Among twenty-two horned beetle species three were maintained at the Tokyo University of Agriculture and Technology: Trypoxylus dichotomus (adult male length, 40–80 mm; adult female length, 40–60 mm (Bouchard et al. 2014)), Dynastes hercules (♂ 50–170 mm; ♀ 40–60 mm (Bouchard et al. 2014)), and Endebius gideon (Linnaeus, 1767) (body length, 35–75 mm (Bouchard et al. 2014)). The other horned beetle species Chalcosoma atlas (Linnaeus, 1758) (60–130 mm (Bouchard et al. 2014)), were purchased from Fujikon Co., Ltd. (http://www.fujikon.net/), Lumberjack Co., Ltd. (http://www.lumber-j.com/) and various insect-dealing internet sites (for example http://64auc.sakura.ne.jp/):Chalcosoma chiron (Oliver, 1789) (50–130 mm (Kohiyama 2014)), Megasoma elephas (Fabricius, 1775) (58–130 mm (Kohiyama 2014)), Megasoma actaeon (Linnaeus, 1758) (50–135 mm (Kohiyama 2014)), Dynastes tityus (Linnaeus, 1763) (40–60 mm (Bouchard et al. 2014)), Dynastes neptunus (Quensel, 1817) (50–150 mm (Kohiyama 2014)), Dynastes satanas (Moser, 1909) (55–115 mm (Kohiyama 2014)), Eupatorus birmanicus (Arrow, 1908) (40–50 mm (Kohiyama 2014)), Haploscapanes Barbarossa (Fabricius, 1775) (26–56 mm (Kohiyama 2014)), Allomyrina pfeifferi (Redtenbacher, 1867) (27–40 mm (Kohiyama 2014)), Augosoma centaurus (Fabricius, 1775) (33–80 mm (Kohiyama 2014)), Oryctes rhinoceros (Linnaeus, 1758) (30–45 mm (Unno 2006)), Eophileurus chinensis (Faldermann, 1835) (18–25 mm (Kohiyama 2014)), Golofa aegeon (Hope, 1837) (58 mm (Imamori 2011)), Golofa pizarro (Hope, 1837) (25–40 mm (Unno 2006)), Endebius florensis (Lansberge, 1879) (77 mm (Imamori 2011)), Eupatorus gracilicornis (50–90 mm (Kohiyama 2014)), Cyclocephala complanata (Burmeister, 1847) (16 mm), and Hexodon latissimus (Olivier, 1789) (28 mm). The body weight of an adult male D. hercules was 22 g and adult female weighed 18 g, whereas the T. dichotomus adult male weighed only 8 g (Suppl. material 1).

Dynastes hercules are naturally found in the Neotropical region of southern Mexico to Bolivia and Trinidad, Guadeloupe, and Dominica in the West Indies (Ratcliffe and Cave 2015). Dynastes hercules are among the largest beetles and are the most recognizable insect species in the world. Adults fly mostly at night, especially during the first two hours after sunset (Bouchard et al. 2014). Trypoxylus dichotomus are the largest and best-known beetles in Japan. Adult males of this species use their long, forked horns in battles with rival males. Males are active all summer long, whereas females die soon after laying their eggs (Bouchard et al. 2014).

All larvae of horned beetles were fed on breeding fermentation mats (DEBURO Pro) that were purchased from Fujikon Co., Ltd. (http://www.fujikon.net/). All adults were fed on breeding jelly (DORCUS JELLY) containing animal protein, trehalose, collagen, and banana flavor (Fujikon Co., Ltd).

Synchronous movements and gear-like structures were characterized and compared between adult beetles and adults of various other insect species. Experiments were performed with five males and five females of T. dichotomus and D. hercules, and comparisons were made with single adult males of all other horned beetle species. Sixteen adult specimens of other beetle species (not Dynastinae) were collected from the field or purchased from insect-dealers. Sexes of these non-horned beetles were not recorded. Mouthparts of insect specimens were softened by immersion in water all day prior to observations of the synchronous movements and the gear-like structures of mandibles.

