Distinct kinematics and micromorphology for symmetrical rowing and sliding on water in ripple bugs and water striders

Distinct kinematics and micromorphology for symmetrical rowing and sliding on water in ripple bugs and water striders | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article Distinct kinematics and micromorphology for symmetrical rowing and sliding on water in ripple bugs and water striders Sang Yun Bang, Woojoo Kim, Jeongseop Lee, Jinseok Park, Versha Khare, and 2 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-6903117/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 29 Dec, 2025 Read the published version in Scientific Reports → Version 1 posted 10 You are reading this latest preprint version Abstract The evolution of semiaquatic bugs led to at least two independent origins of symmetrical rowing: a drag-based strategy in Veliidae from fast-flowing waters and a surface-tension-based strategy in Gerridae from still or slow-flowing waters. However, the leg micromorphology and motion patterns underlying these strategies remain underexplored. Using scanning electron microscopy and high-speed video, we compared Rhagovelia distincta (Veliidae), which uses midleg pretarsal fans as oar-like blades, with Gerris latiabdominis (Gerridae), which relies on long, hairy midlegs to generate surface-tension-based thrust. R. distincta performed short, high-frequency strokes and actively controlled its fans, which function as “leaky paddles” exploiting drag, and potentially lift, forces. Fan anatomy further suggests nano-structural adaptations for enhanced mechanical performance. R. distincta also engaged its midleg tarsi in surface-tension-based thrust. In contrast, G. latiabdominis exhibited longer stroke durations and kinematics suited to surface-tension-based propulsion. Both species shared key micromorphological features: ventral longitudinal rows and gaps of midleg setae, with posterior rows particularly robust and nano-grooved in G. latiabdominis. Additionally, both formed ventral “beam-like” structures from overlapping flat-tipped setae on hindlegs—and less prominently on forelegs—used for support and sliding. These findings generate new hypotheses for refining models of locomotion on water surface by insects. Biological sciences/Ecology/Behavioural ecology Biological sciences/Zoology/Animal behaviour Biological sciences/Zoology/Biomechanics Biological sciences/Zoology/Entomology Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Figure 8 Figure 9 Introduction Locomotion on the water surface presents a unique set of physical challenges that have driven repeated and diverse evolutionary solutions among insects. The semiaquatic bugs (Gerromorpha) provide an example of adaptive radiation into this novel environment. These insects evolved distinct morphological, behavioral, and anatomical traits that enable movement on the air-water interface. These adaptations reflect multiple, lineage-specific strategies for solving similar functional problems–offering a model system for studying how alternative solutions involving morphology and behavior evolve under ecological constraints of physical environment–the water surface. However, the details of the co-evolutionary and functional interplay between leg microstructures, stroke kinematics, and thrust mechanics across major independent lineages of this adaptive radiation remain poorly understood. To provide more insights into this question, we compare two major independently evolved solutions for symmetrical rowing in Gerromorpha. Here, we focus on how alternative physical mechanisms–drag-based versus surface-tension-based thrust–are realized through contrasting yet functionally convergent morphologies. Notably, two lineages—Veliidae (a polyphyletic family typically found in fast-flowing streams) and Gerridae (typically inhabiting slow or still waters)—have independently evolved symmetrical backward rowing by midlegs for forward thrust (Andersen 1982; Crumiere et al. 2016; Khila et al. 2014). While both rely on midlegs for thrust and fore- and hindlegs for support and sliding, their physical modes of thrust generation differ: Veliidae exploit hydrodynamic drag (and potentially lift), while Gerridae generate thrust primarily through surface tension forces (Andersen 1976; Steinmann et al. 2021). Despite their independent origins and these contrasting mechanisms, the leg kinematics and microstructural adaptations that support these behaviors have not been systematically compared. Within Veliidae, species such as those in the genus Rhagovelia possess specialized midleg pretarsal structures known as swimming fans, which function as oars (Andersen 1976, 1982; Santos et al. 2017; Ortega-Jimenez & Bhamla 2021; Ortega-Jimenez et al. 2024). These structures are assumed to be actively controlled via a claw retractor muscle (Andersen 1976), though recent observations of isolated fans spreading in water have led to the hypothesis of passive elasto-capillary spreading (Ortega-Jimenez & Bhamla 2021; Ortega-Jimenez et al. 2024). The nature of fan manipulation in intact, behaving animals remains unresolved. Similarly, the ventral microstructures on midlegs (involved in thrust) and on fore- and hindlegs (involved in support and sliding) have not been systematically examined in Rhagovelia, though their role in generating surface tension-based forces via water dimples is likely. In contrast, Gerridae species such as Gerris latiabdominis do not possess swimming fans but instead have elongated, hairy midlegs that generate thrust through the creation of asymmetric dimples on the water surface. These midlegs, along with highly hydrophobic ventral surfaces, are critical for surface-tension-based propulsion (Steinmann et al. 2021). While some aspects of locomotory performance have been compared (Crumiere et al. 2016), detailed kinematic analyses and high-resolution comparisons of leg micromorphology between Gerridae and Veliidae are lacking. Prior studies have presented images of leg hair arrangements (Andersen 1982; Perez-Goodwyn 2008), but a focused comparative analysis across functional leg regions has not been conducted since Andersen (1976). Following the framework set by Andersen (1976, 1982) and Crumiere et al. (2016), we compare Rhagovelia distincta (Veliidae) and Gerris latiabdominis (Gerridae), two small-bodied representatives of their respective clades. We examine their midleg microstructures and stroke kinematics to evaluate how different physical thrust mechanisms are supported by contrasting anatomical features. We further investigate the hypothesized passive versus active mechanisms of fan control in R. distincta . Finally, we describe and compare fore- and hindleg microstructures used in support and sliding to identify potential convergences across taxa with different thrust mechanisms. Our results provide a foundation for future work on the functional evolution of water-surface locomotion in Gerromorpha. Results Behavioral observations: leg use at rest R. distincta body is supported on foreleg and hindleg tarsus with minimal contribution from the midleg tarsal tips (Figure 1A1, B1, C1, D1; Figure S1). Occasionally, a very small portion of the midleg`s fan is protruded from the tarsal tip into the water body through the surface it`s in contact with (Figure 1B1; Video S3 Part 4). G. latiabdominis body is supported on foreleg tarsus, hindleg tibia and tarsus, and midleg intermediate-distal tibia and tarsus (Figure 2A1, B1, C1, D1; Video S2), which create dimples without piercing the water surface (dimples cast shadows on the bottom of the container; Figure 2C1, D1; Video S2 Part 4). Behavioral observations: leg use at locomotion During the initial thrust phase (Figure 1), R. distincta moves its midlegs forward above the water, placing the tarsi on the surface (Figure 1C2), which initiates fan extension into the water (Figure 1B2). Alternatively, the midlegs may advance while the tarsal tips remain in contact with the surface and a small portion of the fan protrudes underwater (Figure 1B6; Figure S2D; Video S3 Part 4). As the fan rapidly protracts (Figure 1B2; Figure S2C, D), the tarsus is simultaneously pressed downward (Figure 1C3; Figures S2B, S5) and the midlegs are pushed backward (Figure 1C4, C5). This generates growing anteroposterior asymmetrical dimples (Figure 1B4), visible as shadows expand from miniscule circles at the tarsal tips to ovals extending from the tibiotarsal joints (Figure 1D4). Contrast to G. latiabdominis (Figure 2D), R. distincta shows an expanded anterior dimple region (Figure 1C3), likely caused by water displaced beneath the surface by the fan (Figure 1B4; Figure S2G, H). Strong strokes can produce surface waves (Figure 1C4, C5; 17 of 99 strokes; Video S1 Part 1), but all strokes transition into a passive sliding phase, during which the midlegs either disengage from the surface or trail behind with minimal fan protrusion (Figure 1B6; Video S1). Observed midleg disengagement suggests adhesive forces and surface tension are overcome in this process (Video S3). Throughout, the fore- and hindleg tarsi remain in contact with the surface, providing support during sliding. Duration of fan protraction ranged between 6­­–23 ms (12.3 ± 3.4 ms; n = 69) while the duration of fan retraction ranged between 3–15 ms (8.4 ± 2.6 ms; n = 73) (Figures S3 and S4). Shorter durations were often associated with strokes beginning or ending with partially protracted fans, where 2–4 distal setae tips protruded into the water. Fully protracted fans had an average area of 0.89 ± 0.04 mm 2 , radius of 0.85 ± 0.01 mm, and protracted angle of 139.13 ± 3.64° (n = 6; Figure S3). The longitudinal axis of wetted midleg and protracted fan typically lay in the same plane, slanted at 80.2 ± 6.0° (n = 26; Figure S3E). Analysis of 110 slow-motion videos (each capturing 1–4 fan manipulations) showed that R. distincta actively controls the timing, extent, and duration of fan protraction and retraction, regardless of midleg position on the water surface (Figure S4; Video S3). G. latiabdominis generated thrust without piercing the surface (Figure 2), using midlegs slightly rotated so that anteroventral gap-row microstructures pressed backward against the water (Figure S2; 27–36 ms into the stroke). Backward movement of the midlegs produced anteroposterior asymmetrical dimples and backward-moving surface waves (Figure 2D; Figure S2E, F; Video S2). Midleg disengagement occurred when the wetted midleg aligned nearly parallel to the direction of body movement and proceeded gradually from proximal to distal segments. Some instances of disengagement appeared smooth (Video S2), while others suggested adhesive forces and surface tension were overcome during the process (Video S2; Figure S2E, 95 ms). Hindlegs contributed weakly to thrust, as indicated by faintly asymmetrical shadows and small wave bows during the initial stroke phase (Figure 2D4). Hindleg tarsi and tibiae provided support during sliding, shown by shadows aligned with the body axis (Figure 2D5, D6). Forelegs also supported the body, particularly near the end of the stroke, as indicated by prominent shadows (Figure 2D6), except at mid-stroke when support was reduced (Figure 2D4). Kinematics profiles during a stroke In both species (Figure 7), the midleg femur angle increased gradually during the stroke, but more steeply in R. distincta than in G. latiabdominis . The initial angle was more acute in R. distincta (~20°) compared to G. latiabdominis (~60°), with both reaching a final angle of ~120° (Figure 7A, B). Peak femur angular velocity was higher in R. distincta and occurred mid-stride at a femur angle of 85°, whereas in G. latiabdominis it peaked at 100° (Figure 7A, B). In G . latiabdominis , the femur–tibia and tibia–tarsus angles remained relatively small throughout the stroke. In contrast, R. distincta showed more pronounced changes: the femur–tibia angle increased from 30° to 60° (Figure 7C), and the tibia–tarsus angle ranged from 8° to 16° (Figure 7D). These patterns suggest that R. distincta engages in coordinated rotations at the coxa–femur, femur–tibia, and—though to a lesser extent—tibia–tarsus joints, while in G. latiabdominis , motion is largely concentrated at the coxa–femur joint. In both species, leg velocity during slower (longer) strokes remained below the theoretical critical velocity (~0.23 m/s; Figure 7E) above which gravity-capillary waves are generated and may contribute to thrust (Reaver et al. 2019). In faster (shorter) strokes, leg velocity exceeded this threshold within ~5 ms and reached higher peak values in G. latiabdominis than in R. distincta (Figure 7E). In R. distincta , the leg velocity vector was briefly aligned with the body movement axis only during mid-stroke, when the femur was approximately perpendicular to the trajectory (Figures 9G and 7G). For the remaining stroke, the fan–acting as an oar blade–moved in a direction misaligned with the body axis and not perpendicular to its movement direction. In contrast, in G. latiabdominis , the leg velocity vector remained nearly parallel to the body movement axis (Figures 9E and 7F) and nearly perpendicular to the wetted leg section (Figures 9G and 7H) for a much larger portion of the stroke (rectangles on x-axis). In both species, the highest net force per stroke, indicated by peak body acceleration (Figure 7J), coincided with intervals of high “effectiveness” indices (Figure 7F–H; SI Part 5). Kinematics comparisons of a stroke A small subset of G. latiabdominis strokes showed notably higher average leg linear velocities, body velocities, and peak accelerations (Figure 8). However, when all strokes were analyzed together, there were no significant differences between the two species in average midleg velocity (Figure 8B2; Table S4), final body velocity (Figure 8A1; Table S4), or maximal body acceleration (Figure 8A2; Table S4). Strokes by R. distincta were shorter in duration (Figure 8C3; Table S4) and involved faster angular femur movements (Figure 8B1; Table S4), and traveled a shorter distance across the water surface during each stroke compared to G. latiabdominis (Figure 8C1, C2), notably as R. distincta has shorter legs (Tables S1 and S2) and additional angular movements were observed at the tibia and tarsus (Figure 7C, D). Since the kinematic variables are intercorrelated (Figure S26D), estimates of their individual effects on body speed (Figure S26A–C) are not independent. To address this, we performed a principal component analysis and extracted two components (Table 1). Strokes by R. distincta were characterized by shorter duration and higher femur angular velocity (higher RC2 values), along with shorter distance and slower leg speed on the water surface (lower RC1 values) (Figure 8E), distinguishing them from G. latiabdominis . Ventral microstructures on legs Overview The legs of R. distincta had a less dense hair layer than those of G. latiabdominis . We identified 17 setae (hair) types: five shared by both species, five unique to R. distincta , and seven unique to G. latiabdominis (Figure 3; Figure S10; Table S3). Our analysis focused on the ventral microstructures of leg segments that interact with the water surface, and respective nanometer sized details specifically in R. distincta (Figures 4 and 5; Figures S11–S23). Ventral microstructures for thrust generation The swimming fan of R. distincta (Figure 4B1, B2; Figures S11–S14) consists of anterior and posterior claws and a fan made up of 17 (Figures S8, S9) to 21 (Figure 4B2; Figure S11) setae. Each seta bears setulae along its axis at 8–12 µm intervals, forming a feather-like structure (Figure 4B3, B4; Figures S12, S13). When protracted, the distance between adjacent setae ranges from ~20 to ~100 µm, and between setulae from several to ~20 µm, with typical setula spacing of 4–10 µm (Figure S13A, B). The fan is anchored at the inner proximal corner of the cleft between the two lobes of tarsomere 3 (Figure 4B2). The surfaces of the setae, setulae, and claws lack nanogrooves. The setae resemble flat beams or boards (Figure 4B5; Figure S12C) with near-elliptical (Figure 4B6; Figure S12D) or slightly triangular (Figure S12E) cross-sections, measuring 2–4 µm × 7–10 µm. The orientation of the narrow edges suggests that they face the water during thrust generation (Figure 4B5; Figures S12, S13). The setulae are also flat, 1–2 µm wide and 300–700 nm thick, with a hollow center (Figure S12H), and their orientation further suggest that they press against the water with the narrow edges. Transverse sections of fan setae show lamellar outer layers and a central hollow (~700 nm in diameter) containing internal rods (~400 nm in diameter) (Figure 4B6; Figure S12). Claw cross-sections are 1.5–2.5 µm thick and consist of external lamellar layers and multiple internal layers with nanofibers (100 nm) and cluster of nanofibers (200–400 nm) running in various directions (Figure 4B8; Figure S14). Although initial observations with fully dissected fan indicate passive spreading, further observations with the fan and claws intact inside the cleft indicate that fan movement is not passive (SI Part 3B). In fact, we were able to induce fan protraction by pulling the ut tendon attached to the base of the fan-and-claw structure (Figures S8, S9). Along the ventral edge of the posterior lobe, a structure composed of three rows of H2 setae—spaced 6–8 µm apart within each row and separated by two 2.5–5 µm wide gaps—forms a band that presses against the water surface without breaking it during a stroke (Figure 4B9; Figure S16). This pattern resembles the “gaps and rows” arrangement seen on the tarsus of G. latiabdominis (Figure 5B). Similarly, the ventral edge of the anterior lobe, which also contacts the water surface during a stroke, is lined with a band of 3–4 rows of H1 setae (Figure 4B9; Figure S16); however, only lateral views were available, limiting precise row counts. On ventral tarsomere 2, which also interacts with the water surface without breaking it (Figure S5), we observed three rows: a posterior row of H1 setae and two anterior rows of Sp setae, separated by gaps (Figure 4B12, B13; Figure S17). In G. latiabdominis , a “gaps and rows” arrangement of setae was observed along ventral midleg sections that interact with the water surface (Figure 5B; Figures S20 and S21). This pattern resembles the structure on the ventral edge of the lobe in R. distincta (Figure 4B9). It is especially prominent on the tarsus, where a main ventral gap separates a row of cuspidate setae ( C ) from a posterior row of thorn-like T2 setae, with a second T2 row positioned further posteriorly to form a T2–T2 gap (Figure 5B4–B9). Anterior to the C row, a row of M1 setae creates a C–M1 gap (Figure S20E). The tips of these three setae types bend distally along the leg`s longitudinal axis, particularly the cuspidate setae, whose long, flat distal sections may contact each other. In contrast, setae on the lateral and dorsal leg surfaces are relatively straight (Figure S21B). On tarsomere 2, a second, less regularly arranged anterior M1 row creates an M1–M1 gap (Figure S20C, E). This arrangement is less distinct on the tibia, where T2 setae exhibits intermediate morphology between T2 and M1 , and the C setae are replaced by M2 (Figure 5B3). At the tip of the tarsus, the pattern disappears, replaced by a ventral concentration of grass-blade-like setae ( g ) (Figure 5B12). Ventral microstructures