Acoustics, context and function of vibrational signalling in a lycaenid butterfly–ant mutualism

Please download to get full document.

View again

of 14
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Categories
Published
Acoustics, context and function of vibrational signalling in a lycaenid butterfly–ant mutualism
   ANIMAL BEHAVIOUR, 2000, 60, 13–26doi: 10.1006/anbe.1999.1364, available online at http://www.idealibrary.com on ARTICLES Acoustics, context and function of vibrational signalling in alycaenid butterfly–ant mutualism MARK A. TRAVASSOS & NAOMI E. PIERCE Museum of Comparative Zoology, Harvard University (Received 8 April 1999; initial acceptance 29 July 1999;final acceptance 17 November 1999; MS. number: A8282)  Juveniles of the Australian common imperial blue butterfly, Jalmenus evagoras, produce substrate-bornevibrational signals in the form of two kinds of pupal calls and three larval calls. Pupae stridulate in thepresence of conspecific larvae, when attended by an ant guard, and as a reaction against perturbation.Using pupal pairs in which one member was experimentally muted, pupal calls were shown to beimportant in ant attraction and the maintenance of an ant guard. A pupa may use calls to regulate levelsof its attendant ants and to signal its potential value in these mutualistic interactions. Thereforesubstrate-borne vibrations play a significant role in the communication between J. evagoras and itsattendant ants and pupal calls appear to be more than just signals acting as a predator deterrent.Similarly, caterpillars make more sound when attended by Iridomyrmex anceps , suggesting that larval callsmay be important in mediating ant symbioses. One larval call has the same mean dominant frequency,pulse rate, bandwidth and pulse length as the primary signal of a pupa, suggesting a similarity infunction.  2000 The Association for the Study of Animal Behaviour  From the foot-drumming of a banner-tailed kangaroo rat,  Dipodomys spectabilis , in the presence of a snake(Randall& Matocq 1997) to the coordinated group chorusing of nymphal treehoppers (Cocroft 1996), vibrational signal- ling is a widespread form of communication, functioningprimarily in defence, mate attraction and displays of aggression. Instances of such communication betweenunrelated species, however, are relatively rare. Neverthe-less, recent work indicates that vibrational communi-cation may play a vital role in butterfly–ant mutualisms,wherein caterpillars and pupae use an intricate combi-nation of chemical, behavioural and secretory signals tomaintain a retinue of ants that protect them from pred-ators and parasites.DeVries (1990)showed that cater-pillars that interact with ants are capable of producingvibratory ‘songs’. These larvae are found exclusively inthe butterfly families Lycaenidae and Riodinidae. In com-paring ant attendance levels between control larvae of theriodinid Thisbe irenea and larvae that had been experi-mentally muted, calling T. irenea caterpillars were tendedby more ants, indicating that one function of riodinidcalls is ant attraction. DeVries concluded: ‘under selectionfor symbiotic associations, the calls of one insect specieshave evolved to attract other, distantly related insectspecies’(DeVries 1990, page 1106). Both larvae and pupae of some species of Lycaenidaecan produce sound(Dodd 1916;Downey 1966;DeVries 1990). Larvae produce vibrations that are primarily sub-strate borne, although they may have a slight airbornecomponent(DeVries 1991a;M. Travassos, personal observation), whereas pupae produce signals with bothvibrational and airborne components (Hoegh-Guldberg1972;Downey & Allyn 1978). In pupae, a file of teeth on the anterior side of the sixth abdominal segment gratesagainst an opposing plate on the posterior side of the fifthabdominal segment (Downey 1966). Such a plate may bemade up of either tubercles, reticulations, or ridges. Themechanism of sound production in lycaenid larvae, how-ever, has proved more elusive.Hill (1993)proposed apossible stridulatory organ similar to that found in pupae:a file of teeth and an opposing plate. Caterpillars in thefamily Riodinidae, the sister taxon to the Lycaenidae(Kristensen 1976;D. Campbell, A. Brower, N. Pierce, unpublished data), signal by beating vibratory