009

Abstract Interspecific relationships among insects areoften mediated by chemical cues, including non-volatilecuticular compounds. Most of these compounds are hydro-carbons that necessitate the use of solvents for theirextraction, identification, and manipulation during behav-ioral assays. The toxicity of these solvents often precludesthe removal and reapplication of hydrocarbons from and tolive insects. As a consequence, dummies often are used inbehavioral assays, but their passivity can bias the behaviorof the responding insects. To overcome these limitations,we propose a method where cuticular compounds areextracted from live ants by placing them into glass vialshalf-filled with tepid water (ca. 34°C) and vigorouslyshaking the vials to form an emulsion whose supernatantcan be analyzed and/or reapplied to other ants. Wedemonstrate that cuticular compounds can be extractedfrom workers of the red fire ant, Solenopsis saevissima, andreapplied to the cuticle of workers from a sympatricspecies, Camponotus blandus (both Hymenoptera: Formi-cidae), while keeping the ants alive. Gas chromatographic-

mass spectrometric analysis and behavioral assays wereused to confirm the successful transfer of the behaviorallyactive compounds.

Keywords Ants . Behavioral assay . Behavioral ecology .

Camponotus blandus . Chemical composition .

Chemical ecology . Cuticular hydrocarbonsGas chromatography-mass spectrometry . Live dummies .

Solenopsis saevissima . Methods

Introduction

Interspecific relationships among insects are often mediatedby chemical cues, which generally are complex blends ofcompounds. Non- or weakly volatile cuticular compounds,whose primary role is to protect individuals from desicca-tion, are frequently involved in intra- and inter-specificrecognition mechanisms. They are particularly important insocial insects because they mediate conflicts and coopera-tive relationships (see Lenoir et al. 2001, for review).

Except for solid phase injection, which consists inplacing a part of an insect into a glass capsule that is thenplaced directly into the gas chromatograph injector andcrushed (Bagnères and Morgan 1990; Morgan 1990), moststudies that examine the role of cuticular compoundsrequire the use of solvents to extract and identify them.Also, solvents can be used to remove, and then reapplycuticular compounds for bioassays (Ruther et al. 2002;Dani et al. 2005). The use of solvents like hexane orpentane creates biases in bioassays because they are toxicand sometimes lethal. Consequently, solvent extractsgenerally are applied onto dummies made of glass pellets(Wagner et al. 2000; Akino et al. 2004), filter paper(Colazza et al. 2007), or dead insects whose cuticular

O. RouxLaboratoire d’Ecologie Fonctionnelle (UMR-CNRS 5245),Université Paul Sabatier,118 route de Narbonne,31062 Toulouse cedex 09, France

O. Roux (*) : J.-M. Martin :A. DejeanÉcologie des Forêts de Guyane (UMR-CNRS 8172),Campus agronomique, BP 709,97379 Kourou cedex, Francee-mail: oroux@cict.fr

N. T. GhomsiLaboratoire de Chimie et de Biochimie des Interactions,Centre Universitaire JF Champollion,Place de Verdun,81012 Albi cedex 9, France

J Chem Ecol (2009) 35:904–912DOI 10.1007/s10886-009-9673-x

compounds were washed off prior to the assay (Ruther et al.2002; Roux et al. 2007). However, lifeless dummies areinadequate during bioassays when the range of responsesfrom each interacting individual depends on the behavior ofthe others (Gamboa et al. 1991; Roulston et al. 2003).

A good way to avoid this problem consists of insertinglive insects into a glass vial whose inner walls werepreviously coated with an extract. Although this methodhas been used successfully with honey bees, which rubtheir wings, thorax, and abdomen on the coated insides ofthe walls (Dani et al. 2005), the technique is not suitable fortransferring semiochemicals to non-flying insects that walkon the walls.

There are non-destructive, nontoxic extraction techni-ques, such as solid phase micro extraction (SPME), whichconsists of gently rubbing the insects with a glass fibercoated with a membrane with an affinity for cuticularcompounds. The fiber then is introduced into the gaschromatograph injector where the compounds are desorbed.Results obtained through this method are similar to thoseobtained with solvent extracts (Monnin et al. 1998;Turillazzi et al. 1998; Tentschert et al. 2002). Turillazzi etal. (1998) adapted this method by replacing the fiber with aclean piece of cotton wool or filter paper. After rubbing theinsects with substrate, the compounds are rinsed from thecotton wool or filter paper with a solvent, and thenanalyzed. These extractions by “abrasion” allow the insectsto be kept alive so as to monitor variations in their cuticularprofiles at different stages of their lives; it provides resultssimilar to those of solvent extracts (with some variations inrelative proportions: Turillazzi et al. 1998). In addition tothe cost and fragility of SPME fibers, it is difficult to usethese techniques on small and/or fragile insects; moreover,the analyses must follow the extraction as quickly aspossible because the time during which compounds can beconserved is limited. Adaptation with cotton wool allowssome of these disadvantages to be avoided, but cotton woolis slightly less sensitive and can catch on spurs on thecuticle.

