- Open Access
Skeletomuscular adaptations of head and legs of Melissotarsus ants for tunnelling through living wood
© The Author(s). 2018
- Received: 1 June 2018
- Accepted: 2 August 2018
- Published: 14 August 2018
While thousands of ant species are arboreal, very few are able to chew and tunnel through living wood. Ants of the genus Melissotarsus (subfamily Myrmicinae) inhabit tunnel systems excavated under the bark of living trees, where they keep large numbers of symbiotic armoured scale insects (family Diaspididae). Construction of these tunnels by chewing through healthy wood requires tremendous power, but the adaptations that give Melissotarsus these abilities are unclear. Here, we investigate the morphology of the musculoskeletal system of Melissotarsus using histology, scanning electron microscopy, X-ray spectrometry, X-ray microcomputed tomography (micro-CT), and 3D modelling.
Both the head and legs of Melissotarsus workers contain novel skeletomuscular adaptations to increase their ability to tunnel through living wood. The head is greatly enlarged dorsoventrally, with large mandibular closer muscles occupying most of the dorsal half of the head cavity, while ventrally-located opener muscles are also exceptionally large. This differs from the strong closing: opening asymmetry typical of most mandibulated animals, where closing the mandibles requires more force than opening. Furthermore, the mandibles are short and cone-shaped with a wide articulatory base that concentrates the force generated by the muscles towards the tips. The increased distance between the axis of mandibular rotation and the points of muscle insertion provides a mechanical advantage that amplifies the force from the closer and opener muscles. We suggest that the uncommonly strong opening action is required to move away crushed plant tissues during tunnelling and allow a steady forward motion. X-ray spectrometry showed that the tip of the mandibles is reinforced with zinc. Workers in this genus have aberrant legs, including mid- and hindlegs with hypertrophied coxae and stout basitarsi equipped with peg-like setae, and midleg femura pointed upward and close to the body. This unusual design famously prevents them from standing and walking on a normal two-dimensional surface. We reinterpret these unique traits as modifications to brace the body during tunnelling rather than locomotion per se.
Melissotarsus represents an extraordinary case study of how the adaptation to – and indeed engineering of – a novel ecological niche can lead to the evolutionary redesign of core biomechanical systems.
Social insects can shape their living environment dramatically through the coordinated efforts of nestmates working in unison over time. A wide range of ants (as well as termites) are known to construct extensive networks of underground tunnels and chambers. Within these, some ant species host symbionts such as cultivated fungi or sap-sucking insects that provide honeydew rewards [1, 2]. Such long-lived nests are excavated in soil using the mandibles. Arboreal nests seldom reach the same physical scales, except when host plants provide pre-existing living spaces. In other cases, tunnels are usually carved in dead tissues of standing or fallen trees, with only a minority of ant species able to chew through healthy living wood. The latter generally involve chewing an entrance hole or short tunnel to a pre-existing cavity. While many arboreal ants obtain trophic benefits from scale insects, only relatively few species keep their partners inside the nest chambers.
Video of walking attempts of a Melissotarsus sp. worker taken out of the tunnels of its nest inside a living tree (Nardouwsberg, Western Cape, South Africa). (MP4 10095 kb)
It is likely that the Melissotarsus head is specialized for chewing. The largest muscles within the mandibulated insect head are those that operate the mandibles . Each mandible is controlled by one set of “closer” and one set of “opener” muscles. These muscles act antagonistically by pulling opposite extremes of the mandibular base, causing mandibles to swing towards or away from the mouth (closing and opening motions respectively). These muscles have been investigated in various ants that use mandibles to hunt. In particular, trap-jaw ants have mandibular muscles that generate phenomenal closing speeds, either by direct action or the use of power amplification mechanisms (reviewed in ). However, much less attention has been given to ants having mandibles designed for sustained strength, e.g. cracking seeds, biting enemies, dismembering prey, and tunnelling. Wood-boring insects, including larvae and/or adults of many Coleoptera, Lepidoptera, Isoptera and Hymenoptera (e.g. woodwasps, carpenter ants), usually tunnel through dead or decayed wood [9, 10], or chew very short galleries to reach the center (pith) of living branches (e.g. Gesomyrmex ants) . Compared to decayed dry wood, chewing healthy wood requires more strength because moisture keeps wood fibres elastic as opposed to brittle . Melissotarsus nests apparently extend throughout entire trees . In social insects, large nests are attributed to division of labour involving numerous participants, but the superior morphological characteristics of individual participants also need proper emphasis.
