Animals show various defensive reactions against their predators. One such reaction, autotomy, is a behavior in which animals cut off a part of their body to defend themselves. The most prevalent benefit of autotomy is predation avoidance. For example, an autotomized appendage can distract the predator through movement or may itself be consumed by the predator (reviewed in Maginnis 2006, Fleming et al. 2007). Moreover, if an appendage is wounded, an animal can choose to autotomize it rather than carry around a useless limb that may be an encumbrance or even attract the attention of a predator (Shimizu and Masaki 1993, Johnson and Jakob 1999, McCarthy and Dickey 2002, Emberts et al. 2017). Therefore, autotomy is observed in many animal taxa (Fleming et al. 2007).

Despite its immediate benefits, autotomy has costs such as reduced locomotion, impaired feeding behavior, and decreased mating success, particularly in cases where the lost limbs fail to regenerate completely (Lakes and Mücke 1989, Bateman and Fleming 2008). For instance, in the cricket Gryllus bimaculatus, loss of a limb by autotomy results in a decrease in future escape speed and mating success (Bateman and Fleming 2005, 2006a). In the cricket Acheta domestica, the autotomy of a hind limb similarly results in a significant reduction in escape speed and jumping ability and therefore makes the animals more susceptible to predation (Bateman and Fleming 2006b). Furthermore, in D. f. nigrofasciatus, individuals that experienced autotomy at a nymphal stage shortened their development time to an imago by reducing body size (Matsuoka and Ishihara 2010).

In Orthoptera, each limb plays an important role in mating posture (Loher and Rence 1978). In some ground crickets, particularly members of the tribe Pteronemobiini within the subfamily Nemobiinae, males possess a specialized hind tibial spur. During mating, females chew on this spur to ingest nutritious haemolymph (e.g., Fedorka et al. 2004, Hall et al. 2008). Therefore, hind limb autotomy in males may impose a mating cost by reducing their attractiveness or suitability as mates. The forelimbs function as ears, and female crickets rely on the male’s calling song to locate and approach him for mating (e.g., Otte 1992, Brown et al. 1996). These examples suggest that the costs of autotomy in ground crickets may depend strongly on the functional importance of the lost limb.

In the majority of Orthoptera, hind limbs are most likely to be grasped by a predator due to them being the largest and most obvious (Bateman and Fleming 2008). Therefore, it is assumed that the hind limbs of crickets are autotomized most easily among the limbs, and that the middle limbs are autotomized more easily than the forelimbs (Bateman and Fleming 2005, 2006c, 2008). In G. bimaculatus, autotomy of the forelimbs, which bear the hearing organs, takes more time than that of hind limbs (Bateman and Fleming 2006c). In addition, for individuals that experienced autotomy, autotomy of the remaining limbs might take more time because additional autotomy leads to greater costs (Bateman and Fleming 2005). From these observations, Bateman and Fleming (2005, 2006c) predicted that autotomy of the hind limbs would be easiest and that autotomy of the forelimbs would be most difficult.

In the present study, we examined whether a small cricket species, such as D. f. nigrofasciatus (Matsumura, 1904), is more likely to autotomize even its forelimbs because smaller species may rely more heavily on rapid autotomy when other defensive responses, such as biting, are limited. As an indicator of body size, head width is approximately 1.4–1.8 mm in D. f. nigrofasciatus (Matsuoka and Ishihara 2010), compared with 4.7–6.5 mm in G. bimaculatus (Gouws et al. 2011). Therefore, we attempted to measure the time adult D. f. nigrofasciatus took to autotomize a forelimb, a middle limb, or a hind limb. Because forelimbs are functionally important and potentially more costly to lose—especially in females—we predicted, following the reasoning of Bateman and Fleming (2005, 2006c), that they would take longer to autotomize than the other limbs. To further explore the potential costs associated with forelimb autotomy, we also compared the mating success, egg number, and hatching success of intact females, females with a single forelimb autotomized, and females with both forelimbs autotomized.

The band-legged ground cricket Dianemobius fascipes nigrofasciatus was used in this study. We follow Benediktov and Storozhenko (2018) in treating this taxon as a subspecies of Dianemobius fascipes (Walker, 1869).

