Adults of A. bifasciatus of both sexes were supplied by the commercial insectary Bioplanet, Cesena, Italy. Insects were maintained in the laboratory in vials in an environmental chamber (14 h photoperiod, temperature of 23°C ± 1°C, and relative humidity of 60% ± 10%) and were fed with small drops of honey on plastic strips. Adults of both sexes were used in the experiments. Ten females and ten males of A. bifasciatus were weighted with the Sartorius MSE2.7S ultra microbalance.

Eggs of H. halys were supplied by the commercial insectary Bioplanet, Cesena, Italy. Eggs of N. viridula were obtained from insects reared in the laboratory of the Department of Scienze Agrarie, Alimentari e Ambientali of the University of Perugia (Italy). N. viridula was reared in an environmental chamber (14 h photophase, temperature of 25°C ± 1°C; RH of 70% ± 10%) inside net cages (300 mm × 300 mm × 300 mm) (Vermandel, Hulst, Netherlands). Sunflower [Helianthus annus L. (Asteraceae)] seeds, pods of French bean [Phaseolus vulgaris L. (Fabaceae)] and cabbage [Brassica oleracea L. (Brassicaceae)] leaves were used to feed nymphs and adults.

To measure the arolium width, the pretarsi of dried specimens of both sexes were soaked in a 5% solution of lactic acid for 24 h to promote arolium expansion (Gladun and Gumovsky, 2006). The soaked pretarsi with spread arolia were then washed in water and fixed in 5% glutaraldehyde solution for 10–15 min, rinsed in water and subsequently mounted in glycerine. Arolia were observed and photographed with a microscope Leika DMLB connected to a Koppace FHD Camera V 2.0. Pictures were analysed using the software ImageJ (Sun-Java, United States). Legs of A. bifasciatus males and females were observed and photographed with the same microscope and camera.

The tarsi together with the pretarsal attachment devices (arolium and claws) of both sexes of A. bifasciatus and the eggs of H. halys and N. viridula were studied in a scanning electron microscope (SEM) Hitachi S-4800 (Hitachi High-Technologies Corp., Tokyo, Japan), equipped with a Gatan ALTO 2500 cryo-preparation system (Gatan Inc., Abingdon, United Kingdom). For details of sample preparation and mounting for cryo-SEM, see Gorb and Gorb (2009). Whole mounts of insect tarsi and eggs were sputter-coated in frozen conditions with gold-palladium (thickness 10 nm) and examined at 3 kV acceleration voltage and temperature of −120°C at the cryo-stage within the microscope.

The behaviour of the female of A. bifasciatus ovipositing on eggs of H. halys has been observed and recorded using a stereomicroscope Leica MZ6 connected to a high-speed camera Imaging source DMK37BUX287. Three egg-laying females were observed and each one was recorded ovipositing on at least five individual eggs. The portion of the tarsi of forelegs, midlegd and hindlegs in contact with the egg chorion during the different oviposition steps was observed.

Artificial substrates represented by polishing paper with different asperity sizes (average diameter of the abrasive particles −0.3, 1, 3, 9, 12, 35 and 125 µm) and glass with different wettability [hydrophobic glass with a water contact angle of 112.62 ± 3.13° (mean ± SD) and hydrophilic glass with a water contact angle of 32.49 ± 4.17° (mean ± SD)] were used in the centrifugal force tester experiments. Hydrophobic glass has been prepared by treatment of a glass plate by vapour phase deposition silanization method using dichlorodimethyl-silane. A glass disk was rinsed and sonicated in an ultrasound bath (Bandelin electronic, Berlin, Germany) for 5 min with ethanol (70%) and distilled water. This was repeated several times and remained liquid on the substrate was blown by compressed air. Cleaned glass was subjected to air plasma treatment (Diener electronic, Ebhausen, Germany) and then it was vacuum pumped (BÜCHI Labortechnik, Flawil, Switzerland) in a desiccator together with a glass vial containing 200 μl of dichlorodimethylsilane (Merck Schuchardt, Hohenbrunn, Germany). The vacuum pump was disconnected from the desiccator once the silane started to boil, and the desiccator was left closed for 5 h to achieve sufficient deposition of the silane on the glass. The surface-modified glass was rinsed thoroughly with isopropanol and distilled water and was blown dry by compressed air.