Another non-destructive, nontoxic extraction methodinvolves rinsing individuals with water and using theaqueous extract in behavioral assays. Andrews (1911)noted that water-washed termites were attacked by nest-mates as well as by alien individuals. Fresneau (1980)successfully transferred recognition signals in Pachycon-dyla villosa, a ponerine ant. By rinsing workers belongingto two different colonies, and then soaking them in bathsalready used to wash alien individuals, he was able toinduce aggressiveness among nestmates and tolerancebetween alien workers. This method was used by Dejeanet al. (1990) to test learning behavior in P. villosa workersthat adopt a posture of prudence when encountering termitesoldiers, but not termite workers. Rinsed termite workers

soaked in the “bath” water of soldiers triggered a posture ofprudence in ants; whereas, rinsed termite soldiers soaked inthe “bath” water of workers did not induce any prudence,so that termite soldiers viciously bit the ants. The authorsconcluded that compounds, here serving as kairomones,were transferred (after emulsion) between the termitesoldiers and workers.

Henderson et al. (1990) wanted to test several methodsfor removing nestmate discrimination pheromones and thensuppressing aggressive behavior between alien Formicamontana. The methods used included washing the ants withdistilled water; however, this technique yielded poor resultsin comparison to hexane or pentane washes because waterdoes not wash as well as “classical solvents.” Indeed, aftera water-based wash, many cuticular compounds remain onthe insect cuticle, so this method does not always allow theaggressive behavior of non-nestmate individuals to becompletely inhibited (O.R. pers. obs.). In these reports,the insects were rinsed by vigorously shaking them in vialshalf-filled with water to emulsify their cuticular hydro-carbons. The insects, at first numbed, recovered after a fewminutes and could subsequently be used in bioassays.However, no chemical analyses then were performed toverify which compounds were transferred.

In this study, we tested the hypothesis that cuticularcompounds can be extracted and transferred between antsby using water followed by analysis of the chemicalcomposition of the extracts and behavioral assay of theants’ responses in the laboratory.

Methods and Materials

Insects The study was conducted with Camponotusblandus (Smith 1858) (Formicinae) and Solenopsis sae-vissima (Smith 1855) (Myrmicinae) (both Hymenoptera:Formicidae), two Neotropical species found frequently inopen habitats. In October 2007 and February 2009, wegathered workers belonging to one colony from eachspecies located about 20 m from each other along a foresttrail situated near the Petit Saut dam (French Guiana).

Cuticular Compound Extraction Cuticular compoundswere extracted by using two methods. First, twelveindividuals from each species were freeze-killed, and thecompounds extracted by placing the ants into 500 µl ofhexane in a 2 ml glass vial for 5 min. The solvent wasevaporated under a nitrogen stream, and the sample wasre-suspended in 20 µl of hexane. Second, we used amethod designed to allow the insects to be kept alivewherein thirty individuals were placed into a 30 mlglass vial along with about 20 ml of ultra pure water(34°C) and hand shaken vigorously for 5 min with the

J Chem Ecol (2009) 35:904–912 905

aim of emulsifying the cuticular hydrocarbons. The antswere removed, and the aqueous emulsion was extractedwith 500 µl of hexane by shaking the vials for 10 secbefore allowing the contents to settle for 1 min. Thehexane supernatant was transferred to a 2 ml glass vial,removed under a nitrogen stream, and the residue re-suspended in 20 µl of hexane for gas chromatographic-mass spectrometric (GC-MS) analysis.