In this study, we used complementary data from serial histological sections, SEM imaging, micro-CT 3D reconstruction, and X-ray (EDX) spectrometry, to test the hypothesis that the Melissotarsus mandible and leg systems have been modified for wood-tunnelling. We reinterpret the characteristic leg morphology of Melissotarsus workers as specialization for anchoring the body during tunnelling, to resist powerful backward forces exerted by the mandibles.
Colonies of Melissotarsus spp. containing many workers, queens, brood and diaspidids were sampled in Gorongosa National Park, Mozambique (August 2016) and Nardouwsberg, Western Cape, South Africa (May 2017). Inhabited branches were transported to Paris. Extremities of the branches were kept moist, allowing to keep ants and diaspidids alive for several weeks.
Despite the compelling morphological distinctiveness of the genus Melissotarsus as a whole, species-level taxonomy is confusing and challenging given the high degree of conservatism in worker morphology across Africa and Madagascar. The genus was revised by Bolton (1982) who recognised three species in Africa . However, given that specimens were available from only few localities, Bolton noted that species boundaries are difficult to ascertain: there is either one widely distributed species, or more than three species. Without a modern taxonomic revision, it is currently impossible to determine the genuine species identity of the material used in this study. Consequently, unique identifiers are given for all specimens scanned (see Additional file 2). Voucher specimens are held at OIST. We refrain from applying any species name and use Melissotarsus instead, emphasizing that all the morphological features studied here apply to the genus as a whole.
Messor barbarus workers were collected in Montpellier (Southern France) and Lisbon (Portugal), and used as a morphologically unspecialized Myrmicinae for comparison.
Melissotarsus worker heads were fixed in cold 2% glutaraldehyde buffered at pH 7.3 with 50 mM sodium cacodylate and 150 mM saccharose. Postfixation was done in 2% osmium tetroxide in the same buffer. Dehydration was achieved in a graded acetone series, and tissues were embedded in Araldite. Serial sections of 1 μm were made with a Leica EM UC6 ultramicrotome, then stained in methylene blue and thionin and viewed with an Olympus BX-51 microscope. Three heads provided 630 longitudinal, 610 transversal and 400 frontal sections.
To check for the eventual presence of zinc in the mandibular tip, we analyzed double stained thin sections (prepared as described above) in a Jeol ARM-200F electron microscope equipped with a probe aberration corrector, operated at 200 kV. The microscope is equipped with an energy dispersive X-ray (EDX) spectrometer to perform EDX measurements at a collection angle of 0.98 sr and with a 100 mm2 detection area.
Sarcomere length measurement
On the frontal sections, some closer muscle fibres were half-relaxed and half-stretched, due to slow penetration of the fixative (W. Gronenberg, personal communication). Sarcomeres were measured in both the relaxed and stretched parts of 15 fibres, allowing us to infer that closer fibres were relaxed on the transversal and most of the frontal sections. Opener fibre sarcomeres, only measurable on the frontal sections, were all stretched: measured lengths were multiplied by the empirical ratio relaxed/stretched calculated from the half-relaxed half-stretched closer fibres. To get accurate measurements, five sarcomeres at a time were measured on several fibres of various regions of the head, yielding a single sarcomere length value per fibre.
Scanning Electron microscopy
Whole unpinned workers fixed in 96% ethanol were air-dried and point-mounted in a natural position. Mounted specimens were coated with gold-palladium and imaged using a Hitachi S4700 field emission scanning electron microscope at a voltage of 5–10 kV.
X-ray micro-computed tomography
Micro-CT scans were performed using a Zeiss Xradia 510 Versa 3D X-ray microscope operated with the Zeiss Scout-and-Scan Control System software (version 11.1.6411.17883) at the Okinawa Institute of Science and Technology Graduate University, Japan. Material of Melissotarsus sp. and Messor barbarus was initially preserved and stored in 90% ethanol. Prior to the scanning procedure the specimens were stained in a 2 M iodine solution for 24 h, and subsequently transferred into microtubes filled with 99% ethanol. Scan settings were selected accordingly to yield optimum scan quality. Full 360 degree rotations were based on 1601 projections. The resulting scans have resolutions of 990 × 1013 × 988 pixels and an overview of the specimens used and scanning parameter settings is provided (see Additional file 2: Table S1). Post-imaging 3D reconstruction was done with the Zeiss Scout-and-Scan Control System Reconstructor software (version 11.1.6411.17883), and the output files saved in DICOM format.
