Introduction

Death feigning is defined as immobilization in response to external stimuli (e.g., Miyatake et al. 2004; Ruxton et al. 2018). The adaptive significance of death feigning in insects has been much studied, and the behavior is thought to have evolved primarily as a predator avoidance strategy (e.g., Humphreys and Ruxton 2018; Skelhorn 2018). For example, in the red flour beetle, Tribolium castaneum, it has been reported that the frequency of predation by Adanson’s jumper spider, Hasarius adansoni, was lower for individuals of the strain with a longer death-feigning duration than for individuals of the strain with a shorter death-feigning duration (Miyatake et al. 2004, 2009). This behavior of temporary immobility without postural change is called freezing, while temporary immobility with postural change is called death feigning or thanatosis (Sakai 2021).

Postures of death feigning are sometimes unique and vary not only among species (Miyatake 2021), but also between local populations (Ohba and Matsuda 2021). Furthermore, even within a species, the posture varies according to developmental stage; for example, the immobile posture of the red flour beetle differs between adults and larvae, with larvae in a straight position, while adults turn over on their backs and fold their legs (Miyatake et al. 2008; Matsumura et al. 2017). However, to our knowledge, no other study has reported on variations of death-feigning postures during larval instars.

Death feigning has been found in 10 of the 31 orders of insect taxa (Miyatake 2021): Odonata (Gyssels and Stoks 2005), Plecoptera (Moore and Williams 1990), Orthoptera (Nishino and Sakai 1996; Honma et al. 2006), Phasmatodea (Godden 1972), Neuroptera (Sendova-Franks et al. 2020), Mantodea (Edmunds 1972; Lawrence 1992), Hemiptera (Holmes 1906; Villet 1999; Kang et al. 2016; Ohba and Matsuda 2021), Coleoptera (Bleich 1928; Konishi et al. 2020), Lepidoptera (Tojo et al. 1985; Dudley 1989; Larsen 1991) and Hymenoptera (Hölldobler and Wilson 1990; van Veen et al. 1999; King and Leaich 2006; Cassill et al. 2008; Cardoso et al. 2024).

We found that when larvae of two species of Mecoptera insects are stimulated, they adopted death feigning in different postures depending on their larval instar. This paper is the first report of death feigning in Mecoptera. We also observed the frequency and duration of death feigning in the two species of Mecoptera in the laboratory. We identified two types of death-feigning posture in the scorpionfly larvae: one where they stayed straight, and the other where they curled up into a ball shape, and also found that the frequency of these postures varied with larval stages.

Material and Methods

Insects

Two scorpionfly species (Mecoptera: Panorpidae), Panorpa japonica and P. pryeri were used in these experiments (Fig. 1). These two species have a life cycle during larval development in which they spend the day underground and are active on the surface at night. Predators of the larvae are unknown. The larvae pupate in the soil and then emerge above ground in early spring.

Fig. 1
figure 1

Adults of two Mecoptera species used in this study: A: Panorpa japonica and B: Panorpa pryeri

Full size image

In the present study, 54 larvae were derived from adults collected in April 2020 at a bank of the Asahikawa River, Okayama City (34° 6′ 8′′ N, 133° 9′ 3′′ E) for P. japonica. For P. pryeri, larvae were derived from adults collected in June 2022 at Koguri-yama Mountain, Hirosaki City, Aomori Prefecture (40° 32′ 44′′ N, 140° 28′ 39′′ E). Fifty-five larvae of P. pryeri were used in the experiment, each larva was reared in a Petri dish (9 cm diameter × 2 cm high) lined with a moderately moistened nonfat cotton. The food was commercially available dried sewage worms.

Observation of Death Feigning

Individual scorpionfly larvae were placed in a Petri dish (9 cm diameter, 1.5 cm height) and lined with filter paper and stimulated with a bristle handle end brush (Pentel, ZBS 1–15, Tokyo). If the larvae did not move after 5 s after stimulation while maintaining the "straight" or "ball" posture described below, they were considered to have entered death feigning; those that moved within 5 s were not considered to be in death feigning. The duration of death feigning from the start to the end (when the larvae started moving again) was measured with a stopwatch, and if the larvae did not become inert after 10 or more stimulations, it was recorded as not feigning death. Both species exhibited two types of postures while feigning death, named “straight” and “ball” (Fig. 2).