papillaeagainst epicranial granulations; lycaenids lack such astructure (e.g.Cottrell 1984). Correspondence: N. E. Pierce, Museum of Comparative Zoology, Harvard University, 26 Oxford Street, Cambridge, MA 02138,U.S.A. 0003–3472/00/070013+14 $35.00/0  2000 The Association for the Study of Animal Behaviour  13  At least half of all lycaenids interact with ants, varyingfrom facultative interactions where juveniles are foundwith or without ants, and often with many species,to obligate ones where juveniles cannot live withoutants and usually associate with only one or two closelyrelated species(Pierce 1987;Fiedler 1991). Not only are myrmecophilous lycaenids protected against ants them-selves, which might otherwise be threatening predators(Malicky 1970), but it has been shown experimentally that attendant ants also protect juveniles from predatorsand parasites (e.g.Pierce et al. 1987;Fiedler & Maschwitz 1988a;DeVries 1991c;Cushman et al. 1994;Wagner 1994).Dodd (1916)andDeVries (1990)suggested that lycaenid caterpillar calls, which are essentially substrate-borne vibrations, may be important in mediating antinteractions, because ants, although nearly deaf to air-borne sounds, are sensitive to vibration(Fielde & Parker1904;Ho ¨ lldobler & Wilson 1990). Lycaenid larvae areable to produce several different vibrational signals.DeVries (1991a)noted that some lycaenid sounds havetwo components: a low background sound accompaniedby a louder pulsing. Leptotes cassius , for example, has ‘ aticking background and an irregular, galloping series of trills ’ (DeVries 1991a,page 17). No experimental work has examined the function of lycaenid larval calls.Research on the function of pupal calls has been incon-clusive. Functional explanations for lycaenid pupal sig-nals have speculated about their role in defence(Hinton1948;Downey & Allyn 1978), conspecific attraction in the formation of aggregations (Prell 1913), andmyrmecophily(Downey 1966;Elfferich 1988;Brakefield et al. 1992). The acoustic characteristics of such signalsare better known. Lycaenid pupae produce three distinctsignals, each distinguishable by amplitude level (Downey& Allyn 1978). The loudest pulse, the primary signal, isalso the longest. Secondary signals are briefer and areoften found in pulse trains. Tertiary signals have onlybeen recorded in large pupae and consist of irregular clicktrains not much louder than background noise. AlthoughDowney (1966)andHoegh-Guldberg (1972)reported a correlation between primary signal production andmovement between the fifth and sixth abdominal seg-ments, no observations of abdominal movement havebeen made regarding the production of secondary andtertiary signals.We analysed the acoustic properties, context and func-tion of sound production in juveniles of the Australianlycaenid Jalmenus evagoras . Larval and pupal J. evagoras associate with ants of several species in the genus Irid-omyrmex . In return for producing nutritious secretions,caterpillars and pupae receive protection against pred-ators and parasites (Pierce et al. 1987). Jalmenus evagoras larvae have several mechanisms to attract and appeaseattendant ants (reviewed inKitching 1983), including adorsal nectary organ (DNO) on the seventh abdominalsegment that produces a sugary secretion that the antsimbibe, and surface epidermal glands called perforatedcupola organs (PCOs) scattered along the length of thecuticle that are thought to produce substances importantin both appeasement and reward. PCOs are also found onpupae, which lack a functional DNO. On the eighthabdominal segment, larvae have a pair of eversible ten-tacle organs (TOs), believed to release a volatile chemicalthat alerts attendant ants if a larva is alarmed or the DNOis depleted(Henning 1983;Fiedler & Maschwitz 1988b). Late-instar larvae and pupae of  J. evagoras stridulatewhen disturbed(Pierce et al. 1987). Like other lycaenids (Downey 1966), pupae have a file-and-plate stridulatoryorgan between the fifth and sixth abdominal segment.The pupal plate extends the length of the intersegmentalregion and has a series of ridges against which the teethgrate.DeVries (1991a, page 17) found that pupae produce7.5 ‘ metallic click-like pulses ’ per second, each with afrequency of 2300 Hz, and that larvae of  J. evagoras produce a drumming