In the Biopac force tester experiments, a series of artificial substrates mimicking the host insect egg clutches were prepared using glass beads having different diameters (0.53 ± 0.03, 0.85 ± 0.02, 1.35 ± 0.02, 1.98 ± 0.04, and 3.06 ± 0.01 mm) and a water contact angle of 92.02 ± 1.03° (mean ± SD). Similar substrates as glass dummies of the host eggs were used in other behavioural investigations (Conti et al., 2003). Groups of glass beads with the same diameter were glued (Super Attack, Loctite Henkel Adhesives, Düsseldorf, Germany) on a glass microscope slide forming a single layer to mimic an egg clutch. An untreated microscope slide was used as flat hydrophilic glass substrate [water contact angle of 27.02 ± 0.64° (mean ± SD)].

The roughness of different polishing paper with asperity sizes of 0.3, 1, 3, 9, 12, 35 and 125 μm was characterized using the white light interferometer (Salerno et al., 2017). Roughness data are reported in Salerno et al. (2022).

The wettability of surfaces used in the experiments (smooth hydrophilic glass, smooth hydrophobic glass, glass beads, eggs of H. halys and N. viridula) was characterized by determining the contact angles of water (aqua millipore, droplet size: 1 μl for glass substrate and 0.5 μl for glass beads and eggs, sessile drop method) using a high-speed optical contact angle measuring instrument OCAH 200 (Dataphysics Instruments GmbH, Filderstadt, Germany). Ten measurements (n = 10) were performed for each substrate.

To test the attachment ability of A. bifasciatus to artificial substrates represented by polishing paper with different asperity sizes and flat hydrophilic and hydrophobic glass, a centrifugal force tester (Gorb et al., 2001) was used. Males were not tested on polishing paper with 35 and 125 μm asperity sizes because, due to their small dimension, the optic sensor could not detect them on these rough surfaces.

The centrifuge is constituted of a metal drum covered by a substrate disc to be tested. The metal drum is driven by a computer-controlled motor. Just above the disc, the fibre-optic sensor monitored by the computer is placed. After positioning the insect on the horizontal disc, the centrifuge drum was allowed to begin the rotation at a speed of 50 rev min⁻¹ (0.883 rev s⁻¹). The position of the insect on the drum was monitored by using a combination of a focused light beam and a fibre-optical sensor. The drum speed was continuously increased until the insect lost its hold on the surface under centrifugal force. The rotational speed at contact loss, the last position of the insect on the drum (radius of rotation), and the insect weight were used to calculate the maximum frictional component of the attachment force. In total, 38 females and 25 males of A. bifasciatus were tested. In random order, each specimen was tested on four or five different surfaces and on hydrophilic glass as a reference.

This force measuring experimental set-up was used for testing the insect attachment ability to artificial substrates with glass beads of different sizes, eggs of H. halys and N. viridula and flat hydrophilic glass. Forces were measured with the load cell force sensor FORT-10 (10 g capacity; World Precision Instruments Inc., Sarasota, FL, United States) connected to a force transducer MP 160 (Biopac Systems Ltd., Goleta, CA, United States) (Gorb et al., 2010). The sensor was kept in the horizontal position and the insect thorax was attached using a small drop of glue (Super Attack, Loctite Henkel Adhesives, Düsseldorf, Germany) to the tip of a vertical pin fixed to the sensor. Experimental insects were anaesthetized with carbon dioxide for 60 s before gluing their thorax to the pin and allowed to recover undisturbed for 5 min before testing them. The force sensor was fixed to a motorised micromanipulator (Narishige MMO-203) and moved in such a way that the insect was in contact with all six legs to the tested surface. Afterwards, the force sensor was moved in a direction perpendicular to the surface until the insect detached from it. Data were recorded using AcqKnowledge 5.0 software (Biopac Systems Ltd., Goleta, CA, United States). Pull-off force values were estimated using the force-time curves as the maximum of the detected forces. 30 females and nine males were tested.

