Influence of surface free energy of the substrate and flooded water on the attachment performance of stick insects (Phasmatodea) with different adhesive surface microstructures Free

The ability of organisms to attach to and detach from surfaces is often essential for their survival. Attachment systems used for terrestrial locomotion evolved convergently in different animal lineages; for example, in squamates, amphibians and arthropods (Gorb, 2001; Scherge and Gorb, 2001; Persson, 2003; Federle et al., 2000; Federle, 2006; Federle et al., 2006; Büscher et al., 2019; Büscher and Gorb, 2021). Organisms of great interest for this study are Phasmatodea, colloquially known as stick insects. They are terrestrial herbivorous insects that inhabit a high diversity of habitats worldwide (Bedford, 1978; Brock et al., 2022). These insects are strongly adapted to plants, as visible in their cryptic mimicry, i.e. their leaf- and twig-like appearance, and presumably co-evolved with them (Wang et al., 2014; Bank et al., 2021a,b). This co-evolution shaped the attachment systems of phasmids, enabling them to attach to a wide variety of plant surfaces (Büscher et al., 2018a,b, 2019; Büscher and Gorb, 2019).

Their tarsal attachment system consists of three main units. The first unit is the euplantulae, which, depending on the species, are situated on the first to fourth or first to fifth tarsomeres (Beutel and Gorb, 2006). They are covered with a diversity of different species-specific adhesive microstructures (AMs) (Büscher et al., 2018a,b, 2019). It has been previously assumed that these microstructures evolved as adaptations to the substrates the animals are exposed to (Büscher and Gorb, 2017; Büscher et al., 2018a; Cumming et al., 2021). The euplantulae as such enable a trade-off between generating stationary attachment and propulsion for locomotion (Gorb et al., 2000; Jiao et al., 2000; Federle, 2006; Bußhardt et al., 2012; Büscher et al., 2018a).

The second unit involved in attachment is the two pretarsal claws, used for interlocking with rough surface asperities, and an arolium in between them as the third unit. The latter is rather soft and used to increase the contact area to the surface and by this to enhance the attachment forces (Clemente and Federle, 2008). Phasmids rely on ‘wet adhesion’, meaning that they produce an adhesive fluid that influences attachment (Dirks et al., 2010). Hermarchus leytensis has been shown to possess pores on the pads (Gottardo and Vallotto, 2014), through which the fluid is secreted. It then interacts with the surface and increases the contact area, resulting in stronger adhesion.

In general, the forces involved in insect adhesion can be divided into two groups: those that arise during dry interactions – the van-der-Waals forces, chemical bonds and friction; and those caused by the adhesive fluid produced by insects – the capillary and viscous forces (Stefan, 1875; Johnson, 1998; Dirks, 2014; Ditsche and Summers, 2014; Popov, 2016; Büscher and Gorb, 2021; Meng et al., 2021). Chemical analysis of the adhesive fluid of insects indicated an emulsion composed of rather different constituents, such as unipolar carbohydrates, fatty acids, polar proteins and peptides, resulting in stronger adhesion to substrates with different surface energies and polarities (Attygall et al., 2000; Vötsch et al., 2002; Dirks et al., 2010; Dirks and Federle, 2011; Betz et al., 2018). Studies investigating the components of adhesive secretions have shown that they can differ greatly between groups of insects (Attygall et al., 2000; Vötsch et al., 2002; Dirks et al., 2010; Dirks and Federle, 2011; Betz et al., 2018). These differences in composition can therefore influence the attachment of the insects.

Attachment of phasmids is constrained by many factors, for example roughness, mechanical and chemical properties and inclination of the surface, of which some have been investigated already. Studies regarding the effect of surface roughness on phasmid attachment have shown that the AMs determines the attachment performance on different rough surfaces (Bußhardt et al., 2012; Büscher and Gorb, 2019). The animals examined in this study possess smooth (Medauroidea extradentata), nubby (Sungaya inexpectata) or hexagonal (Epidares nolimetangere) AMs. Species with smooth AMs on their euplantulae generate stronger adhesion on smooth substrates compared with species with nubby AMs, which perform better on micro-rough surfaces (Bußhardt et al., 2012; Büscher and Gorb, 2019). The function of hexagonal AMs has not been previously studied in phasmids, but has in other animals and artificial systems (Federle et al., 2006; Varenberg and Gorb, 2009; Drotlef et al., 2013).