call resembling a ‘ khen-khen-khen-khen ’ at a rate of 7 calls/s and a mean frequency of 1700 Hz. Our study aimed to investigate the parametersof these calls in greater detail and to explore their func-tional significance. We focus on the vibrational compo-nents of these sounds, measured with an accelerometer.Throughout, except where noted otherwise, we use theterms ‘ sounds ’ and ‘ calls ’ for simplicity in describingthese substrate-borne vibrations. Nothing is known of thesensitivity of lycaenid larvae and pupae to airborne versussubstrate-borne signals; ants are nearly deaf to airbornesound, but sensitive to vibrations(Fielde & Parker 1904; Ho ¨ lldobler & Wilson 1990). GENERAL METHODS The study was conducted in the Museum of ComparativeZoology Laboratories from June 1996 to February 1997.We collected J. evagoras eggs from field sites in Ebor,New South Wales (30  24  S, 152  19  E), Mount Nebo,Queensland (27  24  S, 152  47  E), and Canberra, AustralianCapital Territory (35  21  S, 148  56  E), Australia. Queen-right colonies of  Iridomyrmex anceps maintained in thelaboratory were collected from Mount Nebo, Canberra,and Griffith, Australia (27  33  S, 153  3  E). Ant colonieswerefedanartificialdiet(Bhatkar&Whitcomb1970)andchopped crickets daily. Larvae of  J. evagoras were rearedon Acacia melanoxylon and A. irrorata raised from seedpurchased from the Queensland Forestry Department.All calls were recorded in an experimental arenameasuring 132  66 cm and 132 cm high with two verti-cal surfaces covered in black construction paper to reducethe effects of external stimuli such as sunlight. To allowfor manipulation of the set-up, the other two verticalsurfaces remained uncovered. The room was maintainedat a constant temperature of 22  C and had fluorescentlighting overhead.Because of their low amplitude, lycaenid calls are diffi-cult to analyse. AlthoughHoegh-Guldberg (1972)con-cluded that there was no resonance from using recordingvials to amplify airborne signals,Downey & Allyn (1978)showed that this method, and the use of a directionalmicrophone, introduces artefacts such as standing wavesand frequencies.DeVries (1991b)used a particle velocitymicrophone and amplifier attached to a paper or mylarmembrane that acted as a recording stage. He placedlarvae or pupae on the membrane and recorded theirsubstrate-borne vibrations. 14 ANIMAL BEHAVIOUR, 60, 1  We used a different approach. We recorded the vi-brations using two accelerometers (BU-3170 andBU-1771, Knowles Electronics Inc., Itasca, Illinois). Theseaccelerometers have sensitivity ranges of 20 – 3000 and50 – 3000 Hz, respectively.Downey & Allyn (1978)deter-mined that pupal calls fall between 400 Hz and 5000 HzandDeVries (1991a)found that lycaenid caterpillar callshave a mean frequency of about 1 kHz. Initial testsindicated that pupal calls of  J. evagoras fell within thelower half of this range. Each accelerometer weighed0.28 g, and measured 7.92  5.59  4.14 mm. Recordedvibrations were amplified with an Archer Mini-Amplifierand recorded on a Nagra IV-SJ tape recorder. We ampli-fied the calls further on the tape recorder by 40 dB. Onechannel recorded vibrational signals from the accelerom-eter, the other spoken behavioural observations. For theexperiments we placed larvae on Acacia branches thatwere 1 – 5 mm in diameter. An accelerometer was firmlytaped to the branch so that it was in close contact withthe plant surface and oriented so that its axis of acceler-ation was normal to the plant surface. For experimentswith pupae, an accelerometer was firmly taped to awooden stick on which a caterpillar had pupated, its axisof acceleration normal to the stick ’ s surface. CALL ANALYSIS Methods We examined the context of larval and pupal calls (seebelow) and then used the samples of vibrational signalsobtained from these experiments to analyse the acousticrepertoire of  J. evagoras . We examined these samples withCanary 1.2b 1994, a sound analysis program produced bythe Cornell Laboratory of Ornithology. We defined thebeginning and end of a call with respect to the back-ground noise level. To control for differences in recordingquality and degree of filtering, we used a uniform bright-ness and contrast setting for all spectrograms. However,the signal-to-noise ratio may not have been