Introduction

Since Darwin, sexual selection has distinguished between inter- and intra-sexual selection (Darwin 1871; Anderson 1994). In intra-sexual selection, various behaviors of male combat using developed weapon traits (Emlen and Oring 1977; Shuster and Wade 2003; Emlen 1997), such as horns in deer and beetles, mandibles in stag beetles, forceps in earwigs, and claws in crabs and shrimps, have evolved (Emlen 2008). In cases where males of a certain species engage in contests over access to females or territories against conspecific males, it is predicted that the weapon traits used in such contests evolve, and individuals with higher competitive abilities are expected to gain disproportionate access to fertilization (Emlen 2008). For example, male Australian quacking frogs (Crinia georgiana) wrestle over territory ownership using longer forelimbs compared to females (Buzatto et al. 2015).

Previous experiments of male competition have revealed the winner/loser effect on subsequent contests, in which winners have a higher probability of winning, whereas losers become more likely to lose (Hsu et al. 2006). In addition, winner males mate more often than loser males (Zeng et al. 2018; Filice and Dukas 2019; Teseo et al. 2016; Kim et al. 2018; Kola et al. 2021; Toyoshima and Matsuo 2023). In the broad-horned flour beetle (Gnatocerus cornutus), the experience of losing a fight increases the number of sperm transferred to females, but there is no change in sperm number in a male owing to winning (Okada et al. 2010). In fruit flies (Drosophila melanogaster), loser males are less likely to mate successfully than winner males, but have been shown to produce more offspring (Filice and Dukas 2019). Also, the regulation of sperm quantity and quality by winning or losing may vary depending on the species’ ecology and mating system (Tuni et al. 2016). However, few studies have experimentally tested the impact of fight outcomes on male offspring fitness (but see Tuni et al. 2016; Filice and Dukas 2019).

We found that males of the giant mealworm Zophobas atratus (Coleoptera: Tenebrionidae), bite each other’s hind legs in combat (see Video). As far as we know, such a combat behavior has been reported for only one other animal species, the Namib tenebrionid Onyrnacris plana (Enders et al. 1998), in which sequences of combat behavior and sexual dimorphism were examined. The Tenebrionidae family inhabit dry and tropical regions and is highly variable in body size, with sizes ranging from one mm to almost 80 mm (Kergoat et al. 2014). However, why males bite their rival’s hind leg remain unclear, and nobody has examined how biting the legs of the rival male affects the fitness of the winner and loser of the combat.

In this study, we assessed sexual differences in body parts to test whether male morphology is adapted to this particular fighting behavior. Next, we examined the traits of winner and loser males in the leg-biting combat of this species through laboratory observations. We measured the number of eggs laid, and the number of hatches as indicators of their fitness, and measured the duration of mating. During combat, a male injuries the opponent male (Kappeler 1997; Eberhard 1998; Lailvaux et al. 2004; Rojas et al. 2012; Emberts et al. 2021). In the leg-biting combat, there is a possibility of injuring the hind legs. This species has a mating behavior in which the male mounts the female and inserts the copulatory apparatus into the female while supporting her body with his hind legs. We predicted that males with injured hind legs would have difficulty maintaining the proper mating posture, as they are unable to support his body during copulation, shortening the duration of mating, and thus reducing the number of eggs laid and hatched. Furthermore, the intensity of male combat generally increases when there is no size difference between the combatants (e.g., Miyatake 1993; Enders et al. 1998; Tedore and Johnsen 2015; Chen et al. 2020). Thus, we first examined whether the degree of hind leg damage varies with the intensity of combat, and secondly, we investigated whether an increase or decrease in male fitness was affected by a difference in body size between contesting males.