Another factor of substrates that influences adhesion is the surface chemistry, measured by the surface free energy (SFE). Substrates with a high SFE are hydrophilic, have a low contact angle (θ<90 deg) with water and therefore show a good wettability of the surface. A low SFE indicates that the substrate is hydrophobic, has a high contact angle (θ>90 deg) with water and shows poor wettability of the surface. Plants are particularly important for the herbivorous Phasmatodea as substrates for attachment. Along with other purposes, low SFE of plant surfaces is one strategy to avoid attachment of herbivorous insects (Gorb and Gorb, 2009). Plant surfaces often have a low SFE, whereby it has an influence on attachment (Voigt and Gorb, 2009; Gorb and Gorb, 2017). Some insectivorous plants exploit the interaction between high SFE and a humidity-induced water film on the attachment. This has been shown by Labonte et al. (2021) for the surface of the pitcher plant Nepenthes fusca, which has a high SFE, enabling it to effectively hold a water film to strongly reduce the attachment forces of insects. Insect anti-adhesive effect arises from the combination of surface chemistry and texture, but it is primarily driven by the substrate roughness, and less by the surface chemistry (England et al., 2016).

In the present study, we investigated the influence of three different substrates with different SFE (glass, silanized glass and Teflon) on the pull-off and traction forces of three stick insect species with different AMs. Furthermore, we analyzed the pull-off and traction forces on all three substrates with and without water film to investigate the effect of water presence and SFE regarding the attachment performance. Specifically, we asked the following questions. (i) How does SFE of the substrate influence the pull-off and traction forces of the three stick insect species? (ii) How does the water film influence the pull-off and traction forces? (iii) How does the combination of specific SFE and the presence of the water film affect the attachment performance of the three species? (iv) How do pull-off and traction forces differ between the species with nubby, hexagonal and smooth AMs on their euplantulae?

The species in this study were selected according to the AMs on their euplantulae (see Büscher et al., 2019). Sungaya inexpectata Zompro 1996 (Fig. 1A) possesses nubby AMs. The euplantulae of Epidares nolimetangere (de Haan 1842) (Fig. 1B) exhibit hexagonal patterned (plateaus) AMs. Medauroidea extradentata Brunner von Wattenwyl 1907 (Fig. 1C) has smooth AMs (Fig. 2). All individuals used were obtained from the laboratory cultures of the Department of Functional Morphology and Biomechanics (Kiel University, Germany). The insects were fed with blackberry leaves ad libitum and kept on a regular day and night cycle. Only adult individuals with all six legs intact were selected. The adhesive system of all individuals was tested on a glass plate prior to the experiments: if they were unable to show any pull-off forces, they were not selected for further experiments.

All specimens were attached to a force transducer (S. inexpectata and M. extradentata: 100 g capacity, FORT100, World Precision Instruments Inc., Sarasota, FL, USA; E. nolimetangere: 25 g capacity, FORT25, World Precision Instruments Inc.) by gluing a hair onto the metanotum of the stick insects with beeswax. The force transducers were connected to a BIOPAC Model MP100 and a BIOPAC TCI-102 System (BIOPAC Systems, Inc., Goleta, CA, USA). Force–time curves were recorded with AcqKnowledge 3.7.0 software (BIOPAC Systems Inc.).

The highest peak of the obtained graph was considered as the maximum attachment force for further processing (Fig. 3C,D). For the recording of the traction forces, the force transducer was aligned horizontally at the same height as the insect. The substrate was then pulled back from the animal (Fig. 3B). For measuring the pull-off force, the force transducer was positioned above the insect and the substrate was pulled away from the animal in the direction orthogonal to the sensor (Fig. 3A). In both experiments, the surface was moved manually at a steady speed with a winch; the speed for the pull-off force measurements was around 0.5–1.0 cm s⁻¹ and that for the traction force measurements was around 1.0–3.0 mm s⁻¹. The irregular increases of the traction and pull-off forces prior to the maximum forces (Fig. 3C,D) represent the normal stick and slip response of the stick insects (Büscher et al., 2020).

For each species, pull-off and traction force measurements were conducted. A total of six different substrates were used: three surfaces with different SFE (see below), each of them in the dry condition and with a thin water film. The measurements were performed with the same individuals per species (S. inexpectata N=13, E. nolimetangere N=24, M. extradentata N=11), which were measured 3 times in a row. The mean value of the three measurements was calculated to reduce the intra-individual variance. The sequence of individuals was randomized for every substrate. In order to hold a stable thin water film on the surface for the wet measurements, an additional border of dental wax was added (Fig. 3A,B). This edge produced a water film (approximately 2–3 mm thick) that covered the entire surface. During the measurements on the surfaces with a thin water film, it was always ensured that the entire adhesive organ of the phasmid was covered by water. Ditsche and Summers (2014) describe such conditions as immersed conditions.

The sex of the specimens was kept the same per species, to ensure the same AMs and body proportions within the groups. For S. inexpectata (mean±s.d. body mass 2.79±0.88 g) and M. extradentata (body mass 1.10±0.04 g), females were used, whereas for E. nolimetangere (body mass 0.78±0.08 g), males were used. Ambient humidity in the laboratory was around 45–58% at a temperature of 21–23°C.

All individuals were tested in all treatments. The individuals did not show any signs of exhaustion during the series of measurements. Each individual contributed equally to the experiments.