The friction force of A. bifasciatus on polishing paper with different asperity sizes normalised to the insect friction force recorded on smooth hydrophilic glass was analysed separately for females and males using the one-way analysis of variance (ANOVA). For data obtained on polishing paper with asperity sizes of 0.3, 1, 3, 9, and 12 μm, the male and female normalised friction forces were compared using the Student t-test for independent samples.

The comparisons of the safety factor (the force divided by the insect weight) and force values of A. bifasciatus females and males on smooth hydrophilic and hydrophobic glass (friction force) as well as on eggs of H. halys and N. viridula (pull-off force) were conducted using the repeated measures analysis of variance (ANOVA) considering the insect sex as a factor.

The safety factor and pull-off force of A. bifasciatus females and males on a flat glass surface and glass beads with different diameters were analysed using the two-way analysis of variance (ANOVA), considering the insect sex and the different substrates as factors. Before the analysis, the data were subjected to Box-Cox transformations, to reduce data heteroscedasticity (Sokal and Rohlf, 1998).

Data on arolium width of forelegs, midlegs and hindlegs in males and females of A. bifasciatus were analysed using the two-way analysis of variance (ANOVA), considering the insect sex and different legs as factors.

The ANOVA analyses were followed by the Tukey unequal N HSD post hoc test for multiple comparisons between means (Statistica 6.0, Statsoft Inc., 2001).

The tarsi of A. bifasciatus are composed of five tarsomeres and show sexual dimorphism, especially in the midlegs (Figure 1). These last have straight tarsi in the male and all tarsomeres have the same width, while in the female the tarsi of the medium legs show a curved shape and the first tarsomere is wider than the others (Figure 1). Observations under cryo-SEM reveal that the forelegs and hindleg tarsi in males (Figures 2A,B,D) and females (Figure 3) are similar and show no special differentiation except for a cleaning structure (comb) located in the proximal portion of the forelegs (Figures 2A,B, 3A,C). Articulated setae and microtrichia are visible on the inner side of the tarsi (Figures 2C, 3B,E) and spurs (S) are present at the distal tibia in all the legs of both sexes (Figures 1–3). The tarsi of the midleg in the male show setae (arrowheads in Figure 2) and microtrichia (arrows in Figure 2) along their inner side (Figure 2C), whereas in the female midleg tarsi the inner side of each tarsomere is characterised by a central row of microtrichia and two lateral rows of big and thick pegs, which are set one close to the other (Figures 4A–F). The pegs are articulated, pointed in their apical portion and slightly curved towards the inner portion of the tarsomere, with longitudinal grooves. They are about ten in the first tarsomere forming two rows facing one to the other, about eight in the second tarsomere, four in the third one, and two in the fourth one, while no peg is present in the last tarsomere (Figure 4B). In the first and second tarsomeres hosting the higher number of pegs, there are more pegs in the posterior row than in the anterior one (Figures 4C,D).

The pretarsal attachment devices of A. bifasciatus are similar in both sexes and are constituted of a smooth pad, the arolium (A) and a pair of claws (C) (Figure 5). The arolium dorsally is articulated to the fifth tarsomere through the manubrium (MA), a subtriangular sclerite with five or six long hairs along its distal margin and one distal hair (DH) (Figures 5A,B,D). The arolium is supported by dorsal sclerites named dorsal plates (DP) in continuity with the arcus (ARC) surrounding the arolium proximally and ventrally (Figures 5A–D). The arolium can be inverted and nearly invisible (Figures 5E,F) or everted [completely (Figures 5B–D) or partly spread (Figure 5A)]. When the arolium is everted, it is formed by two big lobes separated by a dorsal deep median cleft (Figures 5B–D). The surface of the everted arolium is not smooth but wrinkled (Figure 5C). Ventrally a quadrangular unguitractor plate (UP) with microtrichia and scales along with its lateral surface inserts into the last tarsomere and is in continuity with planta (P), whose distal portion extends in the arolium (Figure 5F). Two claws covered with some setae are located laterally to the hairy planta (Figure 5F). When the arolium is everted, the claws are rotated towards the last tarsomere (Figures 5C,D). When the arolium is inverted, the claws are kept along the hairy planta (Figure 5F).

