BEHAVIOR Stridulation by haematoloma (: ): Production Mechanism and Associated Behaviors

1 2,3 1 2 ARIEL F. ZYCH, R. W. MANKIN, JAMES F. GILLOOLY, AND EVERETT FOREMAN

Ann. Entomol. Soc. Am. 105(1): 118Ð127 (2012); DOI: http://dx.doi.org/10.1603/AN11048 ABSTRACT The Hemiptera displays a notable diversity of vibratory communication signals across its various families. Here we describe the substrate and airborne vibrations (sounds), the mechanism of production, and associated behaviors of Herrich-Schaeffer, a member of the family Rhopalidae. Adult males and females both produce short, stereotyped sound bursts by anteriorÐ posterior movement of abdominal tergites I and II against a stridulitrum located on the ventral surface of the metathoracic wing. Sound bursts are produced by a single adult male or female when physically touched by another adult, and are strongly associated with being crawled on by the approaching individual, but are not produced in response to contact with other or when pinched with forceps. The propensity to produce sounds when crawled upon decreases during the mating season. These sound bursts by J. haematoloma likely are communication signals. Rhopalidae has been signiÞcantly absent from the vibratory communication literature until now. Although the sounds are produced using a mechanism common to vibratory communication systems in closely related Het- eropteran Hemiptera, the sounds in these other species function primarily in courtship or in motherÐ daughter interactions, which suggests that the functions of stridulation and the behavioral contexts have diversiÞed in the .

KEY WORDS communication, mating, stridulation, Heteroptera

Acoustic and vibratory communication is incredibly Rhopalidae (Hemiptera: Heteroptera), Jadera haema- diverse and has arisen multiple times during the evo- toloma Herrich-Schaefer, and describe for the Þrst lution of Hemiptera. The evidence in support of mul- time the mechanism and behavior associated with tiple evolutionary origins of vibratory communication vibrations in this family. in this group is quite extensive because of the char- Jadera haematoloma is an abundant rhopalid acteristic morphology of diverse vibration-producing throughout the continental United States that has structures in several families (Ashlock and Lattin 1963, been well-studied as an example of rapid evolution Schaefer 1980, Schaefer and Pupedis 1981, Polhemus associated with host plant shift (Carroll and Boyd 1994, Tishechkin 2006). Hemiptera occupy a wide 1992). Jadera sp. play a role in reducing the seed array of niches and communicate in many different productivity of a species of (Koelreuteria ecological and behavioral arenas (Cocroft and Rodri- elegans Laxmann) classiÞed as a Class II Invasive by guez 2005). Comparisons of the structures used for the Florida Exotic Pest Plant Council (Carroll et al. vibratory communication among distantly related 2003, FLEPPC 2009). Rhopalids also are pests in coun- Hemipteran taxa may improve understanding of the tries where Sapindales such as Lychee and Longan are behavioral and ecological pressures that drive acoustic cultivated (Waite and Hwang 2002). convergence or diversiÞcation in . However, The opportunity to investigate rhopalid-produced elaborating Hemiptera as a clade for studying acoustic vibrations came from a chance discovery that groups evolution requires a good record of the diversity of of J. haematoloma placed into a bucket for transfer signals and signal-producing structures, a record that from a Þeld site to a laboratory produced audible is still incomplete. Here we investigate the vibrations sounds. A search of the literature revealed no infor- produced by a well-studied member of the family mation about sound production by this species, and only minimal information about substrate or airborne The use of trade, Þrm, or corporation names in this publication does communication in Rhopalidae. In spite of their eco- not constitute an ofÞcial endorsement or approval by the United logical and economic importance, a single recording of States Department of Agriculture, Agricultural Research Service of any product or service to the exclusion of others that may be suitable. a male hyoscyami L. is the only documenta- The USDA is an equal opportunity provider and employer. tion of sounds by species in this group (formerly Cori- 1 Department of Biology, University of Florida, Gainesville, FL zus hyoscyami L., Gogala 1990). Consequently, a study 32611-8525. was initiated to: 1) record and analyze the sounds in 2 USDAÐARS Center for Medical, Agricultural, and Veterinary En- tomology, Gainesville, FL 32608. a behavioral context, 2) describe the mechanism of 3 Corresponding author, e-mail: [email protected]. sound production, and 3) explore whether sounds January 2012 ZYCH ET AL.: VIBRATORY COMMUNICATION IN J. haematoloma 119 produced may be used for communication among aggregations by using a bag net beaten against members of the same species or as aposematic threats branches containing seasonal aggregations (Zych to potential predators. 2010). IdentiÞcation was veriÞed with the help of The body parts used by other closely-related members of the Florida State Collection of Arthropods Hemiptera to produce vibrations provide clues to the in Gainesville, FL (Slater and Baranowski 1978, Schuh mechanism that may be used by J. haematoloma. Rho- and Slater 1995). Groups of up to 100 adults and palidae lies within the infra-order , nymphs were housed in 2-liter plastic buckets with a diverse group that includes , Pentatomi- screen lids and given fresh host leaves and water from dae, and other families with species known soaked wicks. Host seeds were omitted from enclo- to communicate through vibrations (Henry 1997). In sures to accurately reßect seed availability in the Þeld the Pentatomomorpha, a tergal plate formed by the at the time of collection (no intact seeds were present fusion of abdominal tergites I and II commonly is at the collection site). The buckets were held in a associated with vibratory communication that may be growth chamber maintained at 26ЊC with a photope- used in Rhopalidae (reviewed by Virant-Doberlet and riod of 14:10 (L:D) h to simulate summer temperature Cˇ okl 2004, Gogala 2006). Two mechanisms of the ter- and photoperiod. Individuals were housed for up to 2 gal plate have been proposed as the vibrational mech- wk before being returned to the Þeld and replaced by anism. The Þrst is as a plectrum used in conjunction new Þeld collections. Acoustic vibratory, and video with a wing stridulitrum (Leston et al. 1954, Ashlock recordings of groups of individuals were conducted at and Lattin 1963, Schuh and Slater 1995). The second 26ЊC and 60% RH in a vibration-shielded anechoic is as a “tymbal”; a bi-stable plate that pops in and out chamber (Mankin et al. 1996) at the Center for of two stable conÞgurations, similar to that used by Medical, Agricultural, and Veterinary Entomology, and (Gogala et al. 1974, Gogala Gainesville, FL. 2006). The tymbal mechanism was Þrst hypothesized Sound Recording and Analysis. To fully character- for , where Gogala 2006 and colleagues ize the vibrations produced by J. haematoloma, it was (showed that wax application between tergites I and important to record and describe both substrate and II silenced low-frequency signals. Subsequent repli- airborne (acoustic) vibrations, as well as identify vari- cations of this technique have turned up conßicting ation among individuals. To capture both acoustic and results [Lawson and Chu 1971, Numata et al. 1989]). vibratory recordings, nine tests were conducted with Other vibrations in Pentatomomorpha have been at- 10Ð12 individuals placed in a cylindrical cage (6 cm tributed to a tymbal mechanism without direct obser- diameter by 10 cm height) made of 1-mm metal screen vation or manipulation of the tergum (Schaefer 1980, mesh for 30 min. An accelerometer (Bru¨el and Kjaer Virant-Doberlet and Cˇ okl 2004). Using J. haema- [B&K], Naerum Denmark) for recording substrate- toloma, it is possible to address not only the persis- borne vibrations (Wenninger et al. 2009) was clipped to tence of tergal plate involvement in sound production the top edge of the screen lid. These accelerometer in Pentatomomorpha, but also determine whether it recordings were coupled with acoustic recordings by functions as a tymbal or a plectrum. using a B&K microphone (Mankin et al. 2000) horizon- Before the sounds and sound-associated morphol- tally positioned atop a foam block 1 cm from the cage. ogy described in this report could be contributed to Both acoustic and substrate-borne vibrations were re- the body of literature on the diversity of insect signals, corded at the same time to determine whether they were it was important to distinguish between random or produced synchronously and to determine how each incidental sounds and those that may serve a role in differed from the other. All recordings were digitized communication. To disprove that sounds are pro- and saved on a computer using a commercially available duced randomly, we Þlmed interactions and then speech analysis system (Wenninger et al. 2009). asked whether sounds were associated with speciÞc To determine whether sex or body length inßu- behaviors and participants. We then identiÞed intra- enced variation in the sounds produced, a separate set and inter-speciÞc interactions that may play a role in of recordings were made of adults of known sex and communication, such as a defensive response to pred- body length. Body length was measured as the tip of ator threats (Masters 1979) or as an attractive signal to the clypeus to the posterior edge of the last abdominal conspeciÞcs (Wenninger et al. 2009). This behavioral segment. Nineteen adults whose sex and body length context will play an important role in our future un- were determined (Carroll and Loye 1987) were indi- derstanding of how vibrations are used for communi- vidually placed in the anechoic chamber setup with an cation in the Rhopalidae and other related families. individual who was silenced by waxing (see Experi- mental Manipulation of Stridulatory Apparatus be- low). We recorded three bouts of sound production Materials and Methods for each individual to associate sound characteristics J. Haematoloma Collection. Adult and immature J. with body length measurements. haematoloma were Þeld collected in July and August We analyzed the microphone and accelerometer 2009 beneath large (Ͼ6-m crown height) specimens signals by using Raven Pro version 1.3 (Cornell Lab of of the local host, Golden Raintree, Koelreuteria panicu- Ornithology 2008). The Raven software provided dis- lata v. bipinnata Laxm, in northwest Gainesville (Ala- plays of oscillograms and spectrograms and measured chua County) and on the University of Florida cam- the temporal patterns and amplitudes of individual pus. Adults were captured from large canopy sound impulses and groups (bursts) of impulses 120 ANNALS OF THE ENTOMOLOGICAL SOCIETY OF AMERICA Vol. 105, no. 1