for support and sliding In R. distincta , specialized microstructures involved in support and sliding were found on the ventral tarsi of the forelegs (Figure 4C), hindlegs (Figure 4A), and at the distal tips of the midleg tarsus (Figure 4B10, B11). These structures consist of one or more rows of Sp setae with flattened, bent tips that overlap to form a “beam-like” surface ~20–25 µm below the leg cuticle, oriented toward the water. Two Sp rows were observed on the ventral forelegs (Figure 4C3, C4; Figure S18) and 1 on the hindlegs (Figure 4A1; Figure S19), with the flattened tips especially prominent near the leg ends (Figures 4A5, A6; 4C4). Sp rows are flanked—particularly posteriorly—by H1 setae on the forelegs (Figure 4C; Figure S18) and M2 setae on the hindlegs (Figure 4A; Figure S19). At the tips of the midlegs, which also provide support, a dense cluster of Sp setae—likely modified H1 types—was observed, with long, flat, overlapping tips (Figure 4B11; Figure S15F, G). In G. latiabdominis , specialized microstructures involved in support and sliding were found on the ventral forelegs and hindlegs. On the ventral foreleg tarsus, a band of grass-blade setae ( g ) with overlapping flat, bent distal tips was observed ~15–20 µm from the leg cuticle, facing the water surface (Figure 5C4; Figure S22). An entangled cluster of web setae ( W ) was also present, especially near the tibiotarsal joint (Figure 5C2; Figure S22). On the ventral side of the hindleg tibia and tarsus, 2–3 rows of leaf-blade setae ( L ) were arranged in an orderly manner (Figure 5A; Figure S23). Their overlapping distal tips formed a “beam-like” surface with nanogrooves, ~20 µm above the water surface (Figure 5A2; Figure S23), and were accompanied by a posterior row of large thorn setae ( T3 ) (Figure 5A6, A8; Figure S23A, E). The ventral sides of joints were covered with bundles of L and leaf-like l setae (Figure 5A1, A6), with l setae also present at the tarsal tips (Figure 5A9). Contact angle on midleg sections used in thrust generation Midleg tarsal surfaces in G. latiabdominis were generally more hydrophobic than those in R. distincta (Figure 6). In R. distincta , water droplets rapidly lost their spherical shape after contacting the hair layer, spreading 10–40% and showing increased shape indices. Contact angles on the dorsal and ventral midleg surfaces progressively decreased from approximately 130.3° and 75.5°, respectively, to as low as 21.9° as droplets collapsed, indicating relatively high surface wettability (Figure 6A, B). This effect was especially pronounced on the ventral side, where the swimming fan and associated microstructures interact with water during locomotion (Figure 6B, C); Figure S24). In contrast, droplets on the dorsal and ventral midleg surfaces of G. latiabdominis retained their spherical shapes with minimal spreading (shape index ≈ 1.5), and contact angles remained high throughout dissipation, ranging from 132.1° to 109.3°, consistent with strong hydrophobicity (Figure 6D, E; Figure S24; SI Part 4). Table 1. Principal Component Analysis of behavioral variables. PCA was conducted on seven behavioral variables from R. distincta (n = 21) and G. latiabdominis (n = 12). The table shows eigenvalues, percentage of variance explained, and loadings for the first two rotated components (RC1 and RC2), based on the fa.parallel and principal functions from the psych R package. Loadings with absolute values greater than 0.75 are shown in bold. Related results are illustrated in Figure 8D. RC1 Midleg`s & Body`s Movements PC RC2 Midleg Angular Speed & Duration PC Linear velocities & accelerations: Average midleg velocity (mm/s) 0.80 0.53 Maximum acceleration (mm/s 2 ) 0.77 0.48 Final body velocity (mm/s) 0.77 0.60 Linear midleg movement distances: Midleg`s stroke amplitude (mm) 0.96 -0.22 Distance traveled by wetted midleg (mm) 0.96 -0.20 Angular midleg (femur) speed and duration: Average angular velocity (degrees/s) 0.03 0.93 Stroke duration (s) 0.04 -0.98 Eigenvalue 3.67 2.77 Variance % 52% 40% Discussion Observations of live R. distincta suggest that thrust during a stroke results from a combination of two forces: hydrodynamic forces generated by the oar-like motion of the fan, as proposed in previous studies (Andersen 1976; Ortega-Jimenez et al. 2024; Santos et al. 2017), and additional capillary forces arising from an anteroposteriorly asymmetrical dimple beneath the tarsus, consistent with surface-tension-based mechanisms described in Gerridae (Steinman et al. 2021). In contrast, behavioral evidence from G. latiabdominis aligns with the surface-tension-based thrust mechanism characterized in detail for Aquarius paludum (Steinman et al. 2021). The observed differences between these species appear closely tied to variations in leg microstructures, material properties, and motion kinematics that reflect their distinct ecological contexts and thrust-generation strategies. In R. distincta , several microstructural features suggest specialization for hydrodynamic thrust. The hydrophilic properties of the claw likely facilitate surface penetration during stroke initiation, as the fan and claw extend downward from the cleft's internal compartment. Similar surface textures between the claw and fan setae indicate that the fan may also be hydrophilic, enhancing its ability to submerge through the water surface. The fan’s setae and setulae are oriented to press against the water with their narrow edges, a configuration that minimizes deformation under hydrodynamic forces, as predicted by beam theory (Bhavikatti 2010; SI Part D). The fan setae’s internal architecture, comprising a hollow core with columnar nanofibers, and the claw’s lamellar structure, resembles engineered designs like sandwich and lamellar composites (Zenkert 1995; Wu and Zhu 2021), which are known to enhance stiffness and fatigue resistance. This likely enables the fan to function effectively as an oar-blade. The nanofiber arrangement within the fan setae provides both flexibility and mechanical resilience—qualities vital for high-frequency paddling in fast-flowing water. The central hollow core may serve to reduce weight and accommodate controlled deformation, further improving maneuverability. Additionally, the H1 and H2 setae rows positioned at the cleft entrance likely act as a barrier against water intrusion while assisting fan deployment through elastocapillary interactions. Based on the observed stroke speed, setula spacing, and thickness, we estimate the fan’s leakiness to range from ~0.3 to 0.6 (SI Part 7), suggesting it behaves as a “leaky paddle” rather than a solid blade. This is similar to bristled appendages in copepods and barnacle larvae (Koehl 1993; Lamont & Emlet 2018). Comparable fan-like structures with potential “leaky paddle” functionality are found in other Veliidae genera such as Tetraripis and Trochopus . In contrast, “Veliidae” species that depend primarily on surface-tension-based thrust, like Velia sp. (Andersen 1976), tend to exhibit more developed ventral "gaps and rows" arrangements and less-developed fan structures, supporting the hypothesis of divergent functional adaptations. In G. latiabdominis , a different set of microstructural adaptations supports thrust generation primarily through surface tension. The species exhibits denser and more hydrophobic setae on its midlegs than R. distincta , likely reflecting its reliance on surface-tension-based propulsion. Deep, asymmetrical dimples beneath the wetted portion of the leg, along with prominent bow waves, contribute to increased thrust. Interestingly, even the lower hydrophobicity observed in R. distincta can still support thrust via surface tension through dimple formation, though to a lesser degree. This is evidenced by the net thrust outputs, which are significantly lower in R. distincta (100–200 µN) than in G. latiabdominis (400–500 µN; Figure S25). Across both species, leg surfaces involved in surface-tension-based thrust display linear arrangements of distally bending setae forming "gaps and rows." These are absent from other leg surfaces and may serve specialized functions, potentially related to air retention during dimple formation, which prevents surface penetration and facilitates thrust. Theoretical work (Uesugi et al. 2020; Uesugi 2021) suggests such arrangements can trap air and maintain smooth water contact. Our observations support this, as these structures appear to function like pressurized air pockets that resist water surface breakage under thrust loads. Additionally, the smoother longitudinal gaps may reduce adhesion during stroke recovery. This arrangement is more pronounced in G. latiabdominis , where the setae also exhibit nanogrooves known to enhance hydrophobicity (Feng et al. 2007; Gao & Jiang 2004). The posterior concentration of thicker T2 and T3 setae in mid- and hindlegs may be a specific adaptation to withstand the higher pressure during backward strokes. These structural reinforcements are aligned with previous findings on jumping and propulsion in surface-dwelling insects (Koh et al. 2015; Yang et al. 2016). Kinematic data further underscores how thrust generation mechanisms are integrated with species-specific movement strategies. In R. distincta , hydrodynamic thrust is supported by a midleg stroke that begins from a more acute femur angle and involves backward rotations across multiple leg joints. The resulting movement vector deviates from the body axis, potentially enabling lift-like forces, akin to paddling strategies in human kayaking (Jackson 1995; Michael et al. 2009) and animal locomotion (Johansson and Norberg 2001). This configuration would be inefficient for surface-tension-based propulsion but is well-suited to fast-flowing environments where high-frequency strokes with minimal surface contact time are advantageous. The ability of R. distincta to actively control fan protraction and retraction (Andersen 1976; SI Parts 2 and 3B) could allow for greater flexibility during maneuvering and stroke timing, in contrast to the passive deployment described in recent studies (Ortega-Jimenez 2021, 2024). Conversely, G. latiabdominis initiates midleg movement at a less acute femur angle, with primary rotation occurring at the coxa–femur joint, and maintains a nearly straight femur–tibia segment. This configuration enhances wetted surface area, allowing for more efficient surface-tension-based thrust. The leg stroke is longer in duration and follows a nearly parallel trajectory to the body axis, facilitating the formation of asymmetrical dimples critical to curvature force production. The backward movement of the leg, nearly perpendicular to its axis, promotes directional asymmetry in the dimple, optimizing propulsion in still or slow-moving waters. Both species share a distinct ventral setal arrangement that likely contributes to standing and sliding on the water surface. Flattened, overlapping setae form a beam-like structure along the underside of wetted legs, similar to those described in Gerris and Aquarius (Andersen 1976; Perez-Goodwyn 2008). This configuration minimizes surface penetration and drag, offering support without water breakage. The beam’s alignment with body motion also likely facilitates sliding and steering. In the heavier G. latiabdominis , these beams feature hydrophobic nanogrooves that enhance their supporting function. In contrast, the lighter R. distincta lacks such grooves, suggesting lower support demands. Moreover, the spoon-like setae at the tips of R. distincta 's midlegs may help resist displacement by currents, offering a stabilizing advantage in fast-flowing habitats. Conclusion Our results show that two representative species from distinct clades of semiaquatic bugs—each having independently evolved a symmetrical rowing stroke (Andersen 1982; Khila et al. 2014)—rely on different thrust-generation mechanisms, each requiring distinct leg movement patterns, leg lengths, and ventral microstructures. These adaptations enable R. distincta to generate thrust through a combination of hydrodynamic drag and surface tension, while G. latiabdominis relies exclusively on surface-tension-based thrust. The results suggest that the fan in Rhagovelia functions as a “leaky paddle,” producing thrust via hydrodynamic drag. Our study provides new insights into the morphology and anatomy of Rhagovelia `s fan and claw, structures essential to its hydrodynamic thrust mechanism (Andersen 1976; Ortega-Jimenez & Bhamla 2021; Ortega-Jimenez et al. 2024; Santos et al. 2017). In addition, we document ventral setal arrangements involved in surface-tension-based thrust—features that have received little attention in this genus. We also identify shared microstructures in both species used for surface-tension-based thrust, support, and sliding, and we propose hypotheses for the functional roles of these features based on our observations. Many precedent studies on water striders` setae (Gerridae) treat leg micromorphology superficially, and theoretical models often assume a uniform distribution of simple conical hairs on the ventral leg surface. Our findings challenge this simplification by providing detailed morphological data and offering testable hypotheses that can be translated into formal fluid dynamics models. These will enable a quantitative assessment of how specific microstructures contribute to thrust, support, and sliding. Finally, as the observed leg movement patterns and microstructures appear to co-evolve with thrust mechanisms, leg morphology, and habitat, they offer a foundation for phylogenetic studies across taxa. Such studies, especially if they incorporate allometric scaling (Kim et al. 2022, 2024), can shed light on the evolutionary processes behind micromorphological adaptations in semiaquatic bugs. To fully understand these traits, future research should also explore the developmental and genetic basis of setal diversity and arrangement, building on the approaches of Santos et al. (2017) and Finet et al. (2018, 2022). Methods Field sites and study species In January and February of 2020 and 2023, specimen collections and detailed observations of Rhagovelia distincta (body weight: 4–14 mg; Table S1) were made at the Southwestern Research Station, Arizona, USA (SWRS; 31°53′3′′N, 109°12′21′′W). In August and September of 2020, specimen collections and detailed observations of Gerris latiabdominis (16–19 mg; Table S2) were made at Gwanak Mountain, Korea (37°26′42′′N, 126°57′51′′E) and Seoul National University, Korea (37°28′57′′N, 126°96′04′′E), respectively. Each individual was weighed (GEM20 High Precision Digital Milligram Jewelry Scale, Smart Weigh, 0.001 g). Videographic and photographic observations We filmed four types of high speed and standard videos with Sony RX10-III at 959.04 frames per second (fps) and with Chronos 2.1-HD at 1000–4000 fps of individuals in acrylic containers (18 x 18 cm, filled with water): - Type 1: directly from above (85 and 62 movies collected from six and seven individuals of R. distincta and G. latiabdominis , respectively). - Type 2: from the side of various angles (below surface, surface level, and above surface) (249 and 87 movies of R. distincta and G. latiabdominis , respectively). - Type 3: directly from below with light source positioned directly above the container (20 and 60 movies from two and five individuals of R. distincta and G. latiabdominis , respectively) to visualize the shadows on the bottom of the container; shadows correspond to dimples under legs on the water surface. Two variables were extracted from the video types 2 and 3: - Wetted midleg length (mm): maximal leg section in contact with the water surface in the middle of fast thrust strokes when midleg angle to body axis approximates 90°. The wetted midleg consisted of tarsus in R. distincta (6 individuals) and proximal tibia to tarsal tip in G. latiabdominis (6 individuals). - Swimming fan area (mm 2 ): six still frames from six different type 2 clips of R. distincta were used to evaluate the fan surface area (Figure S3). Microscopic observations of leg microstructures Using optical microscopy, we observed the morphology and behavior of the fan in specimens of R. distincta . Using Scanning Electron Microscopy (SEM), we visualized the hair structures on leg sections that interact with water (SI Part 3). Contact angle measurements Contact angle (in degrees; °), height and width (mm) of small droplets on the surface of ventral and dorsal microstructures of tarsomere 3 and tarsomere 2 of R. distincta and G. latiabdominis , respectively, and on the tarsal claw of R. distincta , were measured (with ImageJ 1.53t) in frames of high-speed video (2000 fps; Chronos 2.1-HD Camera, Kron Technologies). The specimens, sprayed with water, were mounted on a micromanipulator (MM-3, Narishige, Japan) parallel to the camera (Video S4). Kinematic profiles of a stroke Detailed kinematic analyses of symmetric strokes by R. distincta and G. latiabdominis were restricted to data extracted with Tracker (https://physlets.org/tracker/) from selected videos: 21 and 12 strokes from six and six individuals for R. distincta and G. latiabdominis , respectively. Cartesian ( x , y ) coordinates of 10 and 9 points on the insect body for R. distincta and G. latiabdominis , respectively (midleg tips were not digitized in G. latiabdominis due to resolution issues) were digitized and subsequently smooth-splined using the “ stats ” package (Everitt & Hothorn 2010; R Core Team 2023) (df = 5 and smoothing parameter = 0.5). To compare the two species, we focused on five aspects (Figure 9) of leg movements during a stroke and extracted the following kinematic variables for each frame, or two consecutive frames, through the thrust phase of a stroke: - Midleg femur angl e (degrees): the angle between the body axis and the femur at the coxae (Figure 9B) was calculated at each frame. The coxal joint is where the major leg angular movement is performed in both species. - Femur-tibia angle (degrees): the angle between the femur and the tibia at the femorotibial joint was calculated at each frame (Figure 9C). - Tibia-tarsus angle (degrees): the angle between the tibia and the tarsus at the tibiotarsal joint was calculated at each frame (Figure 9D). - Midleg angular velocity (degrees/s): calculated by dividing the between-frame difference in midleg femur angles by the latency between the two consecutive frames (i.e., 1/fps). - Leg velocity (mm/s) ( U ): The linear velocity in horizontal plane of the midpoint of the wetted midleg length. It is calculated for each pair of two consecutive frames via dividing the distance traveled between by the latency by the two consecutive frames. Three proxies of “effectiveness” of midleg application for thrust generation during a stroke were calculated (Figure 9) using basic trigonometry and vector algebra: - “Effectiveness” of leg velocity vector`s direction (Figure 9E) (proportion; range 0–1): we determined the proportion of the leg velocity vector (and of R. distincta `s fan protracted under the leg; green vector) employed along the direction parallel to the body movement (blue or violet vectors). Values closer to ‘1’ indicate “more effective” employment of legs on water surface because the backward leg velocity vector is near-parallel to the body axis line ( ) resulting in anteroposterior asymmetry of the dimple crucial for curvature force (i.e., surface tension) contribution to thrust (Steinman at al 2021). Positive values indicate backward velocity vector (blue) that contributes to forward thrust, while negative values indicate forward vector (violet; when legs are dragged along body movement). - “Effectiveness” of leg length`s use (Figure 9F) (proportion; range 0–1): evaluation of the relative length of wetted leg projection (blue) on the line perpendicular to the body movement axis (relative to the actual wetted leg length marked green). In G. latiabdominis , it may be approximately viewed as the effective proportion of the total wetted midleg length (blue) that pushes the surface dimple directly backwards along the leg velocity vector parallel to the body movement direction (blue vector in Figure 9E). Values closer to ‘1’ indicate “more effective” employment of the midleg length pushing the dimple backwards; they also indicate that the fan surface in R. distincta is approximately perpendicular to the body axis ( ), under the assumption that tarsus on the water surface lies approximately within the near-vertical plane with the surface of the R. distincta`s fan under water. - “Effectiveness” of wetted leg orientation (Figure 9G) (degrees): angle θ indicates the orientation of wetted midleg`s main axis (as well as the plane of the fan protracted under the leg, assuming fan surface`s plane includes the longitudinal axis of wetted leg) relative to the leg velocity vector. Angles closer to ‘90°’ indicate “more effective” employment of the full wetted midleg length in pushing the water surface dimple along the leg velocity vector and creating dimple asymmetry along the velocity vector. Under the assumption that longitudinal axis of wetted leg on the water surface approximately lies within the plane of the R. distincta fan`s surface protracted under water, θ values closer to 90° indicate that the angle between the fan surface and the fan movement direction is near perpendicular and hydrodynamic drag from the fan pushing backwards contributes to thrust. We additionally extracted three variables from the body movements : - Body velocity (mm/s): distance (mm) traveled by the body center (position derived from the average of head and abdomen tip positions) between consecutive frames divided by the latency between the two consecutive frames (1/fps). - Body acceleration (mm/s 2 ): rate of change of body velocity derived from each pair of consecutive body velocity values divided by the latency between the two frames. - Net force (µN): body acceleration (mm/s 2 ) multiplied by insect body mass (mg) and by 0.001 for unit conversion. It represents a horizontal vector of net thrust force during a stroke. Kinematic characterization of a stroke The following kinematic variables were extracted from 21 strokes of six individuals of R. distincta and 12 strokes of six individuals of G. latiabdominis (1 value per stroke): - Distance traveled by wetted midleg (mm): sum of frame-by-frame distances of wetted midleg midpoint was measured along the actual trajectory of the midpoint during the thrust phase of a stroke. - Midleg`s stroke amplitude (mm): direct straight-line distance from the initial (beginning of thrust stroke) to the final (end of thrust stroke; when midleg`s velocity vector is no longer opposite to the body velocity vector) positions of wetted midleg midpoint was measured. - Average angular velocity (degrees/s): mean of all frame-by-frame midleg angular velocities calculated within the thrust phase of a stroke. - Average midleg velocity (mm/s): mean of all frame-by-frame leg velocities calculated within the thrust phase of a stroke. - Leg velocity at maximum body acceleration (mm/s): leg velocity that corresponds to the maximum body acceleration (i.e., maximum horizontal net force) within the thrust phase of a stroke. - Maximal body acceleration (mm/s 2 ): maximum value of body acceleration within the thrust phase of a stroke. - Maximal net thrust force (µN) : maximum value of body force within the thrust phase of a stroke. - Final body velocity (mm/s): final value of body velocity (between the last two consecutive frames) in the thrust phase of a stroke. - Midleg thrust duration (ms): latency from the initiation of midleg thrust movements to the moment of their disengagement Statistical analyses All analyses were performed in R version 4.3.1 (R Core Team, 2023). We used linear mixed-effects models (with “individual” as the random factor; “ lme4 ” package by Bates et al . 2015; “ lmerTest ” package by Kuznetsova et al. 2017) to compare the effects of average leg velocity , stroke duration , and distance traveled by wetted midleg on the final body velocity between the two species (interaction with categorical variable “ species ”). However, these independent variables were correlated, and the statistical models would not allow proper evaluations of these effects. Therefore, we extracted principal components using functions fa.parallel and principal from the “ psych ” R package (Revelle 2023) from the pooled data for both species considering all seven kinematic variables: body velocity , maximum acceleration , average angular leg velocity , average leg velocity , midleg`s stroke amplitude , distance traveled by wetted midleg , and stroke duration . We focused on the variables with loading values > 0.75 (Rutherford et al. 1988). Declarations Funding statement This work was supported by the National Research Foundation of Korea(NRF) grant funded by the Korea government(MSIT)(IRIS RS-2024-00343461; RS-2025-00514508); DGIST Start-up Fund Program nr 20200810 and individual mid-career grant 2022R1A2C1006090 of the Ministry of Science, ICT and Future Planning of Korea. Data availability All data generated or analyzed during this study are available in supplementary files. References Andersen NM, A comparative study of locomotion on the water surface in semiaquatic bugs (Insecta, Hemiptera, Gerromorpha), Videnskabelige Meddelelser fra Dansk Naturhistorisk Forening 1976; 139: 337–396. Andersen NM, Phylogenetic inference as applied to the study of evolutionary diversification of semiaquatic bugs (Hemiptera: Gerromorpha), Systematic Zoology 1979; 28: 554–578. Andersen NM, The semiaquatic bugs: Phylogeny, adaptations, biogeography, and classification, Scandinavian Science Press 1982. Andersen NM, Cladistic inference and evolutionary scenarios: locomotory structure, function, and performance in water striders, Cladistics 1995; 11: 279–295. Armisén D, Refki PN, Crumière AJJ, et al., Predator strike shapes antipredator phenotype through new genetic interactions in water striders, Nature Communications 2015; 6: 8153. Armisen D, Viala S, da Rocha Silva Cordeiro I, et al., Transcriptome-based phylogeny of the semi-aquatic bugs (Hemiptera: Heteroptera: Gerromorpha) reveals patterns of lineage expansion in a series of new adaptive zones, Molecular Biology and Evolution 2022; 39(11): msac229. Bates D, Kliegl R, Vasishth S, et al., Parsimonious mixed models, arXiv preprint 2015; arXiv:1506.04967. Bhavikatti SS, Mechanics of solids, New Age International (P) Ltd., Publishers 2010. Bush JWM, Hu DL, Prakash M, The integument of water-walking arthropods: form and function, Advances in Insect Physiology 2008; 34: 117–192. Cheer AY, Koehl MAR, Paddles and rakes: fluid flow through bristled appendages of small organisms, Journal of Theoretical Biology 1987; 129(1): 17–39. Cianferroni F, Santini G, Comparative description of the hairy structures in two endemic Velia species (Insecta: Hemiptera: Heteroptera: Veliidae): V. currens (Fabricius, 1794) and V. gridellii Tamanini, 1947, Acta Zoologica 2014; 95: 347–357. Crisp DJ, The stability of structures at a fluid interface, Transactions of the Faraday Society 1950; 46: 228–235. Crisp DJ, Thorpe WH, The water-protecting properties of insect hairs, Discussions of the Faraday Society 1948; 3: 210–220. Crumière AJJ, Santos ME, Sémon M, et al., Diversity in morphology and locomotory behavior is associated with niche expansion in the semi-aquatic bugs, Current Biology 2016; 26(24): 3336–3342. Damgaard J, Phylogeny of the semiaquatic bugs (Hemiptera-Heteroptera, Gerromorpha), Insect Systematics & Evolution 2008; 39(4): 431–460. Damgaard J, Evolution of the semi-aquatic bugs (Hemiptera: Heteroptera: Gerromorpha) with re-interpretation of the fossil record, Acta Entomologica Musei Nationalis Pragae 2008; 48: 251–268. Everitt BS, Hothorn T, A handbook of statistical analyses using R, Chapman & Hall/CRC 2010. Feng XQ, Gao X, Wu Z, et al., Superior water repellency of water strider legs with hierarchical structures: experiments and analysis, Langmuir 2007; 23: 4892–4896. Finet C, Decaras A, Armisén D, et al., The achaete–scute complex contains a single gene that controls bristle development in the semi-aquatic bugs, Proceedings of the Royal Society B 2018; 285(1892): 20182387. Finet C, Decaras A, Rutkowska M, et al., Leg length and bristle density, both necessary for water surface locomotion, are genetically correlated in water striders, Proceedings of the National Academy of Sciences 2022; 119(9): e2119210119. Gao X, Jiang L, Water-repellent legs of water striders, Nature 2004; 432: 36. Hu DL, Bush JWM, The hydrodynamics of water-walking arthropods, Journal of Fluid Mechanics 2010; 644: 5–33. Hu DL, Chan B, Bush JWM, The hydrodynamics of water strider locomotion, Nature 2003; 424(6949): 663–666. Jackson PS, Performance prediction for Olympic kayaks, Journal of Sports Sciences 1995; 13: 239–245. Ji XY, Wang WJ, Feng XQ, Role of flexibility in the water repellency of water strider legs, Physical Review E 2012; 85: 021607. Johansson C, Norberg UML, Lift-based paddling in diving grebe, Journal of Experimental Biology 2001; 204: 1687–1696. Khila A, Abouheif E, Rowe L, Comparative functional analyses of Ultrabithorax reveal multiple steps and paths to diversification of legs in the adaptive radiation of semi-aquatic insects, Evolution 2014; 68: 2159–2170. Kim W, Amauger J, Ha J, et al., Two different jumping mechanisms of water striders are determined by body size, Proceedings of the National Academy of Sciences 2023; 120: e2219972120. Kim W, Lee JH, Pham TH, et al., Physics of sliding on water explains morphological and behavioral allometry across a wide range of body sizes in water striders (Gerridae), Proceedings of the Royal Society B 2024; 291: 20241357. Koehl MAR, Hairy little legs: feeding, smelling and swimming at low Reynolds numbers, Contemporary Mathematics 1993; 141: 33–64. Koh JS, Yang E, Jung GP, et al., Jumping on water: surface tension–dominated jumping of water striders and robotic insects, Science 2015; 349: 517–521. Kuznetsova A, Brockhoff PB, Christensen RHB, lmerTest package: tests in linear mixed effects models, Journal of Statistical Software 2017; 82: 1–26. Lamont EI, Emlet RB, Permanently fused setules create unusual folding fans used for swimming in cyprid larvae of barnacles, Biological Bulletin 2018; 235: 185–194. Michael JS, Smith R, Rooney KB, Determinants of kayak paddling performance, Sports Biomechanics 2009; 8: 167–179. Molnar K, Vas LM, Czigany T, Strength of electrospun single nanofibers through modeling tensile behavior of the nanofibrous mat, Composites Part B: Engineering 2012; 43: 15–21. Ortega-Jimenez V, Bhamla S, Extreme maneuvering and hydrodynamics of Rhagovelia water striders, APS Division of Fluid Dynamics Meeting Abstracts 2021; E13-007. Ortega-Jimenez V, Bhamla S, Ultrafast elastocapillary fans control agile maneuvering in ripple bugs and robots, bioRxiv preprint 2024. Perez Goodwyn PJ, Voigt D, Fujisaki K, Skating and diving: changes in functional morphology of the setal and microtrichial cover during ontogenesis in Aquarius paludum Fabricius (Heteroptera, Gerridae), Journal of Morphology 2008; 269: 734–744. R Core Team, R: A language and environment for statistical computing, R Foundation for Statistical Computing 2023. Reaver NGF, Kaplan DA, Mattson RA, et al., Hydrodynamic controls on primary producer communities in spring-fed rivers, Geophysical Research Letters 2019; 46: 4715–4725. Refki PN, Khila A, Key patterning genes contribute to leg elongation in water striders, EvoDevo 2015; 6: 14. Revelle W, How to use the psych package for regression and mediation analysis, Northwestern University 2023. Rutherford GSW, Hair JF, Anderson RE, et al., Multivariate Data Analysis with Readings, The Statistician 1988; 37: 484. Santos ME, Le Bouquin A, Crumière AJJ, et al., Taxon-restricted genes at the origin of a novel trait allowing access to a new environment, Science 2017; 358: 386–390. Steinman T, Cribellier A, Casas J, Singularity of the water strider propulsion mechanisms, Journal of Fluid Mechanics 2021; 915: A118. Sun P, Zhao M, Jiang J, et al., The study of dynamic force acted on water strider leg departing from water surface, AIP Advances 2018; 8: 015228. Uesugi K, Mayama H, Morshima K, Proposal of a water-repellency model of water strider and its verification by considering directly measured strider leg-rowing force, Journal of Photopolymer Science and Technology 2020; 33: 185–192. Uesugi K, Water-repellency model of the water strider Aquarius paludum paludum by the curved structure of leg micro-hairs, Journal of Photopolymer Science and Technology 2021; 34: 393–399. Wang Q, Yao X, Liu H, et al., Self-removal of condensed water on the legs of water striders, Proceedings of the National Academy of Sciences 2015; 112: 9247–9252. Wei PJ, Chen SC, Lin JF, Adhesion forces and contact angles of water strider legs, Langmuir 2009; 25: 1526–1528. Wingert MC, Jiang Z, Chen R, et al., Strong size-dependent stress relaxation in electrospun polymer nanofibers, Journal of Applied Physics 2017; 121: 0151031-6. Wu X, Zhu Y, Gradient and lamellar heterostructures for superior mechanical properties, MRS Bulletin 2021; 46: 244–249. Xue Y, Yuan H, Su W, et al., Enhanced load-carrying capacity of hairy surface floating on water, Proceedings of the Royal Society A 2014; 470: 20130832. Yang E, Son JH, Lee SI, et al., Water striders adjust leg movement speed to optimize takeoff velocity for their morphology, Nature Communications 2016; 7: 13698. Ye Z, Zhu G, Damgaard J, et al., Phylogeography of a semiaquatic bug, Microvelia horvathi (Hemiptera: Veliidae): An evaluation of historical, geographical and ecological factors, Scientific Reports 2016; 6: 21930. Zenkert D, An Introduction to Sandwich Construction, Chameleon Press 1995. Zhang X, Zhang C, Ma P, Mechanical properties of hollow polyester monofilament: Compression and tension behaviors, Journal of Engineered Fibers and Fabrics 2019; 14: 1–6. Additional Declarations No competing interests reported. Supplementary Files VideoS1.mp4 VideoS2.mp4 VideoS3.mp4 VideoS4.mp4 Supplementaldata1.xlsx Supplementaldata2.xlsx Supplementaryinformation.docx Cite Share Download PDF Status: Published Journal Publication published 29 Dec, 2025 Read the published version in Scientific Reports → Version 1 posted Editorial decision: Revision requested 29 Jul, 2025 Reviews received at journal 18 Jul, 2025 Reviews received at journal 15 Jul, 2025 Reviewers agreed at journal 07 Jul, 2025 Reviewers agreed at journal 07 Jul, 2025 Reviewers invited by journal 07 Jul, 2025 Editor assigned by journal 07 Jul, 2025 Editor invited by journal 18 Jun, 2025 Submission checks completed at journal 17 Jun, 2025 First submitted to journal 16 Jun, 2025 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-6903117","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":481837169,"identity":"a3dfa1eb-89ba-4258-89ad-011a846b9669","order_by":0,"name":"Sang Yun Bang","email":"","orcid":"","institution":"Seoul National University","correspondingAuthor":false,"prefix":"","firstName":"Sang","middleName":"Yun","lastName":"Bang","suffix":""},{"id":481837170,"identity":"ce2c2999-29ef-4f66-85bd-1dc8acc4e683","order_by":1,"name":"Woojoo Kim","email":"","orcid":"","institution":"Seoul National University","correspondingAuthor":false,"prefix":"","firstName":"Woojoo","middleName":"","lastName":"Kim","suffix":""},{"id":481837171,"identity":"6eebcdac-b3e1-4ab8-805d-f37cb2a1bad6","order_by":2,"name":"Jeongseop Lee","email":"","orcid":"","institution":"Seoul National University","correspondingAuthor":false,"prefix":"","firstName":"Jeongseop","middleName":"","lastName":"Lee","suffix":""},{"id":481837172,"identity":"f94443e3-247a-4e29-a1eb-4d74fd1ff17b","order_by":3,"name":"Jinseok Park","email":"","orcid":"","institution":"Seoul National University","correspondingAuthor":false,"prefix":"","firstName":"Jinseok","middleName":"","lastName":"Park","suffix":""},{"id":481837173,"identity":"7830f8d6-f22d-477a-931e-80dee2de3ebd","order_by":4,"name":"Versha Khare","email":"","orcid":"","institution":"Seoul National University","correspondingAuthor":false,"prefix":"","firstName":"Versha","middleName":"","lastName":"Khare","suffix":""},{"id":481837174,"identity":"c5b68ce0-7449-4b6f-8d83-de2b83fd1501","order_by":5,"name":"Sang-im Lee","email":"","orcid":"","institution":"DGIST","correspondingAuthor":false,"prefix":"","firstName":"Sang-im","middleName":"","lastName":"Lee","suffix":""},{"id":481837175,"identity":"b834be0a-d6e7-499c-b261-909d9abffca7","order_by":6,"name":"Piotr Grzegorz Jablonski","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAABK0lEQVRIiWNgGAWjYDACZhDBw5DAL8HABmLKQETgUli1MDaAtEjOgGjhgYjg08IAUZBgcAOmhQGmBQcwZ2c+/rhCxi7P+HbzsQcfamx4ICJ/7PLk3bkTGD7uqUXXYtnMlth4hie52OzOsXTDGcfSeMAiZ9uSiw0P825gnPHsOLoWg8M8ho0NPMyJ227kmEnzNhzmAYs0NjAnbmzm3cDMc+AYDi31iZtn5H+T/gvT0vCnnpCWw4kbJHLYpBnhWtgOJ85nBmupweaXmQ08x4sl7hwzk+yB+mVmY9vxxA1ALQdnHDiAEWL8hw98bOypzuOf3fxM4keNjRxYpOFPdeL8/rMbH3w4UIfhMBDB2IMuAmYAzQeiw1i1MPzApkW+AUxh2DIKRsEoGAUjDgAAs7p1aqLMI24AAAAASUVORK5CYII=","orcid":"","institution":"Seoul National University","correspondingAuthor":true,"prefix":"","firstName":"Piotr","middleName":"Grzegorz","lastName":"Jablonski","suffix":""}],"badges":[],"createdAt":"2025-06-16 08:08:46","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-6903117/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-6903117/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1038/s41598-025-28453-0","type":"published","date":"2025-12-29T15:58:12+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":86396824,"identity":"3afe1a2d-d682-4786-91d8-14b221d82056","added_by":"auto","created_at":"2025-07-10 08:04:11","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":1697716,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003e\u003cstrong\u003eRhagovelia distincta\u003c/strong\u003e\u003c/em\u003e\u003cstrong\u003e during the thrust phase. \u003c/strong\u003e(A1–A6) Side view above the water surface, showing interactions between the leg and water surface. (B1–B6) Side view below the surface, highlighting the motion of the swimming fan during the thrust phase of a stroke. (C1–C6) Top view capturing body and leg positions throughout the thrust phase. (D1–D6) Bottom view from beneath the container, showing shadows cast by the body and water-surface dimples. Abbreviations: Fe – femur; Tb – tibia; T1–T3 – tarsomeres 1–3; WL – wetted midleg length; Wb – wave bow. Panels B4 and C3: blue and orange annotations indicate interpreted differences in dimple shape based on comparisons with \u003cem\u003eG. latiabdominis\u003c/em\u003e (see Figure 2 and Figure S2).