completelycomparable for all recordings used. For each call, wemeasured four properties. We calculated the dominantfrequency as the average of the upper and lower fre-quency bounds of a call. The bandwidth of a call con-sisted of the difference between these upper and lowerbounds. We measured the pulse length as the duration of a call, and we calculated the pulse rate, measured only forcalls in pulse trains, as the inverse of the time until thenext pulse was produced. All four characteristics weremeasured for each call sample. For each subject, weaveraged the parameter values for each call. The callcharacteristics of 11 pupae were derived from a total of 108 calls, while the call characteristics of nine larvae werecalculated from 125 calls.To determine the relative amplitudes of a pair of calls,we used recordings of a subject producing both call typesin the same trial. Whenever possible, we made 10 peak-to-peak voltage measurements for each call type and thenaveraged these for each subject. Amplitude differencesbetween a subject ’ s different calls were measured in deci-bels. The relative amplitudes of the calls of 10 pupae werederived from a total of 144 signals, while the relativeamplitudes of the calls of 13 larvae were calculated from294 signals.Counts given with each call refer to the numberof subjects sampled for each particular call type. Weperformed pooled comparisons between parameters of different calls with Mann – Whitney U  tests. Comparisonsbetween the peak-to-peak voltages of pairs of differentcalls were made with Wilcoxon signed-ranks tests. Indetermining significance, ties were taken into account.We calculated relationships between pairs of variables asPearson ’ s correlations. Results Pupae of  J. evagoras produced two types of substrate-borne vibrations (Table 1), which matchedDowney & Allyn ’ s (1978)description of primary and secondary sig-nals; no tertiary signals were detected. Primary signals( N  =11) had a higher amplitude than secondary signals( N  =5; Wilcoxon test: T  =55, N  =10, P< 0.01), which weretypically found interspersed between primary signals(Fig. 1a) or in pulse trains. Both kinds of signals were alsoproduced as single pulses. They did not differ signifi-cantly in mean dominant frequency (Mann – Whitney U  test: U  =22.00, N  1 =11, N  2 =5, NS) or bandwidth ( U  =13.00, N  1 =11, N  2 =5, NS).Larvae produced three kinds of calls. The call with thehighest amplitude sounded like a grunt ( N  =9) and hadthe longest pulse length and the highest mean dominantfrequency of the three larval calls (Fig. 1b). Jalmenusevagoras caterpillars also produced a lower-amplitude,lower-frequency call ( N  =6) resembling the sound of a cat Table 1. Call characteristics for each juvenile call (mean ± SE)CallFrequency(Hz)Bandwidth(Hz) Pulses/sPulse length(s)Relative amplitude(dB)PupaePrimary signal 849.2 ± 31.0 1435.4 ± 62.8 1.76 ± 0.23 0.082 ± 0.008 StandardSecondary signal 772.6 ± 90.7 1098.3 ± 170.8 9.24 ± 2.54 0.033 ± 0.004 − 5.9 ± 1.2LarvaeGrunt 754.4 ± 34.6 1361.4 ± 75.0 2.01 ± 0.36 0.106 ± 0.018 StandardDrum 471.7 ± 79.5 831.7 ± 129.4 8.29 ± 0.33 0.040 ± 0.005 − 2.7 ± 0.9Hiss 444.1 ± 39.2 778.2 ± 31.6 6.39 ± 0.47 0.050 ± 0.007 − 9.8 ± 1.5 15TRAVASSOS & PIERCE: VIBRATIONAL SIGNALLING  Figure 1. Spectrogram (top) and waveform (bottom) of (a) the primary and secondary signals produced by J. evagoras  pupae, and a train of (b) grunts, (c) drums and (d) hisses produced by J. evagoras  larvae. Drawings by Christopher Adams. 16 ANIMAL BEHAVIOUR, 60, 1  purring or a low-pitched drumming(Fig. 1c). The lowest- amplitude call ( N  =3) sounded like a rapid ‘ hiss-hiss-hiss ’ and, like the drum call, was found only in pulse trains(Fig. 1d). There was no significant difference betweenthe mean dominant frequency (Mann – Whitney U  test: U  =7.00, N  1 =6, N  2 =3, NS), pulse length ( U  =4.50, N  1 =6, N  2 =3, NS), or bandwidth ( U  =9.00, N  1 =6, N  2 =3, NS) of the drum and hiss calls. The drum call, however, had ahigher pulse rate ( U  =0.00, N  1 =6, N  2 =2, P< 0.05) and was5.9 dB louder than the hiss call ( N  =8), a significantdifference (Wilcoxon test: T  =28, N  =7, P< 0.05). Both of these calls differed significantly from the grunt withrespect to all four parameters.The larval grunt call and the pupal