Fig. 2
figure 2

Death-feigning postures, straight (A) and ball (B). Examples of P. japonica are shown

Full size image

Measurements of death feigning were taken at five larval instar stages: 1st, 2nd, 3rd, early 4th, and late 4th instars when the larva had matured to a white body color. Death-feigning posture and duration of death feigning were compared for each larva. In this experiment, death feigning was observed in 39 P. japonica and 44 P. pryeri, and five replicate measurements were taken for all individuals.

Statistics

A generalized linear mixed model was used to test the respective proportions of individuals that immobilized in a straight posture, those who immobilized in a balled posture, and those who did not immobilize. In this analysis, we used an item response tree (IRTree) model (López-Sepulcre et al. 2015). This method argues that instead of being analyzed using traditional nonparametric statistics or a series of separate analyses split by response categories, this kind of data can be more holistically analyzed using a generalized linear mixed model (GLMM) framework extended to binomial response trees (Lopez-Sepulcre et al. 2015). Whether the larvae of the scorpionfly feigned death or not was considered as "node 1," and the posture of death feigning (straight or ball) was considered as "node 2." The presence or absence of each event was analyzed as binary data with a binomial distribution (López-Sepulcre et al. 2015). In this analysis, species, experiment, larval instar, and posture were used as a fixed effect, and larval ID and observation were used as a random effect.

In the analysis of the duration of death feigning, the GLMM with a gamma distribution was used. Other statistical models in this analysis were similar to those for death-feigning frequency. If a significant effect was found in each analysis, we conducted multiple comparison tests with Bonferroni correction.

All analyses were conducted in R ver.4.1.0 (R Core Team 2017) and using the statistical packages lme4 (Bates et al. 2015) and car (Fox and Weisberg 2019).

Results

Table 1 shows results of IRTree GLMM for test of effects of species, experiment, and larval instar on frequency of death feigning. No significant effect of larval instar was observed in the test results of node 1 (i.e., whether they feigned death: the percentage of white bars vs. colored bars in Fig. 3). The estimate of developmental stage was negative (-0.29), indicating a trend that the number of individuals, which did not feign death, increased (Fig. 3), but the effect was not statistically significant (p = 0.09). Moreover, the effect of species was not significant. In the results for Node 2 (i.e., the death-feigning pose: the percentage of black bars vs gray bars in Fig. 3), the negative estimate of larval stage (-1.00) indicates that the proportion of "straight" death feigning individuals increased with development, and the effect was statistically significant (p = 0.02). The effect of the species-larval stage interaction was also statistically significant (p < 0.05), with an increasing proportion of P. pryeri individuals with "ball" death feigning with their development (Fig. 3B).

Fig. 3
figure 3

Proportions of each pose of death-feigning in P. japonica (A) and P. pryeri (B), respectively. Black and grey bars show ball and straight poses death feigning. White bar shows no response to stimulus

Full size image
Table 1 IRTree GLMM for test of effects of species, experiment, and larval instar on frequency of death feigning. Results with p-value ≤ 0.05 are given in bold
Full size table

Figure 4 shows the results of duration of death feigning. Comparing the duration of death feigning between the two species regardless of their age revealed that P. pryeri larvae showed significantly longer duration than P. japonica larvae (Fig. 4A, Table 2). Moreover, death feigning in the form of a ball had a significantly longer duration than in a straight position, regardless of species (Fig. 4B, Table 2). In these results, the effects of the interaction between species and posture were not significant (Table 2). Figure 4 C shows the death-feigning duration of larvae in each instar of the two species. The results of statistical analysis showed a significant interaction between species and age with duration of death feigning (Fig. 4C, Table 2). In both species, the duration of death feigning increased with age up to the third instar. In older P. pryeri larvae the duration of death feigning continued to increase with age, while it decreased in P. japonica larvae showed the duration of death feigning decreased at the fourth instar (Fig. 4C, Table 2).