\u003c/p\u003e","description":"","filename":"image1.png","url":"https://assets-eu.researchsquare.com/files/rs-6903117/v1/68656e2ec0bd9489d4a26be8.png"},{"id":86396821,"identity":"26197f48-0545-4ace-be56-2f703261d3f7","added_by":"auto","created_at":"2025-07-10 08:04:11","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":1462237,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003e\u003cstrong\u003eGerris latiabdominis\u003c/strong\u003e\u003c/em\u003e\u003cstrong\u003e during the thrust phase.\u003c/strong\u003e (A1–A6) Side view above the water surface, showing midleg motion during thrust. (B1–B6) Side view above the water, focused on midleg interaction with the water surface. (C1–C6) Top view capturing body and leg positions throughout the thrust phase. (D1–D6) Bottom view from beneath the container, showing shadows cast by the body and water-surface dimples. Abbreviations: Tb – tibia; T1–T2 – tarsomeres 1 and 2; WL – wetted midleg length; Wb – wave bow.\u003c/p\u003e","description":"","filename":"image2.png","url":"https://assets-eu.researchsquare.com/files/rs-6903117/v1/bd08d760bbe5f9f2e2e6e074.png"},{"id":86396822,"identity":"300069c5-1ad7-446c-8c5d-9abc4ed8d0df","added_by":"auto","created_at":"2025-07-10 08:04:11","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":441586,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eSchematic drawings of different types of setae found on leg sections of \u003c/strong\u003e\u003cem\u003e\u003cstrong\u003eRhagovelia distincta\u003c/strong\u003e\u003c/em\u003e\u003cstrong\u003e and \u003c/strong\u003e\u003cem\u003e\u003cstrong\u003eGerris latiabdominis\u003c/strong\u003e\u003c/em\u003e\u003cstrong\u003ethat interact with the water surface. \u003c/strong\u003e(A) Microsetae, \u003cem\u003em\u003c/em\u003e. (B) Macrosetae 2, \u003cem\u003eM2\u003c/em\u003e. (C) Grooming comb, \u003cem\u003eG\u003c/em\u003e. (D) Cuspidate setae, \u003cem\u003eC\u003c/em\u003e. (E) Stumped setae, \u003cem\u003eS\u003c/em\u003e. (F) Macrosetae 2, \u003cem\u003eM2\u003c/em\u003e. (G) Hook setae 1, \u003cem\u003eH1.\u003c/em\u003e(H) Hook setae 2, \u003cem\u003eH2\u003c/em\u003e. (I) Spoon setae, \u003cem\u003eSp\u003c/em\u003e. (J) Obtuse setae, \u003cem\u003eO\u003c/em\u003e. (K–L) Macrosetae 1, \u003cem\u003eM1\u003c/em\u003e. (M) Thorn setae 1, \u003cem\u003eT1\u003c/em\u003e. (N) Thorn setae 2, \u003cem\u003eT2\u003c/em\u003e. (O) Thorn setae 3, \u003cem\u003eT3\u003c/em\u003e. (P) Web setae, \u003cem\u003eW\u003c/em\u003e. (Q) Leaf-blade setae, \u003cem\u003eL\u003c/em\u003e. (R) Leaf-like setae, \u003cem\u003el\u003c/em\u003e. (S) Grass-blade setae, \u003cem\u003eg\u003c/em\u003e. Alphabetical labels correspond to Table S3. Setae found in both species are labeled in \u003cstrong\u003eblack\u003c/strong\u003e; those found only in \u003cem\u003eR. distincta\u003c/em\u003eare in \u003cstrong\u003eblue\u003c/strong\u003e; and those found only in \u003cem\u003eG. latiabdominis\u003c/em\u003e are in \u003cstrong\u003egreen\u003c/strong\u003e. Note that scales vary between panels and are specified for each seta type. Descriptions of all seta types are provided in Supplementary Information Part 3C.\u003c/p\u003e","description":"","filename":"image3.png","url":"https://assets-eu.researchsquare.com/files/rs-6903117/v1/bae70806b7fd689951e59b62.png"},{"id":86396833,"identity":"6e95aa33-ef17-4d92-98ae-2faab1099a6d","added_by":"auto","created_at":"2025-07-10 08:04:12","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":1854829,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eScanning electron microscopy summary of leg microstructures in \u003c/strong\u003e\u003cem\u003e\u003cstrong\u003eRhagovelia distincta\u003c/strong\u003e\u003c/em\u003e\u003cstrong\u003e. \u003c/strong\u003e(A) Hindleg: (A1) a row of spoon setae, \u003cem\u003eSp\u003c/em\u003e, on proximal ventral tarsomere 2; (A2) anterio-ventral view of the tarsal joint (between tarsomeres 1 and 2), showing overlapping \u003cem\u003eSp\u003c/em\u003esetae with flattened tips, flanked by macrosetae 2, \u003cem\u003eM2\u003c/em\u003e; (A3–A5) anterio-ventral views from proximal tarsomere 2 to the tarsal tip, showing a progressively beam-like structure formed by overlapping \u003cem\u003eSp\u003c/em\u003e setae on the ventral side; (A6) a row of spoon setae, \u003cem\u003eSp,\u003c/em\u003e on distal tarsomere 2; (A7) posterior-lateral view of distal tarsomere 2 with a row of spoon setae, \u003cem\u003eSp\u003c/em\u003e, along the water-interacting ventral side (yellow shading in A2–A7). (B) Midleg: (B1) schematic of the pretarsal swimming fan used for hydrodynamics-based thrust; (B2) anterio-ventral view of tarsomere 3 showing the fan and anterior claw extending from the cleft between two lobes (posterior claw not visible); (B3) protracted fan in water showing hierarchical structure of setae and setulae; (B4) anterior view of folded fan, highlighting relative thickness of setae and setulae; (B5) “board-like” cross-sectional shape of fan; (B6) cross-section of a fan seta showing internal layers, hollow core with pillars, and outer layers; (B7) surface of the anterior claw extruding from the cleft surrounded by hook setae (anterior lobe`s \u003cem\u003eH1\u003c/em\u003e setae visible); (B8) cross-section of the claw; (B9) Ventral edges of the cleft with rows of \u003cem\u003eH1 \u003c/em\u003e(anterior lobe) and \u003cem\u003eH2 \u003c/em\u003e(posterior lobe) setae; (B10) anterior view of the distal portion of tarsomere 3; (B11) close-up of the anterior lobe tip showing long, flattened modified \u003cem\u003eH1 \u003c/em\u003esetae resembling spoon setae, \u003cem\u003eSp\u003c/em\u003e, and internal cleft wall lined with \u003cem\u003eH2\u003c/em\u003e; tip of anterior claw also visible (posterior lobe removed); (B12) lateral view of tarsomere 2; (B13) ventral view of tarsomere 2 showing orderly rows of \u003cem\u003eSp \u003c/em\u003eand \u003cem\u003eH1 \u003c/em\u003esetae. (C)\u003cstrong\u003e \u003c/strong\u003eForeleg: (C1) posterior view showing sparse cuspidate setae, \u003cem\u003eC\u003c/em\u003e, dorsal macrosetae 2, \u003cem\u003eM2\u003c/em\u003e, and ventral hook setae, \u003cem\u003eH1\u003c/em\u003e; (C2) anterio-ventral view showing microsetae, \u003cem\u003em\u003c/em\u003e, ventral hook setae,\u003cem\u003e H1\u003c/em\u003e, and a row of spoon setae, \u003cem\u003eSp\u003c/em\u003e along the ventral water-interacting surface; (C3) close-up of spoon setae, \u003cem\u003eSp\u003c/em\u003e, near the claw base; (C4) array of spoon setae, \u003cem\u003eSp,\u003c/em\u003e at the tarsal tip. Color-shaded regions in SEM panels denote ventral (water-interacting) leg surfaces. Additional SEMs are provided for hindlegs in Figure S19, midlegs in Figures S11–S17, and forelegs in Figure S19.\u003c/p\u003e","description":"","filename":"image4.png","url":"https://assets-eu.researchsquare.com/files/rs-6903117/v1/63effae5f71469f07ea04660.png"},{"id":86396834,"identity":"f74ae378-f19a-4d7c-b9a2-2e0330c598f6","added_by":"auto","created_at":"2025-07-10 08:04:12","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":2725183,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eScanning electron microscopy summary of leg microstructures in \u003c/strong\u003e\u003cem\u003e\u003cstrong\u003eGerris latiabdominis\u003c/strong\u003e\u003c/em\u003e\u003cstrong\u003e. \u003c/strong\u003e(A) Hindleg: (A1) ventral view of femorotibial joint covered with leaf-blade setae, \u003cem\u003eL\u003c/em\u003e; (A2) ventral proximal tibia showing overlapping distal sections of \u003cem\u003eL\u003c/em\u003esetae forming a “beam-like” structure believed to support the insect on water; (A3) same as A2, but from a different preparation; \u003cem\u003eL\u003c/em\u003e setae appear more randomly bent due to cleaning and drying procedures. A row of large thorn setae 3, \u003cem\u003eT3\u003c/em\u003e, runs posterior to the \u003cem\u003eL\u003c/em\u003e setae, with macrosetae 2, \u003cem\u003eM2\u003c/em\u003e, present on the leg`s posterior side; (A4) lateral view of distal tibia with dense microsetae, \u003cem\u003em\u003c/em\u003e, macrosetae 1, \u003cem\u003eM1\u003c/em\u003e, thorn setae 1, \u003cem\u003eT1\u003c/em\u003e, and a ventral row of thorn setae 3, \u003cem\u003eT3,\u003c/em\u003e on the water-interacting surface; (A5) ventral view of tarsomere 1 with a longitudinal row of leaf-blade setae, \u003cem\u003eL,\u003c/em\u003e and an adjacent row of thorn setae 3, \u003cem\u003eT3\u003c/em\u003e; (A6) ventral view of tibiotarsal joint with arrays of leaf-like setae, \u003cem\u003el,\u003c/em\u003e and thorn setae 3, \u003cem\u003eT3,\u003c/em\u003eextending from distal tibia; (A7) ventral view of tarsomere 2 with a continuing row of leaf-blade setae, \u003cem\u003eL\u003c/em\u003e, and an adjacent row of thorn setae 3, \u003cem\u003eT3.\u003c/em\u003e(A8) close-up of ventral overlapping \u003cem\u003eL\u003c/em\u003e setae forming a flat “beam-like” surface with nano-grooves running longitudinally; (A9) ventral view of tarsal tip with an array of leaf-like setae, \u003cem\u003el\u003c/em\u003e. (B) Midleg: (B1) overview of the \"gaps and rows\" arrangement on ventral midleg segments involved in thrust generation (purple shading). Top panel: full tibia and tarsus; middle panel: tarsus with one main longitudinal gap and two less distinct gaps; bottom panel: clearer visualization of the three gaps, each flanked by linear setal rows. (B2) ventral view of intermediate tibia with macrosetae 2, \u003cem\u003eM2\u003c/em\u003e, flanked by rows of macrosetae 1, \u003cem\u003eM1\u003c/em\u003e, separated by noticeable gaps; (B3) distal tibia with continuing macrosetae 2,\u003cem\u003e M2\u003c/em\u003e, and adjacent rows of thorn setae 2, \u003cem\u003eT2\u003c/em\u003e, which resemble \u003cem\u003eM1\u003c/em\u003e; (B4–B6) anteroventral view of tarsomere 1, with a row of cuspidate setae, \u003cem\u003eC\u003c/em\u003e, adjacent rows of thorn setae 2, \u003cem\u003eT2\u003c/em\u003e, and a gap in between; (B7–B9) ventral view of tarsomere 2 with a continuing row of cuspidate setae, \u003cem\u003eC,\u003c/em\u003e and two posterior rows of thorn setae 2, \u003cem\u003eT2;\u003c/em\u003evisible are the main gap between \u003cem\u003eC \u003c/em\u003eand \u003cem\u003eT2\u003c/em\u003e, and a narrow gap between the two \u003cem\u003eT2 \u003c/em\u003erows; (B10) anterior view of tarsomere 2 (B11) clear view of the main gap between \u003cem\u003eC \u003c/em\u003eand \u003cem\u003eT2 \u003c/em\u003erows; (B12) tarsal tip with an array of grass-blade setae, \u003cem\u003eg\u003c/em\u003e. (C)\u003cstrong\u003e \u003c/strong\u003eForeleg: (C1) grooming comb, \u003cem\u003eG,\u003c/em\u003e and stumped setae, \u003cem\u003eS\u003c/em\u003e, on dorsal tibiotarsal joint; (C2) anterior-lateral view of proximal tarsus with stumped setae, \u003cem\u003eS,\u003c/em\u003e on lateral side, and overlapping grass-blade,\u003cem\u003e g,\u003c/em\u003e and web setae, \u003cem\u003eW,\u003c/em\u003e on the ventral water-interacting surface; (C3) posterior-lateral view of distal tarsus with grass-blade,\u003cem\u003e g,\u003c/em\u003e and web setae, \u003cem\u003eW\u003c/em\u003e; (C4) close-up of overlapping distal tip of grass-blade, \u003cem\u003eg\u003c/em\u003e. Color-shaded regions in SEM panels denote ventral (water-interacting) leg surfaces. Additional SEMs are provided for hindlegs in Figure S23, midlegs in Figures S20 and S21, and forelegs in Figure S22.\u003c/p\u003e","description":"","filename":"image5.png","url":"https://assets-eu.researchsquare.com/files/rs-6903117/v1/d517d306478c8f928e5482f3.png"},{"id":86396829,"identity":"2f549394-903e-4faf-bd4d-917b0b0ee5f4","added_by":"auto","created_at":"2025-07-10 08:04:12","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":627610,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eContact angle and droplet shape on distal tarsus of \u003c/strong\u003e\u003cem\u003e\u003cstrong\u003eRhagovelia distincta \u003c/strong\u003e\u003c/em\u003e\u003cstrong\u003eand \u003c/strong\u003e\u003cem\u003e\u003cstrong\u003eGerris latiabdominis\u003c/strong\u003e\u003c/em\u003e\u003cstrong\u003e. \u003c/strong\u003eTop panels are example images of water droplets on different leg surfaces. Middle panels contain corresponding SEM images showing dorsal hair layers (A, D), ventral hair layers (B, E), and the surface of \u003cem\u003eR. distincta\u003c/em\u003e`s claw (C). Bottom panels are time-course changes (relative to droplet disappearance, set as 100%) in droplet shape index (width/height, yellow squares) and contact angle (degrees) measured on the proximal (orange diamonds) and distal (purple circles) sides of the droplet. (A) Contact angle on \u003cem\u003eR. distincta\u003c/em\u003e`s dorsal midleg tarsus (tarsomere 3). (B) Contact angle on \u003cem\u003eR. distincta\u003c/em\u003e`s ventral midleg tarsus (tarsomere 3). (C) Contact angle on \u003cem\u003eR. distincta\u003c/em\u003e`s claw. (D) Contact angle on \u003cem\u003eG. latiabdominis\u003c/em\u003e`s dorsal midleg tarsus (tarsomere 2). (E) Contact angle on \u003cem\u003eG. latiabdominis\u003c/em\u003e`s ventral midleg tarsus (tarsomere). For additional details, see Figure S24 and commentary in SI Part 4, and Video S4.\u003c/p\u003e","description":"","filename":"image6.png","url":"https://assets-eu.researchsquare.com/files/rs-6903117/v1/eaada4db3e991ccdd29a2f8c.png"},{"id":86396823,"identity":"2760b583-e8b3-4feb-80b9-10c158eaa774","added_by":"auto","created_at":"2025-07-10 08:04:11","extension":"png","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":415226,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eKinematic profiles of \u003c/strong\u003e\u003cem\u003e\u003cstrong\u003eRhagovelia distincta \u003c/strong\u003e\u003c/em\u003e\u003cstrong\u003eand \u003c/strong\u003e\u003cem\u003e\u003cstrong\u003eGerris latiabdominis\u003c/strong\u003e\u003c/em\u003e\u003cstrong\u003e during a stroke. \u003c/strong\u003e(A) Midleg femur angle (degrees ± SD): angle between femur and body axis; (B) Midleg angular velocity (degrees/s ± SD), derived from A. (C) Femur-tibia angle (degrees ± SD). (D) Tibia-tarsus angle (degrees ± SD). (E) Leg velocity (mm/s ± SD) along the direction of the wetted midleg trajectory on the water surface. (F) “Effectiveness” of leg velocity vector`s direction. (G) “Effectiveness” of leg length`s use. (H) “Effectiveness” of wetted leg orientation. (I) Body velocity (mm/s ± SD) along the body movement axis. (J) Body acceleration (mm/s\u003csup\u003e2\u003c/sup\u003e ± SD), derived from I. Insets in (A–D) illustrate how each angle was measured: dark blue indicates fast strokes of \u003cem\u003eR. distincta\u003c/em\u003e, light blue indicates slow strokes of \u003cem\u003eR. distincta\u003c/em\u003e, dark green indicates fast strokes of \u003cem\u003eG. latiabdominis\u003c/em\u003e, and light green indicates slow strokes of \u003cem\u003eG. latiabdominis\u003c/em\u003e. Period of high “effectiveness” in (F and H) and (G) illustrate range of ± 0.1 and ± 10°, respectively, from each maximal “effectiveness” observed. Gray-shaded regions indicate time intervals where data from all individuals were included in the average; standard deviation is shown only when n \u0026gt; 4. Sample sizes: 21 strokes from six individuals of \u003cem\u003eR. distincta\u003c/em\u003e and 12 strokes from six individuals of \u003cem\u003eG. latiabdominis\u003c/em\u003e.\u003cbr\u003e\n\u003c/p\u003e","description":"","filename":"image7.png","url":"https://assets-eu.researchsquare.com/files/rs-6903117/v1/2c342b977f1f69944f1eccb8.png"},{"id":86396825,"identity":"06cab1db-d04c-4a78-b127-fd5459fd3528","added_by":"auto","created_at":"2025-07-10 08:04:11","extension":"png","order_by":8,"title":"Figure 8","display":"","copyAsset":false,"role":"figure","size":296580,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eComparative kinematics between \u003c/strong\u003e\u003cem\u003e\u003cstrong\u003eRhagovelia distincta \u003c/strong\u003e\u003c/em\u003e\u003cstrong\u003eand \u003c/strong\u003e\u003cem\u003e\u003cstrong\u003eGerris latiabdominis \u003c/strong\u003e\u003c/em\u003e\u003cstrong\u003eduring a single thrust stroke. \u003c/strong\u003eEach data point represents one of 21 (\u003cem\u003eR. distincta\u003c/em\u003e) and 12 (\u003cem\u003eG. latiabdominis\u003c/em\u003e) stroke observations from six individuals per species.\u003cstrong\u003e \u003c/strong\u003e(A) Body movement variables: (A1) final body velocity (mm/s); (A2) maximal body acceleration (mm/s\u003csup\u003e2\u003c/sup\u003e). (B) Midleg movement velocity variables: (B1) average femur angular velocity (degrees/s); (B2) average leg velocity on the water surface (mm/s). (C) Midleg movement distance and time variables: (C1) stroke amplitude of the midleg (mm); (C2) distance traveled by the wetted midleg (mm); (C3) stroke duration (s). (D) Principal Component Analysis (PCA): scatterplot of strokes based on two principal components (Table 1) extracted from variables in panels A–C. Ellipses represent 95% confidence intervals: \u003cem\u003eR. distincta\u003c/em\u003e in blue, \u003cem\u003eG. latiabdominis \u003c/em\u003ein green. RC1 corresponds primarily to distance and speed; RC2 to stroke duration and angular velocity. Shaded bands on x and y axes illustrate the ranges of observed values for each species.\u003c/p\u003e","description":"","filename":"image8.png","url":"https://assets-eu.researchsquare.com/files/rs-6903117/v1/e8d9db54f9d0cf451e5cfc9d.png"},{"id":86396835,"identity":"d2653fed-f43b-448b-a1bc-05f67acdbdd7","added_by":"auto","created_at":"2025-07-10 08:04:12","extension":"png","order_by":9,"title":"Figure 9","display":"","copyAsset":false,"role":"figure","size":361646,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eGraphical definitions of variables extracted from video recordings of thrust-generating strokes. \u003c/strong\u003e(A) Digitized tracking points on the insect body and midleg. (B–D) Joint angles: (B) femur angle relative to the body axis, (C) femur-tibia angle, and (D) tibia-tarsus angle. (E–G) Indices of “effectiveness” for midleg orientation and movement: (E) effectiveness of leg velocity direction, (F) effectiveness of leg length`s use, (G) effectiveness of wetted leg orientation. Further details are provided in the ‘Materials and Methods’ section.\u003c/p\u003e","description":"","filename":"image9.png","url":"https://assets-eu.researchsquare.com/files/rs-6903117/v1/6ec540213a6cac63d7f1a25b.png"},{"id":99545558,"identity":"cea7f095-9499-4aec-8c11-8da38d81f65a","added_by":"auto","created_at":"2026-01-05 16:08:34","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":12749858,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6903117/v1/26d7949b-73f0-47a8-a274-31dd0b111052.pdf"},{"id":86396820,"identity":"333d959a-a514-4801-bf42-aa54b12bedd0","added_by":"auto","created_at":"2025-07-10 08:04:11","extension":"mp4","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":3490713,"visible":true,"origin":"","legend":"","description":"","filename":"VideoS1.mp4","url":"https://assets-eu.researchsquare.com/files/rs-6903117/v1/72cb947116fc886dbc352372.mp4"},{"id":86396832,"identity":"f55c0e93-6de1-408c-936f-d2b522ddc07d","added_by":"auto","created_at":"2025-07-10 08:04:12","extension":"mp4","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":3356009,"visible":true,"origin":"","legend":"","description":"","filename":"VideoS2.mp4","url":"https://assets-eu.researchsquare.com/files/rs-6903117/v1/41910990af7053b54452247e.mp4"},{"id":86396826,"identity":"163c89b3-bdbd-4481-989b-b8e5bd38b459","added_by":"auto","created_at":"2025-07-10 08:04:11","extension":"mp4","order_by":3,"title":"","display":"","copyAsset":false,"role":"supplement","size":7090566,"visible":true,"origin":"","legend":"","description":"","filename":"VideoS3.mp4","url":"https://assets-eu.researchsquare.com/files/rs-6903117/v1/593f1dd164d94d19b5915abf.mp4"},{"id":86396827,"identity":"cfea47dc-f3a2-48bc-9792-694e70231ec3","added_by":"auto","created_at":"2025-07-10 08:04:11","extension":"mp4","order_by":4,"title":"","display":"","copyAsset":false,"role":"supplement","size":814642,"visible":true,"origin":"","legend":"","description":"","filename":"VideoS4.mp4","url":"https://assets-eu.researchsquare.com/files/rs-6903117/v1/f37de13c5cd76c9be1e1443c.mp4"},{"id":86396828,"identity":"5e75a017-4d61-41b5-8a72-b7e072c44a0a","added_by":"auto","created_at":"2025-07-10 08:04:11","extension":"xlsx","order_by":5,"title":"","display":"","copyAsset":false,"role":"supplement","size":3720710,"visible":true,"origin":"","legend":"","description":"","filename":"Supplementaldata1.