Learning the Language of Insects- and How to Talk Back - Entomology

Made in United States of AmericaRepiinted from AMERICAN ENTOMOLOGISTVol 39, No 4, Wmter 1993@ 1993 Entomological Society of AmericaLearning the Language of InsectsandHow to Talk BackrIIHAVE BEEN LUCKY over my careerthus far to have been involved in researchon insect pheromones, which inmany ways epitomize the complexity andpower of insect communication, and, for entomologists,present much promise as ameans of manipulating insect behavior.However, in successfully doing our jobs asresearchers, we must not only learn the languageof insects, but we also should pay attentionto communication between insectsand ourselves, and ultimately to communicationamong fellow r esearcher s and society.I favor the definition of communicationoutlined by E. 0. Wilson (1974), whowrote, "communication occurs when theaction of or cue given by one organism isperceived by and thus alters the probabilityof behavior in another organism in a fashionadaptive to either one or both of theparticipants " Communication has sometimesbeen defined more narrowly, in termsof the intention of the sender-i.e., whetheror not the signal that is sent was meant tobenefit the sender. Apart from it being impossibleto prove experimentally the teleologicalassumption that we can actuallyknow the goal of the sender, in this definitiononly intraspecific information flowis considered to be communication Anemission from one individual, interceptedby third parties, for example, from anotherspecies nearby, and causing a behavioralresponse in the eavesdroppers would not beconsidered communication because theemission was not "meant" to be received bythat third party.I do not believe this is a particularlyhelpful way to view communication, becauseit precludes the possibility that humanscan communicate with insects andvice versa. Clearly, we do communicatewith insects, and they do respond so reliablyto us With Wilson's framework, insects'emissions do alter entomologists'probabilities of behavior, as do ours theirsThis is one of the things that makes entomologysuch an exciting discipline, andsuch a promising area for insect behaviorresearch to have a positive effect on societyI was fortunate to have begun my researchcareer working with WendellThomas C. Bake?Roelofs and Ring Card6 Whenever a pheromoneresearch project stalled out orreached an apparent impasse for one reason01 another because the insects were notresponding to our test pheromone blends,Wendell and Ring would insist upon goingback through all the data with the idea thatthe insects were telling us something important,only we were too stupid to find theright way to listen! The lesson learned, oncewe got around the impasse and identifiedthe pheromone, was that learning the languageof insects requires that we do listencarefully, then, in our own halting way,persist in trying to talk back.When communication occurs, there isinformation flow between sender and receiver,through a communication channelin the environment, as depicted by Shannon&Weaver (1949) Apart from the meaning,or semantic value of the information, theamount of information that reaches thereceiver after the signal travels through thechannel depends, in a statistical sense only,on the reduction in an obser ver 's (receiver's)uncertainty about the occurrence of asignal, given a finite set of possible signals.The channel itself can be flooded withvarious amounts of noise (Fig. I), that mayinfluence the remaining uncertainty andhence reduce the amount of received information.Also, even without noise the channelcan, through its quirky characteristics,alter the emitted transmission and imposeupon it a selective attenuation, such that bythe time it reaches the receiver it has beendistorted. Despite these problems, andwithout teleologically trying to infer the"message" in the signal, it is possible tofactor out the key components-i.e., theessence-of the emission by carefullyanalyzing the behavior of the sender, thestructure of the emission, and the behaviorof the receiver. After all, many parasitoidshave learned over evolutionary time to dothis with many of their insect hosts So havepredators using communication signals aggressively,such as bolas spiders and firefliesthat attract and kill prey by signaling in thepheromonal or light-flash "languages" oftheir prey Many phytophagous insects aswell have been selected to locate plantparts, 01 whole plants, by using a narrowrange of possible emissions that indicatethat a favorable host is present. Understandingthe minimal set of cues that influencesbehavior was the essence of theethological school of behavior studies, asexemplified by Niko Tinbergen's work onbaby herringgulls'responses to the red spoton their mothers' beaks or his work ondigger wasps' homing abilities using landmarksInformation flows to us from ourinsect subjects. Our job is to learn what itconsists ofCommunication in the three major modalities-light,sound (pressure disturbances),and chemicals-has many commoncharacteristics There is a main carriel frequency,which is the dominant wavelengthin light, the main tone in a sound emission,and the blend of chemicals in a semiochemicalsemission (Fig. 2 A-C). In addition, researchin all modalities has shown that it isimportant that the emission at a particularcanier frequency be amplitude-modulated,which means that its intensity be variedconsiderably for it to evoke a response in thereceivers (Fig. 2 D-F) In communicationusing light emissions, this entails flashingor varying the intensity in some formin space and time (Fig. 2E). In sound