Materials and methods

Study species

A base stock of Z. atratus purchased from a breeder (Pro-Worm; Tokyo, Japan) was used in this study. They were reared cumulatively for one year at Okayama University. The larvae of Z. atratus were reared in a group (n = 500) and fed raw carrots and tropical fish food (Neopros; Kyorin, Himeji, Japan). Larvae were isolated to facilitate emerging because this species pupates only under isolation (Rumbos and Athanassiou 2021). Although emerging from isolation is possible at a body weight of 300 mg, in this experiment larvae were isolated at a body weight of 400 mg or more to increase the rate of successful eclosion (Quennedey et al. 1995). All larvae and adults were reared in a chamber kept in 16:8 L:D, temperature 25 ± 2 ℃, and humidity 60–70%.

Morphological measurements

The emerged adults (males: n = 38, females: n = 53) were reared in groups for one month under the same conditions as the larval stage. The adults were frozen (− 18 to − 24 ℃) before the mating bouts and measured. Measurements included antenna length (right and left), head width (fore, mid-, and hind), head length, mandible length and width (right and left), mandible teeth width (right and left), maxilla length (right and left), labial palp length (right and left), prothorax width (fore and mid-), prothorax length, elytra width, elytra length, and weight (Fig. 1). Each trait was photographed with a digital camera (DP23, Tokyo, Japan) attached to a stereomicroscope (SZX10, Olympus, Tokyo, Japan), and the length and width of each organ were measured (in 0.001 mm increments) with the built-in measurement software (DP23, Olympus; Tokyo, Japan).

Fig. 1
figure 1

Measurement sites for sexual dimorphism experiments. The white line indicates the part measured. AL: antenna length; FHW: forehead width; MHW: mid-head width; HHW: hindhead width; HL: head length; ML: mandible length; MW: mandible width; MTW: mandible teeth width; MxL: maxilla length; LPL: labial palp length; FPTW: foreprothorax width; MPTW: mid-prothorax width; HPTW: hind prothorax width; PTL: prothorax length; EW: elytra width; EL: elytra length. Initials L and R (e.g., LAL and RAL) indicate left and right. HL is from FHW middle point to HHW middle point. PTL is from FPTW middle point to prothorax end. EL is from EW middle point to elytra end

Full size image

Male-male combat

Adults were reared in isolation for 3–4 months after emerging (average adult life span: 6 months) under the same lighting, temperature, and humidity conditions as described above for the larval stage and were fed the same type of tropical fish food (about 0.05 g) and cut raw carrots (about 0.5 g) once every two days. These isolated individuals were used for the fighting experiment (contested males).

Fighting was observed in the same chamber as rearing, a plastic container (156 × 149 × 65 mm; As One, Osaka, Japan) with a piece of filter paper (Advantec, Tokyo, Japan) on the bottom (Fig. 2). A loser in this experiment was defined as a male who was bitten on a hind leg and ceased fighting and escaped during the experiment. The combat duration was measured as the time from the start of combat until the two individuals separated. Observation of combat sometimes continued for three hours. If the fight was not settled within the observation time, the data were excluded from the analysis.

Fig. 2
figure 2

Experimental design. Male-male combat: the males fought in two different pairs, and the duration of the struggle was measured. The pairs have the same body size (upper graphs) or a different body size (lower graphs). A mating experiment was conducted one day after the end of the male combat. Mating experiment: Each pair (winner and female, loser and female, uncontested and female) were observed mating, and the duration of mating was measured. The number of eggs and hatching of these eggs (over a 15-d period) was counted. After the mating experiments, females were moved to new containers and reared individually. Eggs laid were counted once every 3 days. The number of hatches was counted every day

Full size image

In this experiment, a difference of ≤ 0.2 mm in mid-prothorax width was defined as a fighting group (n = 38) with no difference in body size (same body-size pair), while a difference of ≥ 0.2.5 mm was defined as a group (n = 36) with a difference in body size (different body-size pair). Males with a pronotum width of 7.05–8.05 mm were used in fights with the same body-size opponent. The males paired in fights with different body sizes were those with pronotum widths of 6.95–10.0 mm. The males were the same age as the rival in the male combat. These winners and losers, and uncontested males that had not experienced fighting were used in the following mating experiment.