3D reconstruction of image stacks was first visualized with Drishti 2.6.3 , a software that only uses voxel intensity to build 3D models. Transfer Function Editor, Point/String Light, Clipping Plane and Viewport tools were used for internal and external visualization of the worker and queen head, mesosoma and legs. Then, active voxel designation (i.e. segmentation) of the reconstructed image stacks was performed with ITK-SNAP 3.6.0  for one half of the Melissotarsus worker head. The ‘region competition’ algorithm was used for 3D automatic segmentation, which was followed by manual segmentation to correct the boundaries of some structures, for example between the cuticle and the muscles involved in mandible closing. Muscles and skeletal structures were annotated by homology relative to Apis mellifera following Snodgrass (1956) . The resulting segmentation was exported as vtk mesh files. A transparent rendering of the full head was created with the Isosurface tool in Amira software (version 6.3.0), and then exported as a ply mesh file. Meshes were opened in ParaView (version 5.4.1) for visualization, snapshots, and animation. Mandible, apodemes and mandible muscles were segmented similarly for one half of the head of a Messor barbarus worker, then analysed with the same workflow.
To complement the information from virtual 3D models, we dissected worker heads. To assess musculature, heads fixed in 80% ethanol were cut open with a razor blade at various angles. For the assessment of cuticular skeleton, individuals were put in a 12% potassium hydroxide (KOH) solution overnight to remove all the soft tissues. These preparations were examined under a Wild M5 stereomicroscope.
Mandible muscles and apodemes
The opener muscles are considerably enlarged (4.6% of the head volume compared to 1.1% in Messor barbarus). They fill the ventral enlargement of the head, with fibres originating along its concave floor, posteroventral wall of the head, and the large ventromedial phragma that runs longitudinally across the floor of the enlarged ventral cavity. Unlike what is observed in other ants, we found large sclerotized opener apodemes, each consisting of a single broad blade-like process oriented horizontally and parallel to the floor of the head (Fig. 5). Like the closer apodemes, these opener apodemes provide a large surface area for the insertion of the unusually large opener muscles.
The brain is a large central organ whose position conflicts with mandible muscle fibres that cross the head anteroposteriorly. More precisely, the lateral optic lobes are a direct constraint on the volume of the closer muscles and the geometry of the closer apodeme. In Melissotarsus workers, optic lobes are reduced along with eye size. In addition, space between the brain and the muscles is minimal, with fibres passing one micron away over and under the optic lobes. Importantly, the Melissotarsus head also contains multiple clusters of hypostomal silk glands (Fig. 4), as well as pheromone-producing intramandibular and mandibular glands, enzyme-producing propharyngeal glands, and lipid-metabolizing postpharyngeal glands. These glands are additional constraints on the geometry and volume of the mandibular apparatus.
For both closer and opener muscles we distinguished direct fibres inserting on the apodemes, as well as indirect fibres connecting the apodemes via a membranous filament. Closer direct fibres had longer sarcomeres than closer indirect fibres (6.6 ± 1.0 μm vs 5.6 ± 0.6 μm, Wilcoxon test, p < 0.001). Direct fibre sarcomeres were also wider than indirect fibre sarcomeres for the opener muscles (8.4 ± 0.6 μm vs 7.4 ± 0.5 μm, Wilcoxon test, p < 0.01). Opener muscle sarcomeres were longer in both direct and indirect fibres (Wilcoxon test, p < 0.001 for both).
Mandible shape and articulation
Adaptations of mid and hind legs
Head modifications as adaptations for chewing through live wood
The dorsoventrally enlarged head of Melissotarsus workers is packed with dorsal closer and ventral opener muscles. Ventral enlargement of the head provides space for large opener muscles hitherto unreported in other ants. In addition, the increased number of opener fibres in Melissotarsus connect to a spatula-shaped apodeme attached to the expanded outer process of mandibles. Such large sclerotized opener apodeme is not found in other insects [21, 22]. The unusually elaborate mandible opener mechanism in Melissotarsus apparently helps to disengage the mandibles from the elastic wood fibres while tunnelling – like a lumberjack removing his axe from a trunk after each stroke. Without this opening strength, workers would need to walk backwards to reopen their mandibles and chewing efficiency would decrease. It is the strong antagonistic contraction of opener and closer muscles that lies behind efficient tunnelling in live wood.