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-6903117/v1/19b4bff26d7e8f59c3e035cc.xlsx"},{"id":86398023,"identity":"04f1a820-1024-40f4-8fb4-c0dcdf909e90","added_by":"auto","created_at":"2025-07-10 08:12:12","extension":"xlsx","order_by":6,"title":"","display":"","copyAsset":false,"role":"supplement","size":3071792,"visible":true,"origin":"","legend":"","description":"","filename":"Supplementaldata2.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-6903117/v1/a27cef11ae7e2966a297c2c6.xlsx"},{"id":86396837,"identity":"d6cf3df5-2d22-4f63-ba39-922a3345dea2","added_by":"auto","created_at":"2025-07-10 08:04:26","extension":"docx","order_by":7,"title":"","display":"","copyAsset":false,"role":"supplement","size":258247078,"visible":true,"origin":"","legend":"","description":"","filename":"Supplementaryinformation.docx","url":"https://assets-eu.researchsquare.com/files/rs-6903117/v1/7c7f44ebe3cba64499b52025.docx"}],"financialInterests":"No competing interests reported.","formattedTitle":"Distinct kinematics and micromorphology for symmetrical rowing and sliding on water in ripple bugs and water striders","fulltext":[{"header":"Introduction","content":"\u003cp\u003eLocomotion on the water surface presents a unique set of physical challenges that have driven repeated and diverse evolutionary solutions among insects. The semiaquatic bugs (Gerromorpha) provide an example of adaptive radiation into this novel environment. These insects evolved distinct morphological, behavioral, and anatomical traits that enable movement on the air-water interface. These adaptations reflect multiple, lineage-specific strategies for solving similar functional problems\u0026ndash;offering a model system for studying how alternative solutions involving morphology and behavior evolve under ecological constraints of physical environment\u0026ndash;the water surface. However, the details of the co-evolutionary and functional interplay between leg microstructures, stroke kinematics, and thrust mechanics across major independent lineages of this adaptive radiation remain poorly understood. To provide more insights into this question, we compare two major independently evolved solutions for symmetrical rowing in Gerromorpha. Here, we focus on how alternative physical mechanisms\u0026ndash;drag-based versus surface-tension-based thrust\u0026ndash;are realized through contrasting yet functionally convergent morphologies.\u003c/p\u003e\n\u003cp\u003eNotably, two lineages\u0026mdash;Veliidae (a polyphyletic family typically found in fast-flowing streams) and Gerridae (typically inhabiting slow or still waters)\u0026mdash;have independently evolved symmetrical backward rowing by midlegs for forward thrust (Andersen 1982; Crumiere et al. 2016; Khila et al. 2014). While both rely on midlegs for thrust and fore- and hindlegs for support and sliding, their physical modes of thrust generation differ: Veliidae exploit hydrodynamic drag (and potentially lift), while Gerridae generate thrust primarily through surface tension forces (Andersen 1976; Steinmann et al. 2021). Despite their independent origins and these contrasting mechanisms, the leg kinematics and microstructural adaptations that support these behaviors have not been systematically compared.\u003c/p\u003e\n\u003cp\u003eWithin Veliidae, species such as those in the genus Rhagovelia possess specialized midleg pretarsal structures known as swimming fans, which function as oars (Andersen 1976, 1982; Santos et al. 2017; Ortega-Jimenez \u0026amp; Bhamla 2021; Ortega-Jimenez et al. 2024). These structures are assumed to be actively controlled via a claw retractor muscle (Andersen 1976), though recent observations of isolated fans spreading in water have led to the hypothesis of passive elasto-capillary spreading (Ortega-Jimenez \u0026amp; Bhamla 2021; Ortega-Jimenez et al. 2024). The nature of fan manipulation in intact, behaving animals remains unresolved. Similarly, the ventral microstructures on midlegs (involved in thrust) and on fore- and hindlegs (involved in support and sliding) have not been systematically examined in Rhagovelia, though their role in generating surface tension-based forces via water dimples is likely.\u003c/p\u003e\n\u003cp\u003eIn contrast, Gerridae species such as \u003cem\u003eGerris latiabdominis\u003c/em\u003e do not possess swimming fans but instead have elongated, hairy midlegs that generate thrust through the creation of asymmetric dimples on the water surface. These midlegs, along with highly hydrophobic ventral surfaces, are critical for surface-tension-based propulsion (Steinmann et al. 2021). While some aspects of locomotory performance have been compared (Crumiere et al. 2016), detailed kinematic analyses and high-resolution comparisons of leg micromorphology between Gerridae and Veliidae are lacking. Prior studies have presented images of leg hair arrangements (Andersen 1982; Perez-Goodwyn 2008), but a focused comparative analysis across functional leg regions has not been conducted since Andersen (1976).\u003c/p\u003e\n\u003cp\u003eFollowing the framework set by Andersen (1976, 1982) and Crumiere et al. (2016), we compare \u003cem\u003eRhagovelia distincta\u003c/em\u003e (Veliidae) and \u003cem\u003eGerris latiabdominis\u003c/em\u003e (Gerridae), two small-bodied representatives of their respective clades. We examine their midleg microstructures and stroke kinematics to evaluate how different physical thrust mechanisms are supported by contrasting anatomical features. We further investigate the hypothesized passive versus active mechanisms of fan control in \u003cem\u003eR. distincta\u003c/em\u003e. Finally, we describe and compare fore- and hindleg microstructures used in support and sliding to identify potential convergences across taxa with different thrust mechanisms. Our results provide a foundation for future work on the functional evolution of water-surface locomotion in Gerromorpha.\u003c/p\u003e"},{"header":"Results","content":"\u003cp\u003e\u003cstrong\u003eBehavioral observations: leg use at rest\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eR. distincta\u0026nbsp;\u003c/em\u003ebody\u003cem\u003e\u0026nbsp;\u003c/em\u003eis supported on foreleg and hindleg tarsus with minimal contribution from the midleg tarsal tips (Figure 1A1, B1, C1, D1; Figure S1). Occasionally, a very small portion of the midleg`s fan is protruded from the tarsal tip into the water body through the surface it`s in contact with (Figure 1B1; Video S3 Part 4). \u003cem\u003eG. latiabdominis\u0026nbsp;\u003c/em\u003ebody is supported on foreleg tarsus, hindleg tibia and tarsus, and midleg intermediate-distal tibia and tarsus (Figure 2A1, B1, C1, D1; Video S2), which create dimples without piercing the water surface (dimples cast shadows on the bottom of the container; Figure 2C1, D1; Video S2 Part 4).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eBehavioral observations: leg use at locomotion\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eDuring the initial thrust phase (Figure 1), \u003cem\u003eR. distincta\u003c/em\u003e moves its midlegs forward above the water, placing the tarsi on the surface (Figure 1C2), which initiates fan extension into the water (Figure 1B2). Alternatively, the midlegs may advance while the tarsal tips remain in contact with the surface and a small portion of the fan protrudes underwater (Figure 1B6; Figure S2D; Video S3 Part 4). As the fan rapidly protracts (Figure 1B2; Figure S2C, D), the tarsus is simultaneously pressed downward (Figure 1C3; Figures S2B, S5) and the midlegs are pushed backward (Figure 1C4, C5). This generates growing anteroposterior asymmetrical dimples (Figure 1B4), visible as shadows expand from miniscule circles at the tarsal tips to ovals extending from the tibiotarsal joints (Figure 1D4). Contrast to \u003cem\u003eG. latiabdominis\u003c/em\u003e (Figure 2D), \u003cem\u003eR. distincta\u0026nbsp;\u003c/em\u003eshows an expanded anterior dimple region (Figure 1C3), likely caused by water displaced beneath the surface by the fan (Figure 1B4; Figure S2G, H). Strong strokes can produce surface waves (Figure 1C4, C5; 17 of 99 strokes; Video S1 Part 1), but all strokes transition into a passive sliding phase, during which the midlegs either disengage from the surface or trail behind with minimal fan protrusion (Figure 1B6; Video S1). Observed midleg disengagement suggests adhesive forces and surface tension are overcome in this process (Video S3). Throughout, the fore- and hindleg tarsi remain in contact with the surface, providing support during sliding.\u003c/p\u003e\n\u003cp\u003eDuration of fan protraction ranged between 6\u0026shy;\u0026shy;\u0026ndash;23 ms (12.3 \u0026plusmn; 3.4 ms; n = 69) while the duration of fan retraction ranged between 3\u0026ndash;15 ms (8.4 \u0026plusmn; 2.6 ms; n = 73) (Figures S3 and S4). Shorter durations were often associated with strokes beginning or ending with partially protracted fans, where 2\u0026ndash;4 distal setae tips protruded into the water. Fully protracted fans had an average area of 0.89 \u0026plusmn; 0.04 mm\u003csup\u003e2\u003c/sup\u003e, radius of 0.85 \u0026plusmn; 0.01 mm, and protracted angle of 139.13 \u0026plusmn; 3.64\u0026deg; (n = 6; Figure S3). The longitudinal axis of wetted midleg and protracted fan typically lay in the same plane, slanted at 80.2 \u0026plusmn; 6.0\u0026deg; (n = 26; Figure S3E). Analysis of 110 slow-motion videos (each capturing 1\u0026ndash;4 fan manipulations) showed that \u003cem\u003eR. distincta\u003c/em\u003e actively controls the timing, extent, and duration of fan protraction and retraction, regardless of midleg position on the water surface (Figure S4; Video S3).\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eG. latiabdominis\u003c/em\u003e generated thrust without piercing the surface (Figure 2), using midlegs slightly rotated so that anteroventral gap-row microstructures pressed backward against the water (Figure S2; 27\u0026ndash;36 ms into the stroke). Backward movement of the midlegs produced anteroposterior asymmetrical dimples and backward-moving surface waves (Figure 2D; Figure S2E, F; Video S2). Midleg disengagement occurred when the wetted midleg aligned nearly parallel to the direction of body movement and proceeded gradually from proximal to distal segments. Some instances of disengagement appeared smooth (Video S2), while others suggested adhesive forces and surface tension were overcome during the process (Video S2; Figure S2E, 95 ms). Hindlegs contributed weakly to thrust, as indicated by faintly asymmetrical shadows and small wave bows during the initial stroke phase (Figure 2D4). Hindleg tarsi and tibiae provided support during sliding, shown by shadows aligned with the body axis (Figure 2D5, D6). Forelegs also supported the body, particularly near the end of the stroke, as indicated by prominent shadows (Figure 2D6), except at mid-stroke when support was reduced (Figure 2D4).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eKinematics profiles during a stroke\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eIn both species (Figure 7), the midleg femur angle increased gradually during the stroke, but more steeply in \u003cem\u003eR. distincta\u003c/em\u003e than in \u003cem\u003eG. latiabdominis\u003c/em\u003e. The initial angle was more acute in \u003cem\u003eR. distincta\u003c/em\u003e (~20\u0026deg;) compared to \u003cem\u003eG. latiabdominis\u003c/em\u003e (~60\u0026deg;), with both reaching a final angle of ~120\u0026deg; (Figure 7A, B). Peak femur angular velocity was higher in \u003cem\u003eR. distincta\u003c/em\u003e and occurred mid-stride at a femur angle of 85\u0026deg;, whereas in \u003cem\u003eG. latiabdominis\u003c/em\u003e it peaked at 100\u0026deg; (Figure 7A, B). In G\u003cem\u003e. latiabdominis\u003c/em\u003e, the femur\u0026ndash;tibia and tibia\u0026ndash;tarsus angles remained relatively small throughout the stroke. In contrast, \u003cem\u003eR. distincta\u003c/em\u003e showed more pronounced changes: the femur\u0026ndash;tibia angle increased from 30\u0026deg; to 60\u0026deg; (Figure 7C), and the tibia\u0026ndash;tarsus angle ranged from 8\u0026deg; to 16\u0026deg; (Figure 7D). These patterns suggest that \u003cem\u003eR. distincta\u003c/em\u003e engages in coordinated rotations at the coxa\u0026ndash;femur, femur\u0026ndash;tibia, and\u0026mdash;though to a lesser extent\u0026mdash;tibia\u0026ndash;tarsus joints, while in \u003cem\u003eG. latiabdominis\u003c/em\u003e, motion is largely concentrated at the coxa\u0026ndash;femur joint.\u003c/p\u003e\n\u003cp\u003eIn both species, leg velocity during slower (longer) strokes remained below the theoretical critical velocity (~0.23 m/s; Figure 7E) above which gravity-capillary waves are generated and may contribute to thrust (Reaver et al. 2019). In faster (shorter) strokes, leg velocity exceeded this threshold within ~5 ms and reached higher peak values in \u003cem\u003eG. latiabdominis\u0026nbsp;\u003c/em\u003ethan in \u003cem\u003eR. distincta\u0026nbsp;\u003c/em\u003e(Figure 7E). In \u003cem\u003eR. distincta\u003c/em\u003e, the leg velocity vector was briefly aligned with the body movement axis only during mid-stroke, when the femur was approximately perpendicular to the trajectory (Figures 9G and 7G). For the remaining stroke, the fan\u0026ndash;acting as an oar blade\u0026ndash;moved in a direction misaligned with the body axis and not perpendicular to its movement direction. In contrast, in \u003cem\u003eG. latiabdominis\u003c/em\u003e, the leg velocity vector remained nearly parallel to the body movement axis (Figures 9E and 7F) and nearly perpendicular to the wetted leg section (Figures 9G and 7H) for a much larger portion of the stroke (rectangles on x-axis). In both species, the highest net force per stroke, indicated by peak body acceleration (Figure 7J), coincided with intervals of high \u0026ldquo;effectiveness\u0026rdquo; indices (Figure 7F\u0026ndash;H; SI Part 5).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eKinematics comparisons of a stroke\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eA small subset of \u003cem\u003eG. latiabdominis\u003c/em\u003e strokes showed notably higher average leg linear velocities, body velocities, and peak accelerations (Figure 8). However, when all strokes were analyzed together, there were no significant differences between the two species in average midleg velocity (Figure 8B2; Table S4), final body velocity (Figure 8A1; Table S4), or maximal body acceleration (Figure 8A2; Table S4). Strokes by \u003cem\u003eR. distincta\u003c/em\u003e were shorter in duration (Figure 8C3; Table S4) and involved faster angular femur movements (Figure 8B1; Table S4), and traveled a shorter distance across the water surface during each stroke compared to \u003cem\u003eG. latiabdominis\u003c/em\u003e (Figure 8C1, C2), notably as \u003cem\u003eR. distincta\u003c/em\u003e has shorter legs (Tables S1 and S2) and additional angular movements were observed at the tibia and tarsus (Figure 7C, D). Since the kinematic variables are intercorrelated (Figure S26D), estimates of their individual effects on body speed (Figure S26A\u0026ndash;C) are not independent. To address this, we performed a principal component analysis and extracted two components (Table 1). Strokes by \u003cem\u003eR. distincta\u003c/em\u003e were characterized by shorter duration and higher femur angular velocity (higher RC2 values), along with shorter distance and slower leg speed on the water surface (lower RC1 values) (Figure 8E), distinguishing them from \u003cem\u003eG. latiabdominis\u003c/em\u003e.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eVentral microstructures on legs\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eOverview\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe legs of \u003cem\u003eR. distincta\u003c/em\u003e had a less dense hair layer than those of \u003cem\u003eG. latiabdominis\u003c/em\u003e. We identified 17 setae (hair) types: five shared by both species, five unique to \u003cem\u003eR. distincta\u003c/em\u003e, and seven unique to \u003cem\u003eG. latiabdominis\u003c/em\u003e (Figure 3; Figure S10; Table S3). Our analysis focused on the ventral microstructures of leg segments that interact with the water surface, and respective nanometer sized details specifically in \u003cem\u003eR. distincta\u003c/em\u003e (Figures 4 and 5; Figures S11\u0026ndash;S23).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eVentral microstructures for thrust generation\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe swimming fan of \u003cem\u003eR. distincta\u003c/em\u003e (Figure 4B1, B2; Figures S11\u0026ndash;S14) consists of anterior and posterior claws and a fan made up of 17 (Figures S8, S9) to 21 (Figure 4B2; Figure S11) setae. Each seta bears setulae along its axis at 8\u0026ndash;12 \u0026micro;m intervals, forming a feather-like structure (Figure 4B3, B4; Figures S12, S13). When protracted, the distance between adjacent setae ranges from ~20 to ~100 \u0026micro;m, and between setulae from several to ~20 \u0026micro;m, with typical setula spacing of 4\u0026ndash;10 \u0026micro;m (Figure S13A, B). The fan is anchored at the inner proximal corner of the cleft between the two lobes of tarsomere 3 (Figure 4B2). The surfaces of the setae, setulae, and claws lack nanogrooves. The setae resemble flat beams or boards (Figure 4B5; Figure S12C) with near-elliptical (Figure 4B6; Figure S12D) or slightly triangular (Figure S12E) cross-sections, measuring 2\u0026ndash;4 \u0026micro;m \u0026times; 7\u0026ndash;10 \u0026micro;m. The orientation of the narrow edges suggests that they face the water during thrust generation (Figure 4B5; Figures S12, S13). The setulae are also flat, 1\u0026ndash;2 \u0026micro;m wide and 300\u0026ndash;700 nm thick, with a hollow center (Figure S12H), and their orientation further suggest that they press against the water with the narrow edges. Transverse sections of fan setae show lamellar outer layers and a central hollow (~700 nm in diameter) containing internal rods (~400 nm in diameter) (Figure 4B6; Figure S12). Claw cross-sections are 1.5\u0026ndash;2.5 \u0026micro;m thick and consist of external lamellar layers and multiple internal layers with nanofibers (100 nm) and cluster of nanofibers (200\u0026ndash;400 nm) running in various directions (Figure 4B8; Figure S14). Although initial observations with fully dissected fan indicate passive spreading, further observations with the fan and claws intact inside the cleft indicate that fan movement is not passive (SI Part 3B). In fact, we were able to induce fan protraction by pulling the \u003cem\u003eut\u0026nbsp;\u003c/em\u003etendon attached to the base of the fan-and-claw structure (Figures S8, S9).