Indirect closing of the elytra in a cockchafer, Melolontha hippocastani F. (Coleoptera: Scarabaeidae)

Experimental studies on the flight of beetles are concentrated on their wings: their trajectories (Stellwaag, 1914; Brackenbury, 1994), folding and unfolding (Haas and Beutel, 2001), aerodynamics (Schneider, 1987), elastic structures (Haas et al., 2000), maneuverability and electrical activity of fibrillar muscles (Leston et al., 1965; Schneider and Krämer, 1974; Bauer and Gewecke, 1985; Burton, 1971) and technical modeling (Syaifuddin et al., 2006). Experimental studies on the elytra were rather scarce. Earlier studies investigated the position of the elytra during the flight, aerodynamic forces acting on the open elytra, coupling between the meso- and metathoracic flight systems (Nachtigall, 1964; Schneider and Meurer, 1975; Schneider and Hermes, 1976; Schneider, 1986). We measured trajectories and axes of rotation of the elytra during opening and closing (Frantsevich et al., 2005).

Anatomists have investigated the muscular system and the elytral pivot and sometimes posed hypotheses about the role of certain muscles (Straus-Duerkheim, 1828; Bauer, 1910; Stellwaag, 1914; Larsén, 1966) but they have not described any experiments in confirmation. Recent anatomists describe, among others, the muscles in the mesothorax to reconstruct the phylogeny. They pay no attention to functions of the muscles (Beutel and Komarek, 2004; Friedrich et al., 2009). There exist several studies on the structures locking the elytra together or with other body parts (reviewed in Frantsevich et al., 2005) but they concern immobile elytra. Lastly, we mention my previous observation on a peculiar click of the elytra during righting in histerids (Frantsevich, 1981).

Here, we elucidate the role of muscles that actuate the elytra using (1) video recordings of tethered flying beetles with or without loads applied to the thoracic segments; (2) video recordings of tethered flying beetles wherein some parts of sclerites were excised or certain muscles were cut across; and (3) animation, i.e. passive motion in a dead specimen. The model animals were cockchafers, Melolontha hippocastani and Melolontha melolontha. We show below that closing is performed indirectly by pressure of the hind edge of the pronotum onto the lateral apophysis of the root of the elytron. Experiments on surgery and animation of mesothoracic muscles will be described in the next article (L.F., submitted).

Cockchafers were collected in the field near Kiev, Ukraine. Melolontha hippocastani Fabricius 1801 was stored at 5°C and used in flight experiments; M. melolontha Linnaeus 1758 was stored at −18°C, together with the former species, for animation experiments.

A specimen was killed by freezing. After thawing, we removed the abdomen and, sometimes, cut the thorax down the medial line. The specimen was fixed in 70% ethyl alcohol and dissected under the stereomicroscope MBS-9 (Soviet Union). Specimens with the medial cut were halved. Steps of dissection were photographed with a camera Olympus C-2 Zoom (Olympus Corporation, Tokyo, Japan) adjusted to the ocular.

We used two sites for tethering: (1) the ventrite behind the middle coxae, which provided the firm basis for the pterothorax-fixed reference, and (2) the pronotum – only for demonstration of the flight. The holder was glued with the cyanoacrilate glue. Middle legs were clipped in the first case to prevent grasping to a holder. For indication of the spatial orientation of thoracic segments, thin rods were glued to the scutellum normally to its surface or to the pronotum, normally or tangentially.

A beetle was illuminated with a projector containing a halogen 50 W lamp. The video camera Panasonic NV-A3 (Matsushita Electric Industrial Co., Osaka, Japan) with a headpiece lens provided the frame rate of 25 s⁻¹ with a frame exposure of 2 ms. This rate was enough for rather slow opening or closing of the elytra. Selected episodes were grabbed with the video plate ATI Fury Pro (ATI Technologies Inc., Beverley, MA, USA) and the software Video In 6.3 and Video Editor 6.2 supplied with the plate. Films were digitized in the format MS-MPEG4 V3. During three hours of recording, 353 behavioral acts in 53 live specimens were recorded for the whole project. The quicker method was filming in the movie mode with the photocamera named above. It was used for a macro view with a headpiece lens or for filming under the microscope. 182 behavioral acts in 34 live specimens were recorded.

A thawed dead specimen with clipped legs was glued by the ventrite to a pedestal. We used two types of preparations: (1) a whole animal with the free prothorax actuated with a handle glued tangentially to the pronotum, or (2) the prothorax was removed; thus, obtaining the pterothoracic preparation. Elytra were left intact or unlocked.

