Plethodontid salamanders inhabit terrestrial, scansorial, arboreal and troglodytic habitats in which clinging and climbing allow them to access additional food and shelter as well as escape from unfavorable temperature and moisture conditions and ground-dwelling predators. Although salamanders lack claws and toe pads found in other taxa, they successfully cling to and climb on inclined, vertical and inverted substrates in nature. Maximum cling angle was tested on smooth acrylic, and the relationship between cling angle, body mass and surface area of attachment (contact area) was investigated. This study found that many salamander species can cling fully inverted using only a portion of their ventral surface area to attach. Salamanders fall into three functional groups based on mass and maximum cling angle: (1) high-performing, very small salamanders, (2) moderately high performing small and medium-sized salamanders and (3) low-performing large salamanders. They show significant differences in maximum cling angle, even between species of similar mass. In species of similar mass experiencing significantly different detachment stress (resulting from significantly different contact area), differences in morphology or behavior affect how much body surface is attached to the substrate. High performance in some species, such as Desmognathus quadramaculatus, is attributable to large contact area; low performance in a similarly sized species, Ensatina eschscholtzii, is due to behavior that negatively impacts contact area. There was no clear evidence of scaling of adhesive strength with increasing body size. Salamander maximum cling angle is the result of morphology and behavior impacting the detachment stresses experienced during clinging.

Animals may climb to avoid competition, locate additional food, escape ground-dwelling predators, or find more suitable shelter or microclimate. In addition, moving through mountainous or forested habitats requires animals to be able to climb over obstacles such as rocks and fallen trees. The ability to climb, defined as locomotion on substrates with vertical or steeply sloping surfaces (Cartmill, 1985), requires an animal to not only be able to generate and maintain attachment on the substrate (clinging) but also to remove and replace components of that attachment surface, such as their feet, body and tail, as they move upward (climbing). As a result, the study of clinging can help shed light on some of the constraints of climbing capabilities of organisms and the effect of surface features on animal performance. Here we test maximum cling angle on smooth surfaces in plethodontid salamanders, a large family of semi-aquatic, terrestrial and scansorial amphibians whose clinging and climbing abilities are poorly studied. We measure the amount of surface area used for attachment when clinging (called hereafter ‘contact area’) by salamander species differing in body size, morphology and ecological niche, to better understand how clinging is accomplished and investigate functional constraints that may determine performance.

Salamander habitat use

Plethodontid salamanders, the largest and most diverse group of salamanders, access and occupy elevated and inclined surfaces in their natural habitats (McEntire, 2016). Species from the genera Bolitoglossa, Chiropterotriton, Dendrotriton, Ixalotriton, Nototriton, Pseudoeurycea and Thorius live in arboreal neotropical habitats in Mexico, and Central and South America (McEntire, 2016; Wake, 1987). In the USA, salamanders of the genus Aneides occupy arboreal habitats seasonally, and Aneides vagrans has been documented climbing to 93 m, living year-round in the fern mats of old-growth redwood trees (McEntire, 2016; Spickler et al., 2006). Scansoriality is even more widespread within the family. In addition to the scansorial cave-dwellers Chiropterotriton magnipes and some Hydromantes species, up to 45% of terrestrial and semi-aquatic species have been documented climbing on temporal scales ranging from short periods of nocturnal foraging up plant stems to year-round occupation of tree trunks, rock crevices, tallus slopes,and cave walls (Bradley and Eason, 2018; Camp et al., 2013; Crawford and Peterman, 2013; Gorman and Camp, 2006; Huheey and Brandon, 1973; Lunghi et al., 2017; McEntire, 2016; Spickler et al., 2006; Waldron and Humphries, 2005).

Ability to climb allows salamanders to distribute across habitats in three dimensions. Climbing may allow smaller salamanders to avoid predation and competition for food. In the Appalachian Mountains, smaller species of the genus Desmognathus are found clinging to and climbing up wet rock faces, with predatory Desmognathus quadramaculatus below (Crawford and Peterman, 2013). In the Pacific Northwest, where arboreal Aneides species live, Dicamptodon salamanders prey on other amphibian species (Bury, 1972), and climbing may provide one means of avoiding predation. Climbing rock faces, tree trunks and cave walls may allow salamanders to find nest sites in cracks (Lunghi et al., 2014, 2015; Myer, 1958; Spickler et al., 2006; Waldron and Humphries, 2005). Eurycea and Plethodon species have been documented climbing plant stems and tree trunks at night to forage for insects (Jaeger, 1978; Legros, 2013). In temperate regions, scansorial habitats may provide more suitable temperature and moisture conditions, as in the seasonal occupation of caves by species such as some Hydromantes species and Eurycea lucifuga (Forsman and Swingle, 2007; Gorman and Camp, 2006; Lunghi et al., 2017; Spickler et al., 2006; Wake, 2014). Plethodontid salamanders are lungless and dependent on microclimatic conditions to maintain the diffusion of oxygen across the moist skin surface; the ability to access more suitable temperature and moisture conditions could exert a strong selection pressure on climbing ability (Feder, 1983).

Clinging versus climbing

The relationship between passive cling ability and active climbing performance is complicated by differences in the kinematics, timing and body surfaces involved between a stationary and a dynamic movement. Differences between maximum cling angle and maximum climbing angle and velocity can be driven by the interaction between dynamic attachment surfaces and the substrate characteristics, like moisture (Stark and Yanoviak, 2018). In some cases, animals are able to cling to surfaces at angles where they cannot climb; in others, rapid movement up a surface enables climbing where clinging would be impossible (Emerson and Diehl, 1980; Hanna and Barnes, 1991; Stark et al., 2015). Nonetheless, study of a species’ capacity for attachment to surfaces and of the factors that determine attachment success develops our understanding of the attachment mechanism and strength. It can also inform predictions about climbing capabilities. Clinging on inclined (<90 deg to horizontal), vertical (90 deg), overhanging (>90 deg) and inverted (180 deg) surfaces requires animals to create attachment points that can resist shear and normal forces, the directional components of the force of gravity acting on the animal's body mass, acting to remove the animal from the surface (Fig. 1).

Fig. 1.

Salamander clinging free body diagrams. Free body diagrams indicating forces during stationary clinging on an inclined (0 to 89 deg), vertical (90 deg), overhanging (91 to 179 deg) and inverted (180 deg) surface. Forces indicated are the force of gravity (FG), and its component forces, depending on the angle of the substrate: the compressive force (FC), the shear force (FS) and the tensile force (FT). Also shown are the equal and opposite forces resisting the force of gravity: the normal force (FN), the viscous force (FV) and the adhesive force (FA). The angle of cling is indicated by θ.

Fig. 1.

Salamander clinging free body diagrams. Free body diagrams indicating forces during stationary clinging on an inclined (0 to 89 deg), vertical (90 deg), overhanging (91 to 179 deg) and inverted (180 deg) surface. Forces indicated are the force of gravity (FG), and its component forces, depending on the angle of the substrate: the compressive force (FC), the shear force (FS) and the tensile force (FT). Also shown are the equal and opposite forces resisting the force of gravity: the normal force (FN), the viscous force (FV) and the adhesive force (FA). The angle of cling is indicated by θ.

Mechanisms of clinging

Attachment can be generated in multiple ways, but two major categories of attachment are those created by interlocking the surface between the animal and the substrate, as in gripping with claws or toes, and those creating bonds or seals between the animal and the substrate that are too strong to be broken by the animal's body weight, as in wet adhesion, fibrillar adhesion and suction (Cartmill, 1985; Gorb, 2008; Nachtigall, 1974). Many animals use more than one of these attachment mechanisms: interlocking with claws or bristles, fibrillar adhesion, suction or wet adhesion. Many species of anoles and geckos have claws that are effective on roughened or soft, penetrable substrates and fibrillar adhesive pads that work on smooth substrates (Autumn et al., 2002; Autumn, 2006; Crandell et al., 2014; Russell and Higham, 2009; Zani, 2000). Frogs have sufficient dexterity to grip with the toes, their body is covered in sticky mucus, and in scansorial species, their toes end in specialized wet-adhesive pads (Emerson and Diehl, 1980; Endlein et al., 2013a,b; Federle et al., 2006; Green, 2008; Green and Simon, 1986; Hanna and Barnes, 1991). Suction attachment exists not only in aquatic vertebrates, some of which use suction in extreme clinging and climbing behavior, but also in certain species of bats (Beckert et al., 2015; Christy and Maie, 2019; Fulcher and Motta, 2006; Green and Barber, 2009; Maie et al., 2012; Riskin and Fenton, 2011; Riskin and Racey, 2010; Wainwright et al., 2013). In wet and dry adhesion as well as in suction, the strength of the attachment is partially dependent on the contact area, unlike in interlocking mechanisms (Stefan, 1874; Vogel, 2003).