(Mankin et al. 2008), the dominant frequencies of conducted in November 2010, after mating had begun, bursts, and the maximum frequency ranges of bursts. with the following modiÞcations. To capture encoun- Sound-associated Behaviors. To identify behaviors ters that could result in mating, no acclimation period that may be associated with sound production, we was possible, as individuals typically paired within Þlmed Þeld-caught individuals while recording several minutes after being placed in the containers sounds, and asked whether or not sounds occurred together. Because of time constraints, only seven more frequently when certain behaviors occurred. groups were recorded and analyzed. Because individ- Preliminary observations showed that sounds were uals moved considerably less upon being paired, two not produced outside of encounters between adults. trials resulted in fewer than ten encounters. Behavioral analysis was therefore restricted to en- Statistical Analyses of Behavioral Encounters. We counters between individuals. Groups of four to ten pooled encounters across interaction trials to total 100 adults were placed in ßat rectangular plastic “inter- encounters in the nonmating season in 2009, and 63 action” containers (5 by 8 by 2 cm) with one of the two encounters in the mating season in 2010. Trials were widest surfaces of the container replaced with a Þne not pooled between seasons. This pooling was neces- mesh screen that faced a B&K microphone as de- sary to achieve adequate numbers of targeted behav- scribed above. Interaction containers were held ver- iors to associate with sound production. Encounters tically with one clear surface facing a SONY HD were categorized by the presence or absence of target Handycam video camera, illuminated from 0.5 m behavior, the approached sex, and the approaching above using a 120 W incandescent light bulb. Filming sex. During the mating-season encounters were fur- took place through the nonmesh face of the container ther categorized by whether or not the encounter at normal speed by using the HD macro setting. Fresh resulted in mating between the two participants. A Þeld-collected adult individuals were used for each two-way contingency table with sound (ϩ)orno recording session, and were allowed to acclimate for sound (Ϫ) as the response variable was created for 30 min before the start of recording. Sound and video each of these encounter variables. A Fisher ␹2 (chi- recording continued for 60 min. squared) test then was used to determine if there was To standardize among group interaction trials, we a signiÞcant difference in the number of encounters restricted this analysis to the Þrst ten encounters after featuring a speciÞc behavior or participant sex com- the 30-min acclimation period. Encounters were char- bination that resulted in signaling. Two-tailed ␹2 val- acterized as the period starting when two individuals ues are reported because they are a more conservative at least one body-length apart initiate physical contact representation of the difference between two groups with any appendage, and ending the moment they than a one-tailed test (Fisher 1922). Data analysis was move to greater than one body length apart. Encoun- completed using JMP statistical software (SAS Insti- ters were then characterized according to the relative tute 2008). speed of individuals entering into the encounter, the Sound Production Mechanism. Using three ap- sex of participants, the presence or absence of speciÞc proaches, we conÞrmed that J. haematoloma pro- behaviors, and whether or not sounds were produced duced sound by use of a plectrum and stridulitrum, the (by a signaler). The relative speed of each individual most commonly found sound production mechanism moving into and away from a physical contact was in terrestrial Heteroptera. We Þrst used a series of measured in terms of body length(s), and assigned to high-speed video recordings to identify structures that each participant according to “approaching” and “ap- move speciÞcally in association with sound produc- proached” as the individual moving with the greatest tion and that could function as a plectrum, the basic or least velocity at the start of an encounter (respec- mechanism most common in terrestrial Heteroptera. tively). We categorized two encounter-speciÞc be- After identifying a putative plectrum, we then cap- haviors in addition to recording the sex and speed of tured high-resolution surface images of the articulat- encounter participants. We deÞned “body rocking” as ing side of the hind wings to determine whether there taking place when an individual rolled or rocked its was a stridulitrum that articulated with the putative entire body from side to side around the anteriorÐ plectrum. Lastly, we performed a series of loss-of- posterior axis. When the approaching individual function manipulations to verify that we had correctly crawled completely on top of the approached indi- identiÞed the structures involved in sound produc- vidual, we referred to this as “crawling onto” which tion. These consisted of removing suspected struc- was chosen over “mounting” to emphasize that no tures, or preventing their movement, and then record- effort was made to associate this behavior with cop- ing any sounds produced by those individuals in ulation. The relative speed of each individual moving conspeciÞc interactions. into and away from a physical contact was measured Plectrum Identification. To identify a putative plec- in terms of body lengths per second, and assigned to trum, we searched for body part movements that cor- each participant according to “approaching” and “ap- responded to the rate and behaviors during brief pe- proached” as the individual moving with the greatest riods of sound production. Individuals with wings or least velocity at the start of an encounter. The Þrst removed were Þlmed using a high-speed video camera ten encounters for each group were then pooled to- (250 frames/s) Þtted with a macro lens under infra-red gether totaling 100 scored encounters. illumination (Autumn et al. 2006). Individuals were Ten groups of individuals were Þlmed in September placed in Þve groups of four into small (10 by 8 by 4 2009 before the mating season. Additional tests were cm) clear plastic enclosures, and recorded