A glass surface (G) was used as the hydrophilic surface with high SFE and another glass surface was silanized (S), as described by Voigt and Gorb (2009), to decrease the SFE of the surface. The third surface used was Teflon (T) as a low SFE surface. The physico-chemical properties of the surfaces were characterized by measuring the contact angle of water on the substrate (aqua Millipore, droplet size 1 µl, sessile drop method; n=10 per substrate) using the contact angle measurement instrument OCAH 200 (Dataphysics Instruments GmbH, Filderstadt, Germany). The contact angle was 44.17±7.74 deg for glass surface, 71.69±8.02 deg for the silanized glass surface and 107.89±1.22 deg for Teflon (Fig. 3E,F). The surfaces with a continuous water film are referred to below as glass with water film (GW), silanized glass with water film (SW) and Teflon with water film (TW).

The arithmetic average roughness (R_a) of Teflon was measured with a 3D measurement microscope (Keyence VR 3100, KEYENCE, Neu-Isenburg, Germany). The roughness profile of 10 lines with an approximate length of 1 mm in the horizontal and vertical direction was evaluated.

The statistical analyses were performed using SigmaPlot 12.0 (Systat Software Inc., San José, CA, USA). The data were tested for normal distribution (Shapiro–Wilk test) and homoscedasticity (Levene's test) before subsequent tests. Pull-off and traction forces on the different surfaces with and without water film for each species were compared using Kruskal–Wallis one-way analyses of variance (ANOVA) on ranks and Tukey's post hoc test, as the data were neither normally distributed nor showed homoscedasticity.

Differences between the three species were analyzed with Kruskal–Wallis one-way ANOVA on ranks and Dunn's post hoc test. For all statistical comparisons, the significance level (alpha) was set to 0.05.

To determine the influence of the wet surfaces on the attachment, the attachment on the dry surfaces was assumed to be the maximum attachment (100%). The percentage loss of attachment was calculated by subtracting the remaining attachment (%) on the wet surfaces from 100%. This allowed us to calculate the percentage reduction between the attachment forces on the dry and the wet substrates.

Tarsi of S. inexpectata and M. extradentata were cut from adult females and tarsi of E. nolimetangere from adult males. They were fixed in 1 ml Bouin's fixative for 24 h on a shaker. Afterwards, the samples were dried in an ascending ethanol series, critical point dried (Leica EM CPD300, Leica, Wetzlar, Germany) and sputter-coated with a 10 nm layer of gold–palladium (Leica Bal-TEC SCD500). The samples were mounted on a rotatable specimen holder (Pohl, 2010) and the images were obtained using a scanning electron microscope (TM3000, Hitachi High-technologies Corp., Tokyo, Japan) at 15 kV acceleration voltage. Focus stacks of the part of interest were recorded. They were stitched, merged and processed using the software Photoshop CS5 (Adobe Systems Inc., San Jose, CA, USA).

All species studied possess a tarsus with five tarsomeres as well as a pretarsus with two claws and an arolium in between. The main difference between the species is in the AMs: S. inexpectata has euplantulae on tarsomeres 1–5, each with nubby AMs (Fig. 2A–D); E. nolimetangere possesses euplantulae with hexagonal AMs on tarsomeres 1–5 (Fig. 2E–H); and M. extradentata has smooth euplantulae which are located on tarsomeres 1–4 (Fig. 2I–L).

The measured impact of SFE and water on the pull-off forces of S. inexpectata with nubby AMs revealed significantly larger values on the dry surfaces (mean±s.d. G 104.15±48.14 mN, S 113.04±59.74 mN and T 78.18±36.16 mN) than on the surfaces with a water film (GW 9.36±15.23 mN, SW 37.39±22 mN and TW 16.76±18.28 mN) (Kruskal–Wallis one-way ANOVA on ranks, H=55.31, d.f.=5, P≤0.001, N=13). For the surfaces with a water film, only the pull-off forces between SW and GW were significantly different (Kruskal–Wallis one-way ANOVA on ranks, H=10.496, d.f.=2, P=0.005, N=13) (Fig. 4A).

The same impact was found for both SFE and water on the traction force (G 228.37±159.97 mN, S 148.67±102.11 mN, T 146.17±71.55 mN, GW 32.46±17.47 mN, SW 59.27±24.75 mN and TW 45.41±27.66 mN) (Kruskal–Wallis one-way ANOVA on ranks, H=8.384, d.f.=2, P=0.015; Kruskal–Wallis one-way ANOVA on ranks, H=53.297, d.f.=5, P≤0.001, N=13) (Fig. 4B).

For E. nolimetangere with hexagonal AMs, the impact of the SFE and water resulted in significantly higher pull-off forces on the two dry surfaces G and S than on all other surfaces (G 13.18±2.85 mN, S 14.80±5.64 mN, T 5.14±4.65 mN, GW 9.62±2.04 mN, SW 9.23±1.86 mN and TW 5.10±4.09 mN) (Kruskal–Wallis one-way ANOVA on ranks, H=90.685, d.f.=5, P≤0.001, N=24) (Fig. 4C). The value on GW was significantly higher than those on T and TW. Additionally, the value on SW was significantly higher than that on TW (Kruskal–Wallis one-way ANOVA on ranks, H=90.685, d.f.=5, P≤0.001, N=24).

