Research Article Journal of Orthoptera Research 2024, 33(1): 13-19 Factors related to sound production by the Chinese grasshopper Acrida cinerea during escape TATSURU KUGA!, Ett! KASUYA?*4 1 Graduate School of Systems Life Sciences, Kyusyu University, 744 Motooka, Nishi-ku, Fukuoka 819-0395, Japan. 2 Current address: Hiroshima City Insectarium, 10173 Aza-Fujigamaru, Fukuda-cho, Higashi-ku, Hiroshima 732-0036, Japan. 3 Department of Biology, Kyushu University, 744 Motooka, Nishi-ku, Fukuoka 819-0395, Japan. 4 Current address: Graduate School of Science, Osaka Metropolitan University, Nakamozu, Naka-ku, Sakai, Osaka 599-8531, Japan. Corresponding author: Tatsuru Kuga (tkuga486@gmail.com) Academic editor: Laurel B. Symes | Received 21 January 2023 | Accepted 27 April 2023 | Published 9 January 2024 https://zoobank. org/OBAE7A5B-385B-4BC1 -8456-3669FF03BB44 Citation: Kuga T, Kasuya E (2024) Factors related to sound production by the Chinese grasshopper Acrida cinerea during escape. Journal of Orthoptera Research 33(1): 13-19. https://doi.org/10.3897/jor.33.100865 Abstract Many grasshopper species produce conspicuous sounds while escap- ing from approaching predators; however, they occasionally escape without producing sounds. The Chinese grasshopper, Acrida cinerea, often exhibits noisy escape behavior. Therefore, a field experiment was conducted using A. cinerea to identify factors related to the production of sound during es- cape. This study utilized a predator model with an investigator approaching A. cinerea three times. We examined the relationship between the produc- tion of sound during escape and the following factors: ambient tempera- ture and relative humidity as environmental factors; sex, body length, body weight, and limb autotomy as prey traits; and the repeated approach as a predator trait. The relationships between noisy escape and flight initiation distance (i.e., predator-prey distance when the prey initiates the escape), distance fled (i.e., distance the prey covered during the escape), and the mode of locomotion during escape (i.e., flying or jumping) were also ex- amined. Noisy escape was observed only in males that escaped by flying, whereas the females and males that escaped by jumping invariably escaped silently. Among males that flew, noisy escape was related to ambient tem- perature, limb autotomy, and distance fled. The proportion that produced sound increased in parallel with the ambient temperature and distance fled. This proportion was lower among individuals that had autotomized one of their hind legs. These results indicate that noisy escape behavior is most frequent in healthy male A. cinerea under warm conditions. Keywords antipredator tactics, crepitation, distance fled, flight initiation distance, fly, jump, predator-prey interaction Introduction Many animals exhibit conspicuous behavior when they escape from approaching predators (Edmunds 1974, Ruxton et al. 2018). For example, Thomson's gazelle Eudorcas thomsonii (Ginther, 1884) leaps vertically (Caro 1986), the skylark Alauda arvensis Linnaeus, 1758 sings (Cresswell 1994), and the mountain katydid Acripeza reticulata Guérin-Méneville, 1832 reveals its bright body color (Umbers et al. 2019) when escaping from predators. Intui- tion suggests that such conspicuous behaviors may attract preda- tors’ attention and lead to failure of the escape, unless performed by unpalatable prey as aposematic signals (e.g., Kang et al. 2016). Contrary to this notion, such conspicuous behaviors increase the survival rate of some prey animals (Ruxton et al. 2018). However, due to the lack of experimental evidence, the function of conspicu- ous escape in most prey animals remains to be determined. Many species of grasshoppers produce sounds when they es- cape by flying (Otte 1970). These sounds are considered an an- tipredator defensive strategy (Edmunds 1974, Low et al. 2021). Nevertheless, silent escape is occasionally observed in species that are capable of producing sound in flight (e.g., Acrida cinerea (Thunberg, 1815), Kuga, personal observation). Clarification of the factors related to the production of sounds during escape is necessary to reveal the function of this phenomenon (herein termed noisy escape) in grasshoppers. Previous studies have re- vealed factors related to some types of grasshopper escape strate- gies, particularly the predator-prey distance where the prey initi- ates the escape, termed flight initiation distance (FID), or the dis- tance that the prey covered during the escape, termed distance fled (DF) (Lagos 2017). However, factors related to the noisy escape of grasshoppers remain unknown. This study examined the environmental factors as well as prey and predator traits