for 15-s January 2012 ZYCH ET AL.: VIBRATORY COMMUNICATION IN J. haematoloma 121 intervals encompassing acoustic activity. The speciÞc the remaining 10 males between abdominal tergites I behavioral context in which sounds were observed and II after Gogala (1990). This prevented anterior- from behavioral trials made it possible to attribute posterior movement of the abdominal tergites. Care recorded sounds to speciÞc individuals being Þlmed. was taken not to wax the wings to the thorax so that Because a single individual per interaction produces any role of wing movement in sound production in- signals, Þlm analysis was restricted to encounters in dependent of plectrum movement still could be ob- which the individual producing sounds was isolated served. Individuals were checked after waxing to en- within a frame to ensure that body part movements sure that they still had complete use of all their legs, were correctly attributed to the stridulating individ- wings, and antennae and did not show any noticeable ual. Synchronous high-speed video and sound record- alterations of their behavior. After wax application, ing was not available for this analysis, but video anal- the 10 males were placed in the interaction container ysis was restricted to those 15-s periods of video in together and recorded to determine if muting was which sounds were being continuously produced. complete. When repetitive movements were discovered, the cy- Acoustic Response to Threat Stimuli. To determine cle-rates were calculated using Photron Fastcam whether sounds may be used as defensive signals in Viewer (Photron Limited 2006). This calculation was response to threat, we recorded sounds produced dur- used in place of synchronous high-speed sound and ing Þlmed interactions with , cockroaches, and video recording to determine whether the putative simulated events. We used two types of plectrum moved at a rate sufÞcient to generate the threat stimuli: threats and simulated verte- repetition rates observed during recording sessions. brate predator threats. Two adult individuals were Stridulitrum. Because initial inspection of the high- placed in a large petri-dish Þtted with a screen lid and speed videos suggested that the moving part, or plec- suspended in the same manner as the normal inter- trum, associated with sound production was the plate action trials and were allowed to acclimate for 10 min formed by the fusion of abdominal tergites I and II, we while being Þlmed and recorded. Threat stimuli con- considered whether the undersurface of the metatho- sisted of either four large (Ͼ5 mm in length) carpenter racic wing possessed a stridulitrum that could be used ants (Camponotus sp.) or two American cockroaches in conjunction with the moving plectrum to produce (Periplaneta americana L.) of Ϸ10-mm body length. sound. We used Scanning Electron Microscopy After the acclimation period, the threat stimuli were (SEM) to capture a surface image of the underside of introduced to the container, and interactions and the metathoracic wing. Images were obtained using sounds were recorded for 30 min. This procedure was a variable-pressure scanning electron microscope repeated twice with new adults for each arthropod (Zeiss Evo MA10, Peabody, MA) pressurized at 70 Pa. “threat.” The left metathoracic wing was removed from freeze- Because these insects are highly chemically pro- killed specimens by using forceps and then mounted tected (Aldrich et al. 1990a), it is possible that other on carbon tape dorsal side down. The remaining wings invertebrates do not pose a serious threat. Therefore, were removed using forceps, and the whole body simulated vertebrate predator threats were conducted placed venter-side down for images of the thorax and using isolated adult males in the sound chamber while abdominal tergites I and II. One male and one female recording with a microphone. Adult males were used were included in the SEM-image set that were typical because the interaction trials suggested they were the of the distribution of body lengths observed in the most likely to produce sounds in conspeciÞc encoun- population. The images were similar, so only images of ters. Six males were individually “attacked” by tugging males are displayed in the Þgures. SEM images of the on the legs and antennae with forceps on a stage in underside of the metathoracic wing and abdominal front of a recording microphone. A separate set of six tergites were completed at the Thermal Ionization adult males were individually picked up between the Mass Spectrometry and Scanning Electron Micro- thumb and index Þnger and rolled between the Þngers scope Laboratory in the University of Florida Depart- in front of the microphone four times for Ϸ30 s each. ment of Geological Sciences. Experimental Manipulation of the Stridulatory Ap- Results paratus. To verify that the structures identiÞed using high-speed video analysis and SEM imaging are es- Sounds. Vibrations were detectable both in the sub- sential for sound production, we silenced individuals strate and as audible sound; each type of vibration was by removing the stridulitrum and preventing move- produced synchronously. These vibrations consisted ment of the plectrum. Twenty males were collected of repetitions of a single short, stereotyped burst (cor- that had previously been observed producing signals. responding to a phrase in Eliopoulos 2006) without Ten were chilled for 10 min in an airtight container at any frequency modulation. Burst rate was highly vari- Ϫ10ЊC, and removed fore and metathoracic wings by able among bouts (up to 40 per s), however, the pulling at the wing base slowly. After wing removal, structure of each discrete burst remained Þxed for an individuals were allowed to rest and come to room individual. The frequency range of each acoustic burst temperature for 1 hr before being placed together in was contained within 0.5 and 12 kHz (e.g., Fig. 1). The the interaction containers and recorded. Because Þrst dominant frequency band spanned 1Ð5.5 kHz. The wing removal did not result in complete silencing, second higher frequency band spanned 7.5Ð10 kHz. The melted parafÞn wax was applied beneath the wings of frequency ranges of sounds and associated substrate- 122 ANNALS OF THE ENTOMOLOGICAL SOCIETY OF AMERICA Vol. 105, no. 1