The impact of SFE and water on the traction force also revealed significantly higher values on the dry surfaces G and S than on all other surfaces (G 50.68±19.35 mN, S 47.30±16.11 mN, T 20.94±10.35 mN, GW 12.40±6.43 mN, SW 23.25±7.12 mN and TW 6.63±3.87 mN) (Kruskal–Wallis one-way ANOVA on ranks, H=87.523, d.f.=5, P≤0.001, N=24, q=5.65) The traction forces on the surfaces GW, SW and T were significantly higher compared with that on TW (Kruskal–Wallis one-way ANOVA on ranks, H=87.523, d.f.=5, P≤0.001, N=24; Kruskal–Wallis one-way ANOVA on ranks, H=18.179, d.f.=2, P≤0.001, N=24) (Fig. 4D).

The impact of SFE and water on the pull-off force of M. extradentata (smooth AMs) revealed significantly higher values on the two dry surfaces G and S when compared with the wet surfaces (G 56.75±26.15 mN, S 61.11±12.26 mN, T 37.75±22.42 mN, GW 16.12±12.26 mN, SW 12.85±7.50 mN and TW 12.55±9.04 mN) (Kruskal–Wallis one-way ANOVA on ranks, H=38.597, d.f.=5, P≤0.001, N=11). No significant differences could be observed between the pull-off forces on T and the three wet surfaces (Fig. 4E).

The same results were obtained when comparing the traction forces (G 76.53±35.28 mN, S 78.60±29.47 mN, T 64.27±30.06 mN, GW 30.08±28.28 mN, SW 30.18±14.64 mN and TW 25.65 1±6.06 mN) (Kruskal–Wallis one-way ANOVA on ranks, H=30.127, d.f.=5, P≤0.001, N=11) (Fig. 4F).

The pull-off force safety factors of the species M. extradentata and S. inexpectata on the dry substrates G and S were significantly higher than the values for E. nolimetangere (S. inexpectata: G 3.92±1.74, S 4.26±1.90, T 3.06±1.75; M. extradentata: G 5.06±1.78, S 5.70±1.49, T 3.47±1.58; E. nolimetangere: G 1.69±0.36, S 1.89±0.74, T 0.70±0.64) (Fig. 5A). The values of S. inexpectata and M. extradentata on T and of E. nolimetangere on S were significantly different to the value of E. nolimetangere on T (Kruskal–Wallis one-way ANOVA on ranks, H=103.599, d.f.=8, P≤0.001; S. inexpectata N=13, E. nolimetangere N=24, M. extradentata N=11).

On the wet substrates, the pull-off force safety factors only showed significant differences between the values of S. inexpectata on SW and those of S. inexpectata on GW and E. nolimetangere on TW (S. inexpectata GW 0.48±0.74, SW 1.69±1.14, TW 0.60±0.67; M. extradentata GW 1.40±0.95, SW 1.56±1.36, TW 1.18±0.84; and E. nolimetangere GW 1.30±0.32, SW 1.18±0.15, TW 0.68±0.54) (Kruskal–Wallis one-way ANOVA on ranks, H=34.934, d.f.=8, P≤0.001, N=13/24/11) (Fig. 5A).

Comparing the performance of the species on the same substrates in dry and wet conditions showed that the values of S. inexpectata on G and T and those of M. extradentata on S were significantly different to those of their wet counterparts (Kruskal–Wallis one-way ANOVA on ranks, H=193.094, d.f.=17, P≤0.001, N=13/24/11).

All three species exhibited higher traction force safety factors on the dry surfaces than on the wet surfaces (S. inexpectata: G 8.57±4.88, S 5.59±4.15, T 5.83±3.67, GW 1.35±0.84, SW 2.57±1.20, TW 1.64±1.08; E. nolimetangere: G 6.65±3.01, S 6.17±2.52, T 2.92±1.74, GW 2.01±1.24, SW 2.64±1.38, TW 1.11±0.65; and M. extradentata: G 6.98±2.76, S 7.54±1.98, T 6.17±2.34, GW 2.74±2.79, SW 3.32±2.13, TW 2.40±1.40) (Fig. 5B), whereby only the values of S. inexpectata on G and T and the values of E. nolimetangere on G were significantly different to those for their wet counterpart (Kruskal–Wallis one-way ANOVA on ranks, H=172.383, d.f.=17, P≤0.001, N=13/24/11).

Comparing the dry surfaces, all values, except those of S. inexpectata on S and T, were significantly higher than the value for E. nolimetangere on T (Kruskal–Wallis one-way ANOVA on ranks, H=39.575, d.f.=8, P≤0.001, N=13/24/11).

For the wet surfaces, only the values of the three species on SW were significantly higher than the values of E. nolimetangere on TW (Kruskal–Wallis one-way ANOVA on ranks, H=32.172, d.f.=8, P≤0.001, N=13/24/11).