communication, amplitude modulation involves"chirping" or "trilling" with long orshort periods of silence, rather than emittinga continuous tone (Fig. 2D) Inchemical communication, amplitude modulationusually is provided by the turbulentshearing of the emission as it leaves thesource to create filaments of odor similar towhat we can visualize within smoke plumes(Fig 2F). However, some rhythmic pulsingof chemical emissions can be imparted bythe emitter, such as arctiid or lymantriidmoths emitting pheromone, and these can,in some cases, stand out against the turbulence-inducedmodulation for a half a meteror so downwind (Conner et a1 1980) Suchpulsing thus far has been found to have noextra signal value and also does not increasethe response of males compared to a continuouslyemitted stream of pheromone.No signal from insects has yet beenfound that is frequency modulated and hascommunication value Frequency modulated(FM) signals would involve changingcolor, sweeping through a series of tones, orchanging one's chemical blend However,bats hunting for insect prey are FM emittersthat with each ultrasonic chirp sweepdownward from high to low tones and fromthe particular narrow range of tones of thereflected signal, gather information aboutthe size and wing-beat frequency of theirprey These bats also modulate the amplitudeof their cries by sending discrete chirpswith a silent period that aids in receivingechoes, but there are species that emit AM(amplitude modulated) cries only, with nofrequency sweeps The moths that arehunted have not been shown to beresponsive to the FM modulations of thecries, but rather react defensively by tryingto evade the bats in response to a wide rangeof ultrasonic tones across the frequencyrange of the sweepsBat-moth communication brings up theaspect of reflected versus emitted signals inthe three modalities Bats are providingtheir own sonic lanterns, 01 strobe lights, ifyou will, with which they peer into thedarkness for moving prey items They relyon reflected sound for information to flowto them about prey. Most visually dependentinsect predators and parasitoids relyon ambient light from the sun to reflect offprey items and provide information flow.Likewise, most visual communication systemsuse reflected, ambient light, except inthe notable cases of firefly communication,in which emitted signals are the rule.Emitted, rather than reflected, sounds arethe rule for insects that communicate bymeans of pressure disturbances. However,some species do enhance then resonance byreflecting then emissions off of the sides ofWinter 1 99 3Fig. 1,. Diagram of information flow that results in communication. The signal is depicted asfluctuating in amplitude (wavy purple line), whereas noise in the communication channel is depictedas being unvarying in amplitude (straight lines), but of a different carrier frequency (yellow instead ofpurple). The signal is able to be received and responded to even with the background noise presentdue to the different carrier frequency and its amplitude modulation relative to the noise (afterShannon 6- Weaver 1949)specially constructed burrows, or off ofleaves enlisted to increase the resonance oftheir emissions for the same purpose(Prozesky-Schulze et a1 1975)Semiochemical communication in insectsinvolves both emitted and reflectedcompounds. Obviously "you are what youeat" means that all atoms emitted fromwithin one's body must have at some pointbeen assimilated from an external source,and, therefore, could be considered "reflected."However, some molecules areingested or adsor bed from surroundingsand then reemitted as a signal in more-orlessunaltered form, as shown by TomEisner in his work on the defensive compoundsof many insects (see, for example,Eisner 1970 for review) The results ofmany studies on courtship pheromonesemitted by male moths have shown thatthese pheromones also are comprised ofmainly plant-derived, little-altered molecules.Other semiochemicals are mainlyself-generated, synthesized emissions, asexemplified by the lepidopteran sex pheromonesmanufactured from building-blockacetate molecules, as highlighted by thework of Lou Bjostad and Wendell Roelofs(see Bjostad & Roelofs 1987 for review)Other common properties of communicationin any of these modalities include thefact that sustained reception of intact signalsis affected both by background noise inthe channel and by adaptation of the sensorypathways, whether the system involveslight, sound, 01 chemicals If some ofthe sensory pathways in the receiver havebeen made dysfunctional from adaptationdue to background noise, then the signalthat reaches the brain may have beenseverely altered and rendered ineffective inevoking behaviorAnother major commonality amongcommunication systems in the three mainmodalities is the way in which communicationsystems are dissected by researchersin older to distill the minimal set of signalsimportant in evoking a response in receiversFirst, the natural sequence of behaviorsneeds to be observed and analyzed for theresearcher to get a feeling as to which partsof the sender-r eceiver inter action appear tobe important in evoking a response, or inother words, in successfully communicating,as indicated by the response. Then thenaturally emitted signal needs to be capturedand played back for detailed analysisIn visual communication, relative reflectancesof insect body par ts such as wings canbe gained by obtaining them and placingthem in spectrophotometers to determinethe major hues 01 carrier frequencies thatare preferentially reflected and absorbed bypigments struck by ambient, white sunlightThe emission spectrum of fireflies canbe immediately captured and measured ona spectrophotometer and displayed to findthe major peak emission frequency orwavelength of light Temporal analysis ofthe tempos at which wing reflectances areamplitude modulated 01 firefly flashes arepulsed, as well as the durations of theflashes and pulses can be recor ded on film orvideotape and measured during playback.