With the entire ventral surface of the salamander coated in a thin layer of mucus, wet adhesion is the most likely mechanism of attachment on smooth substrates where attachment by gripping is impossible. Salamanders do not have claws, and inspection of foot and toe surfaces shows no sign of suction cups or the hierarchical series of attachment structures commonly found in specialized clinging and climbing species, such as geckos, anoles, tree frogs, bats, clingfish and insects (Autumn et al., 2002; Autumn, 2006; Crandell et al., 2014; Emerson and Diehl, 1980; Federle, 2002; Green and Alberch, 1981; Federle, 2006; Federle et al., 2006; Hanna and Barnes, 1991; Labonte and Federle, 2015; Riskin and Racey, 2010; Tian et al., 2006; Wainwright et al., 2013). Microscopic examination of foot and toe morphology in five species from the genus Bolitoglossa found no evidence of toe pad structures; instead, the foot is a smooth surface covered in mucus glands (Green and Alberch, 1981). At angles below 90 deg, the animal–mucus–substrate interface will resist shear according to the properties of the mucus layer and its thickness (Fig. 1) (Vogel, 2003). Clinging at 90 deg, the entire body weight of the salamander is supported by this resistance of the mucus to shearing forces caused by gravity (Fig. 1). At overhanging angles greater than 90 deg, the attachment of the animal is dependent on both the viscous and adhesive properties of the mucus layer, which will experience increasing normal forces and decreasing shear forces with increasing angle (Fig. 1). Finally, in inverted attachment (at 180 deg) the full body weight of the salamander is supported by the adhesive properties of the mucus layer resisting the normal stress caused by gravity (Fig. 1). Adhesive strength is generated either by the surface tension of the fluid layer (capillarity) or by its viscosity (as in Stefan adhesion) and is dependent on contact area (Barnes et al., 2006; Emerson and Diehl, 1980; Federle, 2002; Stefan, 1874; Vogel, 2003).

Scaling of clinging

Body mass and dimensions can affect clinging performance. Under isometric scaling, in a clinging organism, surface area scales as with mass or volume to the two-thirds power, and detachment force scales as mass (Schmidt-Nielsen, 1975). If salamander ventral surface area and body mass scale isometrically, either ontogenetically or through evolutionary time, large salamanders will experience large detachment forces with relatively less surface area for adhesive attachment (Adams and Nistri, 2010; Jaekel and Wake, 2007). This is due to increasing gravitation force acting on their body mass. Large salamanders would be expected to detach at lower angles. Increasing the substrate angle will increase the detachment forces acting in shear (at angles of 90 deg and below) or in shear and tension (at angles of 90–180 deg) on the skin–mucus–substrate interface.

Maximum cling angle could be improved in at least three ways: (1) morphological changes to increase the ventral surface area of the salamander through allometric scaling of surface area relative to body mass, (2) chemical changes in the mucus layer to increase its adhesive effectiveness per unit area, or (3) behavioral changes by the salamander to respond to increasing detachment forces by increasing the contact area with the substrate. Behavioral changes to increase contact area would only be possible if the salamander was not already operating at the functional limit of its ventral surface area; if the salamander is already fully sprawling, postural changes will not add contact area. All three of these mechanisms (morphological, chemical and behavioral) could be operating at once to maximize cling angle.

The signature of a change in morphology to increase attachment surface might take the form of a redistribution of the same total body mass to increase area in the relevant dimension, or an increase in the area of specific specialized attachment surfaces, like the feet and tail. Allometric scaling of foot surface area has been found in several species of cave-dwelling plethodontids in the genera Hydromantes and Chiropterotriton but has been ruled out in the arboreal and web-footed species of Bolitoglossa (Adams and Nistri, 2010; Adams et al., 2017; Jaekel and Wake, 2007; Salvidio et al., 2015). Investigation into the relationship between habitat, particularly arboreality, and morphology found no evidence of an arboreal phenotype (Baken and Adams, 2019). One suggestion is that webbed feet in Bolitoglossa could be providing enhanced attachment via suction (Alberch, 1981), although the morphology of bolitoglossan feet lack several of the functional features found in other biological suction cups, such as a fleshy lip surrounding a round ‘cup’ to prevent fluid leakage, and a concave internal structure in which the low pressure region can develop (Beckert et al., 2015; Maie et al., 2012; Nachtigall, 1974). A broadly comparative study of 225 species of clinging and climbing vertebrates and invertebrates found that across distantly related taxa, larger species were more likely to develop increased surface area to increase attachment, but within closely related taxa, larger species were more likely to enhance attachment by increasing stickiness (Labonte et al., 2016).

In this study, we tested smooth-substrate cling performance in plethodontid, and as an outgroup comparison, ambystomatid salamanders. Salamanders lack the morphological specializations found in other scansorial amphibian species, such as claws and toe pads (Green and Alberch, 1981), and probably cling using wet adhesion. As a result, individual maximum cling angle may be limited by the ratio of surface area to body mass, unless changes in body shape, clinging behavior or adhesive effectiveness occur with increasing body size. We measured the amount of contact area used by salamanders when clinging maximally, and investigated the links between morphology, detachment stress and cling angle. We examined whether salamanders respond to increasing cling angle behaviorally by crouching to increase contact area between their ventral surface and the substrate. Using pairwise comparisons between size-matched species, we investigated how species with similar detachment forces use different morphology, adhesive strength and behavioral strategies that affect maximum cling angle (Table S2).

Animals

Animals were collected from natural populations in Chiapas, Mexico, and California and North Carolina, USA. Salamanders were housed individually in plastic enclosures on a substrate of damp unbleached paper towels at 16–20°C on a 12 h:12 h light:dark schedule. Ambient moisture levels were standardized across species and trials by controlling the amount of water added to paper towels in the enclosure to maintain a level of 84±10% humidity. Species were fed on a diet of vitamin-dusted crickets or fruit flies, depending on size. A total of 233 individuals from 20 species [Ambystoma gracile (Baird 1857), Ambystoma maculatum (Shaw 1802), Aneides aeneus (Cope & Packard 1881), Aneides flavipunctatus (Strauch 1870), Aneides lugubris (Hallowell 1849), Aneides vagrans Wake and Jackman 1998, Batrachoseps attenuatus (Eschscholtz 1833), Bolitoglossa franklini (Schmidt 1936), Desmognathus aeneus Brown and Bishop 1947, Desmognathus ocoee Nicholls 1949, Desmognathus quadramaculatus (Holbrook 1840), Ensatina eschscholtzii Gray 1850, Eurycea guttolineata (Holbrook 1838), Eurycea lucifuga Rafinesque 1822, Eurycea wilderae Dunn 1920, Hydromantes platycephalus (Camp 1916), Plethodon elongatus Van Denburgh 1916, Plethodon metcalfi Brimley 1912, Pseudoeurycea leprosa (Cope 1869) and Pseudotriton ruber (Sonnini de Manoncourt & Latreille 1801)] were used in the study (Table 1). All procedures were approved by the Institutional Animal Care and Use Committee (IACUC) of the University of South Florida.

Table 1.