that may be related to the noisy escape of the Chinese grasshopper A. cinerea (Fig. 1A). In Japan, this grasshop- per often produces conspicuous sounds while escaping from an approaching human by flying (Takaie 1998). We focused on en- vironmental factors (i.e., ambient temperature and relative hu- midity), prey traits (i.e., sex, autotomy of hind limb, body length, and body weight), and a predator trait (i.e., repeated approach- es). It has been shown that temperature, sex, limb autotomy, and Copyright Tatsuru Kuga & Eiiti Kasuya. This is an open access article distributed under the terms of the Creative Commons Attribution License (CC BY 4.0), which permits unre- stricted use, distribution, and reproduction in any medium, provided the original author and source are credited. JOURNAL OF ORTHOPTERA RESEARCH 2024, 33(1) 14 repeated approaches affect the escape behavior of orthopteran in- sects (Lagos 2017). Moreover, the body size of invertebrates can af- fect their escape behaviors (Bateman and Fleming 2015). Thus, we included body length and weight in the prey traits. Humidity was included as an environmental factor because it affects the flight behavior of some insects (Belton 1986, Parmezan et al. 2021). As a function of conspicuous escape, the sudden disappearance of conspicuous behavior during the escape may confuse predators about the location of the prey and deter predators from search- ing for it (Edmunds 1974, Loeffler-Henry et al. 2018). Such con- spicuous behavior is termed flash behavior. Loeffler-Henry et al. (2021) hypothesized that flash behavior is effective for prey with longer FID because an approaching predator located far away will be unaware of the prey’s appearance at rest when it does not be- have conspicuously. Evidence obtained from an experiment using computer-generated prey and a human predator model supported this hypothesis (Loeffler-Henry et al. 2021). If the noisy escape of A. cinerea is a flash behavior, it may be observed more frequently in individuals with longer FID. Thus, the relationship between sound production and FID was investigated to see if noisy escape is consistent with flash behavior. Another function of conspicuous escape is to send a signal to predators that the prey has a good ability for escape and de- ter them from approaching the prey (Vega-Redondo and Hasson 1993). This signal is termed a pursuit-deterrent signal. The lizard Psammodromus algirus (Linnaeus, 1758) escapes farther away when it produces sounds during escape attempts (Martin and Lopez 2001). If the noisy escape of A. cinerea is a pursuit-deterrent signal, it may be observed more frequently in individuals with longer DF, similar to P. algirus. Therefore, the relationship between noisy es- cape and DF of A. cinerea was examined. Many grasshopper species, including A. cinerea, escape via two locomotion modes: flying and jumping (Forsman 1999, Maeno et al. 2019). The relationship between locomotion modes and sound production was also examined. Materials and methods Study animals and study sites.—Acrida cinerea is commonly found in Japan and characteristically produces sound during flight (Or- thopterological Society of Japan 2006). The grasshoppers produce this sound (crepitation) by clapping their hindwings (Kuga and Kasuya 2021). We conducted field experiments with adult A. ci- nerea at three grassland sites at Kyushu University, Fukuoka, Ja- pani(33°35-42aN130° 13 085F2 33°35:33 Ni 130"13.10°ER;tand 33 °35'33"N, 130°13'07"E, Fig. 1B). Many adults were observed at the sites during the experimental period (August 1-September 25, 2017). Grasshoppers were identified following the taxonomic key of the Orthopterological Society of Japan (2006). Experimental procedure.—The escape behavior of grasshoppers is often induced by the approach of an investigator (Cooper Jr. 2006, Butler 2013, Bateman and Fleming 2014, Collier and Hodgson 2017, Maeno et al. 2019). In the present study, utilizing a predator model, an investigator approached A. cinerea to provoke escape. A high-speed digital camera (Casio EX-ZR1700, Tokyo, Japan, frame rate: 120 frames/s) and non-high-speed digital camera (PENTAX WG-1, Tokyo, Japan, frame rate: 30 frames/s [August 1-8, 2017]; Sony DSC-WX170, Tokyo, Japan, frame rate: 60 interlaced-fields/s [August 9-September 25, 2017]) were attached to the waist of the investigator during experiments to record the escape behavior of T. KUGA AND E. KASUYA A. cinerea. The high-speed digital camera was placed next to the non-high-speed digital camera. The appearance of the investigator remained unchanged during the approaches to avoid potential ef- fects on escape behavior. At each site, the investigator recorded the behavior of an indi- vidual