The negative effect of water on the pull-off force was stronger for M. extradentata with smooth AMs and S. inexpectata with nubby AMs than for E. nolimetangere with hexagonal AMs on all substrates. Medauroidea extradentata exhibited a percentage loss in pull-of force of 73.07% on G versus GW, 69.83% on S versus SW and 40.04% on T versus TW, followed by S. inexpectata with 60.57% loss on G versus GW, 51.91% on S versus SW and 72.02% on T versus TW. The hexagonal AMs of E. nolimetangere showed a loss of attachment performance on G versus GW of 23.11%, on S versus SW of 31.90% and on T versus TW of 3.52% (Fig. 6A).

The negative influence of water on the traction force on wet surfaces compared with dry surfaces was less pronounced than that for pull-off force among the three species. Sungaya inexpectata exhibited a percentage loss in traction force on G versus GW of 82.90%, on S versus SW of 46.75% and on T versus TW of 64.82%. Epidares nolimetangere showed a percentage loss in traction force on G versus GW of 63.52%, on S versus SW of 46.75% and on T versus TW of 45.80%. Medauroidea extradentata produced a percentage decrease of traction force on G versus GW of 57.44%, on S versus SW of 54.44% and on T versus TW of 24.12% (Fig. 6B).

As with the pull-off force safety factors, all species exhibited higher adhesive strength on dry surfaces than on wet surfaces (M. extradentata: G 10.72±4.94 kPa, GW 3.04±2.32 kPa, S 11.54±5.22 kPa, SW 2.42±1.42 kPa, T 7.13±4.23 kPa, TW 2.34±1,66 kPa; S. inexpectata: G 5.02±2.32 kPa, GW 0.48±0.76 kPa, S 5.45±2.88 kPa, SW 1.86±1.06 kPa, T 3.79±1.78 kPa, TW 0.81±0.88 kPa; and E. nolimetangere: G 3.10±0.67 kPa, GW 2.26±0.48 kPa, S 3.48±1.33 kPa, SW 2.17±0.44 kPa, T 1.14±1.12 kPa, TW 1.20±0.96 kPa).

In contrast to the mass-corrected data, the adhesion pad surface area-corrected data showed the following relationships. Between the dry surfaces, the values of S. inexpectata on G and S were not significantly different to the values of E. nolimetangere on G and S (Kruskal–Wallis one-way ANOVA on ranks, H=90.921, d.f.=8, P≤0.001, but Dunn's post hoc test P>0.05, N=13/24/11) (Fig. 7A). In contrast, the value of E. nolimetangere on G was significantly higher than the value of E. nolimetangere on T, and the value of M. extradentata on S was significantly higher than the value of S. inexpectata on T (Kruskal–Wallis one-way ANOVA on ranks, H=90.921, d.f.=8, P≤0.001, N=13/24/11).

When comparing the values on the wet surfaces, the performance of S. inexpectata on GW and TW differed noticeably between the mass-corrected and adhesion pad surface area-corrected data. For adhesive strength, only the values of S. inexpectata on SW and TW and of E. nolimetangere on TW were not significantly different to the value of S. inexpectata on GW. In addition, the values of E. nolimetangere on GW and SW and of M. extradentata on GW were significantly different to the values of S. inexpectata on TW, whereas the value of S. inexpectata on SW was no longer significantly different (Kruskal–Wallis one-way ANOVA on ranks, H=46.462, d.f.=8, P≤0.001, but Dunn's post hoc test P>0.05, N=13/24/11) (Fig. 7A).

Comparison between the performance of the species on the same substrates in dry and wet conditions showed that only the value of S. inexpectata on S was significantly different (Kruskal–Wallis one-way ANOVA on ranks, H=193.051, d.f.=17, P≤0.001, N=13/24/11) (Fig. 7A).

As with the traction force safety factors, all species exhibited higher shear strength on dry surfaces than on wet surfaces (M. extradentata: G 14.22±6.17 kPa, GW 5.68±5.34 kPa, S 14.84±5.56 kPa, SW 5.70±2.76 kPa, T 12.14±5.68 kPa, TW 4.84±3.03 kPa; S. inexpectata: G 11.01±7.71 kPa, GW 1.56±0.84 kPa, S 7.17±4.92 kPa, SW 2.86±1.19 kPa, T 7.05±3.45 kPa, TW 2.19±1.33 kPa; and E. nolimetangere: G 11.93±4.55 kPa, GW 3.46±2.18 kPa, S 11.13±3.79 kPa, SW 5.05±2.76 kPa, T 5.05±2.80 kPa, TW 1.92±1.08 kPa) (Fig. 7B).

When comparing the performance on the dry surfaces, only a few differences were detected. The performance of M. extradentata on S was significantly different to the values of S. inexpectata on S and T. The performance of S. inexpectata on G was not significantly different to the value of E. nolimetangere on T (Kruskal–Wallis one-way ANOVA on ranks, H=52.594, d.f.=8, P≤0.001, but Dunn's post hoc test P>0.05, N=13/24/11) (Fig. 7B).

