Silks play an important role in the life of various arthropods. A highly neglected prerequisite to make versatile use of silks is sufficient attachment to substrates. Although there have been some studies on the structure and mechanics of silk anchorages of spiders, for insects only anecdotal reports on attachment-associated spinning behaviour exist. Here, we experimentally studied the silk attachment of the pupae and last instar caterpillars of the tea bagworm Eumeta minuscula (Butler 1881) (Lepidoptera, Psychidae) to the leaves of its host plant Ilex chinensis. We found that the bagworms spin attachment discs, which share some structural features with those of spiders, like a plaque consisting of numerous overlaid, looped glue-coated silk fibres and the medially attaching suspension thread. Although the glue, which coats the fibres, cannot spread and adhere very well to the leaf surface, high pull-off forces were measured, yielding a mean safety factor (force divided by the animal weight) of 385.6. Presumably, the bagworms achieve this by removal of the leaf epidermis prior to silk attachment, which exposes the underlying tissue that represents a much better bonding site. This ensures a reliable attachment during the immobile, vulnerable pupal stage. This is the first study on the biomechanics and structure of silk attachments to substrates in insects.
Natural silks are bio-fibres that fascinate through their outstanding material properties and versatility. Silk spinning behaviour has independently evolved in different clades of arthropods, including insects, such as moths, caddisflies, weaver ants and web spinners; arachnids, like spiders, mites and pseudoscorpions; and some crustaceans [1–4]. Silks are composed of large structural proteins (so-called fibroins or spidroins) that form long fibres, and are produced in specialized ectodermal glands [1,2,5–7]. Silks can fulfil a broad bandwidth of functions, above all cocoon and egg sac formation, and, further, locomotion (abseiling, ballooning, jumping stabilization and floating), prey capture, shelter, defence or as part of courtship and mating rituals [1,3].
For most purposes, the silk threads must be attached to substrates, and although attachment is a crucial step in animal constructions, it is severely under-studied . The attachment of silk threads to substrates is usually achieved by a secretion coating that solidifies after extrusion [9–11]. In derived spiders (Araneomorphae) the structural threads lack the glue coat , and are anchored by means of auxiliary glue-coated nano-threads, applied in a distinct pattern [11,13–15]. Such, so-called, attachment discs provide strong attachment on a variety of surfaces [11,16,17], and presumably recruit high pull-off resistance by self-stabilizing mechanisms [11,15,16,18–20].
Lepidopteran caterpillars may also produce attachment points, in which the threads are laid down in multiple loops [21–23]. This may be a means to enhance the contact area with the substrate. However, in these cases it is the structural thread itself that adheres due to its sericin coat. Direct thread attachments might be prone to tensile loading under steep angles, because stress concentration increases with the increase in the pull-off angle in thin, flexible adhesive structures [16,24,25]. Furthermore, many plants exhibit cuticular waxes and corrugated surfaces, which strongly reduce the wettability of fluids [26–28]. Therefore, the gluing of silk threads onto plant surfaces seems challenging. In the literature, this issue is barely addressed. Although there is some information on the mechanics of silk thread anchorages in spiders [11,16,17], as well as other thread-like attachment devices, like mussel byssus [29–31], we are not aware of any study on the mechanics of silk attachments to substrates in insects. This is surprising, given the important role silk plays for the biology of various arthropods, and the enormous interdisciplinary and commercial interest in such bio-fibres.
Here, we studied the attachment of the armoured cases of the tea bagworm Eumeta minuscula (Butler 1881) (Lepidoptera, Psychidae) to leaves by means of silk threads. Larvae of bagworms typically cover themselves with a protective case of silk and plant materials, similar to the larvae of caddisflies. The case protects both the larvae and the pupae from predation and unfavourable abiotic conditions . Silk is furthermore used for aerial dispersal (ballooning) in the freshly hatched larvae [32,33]. During pupation, the bags are suspended on overhanging surfaces to avoid predation. We were interested, in particular, how the bagworms achieve strong anchorage by choosing their attachment sites on the leaves and how they form the bonding. Furthermore, we determined the perpendicular pull-off forces of the cases from their substrate, in order to reveal how strong the attachment is. The aim was to contribute to the exploration of the means of silk thread attachments present in arthropods.
2.1. Attachment disc structure and glue–leaf interaction
The cases of E. minuscula consist of a silk lining with attached plant pieces, and house the caterpillars and pupae. Cases are 20–30 mm in length. They are suspended on the lower leaf surfaces, and attached by a short bundle of silk threads, that unravels near the leaf substrate and forms a 2–3 mm wide patch (attachment disc) of looped agglutinated threads that radially spread from the stalk (figure 1). The silk threads consist of a pair of fibres (figure 2c), of which each one has an approximate diameter of 2 µm. The glue occurs in local patches, in which the fibres are embedded (figure 2b,e–g), suggesting that it is added after the silk extrusion.