\u003c/p\u003e\n\u003cp\u003eAlong the ventral edge of the posterior lobe, a structure composed of three rows of \u003cem\u003eH2\u003c/em\u003e setae\u0026mdash;spaced 6\u0026ndash;8 \u0026micro;m apart within each row and separated by two 2.5\u0026ndash;5 \u0026micro;m wide gaps\u0026mdash;forms a band that presses against the water surface without breaking it during a stroke (Figure 4B9; Figure S16). This pattern resembles the \u0026ldquo;gaps and rows\u0026rdquo; arrangement seen on the tarsus of \u003cem\u003eG. latiabdominis\u003c/em\u003e (Figure 5B). Similarly, the ventral edge of the anterior lobe, which also contacts the water surface during a stroke, is lined with a band of 3\u0026ndash;4 rows of \u003cem\u003eH1\u003c/em\u003e setae (Figure 4B9; Figure S16); however, only lateral views were available, limiting precise row counts. On ventral tarsomere 2, which also interacts with the water surface without breaking it (Figure S5), we observed three rows: a posterior row of \u003cem\u003eH1\u003c/em\u003e setae and two anterior rows of \u003cem\u003eSp\u003c/em\u003e setae, separated by gaps (Figure 4B12, B13; Figure S17).\u003c/p\u003e\n\u003cp\u003eIn \u003cem\u003eG. latiabdominis\u003c/em\u003e, a \u0026ldquo;gaps and rows\u0026rdquo; arrangement of setae was observed along ventral midleg sections that interact with the water surface (Figure 5B; Figures S20 and S21). This pattern resembles the structure on the ventral edge of the lobe in \u003cem\u003eR. distincta\u003c/em\u003e (Figure 4B9). It is especially prominent on the tarsus, where a main ventral gap separates a row of cuspidate setae (\u003cem\u003eC\u003c/em\u003e) from a posterior row of thorn-like \u003cem\u003eT2\u003c/em\u003e setae, with a second \u003cem\u003eT2\u003c/em\u003e row positioned further posteriorly to form a \u003cem\u003eT2\u0026ndash;T2\u003c/em\u003e gap (Figure 5B4\u0026ndash;B9). Anterior to the \u003cem\u003eC\u003c/em\u003e row, a row of \u003cem\u003eM1\u003c/em\u003e setae creates a \u003cem\u003eC\u0026ndash;M1\u003c/em\u003e gap (Figure S20E). The tips of these three setae types bend distally along the leg`s longitudinal axis, particularly the cuspidate setae, whose long, flat distal sections may contact each other. In contrast, setae on the lateral and dorsal leg surfaces are relatively straight (Figure S21B). On tarsomere 2, a second, less regularly arranged anterior\u003cem\u003e\u0026nbsp;M1\u0026nbsp;\u003c/em\u003erow creates an \u003cem\u003eM1\u0026ndash;M1\u003c/em\u003e gap (Figure S20C, E). This arrangement is less distinct on the tibia, where \u003cem\u003eT2\u003c/em\u003e setae exhibits intermediate morphology between \u003cem\u003eT2\u003c/em\u003e and \u003cem\u003eM1\u003c/em\u003e, and the \u003cem\u003eC\u003c/em\u003e setae are replaced by \u003cem\u003eM2\u003c/em\u003e (Figure 5B3). At the tip of the tarsus, the pattern disappears, replaced by a ventral concentration of grass-blade-like setae (\u003cem\u003eg\u003c/em\u003e) (Figure 5B12).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eVentral microstructures for support and sliding\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eIn \u003cem\u003eR. distincta\u003c/em\u003e, specialized microstructures involved in support and sliding were found on the ventral tarsi of the forelegs (Figure 4C), hindlegs (Figure 4A), and at the distal tips of the midleg tarsus (Figure 4B10, B11). These structures consist of one or more rows of \u003cem\u003eSp\u003c/em\u003e setae with flattened, bent tips that overlap to form a \u0026ldquo;beam-like\u0026rdquo; surface ~20\u0026ndash;25 \u0026micro;m below the leg cuticle, oriented toward the water. Two \u003cem\u003eSp\u003c/em\u003e rows were observed on the ventral forelegs (Figure 4C3, C4; Figure S18) and 1 on the hindlegs (Figure 4A1; Figure S19), with the flattened tips especially prominent near the leg ends (Figures 4A5, A6; 4C4). \u003cem\u003eSp\u003c/em\u003e rows are flanked\u0026mdash;particularly posteriorly\u0026mdash;by \u003cem\u003eH1\u0026nbsp;\u003c/em\u003esetae on the forelegs (Figure 4C; Figure S18) and \u003cem\u003eM2\u003c/em\u003e setae on the hindlegs (Figure 4A; Figure S19). At the tips of the midlegs, which also provide support, a dense cluster of \u003cem\u003eSp\u003c/em\u003e setae\u0026mdash;likely modified \u003cem\u003eH1\u003c/em\u003e types\u0026mdash;was observed, with long, flat, overlapping tips (Figure 4B11; Figure S15F, G).\u003c/p\u003e\n\u003cp\u003eIn \u003cem\u003eG. latiabdominis\u003c/em\u003e, specialized microstructures involved in support and sliding were found on the ventral forelegs and hindlegs. On the ventral foreleg tarsus, a band of grass-blade setae (\u003cem\u003eg\u003c/em\u003e) with overlapping flat, bent distal tips was observed ~15\u0026ndash;20 \u0026micro;m from the leg cuticle, facing the water surface (Figure 5C4; Figure S22). An entangled cluster of web setae (\u003cem\u003eW\u003c/em\u003e) was also present, especially near the tibiotarsal joint (Figure 5C2; Figure S22). On the ventral side of the hindleg tibia and tarsus, 2\u0026ndash;3 rows of leaf-blade setae (\u003cem\u003eL\u003c/em\u003e) were arranged in an orderly manner (Figure 5A; Figure S23). Their overlapping distal tips formed a \u0026ldquo;beam-like\u0026rdquo; surface with nanogrooves, ~20 \u0026micro;m above the water surface (Figure 5A2; Figure S23), and were accompanied by a posterior row of large thorn setae (\u003cem\u003eT3\u003c/em\u003e) (Figure 5A6, A8; Figure S23A, E). The ventral sides of joints were covered with bundles of \u003cem\u003eL\u003c/em\u003e and leaf-like \u003cem\u003el\u003c/em\u003e setae (Figure 5A1, A6), with \u003cem\u003el\u003c/em\u003e setae also present at the tarsal tips (Figure 5A9).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eContact angle on midleg sections used in thrust generation\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eMidleg tarsal surfaces in \u003cem\u003eG. latiabdominis\u003c/em\u003e were generally more hydrophobic than those in \u003cem\u003eR. distincta\u003c/em\u003e (Figure 6). In \u003cem\u003eR. distincta\u003c/em\u003e, water droplets rapidly lost their spherical shape after contacting the hair layer, spreading 10\u0026ndash;40% and showing increased shape indices. Contact angles on the dorsal and ventral midleg surfaces progressively decreased from approximately 130.3\u0026deg; and 75.5\u0026deg;, respectively, to as low as 21.9\u0026deg; as droplets collapsed, indicating relatively high surface wettability (Figure 6A, B). This effect was especially pronounced on the ventral side, where the swimming fan and associated microstructures interact with water during locomotion (Figure 6B, C); Figure S24). In contrast, droplets on the dorsal and ventral midleg surfaces of \u003cem\u003eG. latiabdominis\u0026nbsp;\u003c/em\u003eretained their spherical shapes with minimal spreading (shape index \u0026asymp; 1.5), and contact angles remained high throughout dissipation, ranging from 132.1\u0026deg; to 109.3\u0026deg;, consistent with strong hydrophobicity (Figure 6D, E; Figure S24; SI Part 4).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTable 1. Principal Component Analysis of behavioral variables.\u0026nbsp;\u003c/strong\u003ePCA was conducted on seven behavioral variables from\u003cem\u003e\u0026nbsp;R. distincta\u0026nbsp;\u003c/em\u003e(n = 21) and \u003cem\u003eG. latiabdominis\u003c/em\u003e (n = 12). The table shows eigenvalues, percentage of variance explained, and loadings for the first two rotated components (RC1 and RC2), based on the \u003cem\u003efa.parallel\u003c/em\u003e and \u003cem\u003eprincipal\u003c/em\u003e functions from the \u003cstrong\u003epsych\u0026nbsp;\u003c/strong\u003eR package. Loadings with absolute values greater than 0.75 are shown in bold. Related results are illustrated in Figure 8D.\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\" class=\"fr-table-selection-hover\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 240px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 186px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eRC1\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003eMidleg`s \u0026amp; Body`s\u0026nbsp;\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003eMovements PC\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 197px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eRC2\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003eMidleg Angular\u0026nbsp;\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003eSpeed \u0026amp; Duration PC\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"3\" style=\"width: 623px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cu\u003eLinear velocities \u0026amp; accelerations:\u003c/u\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 240px;\"\u003e\n \u003cp\u003eAverage midleg velocity (mm/s)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 186px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.80\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 197px;\"\u003e\n \u003cp\u003e0.53\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 240px;\"\u003e\n \u003cp\u003eMaximum acceleration (mm/s\u003csup\u003e2\u003c/sup\u003e)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 186px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.77\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 197px;\"\u003e\n \u003cp\u003e0.48\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 240px;\"\u003e\n \u003cp\u003eFinal body velocity (mm/s)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 186px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.77\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 197px;\"\u003e\n \u003cp\u003e0.60\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 240px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cu\u003eLinear midleg movement distances:\u003c/u\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 186px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 197px;\"\u003e\n \u003cp\u003e\u003cem\u003e\u0026nbsp;\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 240px;\"\u003e\n \u003cp\u003eMidleg`s stroke amplitude (mm)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 186px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.96\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 197px;\"\u003e\n \u003cp\u003e-0.22\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 240px;\"\u003e\n \u003cp\u003eDistance traveled by wetted midleg (mm)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 186px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.96\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 197px;\"\u003e\n \u003cp\u003e-0.20\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"3\" style=\"width: 623px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cu\u003eAngular midleg (femur) speed and duration:\u003c/u\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 240px;\"\u003e\n \u003cp\u003eAverage angular velocity (degrees/s)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 186px;\"\u003e\n \u003cp\u003e0.03\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 197px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.93\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 240px;\"\u003e\n \u003cp\u003eStroke duration (s)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 186px;\"\u003e\n \u003cp\u003e0.04\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 197px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e-0.98\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 240px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eEigenvalue\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 186px;\"\u003e\n \u003cp\u003e3.67\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 197px;\"\u003e\n \u003cp\u003e2.77\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 240px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eVariance %\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 186px;\"\u003e\n \u003cp\u003e52%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 197px;\"\u003e\n \u003cp\u003e40%\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e"},{"header":"Discussion","content":"\u003cp\u003eObservations of live \u003cem\u003eR. distincta\u003c/em\u003e suggest that thrust during a stroke results from a combination of two forces: hydrodynamic forces generated by the oar-like motion of the fan, as proposed in previous studies (Andersen 1976; Ortega-Jimenez et al. 2024; Santos et al. 2017), and additional capillary forces arising from an anteroposteriorly asymmetrical dimple beneath the tarsus, consistent with surface-tension-based mechanisms described in Gerridae (Steinman et al. 2021). In contrast, behavioral evidence from \u003cem\u003eG. latiabdominis\u003c/em\u003e aligns with the surface-tension-based thrust mechanism characterized in detail for \u003cem\u003eAquarius paludum\u003c/em\u003e (Steinman et al. 2021). The observed differences between these species appear closely tied to variations in leg microstructures, material properties, and motion kinematics that reflect their distinct ecological contexts and thrust-generation strategies.\u003c/p\u003e\n\u003cp\u003eIn \u003cem\u003eR. distincta\u003c/em\u003e, several microstructural features suggest specialization for hydrodynamic thrust. The hydrophilic properties of the claw likely facilitate surface penetration during stroke initiation, as the fan and claw extend downward from the cleft\u0026apos;s internal compartment. Similar surface textures between the claw and fan setae indicate that the fan may also be hydrophilic, enhancing its ability to submerge through the water surface. The fan\u0026rsquo;s setae and setulae are oriented to press against the water with their narrow edges, a configuration that minimizes deformation under hydrodynamic forces, as predicted by beam theory (Bhavikatti 2010; SI Part D). The fan setae\u0026rsquo;s internal architecture, comprising a hollow core with columnar nanofibers, and the claw\u0026rsquo;s lamellar structure, resembles engineered designs like sandwich and lamellar composites (Zenkert 1995; Wu and Zhu 2021), which are known to enhance stiffness and fatigue resistance. This likely enables the fan to function effectively as an oar-blade. The nanofiber arrangement within the fan setae provides both flexibility and mechanical resilience\u0026mdash;qualities vital for high-frequency paddling in fast-flowing water. The central hollow core may serve to reduce weight and accommodate controlled deformation, further improving maneuverability.\u003c/p\u003e\n\u003cp\u003eAdditionally, the H1 and H2 setae rows positioned at the cleft entrance likely act as a barrier against water intrusion while assisting fan deployment through elastocapillary interactions. Based on the observed stroke speed, setula spacing, and thickness, we estimate the fan\u0026rsquo;s leakiness to range from ~0.3 to 0.6 (SI Part 7), suggesting it behaves as a \u0026ldquo;leaky paddle\u0026rdquo; rather than a solid blade. This is similar to bristled appendages in copepods and barnacle larvae (Koehl 1993; Lamont \u0026amp; Emlet 2018). Comparable fan-like structures with potential \u0026ldquo;leaky paddle\u0026rdquo; functionality are found in other Veliidae genera such as \u003cem\u003eTetraripis\u003c/em\u003e and \u003cem\u003eTrochopus\u003c/em\u003e. In contrast, \u0026ldquo;Veliidae\u0026rdquo; species that depend primarily on surface-tension-based thrust, like \u003cem\u003eVelia sp.\u0026nbsp;\u003c/em\u003e(Andersen 1976), tend to exhibit more developed ventral \u0026quot;gaps and rows\u0026quot; arrangements and less-developed fan structures, supporting the hypothesis of divergent functional adaptations.\u003c/p\u003e\n\u003cp\u003eIn \u003cem\u003eG. latiabdominis\u003c/em\u003e, a different set of microstructural adaptations supports thrust generation primarily through surface tension. The species exhibits denser and more hydrophobic setae on its midlegs than \u003cem\u003eR. distincta\u003c/em\u003e, likely reflecting its reliance on surface-tension-based propulsion. Deep, asymmetrical dimples beneath the wetted portion of the leg, along with prominent bow waves, contribute to increased thrust. Interestingly, even the lower hydrophobicity observed in \u003cem\u003eR. distincta\u003c/em\u003e can still support thrust via surface tension through dimple formation, though to a lesser degree. This is evidenced by the net thrust outputs, which are significantly lower in \u003cem\u003eR. distincta\u0026nbsp;\u003c/em\u003e(100\u0026ndash;200 \u0026micro;N) than in \u003cem\u003eG. latiabdominis\u003c/em\u003e (400\u0026ndash;500 \u0026micro;N; Figure S25).\u003c/p\u003e\n\u003cp\u003eAcross both species, leg surfaces involved in surface-tension-based thrust display linear arrangements of distally bending setae forming \u0026quot;gaps and rows.\u0026quot; These are absent from other leg surfaces and may serve specialized functions, potentially related to air retention during dimple formation, which prevents surface penetration and facilitates thrust. Theoretical work (Uesugi et al. 2020; Uesugi 2021) suggests such arrangements can trap air and maintain smooth water contact. Our observations support this, as these structures appear to function like pressurized air pockets that resist water surface breakage under thrust loads. Additionally, the smoother longitudinal gaps may reduce adhesion during stroke recovery. This arrangement is more pronounced in \u003cem\u003eG. latiabdominis\u003c/em\u003e, where the setae also exhibit nanogrooves known to enhance hydrophobicity (Feng et al. 2007; Gao \u0026amp; Jiang 2004). The posterior concentration of thicker T2 and T3 setae in mid- and hindlegs may be a specific adaptation to withstand the higher pressure during backward strokes. These structural reinforcements are aligned with previous findings on jumping and propulsion in surface-dwelling insects (Koh et al. 2015; Yang et al. 2016).\u003c/p\u003e\n\u003cp\u003eKinematic data further underscores how thrust generation mechanisms are integrated with species-specific movement strategies. In \u003cem\u003eR. distincta\u003c/em\u003e, hydrodynamic thrust is supported by a midleg stroke that begins from a more acute femur angle and involves backward rotations across multiple leg joints. The resulting movement vector deviates from the body axis, potentially enabling lift-like forces, akin to paddling strategies in human kayaking (Jackson 1995; Michael et al. 2009) and animal locomotion (Johansson and Norberg 2001). This configuration would be inefficient for surface-tension-based propulsion but is well-suited to fast-flowing environments where high-frequency strokes with minimal surface contact time are advantageous. The ability of \u003cem\u003eR. distincta\u003c/em\u003e to actively control fan protraction and retraction (Andersen 1976; SI Parts 2 and 3B) could allow for greater flexibility during maneuvering and stroke timing, in contrast to the passive deployment described in recent studies (Ortega-Jimenez 2021, 2024).