Mating experiment

Virgin females kept in isolation under the same conditions as males were used in the mating experiments. The male–female pairs used in the experiment were 104 virgin males (winner n = 36, loser n = 37, and uncontested n = 31) and 104 virgin females. Because extreme differences in female body size may alter the number of eggs laid regardless of the presence or absence of leg-biting behavior, we measured the pronotum width as an alternative indicator of female body size in the mating experiment, as in the fighting experiment, and only females with a pronotum width of 7–8 mm were used. The females used were the same age as each male (winner, loser and uncontested) in mating experiments.

Twenty-four hours after the male-male combat, the recent winners and losers and the uncontested males were allowed to mate with a virgin female in the plastic container as described above, and the mating behavior was observed for 90 min (Fig. 2). The mating behavior of this species is a behavior in which the male mounts the female and inserts the copulatory apparatus into the female while supporting her body with his hind legs. Therefore, the insertion of the copulatory apparatus by the male was observed visually, and the mating duration was measured.

The mated females were isolated in a sample cup (130 × 60 mm, Sanyo; Tokyo, Japan) using folded paper towels (Clinpet Japan; Saijo, Japan) as spawning beds and bedding material, and kept isolated in the same chamber for 15 d (Fig. 2). The females were fed the same tropical fish food and freshly cut carrots as in the case of isolation, and then the females were moved to a new cup (Fig. 2). Three days after transferring the females, the eggs in the new cup were counted to determine the number of eggs laid. Each day thereafter, the eggs that hatched were counted, and the hatched larvae were removed. This was continued until all eggs had hatched, and the total number of eggs hatched for each female over a 15-d period was recorded. If a female died during the measurement, the data were removed from the analysis.

Statistics

Multivariate analysis of variance (MANOVA) was used to examine the sex differences at each measurement site. The covariate for this statistic was mid-prothorax width. Duration of combat, duration of mating, number of eggs laid, and number of hatches were tested using a general linear mixed model (Table 1). The male age was added explanatory variable because the male age can influence on duration of combat, mating, number of eggs laid by females and number of hatches (Okada et al. 2020). Bonferroni post hoc tests were used to compare differences between treatments. The overdispersions in GLMM3 and GLMM4 were checked using check overdispersion () in the install: R packages ("performance"). As the result, no overdispersion was detected in our two GLMMs. All statistics were performed in R (ver. 4.1.1) with p < 0.05 indicating significant differences.

Table 1 Generalized linear mixed models (GLMM) used to compare duration of combat, mating, number of eggs, and hatches of the eggs
Full size table

Results

Sexual dimorphism

Males had significantly greater forehead width, mid-head width, head length, left mandible length, right mandible width, and mandible teeth width (right and left) than females (Table 2). On the other hand, females had a significantly longer left labial palp, and wider elytra than males (Table 2). There were no sexual differences in body weight, hind-head width, antenna length (right and left), left mandible width, right mandible length, maxilla length (right and left), right labial palp length, fore-prothorax width, prothorax length, or elytra length Table 2).

Table 2 Results of MANOVA for each Z. atratus trait (vs. female)
Full size table

Male-male combat

The duration of combat for different body-sized pairs was significantly shorter compared to same body-sized pairs (p = 0.044; Fig. 3). In this experiment, the loss of a hind leg (from tibia to tarsus) of two SL males among seventy-three males was confirmed, although we didn’t confirm male’s hind leg injury. Three pairs did not end their combat within three hours, and these data were not used in the statistics including comparison combating, mating duration, number of eggs laid, and hatching of these eggs. Loser males were smaller than winner males in all combats. The duration of combat between three- month-old males was not different compared to the duration of combat between four- month-old males (p = 0.944).