Of 80 inspected attachment points, 70 were on pre-damaged leaf surfaces and all of them were attached to the middle vein of the leaf. scanning electron microscope (SEM) analysis of silk attachments on pre-damaged and intact leaf surfaces revealed that the glue coat is repelled by the surface of the leaf cuticle. This is indicated by the observation that the previously attached threads were often lifted and detached from the plant substrates (figure 2e–g), which was not observed in pre-damaged leaf surfaces. In pre-damaged leaf surfaces, the cuticle and epidermis of the leaf was removed in large parts and most of the threads were attached to the exposed tissue (figure 2a), which was locally strongly wetted by the glue (figure 2b,c).
2.2. Pull-off resistance
When pulling the bagworm pupae perpendicular off the leaf surfaces, we recorded a relatively constant rise in forces, with small short drops (figure 3a), presumably caused by single threads failing or by fast crack propagation of a part of the attachment disc contact area. The maximal forces ranged between 0.60 N and 3.07 N (1.36 ± 0.85 N, mean ± s.d., n = 21). The bagworms had a weight of 0.40 ± 0.15 g (n = 41), so the resulting safety factor (the pull-off force divided by the body weight) of attachment is 385.6, on average. The attachment failed either at the bonding (10 out of 21 tests), leading to detachment from the leaf surface, or at the stalk (11 out of 21 tests), which led to thread bundle rupture. The tests, in which stalk rupture occurred, yielded a median maximal pull-off force of 1.14 N, and the ones with bonding failure 0.85 N (figure 3b); however, the difference is not significant due to an overall high variation in maximal pull-off forces (Mann–Whitney rank sum test, p = 0.597).
Many leaves exhibit wax layers that prevent formation of a proper bonding to various attachment devices of insects [26–28]. Insects that glue eggs or cocoons onto such surfaces must have evolved counter-strategies to gain a strong bonding. Asparagus beetles, for instance, deploy a surfactant-like proteinaceous secretion that mixes with the waxes and forms a tough composite after solidification . Simuliid flies have been reported to perform scraping and tugging movements with their mouthparts on the substrate before attaching their silk to build a cocoon . This was interpreted as a cleaning procedure , but may also roughen the substrate. Similarly, the bagworm larvae studied here damage the leaf surfaces in preparation of their cocoon attachment. They remove a part of the leaf epidermis and expose the underlying tissue.
Our SEM investigations revealed that the glue coat of the bagworm silk fibres cannot spread very well on the intact leaf surface, and detaches easily, whereas good wetting is observed on the damaged leaf surface. The bagworms preferably choose the main leaf vein as an attachment site, presumably because it offers higher mechanical stability. These parts of the leaf are not only substantially thicker. The tissue in these parts is denser than the mesophyll of the plain leaf sections. In those, the apparent contact area of the silk pad would be reduced, because of lots of cavities present after the removal of the epidermis, due to the spongy structure of the mesophyll. The silk threads presumably cannot follow these high amplitude surface irregularities (figure 2d). Local damaging of the plant tissue also suppresses the growth or wax secretion into the attachment site, which potentially could weaken the bonding over time.
The yielded pull-off strength is enormous and corresponds to a safety factor of 151 up to 771. This ensures a reliable bonding even under strong mechanical impact such as heavy rain and strong winds. During pull-off tests, the silk attachment failed either at the interface to the substrate or in the suspension thread bundle. Both cases occurred evenly and did not result in different pull-off forces, which indicates that the adhesion to the plant is comparable with the strength of the silk bundle structure. This may indicate that the mechanical properties of silk and glue are highly balanced and adapted to the physical properties of the host plant tissue.
In spiders, there have been similar observations of leaf surfaces repelling the glue coat of attachment silk, leading to reduced attachment strength . Spiders have to attach their threads frequently and very quickly during both locomotion and the building of complex webs. Therefore, spiders presumably do not have time for a pre-treatment of the surface. Spiders circumvent this problem by spinning hierarchical silk anchorages that yield a high safety factor and interfacial fracture toughness by their load-distributing structure [11,19]. This ensures safe attachment even on strongly repellent surfaces [11,17]. Furthermore, dense brushes of setae on the spinnerets of spiders may serve to clean surfaces from loose particles, but this has not been studied to date.
The silk anchorages of bagworms somewhat resemble the structure of spider silk anchorages: although the threads are directly applied onto the substrate, not by auxiliary glue fibres, they are similarly forming a patch-like attachment disc consisting of looped, agglutinated threads, that are bundled in the median part. Because the tensile load acts simultaneously on numerous independent fibres in such a structure, the failure of single elements does presumably barely affect the integrity of the whole attachment disc. This is indicated by short local drops frequently occurring during constant pull-off, after which the original forces are quickly recovered and further rise (figure 3a). Such a flaw tolerance has similarly been observed in spider attachment discs, and is regarded as a result of the heterogeneity of the contact . Furthermore, a structure consisting of a medially attached stalk and (more or less) radially spread fibres in the adhesive plaque leads to a concentric stress distribution in the adhesive contact under tension, which reinforces the peel-off resistance during delamination [15,20,36,37].