Maximum cling angle and body size of salamander species

Maximum cling angle and body size of salamander species
Maximum cling angle and body size of salamander species

Maximum cling angle

Animals were removed from their enclosures by hand and placed on a clear, dry sheet of acrylic (Plaskolite Optix, Columbus, OH, USA) fastened on a rotating hinge of adhesive tape to a laboratory table edge at an angle of 0 deg relative to horizontal. Acrylic was selected as the attachment substrate due to its wide use in attachment experiments across other taxa (Federle et al., 2003; Hanna and Barnes, 1991; Riskin and Racey, 2010; Russell and Higham, 2009; Smith, 1991). Humidity, and its effect on amphibian mucus, was standardized (at 84±10%) by controlling the amount of water added to enclosures prior to all trials so that animals were well hydrated across all experiments. Animals were replaced in enclosures between trials to limit desiccation and prevent altered adhesive performance due to drying. All animals were oriented in the same direction, facing away from the point of rotation, resulting in a head-up orientation at any angle between 0 and 180 deg. After a 30 s acclimation period, the acrylic was rotated by hand at a rate of 3 deg s–1 until the animal detached or until it was fully inverted (angle of 180 deg). If the animal achieved an angle of 180 deg, a stopwatch recorded cling time up to 60 s, at which time the trial ended. The angle at which the animal detached was measured, and animals were returned immediately to their enclosures.

Animals were tested in no more than three trials per day with a rest of at least 1 h between trials. Trials in which animals voluntarily jumped off were not analysed. Animals were tested in at least three trials, or up to five in total. If a maximal possible performance of 180 deg cling attachment for at least 60 s occurred three times in a row, no more trials were conducted. Otherwise, five trials were conducted.

Contact area

Contact area trials were conducted separately from maximum cling angle trials described above, due to time spent in developing a mechanism for imaging salamander attachment area. Due to equipment constraints, contact area was not measured continuously, but at prescribed angles (0, 45, 90, 135 and 180 deg).

Salamanders were removed from their enclosures by hand and placed on a clear, dry acrylic sheet (Plaskolite Optix) mounted in a wooden frame that could be smoothly rotated 180 deg. LED strip lights (Commercial Electric, Cleveland, OH, USA) were mounted on the flame-polished edges of the acrylic on all four sides so that the light passed into the plane of the acrylic sheet and was trapped via frustrated total internal reflection (as described in Betts et al., 1980 and Endlein et al., 2013b). When animals were placed in contact with the acrylic, ventral attachment surfaces in intimate contact with the acrylic were illuminated by LED light. All animals were oriented in the same direction, facing away from the point of rotation, resulting in a head-up orientation at any angle between 0 and 180 deg. After a 30 s acclimation period, the acrylic was rotated by hand at a rate of 3 deg s–1 to the next prescribed angle. Trials in which animals voluntarily jumped off were not analysed. The ventral surface area of attachment, called contact area, was imaged at angles of 0, 45, 90 (vertical), 135 and 180 deg (inverted) for each trial until the animal detached.

Animals that were unable to attach to a prescribed angle in five successive trials were scored as unable to attach at that angle. Successive trials were conducted in reverse, beginning with placement and the acclimation period at 0 deg, and then collecting images at angles of 180 deg, and ending with angle 0 deg, to account for the potential effects of fatigue or drying on attachment performance at the later angles within a trial. Subsequent analysis found that the order of angles had no effect on the surface area of attachment or maximum attachment angle. The ventral surface of the attached animals was imaged using an iPhone 6 or iPhone SE (Apple, Cupertino, CA, USA) front-facing camera mounted on a wooden frame perpendicular to the acrylic sheet.

All contact area values reported come from trials in which species achieved maximum prescribed cling angle from this experiment (possible values of 0, 45, 90, 135 or 180 deg). Species' mean cling angle and standard error of the mean (s.e.m.) is reported in each table (Tables 25) derived from the maximum cling angle (among the five prescribed angles) for all individuals, as well as the number of individuals, N, included in that analysis. Minimum contact area refers to a mean±s.e.m. species value generated from the single trial from each individual in which they reached their maximum cling angle (among the five prescribed angles) using the smallest amount of contact area (Table 2). This would represent the highest detachment stresses resisted by that individual at that angle (Tables 3 and 4).

Table 2.

Contact area during clinging at prescribed angles

Contact area during clinging at prescribed angles
Contact area during clinging at prescribed angles
Table 3.

Maximum shear stress of salamander species

Maximum shear stress of salamander species
Maximum shear stress of salamander species
Table 4.

Maximum normal stress of salamander species

Maximum normal stress of salamander species
Maximum normal stress of salamander species
Table 5.

Change in contact area at maximum cling angle

Change in contact area at maximum cling angle
Change in contact area at maximum cling angle

To examine whether species with high maximum cling angle were responding behaviorally to increasing angle by increasing their contact area with the substrate, and thereby increasing their attachment strength, we measured changes in contact area within a cling trial (Table 5). We compared contact area on the horizontal surface (0 deg) with the contact area attachment at maximum cling angle, using the trial in which each individual reached its maximum cling angle with the largest difference in contact area between peak cling and the horizontal condition. This trial was used to generate species means because this dataset would capture either the largest gains or largest losses in surface area during a cling trial for each individual. If behavioral addition of contact area or stress-induced loss of contact area at increasing angles were driving maximum cling angle, it would be present in the data. Contact area values represent the mean±s.e.m. contact area at both 0 deg (horizontal) and the maximum prescribed cling angle for that species (Table 5). Data on E. lucifuga contact area on the horizontal surface were not available, owing to constraints of testing in the field.

One species, E.eschscholtzii, was additionally tested for contact area while sedated. These individuals had been sedated using 1% buffered solution of MS-222 for another experiment occurring simultaneously within the laboratory, and their maximum cling angle and contact area while passively clinging were recorded. These data were not analysed statistically, but differences in maximum cling angle between active and sedated individuals are reported below.

Field tests

Some species (A.aeneus, A.vagrans, A.gracile, E.lucifuga, H.platycephalus and P.elongatus) were tested for maximum cling angle and contact area at the site of animal capture or nearby, using a miniaturized, battery-powered version of the LED-illuminated cling apparatus mounted on two tripods (MeFoto Roadtrip, Benro, North White Plains, NY, USA; G-Geekeep Waterproof Flexible LED Light Strips, Amazon, Seattle, WA, USA). In the field, animals were tested in the same manner as described above, with the caveat that time and resource constraints frequently led to fewer trials or fewer experiments being conducted. Humidity was controlled by placing the animals in a prepared deli cup with damp paper towels (as in laboratory testing) for at least 30 min prior to testing. Hydromantes platycephalus and E. lucifuga were tested in five full trials for maximum cling angle, but a single contact area trial per individual was captured (Tables 15). For this reason, surface area values for H. platycephalus and E. lucifuga are reported only as the minimum contact area to maintain attachment. Field-tested animals were released at the point of capture within 24 h. Additional species were tested in cling trials but full datasets could not be collected owing to time and field constraints (Table S1).

Digitizing images

Images were digitized using ImageJ software (Schneider et al., 2012). The Color Thresholding function was used to select each illuminated contact area on the animal in the image. All selected illuminated contact areas were then added to the ROI Manager and converted to a known contact area via the Measure function. The Measure function was calibrated using a scale bar attached to the acrylic and visible in each image.

Stress calculations

Shear (parallel) and normal (perpendicular) stress were calculated by resolving the vectors based on the angle of the substrate and calculating the shear and normal force acting at that angle trigonometrically (Fig. 1). Force was calculated from the angle and the body mass of the individual animals. Contact area was measured from images of the animals’ illuminated attachment surfaces. Contact area data were imaged at five prescribed angles (0, 45, 90, 135 and 180 deg), as described above, and forces calculated at the matching angles. Stress was calculated as the force divided by the contact area (in kPa).

For calculations of shear stress at angles from 0 to 90 deg (the point of maximum shear stress), the ventral contact area was measured at 45 or 90 deg (depending on each individual animal's maximum performance) (Fig. 1). For calculations of normal stress (or the tensile force being resisted per unit contact area while clinging on overhanging or inverted surfaces) at angles from over 90 to 180 deg, the contact area was measured at 135 or 180 deg (depending on each individual's maximum performance) (Fig. 1). Species’ mean values (±s.e.m.) for maximum shear and normal stress and cling angle were reported (Tables 3 and 4).

Statistical analysis

Maximum cling angle data were tested for normality and homogeneity of variance using a Shapiro–Wilk test and Levene's test, respectively. Individuals’ maximum cling angle data were non-normal and heteroscedastic owing to the extremely high number of 180 deg cling trials from high-performing species. Only the trial in which the animal reaches the maximum cling angle for each individual was analysed. All statistical analyses were conducted in R 3.5.2 (https://www.r-project.org/). Figures were created using ggplot2 (Wickham, 2009).

