Anti-predator defences of a bombardier beetle: is bombing essential for successful escape from frogs?

Introduction

Physical and chemical defences have evolved in many organisms to protect against natural enemies (Edmunds, 1974; Eisner, Eisner & Siegler, 2005). For example, some plant and animal species have developed physical deterrents such as thorns and spines (Edmunds, 1974; Cooper & Owen-Smith, 1986; Eisner, 2003; Sugiura & Yamazaki, 2014; Sugiura, 2016; Ito, Taniguchi & Billen, 2016), while other species produce defensive chemicals, including toxic substances, to prevent themselves from being eaten (Eisner, 2003; Eisner, Eisner & Siegler, 2005; Derby, 2007; Mithöfer & Boland, 2012). Organisms whose defence mechanisms can cause severe injury to their natural enemies have also evolved warning signals, such as conspicuous body colouration or particular sounds (Lev-Yadun, 2001; Ruxton, Sherratt & Speed, 2004; Inbar & Lev-Yadun, 2005; Bonacci et al., 2008; Lev-Yadun, 2009; Bura, Kawahara & Yack, 2016; Sugiura & Takanashi, 2018). In response, predators have evolved specific abilities to avoid such well-defended prey by recognising warning colouration or detecting chemical signals (Edmunds, 1974; Endler, 1991; Ruxton, Sherratt & Speed, 2004; Skelhorn & Rowe, 2006; Williams et al., 2010).

Adult bombardier beetles (Coleoptera: Carabidae: Brachinini) bomb, i.e., discharge noxious chemicals from the tip of their abdomens at temperatures of approximately 100 °C, when they are disturbed or attacked (Aneshansley et al., 1969; Dean, 1979; Eisner, 2003; Eisner, Eisner & Siegler, 2005; Arndt et al., 2015). Such ejection of hot chemicals is only known in the coleopteran family Carabidae. Previous studies have investigated how bombardier beetles successfully defend against predators (Eisner, 1958; Eisner & Meinwald, 1966; Eisner & Dean, 1976; Dean, 1980a; Eisner, Eisner & Aneshansley, 2005; Eisner et al., 2006). Bombardier beetles can aim their abdominal discharge in virtually any direction, spraying various parts of their own bodies (e.g., legs and dorsal surface) with the toxic chemicals (Eisner & Aneshansley, 1999). Dean (1980b) reported that predators displayed intense responses to the unheated chemical discharges of bombardier beetles in experiments. This suggests that the cooled chemicals coating the beetles’ body surfaces function as the primary defence against predators. Successful defence mediated by chemicals on the body surfaces of beetles may reduce the costs of spraying (bombing). Further research is needed to clarify the relative importance of chemical toxicity and heat for overall successful anti-predatory defence.

Frogs and toads are important predators of carabid beetles (Larochelle, 1974a; Larochelle, 1974b). However, bombardier beetles have rarely been found in the gut contents and faeces of frogs and toads (Larochelle, 1974a; Larochelle, 1974b; Sarashina, Yoshihisa & Yoshida, 2011; except Mori, 2008), suggesting that bombing prevents toads and frogs from swallowing and ingesting beetles (Eisner & Meinwald, 1966; Dean, 1980a; Eisner, 2003; Sugiura & Sato, 2018). Still, only a few studies have investigated the factors that cause anuran predators to stop preying on bombardier beetles (Dean, 1980b). Elucidating these ecological factors would contribute to a better understanding of the evolution of anti-predatory defences in insects.