Carrier Frequency3 4 6 6 7 8Frequency (kHz)8 .Wavelength (nm)ChainlengthSoundLightAmplitude ModulationDTeleogryllus commodus 8 calling songE/ .. .: ;. : --Photuris versicolor

odorous papers as effectively as they wereto the natural extract We found that thecombination of two particular fractions,the EAG-active one plus another one,increased the attraction compared to theone fraction alone. Analysis of the compoundsin these fractions revealed that theEAG-active one contained one compound,(E)-ethyl cinnamate The other fractioncontained two compounds that wereboth stereoisomers of methyl jasmonate' (Nishida et a1 1982)The verification that one of these compounds,(2)-methyl epijasmonate, was thesecond pheromone component and that@)-ethyl cinnamate was also a pheromonecomponent came from obtaining syntheticsamples of the compounds and presentingthem either singly or as a blend to femaleswho responded by being attracted to theblend Synthetic samples of other compoundspresent in the emission were alsoobtained and presented to the females, yeteven though some were much more abundantthan the two pheromone components,they did nothing to influence thebehavior, and were judged not to be involvedin communication. Therefore, thekey to proving that communication hasoccurred lies in the manipulation of the signalusing synthesized components, alongwith requiring that a significant behavioralresponse be observed in response to theaddition of any of the components Toprove that communication occurs in anymodality, we must prove that we can talkback to the insectsInterestingly, the @)-methyl epijasmonatewas present in such small quantitiesthat we could not get enough for GC-MS analysis However, for months we feltthat the odor from this fraction smelled sofamiliar that if we could locate the othersource of it we might get enough to be ableto identify it This component had a slightlyherbal, slightly floral, slightly fruity odor,yet we could not remember exactly wherewe had smelled it before! Then one day oneof us, Ritsuo Nishida, sniffed a whole uncutlemon before he prepared it to squeeze itinto his tea, and immediately realized thatthe oriental fruit moth hair-pencil coponentwas in there Eventually, after analysis ofthe air borne volatiles from lemons, the (Z)-methyl epijasmonate proved to be a heretoforeunidentified odorant from lemonand, nanogram-for-nanogram, was themost powerful odorant yet identified fromthis fruit. The odor is familiar to us as thenondescript "fruit-bin'' odor that we arehardly conscious of as we walk past piles oflemons, limes, and other citrus fruits in thegrocery store The fragrance industry that isinvolved with lemon-scented cleaning and1 1 1 2 1 3 1 4 1 5 1 6 1 71 8 1 9 10 11 1Fraction numberl2 ,HFig 3. Gas chromatographzc ~ecordzng (packed column) of male G molesta hazrpenctl extract,showing the relatzue abundance of unknown components and the fractzons taken by researchers fortestzng zn electroantennographzc and behauzoral bzoassays Four compounds that were zdentzfzedfrom the extract zncluded two (compounds 1 and 4 (E)-ethyl cznnamate and (1 R,2S)-(+)-(Z)-methylepzlasmonate, respectzuely), that when added together evoked szgnzficant levels of attractzon offemale G molesta females zn behauzoral assays Compounds 2 and 3 were behauzorally znactzue, eventhough 2 was over 10 tzmes as abundant as the other compounds, and even though 3 dzffers from 4only by uzytue of being a steyeozsome~ To the human nose, 3 zs uzrtually odorless, whereas 4 has apleasant, herbally fruzty aroma lzke that of an uncut lemonpolishing agents became interested in manufacturingthis epime~ to add to theirproducts for communicating to humans thesensation of "lemon-freshness."Another interesting sidelight camewhen Ritsuo Nishida read that methyl jasmonatewas known in the perfume industryas the "queen of aroma" and was one of theoldest and most widely used compounds byperfumers, extracted initially from jasmineplants Further ~esearch by Terry Acree andRitsuo Nishida revealed, however, that ofthe four possible epimers, only lR, 2s-(+)-(Z)-methyl epijasmonate, the compoundinvolved in the courtship pheromone blendand contributing to attraction, was stronglyfragrant to the human nose (Acree et a11985) The other three stereoisomers comprising97% of the "queen of aroma,"proved to be virtually odorless! The beautifulfragrance attributed for years to themethyl jasmonate actually came from animpurity present at about 3% in the methyljasmonate, and this impurity is the orientalfruit moth courtship phe~omone component!Who would have thought that mothand human courtship fragrances would beso closely intertwined, and both could involvea signal gathered initially from plantchemicals? We found subsequently thatmales lacking the hairpencil compound@)-ethyl cinnamate can imbibe it fromsugar water and incorporate it into theirhairpencils for use as a pheromone, anexample of a reflected chemical signal(Nishidaet a1 198 5,Lofstedtet a1 1989).The responses of insects receiving soundemissions must be assessed in the samefashion to understand which elements aremost critical to successful communicationFirst, a discriminating way to measure theresponse must be developed, and an exampleof this is with Teleogryllus oceanicusfemales made to walk on a styrofoampretzel and choose whether to go right orleft at choice-points during their walk. Responsesto natural emissions replayed fromaudio tape were compared in this way andshowed that the females could discriminatefrom among hybrid and even hybrid siblingmale songs, based on slight differences inthe patterns of amplitude modulation(Bentley & Hoy 1974) One useful featureof this assay is that the responder is notallowed to take itself into an area where thesound pressure level from one of the twoloudspeakers becomes greater than theother, thereby biasing the choice withoverall amplitudeThe rule in acoustic communicationthus far, whether it be for crickets, grasshoppers,drosophilid fruit flies, and eveninsects such as leafhoppers and lacewingsthat transmit their pressure disturbancesthrough a substrate such as a plant, has beenthat signals among species differ in patternsWinter 1993215