Statistical tests

Differences in maximum cling angle across species were analysed using a non-parametric Kruskal–Wallis test (Fig. 2).

Fig. 2.

Salamander maximum cling angle and change in contact area during clinging. (A) Maximum cling angle (deg) of species. The enlarged point indicates the average maximum cling angle of species, and black points are the maximum cling performance for each tested individual. Color indicates species (see Table 1). N for each species is also reported in Table 1. (B) Boxplot of average change in contact area (in mm2) during clinging from 0 deg to maximum cling angle (out of five prescribed angles: 0, 45, 90, 135 and 180 deg). Positive values indicate gain of contact area, negative values indicate loss. N for each species is reported in Table 5.

Fig. 2.

Salamander maximum cling angle and change in contact area during clinging. (A) Maximum cling angle (deg) of species. The enlarged point indicates the average maximum cling angle of species, and black points are the maximum cling performance for each tested individual. Color indicates species (see Table 1). N for each species is also reported in Table 1. (B) Boxplot of average change in contact area (in mm2) during clinging from 0 deg to maximum cling angle (out of five prescribed angles: 0, 45, 90, 135 and 180 deg). Positive values indicate gain of contact area, negative values indicate loss. N for each species is reported in Table 5.

The effect of body mass and contact area on maximum cling angle was investigated using phylogenetic generalized least squares (PGLS). Phylogenetic relationships of these species are based on the topology and branch lengths of Bonett and Blair (2017), pruned to include only the taxa from this study using the ape package (Paradis and Schliep, 2018) (Fig. S1). Model selection was conducted using the Akaike information criterion with a correction for small sample sizes using the AICcmodavg package (https://cran.r-project.org/package=AICcmodavg). PGLS was performed using a Brownian motion model of evolution with the nlme package (https://cran.r-project.org/package=nlme). Contact area measurements were corrected for body size by first regressing the natural log of contact area against the natural log of body mass and calculating the residuals (Fig. 3). The effect of body mass and size-corrected contact area on maximum cling angle was then determined (Fig. 4).

Fig. 3.

Effect of body mass on contact area. Scatterplot of average minimum contact area during maximal clinging against average body mass, by species. The grey line shows phylogenetic generalized least squares (PGLS) regression, with formula as indicated in the top left. Color indicates species (see Table 1). N for each species is reported in Table 2. Large, poorly performing species are highlighted with triangular points.

Fig. 3.

Effect of body mass on contact area. Scatterplot of average minimum contact area during maximal clinging against average body mass, by species. The grey line shows phylogenetic generalized least squares (PGLS) regression, with formula as indicated in the top left. Color indicates species (see Table 1). N for each species is reported in Table 2. Large, poorly performing species are highlighted with triangular points.

Fig. 4.

Effect of body masson maximum cling angle. Scatterplot of average maximum cling angle against average body mass, by species. The grey line shows PGLS regression, with formula: y=−17.23x+162.87 (P<0.001). Color indicates species (see Table 1). N for each species is also reported in Table 1 and each individual's maximum cling angle is indicated by the smaller, translucent point. Large, poorly performing species are highlighted with triangular points.

Fig. 4.

Effect of body masson maximum cling angle. Scatterplot of average maximum cling angle against average body mass, by species. The grey line shows PGLS regression, with formula: y=−17.23x+162.87 (P<0.001). Color indicates species (see Table 1). N for each species is also reported in Table 1 and each individual's maximum cling angle is indicated by the smaller, translucent point. Large, poorly performing species are highlighted with triangular points.

A post hoc comparison of pairwise differences between species in maximum cling angle, contact area, shear stress and normal stress was conducted on specifically chosen species matching in body mass (Table S2) using a Wilcoxon rank sum test for non-parametric data and corrected for false discovery rate (Benjamini and Hochberg, 1995). Pairs were selected after the initial analysis of maximum cling angle from among the 20 tested species to examine how differences in morphology or behavior might drive differences in maximum cling angle despite similarity in body size, through contact area and detachment stress.

Clinging behaviors

Salamanders showed a wide range of responses to cling trials, only some of which are useful for maintaining attachment. Small salamander species (D.aeneus, D.ocoee, E.wilderae) frequently showed a tendency to re-orient with their head downhill, and then jump or laterally undulate down the acrylic sheet until they could leap from the edge. For trials where the re-orientation happened at angles close to 90 deg or overhanging angles between 90 and 135 deg, this re-orientation and escape behavior sometimes resulted in a small salamander clinging upside down, attached by its hindlegs and tail to the acrylic, and then actively leaping off. In larger species, particularly P.ruber and B.franklini, some salamanders responded to increasing angle by releasing a large amount of fluid from their cloaca. In the case of P.ruber, several times animals then re-oriented head downward, and laterally undulated down the incline along the path lubricated by this fluid off the edge of the acrylic. Trials in which salamanders used these non-clinging behaviors were not used in the analysis or recorded as completed trials.

Overall performance

Maximum cling angle was significantly different among salamander species (Kruskal–Wallis test, χ2=166.5, d.f.=19, P=0.22×10−15; Table 1, Fig. 2), and the amount of contact area used to attach by a given species varies over a 2-fold range at the maximum cling angle (Table 2). Minimum adhesive contact area used while clinging was significantly correlated with body mass (Table 2, Fig. 3). Maximum cling angle was significantly correlated with body mass and size-corrected contact area (Tables 1 and 2, Fig. 4). Pairwise comparisons of selected species of similar body mass were conducted to compare contact area, shear stress and normal stress to determine whether significant differences in maximum cling angle were correlated with significant differences in contact area, or in the effect of shear and normal detachment stress on maximum cling angle (Table S2). Salamanders fall into three major functional groups based on body mass and maximum cling angle: (1) high-performing, very small salamanders, (2) moderately performing small and medium-sized salamanders, and (3) poorly performing large salamanders (Table 1; Fig. 3).

Small species

The very highest and most consistent cling angles (Table 1) are seen in the smallest species with the lowest mass (D. aeneus, B. attenuatus, D. ocoee, E. guttolineata and E. wilderae, <1 g). The calculated shear stress experienced by these species at peak (vertical) attachment is low (0.07±0.01 to 0.16±0.03 kPa; Table 3, Fig. 5). Normal stress experienced at peak (inverted) attachment is nearly identical to shear stress (Table 4, Fig. 5), and accordingly these species also show little to no loss of contact area between the 0 deg horizontal condition and the 180 deg inverted condition (Table 5, Fig. 2).

Fig. 5.

Maximum shear and normal stress of species. Shear boxplot bar indicates average the maximum shear stress of species while clinging maximally (out of two prescribed angles: 45 and 90 deg). Not shown is an outlying data point for A. flavipunctatus, where a single individual achieved a maximum shear stress of 2.07 kPa. Normal boxplot bar indicates the average maximum normal stress of species while clinging maximally (out of two prescribed angles: 135 and 180 deg). Species (see Table 1) are listed in order of increasing maximum cling performance from left to right. N for each species is reported in Tables 3 and 4.

Fig. 5.

Maximum shear and normal stress of species. Shear boxplot bar indicates average the maximum shear stress of species while clinging maximally (out of two prescribed angles: 45 and 90 deg). Not shown is an outlying data point for A. flavipunctatus, where a single individual achieved a maximum shear stress of 2.07 kPa. Normal boxplot bar indicates the average maximum normal stress of species while clinging maximally (out of two prescribed angles: 135 and 180 deg). Species (see Table 1) are listed in order of increasing maximum cling performance from left to right. N for each species is reported in Tables 3 and 4.