Fig. 1. Spectrogram comparisons of bursts produced by a male J. haematoloma, recorded simultaneously by a microphone (A, airborne sound) and an accelerometer (B, substrate vibration). The two-chirp couplet structure of each burst is less apparent in the substrate vibration. Darker shading indicates greater relative energy. borne vibrations were within the detection range of related with body length (linear, N ϭ 19, F ϭ 17.54, insect vibration sensing organs found in other species of r2 ϭ 0.5078, P ϭ 0.0006, Fig. 3). Though chirp interval plant-dwelling Hemiptera (Cˇ okl 1983, Shaw 1994). was Þxed for an individual (SE Ϯ 0.2 ms within a Within a burst, there was a clear paired couplet selected individual), burst rate varied widely from 1 to structure in the airborne component of the signal that 40 bursts/s. In bouts of sound production showing high consisted of two short, distinct chirps (see terminol- burst rate, the two-chirp couplet was always produced ogy in Eliopoulos 2006). The Þrst chirp consistently faster (less time between bursts), rather than decreas- showed a higher amplitude (power) than the second ing the pulse interval within bursts. (Fig. 1A). This two-chirp couplet, somewhat less ap- Sound-associated Behaviors. All sounds recorded parent in the substrate-borne signal than in the air- took place in a very speciÞc behavioral context: they borne signal (Fig. 1B), was comprised of a series of 4Ð7 were exclusive to encounters between two adults of impulses per chirp, as in the example of Fig. 2. the same species. Sounds were only produced by a The mean values of three sets of 20 bursts per single interaction participant at the start of a physical measured individual (N ϭ 19) were compared for the encounter between two adults. Sounds were pro- following analyses. Chirp interval was positively cor- duced exclusively in this encounter context and were not made outside of periods of physical contact. The individual that produced sounds, referred to here as the approached, was always the individual