The female weight is 0.52 ± 0.04 mg (n = 16) and the male weight is 0.30 ± 0.02 mg (n = 15). The arolium width was greater in females compared to males, but not statistically different among forelegs (males: 183.4 ± 3.6 μm; females: 214.7 ± 4.6 μm), midlegs (males: 167.1 ± 6.7 μm; females: 218.1 ± 6.8 μm) and hindlegs (males: 188.7 ± 7.2 μm; females: 218.9 ± 4.0 μm) (legs: F = 1.9; d.f. = 2, 102; p = 0.15; sex: F = 65.6; d.f. = 1, 102; p < 0.001; legs x sex: F = 2.1; d.f. = 2, 102; p = 0.12).

Behavioural observations of the egg-laying females on the eggs of H. halys (Figure 6) revealed that during oviposition, the female touches the egg chorion with the tarsal tip of the fore- and hindlegs. In these legs, the females use only the pretarsal attachment devices (claws and arolium) to attach to the egg surface and the tarsi are not contacting the egg. Conversely in the midlegs, the ventral surface of the whole tarsus is in close contact with the chorion. The contact of the thick pegs located on the first four tarsomeres with the egg surface is clearly visible (Figure 6D and inset). The developed tibial spur of the midlegs is in contact with the egg surface and forms a wide angle with the tarsus (Figure 6).

The egg clutches of H. halys consist of about 27–28 eggs. Each egg is light green or light blue, has a subspherical shape and a diameter of 1.28 ± 0.01 mm (mean ± SE) (Figure 7A). A crown of aeromicropilar processes is located along the apical egg surface (Figure 7B). The whole chorion surface is rough due to wrinkles and tubercles (Figure 7C). The egg clutches of N. viridula are constituted of about 30–130 eggs. The eggs are cream to yellow in colour and barrel-shaped, with flat tops and a diameter of 0.82 ± 0.01 mm (mean ± SE) (Figures 7D–F). A crown of aeromicropilar processes (processes with a central channel for the sperm passage and respiratory exchange in many Heteroptera species) is present along the apical surface (Figure 7E). The chorion is smooth, with some rough areas around the aeromicropiles (Figure 7F).

The contact angle of water to the eggs of H. halys is 134.9 ± 4.3° (mean ± SE, n = 10) and that to N. viridula eggs is 120.6 ± 1.3° (mean ± SE, n = 15).

In the centrifugal experiments on polishing paper with different asperity size (0.3, 1, 3, 9, 12, 35 and 125 μm; 35 and 125 μm tested only in females) (Figure 8), normalised friction force (normalised to friction force recorded on hydrophilic glass as reference) of females was significantly higher than that of males on all the tested surfaces (0.3 μm: t = 4.0; d.f. = 28; p < 0.001; 1 μm: t = 3.4; d.f. = 28; p = 0.002; 3 μm: t = 2.8; d.f. = 28; p = 0.010; 9 μm: t = 3.1; d.f. = 28; p = 0.005; 12 μm: t = 2.9; d.f. = 28; p = 0.008). In females, the force on the surface with the intermediate asperity size (9 μm) was significantly higher than that recorded on surfaces with the smallest (0.3 and 1 μm) and highest roughness, such as 35 and 125 μm, while the force values on surfaces with asperity size 3 and 12 μm was intermediate (F = 3.8; d.f. = 6, 98; p = 0.002). In males, the force was not significantly different on surfaces with asperity sizes 3, 9, and 12 μm and significantly higher compared to the force on 0.3 μm surface. On the surface with an asperity size of 1 μm, the force value was intermediate (F = 4.1; d.f. = 4, 70; p = 0.005).