Frames for illustrations were grabbed from the films either in the *.mov format replayed by Media Player Classic 6.4.7.5 (GNU, Free Software Foundation, Inc., Boston, MA, USA) or in the *.avi format with the software AVIEdit (AM Software, Moscow, Russia). They were further processed as Adobe Photoshop 5.5 images (Adobe Systems, Inc., San Jose, CA, USA). Films for coordinate measurements were processed frame-by-frame with AVIEdit. Each frame was pasted in an image window of the Sigma Scan Pro software (SPSS Inc., Chicago, IL, USA) for coordinate tracing. Computation of the distance, direction, angle and scaling was performed using custom programs in Turbo Basic 1.3 (Borland International, Inc., Austin, TX, USA). Coordinates for the present article were measured in 1325 frames. Demonstrative movies in the grayscale mode are available as supplementary materials to the present article. Colour originals are available at the website: http://izan.kiev.ua/ppages/frantsevich/.

Anatomy of M. melolontha was illustrated in Straus-Duerkheim (Straus-Duerkheim, 1828). We mention below only details relevant for this study. Fig. 1A shows the medial section across the skeleton.

Note that: (1) articulations of the prothorax both with the head and the mesothorax are of the ball-and-socket type, the sockets lie in the prothorax. No condylar structures exist in these articulations. (2) The endoskeleton includes a sternal apophysis in each thoracic segment and three transversal folds of the tergal cuticle – the prophragma Ph1 in front of the mesotergite, the mesophragma Ph2 in front of the metatergite, and the metaphragma Ph3 at the rear border of the metatergite. (3) Both the meso- and metasternites are fused together in a hard plate, the ventrite (Friedrich et al., 2009).

The elytron is inserted into the articulatory membrane, which divides the mesotergite from the mesopleurite. It articulates to the thorax with a small process amidst the basal edge, the root of the elytron, encircled with four small sclerites, the axillary plates Ax1–Ax4. Ax2 is articulated with the pleural shaft, three others offer insertions for the direct elytral muscles. The dorsal side of the root is formed by a narrow medial apophysis and a broader lateral one.

Two closed elytra are tightly compressed to the body and to each other with the aid of specific structures, the locks (Stellwaag, 1914): a friction lock between two elytra down the suture, a clamp between the anterior part of the sutural edge and the groove on the metanotum, a clamp between the most anterior part of the sutural edge and the scutellum, a clamp between the antero-costal edge and the epipleuron, etc. The named locks must be clipped off to provide freedom for animated elytra. Closed elytra are also locked by the hind edge of the pronotum fitting into the groove on the basal edge of the elytron.

Fig. 1B schematically shows intersegmental muscles of the mesothorax. Labeling of the muscles follows Larsén (Larsén, 1966). We show also the indirect longitudinal wing muscle M61. Direction of fibers in M61 indicates the longitudinal body-fixed axis, universal for all flying beetles. Prothoracical rotator M13 from the pronotum to the mesepisternum is not depicted. Direct wing muscles are listed in Table 1 by earlier publications.

Beetles from the fridge were put into a transparent box. Positively phototropic specimens were selected for tethering. The vast majority of them were able to fly after some recovery period. Out of 87 tethered flying specimens of M. hippocastani, seven flew constantly and were arrested only by touching their legs with a brush. Most of beetles flew in bouts. 162 interbout intervals were counted frame-by-frame in 34 undisturbed beetles. Mean interval comprised 18.6±14.0 s (mean ± s.d., range 3–100 s, median 14.5 s). Duration of wing oscillation was about 1.5 s. Between the flights, a beetle devoid of the ground support always performed righting search movements, i.e. lifted the legs above the dorsum and vigorously moved its prothorax.

The flight includes a preparatory and an accomplishing stage and involves the whole body into this activity (see supplementary material Movie 1). Preparation for flight starts with a deep depression of the prothorax followed by lifting of the linked elytra and later by their abduction and elevation. The abdomen is depressed at this moment to give space for unfolding wings. The wings unfold and begin to oscillate. The spread elytra rise and droop in synchrony with the wings, the mesotergite oscillates too. The legs are held static. The abdomen is constantly elevated during the flight. After the cessation of wing strokes, the wing folding begins, which is accompanied by the depression and adduction of the elytra, depression of the abdomen and obligate elevation of the prothorax. Evidently, during the elevation, the hind edge of the pronotum moves backwards towards the mesotergite (see supplementary material Movie 2). The legs resume their search. The wings in most cases are not folded entirely; the accomplishing folding includes coordinated movements of the abdomen and elytra which have been studied earlier (reviewed in Haas and Beutel, 2001).