of amplitude modulation. For example?leafhoppers, Graminella nigrifrons sendvibrations through plants and then wait tohear an answer. Males then begin movingup the plant, and by a type of trial and errorwith regard to walking up one branch oranother, they arrive at the answering femalewho is stationed near the top (Hunt et al.1992). In their work. Hunt et a1 used a Kaysonoiraph from the'audiology and speechtherapy department at University of Kentucky.To learn the language of an insect,entomologists used an instrument affectingour own ability to communicate. Theydetermined that only a small fraction ofamplitude-modulated vibrational pressuredisturbances is actually important in attractingmatesaIn a similar way, the parts of fireflyemissions that are important to communi,-cation were discerned by using synthesizedsignals, in this case generated from flashlights.In the classic work of James Lloyd(19661, he determined that the carrier frequency,or color? of the light emitted byflying males was not that critical for evokinga response from females resting in thefoliage and vice versa (Fig,, 2B), However,the way in which males flashed their lightthatis, modulated its amplitude-includingthe pulse duration and the interpulseinterval? was in fact critical" The intervalbetween the flash of the male and the (usually)single response pulse from the femalewas also crucial to evoking - an orientationresponse from males, who continued toapproach females that responded by flashingat the appropriate interval (Fig. 2E)Although chemical signals are rarely inany obvious way amplitude modulated bythe actions of the emitter, fluctuations in theintensity of pheromone concentration, infact, do occur and have been shown to becrucial for sustained communication. Theimportant fluctuations in amplitude forchemical signals come from the mechanicalshearing of odor as it is released from asource such as a female gland? creatingsmall-scale turbulence. little eddies, thatdownwind result in pockets of clean airinterspersed with dense strands, or filaments,of pheromone. We have found withthe oriental fruit moth that the intermittentcontact with these filaments is essential tosustaining upwind flight of males.How do we know this? The sex heromoneof the oriental fruit moth consists ofthree com-ponents 6% E8-12:Ac + 3% Z8-12:OH in Z8-12:Ac, as discovered byanalyzing the emissions and responses asdescribed earlier for this species' courtshippheromone (Fig 2C) (Roelofs et a[. 1969,Card6 et al. 1979). The carrier frequencyi,,e,,the blend-was then varied syntheti-cally, along with the overall emission amplitude-ie", the concentration-and a responseprofile was formed from both fieldtrapping tests and wind tunnel studies As inthe case of sound and light communicationsystems? the carrier frequency that turnsout to be optimal for moths is centeredaround the naturally emitted female blend.Also, as is typical for sex pheromones, thediscrimination for carrier frequency is exquisitelyand narrowly tuned in males?which differs markedly from sound systemsin which there is ver y broad tuning and littlediscrimination simultaneously for both theemission quality (carrier frequency) andamplitudeIn the wind tunnel, EAG results showedthat a male oriental fruit moth antenna doesreceive intermittent stimulation from thefilaments of pheromone downwind Evenin the field? we showed that 3, lo? or even 30meters downwind the antenna receives significantfluctuations of pheromone concentrationand successfully registers them? asin the wind tunnel (Fig" 2F) (Baker &Haynes 1989) Then we showed that malesdo need fluctuating stimulation to sustaintheir upwind flight? by using a wind tunnelin which we could create a uniform cloud orfog of pheromone We also pulsed the cloudat either one per second or one per twoseconds, with a similar period of clean airbetween the pulses. The males flew upwindonly when the cloud was pulsed When presentedwith the uniform, constant fog ofpheromone, however, males exhibitedwide crosswind "casting" flight whilekeeping station momentarily after takingoff It appeared as if they adapted quickly tothe uniform cloud and behaved as theywould had they just lost pheromone andflown into clean air (Baker et a1 1985).Further analysis of the behavior of orientalfruit moth males in response to quickexposures to and loss of pheromoneshowed that they respond within 117 s toeither odor-on or odor-off They reversedtheir course across the wind-line-i e.? zigzagged-anaverage of seven times per second,and with a single exposure to clean airfollowing loss of pheromone they flewmore across the wind on the very next