Applying Cuticular Hydrocarbons to the Ants Camponotusblandus Workers were applied with the cuticular com-pounds of S. saevissima workers in two different ways. Inboth cases, about 70 workers from the two species wererinsed separately in 20 ml of water as described previous-ly. The ants then were removed and placed into plasticboxes whose walls were coated with fluon® to preventthem from climbing out, and whose floor was lined withabsorbent paper on which the ants dried. For the firstapplication method, rinsed C. blandus were placeddirectly into the emulsion of S. saevissima cuticularcompounds, shaken for 10 sec, and then left in theemulsion for 5 min. They then were removed and placedagain into a plastic box where they were redried. To seehow effectively the cuticular compounds were transferred,some individuals were freeze-killed, and their cuticularcompounds extracted with hexane prior to GC-MSanalysis. For the second application method, the water-based emulsion was extracted with hexane as previouslydescribed, and the residual hexane was carefully removedand the “bath” water transferred to a clean vial. Only thenwere the rinsed C. blandus placed into the water fromwhich the hexane fraction (containing compounds with agreater affinity for hexane than water) was removed. Thissecond technique was used to test the action of theaqueous emulsion alone.

Chemical Analysis Chemical analysis and the identifica-tion of compounds were conducted with a FinniganTrace GC-MS 2000 Series chromatograph directlycoupled to a mass spectrometer quadrupole detector.The entire system was operated by using the Xcaliburdata system, version 1.2. MS spectra were recorded inthe EI mode (70 eV) over a mass range of 50–550 massunits with two scans per second. One µl was injectedinto the splitless mode with an injector temperature of280°C and a detector temperature of 300°C. An apolarRtx®-5MS-MS capillary column (30 m×0.25 mm i.d.×0.25 µm film thickness, 5% diphenyl and 95% dime-thylpolysiloxane) was used. Elution was carried out withhelium at 1 ml/min. The oven temperature wasprogrammed from 80°C (for 30 sec) to 200°C at20°C/min, from 200°C to 240°C at 3°C/min, from240°C to 300°C at 30°C/min, and left at 300°C for

19 min. As it is important only to check if thecompounds in the water-based extract are the same asin the hexane-based extract, only tentative identificationswere made by comparing them with a spectral database,retention indices, a series of standard linear alkanes, andthe calculation of methyl positions on the basis of massspectra. Then, these tentative identifications of thecompounds were provided for informational purposesonly; some were cross-checked with mass spectra foundin the literature. Then, the alkaloids were identifiedbased on comparisons with the work of MacConnell etal. (1971), Brand et al. (1972) and, more recently, Chenand Fadamiro (2009a, b). Some methylalkanes wereidentified through comparisons with the results obtainedby Nelson et al. (1980) on the cuticular methylalkanes ofsome Solenopsis species.

Behavioral Assays To test the effectiveness of our extractionand application methods, five sets of ants were prepared: (1)Untreated C. blandus workers; (2) untreated S. saevissimaworkers; (3) C. blandus workers with the applicationmethod 1 thought to contain the cuticular extracts of S.saevissima workers; (4) C. blandus workers with applica-tion method 2 consisting of the water-based extract of the S.saevissima workers without compounds soluble in hexane;and (5) “blank” C. blandus workers, i.e., individuals thatwere rinsed in two different baths of clean water.

Behavioral interactions were tested between the fol-lowing workers: (1) untreated C. blandus vs. untreated C.blandus; (2) untreated C. blandus vs. “blank” C. blandus;(3) untreated C. blandus vs. C. blandus with applicationmethod 2; (4) untreated C. blandus vs. C. blandus withapplication method 1; (5) untreated S. saevissima vs. C.blandus with application method 1; and (6) untreated C.blandus vs. untreated S. saevissima. All tested C. blandusworkers belonged to the same colony. For each behavioralinteraction, two workers were placed together for 5 min ina neutral arena (a Petri dish; 60 mm diam and 20 mmdeep) whose walls were coated with Fluon® to prevent theants from climbing out. We scored the interactions asfollows: 1 = antennal contact (physical contact, but noaggressive response; may include trophallaxis), 2 = escape(sudden U-turn and running quickly), 3 = aggressiveness(brief biting by one or both of the workers; may includejerking and opening the mandibles), and 4 = fighting(prolonged aggressiveness, including prolonged biting andstinging [Solenopsis] or venom spraying [Camponotus]).We repeated each kind of confrontation 60 times, keepingthe highest value noted for each series of 5-min observa-tions, and used each worker only once. Because S.saevissima is active around the clock and C. blandus istypically diurnal (Orivel and Dejean 2002), all bioassayswere performed during the day.

906 J Chem Ecol (2009) 35:904–912

Statistical Analyses Levels of aggressiveness betweeninteracting ants were compared by using the Kruskal–Wallis test. A post-hoc test (Dunn’s multiple pairwisecomparisons test) was then performed to isolate the dyadicinteractions that differed from the others. All statisticalanalyses were performed with GraphPad Prism 4.03software (GraphPad Software Inc 2005).