In the present study, the F4 generation, derived from 12 adult D. f. nigrofasciatus females collected in Sakai (Osaka, Japan) in July 2006, was used (Matsuoka and Ishihara 2010). The stock culture was established from these field-collected females, each of which was kept individually and allowed to oviposit. Eggs were maintained at 23°C under a 16L:8D photoperiod. Ten hatchlings from each collected female were reared together until the third molt and individually from the fourth instar to adult emergence at 25°C under 16L:8D. Newly emerged adults were paired while avoiding sib-mating, and successive generations were maintained in the same way (Matsuoka and Ishihara 2010, Matsuoka et al. 2011). For the present experiment, all individuals were reared under ambient temperature and photoperiod conditions. Each cricket was housed individually in a plastic container (80 mm in diameter and 140 mm in depth) with a sheet of filter paper on the bottom. They were fed a pelleted insect diet (Oriental Yeast, Tokyo) and pieces of carrot, with food replaced every two days. A moist cotton plug was placed in each container to provide water.

A total of 167 virgin adult individuals (79 males, 88 females) were selected from the F4 generation. For each individual, either the forelimb, middle limb, or hind limb on the right side was continuously grasped at the femur with tweezers, and we attempted to measure the time until autotomy using a stopwatch. However, autotomy occurred so rapidly that precise latency could not be recorded reliably for individual trials. We therefore treated autotomy time qualitatively and recorded whether autotomy occurred immediately (i.e., within approximately 1 s under direct observation). This manipulation is known to readily induce autotomy (Fleming and Bateman 2007, Matsuoka and Ishihara 2010, Matsuoka et al. 2011). The number of individuals that autotomized each limb type was as follows: forelimbs – 20 males and 31 females; middle limbs – 20 males and 18 females; hind limbs – 39 males and 39 females. Each individual was subjected to only one autotomy trial for one limb type, except for 14 females in which the time to autotomize the second forelimb was also measured after the first was autotomized.

All autotomy tests were conducted during the daytime. For the procedure, each cricket was placed in a cylindrical container (70.5 mm in height and 34 mm in inner diameter), and the designated limb was grasped with tweezers inside the container.

Next, we investigated the effect of loss of forelimbs on mating success. Female crickets were randomly divided into three groups. Insects in the first group (n = 15) were kept intact as controls. The second group (n = 17) had the right forelimb autotomized. The third group (n = 14) had both forelimbs autotomized. Each female was paired with an intact male, and each pair was kept in a separate case (78 mm in diameter, 35 mm in depth) placed in an outdoor shelter under natural conditions of day length and average daily temperature of 27°C (in Osaka Prefecture from August to September 2006). In each case, insect diet (Oriental Yeast, Japan) and pieces of carrot were provided as food and replaced every few days. In addition, a moist cotton plug was provided and replaced every day, which served as a water source and oviposition substrate.

Each pair was monitored daily for the presence and number of eggs and subsequent hatching. The monitoring period lasted for 21 days starting from the day after pairing, which is sufficient to detect oviposition (Matsuoka and Ishihara 2010). Eggs were laid on moist cotton, which was replaced daily after checking for the presence of eggs. Given that eggs were typically laid inside the cotton and that food was constantly available, egg cannibalism was considered unlikely in this study. Eggs were collected and maintained separately for each pair in a plastic container (78 mm in diameter and 35 mm in depth) under natural outdoor conditions.

The effects of autotomy treatment were analyzed in terms of three reproductive outcomes: mating success, number of eggs laid, and hatching ratio. Mating success was defined by whether any eggs hatched, with unfertilized eggs not considered evidence of successful mating. The number of eggs laid was compared among treatment groups, excluding females that did not lay any eggs. The hatching ratio, calculated as the proportion of eggs that hatched out of the total number of eggs laid, was compared among treatment groups using only females that produced at least one hatched egg. These variables were statistically analyzed using generalized linear models. Mating success was analyzed with a binomial error distribution, egg number with a negative binomial error distribution to account for overdispersion, and hatching ratio with a binomial error distribution, followed by Tukey’s post-hoc tests. All analyses were conducted in R (version 4.5.2), treating the number of autotomized forelimbs (0, 1, or 2) as a categorical variable with three levels.