Fig. 3. Relationship between chirp interval, the time between the end of the Þrst and the beginning of the second Fig. 2. Oscillogram of a single burst showing impulse chirp of a burst, and body length of female (triangles) and structure in a microphone recording. Impulses 1Ð5, Þrst male (circles) adult J. haematoloma. Solid line indicates the chirp; impulses 6Ð11, second chirp in the burst. linear regression. January 2012 ZYCH ET AL.: VIBRATORY COMMUNICATION IN J. haematoloma 123

Table 1. Effect of specific combinations of sex and behavior of approaching male and female adults on the occurrence or absence of a signal by approached male and female adults during encounters before beginning of mating season and after the beginning of mating season

Pre- Post- Encounter combinations N ␹2 P N ␹2 P Grouped by sex SϮ by 4 sex when M or F 3 100 6.731 0.0159 63 1.239 0.2657 SϮ by 4 sex when M 3 59 0.299 0.7381 42 1.952 0.1623 SϮ by 4 sex when F 3 41 0.370 1.000 21 2.130 0.1444 3 sex when SϮ byMorF4 100 16.246 Ͻ0.0001 63 16.247 Ͻ0.0001 3 sex when SϮ by M 4 52 6.405 0.0202 27 7.895 0.005 3 sex when SϮ by F 4 48 4.396 0.0489 36 11.197 0.0008 Grouped by behavior SϮ when 3 crawls on 4 100 22.537 Ͻ0.0001 63 10.197 SϮ when 4 rocks body 100 2.912 0.243 63 11.599 0.007