Results

Chemical Analysis and Extraction Methods The twomethods of extraction produced both quantitatively (interms of proportions) and qualitatively similar GC-MSprofiles for both ant species (Fig. 1).

The GC-MS profiles of the two ant species were different(Fig. 1a–d). The GC-MS profiles for C. blandus contained23 compounds, all hydrocarbons, ranging from C27 to C40(Table 1). The S. saevissima profile had 34 compounds; theeight early eluting compounds, representing the bulk of theextracts, were piperidine alkaloids typical of Solenopsisvenom (MacConnell et al. 1971; Brand et al. 1972). Thelater eluting compounds were all hydrocarbons ranging fromC23 to C32. Only heptacosane (peak 26) was common toboth species, but present at trace levels.

Application Methods The gas chromatogram obtained fromrinsed C. blandus treated with the emulsion of S.saevissima cuticular compounds was not a perfect blendof the cuticular signatures of the two species (Fig. 1e). TheS. saevissima alkaloids were transferred poorly to the rinsed

Fig. 1 Gas chromatographic-mass spectrometric traces ofextracts from Camponotusblandus and Solenopsis saevis-sima workers. Analyses wereconducted on extracts preparedby two methods (a–d), as wellas on a hexane-based extractfrom C. blandus workers“made-up” with the emulsion ofS. saevissima cuticular com-pounds (e); the area under thehorizontal line is enlarged fivetimes; * = contamination; seeTable 1 for peak assignments

J Chem Ecol (2009) 35:904–912 907

C. blandus, whereas all of the hydrocarbons from S.saevissima (peaks 9–27) were transferred efficiently, andthe proportions of the transferred hydrocarbon componentswere comparable to those in the native extracts in mostcases (except for compounds 15, 19, 22, and 23, whichwere recovered in smaller proportions than in the extractsof the worker S. saevissima, Table 1).

Behavioral Assays We typically noted the absence ofaggressiveness (level 1; i.e., antennation and trophallax-is; Fig. 2) among untreated C. blandus nestmates that issignificantly different from the high level of aggressive-ness that occurred among untreated C. blandus anduntreated S. saevissima workers (mean aggressivenesslevels: 1±0 and 2.46±0.09, respectively; P<0.001;Fig. 2). The latter triggered some fighting and they mainlyraised and vibrated their gaster (“flagging”), a behaviorcorresponding to the airborne dispersal of venom (seeObin and Vander Meer 1985). As a result, C. blandusworkers ran, jerked away, or fought when contact becamefrequent. Interactions between untreated and “blank” C.blandus (to test the effect of double rinsing) were notsignificantly different from those observed betweenuntreated nestmates (mean aggressiveness levels: 1±0and 1.06±0.04, respectively; NS; Fig. 2). Inversely,interactions between untreated C. blandus workers andthose with the “S make-up” resulted in levels ofaggressiveness significantly higher than the two previoussituations involving C. blandus workers (mean aggres-siveness levels: 1±0 and 2.15±0.13, respectively; P<0.001; Fig. 2). Nevertheless, interactions between untreat-ed C. blandus workers and C. blandus workers with "S-Hmake-up" were not different from interactions withuntreated or “blank” C. blandus workers (Fig. 2). Theinteractions between untreated C. blandus workers and C.blandus workers with application method 1 were notsignificantly different from interactions between untreatedC. blandus workers and untreated S. saevissima workers(mean aggressiveness levels: 2.15±0.13 and 2.46±0.09,respectively; NS; Fig. 2). Also, interactions betweenuntreated S. saevissima workers and C. blandus workerswith application method 1 were not significantly differentfrom those between C. blandus workers and C. blandusworkers with application method 1 (mean aggressivenesslevels: 2±0.15 and 2.15±0.13, respectively; NS; Fig. 2),but they were different from those between untreated C.blandus and untreated S. saevissima workers (meanaggressiveness levels: 2±0.15 and 2.46±0.09, respective-ly; P<0.001; Fig. 2). Untreated C. blandus workersmainly ran away after encountering S. saevissima workersas well as C. blandus workers with the application method1 (level 2 of aggressiveness), but C. blandus workers withthe application method 1 did not run away when they

encountered untreated S. saevissima workers (behavioralobservations not shown).