reversalto begin casting flight. Their responseto contact with pheromone on theother hand, was to fly more directly upwindon the very next reversal after contactingpheromone and reverse more frequently, tocreate an upwind surge Considering thatthe EAG results showed that males willencounter only about two or three filamentsper second if they fly straight upwindin the plume, the 117s reaction time of malesto either pheromone on or off meant thatthere is therefore time between filamentsfor the males' behavior to change to crosswindcasting flight? even while "in" thetime-averaged plume. There is also time forthe male to react by surging upwind to eachfilament. We believe that what we see in azigzagging flight track therefore is a kind ofhybrid response that averages out to bezigzagging-i.e., tracks that are not fullycrosswind casting or fully upwind surg-ing-because the moth is usually in transitionfrom one to the other due to thetiming of the intermittency of the simulation(Baker 1990),Why should such a quick and highlyphasic response evolve? We think it mayhave to do with the challenge faced by maleoriental fruit moths, as well as by males ofsome other insect species that must flyupwind to odor sources in quickly shiftingwind fields (Baker 1990). They mustrespond to every filament caused by smallscaleturbulence by surging upwind, becausethe combination of odor plus winddirection points, in many environmentalsituations? directly toward the source. Conversely,males must begin to stop their upwindsurge and begin casting quickly tokeep their station in clean air in order toavoid plunging very deeply into a largepocket of clean air caused by a large-scalewind shift; they quickly become farther andfarther displaced from the new position ofthe odor if they proceed upwind withoutpheromone, Casting will also increase theirchances of recontacting odor that hasmoved off to one side or the other due tosuch a shift in wind direction caused bylarge-scale turbulence. Thus, it behoovesmales to respond quickly to every pocket ofclean air? even small-scale ones? by castingas soon as it is detected, because it may turnout to be one of the large pockets inevitablylooming ahead (Baker 1990). If, on theother hand? the pocket is only one of themany little ones, the male is still in positionto hit the next filament. Thus, although amale moth flying upwind in response topheromone may appear to be in contactwith the plume, it really is never in contactwith pheromone for any period of time atall. he male is either encountering filamentsfrequently due to small-scale turbulence?in which case the upwind surges arerapidly, reiteratively elicited, or else infrequently?in which wide, infrequent reversalsare made crosswind, which allow it torecontact the odorWith regard to pheromone receptorsystems, the consequence of not meetingthe challenge of recovering quicklyenough after exposure to intermittent pheromonestimulation is receptor adaptation.Adaptation is the reduction in a receptorcell's firing response to a stimulus with

@

the right one is being emitted. For instancein the turnip moth? Ag~otisegetum, whileworking in conjunction with the pheromonegroup at the University of Lund,Sweden, we showed that a differentialadaptation of antennal receptors can occurin response to excessive emission rates ofthe correct pheromone blend, and can explainthe arrestment of upwind flightprogress in response to excessively concentratedsynthetic lures that occurs in manymoth species (Baker et al. 1988). We placedsingle cell preparations in pheromoneplumes 70 cm downwind of the source andexposed the cells to the same point sourcesthat had earlier caused arrestment of upwindflight before the source was reached;we also exposed the cells to lower concentrationsthat had promoted complete flightto the source without arrestment. The highestemission rate of the correct femaleemittedblend had caused males to startflying upwind, but within a few seconds,their upwind progress stopped; they becamearrested while in flight and stayedflying at one distance downwind.Recordings from single neurons on maleantennae pointed out why this might havehappened. First, Van der Pers & Lofstedt(1986) had shown that there are three classesof receptors on Agrotzs segetum antennae,each tuned to one of the three pheromonecomponents, a short chain-length, amedium chain-length? and long chainlengthcompound. Therefore, when a filamentof the three-component pheromoneblend strikes the antenna? cells of the threetypes residing in different hairs send a ratioof firing to cells higher up in the antennallobe. All three cell-types need to be firing inorder to register as "pheromone7' in thehigher centers. At the lower emission ratespromoting sustained upwind flight, thecells responsive to all three components dofire consistently and repeatedly in burstsover a sustained period of time. However, atthe highest emission rate? the one causingarrestment, we found that the receptor cellstuned to the major component becomeadapted quickly and stop firing; the cellsresponding to the other two components,on the other hand, keep on firing properlywithout adaptation (Hansson & Baker1991) What this means is that the ratio offiring from the three cell types would nowbe skewed in favor of the two cell typesresponding to only the minor components,and the odor blend would now appear to beoverloaded with those two minor