Fig. 3
figure 3

Comparison of duration of combat between same body-sized pairs (defined when the body size difference between males was ≤ 0.2 mm) and different body-sized pairs (defined when the body size difference between males was ≥ 0.25 mm). The symbol × in each graph shows the mean value

Full size image

Mating experiment

Males in the five groups, the winner and loser males in the combats between same body-size pairs (SW, SL) different body-size pairs (DW, DL), and uncontested males (U), did not differ in mating duration (Bonferroni, SW-U p = 1.000, SW-SL p = 1.000, SW-DW p = 1.000, SW-DL p = 1.000, SL-U p = 0.528, SL-DW p = 1.000, SL-DL p = 1.000, DW-U p = 0.952, DW-DL p = 1.000, and DL-U p = 0.068; Fig. 4). The mating duration (four months old) was significantly increased compared to the mating duration (three months old) (p < 0.001).

Fig. 4
figure 4

Comparison of mating duration among same body-size pairs (defined as the body-size difference between males is ≤ 0.2 mm) and different body-size pairs (defined as the body size difference between males is ≥ 0.25 mm). Symbol × in each graph shows the mean value. ‘Uncontested’ represents males who didn’t perform combat and mate with a female. Different letters indicate a significant difference at p < 0.05

Full size image

Females mated with SL males laid a significantly fewer eggs compared with females mated with uncontested males (p = 0.0098 Fig. 5). However, SW-U (p = 0.476), SW-SL (p = 1.000), SW-DW (p = 1.000), SW-DL (p = 1.000), SL-DW (p = 0.100), SL-DL (p = 0.061), DW-U (p = 1.000), DW-DL (p = 1.000), and DL-U (p = 1.000) showed no significant difference (Fig. 5). The number of laid eggs (four months old) was significantly increased compared to the number of laid eggs (three months old) (p < 0.001).

Fig. 5
figure 5

Comparison of the number of eggs among same body-size pairs (defined as the body-size difference between males of ≤ 0.2 mm) and different body-size pairs (defined as the body size difference between males of ≥ 0.25 mm). Symbol × in each graph shows the mean value. ‘Uncontested’ indicates a male who didn’t perform combat and mate with a female. Different letters indicate a significant difference at p < 0.05

Full size image

The number of eggs hatches sired by SL males was significantly lower compared to uncontested males (p = 0.021; Fig. 6), but SW-U (p = 0.785), SW-SL (p = 1.000), SW-DW (p = 1.000), SW-DL (p = 1.000), SL-DW (p = 0.194), SL-DL (p = 0.132), DW-U (p = 1.000), DW-DL (p = 1.000), and DL-U (p = 1.000) showed no significant difference (Fig. 6). The number of eggs hatches (four months old) was significantly increased compared to the number of laid eggs (three months old) (p < 0.001).

Fig. 6
figure 6

Comparison of number of egg hatches among same body-sized pairs (defined as the body size difference between males of ≤ 0.2 mm) and different body-sized pairs (defined as the body size difference between males of ≥ 0.25 mm). Symbol × in each graph shows the mean value. ‘Uncontested’ means a male who didn’t combat a male and mate with a female. Different letters indicate a significant difference at p < 0.05

Full size image

Discussion

In the mating experiments, contrary to the hypothesis, males with missing hind legs were still able to insert their own copulatory organs into females, and there was no significant difference in the duration of the insertion of copulatory organ among all groups (Fig. 4). However, the number of eggs was significantly decreased only in females mated with SL males compared to females mated with uncontested males (Fig. 5), and the rate of egg hatches was significantly lower in females mated with SL males than uncontested males (Fig. 6). This could be due to the following two factors: the first is a defect in the insertion of the mating organ due to a hind leg injury in escalated inter-male fighting males. The second factor could be attributed to cryptic female choice (Eberhard 1996).

Generally, when body sizes of males are the same, male-male combat escalates (e.g., Miyatake 1993; Enders et al. 1998; Tedore and Johnsen 2015; Chen et al. 2020). In this experiment, a significant increase in fight duration was also observed in same body-sized pairs, and we observed escalation in male combat (Fig. 3). Second, SL males often showed a mating posture different from that of uncontested males (Fig. 7). This suggests that although the male’s copulatory organ is inserted into the female, the sperm (or spermatophore) may not be delivered to the appropriate location within the female's reproductive organ. Future studies should count sperm inserted into females.