This study aims to investigate the responses of the black-spotted pond frog Pelophylax nigromaculatus (Hallowell) (Anura: Ranidae) to the defensive behaviour of the bombardier beetle Pheropsophus jessoensis (Morawitz). Pheropsophus jessoensis is a bombardier beetle found in East Asia (Ueno, Kurosawa & Sato, 1985; Jung et al., 2012); the beetle is a common inhabitant of farmlands, grasslands, and forest edges in Japan (Habu & Sadanaga, 1965; Ueno, Kurosawa & Sato, 1985; Yahiro et al., 1992; Ishitani & Yano, 1994; Fujisawa, Lee & Ishii, 2012; Ohwaki, Kaneko & Ikeda, 2015; Sugiura & Sato, 2018). Adult Ph. jessoensis eject toxic chemicals (1,4-benzoquinone and 2-methyl-1,4-benzoquinone) at a temperature of approximately 100 °C from their rear ends in response to predator attacks (Video S1; Kanehisa & Murase, 1977; Kanehisa, 1996). Pelophylax nigromaculatus is a true frog inhabiting wetlands and farmlands of East Asia (Liu et al., 2010; Tsuji et al., 2011; Komaki et al., 2015), being one of the most abundant frog species of traditional agricultural landscapes including farmlands, grasslands, and forest edges (Hirai, 2002; Honma, Oku & Nishida, 2006; Tsuji et al., 2011; Matsuhashi & Okuyama, 2015). Using its tongue, Pe. nigromaculatus readily catches and swallows smaller prey (Video S2; Honma, Oku & Nishida, 2006). Pelophylax nigromaculatus is a generalist predator that has been reported to prey on carabid beetles (Maeda & Matsui, 1999; Hirai & Matsui, 1999; Sano & Shinohara, 2012; Sarashina, Yoshihisa & Yoshida, 2011). As Ph. jessoensis and Pe. nigromaculatus co-occur in the same grassland habitats, this frog species is a potential predator of adult Ph. jessoensis. In early June 2016, I offered an adult Ph. jessoensis to an adult Pe. nigromaculatus under laboratory conditions. The frog attacked the beetle, but stopped the attack immediately after its tongue touched the beetle. No bombing sounds were heard, suggesting that the frog ceased its attack before the beetle bombed. Therefore, I hypothesised that bombing is not essential when Ph. jessoensis seeks to avoid being swallowed by Pe. nigromaculatus. To test this hypothesis, I observed Pe. nigromaculatus attacking Ph. jessoensis under laboratory conditions using a digital video camera. Acceptance or rejection of prey was carefully investigated using slow-motion videos. Furthermore, both dead and live beetles were used to test whether bombing is essential for successful defence against predatory attacks by Pe. nigromaculatus. Finally, I discuss the importance of primary and secondary defences in terms of overall anti-predation defence.

Materials and Methods

Results

In the experiment using live adult bombardier beetles (n = 28), 26 frogs (92.9%) rejected the beetles without swallowing them (Fig. 1); 19 frogs (67.9%) stopped attacking the beetles immediately after touching the beetles with their tongues (Fig. 2; Video S3), and seven frogs (25.0%) spat out the beetles after taking the beetles into their mouths (Fig. 3; Video S4). No beetle bombed a frog before being taken into the frog’s mouth. All beetles that were taken into frog mouths bombed (Video S4). Only two frogs (7.1%) were observed to swallow the bombardier beetles (Table 1) after being bombed in the mouth; one of the frogs successfully digested the beetle, but the other frog vomited the beetle 18 min after swallowing it (Table 1). The vomited beetle was still alive. Of the frogs that took the beetles into their mouths, 88.9% (n = 8/9) initially stopped attacking the beetles when their tongues first touched the beetles, but resumed their predatory attack soon thereafter (Fig. 3; Video S4). Frogs that did not swallow beetles consumed other prey (i.e., T. molitor larvae) soon thereafter.

When dead beetles were used (n = 28), 24 frogs (85.7%) rejected the dead beetles without swallowing them (Fig. 1); 20 frogs (71.4%) stopped attacking the beetles after their tongues touched the dead beetles (Video S5), and four frogs (14.3%) spat out the beetles after taking the beetles into their mouths (Fig. 1). Only four frogs (14.3%) swallowed the dead beetles. Similar to the experiment using live beetles, 87.5% of the frogs that took beetles into their mouths (n = 7/8) were initially deterred when their tongues first touched the beetles, but continued with their predatory behaviour soon afterwards. The frogs that did not swallow beetles ate other prey (i.e., T. molitor larvae) soon thereafter.

The proportion of dead beetles swallowed by frogs (14.3%) was higher than that of live beetles (7.1%). However, the GLM results indicated that the frog swallowing rates of live and dead beetles did not significantly differ (Table 1). Whether beetles were swallowed or not was associated with beetle size, but not frog size (Table 1). Beetles were more likely to be swallowed as beetle size decreased (Fig. 4A). The interaction of frog and beetle weight was not significant (Fig. 4A).

The proportion of dead beetles digested by frogs (14.3%) was higher than the proportion of live beetles swallowed (3.6%). The GLMs indicated that the difference between the digestion rates of live and dead beetles was significant (Table 2). Beetle size affected the digestion rate (Table 2, Fig. 4B).

Disussion

Here, I found that Pe. nigromaculatus frequently rejected Ph. jessoensis without attempting to swallow the beetles (Fig. 1). Around 70% of frogs stopped attacking both live and dead beetles immediately after touching the beetles with their tongues. Although the high-speed release of hot noxious chemicals (bombing) protected Ph. jessoensis from digestion by the frog Pe. nigromaculatus (Figs. 1 and 3; Video S4), my findings support the hypothesis that bombing is not essential for Ph. jessoensis to successfully evade being swallowed by Pe. nigromaculatus (Figs. 1 and 3; Video S3). Which factors, then, stopped the frogs from attacking? Three potential reasons can be considered: (1) the frogs recognised the warning colouration of the beetles; (2) the body size of the beetles was too large for the frog to accommodate; and (3) the frogs reflexively avoided the beetles after detecting toxic substances or other deterrent characteristics on the beetles’ body surfaces.