There is no significant difference between the force of males and females on both hydrophilic and hydrophobic surfaces. The attachment ability of females does not differ on both surfaces, while the friction force of males is significantly higher on hydrophilic glass than on hydrophobic glass (sex: F = 0.5; d.f. = 1, 28; p = 0.473; surfaces: F = 13.9; d.f. = 1, 28; p < 0.001; sex x surf.: F = 6.8; d.f. = 1, 28; p = 0.015) (Figure 9A).

On the smooth hydrophilic glass, the safety factor of males is significantly higher than that of females, while there is no significant difference between the safety factor of males and females on smooth hydrophobic glass (Figure 9B). In females, difference between the safety factors on hydrophilic glass and hydrophobic glass is non-significant, while males attach significantly stronger to hydrophilic glass than to hydrophobic glass (Figure 9B) (sex: F = 32.9; d.f. = 1, 28; p < 0.001; surfaces: F = 19.3; d.f. = 1, 28; p < 0.001; sex x surf.: F = 12.7; d.f. = 1, 28; p = 0.001).

In the Biopac force tester experiments with artificial substrates bearing glass beads with different diameters used as dummies of the host eggs, the force values in females are significantly higher than those in males on surfaces constituted of beads with a diameter of 0.5, 0.9 and 1.4 mm, while there is no significant difference between males and females on the surfaces bearing 2.0 and 3.1 mm beads and on smooth glass (Figure 10A). In females, the forces are significantly higher on the surfaces composed of beads with a diameter of 0.5, 0.9 and 1.4 mm compared to the surfaces constituted of 2.0 and 3.1 mm beads. The force values on these last surfaces are higher than that recorded on flat glass (Figure 10A). In males, the forces are significantly higher on the surfaces constituted of beads with a diameter of 0.5, 0.9 and 1.4 mm than on flat glass, while on surfaces with 2.0 and 3.1 mm beads it is intermediate (Figure 10A) (sex: F = 94.6; d.f. = 1, 128; p < 0.001; surfaces: F = 28.1; d.f. = 5, 128; p < 0.001; sex x surf.: F = 2.3; d.f. = 5, 128; p = 0.049).

As for the safety factor, there is no significant difference between males and females for each tested surface (Figure 10B). In females, the safety factor is significantly higher on the surfaces constituted of beads with a diameter of 0.5, 0.9 and 1.4 mm compared to the surfaces with beads of 2.0 and 3.1 mm in diameter. The safety factor on these last surfaces is higher than that recorded on flat glass (Figure 10B). In males, the safety factor is significantly higher on the surfaces constituted of beads with a diameter of 0.5, 0.9 and 1.4 mm than on flat glass, while it was intermediate on surfaces with beads of 2.0 and 3.1 mm in diameter (Figure 10B) (sex: F = 5.0; d.f. = 1, 128; p = 0.028; surfaces: F = 26.8; d.f. = 5, 128; p < 0.001; sex x surf.: F = 2.3; d.f. = 5, 128; p = 0.052).

The force of females is higher than that of males on the eggs of H. halys, whereas force values on the egg surface of N. viridula are similar in the two sexes (Figure 11A). The force of females is significantly higher on the eggs of H. halys than on the eggs of N. viridula, while in males there is no difference in the forces recorded on different egg surfaces (Figure 11A) (sex: F = 13.1; d.f. = 1, 16; p = 0.002; surfaces: F = 13.1; d.f. = 1, 16; p = 0.002; sex x surf.: F = 1.9; d.f. = 1, 16; p = 0.177). There is no significant difference between safety factors of males and females on both egg surfaces (of H. halys and N. viridula) (Figure 11B). In females, the safety factor recorded on the egg surface of H. halys is higher than that recorded on N. viridula egg surface, while there is no difference between the safety factors recorded on both egg surfaces in males (Figure 11B) (sex: F = 2.9; d.f. = 1, 16; p = 0.104; surfaces: F = 13.8; d.f. = 1, 16; p = 0.002; sex x surf.: F = 0.3; d.f. = 1, 16; p = 0.573).