The time course of these events is plotted in Fig. 2. We measured angles in projection on the pitch plane between the tethering handle and direction markers: rods glued to the prothorax and the scutellum, the edge of the elytron, the line from the base to the tip of the abdomen. We take into account only the relative displacement. The pixelization error in 20 separate measurements of the prothoracic rod in the same frame yielded the standard deviation of direction ±0.41 deg. The range of excursions of the prothorax was 24.5 deg., that of the abdomen was 23.5 deg. and vibration of the mesonotum was 14 deg.

Strokes of the elytron occur perpendicularly to the pitch plane, they are not recognized by change of direction of the elytron. We evaluated position of the elytron by the length, in pixels, of the projection of the elytral edge: long if closed or elevated high, short if elevated low. Strokes are apparently slowed down due to an occasional stroboscopic effect of the low frame rate of 25 s⁻¹ combined with the short exposure of 2 ms. Synchrony between the wings, elytra and mesotergite is evident. The downstroke is synchronized with the protraction of the mesotergite.

If depression of the prothorax before the flight seems obligate in order to unlock elytra for a free opening, the obligate elevation a priori seems not so necessary, because the open elytra have space to turn back. We disturbed the elevation, using loads that slowed down this act in the upright tethered beetles.

We applied a weight on a lever glued to the prothorax, which would bend this segment down in an upright beetle tethered by the mesoventrite. The lever was made from an insect pin with a hook at the end. It was glued tangentially to the pronotum. Weights were made of coins whose mass was 1, 2, 3 or 5 g. Upon adding sufficient weight, the prothorax bent down in an upright tethered beetle. Nevertheless, the beetle did not change its flight bout performance. There existed a threshold weight that prevented the flight. This threshold was specific for a tested specimen and ranged from 5 to 10 g in fresh beetles but only to 2 g in aged ones. We analyzed behavior with the sub-threshold weight.

During interbout search periods, the prothorax in an unloaded beetle performed rhythmical and equally slow elevations and depressions. In the loaded upright beetle, elevations were slowed down and depressions became abrupt. The range of the elevation–depression of the prothorax was about 25 deg. Spatial orientation of the tethered unloaded beetle did not influence its flight performance. In a beetle turned upside down, the direction of the load was inverted. The intermittent behavior was the same in the upright and overturned unloaded specimens, while the prothorax in the overturned loaded beetle was almost arrested in its elevated position: the range of the elevation–depression was only 8.8 deg.

The time course of opening and closing in an upright and upside down beetle is plotted in Fig. 3. The flight behavior with a negligible load did not depend on the tilt of a beetle. The elevation of the pronotum was synchronous with the depression of the elytra during closing. An additional load did not hinder the opening in an inverted beetle. It strongly retarded closing in an upright beetle, counteracting elevation of the prothorax whereas in an inverted beetle it facilitated elevation of the prothorax and closing (see supplementary materialMovie 3 and Movie 4). Statistics of the time for opening and closing in 105 flight episodes with or without a sub-threshold load is plotted in Fig. 4. The load significantly (Student's t-test, P₀<0.1%) retarded only the closing and only in upright beetles, wherein it also hindered elevation of the prothorax.

The center of rotation of the ball-and-socket joint between the pro- and mesothorax does not coincide with the contact zone between the segments. We found position of the center, filming in profile movements of the prothorax in a beetle tethered by the ventrite. A paper stripe was glued to the prothorax in the pitch plane. Two distant dots on this stripe, above and below the prothorax, circumscribed arcs during movements of the prothorax. Geometry of these arcs provided evaluation of the center location.

The center of rotation marked with an asterisk in Fig. 1 is situated at the level of the costal edge of the closed elytron, 4–5 mm behind the front butt of the mesothorax, far below the prophragma and insertions of M4 above this phragma.

A beetle tethered by the ventrite was allowed to rise and droop the lever glued to the pronotum. Small pliers were fixed in the open state within the inferior area of the lever excursions. The experimenter could clamp the lever manually in its inferior position during the flight of the insect; thus, fixing the prothorax depressed. 95 episodes of the free or clamped flight were recorded in nine beetles.