componentswhen in fact the blend in the air wasstill correct. Excessive concentration thereforewould cause a perceived change in thequality of the odor as a result of the differentialadaptation of one receptor type. As inthe example of the differential adaptationof one of our visual receptor types, higherordercells in the brain of the male mothwould judge the ratio of inputs to be somethingother than what was present in thereal worldThe idea that by exposing insects toexcessive amounts of odor we can alter theirresponse to their own pheromone by changingthe quality of the received signalthrough adaptation or habituation? allowsus to develop new strategies for insectcontrol As entomologists, especially oneshired by the USDA or the Agricultural ExperimentStations in our respective states,we have an obligation to work to generatenew knowledge that can be applied toimprove the quality of life for the citizens ofour states Taking our knowledge of noiseand how it can affect receptor pathways? wecan try to add noise artificially to a communicationchannel to disrupt communication.In this way? once we have learned thepheromonal language of a particular species,we can talk back, loudly, with veryspecific noise to impede possible informationflow in that species.Noise plays an important role in theevolution of communication systems. Wecan see the importance of natural noise incommunication channels by the ways inwhich palatable prey insects have beenselected to blend into it. In sound communication,clearly the evasive behavior ofmoths responding to bat cries includes turning-tailon the bat and reducing theirsignal to blend, as a "stealth" moth? into thebackground of the open air of the night. Butalso the behavior includes diving into thebushes and blending into the many reflectivesurfaces of the vegetation that create amaze of noise to the strobe-light sounds ofthe bats as they search and close in ontheir prospective prey. Likewise defensivelyblending into the background noise in thelight often involves painting oneself in pigmentsthat match the ambient reflectivefrequency of the vegetation? in the middlewavelengths that appear green to us and toprospective predators. In addition, mothsthat rest on bark in daylight not only havethe appropriate background noise pigments,but also, as Sargent (1968) showed?have been selected behaviorally to choosethe appropriate background shade? regardlessof what they see painted on their bodies.Noise has had a great selective effect therefore?on the behavior of these moths in notcommunicating with predators such asforaging birds.In pheromone systems there is tantalizingevidence that some pheromone blendshave evolved to stand up to different kindsof pheromonal background noise in differ-ent geographic locations that might otherwiseprevent communication^ Linn et al,(1984) found that in the cabbage loopermoth T~ichoplusia ni, there are componentsemitted as part of the blend that areredundant? and can be eliminated withoutsignificantly harming the subsequent responseThey found several important redundantpairs of components that are inthis way mutually replaceable in the signalto evoke the same optimal response Thisredundancy can be likened to our ownprinted language that contains extra? redundantletters, Some can easily be droppedout without affecting the reception of-themessage, and in a noisy channel that containsthese exact same letters, the messagecan still be picked out without error. If theredundancy did not exist. there would notbe a buffer against noise? and now furthernoise in the system would fatally erode theremaining letters' reception and make themessage significantly more difficult to receiveIn T nz, not just any two pairs of compoundsare redundant, but rather only specificpairs can fill-in for each other if one ofthe ;air is missing, If both members of aredundant pair are missing? then the sustainedupwind flight to the source is leduced.We had hypothesized that one waythat one member of the pair might be causedto be effectively absent would be a result ofnoise in the environment from other speciesof calling moths that would cause theadaptation of that particular componentspecificpathway (Baker 1989). OUI ideawas that species such as T. ni that are foundall over the countr y may have a robust set ofcompounds because their rich blend canstand up to the different kinds of noise indifferent geographical regions emanatingfrom females from different species-complexes,Such species could be successful incommunicating anywhere in the country,because the redundancy would thereforeserve as a buffer against environmentalnoise. An example of one redundant pair is12:Ac and Z9-14:Ac, which are two minorpheromone components in T ni; another isZ5-12:Ac and 11-12:Ac.Using the cut-sensillum single cell techniqueto record from antennal neurons?Julie Todd in my laboratory found that thehypothesized redundant pathways are re-presented neuronally . by . cells tuned to theseparticular redundant pairs of compounds(Todd et al. 1992)- That is7 she found antennalneurons that responded to either 12:Ac or Z9-14:Ac, but not the other pheromonecomponents. Also, she found cellsthat responded to either ZS-12:Ac or 11-12:Ac7 but not the other compounds. Wethen hypothesized that cells responding to