Medium-sized species

Species between 1 and 4 g in body mass (A. aeneus, A. flavipunctatus, A. vagrans, E. lucifuga, B. franklini, P. leprosa, H. platycephalus, P. elongatus and P. metcalfi) show a narrow range of detachment at high angles (Table 1; Fig. 3). Most of these species cling between 160 and 175 deg, but three species maintain attachment to 180 deg (B. franklini, A. vagrans and E. lucifuga). Shear stress in these species at 90 deg is more than double that of smaller species, despite the inherent size correction. Medium-sized species that fail to cling at angles less than 180 deg show a similar range of normal stresses to smaller species clinging at 180 deg (Table 4, Fig. 5). In the three species that can cling fully inverted (B. franklini, A. vagrans and E. lucifuga), maximum normal stress at 180 deg is higher (0.12±0.02 to 0.20±0.03 kPa; Table 4, Fig. 5). Of the species that can maintain attachment to 180 deg, B. franklini and A. vagrans show loss of 24–40% of contact area, respectively, with increasing angle (Table 5, Fig. 2).

Large species

Large-bodied salamanders (A. gracile, A. maculatum, A. lugubris, D. quadramaculatus, E. eschscholtzii, P. ruber) showed a lower maximum cling angle overall than small and medium-sized species, with the exception of D. quadramaculatus (Table 1, Fig. 2). Shear stress and normal stress in this large species was similar to that of medium-sized species (Tables 3 and 4, Fig. 5). For D. quadramaculatus, as in a few of the small and medium-sized species, contact area increases with increasing angle (Table 5, Fig. 2). The largest and poorest performing species, A. lugubris, E. eschscholtzii and P. ruber, and the outgroup species A. gracile and A. maculatum cling to angles ranging from 76±5 to 114±13 deg (Table 1). Analysis of contact area during clinging shows that these species experience higher shear stress prior to detachment than other species (Table 3, Fig. 5), but they fail before 135 deg so normal stress could not be measured. Comparison of contact area at 0 deg and the highest angle prior to detachment shows a pattern of decreasing contact area at higher angles in four of the five species (Table 5, Fig. 2). Ensatina eschscholtzii was notable owing to its raised posture at all clinging angles. During a later experiment, when sedated E. eschscholtzii were available, a sedated animal placed on the acrylic not only had a much higher contact area with the substrate, but also could passively cling to much higher angles. Ambystoma gracile, the poorest performing species, showed an average gain of 100% in contact area between 0 deg (horizontal) and the peak measured angle, 45 deg, above which it detached (Table 5, Fig. 2).

Salamander clinging

Plethodontid salamanders are broadly capable of generating attachment to smooth surfaces at a range of angles from horizontal (0 deg) to fully inverted (180 deg). Despite a lack of specialized morphology associated with clinging, performance in salamanders is comparable to or even exceeds that of tree frogs (Barnes et al., 2006; Emerson and Diehl, 1980; Hanna and Barnes, 1991) on a smooth acrylic substrate. Salamander cling performance is most likely explained by the mucus coating on the salamander skin generating wet adhesive forces.

High maximum cling angles cannot be attributed to any consistent behaviors by salamanders to promote or maintain attachment. In fact, salamanders clinging maximally use a variety of body surfaces, resulting in a 2-fold range of contact area (Table 2). Tail, feet and the ventral surface of the abdomen and jaw can all be points of contact during clinging (Fig. 6). Unlike tree frogs (Barnes et al., 2006; Emerson and Diehl, 1980; Hanna and Barnes, 1991) the salamanders show no behavioral tendency to add contact area as the substrate angle increases, and in general lose contact area when at the highest angle compared with a horizontal surface (Table 5, Fig. 2). The high variation in contact area during peak cling performance, and the ability by some species to maintain 180 deg attachment despite a loss of up to 40% of contact area suggests that the contact area needed to maintain attachment is much lower than the total surface area available in most species. Salamanders are not using their entire ventral surface area when clinging maximally (Fig. 6). Some species are ‘overbuilt’ for the task of clinging, in particular small and medium-sized species.

Fig. 6.

Images of clinging salamanders. (A,B,H) Bolitoglossa franklini clinging fully inverted (180 deg) using only feet and tail. (C) Aneides aeneus clinging using ventral surface of jaw. (D) Bolitoglossa franklini clinging using feet, trunk and tail, illuminated by frustrated total internal reflection. (A,E) Bolitoglossa franklini clinging with overlapping or detached feet. (F) Aneides lugubris clinging at 60 deg with foot folded backward. (G) Desmognathus aeneus clinging with detached fore and hindfeet. (I) Eurycea wilderae clinging fully inverted (180 deg). (J) Batrachoseps attenuatus clinging fully inverted (180 deg). (K) Desmognathus quadramaculatus clinging on a vertical surface (90 deg). (L) Sedated Ensatina eschscholtzii clinging fully inverted (180 deg). (M) Ensatina eschscholtzii in stationary cling posture at 20 deg.

Fig. 6.

Images of clinging salamanders. (A,B,H) Bolitoglossa franklini clinging fully inverted (180 deg) using only feet and tail. (C) Aneides aeneus clinging using ventral surface of jaw. (D) Bolitoglossa franklini clinging using feet, trunk and tail, illuminated by frustrated total internal reflection. (A,E) Bolitoglossa franklini clinging with overlapping or detached feet. (F) Aneides lugubris clinging at 60 deg with foot folded backward. (G) Desmognathus aeneus clinging with detached fore and hindfeet. (I) Eurycea wilderae clinging fully inverted (180 deg). (J) Batrachoseps attenuatus clinging fully inverted (180 deg). (K) Desmognathus quadramaculatus clinging on a vertical surface (90 deg). (L) Sedated Ensatina eschscholtzii clinging fully inverted (180 deg). (M) Ensatina eschscholtzii in stationary cling posture at 20 deg.

Feet have been the focus of many studies of adhesion in salamanders and other clinging and climbing organisms (Adams and Nistri, 2010; Adams et al., 2017; Alberch, 1981; Autumn et al., 2002; Autumn, 2006; Barnes et al., 2006; Emerson and Diehl, 1980; Federle et al., 2003; Green and Alberch, 1981; Jaekel and Wake, 2007; Labonte and Federle, 2015; Labonte et al., 2016; Salvidio et al., 2015; Tian et al., 2006). Baken and Adams (2019) found that arboreal salamanders do not differ significantly from terrestrial species in foot morphology or body shape. The feet of salamanders contribute only a small portion of the total contact area relative to the ventral surface area of the body available for attachment. Some trials reveal that salamanders cling fully inverted using only their feet and a small portion of the tail and can cling while their feet are detached, overlapping or partially folded (Fig. 6). A surface commonly used in clinging and climbing is the ventral surface of the jaw (Fig. 6); the head may be pressed to the substrate to reduce torque acting to rotate the head away from the substrate and initiate peeling, the most frequently observed form of detachment in this study.

Clinging in small salamanders

The smallest species, under 1 g in body mass, have the highest maximum cling angle, and are uniformly capable of attaching at 180 deg. Isometric scaling principles dictate that they have higher surface area per unit volume (and mass) than a larger salamander of the same shape, and this should result in lower stresses acting in shear (parallel along the surface) on vertical surfaces and in tension (perpendicular to the surface) on inverted surfaces (Schmidt-Nielsen, 1975). Shear stress experienced by these three species at peak shear (90 deg) is similar to that of larger animals (Table 3, Fig. 5), but normal stress at 180 deg is lowest in these small species (Table 4, Fig. 5), perhaps because at peak cling angle small salamanders cling using the same contact area as they have on the horizontal surface, unlike larger species. In large salamanders, loss of contact area at increasing angles is indicative of increasing shear and normal stresses, and a harbinger of cling failure.

Clinging in medium-sized salamanders

The maximum cling angle of the medium-sized salamanders suggests that for some species, either an exceptionally effective mucus improves performance or subtle behavioral shifts result in increased attachment. As the largest species that can cling fully inverted, the cling performance of B. franklini is of particular note. They show a similar amount and range of contact area, shear and normal force as P. metcalfi, the closest tested species in terms of body size, but they significantly outperform P. metcalfi by 20–30 deg (Tables 1–4, Fig. 1). Their high performance cannot be attributed to behavioral adaptations to increase contact area; they show a similar 2-fold range of contact area during peak cling performance seen in most other species (Table 2), and more loss of contact area than the smallest species with the same maximum cling angle (Table 5, Fig. 2). This species does not behaviorally add contact area at increasing angles by assuming a more crouched posture or adhering additional body parts to the substrate, as is seen in tree frogs (Table 5, Fig. 2; Emerson and Diehl, 1980; Hanna and Barnes, 1991). This could mean that despite losing up to 40% of their surface area, B. franklini retain sufficient contact area to attach at 180 deg. However, the facility with which these salamanders cling, and the large body sizes at which they can successfully cling to 180 deg (11.5 g, in one case) suggest either that B. franklini mucus is chemically and mechanically different from other tested salamanders (i.e. they are stickier), or they are using their attachment surfaces differently than other species, for example by attempting to re-attach when peeling starts, and as a result, they have capacity to resist greater normal stresses. Further investigation into cling behavior, climbing behavior and the chemical and physical properties of mucus in this species may indicate which of these are driving high cling performance.