Not only are mandible muscles very large in Melissotarsus workers, but their output force is maximized through lever effects deriving from the broad base and prominent lateral (outer) process of mandibles. Insect mandibles have a triangular base in which two corners correspond to the mandible hinges and the third corner corresponds to the insertion of the closer muscle . Accordingly, the closer muscles always pull the mandibles at a point much further away from the mandible hinge (the axis of rotation) than the opener muscles (Fig. 7). This is but a very basic lever system: by virtue of the insertion distance of the closer muscle from the fixed hinge, the input force of the closer muscles gets amplified, resulting in a greater output force in terms of biting. Moreover, the same holds for the lateral process where the opener muscle inserts: any elongation of this process will move the insertion of the opener muscle further away from the axis of rotation, increasing the lever effect and amplifying the input force. The broadening of the mandibular base and shorter length of Melissotarsus worker mandibles thus results in an increase in chewing force (both opening and closing actions). To our knowledge, such lever effects resulting from modified mandible geometry have not been described in other wood-chewing insects.
The pointy mandibles of young workers progressively abrade with age , evidence of intense strain while chewing. Together with zinc reinforced tips, widespread in insects (e.g. ), robust design makes Melissotarsus mandibles highly-suited for tunnelling in wood. Given the importance of mandibular sharpness for chewing, the special setae on the mandibles’ surfaces might act as a proprioceptor system to assess mandible wear: in young individuals the sharp tip is longer than any of the setae, so chewing will not exert any pressure on them; as the mandibles wear down, more and more of the setae will stick out and chewing will always mechanically stimulate them, signalling the worker that her mandibles are becoming less effective.
Leg adaptations reflect a trade-off between walking and tunnelling
Melissotarsus workers have developed a specialized morphology suited for their uniquely engineered ecological niche: an obligate mutualist partnership with diaspidid scale insects that feed inside tunnels chewed in healthy wood. Their head is large, packed with silk glands but also huge mandible closer and unusually large opener muscles. The insertion of these muscles and the shape of the mandibles itself maximize the force output for slow but powerful closing and opening motions. The remarkable design of mid and hind legs braces the body while chewing. These morphological adaptations for tunnelling evolved at the expense of normal walking and foraging, an unprecedented situation in ants. Polyphenism allows this extreme specialization of ant workers because the queen caste remains able to disperse by flight and walk on the outside of host trees during the first stages of colony foundation.
We thank Brian L. Fisher for obtaining collecting and export permits for all specimens during Ant Course 2016. We thank Wulfila Gronenberg for sharing his experience regarding sarcomere length measurement. We acknowledge the assistance of An Vandoren in making the histological sections and Alex Vrijdaghs in scanning microscopy. We are grateful to Cédric Van Goethem, Ivo Vankelecom and the Hercules project AKUL/13/19 of KU Leuven for their help in the zinc analysis.
A.K., E.P.E., and F.H.G. were supported by subsidy funding to OIST, and E.P.E. was supported by a JSPS KAKENHI (JP17K15180) grant.
Availability of data and materials
All datasets are deposited on Zenodo, https://doi.org/10.5281/zenodo.1341553.
CP conceived the experiment framework and collected the specimens. CP and AK conducted the behavioural observations. FHG, AK, and EPE collected micro-CT scan data, FHG reviewed the species taxonomy of the genus. JB performed the histology, SEM and TEM-EDX. RAK performed SEM and direct dissections and prepared the habitus illustration. AK analysed the data with the help of RAK, CP, FHG, EPE. AK, RAK, CP, FHG and EPE wrote the paper. All authors read and approved the final manuscript.
Ethics approval and consent to participate
Consent for publication
The authors declare that they have no competing interests.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.