these redundant pairs mght send theiraxons to the same discrete centers in themacroglomer ular complex in the antennallobe of the brain There would be fourmajor centers in the antennal lobe7 onereceiving inputs from cells responsive to themajor component only7 and three other centersreceiving Inputs from particular redundantpairs of receptor cells that reflect thebehavioral redundancy found earlier byLinn and Roe10fs~In collaboration with Bill Hansson at the4University of Lund7 Sweden7 we have nowbegun to see if such redundant-pair-specificcenters exist7 by recording from the cells' and finding out which compounds they aretuned to, then dying the cells with cobalt,(and s~lver-intensifying them to follow themdown to their areas of arborization inthe macrog~omerular complex (MGC)(Hansson et a1 1992) So far7 we have foundthat cells specific for the major componentdo in fact send their axons only to onemajorsubcompartment of the macroglomerularcomplex It will be exciting to see if theother subcompartments that are found inthe MGC are there for the purpose of receivlnginputs from par ticular redundantpairs of minor components.Natural noise does appear to have hadan effect on the ability of insects to communicate>and now by artificially addingselected types of noise we know we can affectplocesses such as mating and oviposition?even feeding The use of pheromonesfor disrupting mating has become commerciallyacceptable? even desirable Forinstance, by placing controlled-release dispensersof three-component oriental fruitmoth pheromone into the environment> thesuccess rate of males locating calling femalescan be reduced to then California> peach growershave accepted this control measure> and apattern that has occurred is that after threesuccessive years of using pheromones, themoth populations plummet to the pointwhere often no control intervention isnecessar ye Included among other species forwhich commercially available disruptantshave been successful are the tomato pinwoImmoth7 the pink bollworm moth7 andgrape berr y moth (Rldgway et a1 1990).Does the pheromone disruption-emittednoise work by confusing males? causingthem to fly upwind7 locate the inappropriatesynthetic sources, and waste timeinvestigating them instead of mating? Ordoes it work by adapting (habituating) themales' pheromone-specific pathways tohampe~ their sensory capabilities? There isevidence that it is a little of both, since pinkbollworm males have been observed underfield conditions visiting disruptant hollowWinter 1993fibers However> evidence for the adaptation-habituationmechanism as a contributingfactor comes from the experimentsof Hollis Flint and coworkers in Arizona(Flint & Merkle 1984) They showed thatwhen noise in the form of the correct SO: SOpheromone blend of the two pink bollwormpheromone components was broadcastinto fields7 traps emitting this sameratio were completely shut down to maleattraction, as were traps emitting a differentseries of blend ratios of the same twocomponents When they used a 9:l ratio ofthese components with which to disruptcommunication7 however, they found thatdisruption again was profound across allsignaling frequencies, including the naturalblend However, this time traps emitting afrequency artificially enriched with thecomponent that predominated in the disruptionblend did manage to capture a smallbut significant number of males., This blendwould never have attracted males in a noisefreeenvironment7 because it is such an extremeoff-ratio. This odd result showed thatmales could be made to fly upwind to a badblend as if it were an optimal one byartificially compensating for the adaptationEnriching the blend with the componentto which the adaptation hadselectively occurred brought back an illusionof the correct balance to the odor"Finally> the steps involved in learning thelanguage of insects must include communicationwith other scientists We musteffectively use our language to pass on ourknowledge about the insects to our colleaguesand to transmit our insights aboutpossible evolutionar y patterns or about theprospects for applying our knowledge forpest management purposes. The insects do7indeed, oftentimes seem to be trying to tellus something, Once we learn what it is7 wemust not be lazy and let their message die byfailing to publish or talk about our findings.The important feature of communicationbetween insects and ourselves is that wemust understand that we are the channel,,The old axiom "publish or perish" isfamiliar to everyone,, But actually7 the importantaspect of this statement for thoseinvolved in insect communication researchis publish, or else the language of insectswill perish,, Their words will not merelyhave fallen on deaf ears7 but even moretragically, they will have ended up in a mutehuman channel.References CitedAcree, T. E , R. Nishida & H. Fukami. 1985.Odor thresholds of the stereoisomers ofmethyl jasmonate J Agric Food Chem 33:425427Baker, T,, C. 1989,, Origin of courtship and sexpheromones of the oriental fruit moth and adiscussion of the role of phytochemicals in theevolution of lepidopteran male scents pp,401418, In C,, H, Chou & G,, R, Waller[eds,], Phytochemical ecology: allelochemicals,mcotoxins, and insect pheromones andallomones, Academia Sinica, Taipei, R, O,,C,1990,, Upwind flight and casting flighc complementaryphasic and tonic systems used forlocation of sex pheromone sources by malemoths, pp 18-25, In K, D~ving, led,,], IS01X, Proceedings 10th International Symposiumon Olfaction and Taste, GCSIAS Oslo,Baker, T. C. & R" T" Cadi. 1979, Courtshipbehavior of the Oriental fruit moth (Grapholithamolesta): experimental analysis andconsideration of the role of sexual selection inthe evolution of courtship pheromones in theLepidoptera, Ann, Entomol, Soc, Am, 72:173-188,Baker, T. C. & K. F. Haynes. 