Clinging in large salamanders

Body size affects cling performance, as in tree frogs, with the largest species having the lowest detachment angles (Table 1, Figs 2 and 4). Failure to cling at angles less than 90 deg can be attributable to the increase in shear stress with increasing angle, which explains cling failures in some species (A. gracile, A. lugubris, E. eschscholtzii, A. maculatum and P. ruber). Although the exact shear stress at the moment of failure was not measured in this experiment, only species with measured shear stresses over 0.25 kPa fail at angles less than 90 deg (Table 3, Fig. 5). In cling failures between 90 and 135 deg, failure is due to peeling, which often initiates at the tip of the jaw or the most uphill portion on the head and rapidly propagates downwards, resulting in detachment. One explanation is that the small normal stress acting on the animal perpendicular to the substrate is sufficient to begin the peeling event. The largest species experience the largest torques; in addition, the largest species show the highest loss of contact area between their horizontal posture and their last measured angle before detachment, which probably contributes to their rapid detachment (Table 5, Fig. 2).

Within large salamanders, there seems to be evidence that the mucus coating differs among species chemically or mechanically in its resistance to shearing and normal forces. Ambystoma gracile has a similar body weight to A. maculatum, but significantly more contact area (Table 2, Table S2). Ambystoma gracile is also the only species that routinely adds substantial contact area between the horizontal and 45 deg condition, which results in lower shear stress than A. maculatum (Tables 3–5, Table S2). However, A. gracile fails at significantly lower angles than A. maculatum (Table 1). Potentially, this could be evidence that species-specific differences in the chemical or physical properties of mucus can determine maximum cling angle beyond the amount of surface area in contact with the substrate.

There is also evidence that body shape and cling behavior determine maximum cling angle in other large species. Desmognathus quadramaculatus is unusual in that it is a large species with high maximum cling angle. They cling better than similarly sized E. eschscholtzii (Table 1; Table S2), with higher contact area and lower shear stress (Tables 2 and 3; Table S2). This can potentially be attributed two main differences between the two species: first, that the posture of D. quadramaculatus is much more sprawled than that of E. eschscholtzii (Fig. 6). In fact, when sedated using MS-222 and placed on the acrylic in a prone posture, E. eschscholtzii not only had a greater contact area, but also had a much higher maximum cling angle than when unsedated (Fig. 6; M.K.O.D. and S.M.D., personal observation). In addition, D. quadramaculatus occupies streambed habitats, and is frequently found beneath rocky bed elements. It has a ventrally flattened profile, which may result in a body mass distributed across a larger ventral surface area than the more cylindrical trunk of E. eschscholtzii (3.96 times as much contact area; Table 2, Table S2). Sprawled posture and shorter limbs may also decrease torques acting on the body at increasing angles, reducing the risk of detachment by peeling. In fact, D. quadramaculatus shows a small net gain in contact area between the horizontal and peak cling angle (Table 5, Fig. 2). Body shape may equip certain species with more ventral surface area, which has the potential to be used to attach, but contact area depends on behavior during clinging for its application.

Scaling of cling performance

A broadly comparative study of 225 species of clinging and climbing vertebrates and invertebrates found that across distantly related taxa, larger species were more likely to develop increased surface area to increase attachment, but within closely related taxa, larger species were more likely to enhance attachment by increasing stickiness (Labonte et al., 2016). We found that larger species cling poorly, and do not show enhanced ability to resist detachment forces, while smaller species cling well at high angles. Salamanders may therefore not conform to the general scaling pattern derived from comparative analyses of clinging and climbing animals in which larger animals show positive allometry or increased stickiness of attachment surfaces (Labonte et al., 2016). Drawing precise conclusions about the scaling of adhesive area or adhesive strength in salamanders is constrained by two challenges. First, salamander adhesion is not limited to specialized adhesive pads, which means that the scaling of adhesive area is confounded between the scaling of available ventral body surface and the behavioral use of these areas by a clinging animal. This study indicates that some species are more constrained by behavior than by body shape, and suggests that for other species, the opposite is true. The second challenge is in determining whether adhesive strength scales with body size across salamanders, or even within high-performing species. Owing to the rotated platform experimental design, any species that detaches probably reached a maximum shear or normal force that caused attachment failure. Any species that clings at 180 deg has not reached its maximum adhesion force yet, either owing to adhesive strength or the contact area over which the normal force is being distributed. Additional experiments are needed to determine maximum attachment strength and stress in order to compare across species independent of cling behavior.

Conclusions

Plethodontid salamanders are capable of clinging fully inverted to smooth substrates across a range of body sizes, morphologies and ecological niches. These experiments support the conclusion that salamander clinging performance is equal to or exceeds that of arboreal and scansorial frogs. Many portions of the body surface can be used in successful clinging, including the ventral surface of the jaw, trunk and tail, as well as the feet. Unlike in tree frogs, species barely responded behaviorally to increases in substrate angle; total contact area and maximum cling angle are largely determined by body shape and posture at 0 deg (horizontal).

Cling performance is impacted by the scaling of contact area to body mass, which determines the stresses acting on the substrate–mucus–skin interface. The smallest salamanders experience the lowest shear and normal stress and are able to resist these stresses to cling fully inverted. In the ambystomatid outgroup, and in the largest tested species of salamanders, detachment occurs at lower angles. Larger salamanders attach with less contact area per unit body mass, experience higher shear and normal stresses, and generally show decreasing amounts of contact area at increasing angles. As a result, large salamanders fail at lower angles.

Species-specific high maximum cling angle provides evidence that some salamanders may have unique morphological, behavioral or chemical traits that can affect their ability to cling. In the case of A. gracile, low performance despite moderate shear stress and high contact area suggest that their mucus is more susceptible to shearing stress than other species. Desmognathus quadramaculatus, in comparison with similarly sized E. eschscholtzii, clings to significantly higher angles, as the result of a more sprawling posture and more ventrally flattened profile that increase contact area. The highest maximum cling angle was found in B. franklini, which appears to be especially sticky, better able to prevent a peeling event, or using unique morphology or behavior, or all three; investigation into the mechanism of attachment and the material properties of secreted mucus may reveal differences in composition and adhesive properties.

We thank J. Scales for providing help in the designing of the original experiments, J. Scales, J. Olberding, C.S. Easterling, S. Bloom, T. Kelsay and C. Brown for useful discussion, and P. Motta, D. Murphy and B. Gemmell for comments on the manuscript. J. Scales, C. Brown, C. Evelyn, J. Gavin Bradley, W. Smith, J. Spickler and J. Beshel shared localities, assisted in animal and data collection, or shared captive animals.

Author contributions

Conceptualization: M.K.O.D., S.M.D.; Methodology: M.K.O.D., S.M.D.; Formal analysis: M.K.O.D.; Investigation: M.K.O.D.; Data curation: M.K.O.D.; Writing - original draft: M.K.O.D.; Writing - review & editing: M.K.O.D., S.M.D.; Visualization: M.K.O.D.; Supervision: S.M.D.; Funding acquisition: M.K.O.D., S.M.D.

Funding

This study was supported by the National Science Foundation (IOS 1350929) grant to S.M.D. Funding for research, travel and equipment was provided in part by the University of South Florida Presidential Fellowship, the Porter Family Foundation University of South Florida, the Fern Garden Club of Odessa, and the Grant in Aid of Research from Highlands Biological Station.