- Delabie JHC. Trophobiosis between Formicidae and Hemiptera (Sternorrhyncha and Auchenorrhyncha): an overview. Neotrop Entomol. 2001;30:501–16.View ArticleGoogle Scholar
- Weber NA. Fungus-Growing Ants. Science. 1966;153:587–604.View ArticlePubMedGoogle Scholar
- Peeters C, Foldi I, Matile-Ferrero D, Fisher BL. A mutualism without honeydew: what benefits for Melissotarsus emeryi ants and armored scale insects (Diaspididae)? PeerJ. 2017;5:e3599.View ArticlePubMedPubMed CentralGoogle Scholar
- Fisher BL, Robertson HG. Silk production by adult workers of the ant Melissotarsus emeryi (Hymenoptera, Formicidae) in south African fynbos. Insect Soc. 1999;46:78–83.View ArticleGoogle Scholar
- Hölldobler B, Obermayer M, Plowes NJR, Fisher BL. New exocrine glands in ants: the hypostomal gland and basitarsal gland in the genus Melissotarsus (Hymenoptera: Formicidae). Naturwissenschaften. 2014;101:527–32.View ArticlePubMedGoogle Scholar
- Delage-Darchen B. Une fourmi de Côte-d’Ivoire: Melissotarsus titubans Del., n. sp. Insect Soc. 1972;19:213–26.Google Scholar
- Snodgrass RE. Principles of insect morphology. New York and London: McGraw-Hill Book Company, inc.; 1935.Google Scholar
- Larabee FJ, Suarez AV. The evolution and functional morphology of trap-jaw ants (Hymenoptera: Formicidae). Myrmecol News. 2014;20:25–36.Google Scholar
- Solomon, JD. Guide to insect borers in North American broadleaf trees and shrubs. Washington, DC: USDA. For Ser Agric Handbk AH-706; 1995.Google Scholar
- Vilhelmsen L, Turrisi GF. Per arborem ad astra: morphological adaptations to exploiting the woody habitat in the early evolution of Hymenoptera. Arthropod Struct Dev. 2011;40:2–20.View ArticlePubMedGoogle Scholar
- Peeters C, Ito F, Wiwatwitaya D, Keller RA, Hashim R, Molet M. Striking polymorphism among infertile helpers in the arboreal ant Gesomyrmex. Asian Myrmecol. 2017;9:1–15.Google Scholar
- Aicher S, Stapf G. Compressive strength parallel to the fiber of spruce with high moisture content. Eur J Wood Wood Prod. 2016;74:527–42.View ArticleGoogle Scholar
- Mony R, Kenne M, Orivel J, Dejean A. Biology and ecology of pest ants of the genus Melissotarsus (Formicidae: Myrmicinae), with special reference to tropical fruit tree attacks. Sociobiology. 2002;40:645–54.Google Scholar
- Bolton B. Afrotropical species of the myrmicine ant genera Cardiocondyla, Leptothorax, Melissotarsus, Messor and Cataulacus (Formicidae). Bull Br Mus Nat Hist. 1982;45:307–70.Google Scholar
- Limaye A. Drishti: a volume exploration and presentation tool. Proc. SPIE 8506, Developments in X-Ray Tomography VIII; 2012. p. 85060X.Google Scholar
- Yushkevich PA, Piven J, Hazlett HC, Smith RG, Ho S, Gee JC, Gerig G. User-guided 3D active contour segmentation of anatomical structures: significantly improved efficiency and reliability. NeuroImage. 2006;31:1116–28.View ArticlePubMedGoogle Scholar
- Snodgrass RE. Anatomy of the honey bee (Ithaca: Comstock Pub. Associates); 1956.Google Scholar
- Hita Garcia F. 3D models of Melissotarsus worker and queen uploaded on SketchFab. 2018. https://skfb.ly/6vFQC. Accessed 29 May 2018.Google Scholar
- Paul J, Gronenberg W. Optimizing force and velocity: mandible muscle fibre attachments in ants. J Exp Biol. 1999;202:797–808.PubMedGoogle Scholar
- Khalife A. Video of surface rendering of internal head structures in Melissotarsus worker ants, specialised for chewing healthy wood. 2018. https://zenodo.org/record/1318495. Accessed 28 July 2018.Google Scholar
- Larabee FJ, Gronenberg W, Suarez AV. Performance, morphology and control of power-amplified mandibles in the trap-jaw ant Myrmoteras (Hymenoptera: Formicidae). J Exp Biol. 2017;220:3062–71.View ArticlePubMedGoogle Scholar
- Li D, Zhang K, Zhu P, Wu Z, Zhou H. 3D configuration of mandibles and controlling muscles in rove beetles based on micro-CT technique. Anal Bioanal Chem. 2011;401:817–25.View ArticlePubMedGoogle Scholar
- Schofield RMS, Nesson MH, Richardson KA. Tooth hardness increases with zinc-content in mandibles of young adult leaf-cutter ants. Naturwissenschaften. 2002;89:579–83.PubMedGoogle Scholar
- Brandão C, Silva R, Delabie J. Neotropical Ants (Hymenoptera) Functional Groups: Nutritional and Applied Implications. In: Parra J, ed. (CRC Press). Insect Bioecology and Nutrition for Integrated Pest Management; 2012. p. 213–236.Google Scholar