1989. Field andlaboratory electroantennographic measurementsof pheromone plume structure correlatedwith oxiental fruit moth behaviour,,Physiol Entomol, 12: 263-2 79,,Baker, T. C", R., T., Cadi & ,J,. R. Miller:. 1980.Oriental fruit moth pheromone componentemission rates measured after collection byglass-surface adsorption, J, Chem, Ecol,. 6:749-758,Baker, T. C.> M,, A. W~llis, K. F. Haynes & I? L,,Phelan, 1985. A pulsed cloud of pheromoneelicits upwind flight in male moths, Physiol,Entomol 10: 257-265,Baker, T. C.,, B, S. Hansson, C. Lofstedt & J.Lofqvist. 1988" Adaptation of antennalneurons in moths is associated with cessationof pheromone-mediated upwind flight, Proc,Natl, Acad,, Sci, U,S,.A,. 85: 9826-9830,,Bentley, D. & R. R,. Hoy. 1974. The neurobiologyof cricket song, Sci Am 231: 33-44,Bjostad, L. B. & W, L. Roelofs. 1987., Sex pheromonebiosynthesis in lepidopterans: desaturationand chain shortening, pp,. 77-120,, In G,,D, Prestwich & G, L, Bloomquist [eds,], Pheromonebiochemistry, Academic, New York.,Card6> A,, M., T. C. Baker & Ro I.Cadi. 1979.Identification of a four-component sex pheromoneof the female oriental fruit moth,Grapholita molesta (Lepidoptera: Tortrici-dae),, J,, Chem, Ecol,, 5: 42 3427,Conner7 W. E.,, T" Eisner, R. K. Vander Meer, A.Guer~ero, D. Ghuingelli & ,J. Meinwald..1980. Sex attractant of an arctiid moth(Utethesia ornatrix): a pulsed chemical signal,.Behav, E~ol,, Sociobiol,, 'E 55-63,Eisner, 1. 1970. Chemical defense againstpredation in arthropods, pp,, 1,57-217,, In E,,Sondheimer & J,, B, Simeone [eds,]? Chemicalecology Academic, New York,,Flint, H. M" & ,J. R. Merkle. 1984. The pinkbollworm (Lepidoptera: Gelechiidae): alterationof male response to gossyplure by releaseof its component Z,Z-isomer., J,, Econ,, Entomol,77: 1099-1104,,Hansson, B. S,, & T. C. Baker. 1991. Differentialadaptation rates in a male moth's sex pheromonereceptor neulons,, Naturwiss '78: 517-520,

Hansson, B. S,,, H. Ljungberg, E. Hallberg & C.Lofstedt. 1992. Functional specialization ofolfactory glomeruli in a moth Science 2 56:1313-1315Hunt R. E., D. P. Fox & K. F. Haynes. 1992.Behavioral response of Graminella nigrifrons(Homoptera: Cicadellidae) to experimentallymanipulated vibrational signals, J. InsectBehav, 5: 1-1 3,Lall, A. B,,, H. H. Seliger, W, H. Biggley & J. E.Lloyd,, 1980. Ecology of colors of fireflybioluminescence Science 210: 560462,Leroy, Y, 1964. Transmission du parametre frequencedans Ie signal acoustique des hybridesF, et P x F, de deux Grillons: Teleogrylluscommodus Walker et T oceanicus Le Guillou(Orthopteres, Ensiferes), C.R, Acad,, Sci.(Paris) 258: 892-895,Linn, C. E., L. B. Bjostad, J. W. Du & W. L.Roelofs. 1984. Redundancy in a chemicalsignal: behavioural responses of male Trichoplusiani to a 6-component sex pheromoneblend J, Chem Ecol11: 1635-1658Lloyd, J. E. 1966. Studies on the flash communicationsystem in Photinus fireflies, Misc, Pub.,Mus, Zool, Univ Mich,, 130: 1-95,Lofstedt, C., N. J. Vickers, W. L. Roelofs & T. C.Baker. 1989. Diet related courtship success inthe O~iental fruit moth, Grapholita molesta(To~tricidae), Oikos 55: 402408,Nishida, R., I.C. Baker & W. L. Roelofs. 1982."Hairpencil" pheromone components ofmale Oriental fruit moths, Grapholithamolesta J Chem. kol 8: 947-959.Nishida, R., T. C. Baker, W L Roelofs & T EAcree. 1985. Orientalfiuitmoth pheromone:Attraction of females by an herbal essence,pp 47-63 In 1 E Acree&D M Soderland[eds 1, Semiochemistry: flavors and pheromonesWalter de Giuyter, BerlinProzesky-Schulze, L., 0 P.M. Rozesky, FAndeison & G.J.J. van der Meiwe. 1975. Useof a self-made sound baffle by a tree cricket.Nature (Lond.) 255: 142-143,Ridgway, R. L., R. M. Silverstein & M. N.Inscoe [eds.]. 1990. Behavioi-modifyingchemicals for insect management. Dekker,New Yo1 k,Roelofs, W. L., A. Comeau & R. Selle. 1969. Sexpheromone of the oriental fruit moth. Nature(Lond.,) 224: 723,Sargent, T. D. 1968. Cryptic moths: effects onbackground selections of painting the circumocularscales. Science 159: 100-101.,Shannon, C, E. & W. Weaver. 1949. Themathematical theory of communication,,University of Illinois Press, UI bana.Todd, J. L., K. E Haynes & T. C. Baker. 1992.Antenna1 neurones specific for redundantpheromone components in normal andmutant Trichoplusia ni males, PhysiolEntomol, 17: 183-192,,Van der Peis, J.N.C. & C. Lofstedt. 1986. Signalresponserelationship in sex pheromonecommunication, pp. 235-241 In T, L, Payne,M, C, Birch, & C..E..J,, Kennedy [eds.,],Mechanisms in insect olfaction. Claiendon,Oxford.Wilson, E. 0. 1974. The insect societies 3rdedition Belknap, Cambridge, MA 0Thomas C Baker is a professor andchair of the Department of Entomologyat Iowa State University, Ames 50011-3222. His maloy research interest is inthe neuroethology of pheromone-meditedflight in moths, as well as developingpheromones and other semiochemicalsfor use in agricultural and urban environmentsCurrent projects include NSF- andUSDA-funded work on Trichoplusia niolfaction and hehothine moth orientation.This article was adapted from theplenary lectu~e presented at the 1991 ESAAnnul Meeting in Reno, NV