Adams
,
D. C.
and
Nistri
,
A.
(
2010
).
Ontogenetic convergence and evolution of foot morphology in European cave salamanders (Family: Plethodontidae)
.
BMC Evol. Biol.
10
,
216
.
Adams
,
D. C.
,
Korneisel
,
D.
,
Young
,
M.
and
Nistri
,
A.
(
2017
).
Natural history constrains the macroevolution of foot morphology in European plethodontid salamanders
.
Am. Nat.
190
,
292
-
297
.
Alberch
,
P.
(
1981
).
Convergence and parallelism in foot morphology in the neotropical salamander genus Bolitoglossa: I. Function
.
Evolution
35
,
84
-
100
.
Autumn
,
K.
(
2006
).
Mechanisms of adhesion in geckos
.
Integr. Comp. Biol.
42
,
1081
-
1090
.
Autumn
,
K.
,
Sitti
,
M.
,
Liang
,
Y. A.
,
Peattie
,
A. M.
,
Hansen
,
W. R.
,
Sponberg
,
S.
,
Kenny
,
T. W.
,
Fearing
,
R.
,
Israelachvili
,
J. N.
and
Full
,
R. J.
(
2002
).
Evidence for van der Waals adhesion in gecko setae
.
Proc. Natl. Acad. Sci. USA
99
,
12252
-
12256
.
Baken
,
E. K.
and
Adams
,
D. C.
(
2019
).
Macroevolution of arboreality in salamanders
.
Ecol. Evol.
9
,
7005
-
7016
.
Barnes
,
W. J. P.
,
Oines
,
C.
and
Smith
,
J. M.
(
2006
).
Whole animal measurements of shear and adhesive forces in adult tree frogs: insights into underlying mechanisms of adhesion obtained from studying the effects of size and scale
.
J. Comp. Physiol. A Neuroethol. Sens. Neural Behav. Physiol.
192
,
1179
-
1191
.
Beckert
,
M.
,
Flammang
,
B. E.
and
Nadler
,
J. H.
(
2015
).
Remora fish suction pad attachment is enhanced by spinule friction
.
J. Exp. Biol.
218
,
3551
-
3558
.
Benjamini
,
Y.
and
Hochberg
,
Y.
(
1995
).
Controlling the false discovery rate: a practical and powerful approach to multiple testing
.
J. R. Stat. Soc. Ser. B
57
,
289
-
300
.
Betts
,
R. P.
,
Duckworth
,
T. A.
,
Austin
,
I. G.
,
Crocker
,
S. P.
and
Moore
,
S.
(
1980
).
Critical light reflection at a plastic/glass interface and its application to foot pressure measurements
.
J. Med. Eng. Technol.
4
,
136
-
142
.
Bonett
,
R. M.
and
Blair
,
A. L.
(
2017
).
Evidence for complex life cycle constraints on salamander body form diversification
.
Proc. Natl. Acad. Sci. USA
114
,
9936
-
9941
.
Bradley
,
J. G.
and
Eason
,
P. K.
(
2018
).
Predation risk and microhabitat selection by cave salamanders, Eurycea lucifuga (Rafinesque, 1822)
.
Behaviour
155
,
841
-
859
.
Bury
,
R. B.
(
1972
).
Small mammals and other prey in the diet of the pacific giant salamander (Dicamptodon ensatus)
.
Am. Midl. Nat.
87
,
524
.
Camp
,
C. D.
,
Wooten
,
J. A.
,
Jensen
,
J. B.
and
Bartek
,
D. F.
(
2013
).
Role of temperature in determining relative abundance in cave twilight zones by two species of lungless salamander (family Plethodontidae)
.
Can. J. Zool.
92
,
119
-
127
.
Cartmill
,
M.
(
1985
).
Climbing
. In
Functional Vertebrate Morphology
(ed.
M.
Hildebrand
,
D. M.
Bramble
,
K. F.
Liem
and
D. M.
Wake
), pp.
73
-
88
.
Harvard University Press
.
Christy
,
R. M.
and
Maie
,
T.
(
2019
).
Adhesive force and endurance during waterfall climbing in an amphidromous gobiid, Sicyopterus japonicus (Teleostei: Gobiidae): ontogenetic scaling of novel locomotor performance
.
Zoology
133
,
10
-
16
.
Crandell
,
K. E.
,
Herrel
,
A.
,
Sasa
,
M.
,
Losos
,
J. B.
and
Autumn
,
K.
(
2014
).
Stick or grip? Co-evolution of adhesive toepads and claws in Anolis lizards
.
Zoology
117
,
363
-
369
.
Crawford
,
J. A.
and
Peterman
,
W. E.
(
2013
).
Biomass and habitat partitioning of Desmognathus on wet rock faces in the Southern Appalachian Mountains
.
J. Herpetol.
47
,
580
-
584
.
Emerson
,
S. B.
and
Diehl
,
D.
(
1980
).
Toe pad morphology and mechanisms of sticking in frogs
.
Biol. J. Linn. Soc.
13
,
199
-
216
.
Endlein
,
T.
,
Barnes
,
W. J. P.
,
Samuel
,
D. S.
,
Crawford
,
N. A.
,
Biaw
,
A. B.
and
Grafe
,
U.
(
2013b
).
Sticking under wet conditions: the remarkable attachment abilities of the torrent frog, Staurois guttatus
.
PLoS ONE
8
,
e73810
.
Endlein
,
T.
,
Ji
,
A.
,
Samuel
,
D.
,
Yao
,
N.
,
Wang
,
Z.
,
Jon P. Barnes
,
W.
,
Federle
,
W.
,
Kappl
,
M.
and
Dai
,
Z.
(
2013a
).
Sticking like sticky tape: tree frogs use friction forces to enhance attachment on overhanging surfaces
.
J. R. Soc. Interface
10
,
20120838
.
Feder
,
M. E.
(
1983
).
Integrating the ecology and physiology of plethodontid salamanders
.
Herpetologica
39
,
291
-
310
.
Federle
,
W.
(
2002
).
An integrative study of insect adhesion: mechanics and wet adhesion of pretarsal pads in ants
.
Integr. Comp. Biol.
42
,
1100
-
1106
.
Federle
,
W.
(
2006
).
Why are so many adhesive pads hairy?
J. Exp. Biol.
209
,
2611
-
2621
.
Federle
,
W.
,
Barnes
,
W. J. P.
,
Baumgartner
,
W.
,
Drechsler
,
P.
and
Smith
,
J. M.
(
2006
).
Wet but not slippery: boundary friction in tree frog adhesive toe pads
.
J. R. Soc. Interface
3
,
689
-
697
.
Federle
,
W.
,
Baumgartner
,
W.
and
Hölldobler
,
B.
(
2003
).
Biomechanics of ant adhesive pads: frictional forces are rate- and temperature-dependent
.
J. Exp. Biol.
207
,
67
-
74
.
Forsman
,
E. D.
and
Swingle
,
J. K.
(
2007
).
Use of arboreal tree vole nests (Arborimus spp.) by amphibians
.
Herpetol. Conserv. Biol.
2
,
113
-
118
.
Fulcher
,
B. A.
and
Motta
,
P. J.
(
2006
).
Suction disk performance of echeneid fishes
.
Can. J. Zool.
84
,
42
-
50
.
Gorb
,
S. N.
(
2008
).
Biological attachment devices: exploring nature's diversity for biomimetics
.
Philos. Trans. R. Soc. A
366
,
1557
-
1574
.
Gorman
,
J.
and
Camp
,
C. L.
(
2006
).
A new cave species of salamander of the genus Hydromantes from California, with notes on habits and habitat
.
Copeia
1953
,
39
.
Green
,
D. M.
(
2008
).
Treefrog toe pads: comparative surface morphology using scanning electron microscopy
.
Can. J. Zool.
57
,
2033
-
2046
.
Green
,
D. M.
and
Alberch
,
P.
(
1981
).
Interdigital webbing and skin morphology in the neotropical salamander genus Bolitoglossa (Amphibia; Plethodontidae)
.
J. Morphol.
170
,
273
-
282
.
Green
,
D. M.
and
Barber
,
D. L.
(
2009
).
The ventral adhesive disc of the clingfish Gobiesox maeandricus: integumental structure and adhesive mechanisms
.
Can. J. Zool.
66
,
1610