To observe the adductor and abductor muscles of the mandible, adult T. dichotomus specimens were fixed and preserved in 90% alcohol. Heads were dissected carefully using a file (Bkong YWE-B pencil type router; Yanase Corp., Hyogo, Japan) and tweezers. Mandible muscles were observed using binocular microscopy (Stemi 2000-C; Carl Zeiss Microscopy Co., Ltd.). We compared apparent volume of photographed adductor and abductor muscles. Ratios of adductor and abductor muscles in T. dichotomus biting type mandibles were compared with those of adult Locusta migratoria (Linnaeus, 1758) (Orthoptera: Acrididae).

Initially, horned heads were separated from dead male T. dichotomus specimens and were soaked in tap water overnight to soften the hardened articulations, and the labium and maxillae were then removed carefully using a file and tweezers. The average length from the tip of the horn to bottom of the head was 29.3 mm (Suppl. material 4). Twenty male heads with intact gear-like structures between right and left mandibles were used for measurements of right and left mandible strengths in break resistance analyses. Similarly, ten male heads with broken gear-like structures were used in tests of strengths according to breakage resistance. Head specimens were then fixed by sandwiching the horns between two plastic erasers (to prevent slipping) using a sponge and a small vice as shown in Fig. 2A. A paper container with a wire was attached to grip the tip of the mandible (Fig. 2B) and metal nuts of 4.5 g were used as weights. Load was applied to the mandibles at a right angle to the rotary direction of movement (Suppl. material 5). The container wire was then hung on the tip of a single mandible and the metal nuts were added stepwise to the container until the mandible was dislocated and fell from the mouth. Gross weights of nuts and the container were then recorded. Statistical analyses were performed using a generalized linear model (GLM) with R (Version 3.2.2). For multiple comparisons, Tukey-Kramer tests were used to identify differences between groups of mandibles.

Figure 2. 

Measurements of mandible strengths of break resistance. A T. dichotomus horns were fixed in a small vice B wire was hung on the tip of the mandible.

If horned beetles use mandibles to carve bark, the tips of the mandibles must be in contact with the tree surface earlier than the projections of the clypeus when the beetle takes the posture for bark carving. Therefore, we determined whether the tips of the mandibles or the projections of the clypeus touch the tree surface first. To this end, paraffin (solidification point about 51–53 °C) was melted, and a small piece of black crayon was melted and mixed in the plastic container to color the paraffin (Sterile No. 2 Square Schale; Eiken Kizai Co., Ltd. Tokyo, Japan), and was then cooled until it hardened. Initially, we determined whether tips of mandibles or the projections of clypeus made pits in the paraffin surface first. But pits on the paraffin were harder to see than scratches on the paraffin. Thus, heads of horned beetles of several genera were manually held at various forward angles (from 0°to 60°) between the horn and paraffin and were used to scratch the paraffin surface. Heads were also placed in the back position and the scars were photographed (Fig. 3).

Figure 3. 

Scarring of flat paraffin surfaces; T. dichotomus heads were placed in the back position to show the widths of scars. A lateral view B frontal view. Only scratches from T. dichotomus are shown.

Videos of bark-carving behaviors and actions of mandibles were recorded using a Sony Handycam HDR-PJ590V digital HD video recorder (Sony Corp., Tokyo, Japan). Perches comprised rotten wood and plates of jelly, which were made from logs that were used for cultivation of mushrooms, and were purchased from Fujikon Co., Ltd. We assumed that the bark-carving behavior observed in laboratory conditions on wood logs is similar to that observed on tree barks in the wild. Videos of bark-carving behaviors were recorded using adult D. hercules and T. dichotomus. The carving behavior of an adult male T. dichotomus was also recorded in a plastic green insect cage. To obtain specimens with broken gear-like structures between mandibles but no damage to articulation, gear regions of mandibles were disabled using a file (Bkong YWE-B pencil type router; Yanase Corp., Hyogo, Japan). The router bit was made from a nail and acrylic resin was used to bind mandibles. All materials were purchased from Fujikon Co., Ltd. (http://www.fujikon.net/) as an insect specimen kit.