\u003c/p\u003e\n\u003cp\u003eConversely, \u003cem\u003eG. latiabdominis\u003c/em\u003e initiates midleg movement at a less acute femur angle, with primary rotation occurring at the coxa\u0026ndash;femur joint, and maintains a nearly straight femur\u0026ndash;tibia segment. This configuration enhances wetted surface area, allowing for more efficient surface-tension-based thrust. The leg stroke is longer in duration and follows a nearly parallel trajectory to the body axis, facilitating the formation of asymmetrical dimples critical to curvature force production. The backward movement of the leg, nearly perpendicular to its axis, promotes directional asymmetry in the dimple, optimizing propulsion in still or slow-moving waters.\u003c/p\u003e\n\u003cp\u003eBoth species share a distinct ventral setal arrangement that likely contributes to standing and sliding on the water surface. Flattened, overlapping setae form a beam-like structure along the underside of wetted legs, similar to those described in \u003cem\u003eGerris\u003c/em\u003e and \u003cem\u003eAquarius\u003c/em\u003e (Andersen 1976; Perez-Goodwyn 2008). This configuration minimizes surface penetration and drag, offering support without water breakage. The beam\u0026rsquo;s alignment with body motion also likely facilitates sliding and steering. In the heavier \u003cem\u003eG. latiabdominis\u003c/em\u003e, these beams feature hydrophobic nanogrooves that enhance their supporting function. In contrast, the lighter \u003cem\u003eR. distincta\u003c/em\u003e lacks such grooves, suggesting lower support demands. Moreover, the spoon-like setae at the tips of \u003cem\u003eR. distincta\u003c/em\u003e\u0026apos;s midlegs may help resist displacement by currents, offering a stabilizing advantage in fast-flowing habitats.\u003c/p\u003e"},{"header":"Conclusion","content":"\u003cp\u003eOur results show that two representative species from distinct clades of semiaquatic bugs\u0026mdash;each having independently evolved a symmetrical rowing stroke (Andersen 1982; Khila et al. 2014)\u0026mdash;rely on different thrust-generation mechanisms, each requiring distinct leg movement patterns, leg lengths, and ventral microstructures. These adaptations enable \u003cem\u003eR. distincta\u003c/em\u003e to generate thrust through a combination of hydrodynamic drag and surface tension, while \u003cem\u003eG. latiabdominis\u003c/em\u003e relies exclusively on surface-tension-based thrust.\u003c/p\u003e\n\u003cp\u003eThe results suggest that the fan in \u003cem\u003eRhagovelia\u003c/em\u003e functions as a \u0026ldquo;leaky paddle,\u0026rdquo; producing thrust via hydrodynamic drag. Our study provides new insights into the morphology and anatomy of \u003cem\u003eRhagovelia\u003c/em\u003e`s fan and claw, structures essential to its hydrodynamic thrust mechanism (Andersen 1976; Ortega-Jimenez \u0026amp; Bhamla 2021; Ortega-Jimenez et al. 2024; Santos et al. 2017). In addition, we document ventral setal arrangements involved in surface-tension-based thrust\u0026mdash;features that have received little attention in this genus. We also identify shared microstructures in both species used for surface-tension-based thrust, support, and sliding, and we propose hypotheses for the functional roles of these features based on our observations.\u003c/p\u003e\n\u003cp\u003eMany precedent studies on water striders` setae (Gerridae) treat leg micromorphology superficially, and theoretical models often assume a uniform distribution of simple conical hairs on the ventral leg surface. Our findings challenge this simplification by providing detailed morphological data and offering testable hypotheses that can be translated into formal fluid dynamics models. These will enable a quantitative assessment of how specific microstructures contribute to thrust, support, and sliding.\u003c/p\u003e\n\u003cp\u003eFinally, as the observed leg movement patterns and microstructures appear to co-evolve with thrust mechanisms, leg morphology, and habitat, they offer a foundation for phylogenetic studies across taxa. Such studies, especially if they incorporate allometric scaling (Kim et al. 2022, 2024), can shed light on the evolutionary processes behind micromorphological adaptations in semiaquatic bugs. To fully understand these traits, future research should also explore the developmental and genetic basis of setal diversity and arrangement, building on the approaches of Santos et al. (2017) and Finet et al. (2018, 2022).\u003cstrong\u003e\u003cbr\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e"},{"header":"Methods","content":"\u003cp\u003e\u003cstrong\u003eField sites and study species\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eIn January and February of 2020 and 2023, specimen collections and detailed observations of \u003cem\u003eRhagovelia distincta\u0026nbsp;\u003c/em\u003e(body weight: 4\u0026ndash;14 mg; Table S1) were made at the Southwestern Research Station, Arizona, USA (SWRS; 31\u0026deg;53\u0026prime;3\u0026prime;\u0026prime;N, 109\u0026deg;12\u0026prime;21\u0026prime;\u0026prime;W). In August and September of 2020, specimen collections and detailed observations of \u003cem\u003eGerris latiabdominis\u0026nbsp;\u003c/em\u003e(16\u0026ndash;19 mg; Table S2) were made at Gwanak Mountain, Korea (37\u0026deg;26\u0026prime;42\u0026prime;\u0026prime;N, 126\u0026deg;57\u0026prime;51\u0026prime;\u0026prime;E) and Seoul National University, Korea (37\u0026deg;28\u0026prime;57\u0026prime;\u0026prime;N, 126\u0026deg;96\u0026prime;04\u0026prime;\u0026prime;E), respectively. Each individual was weighed (GEM20 High Precision Digital Milligram Jewelry Scale, Smart Weigh, 0.001 g).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eVideographic and photographic observations\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe filmed four types of high speed and standard videos with Sony RX10-III at 959.04 frames per second (fps) and with Chronos 2.1-HD at 1000\u0026ndash;4000 fps of individuals in acrylic containers (18 x 18 cm, filled with water):\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cu\u003e- Type 1:\u003c/u\u003e\u003c/strong\u003e directly from above (85 and 62 movies collected from six and seven individuals of \u003cem\u003eR. distincta\u0026nbsp;\u003c/em\u003eand \u003cem\u003eG. latiabdominis\u003c/em\u003e,\u003cem\u003e\u0026nbsp;\u003c/em\u003erespectively).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cu\u003e- Type 2:\u003c/u\u003e\u003c/strong\u003e from the side of various angles (below surface, surface level, and above surface) (249 and 87 movies of \u003cem\u003eR. distincta\u003c/em\u003e and \u003cem\u003eG. latiabdominis\u003c/em\u003e, respectively).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cu\u003e- Type 3:\u003c/u\u003e\u003c/strong\u003e\u0026nbsp; \u0026nbsp;directly from below with light source positioned directly above the container (20 and 60 movies from two and five individuals of \u003cem\u003eR. distincta\u003c/em\u003e and \u003cem\u003eG. latiabdominis\u003c/em\u003e, respectively) to visualize the shadows on the bottom of the container; shadows correspond to dimples under legs on the water surface.\u003c/p\u003e\n\u003cp\u003eTwo variables were extracted from the video types 2 and 3:\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003e- Wetted midleg length\u003c/em\u003e\u003c/strong\u003e\u003cstrong\u003e\u0026nbsp;(mm):\u0026nbsp;\u003c/strong\u003emaximal leg section in contact with the water surface in the middle of fast thrust strokes when midleg angle to body axis approximates 90\u0026deg;. The wetted midleg consisted of tarsus in \u003cem\u003eR. distincta\u0026nbsp;\u003c/em\u003e(6 individuals)\u003cem\u003e\u0026nbsp;\u003c/em\u003eand proximal tibia to tarsal tip in \u003cem\u003eG. latiabdominis\u0026nbsp;\u003c/em\u003e(6 individuals).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003e- Swimming fan area\u0026nbsp;\u003c/em\u003e\u003c/strong\u003e\u003cstrong\u003e(mm\u003csup\u003e2\u003c/sup\u003e):\u0026nbsp;\u003c/strong\u003esix still frames from six different type 2 clips of \u003cem\u003eR. distincta\u003c/em\u003e were used to evaluate the fan surface area (Figure S3).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMicroscopic observations of leg microstructures\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eUsing optical microscopy, we observed the morphology and behavior of the fan in specimens of \u003cem\u003eR. distincta\u003c/em\u003e. Using Scanning \u0026nbsp;Electron Microscopy (SEM), we visualized the hair structures on leg sections that interact with water (SI Part 3).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eContact angle measurements\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eContact angle (in degrees; \u0026deg;), height and width (mm) of small droplets on the surface of ventral and dorsal microstructures of tarsomere 3 and tarsomere 2 of \u003cem\u003eR. distincta\u003c/em\u003e and \u003cem\u003eG. latiabdominis\u003c/em\u003e, respectively, and on the tarsal claw of \u003cem\u003eR. distincta\u003c/em\u003e, were measured (with ImageJ 1.53t) in frames of high-speed video (2000 fps; Chronos 2.1-HD Camera, Kron Technologies). The specimens, sprayed with water, were mounted on a micromanipulator (MM-3, Narishige, Japan) parallel to the camera (Video S4).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eKinematic profiles of a stroke\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eDetailed kinematic analyses of symmetric strokes by \u003cem\u003eR. distincta\u0026nbsp;\u003c/em\u003eand \u003cem\u003eG. latiabdominis\u0026nbsp;\u003c/em\u003ewere restricted to data extracted with \u003cem\u003eTracker\u003c/em\u003e (https://physlets.org/tracker/) from selected videos: 21 and 12 strokes from six and six individuals for \u003cem\u003eR. distincta\u003c/em\u003e and \u003cem\u003eG. latiabdominis\u003c/em\u003e, respectively. Cartesian (\u003cem\u003ex\u003c/em\u003e, \u003cem\u003ey\u003c/em\u003e) coordinates of 10 and 9 points on the insect body for \u003cem\u003eR. distincta\u003c/em\u003e and \u003cem\u003eG. latiabdominis\u003c/em\u003e, respectively (midleg tips were not digitized in \u003cem\u003eG. latiabdominis\u003c/em\u003e due to resolution issues) were digitized and subsequently smooth-splined using the \u0026ldquo;\u003cem\u003estats\u003c/em\u003e\u0026rdquo; package (Everitt \u0026amp; Hothorn 2010; R Core Team 2023)\u003cem\u003e\u0026nbsp;\u003c/em\u003e(df = 5 and smoothing parameter = 0.5).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eTo compare the two species, we focused on five aspects (Figure 9) of leg movements during a stroke and extracted the following kinematic variables for each frame, or two consecutive frames, through the thrust phase of a stroke:\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003e- Midleg femur angl\u003c/em\u003e\u003c/strong\u003e\u003cstrong\u003ee (degrees):\u0026nbsp;\u003c/strong\u003ethe angle between the body axis and the femur at the coxae (Figure 9B) was calculated at each frame. The coxal joint is where the major leg angular movement is performed in both species.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003e- Femur-tibia angle\u003c/em\u003e\u003c/strong\u003e\u003cstrong\u003e\u0026nbsp;(degrees):\u003c/strong\u003e the angle between the femur and the tibia at the femorotibial joint was calculated at each frame (Figure 9C).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003e- Tibia-tarsus angle\u003c/em\u003e\u003c/strong\u003e\u003cstrong\u003e\u0026nbsp;(degrees):\u003c/strong\u003e the angle between the tibia and the tarsus at the tibiotarsal joint was calculated at each frame (Figure 9D).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003e- Midleg angular velocity\u003c/em\u003e\u003c/strong\u003e\u003cstrong\u003e\u0026nbsp;(degrees/s):\u003c/strong\u003e calculated by dividing the between-frame difference in midleg femur angles by the latency between the two consecutive frames (i.e., 1/fps).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003e- Leg velocity\u0026nbsp;\u003c/em\u003e\u003c/strong\u003e\u003cstrong\u003e(mm/s) (\u003cem\u003eU\u003c/em\u003e):\u0026nbsp;\u003c/strong\u003eThe linear velocity in horizontal plane of the midpoint of the wetted midleg length. It is calculated for each pair of two consecutive frames via dividing the distance traveled between by the latency by the two consecutive frames.\u003c/p\u003e\n\u003cp\u003eThree proxies of \u0026ldquo;effectiveness\u0026rdquo; of midleg application for thrust generation during a stroke were calculated (Figure 9) using basic trigonometry and vector algebra:\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e- \u003cem\u003e\u0026ldquo;Effectiveness\u0026rdquo; of leg velocity vector`s direction\u0026nbsp;\u003c/em\u003e\u003c/strong\u003e(Figure 9E) (proportion; range 0\u0026ndash;1): we determined the proportion of the leg velocity vector (and of \u003cem\u003eR. distincta\u003c/em\u003e`s fan protracted under the leg; green vector) employed along the direction parallel to the body movement (blue or violet vectors). Values closer to \u0026lsquo;1\u0026rsquo; indicate \u0026ldquo;more effective\u0026rdquo; employment of legs on water surface because the backward leg velocity vector is near-parallel to the body axis line (\u0026nbsp;) resulting in anteroposterior asymmetry of the dimple crucial for curvature force (i.e., surface tension) contribution to thrust (Steinman at al 2021). Positive values indicate backward velocity vector (blue) that contributes to forward thrust,\u0026nbsp;while negative values indicate forward vector (violet; when legs are dragged along body movement).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003e- \u0026ldquo;Effectiveness\u0026rdquo; of leg length`s use\u003c/em\u003e\u003c/strong\u003e\u003cem\u003e\u0026nbsp;\u003c/em\u003e(Figure 9F) (proportion; range 0\u0026ndash;1): evaluation of the relative length of wetted leg projection (blue) on the line perpendicular to the body movement axis (relative to the actual wetted leg length marked green). In \u003cem\u003eG. latiabdominis\u003c/em\u003e, it may be approximately viewed as the effective proportion of the total wetted midleg length (blue) that pushes the surface dimple directly backwards along the leg velocity vector parallel to the body movement direction (blue vector in Figure 9E). Values closer to \u0026lsquo;1\u0026rsquo; indicate \u0026ldquo;more effective\u0026rdquo; employment of the midleg length pushing the dimple backwards; they also indicate that the fan surface in \u003cem\u003eR. distincta\u003c/em\u003e is approximately perpendicular to the body axis ( ), under the assumption that tarsus on the water surface lies approximately within the near-vertical plane with the surface of the \u003cem\u003eR. distincta`s\u003c/em\u003e fan under water.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003e- \u0026ldquo;Effectiveness\u0026rdquo; of wetted leg orientation\u003c/em\u003e\u003c/strong\u003e (Figure 9G) (degrees): angle \u0026theta; indicates the orientation of wetted midleg`s main axis (as well as the plane of the fan protracted under the leg, assuming fan surface`s plane includes the longitudinal axis of wetted leg) relative to the leg velocity vector. Angles closer to \u0026lsquo;90\u0026deg;\u0026rsquo; indicate \u0026ldquo;more effective\u0026rdquo; employment of the full wetted midleg length in pushing the water surface dimple along the leg velocity vector and creating dimple asymmetry along the velocity vector. Under the assumption that longitudinal axis of wetted leg on the water surface approximately lies within the plane of the \u003cem\u003eR. distincta\u003c/em\u003e fan`s surface protracted under water, \u0026theta; values closer to 90\u0026deg; indicate that the angle between the fan surface and the fan movement direction is near perpendicular and hydrodynamic drag from the fan pushing backwards contributes to thrust.\u003c/p\u003e\n\u003cp\u003eWe additionally extracted three variables from the body movements :\u003c/p\u003e\n\u003cp\u003e- \u003cstrong\u003e\u003cem\u003eBody velocity\u003c/em\u003e\u003c/strong\u003e\u003cstrong\u003e\u0026nbsp;(mm/s):\u0026nbsp;\u003c/strong\u003edistance (mm) traveled by the body center (position derived from the average of head and abdomen tip positions) between consecutive frames divided by the latency between the two consecutive frames (1/fps).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003e- Body acceleration\u003c/em\u003e\u003c/strong\u003e\u003cstrong\u003e\u0026nbsp;(mm/s\u003csup\u003e2\u003c/sup\u003e):\u0026nbsp;\u003c/strong\u003erate of change of body velocity derived from each pair of consecutive body velocity values divided by the latency between the two frames.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003e- Net force\u0026nbsp;\u003c/em\u003e\u003c/strong\u003e\u003cstrong\u003e(\u0026micro;N):\u003c/strong\u003e body acceleration (mm/s\u003csup\u003e2\u003c/sup\u003e) multiplied by insect body mass (mg) and by 0.001 for unit conversion. It represents a horizontal vector of net thrust force during a stroke.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eKinematic characterization of a stroke\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe following kinematic variables were extracted from 21 strokes of six individuals of \u003cem\u003eR. distincta\u0026nbsp;\u003c/em\u003eand 12 strokes of six individuals of \u003cem\u003eG. latiabdominis\u0026nbsp;\u003c/em\u003e(1 value per stroke):\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003e- Distance traveled by wetted midleg (mm):\u003c/em\u003e\u003c/strong\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003esum of frame-by-frame distances of wetted midleg midpoint was measured along the actual trajectory of the midpoint during the thrust phase of a stroke.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003e- Midleg`s stroke amplitude (mm):\u003c/em\u003e\u003c/strong\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003edirect straight-line distance from the initial (beginning of thrust stroke) to the final (end of thrust stroke; when midleg`s velocity vector is no longer opposite to the body velocity vector) positions of wetted midleg midpoint was measured.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003e- Average angular velocity (degrees/s):\u003c/em\u003e\u003c/strong\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003emean of all frame-by-frame midleg angular velocities calculated within the thrust phase of a stroke.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003e- Average midleg velocity (mm/s):\u003c/em\u003e\u003c/strong\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003emean of all frame-by-frame leg velocities calculated within the thrust phase of a stroke.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003e- Leg velocity at maximum body acceleration (mm/s):\u003c/em\u003e\u003c/strong\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003eleg velocity that corresponds to the maximum body acceleration (i.e., maximum horizontal net force) within the thrust phase of a stroke.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003e- Maximal body acceleration (mm/s\u003csup\u003e2\u003c/sup\u003e):\u0026nbsp;\u003c/em\u003e\u003c/strong\u003emaximum value of body acceleration within the thrust phase of a stroke.