P values from FisherÕs two-tailed test performed using a ␹2 contingency table of the presence or absence of signals during N observations of each encounter type (df ϭ 1), sorted by sex of participants or by presence or absence of listed behaviors. 3, approaching; 4, approached; M, male; F, female; ϩ, occurrence; Ϫ, absence; S, signal; Pre-, before beginning of mating season; Post-, after the beginning of mating season. moving with the least relative velocity (body crawled-upon individuals signaled more frequently lengths/s) at the start of the encounter (N ϭ 100). The than in encounters where they were not crawled upon individual moving with the most velocity, the ap- (N ϭ 63, df ϭ 1, ␹2 ϭ 10.197, P ϭ 0.0014), sounds proacher, was always silent; consequently, only ap- occurred in only 73.2% (31/40) of those encounters, proached individuals were signalers. These results not 100% as observed in the nonmating season. Fur- suggest that signals are a direct response to approaches thermore, only 58.3% of mounting encounters that by other individuals and are not used in long or short- resulted in pairing (N ϭ 12) produced sounds. Using range mate attraction. a nominal logistic regression, we compared the re- Nonmating Aggregations (2009 Trials). In trials sponse to being crawled upon during mating and non- conducted with specimens collected from nonfeed- mating seasons and found that the probability of pro- ing, nonmating aggregations in 2009, certain target ducing sounds when crawled upon during the mating behaviors and sex combinations during encounters season was signiÞcantly lower than in the nonmating strongly increased the probability that an encounter season (N ϭ 163, df ϭ 3, ␹2 ϭ 41.023, P Ͻ 0.0001). produced sounds (Table 1). Sounds were produced in In contrast to the encounters in trials using non- 100% of encounters where the approaching individual mating aggregations, the sex of the approached indi- ϭ climbed on top of the signaler (N 19). Approaching vidual was not a signiÞcant predictor of sound pro- males crawled onto other individuals signiÞcantly duction in trials during the mating season (Table 1). more than approaching females, regardless of the sex Whether or not an encounter between a male and a of the signaler (N ϭ 100, df ϭ 1, ␹2 ϭ P ϭ 9.005, 0.0035). female produced sounds during the mating season was Males also crawled onto other males more frequently not related to mating outcome (N ϭ 40, df ϭ 1, ␹2 ϭ than females. Because all encounters where an indi- 0.0049, P ϭ 0.9443). vidual crawled on top of another resulted in sound Sound Production Mechanism. High speed video of production by the crawled-on individual, the male bias in crawling behavior is a reßection of male-bias in individual interactions with wings removed showed the sex of approachers during interactions which pro- rapid (15Ð25 cycles/s) anteriorÐposterior movement duced sounds. Signals also were produced much more of the fused abdominal tergites I and II. During the frequently in encounters where the approached sig- period of rapid tergal movements, all other parts of the naler shook his or her body laterally from side to side, body remained isolated from movement (including a behavior that occurred more frequently when that head, mouthparts, legs, and thorax). This form of ter- individual was crawled upon (N ϭ 100, df ϭ 1, ␹2 ϭ gal movement only was observed in the approached 13.020, P ϭ 0.0062). The strong association between individual for the Þve interactions Þlmed during the being crawled upon and body rocking by the ap- high speed video sessions. The rate of contractions proached individual suggests that the function of (forward-backward) fell within the typical repetition sounds may be to discourage close physical contact. rate observed when bursts are produced (25Hz). The Mating Aggregations (2010 Trials). In the seven anteriorÐposterior movement of the tergal plate was trials observed after mating had begun in natural pop- not accompanied by any deformation, depression, or ulations, pairings took place within the Þrst 15 min of folding of the tergites or corresponding sternites. Con- observation (N ϭ 12). These pairings were initiated sequently, we can be reasonably certain that the tergal with physical contact, rapid mounting by the male, and plate does not function as a tymbal to produce signals, subsequent coupling, similar to the mating sequence as has been suggested by previous authors for this and described by Carroll and Corneli (1995) for Oklahoma other families in the Pentatomomorpha (Gogala et al. and Florida populations of J. haematoloma. Though 1974, 1984; Virant-Doberlet and Cˇ okl 2004). 124 ANNALS OF THE ENTOMOLOGICAL SOCIETY OF AMERICA Vol. 105, no. 1

Fig. 4. Adult male J. haematoloma, SEM image of the ventral side of the left metathoracic wing. Arrow indicates stridulitrum on costal vein; scale bar, 200 ␮m.

SEM images obtained of the undersides of the right with abdominal tergites I and II to produce sounds. and left metathoracic-wings conÞrmed the presence Experimental removal of the wings followed by the of a stridulitrum located on the ventral side of the acoustic analysis revealed that wings are needed to costal vein of the metathoracic wing (Fig. 4). The produce high-frequency components of sounds, but stridulitrum opposes a raised surface on the dorsal not low-frequency components. Metathoracic wing side of abdominal tergite I which, when moved along removal eliminated all higher frequency components the observed anteriorÐposterior Þeld of movement, of signals, and reduced the two-chirp couplet struc- follows the length of the stridulitrum against the grain ture to a single low-frequency chirp (0Ð1000 Hz, Fig. of the stridulitrum teeth (Fig. 5). Stridulitra were 6); however, it did not silence signals completely. found in both male and female adults and were absent Application of wax between the thorax and abdominal on the wing pads of nymphs. tergites I and II successfully silenced individuals by The experimental loss-of-function manipulations preventing anteriorÐposterior movement of abdomi- supported the hypothesis that the wings act in concert nal tergites I and II. These observations suggest that