Discussion

Our study demonstrates that it is possible to extractcuticular compounds from live tropical ants by using atepid water bath, and that the extracted workers remainactive even after being rinsed twice. Also, this methodpermits the cuticular extracts to be reapplied onto other livetropical ants with the aim, as in our study, of transferring apart or all of the cuticular recognition signals.

That ants living in equatorial countries survived ourrinsing treatment is not surprising as they are adapted toliving in areas with heavy rain, particularly ground-nesting species whose nests can be flooded. Whenflooded, the colonies of several species form rafts thatfloat on the flooding waters. This frequently has beenreported for fire ants (Solenopsis invicta, S. saevissima,and Wasmannia auropunctata), Pheidole spp. (all Myrmi-cinae) and the Argentine ant (Linepithema humile;Dolichoderinae) (Hölldobler and Wilson 1990; Morrison1998; Holway et al. 2002; Haight 2006).

Water is a polar solvent and so is not theoreticallyadapted to extracting non-polar compounds. It also is wellknown that cuticular lipids are mainly made up of hydro-carbons, which are not water-soluble compounds. Never-theless, our study has shown that by “relatively vigorously”shaking the glass vial containing the water and the ants,tepid water can be used to extract cuticular hydrocarbons ina way that is similar to using hexane. In fact, these non-polar compounds were not solubilized, but emulsified andfloating on the surface of the water. To explain how theemulsion can transfer cuticular lipids on the ants, Dejean etal. (1990) used the comparison of birds caught in an oilslick, with the cuticular lipids acting as the oily film on thesurface of the water and the live ants as the birds that werecoated with the "pollutant." Even if water cannot extract allof the cuticular lipids (in terms of quantity), it is powerfulenough to extract part of them, and to produce qualitativelythe same extracts as hexane; moreover, the quantityextracted and reapplied is enough to induce the expectedbehaviors. In addition, the extraction of cuticular com-pounds from the “bath” water of S. saevissima workers byadding hexane prior to applying the cuticular compounds ofC. blandus (application method 2) showed that thetransferred signal really is contained in the hexane fractionand could be hydrocarbons.

We point out that the volume of water used in theapplication experiment is less important than the diameterof the vial. Indeed, as we were dealing with an emulsion,

908 J Chem Ecol (2009) 35:904–912

Table 1 Tentative identification and relative quantitiesa of cuticular compounds from Camponotus blandus and Solenopsis saevissima by usingtwo different methods of extraction and for application of S. saevissima cuticular extract to C. blandus workers

Peak Retentionindex

C. blandus S. saevissima Appliedextract

Identification m/z

Hexane H2O Hexane H2O

1 1851 + + tr 2-Methyl-6-(4-undecenyl)piperidine 98-111-124-180-236-251

2 1871 + + + + tr 2-Methyl-6-(undecyl)piperidine 98-238-252

3 2016 + + tr 2-Methyl-6-(4-tridecenyl)piperidine 98-111-124-180-264-278

4 2069 + + + + ++

tr 2-Methyl-6-(4-tridecenyl)piperidine(isomer of peak 3)

98-111-124-180-264-278

5 2081 + + + + tr 2-Methyl-6-(tridecyl)piperidine 98-266-280

6 2094 + + tr 2-Methyl-6-(4,x-tridecadienyl)piperidine 98-111-124-178-220-262-276

7 2271 + + + + tr 2-Methyl-6-(4-pentadecenyl)piperidine 98-154-208-292-307

8 2292 + + tr 2-Methyl-6-(pentadecyl)piperidine 98-294-308

9 2300 + + + n-Tricosane (n-C23) -

10 2332 tr tr tr 11-Methyltricosane (11-meC23) 168/169-196/197

11 2370 + + + 3- Methyltricosane + unknown(3-meC23 + unk.)

56/57-308/309

12 2400 + + + n-Tetracosane (n-C24) -

13 2407 + + + 3,7-Dimethyltricosane (3,7-dimeC23) 57-127-253-323

14 2471 + + + Pentacosene (C25:1) -

15 2500 + + + + + n-Pentacosane (n-C25) -

16 2536 + + + 13- and 11-Methylpentacosane (13 + 11-meC25)

168/167-196/197-224/225

17 2538 tr tr tr 9-Methylpentacosane (9-meC25) 140/141-252/253

18 2557 tr tr tr 5-Methylpentacosane (5-meC25) 84/85-308/309

19 2578 + + + + + 3-Methylpentacosane (3-meC25) 56/57-336/337

20 2590 tr tr tr 5,9-Dimethylpentacosane (5,9-dimeC25) 85-155-253-323

21 2600 tr tr tr n-Hexacosane (n-C26) -

22 2620 + + + + + 3,7- 3,9- and 3,11-Dimethylpentacosane (3,7+3,9+3,11-dimeC25)