Fig. 4
figure 4

Duration of death feigning in two scorpionfly species. A and B show death-feigning duration between each species and posture, respectively. C shows death-feigning duration at each instar larva. In C, black and grey show P. japonica and P. pryeri, respectively. There are significantly differences among different alphabets (Bonferroni correction). In early and late 4th instars, duration of death feignings are longer in P. pryeri than P. japonica. Error bars show standard error

Full size image
Table 2 Results of GLMM for test of effects of species, experiment, and larval instar on death-feigning duration (only the experiment 1)
Full size table

Discussion

We found that the posture and duration of death feigning change as the larval instar progressed in the two Mecoptera species examined. The relationship between age and death feigning in the adult stage has been studied in some insect species. For example, in the fire ant Solenopsis invicta, where interspecific colony struggles are intense, workers, which have softer exoskeletons, choose death feigning in the first few days of life when they are more vulnerable to attack, while older workers are more likely to choose counterattack (Cassill et al. 2008). The braconid parasitoid wasp Heterospilus prosopidis also has a long death-feigning duration in young adults, but the duration shortens with age (Amemiya and Sasakawa 2021). These studies may be examples of changes in predator avoidance behavior as the body hardens. In adult sweet potato weevils Cylas formicarius, adult virgin females have longer death-feigning periods when they are young, and after that the duration decreases with age (Kuriwada et al. 2011). In final instar (6th instar) larvae of the common cutworm Spodoptera litura, the solitary phase showed longer duration of death feigning compared to the gregarious phase (Tojo et al. 1985).

In insects, there are many examples in which predators control behavioral differences depending on the larval instar. For example, larvae of the Colorado potato beetle Leptinotarsa decemlineata show body writhing and falling in response to predator contact more frequently in the 3rd and 4th than in the 2nd instar (Ramirez et al. 2010). On the other hand, in the grasshopper Ageneotettix deorum, 3rd and 4th instar nymphs showed antipredator behavior in response to predator contact, while 5th instar larvae do not (Danner and Joern 2003). Ramirez et al. (2010) suggested that the frequency with which antipredator behavior is adopted, might be controlled by larval predation risk. However, no study has been conducted of alternations of death feigning as instars progress, and this is the first study to quantitatively show that death-feigning posture changes during larval instars. Differences in death-feigning posture among local populations have also been reported in the ferocious water bug Appasus japonicus (Ohba and Matsuda 2021).

Why do death-feigning postures change with age and among species? One possibility is predation avoidance in the species' habitat. Larvae of Panorpidae are found in the soil of mountainous areas (Byers and Thornhill 1983), and the larvae of P. japonica and P. pryeri are also found in soil. When predators such as mammals dig up the surface layer, larvae of this species are at increased risk of predation. In such cases, larger-sized larvae may be able to escape predators by curling up and rolling down the slope. One of the authors has observed that the larvae curl up and roll down on slopes when P. japonica larvae are stimulated during rearing, (RI, personal observation). On the other hand, small larvae are less conspicuous and may be less likely to be detected by predators by becoming immobile rather than curling up to escape. That is, death feigning in small body size larvae may function as hiding, while the posture of death feigning in ball in large body size larvae may function as escaping by rolling down a slope. The different death-feigning poses depending on the larval instars among species may be due to the different types of natural enemies among the two species.

In addition, repeated exposure to external stimuli may have influenced the death feigning. A previous study showed that the death feigning in the Colorado potato beetle L. decemlineata e subsequently changed depending on their experience of repeated external stimulations at a younger age (Acheampong and Mitchell 1997). In this study, stimulation was applied repeatedly from the 1st instar larva to the late 4th instar larva, so repeated experience of external stimulation at a younger age may have influenced the death feigning. In Panorpidae, there are few examples of studies on larval strategies against predators, including physical and chemical defenses, and to our knowledge, only camouflage by larval body color has been suggested (Jiang et al. 2019). This study indicates that Panorpidae larvae may use death feigning as an anti-predator strategy. Further work is, therefore, needed to confirm whether the death feigning of Panorpidae is anti-predator behavior, and to investigate the significance of changes in death-feigning posture during larval instars.