grasshopper during three consecutive escapes. The experi- ments took place from 10:00 to 15:00 each day. An interval be- tween experiments in the same sites was 1h or longer to minimize potential effects of the previous experiment that could influence the results of the next experiment (e.g., disturbance of the grass). Experiments were not conducted during periods of rain. The experimental procedure that was followed at each site in- cluded the three steps below: identification of an individual, three consecutive approaches to the target, and capture of the target. The investigator searched for an individual A. cinerea while walking at one step per second (walking speed, mean + standard deviation [SD] = 36.6 + 1.0 cm/s, n = 20). The walking speed was maintained using the metronome sound from an audio player (MD720J/A, Apple, California, USA; METRONOME STAR app v.2.0.0, 60 beats/ min). The same area in the site was never searched more than once during the experimental procedure. Following the identification of an individual A. cinerea, the investigator approached the target at the same walking speed. The first encounter with a target often occurred while the grasshopper was escaping, and this escape was regarded as the first attempt. The first approach was terminated when the target initiated the escape. Then, markers (wire rings with a diameter of 5 cm) were quickly placed on the investigator's position and the initial location of the target grasshopper at the start of the first escape. The second and third approaches were con- ducted in the same manner immediately after the markers were placed. Following the three consecutive approaches, the investiga- tor captured the grasshopper and placed a marker on the position of the grasshopper at the end of the third escape. When the investigator failed in either the approach step or the capture step, the step of identification was restarted at the same site. The three steps were repeated at that site until the investi- gator accomplished all three steps or searched the whole area of the site for the target grasshopper. All captured grasshoppers were maintained in a laboratory (temperature: 22-26°C; food: mostly Paspalum urvillei Steudel) until the end of the study. Measurements.—The sound produced by A. cinerea is detectable by the human ear. The investigator recorded whether sounds were produced by A. cinerea during the escape attempts. This data re- cording was confirmed using videos captured by the non-high- speed digital camera. Video analyses were conducted using the BORIS v.4.1.11 software (Friard and Gamba 2016). Locomotion modes during escape attempts were classified ac- cording to the video recorded by the high-speed digital camera. Wing flapping after takeoff indicated flying, while lack of wing flapping after takeoff denoted jumping. A preliminary experiment showed that target grasshoppers often escaped outside the camera frame. To confirm the locomotion modes of targets outside the camera frame, the investigator observed the locomotion modes vis- ually while approaching the target in the field. When wing flapping of the target was not recorded in the video but was observed in the field, the locomotion mode of that target was classified as flying. FID and DF were recorded by measuring the distances be- tween markers using a steel tape measure to the nearest 1 cm. FID was measured as the distance between two markers placed on the positions of the grasshopper and the investigator at the initia- JOURNAL OF ORTHOPTERA RESEARCH 2024, 33(1) T. KUGA AND E. KASUYA 15 Fig. 1. Photos of male A. cinerea (A) and its habitat where field experiments were conducted (B). tion of each escape attempt. DF was measured as the distance between the two markers placed at the positions of the grasshop- per at the initiation and end of each escape attempt. The second and third escape attempts were induced immediately after the previous escape. Hence, the markers placed at the positions of the grasshopper at the initiation of the second and third escape attempts were considered to be placed at the positions of the grasshopper at the end of the first and second escape attempts, respectively. These measurements were conducted after capturing the target grasshopper. The ambient temperature (to the nearest 0.01°C) and relative humidity (to the nearest 0.01%) were recorded after the capture of the target. We used a temperature and humidity data logger (Sato- shoji LITE5032P-RH, Kanagawa, Japan) for the recording. During each experiment, the data logger was hung on a tree branch at a height of 140-200 cm. The morphological traits of the individuals were measured in a single day after the end of the final experiment. Body weight to the nearest 0.01 g was measured using an electronic balance de- vice (Sartorius 1416MP8, Gottingen, Germany). The grasshoppers defecated frass, thereby reducing their body weight between the time of collection and that of the measurement. Measurement of body weight at the time of escape was important to examine the relationship with noisy escape. Thus, the total weight of the grass- hopper and its frass, rather than the grasshopper’s weight alone, was measured. Using a digital caliper, body length was measured to the nearest 0.01 mm as the distance from the tip of the head to the end of the forewings (Mitutoyo CD-20C, Kanagawa, Japan). Statistical analyses.