Fig. 7
figure 7

Mating posture of Z. atratus. The left is the mating position of an uncontested male. The right is a male who lost in combat in which the body size of the males was the same. In this figure, the upper insect is a mounting male, and the lower insect is a female

Full size image

In mating systems where females engage in multiple matings with different males, cryptic female choice can occur. Although multiple matings with different males by females of Z. atratus have not been observed, it has been documented in species of the Tenebrionidae family such as Tribolium castaneum, Tenebrio molitor, Bolitotherus cornutus, G. cornutus, and Onymacris unguicularis (de Villiers and Hanra 1991; Conner 1995; Fedina and Lewis 2008; Okada et al. 2015). Therefore, there is a possibility of multiple mating occurring in Z. atratus as well, suggesting the potential presence of cryptic female choice of this species.

Cryptic female choice, such as regulation of the amount of sperm injected into the spermatophore and spermatophore elimination immediately after mating, has been observed in females of T. castaneum (Fedina 2007). It is therefore possible that females in this species also regulate the amount of sperm transported into the organ. However, it is not clear whether this effect is due to leg injury or the female’s choosiness. Because it is also unclear how long this reduction in fitness continues, we need to clarify the mechanism behind this reduction in the future.

The Z. atratus male does not have the exaggerated mandible. Why does this difference occur? In cases where males of a certain species engage in contests over access to females or territories against conspecific males, exaggerated traits used as weapons usually evolve, and individuals with higher competitive abilities are expected to gain disproportionate access to fertilization using these weapons (Emlen 2008). Sexual dimorphism has been suggested to vary in shape and size due to factors such as contest behavior (Caro et al. 2003) and habitat conditions (Butler et al. 2007; de la Cámara et al. 2023). In Z. atratu, sexual dimorphisms were found for forehead width, mid-head width, head length, left mandible length, left mandible tooth width, right mandible width, and right mandible tooth width. Previous studies investigating sexual dimorphism in the Namib tenebrionid O. plana, which engages in leg-biting combat, did not examine the mandibles (Enders et al. 1998), and O. plana does not possess enlarged mandibles. The morphological differences observed between males and females of Z atratus may be due to adaptations to the leg-biting behavior of males, but factors that shape weapons may also be related to the fighting behavior of males of this species (Caro et al.2003).

The shape of the weapon trait varies with habitat (Emlen 2008). Z. atarus has been found living under attic columns and tiles (Tschinkel 1981). This may have led to the species having a wider mandible suitable for enclosed or narrow spaces, rather than the larger and longer mandibles of stag beetles.

The female’s left labial palp length is longer than that of males. Although we did not find any report on the labial palp in the family Tenebrionidae, in the peach fruit moth Carposina sasakii Matsumura (Lepidoptera: Carposinidae), the labial palp has the function of finding suitable sites for egg-laying (Chen and Hua 2016). In Z. atratus, the labial palp may also have the function of finding suitable sites for egg laying. Further, the sexual dimorphism of the female elytra width is likely the result of an adaptation to laying more eggs.

Males of Z. atratus bite the hind legs of other males, reducing the fitness of their rivals. The decrease in fitness of loser males was observed in escalated combats when the males in combat were the same size. The present experiment clearly showed the difference in the reproductive fitness of uncontested males and losers in leg-biting combats. However, the cause of decreased fitness in loser males is not apparent from the results of this experiment. Therefore, further experiments, such as investigating the relationship between bite wounds and fitness or the possibility of cryptic female choice, are required in the future.

Furthermore, the Tenebrionidae family is diverse in species (Kergoat et al. 2014), and its sexual dimorphism varies greatly (Conner 1995; Enders et al. 1998; Okada et al. 2015; Vendl et al. 2018). For example, there are species like G. cornutus, which possess developed mandibles, and B. cornutus, which display prominent horns. From this, it can be said that the Tenebrionidae family is suitable for investigating diversities in sexual dimorphism, and the results of this experiment can offer crucial insights into this phenomenon.