56-127-155-183-225-253-281-351

23 2633 + + + + + 3,7,11-Trimethylpentacosane (3,7,11-dimeC25) 127-197-224-295

24 2666 tr tr tr Unknown branched hydrocarbon 141-281

25 2694 + + + 4,14- and 4,18-Dimethylhexacosane(4,14+4,18-dimeC26)

70-141-197-225-281-351

26 2700 tr tr tr tr tr n-Heptacosane (n-C27) -

27 2728 tr tr tr Unknown 140/141-210/211-280/281

28 2800 tr tr tr n-Octacosane (n-C28) -

29 2844 + + + 4-Methyloctacosane (4-meC28) 70/71-364/365

30 2900 + + + n-Nonacosane (n-C29) -

31 2976 tr tr tr 4-Methylnonacosane + Unknown(4-meC29 + unk.)

70/71-378/379

32 3000 + + + n-Triacontane (n-C30) -

33 3040 tr tr tr Unknown branched hydrocarbon 197-224-253-281

34 3073 + + + + + + + + 4-Methyltriacontane + unknown (4-meC30) 70-155-309-393

35 3100 + + + n-Hentriacontane (n-C31) -

36 3200 tr tr tr n-Dotriacontane(n-C32) -

37 3260 tr tr tr Unknown 197-253-281-421

38 3356 + + + + + + 17-Methyltritriacontane (17-meC33) 252/253

39 3457 + + + 16-Methyltetratriacontane (16-meC34) 238/239-280/281

J Chem Ecol (2009) 35:904–912 909

the smaller the diameter of the vial, the greater the height ofthe layer of the supernatant, facilitating the transfer ofcompounds to the ant cuticle. In the choice of the vial, then,there is a trade-off between diameter and the volume ofwater able to contain, in this case, 70 ants.

The main problem with using live insects in a water orpolar solvent rinse is that, in addition to the expected cuticularhydrocarbons, extracts can contain exocrine gland secretionsand regurgitated and excreted items. However, Vander Meerand Morel (1998) pointed out that solvent extracts of deadinsects also contain such compounds. In our study, cuticularhydrocarbons and alkaloids from the venom gland of S.saevissima were obtained through hexane- and water-basedextractions. Like hydrocarbons, fire ant alkaloids are notwater-soluble and less dense than water (Blum et al. 1958).

Nevertheless, they were poorly transferred to the C. blanduscuticle, probably due to their greater polarity.

We noted that during the interactions involving S.saevissima or C. blandus workers with the applicationmethod 1, untreated C. blandus workers ran away morefrequently in the former than in the latter case (70% vs.36.7%). This can be due to the alkaloids contained in the S.saevissima venom that have a repellent function duringheterospecific encounters (see Obin and Vander Meer 1985,for S. invicta). Indeed, that the alkaloids are not transferredduring the reapplication process could explain at least a partof the difference. Also, because aggressive interactions arepotential sources of injury or death, ants rather initiatefights when they must defend their nest, territories or food(Roulston et al. 2003). So, in a neutral arena devoid of food

Fig. 2 Box plot and mean ± se of aggressiveness responses observedduring the six different types of interactions between Camponotusblandus and Solenopsis saevissima. Mean (♦). The box plots indicatethe median (large horizontal bars), the 25th and 75th percentiles(white squares), and the minimum and maximum values (whiskers).

Statistical comparisons were performed by using the Kruskal–Wallistest (H=186.7; df=5; P<0.001) followed by a Dunn’s post hoc test:different letters indicate significant differences (P<0.001). (n) =number of occurrence for each level of aggressiveness

Table 1 (continued)

Peak Retentionindex

C. blandus S. saevissima Appliedextract

Identification m/z

Hexane H2O Hexane H2O

40 3566 + + + ++

+ + 19-Methylpentatriacontane (19-meC35) 280/281-252/253

and 13,21-Dimethylpentatriacontane(13,21-dimeC35)

196/225/323/351

41 3575 tr tr tr Unknown -

42 3600 + + + n-Hexatriacontane (n-C36) -

43 3786 + + + + + + mixed of unknown branched hydrocarbons 197-253-323-379

44 3808 + + + + + + 17-Methyloctatriacontane (17-meC38) +unknown branched hydrocarbons