Mandibles of adult horned beetles are present on both sides of the head (Figs 1, 4A–C). Two types of mandible tips have been described previously (Rowland and Miller 2012); those in Chalcosoma, Eupatorus, and Trypoxylus do not branch off, are sharp, and turn upwards, and those of Dynastes and Megasoma species are shaped like forks and turn upwards (Fig. 1 and Suppl. пaterial 6). No sexual differences in mandible forms were identified. Because mandible tips fail to make contact with each other, the mandibles of T. dichotomus and D. hercules do not have biting and chewing functions. The videos recorded in this study indicate that T. dichotomus and D. hercules use their mandibles to carve bark and do not use the projection of the clypeus. Use of the projection of the clypeus to carve bark or xylem would obscure the orange hair of the maxilla and labrum under the wood. However, in the images of adult male (from 20–40 s in Suppl. material 2 and 50 s to 1 min and at 30 s in Suppl. material 3), female D. hercules carving wood (from 1 s to 1 min and 10 s in Suppl. material 7) and adult male T. dichotomus beetles carving the green plastic cage (from 1 min and 28 s to 1 min and 35 s in Suppl. material 7), the orange hairs are visible. Moreover, during bark-carving, mandibles were fully opened (Suppl. materials 2, 3, 7) to about 30 degrees from the closed position (Fig. 5A–C), and microscope observations show that the tips of the mandibles were projected further forward than the projection of the clypeus. To confirm these observations, we scratched a flat paraffin surface using the heads of various horned beetle genera at various forward-position angles. In all cases, mandibles, rather than projections of the clypeus, scratched the paraffin. We showed only the scars from paraffin surface scratching experiments with T. dichotomus in Fig. 3.

Figure 4. 

Mandibles of T. dichotomus adults. A A male adult viewed from the left side with a ruler B lateral view of the head from the left side; the clypeus and mandible are indicated by arrows. Antenna and maxillary palp parts were removed to expose the mandible C diagram of the head; the mandible is indicated by a stipple D a head before removing the mandible E a head after removing the mandible F the left side mandible was dissected from the cranium G diagram showing the double articulation of the mandible; the right side diagram is of the cranium H a left side mandible of T. dichotomus with attached muscles. Abbreviations: add = adductor muscle; abd = abductor muscle; d = dorsal articulation (socket); v = ventral articulation (condyle). Scale bar: 1 mm.

Figure 5. 

Engagement of gear-like structures on mandibles of D. hercules during opening and closing; engaged regions of mandibles were exposed using a file. A Lateral view of the head from the left side; dotted lines indicate the position of the transverse section B interior view of the head from the ventral side C diagram of the engagement region; gear-like structures are emphasized by stipple. Abbreviations: R = right compound eye; L = left compound eye. Scale bar: 7 mm.

In further experiments, both mandibles of adult horned beetles moved in complete synchrony. Specifically, manual opening and closing of the right mandible of living adult T. dichotomus and D. hercules beetles were accompanied by synchronous opening and closing of the left mandible. Similarly, movements to the left and right were precisely synchronous in both mandibles, and simultaneous mandible movements were also observed in dead insects (Suppl. material 8). These observations suggest that the mandibles are mechanically connected.

To investigate mechanical connections between left and right mandibles, we dissected the heads of T. dichotomus and D. hercules and observed mandible structures under a stereomicroscope. Mandibles of adult T. dichotomus and D. hercules are single, heavily sclerotized pieces with a dicondylic articulation. The ventral condyle of the mandible is a ball-like structure that fits into a socket like the acetabulum of the cranium, and the dorsal condyle of the cranium fits into an acetabulum of the mandible (Fig. 4D–G). This dicondylic articulation is typical for biting-chewing mouthparts. Because two points of the articulation are central to the axis of the mandibular movement, T. dichotomus and D. hercules mandibles can only move in a single plane. No sexual differences in mandible forms were identified. Mandibular muscles differed from those of typical biting-type insects. In particular, the adductor muscle of biting insects is generally very large and the abductor muscle is small (Chapman 2013, Snodgrass 1993), whereas in T. dichotomus, the adductor muscle is relatively small and the abductor muscle is larger than in other biting-type insects (Fig. 4H). Adductor and abductor muscles of mandibles are compared between T. dichoromus and typical biting type insect L. migratoria in Suppl. material 9. Furthermore, the posterior regions of mandibles that could not be observed from the outside were asymmetrical, and the gear-like structure was hard and engaged between the mandibles. Artificial devices that transmit force include V-belts and gears. The general definition of a gear is “a toothed machine part that engages successively with another toothed part to transmit motion or to change speed or direction.” Although the mandibles of T. dichotomus and D. hercules do not rotate 360°, their structure fulfills this definition, as observed in the mouth cavity after stripping the labium and maxillae. Initially, we suspected that these posterior region structures were molar cusps. However, adults of T. dichotomus and D. hercules could not separate engaged mandible and these gear-like structures remained in contact after death (Suppl. material 8). In contrast, no gear structures of mandibles were observed in larvae of T. dichotomus (Suppl. material 10) and D. hercules (Suppl. material 11), and posterior regions of mandibles in these specimens were separable, indicating that the gear-like structures appear at the adult stage.