Only one specimen demonstrated the complete arrest of the elytra in the raised state after cessation of the flight (see supplementary material Movie 5). Other specimens were able to partially close their elytra in the clamped state, followed by recoil after the release. In order to elucidate coordination of body parts, we recorded three beetles with the straw rods glued normally to the scutellum. The lever and the rod indicated the spatial orientation of the pronotum and mesotergite. Fig. 5 illustrates behavior during the clamp and recoil. Indeed, the mesotergite is not arrested by the clamp but slowly moves forwards whereas during the recoil it promptly returns back.

Mean parameters were compared for seven free and seven clamped flights in the same three specimens. All angles were measured in a projection onto the pitch plane. The range of excursions of the prothorax was 24.8±7.5 deg. (mean ± s.e.m.) and 26.9±4.5 deg. for free and clamped flights, respectively. Before the first stop at 0.43 s after the beginning of the closing, elytra of clamped beetles moved on average about −37.1±6.4 deg., range from −6 deg. to −53 deg., while during the release they moved on average by −26.6±7.4 deg. Duration of release was 0.24±0.02 s. On the background of the immobile prothorax during the partial closing, the mesotergite protracted by 4.2±0.6 deg. The release of the mesotergite was retraction about −7.0±1.5 deg.; despite this, the turn of the elytra was performed by the strong elevation of the prothorax.

A frozen and thawed beetle was glued to a pedestal so that the prothorax and the head of the specimen had freedom to move. The articulatory membrane and several intersegmental muscles were cut in order to avoid hindrance to induced movements from the side of stiff contracted muscles. A handle pointing forward was glued to the pronotum.

The prothorax was depressed with this handle and the elytra were passively open. Elevation of the prothorax by the handle caused the closing of the elytra (see supplementary material Movie 6). This performance was recorded in 17 episodes in three specimens. The same manipulation was successfully tested in a dozen coleopteran species as well (L.F., unpublished).

This result suggests that it is an interaction between the hind edge of the pronotum and the open elytron, which closes this last. In order to locate the interaction site, we excised pieces on the hind edge in several specimens and tested whether the elevation was able to close the elytron. We excised areas in front of the scutellum (Fig. 6A), the root of the elytron (Fig. 6B) and the anepisternum. These experiments were video recorded (see supplementary material Movie 8), each prothorax was fixed, cleared of muscles and photographed. We found out that the first and the last excision did not impair the animated closing whereas the excision in front of the root prevented it.

The site on the root, which must be actuated for the closing, was tested manually with a thin acupuncture pin in an unlocked preparation of the pterothorax with the passively open elytron. We found out that the backward-directed pricking amidst the lateral apophysis readily turned the elytron into its closed position. This performance was recorded in 20 video films (see Fig. 7 and supplementary material Movie 7).

Upon the passive opening, the lateral apophysis rotated forward and protruded ahead relative to its resting position by 0.75 mm at the pinpoint, by 1 mm at the base in a specimen of M. melolontha 28 mm long. We conclude that the pressure of the hind edge of the pronoum onto the lateral apophysis turns the apophysis back, certainly together with the elytron itself. Excursion of the pronotum relative to the mesotergite in live beetles was in the range 1.3–1.5 mm. It is possible to animate the opening of the unlocked elytra by manipulations with the mesotergite or the mesepisterna but reverse motions do not close the spread elytra.

A deep excision in the hind edge of the pronotum was made in a cooled live beetle. The beetle was tethered by the ventrite and tested for the ability to fly and close the elytra. Preliminary experiments on five specimens were only partly successful: some specimens could normally close their elytra. An anatomical inspection revealed that the excision in these specimens was not precise with respect to the root position: either shifted sideward or left residual pieces of the invaginated pronotal edge. The root is covert in an intact beetle. Therefore, we searched for reliable landmarks on the elytra, which indicated the site of operation (Fig. 6C). The root is situated in front of the area between the second and the third stria (longitudinal ribs on the elytron).

Further operation on five beetles was successful in all specimens (see supplementary material Movie 9). The excision was done in front of the root of the right elytron. The beetles were able to open both forcedly closed elytra, flew quite normally but after cessation of the flight only the left elytron returned to the rest position, the right one remained abducted (Fig. 6D–H). Anatomical checks revealed that the space in front of the right root was clear.