primary signal wereremarkably similar. There was no significant differencebetween them in mean dominant frequency (Mann – Whitney U  test: U  =25, N  1 =9, N  2 =11, NS), pulse rate( U  =25, N  1 =7, N  2 =9, NS), pulse length ( U  =15, N  1 =7, N  2 =9, NS), or bandwidth ( U  =45, N  1 =9, N  2 =11, NS).However, the pupal primary signal was several timeslouder than the larval grunt (M. Travassos, personalobservation). CONTEXT OF LARVAL SOUND PRODUCTION Methods We placed larvae on individual Acacia that had beencleared of ants and juvenile J. evagoras . Because calls canonly be detected within a few centimeters of a caterpillar,we applied a band of Tanglefoot (The TanglefootCompany, Grand Rapids, Michigan), a molasses-likesubstance, to the base of the host plant to limit thecaterpillar ’ s movements. We taped an accelerometer to aplant branch near the caterpillar.After an acclimatization period of at least 30 min, werecorded calls in a 5-min control period through onechannel of the tape recorder and spoken observations of the caterpillar ’ s behaviour, coded as either stationary,walking (no feeding-related behaviour) or foraging(which included feeding), on the second channel. Inaddition, we also noted larval TO eversions in most timeperiods.We used a wooden dowel to connect the host plant toan I. anceps colony. We again recorded the caterpillar ’ scalls and behaviour in the 5 min following a worker ant ’ sfirst contact with the caterpillar. Thirty minutes after anant ’ s discovery of the larva, we made a second 5-minrecording.We tested 11 larvae. Each caterpillar acted as its owncontrol in comparisons of call production under differentconditions. Results The presence of ants influenced the rate of larval soundproduction for two of the three types of calls (Fig. 2).Larvae produced significantly more grunts when firstdiscovered by ants (Wilcoxon test: T  =55, N  =10, P< 0.01)and after 30 min of ant attendance ( T  =55, N  =10, P< 0.01)than in the control trial. Grunt call production did notdiffer between the two ant attendance intervals ( T  =30.50, N  =9, NS). A caterpillar produced the hiss call when  I. anceps workers first discovered it. Hiss call productionwas only detected once during the 5-min control period,but increased significantly once ants contacted the larva( T  =34.50, N  =10, P< 0.05). However, after 30 min of antcontact, the hiss call was not produced. In contrast to thegrunt and hiss call, the drum call was not producedsignificantly more when a caterpillar was attended byants (  22 =4.77, N  =9, NS). However, in the first 5 min of ant attendance, there was a positive correlation betweenthe amount of time spent foraging and the number of drum calls produced (Fig. 3). The more time a larva spent foraging when first discovered by ants, the more likely itwas to produce a high number of drum calls. Correspond-ingly, during this time interval, there was a negativecorrelation between the amount of time a caterpillar wasstationary and the number of drum calls produced( r  6 =  0.729, P< 0.05). However, no such associationexisted between walking and drum call production( r  6 =0.224, NS). In the second ant interval, there was nocorrelation between time spent foraging and drum callsproduced ( r  6 =  0.089, NS). There were no significantcorrelations between activity and the production of eitherhiss or grunt calls for any time interval. 100Control:no AntsTreatment    N  o .  o   f   1   0 -  s   i  n   t  e  r  v  a   l  s   i  n  w   h   i  c   h  a  c  a   l   l  w  a  s  p  r  o   d  u  c  e   d   i  n  e  a  c   h   5 -  m   i  n   t  r  e  a   t  m  e  n   t Ant attendance:30 – 35 minAnt attendance:1 – 5 min8642*1000    G  r  u  n   t  s  p  r  o   d  u  c  e   d   i  n  e  a  c   h   5 -  m   i  n   i  n   t  e  r  v  a   l 80604020*DrumHiss* Figure 2. Jalmenus evagoras  larval call production in the presenceand absence of ants (* P< 0.05, Wilcoxon signed-ranks test). 17TRAVASSOS & PIERCE: VIBRATIONAL SIGNALLING
Similar documents
View more...
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks
SAVE OUR EARTH

We need your sign to support Project to invent "SMART AND CONTROLLABLE REFLECTIVE BALLOONS" to cover the Sun and Save Our Earth.

More details...

Sign Now!

We are very appreciated for your Prompt Action!

x