On the wet surfaces, the performance of S. inexpectata on GW decreased, becoming significantly different from the values of E. nolimetangere on SW and M. extradentata on SW and TW. Other significant differences were measured between M. extradentata on SW and S. inexpectata on TW, and between M. extradentata on TW and E. nolimetangere on TW (Kruskal–Wallis one-way ANOVA on ranks, H=42.985, d.f.=8, P≤0.001, N=13/24/11) (Fig. 7B). No significant difference was measured between S. inexpectata on SW and E. nolimentagere on TW (Kruskal–Wallis one-way ANOVA on ranks, H=52.594, d.f.=8, P≤0.001, but Dunn's post hoc test P>0.05, N=13/24/11).

No significant difference was detected when comparing the performance of the species on the same substrates in dry and wet conditions (Kruskal–Wallis one-way ANOVA on ranks, H=177.160, d.f.=17, P≤0.001, N=13/24/11).

All three species were able to exert enough attachment force on all surfaces used in the experiments to hold more than their own weight (SF>1) (Fig. 5). Consequently, the tarsal attachment system of the three species can cope with the different SFE of the surfaces and the presence of water. Considering their habitats, this behavior was expected, as these insects are generally exposed to various natural substrate conditions (Heepe et al., 2017). Nevertheless, differences in the attachment performance were revealed among the three species, suggesting that the presence of water on the substrate and differences in the SFE of the substrate affect the three attachment systems to different extents.

The attachment of animals using wet adhesion on dry substrates is mainly influenced by the interaction between the adhesive system, the adhesive secretion and the substrate surface.

The animals studied in these experiments possess smooth adhesive pads. This implies that the inner structure of the euplantulae and arolium consists of fibrils that are hierarchically organized, giving the attachment system certain flexibility (Perez-Goodwyn et al., 2006; Peisker et al., 2013; Bennemann et al., 2014), which allows the attachment devices to adapt to the surface profile up to a certain amplitude of roughness (Persson and Gorb, 2003; Eimüller et al., 2008; Gorb, 2011). As a result of this flexibility and the smooth surfaces used in our experiments (Teflon surface measured vertically: R_a=5.1±1.14 µm; Teflon surface measured horizontally: R_a=7.14±0.80 µm), the influence of the claws on attachment can be excluded (Büscher and Gorb, 2019).

One important factor that can influence the attachment performance of the species investigated in this study is the difference in their AMs. Because of the different AMs, the actual area of the adhesive contact, as well as the distribution of the adhesive fluid to the substrate, can differ greatly between the species studied here.

The actual adhesive contact area of S. inexpectata is reduced compared to the projected pad surface area as a result of the nubby AMs (Fig. 2C). Although the contact area increases by the deformation of the nubs under load, it remains smaller than that of a smooth-to-smooth contact (Labonte et al., 2014). The nubby AMs can distribute the adhesive fluid efficiently, which supports contact formation and strengthens the adhesion process (Arnold, 1974; Labonte and Federle, 2013; Büscher and Gorb, 2019; Büscher et al., 2020).

One advantage of the hexagonal AMs of E. nolimetangere is the formation of channels that can distribute the produced adhesive secretion evenly along the surface (Fig. 2G). Furthermore, the contact splitting reduces crack propagation as the pad is peeled off and avoids stick-and-slip motion of the pads (Johnson et al., 1971; Arzt et al., 2003; Meng et al., 2021). This has been shown for other insect species (Varenberg and Gorb, 2009) and tree frogs (Hanna and Barnes, 1990; Ba-Omar et al., 2000; Tsipenyuk and Varenberg, 2014; Meng et al., 2021).

Because of the smooth and unstructured AMs of M. extradentata, the actual adhesive area is not reduced (Fig. 2K). This also indicates that the secretion and distribution of the adhesive fluid cannot be supported by an additional structure (Büscher and Gorb, 2021).

Comparing the adhesive performance of the three species with a focus on the mass-corrected data (safety factor; Fig. 5) and adhesive pad surface area-corrected data (adhesive and shear strength; Fig. 7) shows that all species behave approximately the same. In this case, a comparison of the attachment performance between the three species using mass-corrected data is probably a better proxy than using projected adhesive pad surface area-corrected data, as the actual contact area of structured attachment pads is smaller than the measured projected area.

The second important factor influencing the attachment performance is the distribution and composition of the adhesive pad secretion. The presence of fluids in contacts implements capillary forces and viscous forces, increasing attachment strength (Langer et al., 2004; Ditsche and Summers, 2014). Additionally, the fluid fills nanoscale asperities of the substrate, increasing the actual contact area (Gorb et al., 2000). The secretions of phasmids and cockroaches consist of hydrophilic proteins and peptides as well as hydrophobic carbohydrates (Dirks and Federle, 2011; Betz et al., 2018), which together form a two-phasic secretion consisting of the hydrophilic and hydrophobic parts. However, the fluid may increase the distance between the two solid surfaces and thus reduce the contribution of van der Waals forces (Stefan, 1875; Kendall, 1994; Gorb, 2001; Varenberg and Gorb, 2008; Popov, 2016).