The three-dimensional shape of the mandible gear-like structure is very complicated (Fig. 6A, B). Because the gear-like structure with the engagement of mandibles was difficult to represent in two dimensions, we solidified the mandibles of D. hercules by filling with acrylic resin and then filed the solidified mandibles to expose horizontal and ventral sections. As shown in Fig. 6, left and right mandibles have two-gear teeth each, and the engagement point moved with opening and closing of mandibles (Suppl. material 8).

Figure 6. 

Gear-like structures on mandibles of adult horned beetles. To show the mandibular gear-like structure, the head was split along the anterior midline, and the mandibles were laid out and viewed from the inside. A Mandibles of a T. dichotomus male adult B diagram of structures in Fig. 6A; gear-like structures are shown as stipple. Arrows indicate the tip of each tooth. Scale bar: 2 mm.

In further analyses, we observed gear-like structures of mandibles in adults of 20 other species of horned beetle, suggesting that the gear-like structure of mandibles is common to all adult horned beetles (Suppl. material 6).

Many insects have asymmetric mandibles during the larval or adult stages (Snodgrass 1993). Thus, we examined the posterior regions of the mandibles from various other insect species, but found neither synchronous movements nor mandible gear-like structures (Suppl. material 6).

To investigate the roles of the mandible gear-like structures, we measured the break-resistance strengths of mandibles from dead adult T. dichotomus under load. Preliminarily, we noticed that considerable force was necessary to dislocate single sides of intact and engaged mandibles from the cranium using tweezers. Moreover, although two articulations remained in the mandibles, these could be dislocated easily. However, there was a possibility that initial loading may damage the remaining mandibles and its articulations. Therefore to exclude this possibility, we used one intact specimen for only one side mandible dislocation experiments (Fig. 7, Right, Intact and Left, Intact groups). Additionally, to obtain the specimen in which the gear-like connection between mandibles was broken without damage to articulation, we filed the gear-like structure region (Suppl. material 12) and performed further load-dislocation experiments. As shown in Fig. 7 (Left, Intact), a load of 600 g was required to dislocate the left side mandible from the cranium. On the other hand, a load of about 400 g was required to dislocate the right side mandible. Thus, we found that the break-resistance strengths of right and left mandibles differed in the intact state (Fig. 7). Disengaged mandibles were dislocated under smaller loads than those of required to dislocate engaged mandibles (Fig. 7, Suppl. material 4). Although no significant differences were observed between “Right”, “Intact”, and “Broken” mandibles, disengaged right mandibles (Broken) were dislocated under smaller load than those required to dislocate engaged mandibles (Intact). In contrast, there were significant differences between “Left”, “Intact”, and “Broken” mandibles. Moreover, the difference in strength between right and left mandibles was not observed after filing the gear region (Fig. 7, Right, Broken and Left, Broken groups). Finally, unlike man-made single gear structures that do not prevent movement up and down along the axis of rotation, the complicated gear-like structure of horned beetle mandibles prevents such movements during the use of only one mandible side (Fig. 5C).

Figure 7. 

Load strengths of mandibles from a T. dichotomus adult male. Abbreviations: Right = right mandible; Left = left mandible. Intact = right and left mandibles in the intact state; Broken = the gear-like structure was broken before the measurement. Vertical bars indicate standard errors of the mean (SE). Significant differences between mandible break-resistance strengths are indicated by different letters (a, b; Tukey, p < 0.05).