All limbs, including fore, middle, and hind limbs, were autotomized almost instantaneously under direct observation in all individuals (n = 167), regardless of limb position. This was also true for the second forelimb in females that had already autotomized one (n = 14). Because autotomy occurred too rapidly for reliable stopwatch-based quantification, no precise latency values could be recorded, and no formal statistical comparison among limb types was conducted.

Autotomy treatment affected post-mating reproductive outcomes in several ways. Regarding hatching success, no hatching was observed in 3 of 15 pairs in the intact group, 5 of 17 pairs in the single forelimb autotomy group, and 10 of 14 pairs in the double forelimb autotomy group. Females in the double forelimb autotomy group had significantly lower hatching success than those in the intact group (z = 2.63, p = 0.023; Fig. 1). Hatching success did not differ significantly between the intact group and the single forelimb autotomy group (z = -0.61, p = 0.814). Females in the double forelimb autotomy group also had lower hatching success than those in the single forelimb autotomy group (Fig. 1), although this difference was not statistically significant (z = 2.25, p = 0.063).

Autotomy treatment did not significantly affect the number of eggs laid when analyzed using a negative binomial generalized linear model (likelihood ratio test: χ² = 2.71, df = 2, p = 0.258). Nevertheless, the estimated mean number of eggs laid was highest in the single forelimb autotomy group, intermediate in the intact group, and lowest in the double forelimb autotomy group (Fig. 2).

Finally, hatching ratio was reanalyzed using only females that produced at least one hatched egg. The analysis revealed a significant effect of autotomy treatment on hatching ratio (likelihood ratio test: χ² = 17.83, df = 2, p < 0.001). The double forelimb autotomy group showed a significantly lower hatching ratio than both the intact group (z = -3.96, p = 0.0002; Fig. 3) and the single forelimb autotomy group (z = 2.78, p = 0.0149). The difference between the intact group and the single forelimb autotomy group was not statistically significant (z = -2.30, p = 0.0560).

It has been predicted that forelimbs would take longer to autotomize than middle or hind limbs in crickets (Dixon 1989, Bateman and Fleming 2006a). This is because the forelimbs of ensiferan species bear tympanal hearing organs, and autotomized limbs, irrespective of their positions, are not regenerated (Matsuoka’s pers. obs. for D. f. nigrofasciatus, Bateman and Fleming 2008 for G. bimaculatus). In female crickets, the forelimbs are important for detecting male calling song (e.g., Otte 1992, Brown et al. 1996). In addition, during mating, the female’s forelimbs and middle limbs are used to position her body on the male’s dorsum, while the hind limbs remain on the ground to steady her posture (Loher and Rence 1978). Thus, forelimbs may be especially important for females in mating-related contexts. Therefore, if a single forelimb of a female is autotomized, the remaining forelimb is predicted to take more time to be autotomized than the first one because of its increasing importance in successful mating. However, our observations did not reveal any detectable delay in autotomy among limb types or between the first and second forelimbs in D. f. nigrofasciatus.

In the field population of D. f. nigrofasciatus, it was observed that most individuals that experienced autotomy lost their hind limbs rather than their forelimbs or middle limbs irrespective of sex (Matsuoka pers. obs.). This observation seems common to the majority of Orthoptera (Bateman and Fleming 2008). However, in the present experiment, fore- and middle limbs appeared to be autotomized as readily as hind limbs under direct observation. This may be due to the lack of effective protection against predators other than autotomy in smaller species like D. f. nigrofasciatus, compared with larger species such as G. bimaculatus. In fact, field cricket species—which are much larger than ground crickets—have been reported to show less readiness for autotomy of a limb (Bateman and Fleming 2008), and defensive methods such as kicking and biting have been observed in larger species (Bateman and Toms 1998, Field and Glasgow 2001, Wolf et al. 2006). This suggests that smaller species may rely more heavily on autotomy than larger species. It might be more effective for larger species to defend themselves against predators without autotomy. For instance, in porcelain crabs, Petrolisthes spp. (Porcellanidae), increased body size is associated with less reliance on autotomy of a single cheliped (Wasson and Lyon 2005). In the starfish Asterias rubens, large individuals take a longer time to autotomize an arm (Marrs et al. 2000). In small species, speedy loss of the appendage through autotomy may be a key requirement for successful escape because small species have less resistance to predators than larger species (Bateman and Fleming 2008). Therefore, small cricket species such as D. f. nigrofasciatus may readily autotomize limbs irrespective of limb position. These experimental results do not contradict the field observation that most D. f. nigrofasciatus that experience autotomy lose their hind limbs because hind limbs are the largest, and possibly the easiest to grasp, and may also be closest to the predator if the cricket is fleeing. More generally, although the effects of hind limb autotomy have been studied in many cricket species, the effects of forelimb or middle limb autotomy remain poorly understood, and further comparative work across species is needed.