Fig. 5. Adult male J. haematoloma right abdominal tergites I and II. Dashed double-headed arrow indicates Þeld of motion of abdominal segments along anterior-posterior axis, single headed arrow indicates plectrum surface. Th, thorax; scale bar, 200 ␮m. January 2012 ZYCH ET AL.: VIBRATORY COMMUNICATION IN J. haematoloma 125

Fig. 6. Sample spectrograms of bursts produced by a male with wings (A, expanded view in B), and a male with wings removed (C, expanded view in D). Wing removal abruptly silenced all higher-frequency spectra, and eliminated the two-chirp couplet structure. Numbered labels mark the second chirp of each burst in B and the end of each burst in D. the high-frequency two-chirp couplet is generated by various Cicadidae (Drosopoulos et al. 2006), Mem- the forward (Þrst chirp) and backward (second chirp) bracidae (Rodriguez et al. 2004), and (Percy movement of the tergal plate against a metathoracic wing et al. 2004). A potential beneÞt of the broadband stridulitrum. Without wings, the forward-reverse move- frequency range and minimal temporal modulation is ments of the tergal plate no longer generate high-fre- that the signal can be easily identiÞed and interpreted quency sounds by rubbing against the stridulitrum. The within a variety of different substrates. Typically, J. remaining low-frequency percussive signals after wing haematoloma are found aggregating on many different removal corresponded to the impact of the tergal plate substrates, and there is considerable variation in how on the posterior edge of the metanotum. leaves, stems, branches or other structures transmit Acoustic Response to Threat Stimuli. In the series of and attenuate signals compared with airborne trans- arthropod “threat” encounters, sounds were never mission (Mankin et al. 2008). produced in inter-species encounters, between Jadera The tergal plate and wing-stridulitrum mechanism and cockroaches (N ϭ 23) or ants (N ϭ 17). There was found in J. haematoloma has been observed in several apparent release of chemical defense volatiles by each other members of the Pentatomomorpha; adult which caused the roaches and ants to clean (Leston et al. 1954, Jorigtoo et al. 1998) two lygaeid themselves thoroughly. Pinching legs and antennae genera (Piesma and Kleidocerys, Ashlock and Lattin with forceps did not successfully elicit sounds from 1963); Cydnidae; and the (Gogala et al. any males. Smothering between the thumb and index 1974, Schaefer 1980) but is not similar to the tymbal Þnger encouraged sound production by all individuals mechanism observed in . The stridu- (N ϭ 6) so long as constant pressure was applied. lation shown here in absence of deformation of the Defensive chemicals (toasty almond smell) could be tergal plate favors the plectrum hypothesis over the smelled after the smothering treatment. tymbal hypothesis for the role of the tergal plate in vibration production. The Cydnidae, which also pos- sess a metathoracic wing stridulitrum and abdominal Discussion plectrum, differs from Jadera in sound complexity. We have shown that J. haematoloma produces sub- Jadera produced only a single repeated burst, with strate and airborne vibrations in a unique behavioral fewer discrete harmonics detectable through the sub- context by using similar morphological structures used strate, and did not demonstrate the diversity of con- by other closely related Hemiptera. The sounds pre- text-speciÞc songs observed by Gogala et al. (1974) in sented here show superÞcial similarity to other the Cydnidae. High-resolution recordings were un- Hemiptera that have short stereotyped signals without available for comparisons with sounds produced by much signal modulation (Gogala 1984). There was a the Lygaeidae, , , Thau- positive relationship between chirp interval and body mastellidae, and Leptopodidae, which have similar length. Jadera haematoloma sounds lack the complex wing and tergal plate morphology. The recordings by modulation or multiple song types that have been Gogala (1990) of adult male Arhyssus hyoscyami observed in the courtship rituals of some Pentatomi- (Hemiptera: Rhopalidae) very closely resembled dae (Cˇ okl et al. 2001) and strongly differ from the those recorded here; signals were succinct and highly continuous, heavily modulated sounds produced by stereotyped, however, the speciÞed frequency range 126 ANNALS OF THE ENTOMOLOGICAL SOCIETY OF AMERICA Vol. 105, no. 1 was much lower than those recorded here (80 Hz), tures may be conserved, the sounds themselves may possibly because of differences in recording equip- not show phylogenetic signal. Instead, the function of ment. From these observations drawn from the liter- signals and the behavioral context might be more ature, it is clear that a shared sound producing mech- important in shaping signal structure in this group of anism does not necessarily confer acoustic similarity. insects. Further exploration of signal diversity in target The behavioral context of sound production sug- clades, and descriptions of each behavioral context gests that sounds may be signals in response to an would help to address this exciting question. approach by another individual. Sound production began by an individual who was approached by a conspeciÞc moving at greater speed. The speciÞc be- Acknowledgments havioral context of sound production by J. haema- We thank S. Tonia Hsieh for the use and assistance with toloma is unique because the sounding role in an high-speed Þlming equipment, Susan Halbert at the Florida encounter is determined by the relative speed of the State Arthropod Collections, and Ann Heatherington of the individual, regardless of its sex. In most hemipteran UF Department of Geology for assistance with SEM image communication systems, the signaler role is stereo- collection. We also thank H. J. Brockman and C. W. Miller for typed by sex or age as is appropriate signaler identi- help on early drafts of this manuscript. Funding for this Þcation during courtship and mother-offspring inter- research was provided by an NSF graduate research fellow- ship and University of Florida Alumni Fellowship to AFZ. actions (Cocroft 2001, Cocroft and Rodriguez 2005). Because two adult J. haematoloma are already touch- ing when sounds are produced, and because it occurs References Cited in the absence of offspring, it is clear that sounds are Aldrich, J. R., S. P. Carroll, W. R. Lusby, M. J. Thompson, J. P. not used as signals to attract conspeciÞcs from a dis- Kochansky, and R. M. Waters. 1990a. Sapindaceae, cya- tance or to alert offspring to threats. On the contrary, nolipids, and bugs. J. Chem. Ecol. 16: 199Ð210. increased repetition rates associated with being Aldrich, J. R., S. P. Carroll, J. E. Oliver, W. R. Lusby, A. A. crawled upon and simultaneous body rocking behav- Rudmann, and R. M. Waters. 1990b. Exocrine secre- ior suggest that sounds are more common in proximate tions of scentless plant bugs: Jadera, and interactions. The association of sounds with close species (Hemiptera: Heteroptera: Rhopalidae). Biochem. proximity between individuals and with body rocking Syst. Ecol. 18: 369Ð376. suggests these vibrations are used to deter contact Ashlock, P. T., and J. D. Lattin. 1963. Stridulatory mecha- with conspeciÞcs. Decreased acoustic response to be- nisms in the Lygaeidae, with a new American genus of (Hemiptera: Heteroptera). Ann. Entomol. Soc. ing crawled upon during the mating season might Am. 56: 693Ð703. suggest that signals are being used to indicate that a Autumn, K., S. T. Hsieh, D. M. Dudek, J. Chen, C. Chitaphan, female or male is unreceptive to mating. However, the and R. J. Full. 2006. Dynamics of geckos running verti- results show that during male-female encounters, cally. J. Exp. Biol. 209: 260Ð272. sounds did not inßuence mating outcome during the Carroll, S. P., and C. Boyd. 1992. Host race radiation in the mating season. Therefore further investigation into soapberry bug: natural history with the history. Evolution the inßuence of Þeld density, food availability, and 46: 1052Ð1069. rates of cannibalism in the Þeld are needed to address Carroll, S. P., and P. S. Corneli. 1995. Divergence in male the seasonality of this behavior. mating tactics between two populations of the soapberry bug: genetic change and the evolution of a plastic reaction All inter-species interactions with other arthropods norm in a variable social environment. Behav. Ecol. 6: failed to produce sounds. Though handling the insects 46Ð56. caused them to produce sound, all other threat stimuli Carroll, S. P., and J. E. Loye. 1987. Specialization of Jadera failed to elicit signals. Jadera haematoloma is chemi- species (Hemiptera: Rhopalidae) on seeds of Sapin- cally protected, and may only experience real threat daceae (Sapindales), and coevolutionary responses of from larger vertebrate predators (Aldrich et al. 1990a, defense and attack. Ann. Entomol. Soc. Am. 80: 373Ð378. 1990b) which have been shown to respond negatively Carroll, S. P., M. Marler, R. Winchell, and H. Dingle. 2003. to aposematism in this species (Ribeiro 1989). These Evolution of cryptic ßight morph and life history differ- results support the hypothesis that sounds target ences during host race radiation in the soapberry bug, Jadera haematoloma Herrich-Schaeffer (Hemiptera: nearby conspeciÞcs and are not frequently used in Rhopalidae). Ann. Entomol. Soc. Am. 96: 135Ð143. inter-species interactions. Cocroft, R. B. 2001. Vibrational communication and the Our exploration of vibrations produced by J. haema- ecology of group-living, herbivorous insects. Am. Zool. toloma revealed that the vibration-producing struc- 41: 1215Ð1221. tures are similar to other closely related Hemiptera. Cocroft, R. B., and R. L. Rodriguez. 2005. The behavioral The tergal plate plectrum that J. haematoloma uses to ecology of insect vibrational communication. Bioscience produce sounds is shared by several other Heterop- 55: 323Ð334. tera, supporting the hypothesis presented by Schaefer Cˇ okl, A. 1983. Functional properties of vibroreceptors in (1993) that the tergal plate plectrum is largely con- the legs of Nezara viridula [L] (Heteroptera: Pentato- midae). J. Comp. Physiol. A 150: 261Ð269. served in terrestrial Heteroptera with multiple inde- Cˇ okl, A., H. L. McBrien, and J. G. Millar. 2001. Comparison pendent origins of wing stridulating structures. How- of substrate-borne vibrational signals of two stink bug ever, there is still considerable acoustic variation species, Acrosternum hilare and Nezara viridula (Het- among species using the same mechanism as J. haema- eroptera: ). Ann. Entomol. Soc. Am. 94: toloma. Therefore, although sound producing struc- 471Ð479. January 2012 ZYCH ET AL.: VIBRATORY COMMUNICATION IN J. haematoloma 127