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003e- Maximal net thrust force (\u0026micro;N)\u003c/em\u003e\u003c/strong\u003e\u003cem\u003e:\u0026nbsp;\u003c/em\u003emaximum value of body force within the thrust phase of a stroke.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003e- Final body velocity (mm/s):\u0026nbsp;\u003c/em\u003e\u003c/strong\u003efinal value of body velocity (between the last two consecutive frames) in the thrust phase of a stroke.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e- \u003cem\u003eMidleg thrust duration (ms):\u003c/em\u003e\u0026nbsp;\u003c/strong\u003elatency from the initiation of midleg thrust movements to the moment of their disengagement\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eStatistical analyses\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAll analyses were performed in R version 4.3.1 (R Core Team, 2023). We used linear mixed-effects models (with \u003cstrong\u003e\u003cem\u003e\u0026ldquo;individual\u0026rdquo;\u003c/em\u003e\u0026nbsp;\u003c/strong\u003eas the random factor; \u0026ldquo;\u003cem\u003elme4\u003c/em\u003e\u0026rdquo; package by Bates\u003cem\u003e\u0026nbsp;et al\u003c/em\u003e. 2015; \u0026nbsp;\u0026ldquo;\u003cem\u003elmerTest\u003c/em\u003e\u0026rdquo; package by Kuznetsova\u003cem\u003e\u0026nbsp;et al.\u0026nbsp;\u003c/em\u003e2017) to compare the effects of \u003cstrong\u003e\u003cem\u003eaverage leg velocity\u003c/em\u003e\u003c/strong\u003e,\u003cstrong\u003e\u003cem\u003e\u0026nbsp;stroke duration\u003c/em\u003e\u003c/strong\u003e, and \u003cstrong\u003e\u003cem\u003edistance traveled by wetted midleg\u003c/em\u003e\u0026nbsp;\u003c/strong\u003eon the \u003cstrong\u003e\u003cem\u003efinal body velocity\u003c/em\u003e\u003c/strong\u003e between the two species (interaction with categorical variable \u0026ldquo;\u003cstrong\u003e\u003cem\u003especies\u003c/em\u003e\u003c/strong\u003e\u0026rdquo;). \u0026nbsp;However, these independent variables were correlated, and the statistical models would not allow proper evaluations of these effects. Therefore, we extracted principal components using functions \u003cem\u003efa.parallel\u003c/em\u003e and \u003cem\u003eprincipal\u003c/em\u003e from the \u0026ldquo;\u003cem\u003epsych\u003c/em\u003e\u0026rdquo; R package (Revelle 2023) from the pooled data for both species considering all seven kinematic variables: \u003cstrong\u003e\u003cem\u003ebody velocity\u003c/em\u003e\u003c/strong\u003e, \u003cstrong\u003e\u003cem\u003emaximum acceleration\u003c/em\u003e\u003c/strong\u003e, \u003cstrong\u003e\u003cem\u003eaverage angular leg velocity\u003c/em\u003e\u003c/strong\u003e, \u003cstrong\u003e\u003cem\u003eaverage leg velocity\u003c/em\u003e\u003c/strong\u003e, \u003cstrong\u003e\u003cem\u003emidleg`s stroke amplitude\u003c/em\u003e\u003c/strong\u003e, \u003cstrong\u003e\u003cem\u003edistance traveled by wetted midleg\u003c/em\u003e\u003c/strong\u003e, and \u003cstrong\u003e\u003cem\u003estroke duration\u003c/em\u003e\u003c/strong\u003e. We focused on the variables with loading values \u0026gt; 0.75 (Rutherford \u003cem\u003eet al.\u003c/em\u003e 1988).\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eFunding statement\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis work was supported by the National Research Foundation of Korea(NRF) grant funded by the Korea government(MSIT)(IRIS RS-2024-00343461; RS-2025-00514508); DGIST Start-up Fund Program nr 20200810 and individual mid-career grant 2022R1A2C1006090 of the Ministry of Science, ICT and Future Planning of Korea.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData availability\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAll data generated or analyzed during this study are available in supplementary files.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n \u003cli\u003eAndersen NM, A comparative study of locomotion on the water surface in semiaquatic bugs (Insecta, Hemiptera, Gerromorpha), Videnskabelige Meddelelser fra Dansk Naturhistorisk Forening 1976; 139: 337\u0026ndash;396.\u003c/li\u003e\n \u003cli\u003eAndersen NM, Phylogenetic inference as applied to the study of evolutionary diversification of semiaquatic bugs (Hemiptera: Gerromorpha), Systematic Zoology 1979; 28: 554\u0026ndash;578.\u003c/li\u003e\n \u003cli\u003eAndersen NM, The semiaquatic bugs: Phylogeny, adaptations, biogeography, and classification, Scandinavian Science Press 1982.\u003c/li\u003e\n \u003cli\u003eAndersen NM, Cladistic inference and evolutionary scenarios: locomotory structure, function, and performance in water striders, Cladistics 1995; 11: 279\u0026ndash;295.\u003c/li\u003e\n \u003cli\u003eArmis\u0026eacute;n D, Refki PN, Crumi\u0026egrave;re AJJ, et al., Predator strike shapes antipredator phenotype through new genetic interactions in water striders, Nature Communications 2015; 6: 8153.\u003c/li\u003e\n \u003cli\u003eArmisen D, Viala S, da Rocha Silva Cordeiro I, et al., Transcriptome-based phylogeny of the semi-aquatic bugs (Hemiptera: Heteroptera: Gerromorpha) reveals patterns of lineage expansion in a series of new adaptive zones, Molecular Biology and Evolution 2022; 39(11): msac229.\u003c/li\u003e\n \u003cli\u003eBates D, Kliegl R, Vasishth S, et al., Parsimonious mixed models, arXiv preprint 2015; arXiv:1506.04967.\u003c/li\u003e\n \u003cli\u003eBhavikatti SS, Mechanics of solids, New Age International (P) Ltd., Publishers 2010.\u003c/li\u003e\n \u003cli\u003eBush JWM, Hu DL, Prakash M, The integument of water-walking arthropods: form and function, Advances in Insect Physiology 2008; 34: 117\u0026ndash;192.\u003c/li\u003e\n \u003cli\u003eCheer AY, Koehl MAR, Paddles and rakes: fluid flow through bristled appendages of small organisms, Journal of Theoretical Biology 1987; 129(1): 17\u0026ndash;39.\u003c/li\u003e\n \u003cli\u003eCianferroni F, Santini G, Comparative description of the hairy structures in two endemic Velia species (Insecta: Hemiptera: Heteroptera: Veliidae): V. currens (Fabricius, 1794) and V. gridellii Tamanini, 1947, Acta Zoologica 2014; 95: 347\u0026ndash;357.\u003c/li\u003e\n \u003cli\u003eCrisp DJ, The stability of structures at a fluid interface, Transactions of the Faraday Society 1950; 46: 228\u0026ndash;235.\u003c/li\u003e\n \u003cli\u003eCrisp DJ, Thorpe WH, The water-protecting properties of insect hairs, Discussions of the Faraday Society 1948; 3: 210\u0026ndash;220.\u003c/li\u003e\n \u003cli\u003eCrumi\u0026egrave;re AJJ, Santos ME, S\u0026eacute;mon M, et al., Diversity in morphology and locomotory behavior is associated with niche expansion in the semi-aquatic bugs, Current Biology 2016; 26(24): 3336\u0026ndash;3342.\u003c/li\u003e\n \u003cli\u003eDamgaard J, Phylogeny of the semiaquatic bugs (Hemiptera-Heteroptera, Gerromorpha), Insect Systematics \u0026amp; Evolution 2008; 39(4): 431\u0026ndash;460.\u003c/li\u003e\n \u003cli\u003eDamgaard J, Evolution of the semi-aquatic bugs (Hemiptera: Heteroptera: Gerromorpha) with re-interpretation of the fossil record, Acta Entomologica Musei Nationalis Pragae 2008; 48: 251\u0026ndash;268.\u003c/li\u003e\n \u003cli\u003eEveritt BS, Hothorn T, A handbook of statistical analyses using R, Chapman \u0026amp; Hall/CRC 2010.\u003c/li\u003e\n \u003cli\u003eFeng XQ, Gao X, Wu Z, et al., Superior water repellency of water strider legs with hierarchical structures: experiments and analysis, Langmuir 2007; 23: 4892\u0026ndash;4896.\u003c/li\u003e\n \u003cli\u003eFinet C, Decaras A, Armis\u0026eacute;n D, et al., The achaete\u0026ndash;scute complex contains a single gene that controls bristle development in the semi-aquatic bugs, Proceedings of the Royal Society B 2018; 285(1892): 20182387.\u003c/li\u003e\n \u003cli\u003eFinet C, Decaras A, Rutkowska M, et al., Leg length and bristle density, both necessary for water surface locomotion, are genetically correlated in water striders, Proceedings of the National Academy of Sciences 2022; 119(9): e2119210119.\u003c/li\u003e\n \u003cli\u003eGao X, Jiang L, Water-repellent legs of water striders, Nature 2004; 432: 36.\u003c/li\u003e\n \u003cli\u003eHu DL, Bush JWM, The hydrodynamics of water-walking arthropods, Journal of Fluid Mechanics 2010; 644: 5\u0026ndash;33.\u003c/li\u003e\n \u003cli\u003eHu DL, Chan B, Bush JWM, The hydrodynamics of water strider locomotion, Nature 2003; 424(6949): 663\u0026ndash;666.\u003c/li\u003e\n \u003cli\u003eJackson PS, Performance prediction for Olympic kayaks, Journal of Sports Sciences 1995; 13: 239\u0026ndash;245.\u003c/li\u003e\n \u003cli\u003eJi XY, Wang WJ, Feng XQ, Role of flexibility in the water repellency of water strider legs, Physical Review E 2012; 85: 021607.\u003c/li\u003e\n \u003cli\u003eJohansson C, Norberg UML, Lift-based paddling in diving grebe, Journal of Experimental Biology 2001; 204: 1687\u0026ndash;1696.\u003c/li\u003e\n \u003cli\u003eKhila A, Abouheif E, Rowe L, Comparative functional analyses of Ultrabithorax reveal multiple steps and paths to diversification of legs in the adaptive radiation of semi-aquatic insects, Evolution 2014; 68: 2159\u0026ndash;2170.\u003c/li\u003e\n \u003cli\u003eKim W, Amauger J, Ha J, et al., Two different jumping mechanisms of water striders are determined by body size, Proceedings of the National Academy of Sciences 2023; 120: e2219972120.\u003c/li\u003e\n \u003cli\u003eKim W, Lee JH, Pham TH, et al., Physics of sliding on water explains morphological and behavioral allometry across a wide range of body sizes in water striders (Gerridae), Proceedings of the Royal Society B 2024; 291: 20241357.\u003c/li\u003e\n \u003cli\u003eKoehl MAR, Hairy little legs: feeding, smelling and swimming at low Reynolds numbers, Contemporary Mathematics 1993; 141: 33\u0026ndash;64.\u003c/li\u003e\n \u003cli\u003eKoh JS, Yang E, Jung GP, et al., Jumping on water: surface tension\u0026ndash;dominated jumping of water striders and robotic insects, Science 2015; 349: 517\u0026ndash;521.\u003c/li\u003e\n \u003cli\u003eKuznetsova A, Brockhoff PB, Christensen RHB, lmerTest package: tests in linear mixed effects models, Journal of Statistical Software 2017; 82: 1\u0026ndash;26.\u003c/li\u003e\n \u003cli\u003eLamont EI, Emlet RB, Permanently fused setules create unusual folding fans used for swimming in cyprid larvae of barnacles, Biological Bulletin 2018; 235: 185\u0026ndash;194.\u003c/li\u003e\n \u003cli\u003eMichael JS, Smith R, Rooney KB, Determinants of kayak paddling performance, Sports Biomechanics 2009; 8: 167\u0026ndash;179.\u003c/li\u003e\n \u003cli\u003eMolnar K, Vas LM, Czigany T, Strength of electrospun single nanofibers through modeling tensile behavior of the nanofibrous mat, Composites Part B: Engineering 2012; 43: 15\u0026ndash;21.\u003c/li\u003e\n \u003cli\u003eOrtega-Jimenez V, Bhamla S, Extreme maneuvering and hydrodynamics of Rhagovelia water striders, APS Division of Fluid Dynamics Meeting Abstracts 2021; E13-007.\u003c/li\u003e\n \u003cli\u003eOrtega-Jimenez V, Bhamla S, Ultrafast elastocapillary fans control agile maneuvering in ripple bugs and robots, bioRxiv preprint 2024.\u003c/li\u003e\n \u003cli\u003ePerez Goodwyn PJ, Voigt D, Fujisaki K, Skating and diving: changes in functional morphology of the setal and microtrichial cover during ontogenesis in Aquarius paludum Fabricius (Heteroptera, Gerridae), Journal of Morphology 2008; 269: 734\u0026ndash;744.\u003c/li\u003e\n \u003cli\u003eR Core Team, R: A language and environment for statistical computing, R Foundation for Statistical Computing 2023.\u003c/li\u003e\n \u003cli\u003eReaver NGF, Kaplan DA, Mattson RA, et al., Hydrodynamic controls on primary producer communities in spring-fed rivers, Geophysical Research Letters 2019; 46: 4715\u0026ndash;4725.\u003c/li\u003e\n \u003cli\u003eRefki PN, Khila A, Key patterning genes contribute to leg elongation in water striders, EvoDevo 2015; 6: 14.\u003c/li\u003e\n \u003cli\u003eRevelle W, How to use the psych package for regression and mediation analysis, Northwestern University 2023.\u003c/li\u003e\n \u003cli\u003eRutherford GSW, Hair JF, Anderson RE, et al., Multivariate Data Analysis with Readings, The Statistician 1988; 37: 484.\u003c/li\u003e\n \u003cli\u003eSantos ME, Le Bouquin A, Crumi\u0026egrave;re AJJ, et al., Taxon-restricted genes at the origin of a novel trait allowing access to a new environment, Science 2017; 358: 386\u0026ndash;390.\u003c/li\u003e\n \u003cli\u003eSteinman T, Cribellier A, Casas J, Singularity of the water strider propulsion mechanisms, Journal of Fluid Mechanics 2021; 915: A118.\u003c/li\u003e\n \u003cli\u003eSun P, Zhao M, Jiang J, et al., The study of dynamic force acted on water strider leg departing from water surface, AIP Advances 2018; 8: 015228.\u003c/li\u003e\n \u003cli\u003eUesugi K, Mayama H, Morshima K, Proposal of a water-repellency model of water strider and its verification by considering directly measured strider leg-rowing force, Journal of Photopolymer Science and Technology 2020; 33: 185\u0026ndash;192.\u003c/li\u003e\n \u003cli\u003eUesugi K, Water-repellency model of the water strider Aquarius paludum paludum by the curved structure of leg micro-hairs, Journal of Photopolymer Science and Technology 2021; 34: 393\u0026ndash;399.\u003c/li\u003e\n \u003cli\u003eWang Q, Yao X, Liu H, et al., Self-removal of condensed water on the legs of water striders, Proceedings of the National Academy of Sciences 2015; 112: 9247\u0026ndash;9252.\u003c/li\u003e\n \u003cli\u003eWei PJ, Chen SC, Lin JF, Adhesion forces and contact angles of water strider legs, Langmuir 2009; 25: 1526\u0026ndash;1528.\u003c/li\u003e\n \u003cli\u003eWingert MC, Jiang Z, Chen R, et al., Strong size-dependent stress relaxation in electrospun polymer nanofibers, Journal of Applied Physics 2017; 121: 0151031-6.\u003c/li\u003e\n \u003cli\u003eWu X, Zhu Y, Gradient and lamellar heterostructures for superior mechanical properties, MRS Bulletin 2021; 46: 244\u0026ndash;249.\u003c/li\u003e\n \u003cli\u003eXue Y, Yuan H, Su W, et al., Enhanced load-carrying capacity of hairy surface floating on water, Proceedings of the Royal Society A 2014; 470: 20130832.\u003c/li\u003e\n \u003cli\u003eYang E, Son JH, Lee SI, et al., Water striders adjust leg movement speed to optimize takeoff velocity for their morphology, Nature Communications 2016; 7: 13698.\u003c/li\u003e\n \u003cli\u003eYe Z, Zhu G, Damgaard J, et al., Phylogeography of a semiaquatic bug, Microvelia horvathi (Hemiptera: Veliidae): An evaluation of historical, geographical and ecological factors, Scientific Reports 2016; 6: 21930.\u003c/li\u003e\n \u003cli\u003eZenkert D, An Introduction to Sandwich Construction, Chameleon Press 1995.\u003c/li\u003e\n \u003cli\u003eZhang X, Zhang C, Ma P, Mechanical properties of hollow polyester monofilament: Compression and tension behaviors, Journal of Engineered Fibers and Fabrics 2019; 14: 1\u0026ndash;6.\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"scientific-reports","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"scirep","sideBox":"Learn more about [Scientific Reports](http://www.nature.com/srep/)","snPcode":"","submissionUrl":"","title":"Scientific Reports","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Scientific Reports","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"","lastPublishedDoi":"10.21203/rs.3.rs-6903117/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-6903117/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"The evolution of semiaquatic bugs led to at least two independent origins of symmetrical rowing: a drag-based strategy in Veliidae from fast-flowing waters and a surface-tension-based strategy in Gerridae from still or slow-flowing waters. However, the leg micromorphology and motion patterns underlying these strategies remain underexplored. Using scanning electron microscopy and high-speed video, we compared Rhagovelia distincta (Veliidae), which uses midleg pretarsal fans as oar-like blades, with Gerris latiabdominis (Gerridae), which relies on long, hairy midlegs to generate surface-tension-based thrust. R. distincta performed short, high-frequency strokes and actively controlled its fans, which function as “leaky paddles” exploiting drag, and potentially lift, forces. Fan anatomy further suggests nano-structural adaptations for enhanced mechanical performance. R. distincta also engaged its midleg tarsi in surface-tension-based thrust. In contrast, G. latiabdominis exhibited longer stroke durations and kinematics suited to surface-tension-based propulsion. Both species shared key micromorphological features: ventral longitudinal rows and gaps of midleg setae, with posterior rows particularly robust and nano-grooved in G. latiabdominis. Additionally, both formed ventral “beam-like” structures from overlapping flat-tipped setae on hindlegs—and less prominently on forelegs—used for support and sliding. These findings generate new hypotheses for refining models of locomotion on water surface by insects.","manuscriptTitle":"Distinct kinematics and micromorphology for symmetrical rowing and sliding on water in ripple bugs and water striders","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-07-10 08:04:03","doi":"10.21203/rs.3.rs-6903117/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2025-07-29T07:21:23+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-07-18T15:42:01+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-07-15T12:20:47+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"239627961531992450222852089249124954365","date":"2025-07-07T14:16:58+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"65248693320444878567130358056791531364","date":"2025-07-07T09:03:10+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2025-07-07T07:58:49+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2025-07-07T07:55:30+00:00","index":"","fulltext":""},{"type":"editorInvited","content":"","date":"2025-06-18T10:57:11+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2025-06-17T17:13:15+00:00","index":"","fulltext":""},{"type":"submitted","content":"Scientific Reports","date":"2025-06-16T08:02:32+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"scientific-reports","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"scirep","sideBox":"Learn more about [Scientific Reports](http://www.nature.com/srep/)","snPcode":"","submissionUrl":"","title":"Scientific Reports","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Scientific Reports","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"1d09471c-c127-4df7-b3d9-cb815a44edef","owner":[],"postedDate":"July 10th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[{"id":51158985,"name":"Biological sciences/Ecology/Behavioural ecology"},{"id":51158986,"name":"Biological sciences/Zoology/Animal behaviour"},{"id":51158987,"name":"Biological sciences/Zoology/Biomechanics"},{"id":51158988,"name":"Biological sciences/Zoology/Entomology"}],"tags":[],"updatedAt":"2026-01-05T16:06:12+00:00","versionOfRecord":{"articleIdentity":"rs-6903117","link":"https://doi.org/10.1038/s41598-025-28453-0","journal":{"identity":"scientific-reports","isVorOnly":false,"title":"Scientific Reports"},"publishedOn":"2025-12-29 15:58:12","publishedOnDateReadable":"December 29th, 2025"},"versionCreatedAt":"2025-07-10 08:04:03","video":"","vorDoi":"10.1038/s41598-025-28453-0","vorDoiUrl":"https://doi.org/10.1038/s41598-025-28453-0","workflowStages":[]},"version":"v1","identity":"rs-6903117","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-6903117","identity":"rs-6903117","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}