The insect cuticle as well as the adhesive secretion can be influenced by the SFE of the substrate. The hydrophobic cuticle of the adhesive pads could adhere better to a low SFE substrate, as stronger hydrophobic–hydrophobic interactions may occur (Bush et al., 2008; Gundersen et al., 2014). Furthermore, the SFE of the substrate could also influence the effect of the adhesive fluid on adhesive performance, as hydrophilic components of the secretion can interact better with surfaces with high SFE and hydrophobic components can interact better with surfaces with low SFE (Roth and Willis, 1952; Dixon et al., 1990; Büscher and Gorb, 2021).

An interesting result is the low pull-off force safety factor performance of E. nolimetangere on dry surfaces (Fig. 5A). This species possesses the smallest surface area of the arolium when compared with the other two species (E. nolimetangere: 0.13±0.02 mm²; M. extradentata: 0.52±0.05 mm²; S. inexpectata: 0.91±0.13 mm²) (Fig. 2), which is most important for pull-off force generation (Labonte and Federle, 2013, 2015; Labonte et al., 2016), and thus provides the smallest potential adhesive area. Another result that stands out is the significant reduction of the pull-off force of E. nolimetangere on T (Fig. 5A), which could be explained by a reduction of the hydrophobic components of the pad secretion, in comparison to the other two species, whereby the secretion has a reduced influence on the attachment.

Medauroidea extradentata produces the highest pull-off force safety factors (Fig. 5A) as a consequence of the smooth unstructured AMs in combination with the medium-sized adhesive pad (M. extradentata: 5.30±0.82 mm²; E. nolimetangere: 4.25±0.71 mm² and S. inexpectata: 20.74±3.20 mm²). That smooth adhesive structures show good adhesion to smooth surfaces has also been shown by other studies with Phasmatodea and other insect species (Drechsler and Federle, 2006; Bußhardt et al., 2012; Betz et al., 2017; Büscher et al., 2018a).

Similar effects of smooth substrates with different SFE on stick insect adhesion have been shown previously (Labonte et al., 2021). Their studied species Carausius morosus has nubby AMs similar to those of S. inexpectata. The substrates used there had water contact angles of 101, 16 and 5 deg. Labonte et al. (2021) showed that a decrease in contact angle correlates with a reduction in the attachment forces, but the attachment performance was still considerable. Our measurements with S. inexpectata showed a similar tendency (Fig. 5).

Experiments on the effect of SFE on the adhesion and traction of tree frogs that possess hexagonal AMs (Osteopilus septentrionalis and Ranoidea caerulea) showed that a decrease in SFE analogous to that used herein, i.e. from glass to silanized glass and further to Teflon, does not influence the generated forces significantly (Crawford, 2016). Epidares nolimetangere, the species with hexagonal AMs, displayed the same results on glass and silanized glass (Fig. 4C,D). However, the low SFE on Teflon resulted in significantly reduced pull-off forces and traction forces in this species of stick insect. This difference in attachment could be caused by the greater roughness of Teflon (Büscher and Gorb, 2019) and by a possible negative influence of the SFE on the adhesive secretion, potentially reducing its contribution to attachment.

The significant reduction of the adhesion performance of all three types of AM on nearly all surfaces with a water film compared with the dry counterpart can be attributed to the negative effects of water on attachment (Figs 4–7).

In the case of flooded substrates, water has a negative influence on attachment forces (Labonte et al., 2021). Water tends to behave like a separating film between the two solid surfaces and increases the distance between them. Consequently, the van der Waals forces and the chemical bonds are significantly reduced (Hosoda and Gorb, 2012). Additionally, the capillary forces are reduced as a result of the surface tension approaching zero, as it is negatively affected by the thickness of the water film (Vogel, 2003; Gorb, 2001; Ditsche and Summers, 2014). In contrast, viscous forces are increased as a result of the higher viscosity of water compared with air (Stefan, 1875; Ditsche and Summers, 2014).

In addition to the physical forces, water could also influence the function of the adhesive secretion by dissolving its hydrophilic components and thus reducing their contribution to attachment. It is likely that the hydrophobic components of the adhesive secretion are not affected by the water, which indicates that their share in attachment is not influenced (Dirks and Federle, 2011; Betz et al., 2018; Büscher and Gorb, 2021).

To reduce the negative influence of water on attachment, the species try to displace it from the contact surface. Thus, a contact surface is created which is affected least by water. Different strategies have evolved that solve this problem, which are represented in this study by the different AMs (Fig. 2C,G,K) (Hosoda and Gorb, 2012; Büscher et al., 2018a; Büscher and Gorb, 2021).