-
1619
.
Green
,
D. M.
and
Simon
,
M. P.
(
1986
).
Digital microstructure in ecologically diverse sympatric microhylid frogs, genera Cophixalus and Sphenophryne (Amphibia, Anura), from Papua-New Guinea
.
Aust. J. Zool.
34
,
109
-
117
.
Hanna
,
G.
and
Barnes
,
W. J. P.
(
1991
).
Adhesion and detachment of the toe pads of tree frogs
.
J. Exp. Biol.
155
,
103
-
125
.
Huheey
,
J. E.
and
Brandon
,
R. A.
(
1973
).
Rock-face populations of the mountain salamander, Desmognathus ochrophaeus, in North Carolina
.
Ecol. Monogr.
43
,
59
-
77
.
Jaeger
,
R. G.
(
1978
).
Plant climbing by salamanders: periodic availability of plant-dwelling prey
.
Copeia
1978
,
686
-
691
.
Jaekel
,
M.
and
Wake
,
D. B.
(
2007
).
Developmental processes underlying the evolution of a derived foot morphology in salamanders
.
Proc. Natl. Acad. Sci. USA
104
,
20437
-
20442
.
Labonte
,
D.
and
Federle
,
W.
(
2015
).
Scaling and biomechanics of surface attachment in climbing animals
.
Philos. Trans. R. Soc. B Biol. Sci.
370
:
20140027
.
Labonte
,
D.
,
Clemente
,
C. J.
,
Dittrich
,
A.
,
Kuo
,
C.-Y.
,
Crosby
,
A. J.
,
Irschick
,
D. J.
and
Federle
,
W.
(
2016
).
Extreme positive allometry of animal adhesive pads and the size limits of adhesion-based climbing
.
Proc. Natl. Acad. Sci. USA
113
,
1297
-
1302
.
Legros
,
D. L.
(
2013
).
Plant climbing in the northern two-lined salamander, Eurycea bislineata, in Algonquin provincial park, Ontario
.
Can. Field-Naturalist
127
,
67
-
69
.
Lunghi
,
E.
,
Manenti
,
R.
and
Ficetola
,
G. F.
(
2017
).
Cave features, seasonality and subterranean distribution of non-obligate cave dwellers
.
PeerJ
5
,
e3169
.
Lunghi
,
E.
,
Manenti
,
R.
,
Manca
,
S.
,
Mulargia
,
M.
,
Pennati
,
R.
and
Ficetola
,
G. F.
(
2014
).
Nesting of cave salamanders (Hydromantes flavus and H. italicus) in natural environments
.
Salamandra
50
,
105
-
109
.
Lunghi
,
E.
,
Murgia
,
R.
,
De Falco
,
G.
,
Buschettu
,
S.
,
Mulas
,
C.
,
Mulargia
,
M.
,
Canedoli
,
C.
,
Manenti
,
R.
and
Ficetola
,
G. F.
(
2015
).
First data on nesting ecology and behaviour in the Imperial cave salamander Hydromantes imperialis
.
North. West. J. Zool.
11
,
324
-
330
.
Maie
,
T.
,
Schoenfuss
,
H. L.
and
Blob
,
R. W.
(
2012
).
Performance and scaling of a novel locomotor structure: adhesive capacity of climbing gobiid fishes
.
J. Exp. Biol.
215
,
3925
-
3936
.
McEntire
,
K. D.
(
2016
).
Arboreal ecology of Plethodontidae: a review
.
Copeia
104
,
124
-
131
.
Myer
,
C. W.
(
1958
).
Notes on the eggs and larave of Eurycea lucifuga Rafinesque
.
Q. J. Florida Acad. Sci.
21
,
125
-
130
.
Nachtigall
,
W.
(
1974
).
Biological Mechanisms of Attachment
.
Springer-Verlag
.
Paradis
,
E.
and
Schliep
,
K.
(
2018
).
ape 5.0: an environment for modern phylogenetics and evolutionary analyses in R
.
Bioinformatics
35
,
526
-
528
.
Riskin
,
D. K.
and
Fenton
,
M. B.
(
2011
).
Sticking ability in Spix's disk-winged bat, Thyroptera tricolor (Microchiroptera: Thyropteridae)
.
Can. J. Zool.
79
,
2261
-
2267
.
Riskin
,
D. K.
and
Racey
,
P. A.
(
2010
).
How do sucker-footed bats hold on, and why do they roost head-up?
Biol. J. Linn. Soc.
99
,
233
-
240
.
Russell
,
A. P.
and
Higham
,
T. E.
(
2009
).
A new angle on clinging in geckos: incline, not substrate, triggers the deployment of the adhesive system
.
Proc. R. Soc. B Biol. Sci.
276
,
3705
-
3709
.
Salvidio
,
S.
,
Crovetto
,
F.
and
Adams
,
D. C.
(
2015
).
Potential rapid evolution of foot morphology in Italian plethodontid salamanders (Hydromantes strinatii) following the colonization of an artificial cave
.
J. Evol. Biol.
28
,
1403
-
1409
.
Schmidt-Nielsen
,
K.
(
1975
).
Scaling in biology: the consequences of size
.
J. Exp. Zool.
194
,
287
-
307
.
Schneider
,
C. A.
,
Rasband
,
W. S.
and
Eliceiri
,
K. W.
(
2012
).
NIH Image to ImageJ: 25 years of image analysis
.
Nat. Methods
9
,
671
-
675
.
Smith
,
A. M.
(
1991
).
Negative pressure generated by octopus suckers: a study of the tensile strength of water in nature
.
J. Exp. Biol.
157
,
257
-
271
.
Spickler
,
J. C.
,
Sillett
,
S. C.
,
Marks
,
S. B.
and
Welsh
,
H. H.
Jr.
(
2006
).
Evidence of a new niche for a North American salamander: Aneides vagrans residing in the canopy of old-growth redwood forest
.
Herpetol. Conserv. Biol.
1
,
16
-
26
.
Stark
,
A. Y.
and
Yanoviak
,
S. P.
(
2018
).
Adhesion and running speed of a tropical arboreal ant (Cephalotes atratus) on wet substrates
.
R. Soc. Open Sci.
5
,
181540
.
Stark
,
A. Y.
,
Ohlemacher
,
J.
,
Knight
,
A.
and
Niewiarowski
,
P. H.
(
2015
).
Run don't walk: locomotor performance of geckos on wet substrates
.
J. Exp. Biol
218
,
2435
-
2441
.
Stefan
,
M.
(
1874
).
Experiments on apparent adhesion
.
Philos. Mag. J. Sci.
47
,
465
-
466
.
Tian
,
Y.
,
Pesika
,
N.
,
Zeng
,
H.
,
Rosenberg
,
K.
,
Zhao
,
B.
,
McGuiggan
,
P.
,
Autumn
,
K.
and
Israelachvili
,
J.
(
2006
).
Adhesion and friction in gecko toe attachment and detachment
.
Proc. Natl. Acad. Sci. USA
103
,
19320
-
19325
.
Vogel
,
S.
(
2003
).
Comparative Biomechanics
.
Princeton
:
Princeton University Press
.
Wainwright
,
D. K.
,
Kleinteich
,
T.
,
Kleinteich
,
A.
,
Gorb
,
S. N.
and
Summers
,
A. P.
(
2013
).
Stick tight: suction adhesion on irregular surfaces in the northern clingfish
.
Biol. Lett.
9
:
20130234
.
Wake
,
D. B.
(
1987
).
Adaptive radiation of salamanders in Middle American cloud forests
.
Ann. MIssouri Bot. Gard.
74
,
242
-
264
.
Wake
,
D. B.
(
2014
).
The enigmatic history of the European, Asian and American plethodontid salamanders
.
Amphib. Reptil.
34
,
323
-
336
.
Waldron
,
J. L.
and
Humphries
,
W. J.
(
2005
).
Arboreal habitat use by the green salamander, Aneides aeneus, in South Carolina
.
J. Herpetol.
39
,
486
-
492
.
Wickham
,
H.
(
2009
).
ggplot2: Elegant Graphics for Data Analysis
.
New York
:
Springer-Verlag
.
Zani
,
P. A.
(
2000
).
The comparative evolution of lizard claw and toe morphology and clinging performance
.
J. Evol. Biol.
13
,
316
-
325
.

Competing interests

The authors declare no competing or financial interests.

Supplementary information