The females with intact forelimbs and those with both forelimbs autotomized showed the highest and the lowest mating successes, respectively (Fig. 1). One important factor is the deterioration of mating posture due to forelimb loss. In Orthoptera, each limb plays an important role in mating posture (Loher and Rence 1978). Forelimbs also help maintain balance, and losing both may severely hinder the female’s ability to engage in mating behavior effectively. The combined effect of losing these functions may be significant in smaller enclosures, such as those used in our experiments, where the spatial constraints amplify the difficulty of maintaining mating posture. Thus, in our controlled environment, difficulty in maintaining mating posture due to forelimb loss may have played a key role in the reduced mating success.

In crickets, the forelimbs also serve as ears and are important for detecting male calling song (Otte 1992, Brown et al. 1996). However, because directional hearing depends on interaural differences (Hill and Boyan 1976), females with only one forelimb remaining may still have impaired ability to localize a calling male. While hearing is important in mate location under natural conditions, its role in the present experimental setting—where males and females were kept in close proximity—may have been limited. Therefore, the observed reduction in mating success following forelimb loss is more likely attributable to impaired balance or physical instability than to auditory deficiency alone.

Females with a single autotomized forelimb tended to lay more eggs than intact females, whereas females with both forelimbs autotomized tended to lay fewer eggs (Fig. 2). However, autotomy treatment did not have a statistically significant effect on egg number; therefore any interpretation of treatment effects on egg production should be made cautiously. It remains possible that females with a single autotomized forelimb increased egg production as a compensatory reproductive response, as suggested for Gryllus texensis under simulated predation risk (Adamo and McKee 2017), but this interpretation remains speculative in the present study. Likewise, females that autotomized both forelimbs may have experienced difficulty in assuming the mating posture and in successful spermatophore transfer, but this possibility also requires further study.

When hatching ratio was reanalyzed using only females that produced at least one hatched egg, the hatching ratio remained highest in the intact group, followed by the single forelimb autotomy group and the double forelimb autotomy group (Fig. 3). Although the difference between the intact group and the single forelimb autotomy group was not statistically significant, the double forelimb autotomy group showed a significantly lower hatching ratio than both the intact and single forelimb autotomy groups. This suggests that bilateral forelimb loss affects not only the likelihood of successful mating but also reproductive performance among females that produced at least some viable eggs.

This study showed that forelimbs, as well as middle and hind limbs, were autotomized almost instantaneously in both males and females under our experimental conditions. Direct observation also suggested that the second forelimb in females was autotomized just as rapidly as the first. We also elucidated the impact of forelimb autotomy on female reproductive performance through mating success, number of eggs laid, and hatching rates. These results suggest that even forelimbs, which are functionally important in mating and sensory perception, are not strongly retained in D. f. nigrofasciatus, possibly due to the species’ small body size and reliance on rapid autotomy for predator escape. Our results also suggest that forelimb autotomy can affect female reproductive performance, probably by impairing the ability to perceive male courtship calls and to maintain proper mating posture. This suggests that the presence or absence of forelimbs plays a crucial role in the reproductive strategy of crickets. Further observational and experimental studies under natural conditions are needed to verify these findings. We believe that this study contributes to a deeper understanding of the adaptive behaviors and reproductive strategies of crickets.