196-253-267-323-393

tr traces; x unknown position of the double bond; m/z mass-to-charge ratioa +++ = base peak (100%); ++ = peak height>10%; + = peak height<10%

910 J Chem Ecol (2009) 35:904–912

sources, it is in the protagonists’ best interest to avoidconflict. The non-aggressive behavior observed duringinteractions between untreated C. blandus workers and C.blandus workers with the application method 1 could resultfrom the imperfect distribution of the cuticular compoundson the cuticle of the reapplied workers (Ruther et al. 1998).However, during interactions between untreated S. saevis-sima workers and C. blandus workers with the applicationmethod 1, reapplied C. blandus workers ran away less oftenthan in the previous cases. This could be because itsperception of its own "gestalt odor" was altered andpartially matched the odor of S. saevissima workers.Fresneau (1980) found that this may lead to tolerance ofthe alien during intraspecific confrontations.

In summary, we have shown that using water-basedextracts of cuticular compounds is similar to using hexane-based extracts, and that they can be reapplied onto otherindividuals through soaking, all while keeping the individ-uals alive. This method of transferring cuticular compoundsmight be applicable to other insect-insect or even insect-plant interactions.

Acknowledgments We are grateful to Andrea Dejean for proofreadingthe manuscript, to Antoine Steven from the Institut Pasteur de la Guyanefor allowing us the use of a GC-MS, to Felipe Ramon-Portugal fortechnical assistance, and to the staff of the Laboratoire Environnementde Petit Saut for field accommodations. Two anonymous reviewers arethanked for their helpful suggestions, which significantly improved themanuscript. Financial support for this study was provided by theProgramme Amazonie II of the French Centre National de la RechercheScientifique (project 2ID) and the Programme Convergence 2007–2013,Région Guyane from the European Community (project DEGA).

References

AKINO, T., YAMAMURA, K., WAKAMURA, S., and YAMAOKA, R. 2004.Direct behavioral evidence for hydrocarbons as nestmate recog-nition cues in Formica japonica (Hymenoptera: Formicidae).Appl. Entomol. Zool. 39:381–387.

ANDREWS, E. A. 1911. Observation on termites in Jamaica. J. Anim.Behav. 1:193–228.

BAGNÈRES, A. G. and MORGAN, E. D. 1990. A simple method foranalysis of insect cuticular hydrocarbons. J. Chem. Ecol.16:3263–3276.

BLUM, M. S., MURRAY, S., WALKER, J. R., CALLAHAN, P. S., andNOVAK, A. F. 1958. Chemical, insecticidal, and antibioticproperties of fire ant venom. Science 128:306–307.

BRAND, J. M., BLUM, M. S., FALES, H. M., and MACCONNELL, J. G.1972. Fire ant venoms: comparative analyses of alkaloidalcomponents. Toxicon 10:259–271.

CHEN, L. and FADAMIRO, H. Y. 2009a. Re-investigation of venomchemistry of Solenopsis fire ants. I. Identification of novelalkaloids in S. richteri. Toxicon 53:469–478.

CHEN, L. and FADAMIRO, H. Y. 2009b. Re-investigation of venomchemistry of Solenopsis fire ants. II. Identification of novelalkaloids in S. invicta. Toxicon 53:479–486.

COLAZZA, S., AQUILA, G., DE PASQUALE, C., PERI, E., and MILLAR, J.2007. The egg parasitoid Trissolcus basalis uses n-nonadecane, a

cuticular hydrocarbon from its stink bug host Nezara viridula, todiscriminate between female and male hosts. J. Chem. Ecol.33:1405–1420.

DANI, F. R., JONES, G. R., CORSI, S., BEARD, R., PRADELLA, D., andTURILLAZZI, S. 2005. Nestmate recognition cues in the honeybee: differential importance of cuticular alkanes and alkenes.Chem. Senses 30:477–489.

DEJEAN, A., CORBARA, B., and OLIVA-RIVERA, J. 1990. Mise enévidence d’une forme d’apprentissage dans le comportement decapture des proies chez Pachycondyla (=Neoponera) villosa(Formicidae, Ponerinae). Behaviour 115:175–187.

FRESNEAU, D. 1980. Fermeture des sociétés et marquage territorialchez les fourmis ponérines du genre Neoponera. Biol EcolMéditerr 7:205–206.

GAMBOA, G. J., REEVE, H. K., and HOLMES, W. G. 1991. Conceptualissues and methodology in kin recognition research, a criticaldiscussion. Ethol. 88:109–127.