The nubby AMs of S. inexpectata probably penetrate the water film to come into direct contact with the substrate (Arzt et al., 2003; Spolenak et al., 2005; Frost et al., 2018). In addition, it is likely that the water is directed into the hollow cavities created by the nubby structure to remove it from contact. This would further increase the real contact area and thus strengthen the attachment, explaining why the nubby AMs of S. inexpectata showed an intermediate effect on the percentage loss of attachment performance on wet surfaces (Fig. 6).

Previous research on hexagonal attachment structures has shown that they are most efficient in coping with water on surfaces (Barnes et al., 2006; Varenberg and Gorb, 2009; Drotlef et al., 2013; Endlein et al., 2013; Crawford, 2016; Meng et al., 2021; Tramsen et al., 2021). The channels on the pad surface can very effectively displace the water from the contact area, thus considerably increasing the real contact area. This is supported by our pull-off force experiments, as E. nolimetangere showed the lowest percentage loss of adhesion on the wet surfaces (Fig. 6A). However, the results with Teflon must be considered with caution, as the low adhesion already present on Teflon cannot be drastically affected by water (Figs 4C and 6A). What is noticeable, however, is that even on such an unfavorable surface, E. nolimetangere still produced enough adhesion to hold more than their own weight (Fig. 5A).

Medauroidea extradentata possesses smooth AMs and therefore would be expected to show the weakest performance on wet surfaces, as the flat, unstructured AMs cannot displace or penetrate the water film as effectively as those of the other two species. While the percentage loss of attachment indeed showed that M. extradentata produced the lowest attachment (Fig. 6A), it was observed that the attachment, once the pad had made contact, was still sufficient for the insects to hold more than their own weight (Fig. 5A).

The effects of substrate wetness on the attachment performance have also been studied in other insect species. Stark and Yanoviak (2018) showed that wetness influences the normal (pull-off) and shear (traction) adhesion of the smooth tarsal attachment system in the tropical arboreal ant Cephalotes atratus. Wetness reduced the normal and shear adhesion significantly on glass with a high SFE, but the normal adhesion on the surfaces with low SFE made of polycarbonate and polypropylene was unaffected (Stark and Yanoviak, 2018). Medauroidea extradentata, the stick insect with smooth AMs, showed the same results on glass, but the pull-off forces on the lower SFE silanized glass were significantly reduced (Fig. 4E). This discrepancy may be due to the fact that the two species represent different lineages of insects, which could have a different composition of the adhesive secretion and, consequently, different physico-chemical interactions between the pad material, secretion, water and substrate surface.

Labonte et al. (2021) have shown that both wetness and SFE influence the generated friction forces of Carausius morosus. The stick insect produced significantly reduced friction forces on wet smooth surfaces with high SFE compared with dry smooth surfaces with low SFE. Additionally, no significant difference was recorded when the wet and dry surfaces had a low SFE (Labonte et al., 2021). Sungaya inexpectata, which also possesses nubby AMs, like C. morosus, showed nearly the same results: both species displayed significantly reduced traction forces on wet surfaces with high SFE compared with those on dry surfaces with low SFE. However, while C. morosus showed similar traction forces on wet and dry surfaces with low SFE, S. inexpectata exerted significantly reduced friction forces on wet Teflon compared with dry Teflon (Fig. 4B). The difference between the two species may be due to differences in the composition and thickness of the adhesive secretion.

Differences in performance between the pull-off and traction force experiments in all three species were detected (Figs 4–6). During the traction force experiments, the three species showed a more similar performance than during the pull-off force experiments. This could be explained by the fact that traction increases the effectiveness of water displacement from the contact surface regardless of the AM, because the application of traction forces mechanically removes water from the interface more efficiently (Peachey et al., 1991).

This study investigated and compared the effects of water films and surface chemistry on the attachment forces of three stick insect species with different AMs on their pads. We found the following: (1) the decrease in SFE does not significantly influence the pull-off forces and traction forces of the three stick insect species; (2) the presence of water reduces the pull-off and traction forces in all three species; and (3) all three species generate enough attachment force on all surfaces to adhere (SF>1). The AMs influence the effectiveness of attachment generation underwater. The AMs of E. nolimetangere had the strongest effect on attachment, the nubby AMs of S. inexpectata a medium effect and the smooth AMs of M. extradentata the least effect.

This study sheds light on how stick insects have adapted to varying environmental substrate conditions during their evolution. It shows how attachment pads with functional surface microstructures cope with such conditions. Nevertheless, many questions remain regarding the influence of the AMs and third bodies on the attachment performance, such as the ability to displace water, the generation of actual contact underwater, contamination and the effect of the adhesive secretion. Therefore, further investigations of such effects on attachment would be interesting. Furthermore, it would be intriguing to investigate how the actual distribution of the adhesive secretion on the AMs supports the function of each attachment system studied here.

We thank Esther Appel and Dr Alexander Kovalev (Department of Functional Morphology and Biomechanics, Kiel University Germany) for their technical assistance.