Direct wing muscles in the mesothorax of M. melolontha and some other beetles were investigated long ago. We list them in Table 1 together with some other muscles originating from the mesotergite and mentioned below. Most probably, judgement on their function has been derived from their position and from some undescribed animations. The opinion on functions of M36, M40 and M42 (in Polyphaga) was unanimous whereas opinions on the function of M33 and M35 or its analog M43 in Adephaga were diverse. Anyhow, previous authors located actuators of closing exclusively within the mesothorax.

Straus-Duerkheim considered M33 in M. melolontha to be the adductor of the elytron, i.e. the closer muscle (Straus-Duerkheim, 1828). Stellwaag suggested that closing in a stag beetle Lucanus cervus demanded relaxation of M42, M35 and then contraction of M36a, M36b (Stellwaag, 1914). This idea was corroborated by Schneider and Meurer for a dynastine beetle Oryctes boas (Schneider and Meurer, 1975). According to Herbst, closing in chafers demanded relaxation of M33and M40 (Herbst, 1944). Elytra return back driven by gravity and elasticity of the mesonotum, supported by contraction of M36a. This last folds the area of Ax3 and pulls the sutural edge of the elytron mesad.

The animation experiment clearly demonstrated that the elevation of the prothorax in a dead animal is enough for closing of the previously spread elytra. The question is whether this mechanism does work in vivo. Measurements of coordination between body parts during the flight reveal that, during slow opening, the depression of the prothorax precedes the divergence of the elytra. On the contrary, during closing, both convergence and elevation are synchronous. Experimental retardation of the elevation retarded the closing. An arrest of the depressed prothorax in a flying beetle either hindered or prevented the closing.

Mechanical interaction between the elytron and the prothorax may be limited to the contact point between the posterior edge of the pronotum and the lateral apophysis of the root, which rotates forward during the opening. The pronotum presses on the root and turns it backwards. Deprivation of this contact by excision on the pronotal edge prevented closing in animation experiments as well as in vivo. Thus, we can consider the elevation of the prothorax as the indirect and main mechanism of closing in Melolontha and perhaps in other beetles.

The hypothesis about closing by pressure of the pronotal edge upon the root of the elytron was proposed by the present author (Frantsevich, 1981) in his early work on histerid beetles. Despite some errors in reconstruction of the musculature, we believe that we properly described the mechanism of the momentary click of the closing elytra in these beetles: the closing lasted 2–3 ms in the larger Hister unicolor and only 0.5–1 ms in the small Atholus duodecimstriatus. The direct elytral muscles pretending to be closers are so feeble in histerids that M36 and M35 were omitted by me (Frantsevich, 1981), M35 [labeling by Larsén, (Larsén 1966)] was not identified by Beutel and Komarek and announced as probably absent, while their description of M36 corresponds not to M36 but to M35 (Beutel and Komarek, 2004). It is clear that such fast movement needs previous storage of energy, presumably in the isometrically contracting large elevators of the prothorax, and its momentary release after unlocking, the mechanism resembling the jump of a click beetle (Evans, 1973).

The muscles which elevate the prothorax in Melolontha are proposed below judging by their arrangement relative to the center of rotation of the prothorax about the mesothorax: these are M11 and M4, maybe the transsegmental M2 and M8. M4 was defined as the depressor in an elaterid Athous haemorrhoidalis (Evans, 1973). The difference between Athous and Melolontha with respect to the drive of this muscle is in the shape of the prophragma (straight in Melolontha, curved down in Athous) and in the type of the joint between the pro- and mesothorax. It is a ball-and-socket joint in Melolontha with the center of rotation situated in the metathorax, behind and below the insertion of M4 above the prophragma. It is a bicondylic monoaxial joint in Athous, pits for condyles lie at the anterior edge of the mesotergite above the prophragma and the insertion of M4.

The proposed mechanism of the indirect closing does not exclude contribution of other mesothoracic muscles. Partial closing of the elytra after cessation of the flight in a clamped beetle may be explained by the limited protraction of the mesotergite: which together with the bases of the elytra approaches the hind edge of the pronotum. This protraction may be caused either by elevators named above or by M29, a presumable antagonist to M28. To date, we have no experimental evidence about the role of some direct elytral muscles in the closing. Direct M36 and M35 are feeble comparing with large and strong elevators. These last provide enough power for the broad rotation of the elytron during closing, sometimes very prompt, and for compression of wings under the elytron during wing folding after the flight.

I am indebted to Dmytro Gladun for collection of M. melolontha.