The bombardier beetle Ph. jessoensis does have a striking yellow and black pattern on its body that could serve as warning colouration (Fig. 1), although this has not been empirically demonstrated. Anuran predators can avoid toxic prey by recognising certain colours or other morphological characteristics and then ignoring those prey (Brower, Brower & Westcott, 1960; Brower & Brower, 1962; Dean, 1980a; Taniguchi et al., 2005; Ito, Taniguchi & Billen, 2016). In fact, 34.1 and 29.2% of frogs did not seek to attack live or dead beetles, respectively, before the feeding experiments commenced, suggesting that Pe. nigromaculatus may recognise the body pattern and shape of Ph. jessoensis and interpret these as warning signals. However, frogs that ignored Ph. jessoensis did not respond to other prey (i.e., T. molitor larvae). Therefore, the experimental conditions used may not be appropriate for analysing foraging by certain frogs. Alternatively, the yellow-and-black pattern of Ph. jessoensis may serve as disruptive camouflage; flightless Ph. jessoensis walk on soil of forest edges, grasslands, and agricultural fields. Further work is needed to explore the significance of Ph. jessoensis colour as a defensive trait.

GLM analysis indicated that beetle size was correlated with beetle-swallowing frequency of frogs (Table 1). Pelophylax nigromaculatus has been reported to spit out large prey that they were unable to swallow after taking the prey into their mouths (Honma, 2004; Honma, Oku & Nishida, 2006). However, 67.9% of the frogs in the experiment with live beetles and 71.4% of the frogs in the experiment with dead beetles stopped their predatory attacks before taking the beetles into their mouths (Figs. 1 and 2; Video S3, Video S5). Thus, my results do not provide strong evidence that the frogs could not ingest large beetles. Rather, the amount of chemicals on the beetle body surface may increase with beetle size (see below).

The rapid responses of the frog species Pe. nigromaculatus to bombardier beetles (Fig. 2) could be considered a reflex action of the frogs’ tongues (cf. Kumai, 1981a; Kumai, 1981b; Hirakawa, 1989). Frogs are known to use their tongues as a chemical detector (Dean, 1980b; Kumai, 1981a; Kumai, 1981b; Barlow, 1998) as well as a prey-catching tool (Noel et al., 2017). For example, chemical or electrical stimulation of the tongue can generate reflex responses in Pe. nigromaculatus (Kusano & Sato, 1957; Kumai, 1981a; Kumai, 1981b; Suzuki & Nomura, 1985; Takeuchi, Satou & Ueda, 1986; Hirakawa, 1989). Because Pe. nigromaculatus is a generalist predator that can attack a variety of arthropods within its field of view (Hirai & Matsui, 1999; Honma, 2004; Honma, Oku & Nishida, 2006; Sano & Shinohara, 2012; Sarashina, Yoshihisa & Yoshida, 2011), Pe. nigromaculatus may have evolved specific responses to toxic prey to avoid being injured by trying to eat them. The results of this study suggest that the tongues of Pe. nigromaculatus may be able to rapidly detect toxic substances or other characteristics on the body surface of the bombardier beetles, and the frogs subsequently avoid the beetles to prevent themselves from being bombed and injured. Previous studies have focused on how frogs and toads use their tongues to catch prey (Ewert, 1970; Nishikawa & Gans, 1996; Monroy & Nishikawa, 2010; Noel et al., 2017). Few reports have explored how frogs and toads use their tongues to detect toxins in/on potential prey (but see Dean, 1980b). Therefore, the tongue responses that I describe in Pe. nigromaculatus will likely be evident in other frogs such as the tree frog Hyla japonica (Günther) (Taniguchi et al., 2005; Ito, Taniguchi & Billen, 2016; Matsubara & Sugiura, 2017).

Conclusions

In one study, the chemicals produced by bombardier beetles’ bombing did not stimulate the tongues of toads any less intensely than did the heat from the chemical reaction (Dean, 1980b). Other than this study, the relative importance of the toxic chemicals and heat produced by bombing for the successful escape of bombardier beetles from predators has been largely unexplored. My results support the hypothesis that bombing is not essential when bombardier beetles defend themselves against frog attacks. Furthermore, my findings suggest that (cool) toxic chemicals on the beetles’ bodies alone may cause frogs to desist from an attack; thus, chemicals on the body may serve as a primary defence and bombing as a secondary defence. Successful defence by chemicals on the body would reduce the costs associated with bombing, suggesting that beetles may have evolved to use chemicals on the body surface as their primary defence. However, further experiments are required to validate this hypothesis; for example, dead beetles with body surfaces cleaned of chemicals, or palatable prey coated with toxic chemicals, should be offered to frogs.

Many prey animals exhibit multiple anti-predator defences (Edmunds, 1974). Predation pressures imposed by different enemies may encourage prey to diversify defence strategies. Further studies are needed.

Supplemental Information