—The following statistical tests were conduct- ed with R v.4.1.1 (R Core Team 2021) in RStudio v.2021.9.0.351 (RStudio Team 2021). The significance level was set at 0.05. We examined the relationship between noisy escape, sex, and locomotion modes. The frequency of noisy escape was compared between males and females in each of the three consecutive escape attempts using Fisher's exact test. This test was also used to exam- ine sex differences in the frequency of each locomotion mode. Factors related to sound production in the first escape at- tempt were examined using generalized linear models (GLMs) with a quasi-binomial error structure and logit link. The models were fitted to the data of males that flew in the first escape at- tempt because females and males that escaped by jumping did not produce sounds (see Results). The objective variable was sound production (no = 0; yes = 1) in the first escape attempt. The explanatory variables were ambient temperature, humid- ity, body length, body weight, limb autotomy (no = 0; yes = 1), FID, and DF in the first escape attempt. We also fitted models that contained a quadratic term of temperature or humidity as another explanatory variable to the data and examined the pos- sibility that these parameters affect sound production quadrati- cally. There were no significant effects found in these quadratic terms of temperature (coefficient + standard error [SE] = -0.06 + 0.03, t = -1.795, degree of freedom [df] = 125, p = 0.075) and humidity (coefficient + SE = -0.002 + 0.006, t = -0.323, df = 125, p = 0.747). Similarly, there were no significant effects of the quadratic terms of temperature and humidity when the model contained both these terms at the same time (temper- ature: t = -1.755, df = 124, p = 0.082; humidity: t = -0.182, df = 124, p = 0.856). Thus, these quadratic terms were removed from the model. Changes in the frequency of noisy escape through repeated es- cape attempts were tested using the exact McNemar test. Changes in frequency were examined for each of the first and second es- cape attempts and for the second and third escape attempts. We used only the data of males that escaped by flying in the three escape attempts for this and the subsequent statistical tests on re- peated escapes. Factors that affect sound production during repeated escape attempts were examined using GLMs with quasi-binomial error structure and logit link. The objective variable was sound produc- tion in the second or third escape attempt. In the model for sound production during the second escape attempt, the explanatory variables were FID, DE, and sound production in the first escape attempt, as well as FID and DF in the second escape attempt. In the model for sound production during the third escape attempt, the explanatory variables were FID, DE and sound production in the second escape attempt, as well as FID and DF in the third es- cape attempt. JOURNAL OF ORTHOPTERA RESEARCH 2024, 33(1) 16 Results We collected data on three consecutive escape attempts of 136 males and 13 females (Table 1). Sound was produced by approxi- mately 70% of male A. cinerea (first escape: 75%, second escape: 76%, third escape: 71%); most males escaped by flying (first es- cape: 99%, second escape: 96%, third escape: 96%) (Table 1). Although some males escaped by jumping in each of the three escape attempts, they did not produce sounds (Table 1). Female A. cinerea did not produce conspicuous sounds regard- less of the locomotion mode (Table 1). Significant sex differences in sound production were detected in each of the three consecu- tive escape attempts (Fisher's exact test: p < 0.001 for all escape at- tempts). Females escaped by jumping more frequently than males (Fisher's exact test: p < 0.001 for all the escape attempts). For males that flew in the first escape attempt, temperature, limb autotomy, and DF were significantly related to sound pro- duction (Table 2). More males produced sounds under high tem- peratures (temperature when sounds were produced, mean + SD = 30.73 + 2.48°C; n = 102; temperature when