Bagworms presumably need a considerable amount of time to produce such an attachment disc, because multiple loops of single threads are overlaid subsequently (as it has been reported for other caterpillars [21,22]). In spiders, attachment discs can be produced very quickly, because multiple gluey threads simultaneously emerge from multiple spigots on the spinneret [11,15,19,38]. The safety factors of such instant dragline anchorages are much smaller than those recorded here for the bagworms, ranging between 1 and 8 on sycamore leaves . This may indicate that attachment discs on non-damaged leaf surfaces are less robust. These differences are, further, related to the differences in the biological function and different investments in single silk attachments, with their highly dynamic use in spiders and a long-term attachment in bagworms.
Our observations show that in insects there are structural and behavioural adaptations to attach silk to plant surfaces. In contrast to spiders, bagworms pre-treat the attachment substrate to enhance the bonding. The bagworm attachment discs share some intriguing similarities with those of spiders, such as a mushroom-like architecture, and gain enormous attachment strength. For the animal this is crucial, because failure of the bonding, especially during the immobile pupal stage, would be fatal. This is a starting point to investigate common principles of fibre-based attachment systems in arthropods, to gain a better understanding for the biological function and evolution of silks, and to generate biomimetic ideas for the quick, non-invasive technical attachments of cable-like structures to substrates.
5. Material and methods
Pupae and last instar caterpillars of the tea bagworm E. minuscula were collected from Ilex chinensis in Nanjing, China, by cutting off the leaves with the attached bags. For later identification, some individuals were kept in the laboratory until the moths hatched. Identification followed Hirowatari et al. , Sauter & Hättenschwiler  and Inoue et al. . The weight of individual animals was obtained using an AL204 electronic microbalance (Mettler-Toledo Group).
5.2. Force measurements
For the force measurements, a custom-made force transducer was used, which functioned as an array of cantilever beams in a metal T-shaped arrangement . Forces were measured from the deflection of the cantilever beams by foil strain gauges glued to the beams. The resolution for each direction was similar (approx. 1–2 mN). The fundamental resonance frequency for the x- and y-axis was lower (approx. 125–138 Hz) than for the z-axis (approx. 350 Hz). The data were logged, using a data acquisition platform (PXIe-8130, National Instruments, USA) at a sampling rate of 1000 Hz.
Only bagworms attached to the middle leaf vein were taken for measurements. Pull-off forces were measured within 1 h after material collection in the field. Small pieces of leaves with a diameter of about 15 mm, containing the attachment disc, were cut off and mechanically attached to the three-dimensional force sensor platform using thin metal plates and screws (figure 4). The distance between clamps (i.e. non-fixed middle part of the leaf) was 7 mm. The bag of the bagworm was pulled manually with constant speed of approximately 1 mm s−1 from the platform at an angle between 75° and 90° until the attachment disc separated from the leaf or the fibre bundle ruptured. The three-dimensional force data were recorded using an amplifier and data acquisition system. From the three force components, the pull-off force was calculated. A statistical analysis was performed, using SigmaStat (Systat Software Inc., San Jose, CA, USA). The Mann–Whitney rank sum test was used to test between forces reached at the two different failure modes (complete detachment versus thread rupture).
The cocoons were cut off the leaves at the level of their stalks, and the leaves with attachment discs were air dried. Then, desired pieces of the leaves with attachment discs were cut off and attached to the aluminium stubs using conductive carbon-containing double-adhesive tape. The samples were sputter coated with gold–palladium with a thickness of 10 nm (B7341, Agar Scientific Ltd, Stansted, Essex, UK) and studied in the SEM (Sigma FESEM, Carl Zeiss AG, Oberkochen, Germany) at an accelerating voltage of 5 kV.
S.N.G. conceived and designed the study. S.N.G., Z.D., A.J., Z.Z. and N.J. performed the experiments and analysed the data. J.L. determined the moth species used. E.G. determined the plant species used. J.O.W. and S.N.G. wrote the paper, and all authors contributed to revision.
We declare we have no competing interests.
This work was funded by the Friendship Award of the Jiangsu Province, China to S.N.G. and a Macquarie Research Fellowship of Macquarie University, Sydney to J.O.W. This study was also supported by the National Natural Science Foundation of China (51375232) and the Natural Science Foundation of Jiangsu Province (BK20141410), which were both awarded to A.J.
We thank Huan Wang (Nanjing University of Astronautics and Aeronautics, China) for assistance during experiments.
- Received January 4, 2017.
- Accepted February 7, 2017.
- © 2017 The Author(s)
Published by the Royal Society. All rights reserved.