Cornell Lab of Ornithology. 2008. Raven interactive sound Masters, W. M. 1979. Insect disturbance stridulation: its de- analysis software v1.3. Cornell Lab of Ornithology, Ithaca, fensive role. Behav. Ecol. Sociobiol. 5: 187Ð200. NY. Numata, H., K. Masahiro, F. Hisashi, and T. Hidaka. 1989. Drosopoulos, S., E. Eliopoulos, and P. Tsakalou. 2006. Is Sound production in the bean bug, Riptortus clavatus migration responsible for the peculiar geographic distri- Thunburg (Heteroptera: ). Appl. Entomol. Zool. bution and speciation based on acoustic divergence of 24: 169Ð173. two cicadas in the Aegean Archipelago?, pp. 219Ð226. In Percy, D. M., G. S. Taylor, and M. Kennedy. 2004. Psyllidae: S. Drosopoulos and M. Claridge (eds.), Insect sounds and acoustic diversity, mate recognition, and phylogenetic communication: physiology, behavior, ecology and evo- signal. Invertebr. Syst. 20: 431Ð445. lution. Taylor & Francis Group, Boca Raton, FL. Photron Limited. 2006. Photron Fastcam Viewer for high Eliopoulos, E. 2006. Sound and techniques in sound analy- speed digital imaging. Photron Limited, San Diego, CA. sis, pp. 11Ð33. In S. Drosopoulos and M. F. Claridge (eds.), Polhemus, J. T. 1994. Stridulatory mechanisms in aquatic Insect sounds and communication: physiology, behaviour, and semiaquatic Heteroptera. J. N.Y. Entomol. Soc. 102: ecology, and evolution. Taylor & Francis, Boca Raton, FL. 270Ð274. Fisher, R. A. 1922. On the interpretation of chi-squared Ribeiro, S. T. 1989. Group effects and aposematism in Jad- from contingency tables, and the calculation of p. J. R. era haematoloma (Hemiptera: Rhopalidae). Ann. Ento- Stat. Soc. 85: 87Ð94. mol. Soc. Am. 82: 466Ð475. [FLEPPC] Florida Exotic Pest Plant Council. 2009. Rodriguez, R. L., L. E. Sullivan, and R. B. Cocroft. 2004. FLEPPC 2009 list of invasive plant species Ð Fall 2009. Vibrational communication and reproductive isolation in FLEPPC, University of Georgia, Athens, GA. the Enchenopa binotata species complex of Gogala, M. 1984. Vibration producing structures and songs (Hemiptera: Membracidae). Evolution 58: 571Ð578. of terrestrial Heteroptera as systematic character. Biolo- SAS Institute 2008. JMP statistical computing software, ver- ski Vestnik (Ljubljana). 32: 19Ð36. sion 2.0. SAS Institute, Cary, NC. Gogala, M. 1990. Distribution of low frequency vibrational Schaefer, C. W. 1980. The sound producing structures of songs in local Heteroptera Scolopia. 1: 125Ð132. some primitive (Hemiptera: Heterop- Gogala, M. 2006. Vibratory signals produced by Heterop- tera). J. N.Y. Entomol. Soc. 88: 230Ð235. tera Ð Pentatomorpha and , pp. 275Ð295. In Schaefer, C. W. 1993. Origins of the New World (Hemiptera: Rhopalidae). Ann. Entomol. Soc. Am. 86: S. Drosopoulos and M. Claridge (eds.), Insect sounds and 127Ð133. communication: physiology, behavior, ecology and evo- Schaefer, C. W., and R. T. Pupedis. 1981. A stridulatory lution. Taylor & Francis Group, Boca Raton, FL. device in certain Alydinae (Hemiptera: Heteroptera: Aly- Gogala, M., A. Cˇ okl, K. Draslar, and A. Blazevic. 1974. Sub- didae). J. Kans. Entomol. Soc. 54: 143Ð152. strate-borne sound communication in Cydnidae (Het- Schuh, R. T., and J. A. Slater. 1995. True bugs of the world eroptera). J. Comp. Physiol. 94: 25Ð31. (Hemiptera: Heteroptera). Cornell University Press, Henry, T. J. 1997. Phylogenetic analysis of family groups Ithaca, NY. within the infraorder Pentatomomorpha (Hemiptera: Shaw, S. R. 1994. Detection of airborne sound by a cock- Heteroptera), with emphasis on the . Ann. roach Ôvibration detectorÕ: a possible missing link in insect Entomol. Soc. Am. 90: 275Ð301. auditory evolution. J. Exp. Biol. 193: 13Ð47. Jorigtoo, N., C. W. Schaefer, and J. A. Lockwood. 1998. Slater, J. A., and R. M. Baranowski. 1978. How to know the Stridulatory apparatus of Piesma Le Peletier & Serville true bugs (Hemiptera-Heteroptera). William C. Brown (Hemiptera: Piesmatidae). Ann. Entomol. Soc. Am. 91: Company, Dubuque, IA. 848Ð851. Tishechkin, D. Y. 2006. Acoustic characters in the classiÞca- Lawson, F. A., and J. Chu. 1971. A scanning electron mi- tion of higher taxa of (Hemiptera), pp. croscopy study of stridulating organs in two Hemiptera. 99Ð110. In S. Drosopoulos and M. Claridge (eds.), Insect J. Kans. Entomol. Soc. 44: 245Ð253. sounds and communication: physiology, behavior, ecology Leston, D., J. G. Pendergrast, and T.R.E. Southwood. 1954. and evolution. Taylor & Francis Group, Boca Raton, FL. ClassiÞcation of the terrestrial Heteroptera (Geocori- Virant-Doberlet, M., and A. Cˇ okl. 2004. Vibrational com- sae). Nature 174: 91Ð94. munication in insects. Neotrop. Entomol. 33: 121Ð134. Mankin, R. W., D. Shuman, and J. A. Coffelt. 1996. Noise Waite, J. K., and J. S. Hwang. 2002. Pests of litchi and longan, shielding of acoustic devices for insect detection. J. Econ. pp. 331Ð360. In J. E. Pena, J. L. Sharp, and M. Wysoki Entomol. 89: 1301Ð1308. (eds.), Tropical fruit pests and pollinators: biology, eco- Mankin, R. W., J. Brandhorst-Hubbard, K. L. Flanders, M. nomic importance, natural enemies and control. CAB Zhang, R. L. Crocker, S. L. Lapointe, C. W. McCoy, J. R. International Oxford, Oxford, United Kingdom. Fisher, and D. K. Weaver. 2000. Eavesdropping on in- Wenninger, E. J., D. G. Hall, and R. W. Mankin. 2009. sects hidden in interior structures of plants. J. Econ. Vibrational communication between the sexes in Dia- Entomol. 93: 1173Ð1182. phorina citri (Hemiptera: Psyllidae). Ann. Entomol. Soc. Mankin, R. W., M. T. Smith, J. M. Tropp, E. B. Atkinson, and Am. 102: 547Ð555. D. Y. Jong. 2008. Detection of Anoplophora glabripennis Zych, A. F. 2010. Seasonal group behavior of adult Jadera (Coleoptera: Cerambycidae) larvae in different host haematoloma (Hemiptera: Rhopalidae) in Central Florida trees and tissues by automated analyses of sound-impulse following host seed exhaustion. Fla. Entomol. 93: 644Ð646. frequency and temporal patterns. J. Econ. Entomol. 101: 838Ð849. Received 4 March 2011; accepted 28 September 2011.