sounds were not produced, mean + SD = 29.06 + 3.35°C; n = 32). Autotomized Table 1. Numbers of individuals in the locomotion modes and sound production in three escape attempts. Sex Attempt Locomotion mode Sound production No Yes Male First Fly 32 102 Jump 2 0 Second Fly 2 103 Jump 6 0 Third Fly 35 96 Jump 5 0 Female First Fly 6 0 Jump 7 0 Second Fly 5 0 Jump 8 0 Third Fly 3 0 Jump 10 0) Table 2. Result of GLM on sound production (no = 0; yes = 1) in the first escape attempt. The error structure was quasi-likelihood (“quasibinomial” in GLM function of R), and the link function was logit. The model contained all the explanatory variables at the same time. Only data of males that escaped by flying in the first escape attempt were included in the analysis. Definitions/Ab- breviations: Autotomy, the occurrence of autotomy of the hind leg (no = 0; yes = 1); Coefficient, estimated value of the coefficient; p, p-value of the statistical test on the coefficient; SE, standard error of the estimate of the coefficient; t, value of t-statistics (df = 126). Explanatory variable Coefficient SE t p (Intercept) -4.919 8.952 -0.550 0.584 Temperature 0.266 0-121 2:196 0.030 Humidity -0.059 0.049 -1.216 0.226 Body length 0.130 pote 7 0.735 0.464 Weight -17.453 9.323 -1.872 0.064 Autotomy -1.865 0.818 —2.281 0.024 FID -—0.006 0.008 -0.757 0.450 DF 0.012 0.004 Deo ad. <0.001 T. KUGA AND E. KASUYA males exhibited noisy escape less frequently than intact males (53%, n = 17 vs. 79%, n = 117, respectively). Sound was pro- duced more frequently by males with longer DF (Fig. 2A). No significant relationships were observed between sound produc- tion and body length, body weight, humidity, or FID (Table 2; Fig. 2B for FID). Some of the males that escaped by flying in all three consecu- tive escape attempts showed both noisy and silent flight (Fig. 3). There was no significant change in the frequency of noisy escape between the first and second escape attempts (exact McNemar test: p = 0.392) and between the second and third escape attempts (ex- act McNemar test: p = 0.327). In the second escape attempt, sound production was signifi- cantly related to DF (Table 3). Similar to the first escape attempt, in the second attempt, males that produced sounds flew further than those that did not produce sounds (Fig. 2C). In the second attempt, FID was not significantly related to sound production (Table 3; Fig. 2D). FID, DE and sound production in the first at- tempt also showed no significant relationship to sound produc- tion in the second attempt (Table 3). Sound production in the third escape attempt was significantly related to DF in the third escape attempt and sound production in the second escape attempt (Table 3). In the third attempt, males that produced sound during their escape showed longer DF than those that escaped without sound (Fig. 2E). The frequency of noisy escape in the third attempt was higher in males that produced sound in the second attempt than in those that escaped silently in the second attempt. There were no significant relationships be- tween sound production in the third attempt and the third FID (Fig. 2F), second FID, or second DF (Table 3). Table 3. Result of GLM on sound production (no = 0; yes = 1) in repeated escape attempts. The error structure was quasi-likelihood (“quasibinomial” in GLM function of R), and the link function was logit. The model contained all the explanatory variables at the same time. Only data of males that escaped by flying in all three consecutive escape attempts were included in the analyses. Definitions/Abbreviations: Coefficient, estimated values of the co- efficient; DF1, DF2, and DF3, DF in the first, second, and third es- cape attempts, respectively; FID1, FID2, and FID3, FID in the first, second, and third escape attempts, respectively; SE, standard error of the estimate of the coefficient; Sound1, Sound2, and Sound3, sound production (no = 0; yes = 1) in the first, second, and third escape attempts, respectively; p, p-value of the statistical test on the coefficient; t: value of t-statistics (df = 120). Objective Explanatory Coefficient SE t p variable variable Sound2 (Intercept) -1.180 0.695 -1.698 0.092 Sound1 1.014 0.584 1.737 0.085 FID1 0.020 0.010 1.928 0.056 FID2 -0.014 0.010 -1.332 0.185 DFI1 -0.002 0.003 -0.545 0.587 DF2 0.009 0.004 2.602 0.010 Sound3 (Intercept) -0.985 0.634 -1.554 0.123 Sound2 2.081 0.552 3.774 <0.001 FID2 -0.005 0.008 -0.601 0.549 FID3 -0.002 0.009 -0.167 0.867 DF2 -0.002 0.003 -0.527 0.599 DF3 0.006 0.003 1.990 0.049 JOURNAL OF ORTHOPTERA RESEARCH 2024, 33(1) T. KUGA AND E. KASUYA 17 A C E 700 O pas Lit _. 600 = = E 500 oa = 2 L£ 400 “A rE Ps ie o00 ! a O Q 200 —_ 100 = ee 0 No Yes No Yes No Yes (3735 _=