HAIGHT, K. L. 2006. Defensiveness of the fire ant, Solenopsis invicta,is increased during colony rafting. Insect. Soc. 53:32–36.

HENDERSON, G., ANDERSEN, J. F., PHILLIPS, J. K., and JEANNE, R. L.1990. Internest aggression and identification of possible nestmatediscrimination pheromones in polygynous ant Formica montana.J. Chem. Ecol. 16:2217–2228.

HÖLLDOBLER, B. and WILSON, B. 1990. The ants. Harvard UniversityPress, Cambridge

HOLWAY, D. A., LACH, L., SUAREZ, A. V., TSUTSUI, N. D., and CASE,T. J. 2002. The causes and consequences of ant invasions. Annu.Rev. Ecol. Syst. 33:181–233.

LENOIR, A., D’ETTORRE, P., ERRARD, C., and HEFETZ, A. 2001.Chemical ecology and social parasitism in ants. Annu. Rev.Entomol. 46:573–599.

MACCONNELL, J. G., BLUM, M. S., and FALES, H. M. 1971. Thechemistry of fire ant venom. Tetrahedron 27:1129–1139.

MONNIN, T., MALOSSE, C., and PEETERS, C. 1998. Solid-phasemicroextraction and cuticular hydrocarbon differences related toreproductive activity in queenless ant Dinoponera quadriceps. J.Chem. Ecol. 24:473–490.

MORGAN, E. D. 1990. Preparation of small samples for chromatog-raphy of insect pheromones. Anal. Chim. Acta 236:227–235.

MORRISON, L. W. 1998. A review of Bahamian ant (Hymenoptera:Formicidae) biogeography. J. Biogeography 25:561–571.

NELSON, D. R., FATLAND, C. L., HOWARD, R. W., MCDANIEL, C. A.,and BLOMQUIST, G. J. 1980. Re-analysis of the cuticularmethylalkanes of Solenopsis invicta and S. richteri. InsectBiochem. 10:409–418.

OBIN, M. S. and VANDER MEER, R. K. 1985. Gaster flagging by fireants (Solenopsis spp.): functional significance of venom dispersalbehavior. J. Chem. Ecol. 11:1757–1768.

ORIVEL, J. and DEJEAN, A. 2002. Ant activity rhythms in a pioneervegetal formation of French Guiana (Hymenoptera: Formicidae).Sociobiology 39:65–76.

ROULSTON, T. H., BUCZKOWSKI, G., and SILVERMAN, J. 2003.Nestmate discrimination in ants: effect of bioassay on aggressivebehavior. Insect. Soc. 50:151–159.

ROUX, O., GERS, C., TENE-GHOMSI, J. N., ARVANITAKIS, L., BORDAT,D., and LEGAL, L. 2007. Chemical characterization of contactsemiochemicals for host-recognition and host-acceptance by thespecialist parasitoid Cotesia plutellae (Kurdjumov). Chemoecol.17:13–18.

RUTHER, J., SIEBEN, S., and SCHRICKER, B. 1998. Role of cuticularlipids in nestmate recognition of the European hornet Vespacrabro L. (Hymenoptera, Vespidae). Insect Soc. 45:169–179.

RUTHER, J., SIEBEN, S., and SCHRICKER, B. 2002. Nestmaterecognition in social wasps: manipulation of hydrocarbonprofiles induces aggression in the European hornet. Naturwissen-schaften 89:111–114.

J Chem Ecol (2009) 35:904–912 911

TENTSCHERT, J., BESTMANN, H. J., and HEINZE, J. 2002. Cuticularcompounds of workers and queens in two Leptothorax antspecies—a comparison of results obtained by solvent extraction,solid sampling, and SPME. Chemoecol. 12:15–21.

TURILLAZZI, S., SLEDGE, M. F., and MONETI, G. 1998. Use of asimple method for sampling cuticular hydrocarbons from livesocial wasps. Ethol. Ecol. Evol. 10:293–297.

VANDER MEER, R. K. and MOREL, L. 1998. Nestmate recognition inants, pp. 79–103, in R. K. Vander Meer, M. D. Breed, K. E.Espelie, and M. L. Winston (eds.). Pheromone communication insocial insects: Ants, wasps, bees, and termites. Westview, Boulder.

WAGNER, D., TISSOT, M., CUEVAS, W., and GORDON, D. M. 2000.Harvester ants utilize cuticular hydrocarbons in nestmate recog-nition. J. Chem. Ecol. 26:2245–2257.

912 J Chem Ecol (2009) 35:904–912