Terrestrial Force Production by the Limbs of a Semi-Aquatic Salamander Provides Insight Into The

Home , Appendage, Aquatic locomotion, Digit (anatomy), Hindlimb, Largemouth bass

bioRxiv preprint doi: https://doi.org/10.1101/2021.05.01.442256; this version posted May 2, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

1 Terrestrial force production by the limbs of a semi-aquatic salamander provides insight into the

2 evolution of terrestrial locomotor mechanics

4 Sandy M. Kawano1*, and Richard W. Blob2

6 1Department of Biological Sciences, The George Washington University, Washington, D.C.

7 20052, U.S.A.

8 2Department of Biological Sciences, Clemson University, Clemson, SC 29634, U.S.A.

10 *Corresponding author. E-mail: [email protected]

12 Running title: Kinetics of fins and limbs on land

13 Keywords: biomechanics, ground reaction force, terrestrial locomotion, salamander, fish, water-

14 to-land transition

16 # of tables: 7 # of figures: 3

17 Total word count (Intro, Results, Discussion, Acknowledgements, Funding, Conclusion, Figure

18 legends): 1107 + 4012 + 884 = 6003

20 Summary statement (15-30 words): Semi-aquatic salamanders had limb mechanics that were

21 intermediate in magnitude yet steadier than the appendages of terrestrial salamanders and

22 semi-aquatic fish, providing a framework to model semi-aquatic early tetrapods.

Page 1 of 34

23 ABSTRACT

24 Amphibious fishes and salamanders are valuable functional analogs for vertebrates that

25 spanned the water-to-land transition. However, investigations of walking mechanics have

26 focused on terrestrial salamanders and, thus, may better reflect the capabilities of stem

27 tetrapods that were already terrestrial. The earliest tetrapods were aquatic, so salamanders that

28 are not primarily terrestrial may yield more appropriate data for modelling the incipient stages of

29 terrestrial locomotion. In the present study, locomotor biomechanics were quantified from semi-

30 aquatic Pleurodeles waltl, a salamander that spends most of its adult life in water, and then

31 compared to a primarily terrestrial salamander (Ambystoma tigrinum) and semi-aquatic fish

32 (Periophthalmus barbarus) to evaluate whether walking mechanics show greater similarity

33 between species with ecological versus phylogenetic similarities. Ground reaction forces

34 (GRFs) from individual limbs or fins indicated that the pectoral appendages of each taxon had

35 distinct patterns of force production, but hind limb forces were comparable between the

36 salamanders. The rate of force development (‘yank') was sometimes slower in P. waltl but

37 generally comparable between the three species. Finally, medial inclination of the GRF in P.

38 waltl was intermediate between semi-aquatic fish and terrestrial salamanders, potentially

39 elevating bone stresses among more aquatic taxa as they move on land. These data provide a

40 framework for modelling stem tetrapods using an earlier stage of quadrupedal locomotion that

41 was powered primarily by the hind limbs (i.e., “rear-wheel drive”), and reveal mechanisms for

42 appendages to generate propulsion in three locomotor strategies that are presumed to have

43 occurred across the water-to-land transition in vertebrate evolution.

Page 2 of 34

44 INTRODUCTION

45 The evolutionary invasion of land was a seminal event in vertebrate history that has received

46 intense study, with integrative approaches filling major gaps in our understanding of how this

47 transition transpired (Ashley-Ross et al. 2013). Skeletal remains and trackways in the fossil

48 record can provide some of the most direct evidence about such historical events, yet also

49 present challenges given the limits to which fossils preserve behavior and functional

50 performance (see (Gatesy et al., 1999; Maidment et al., 2014); and references therein).

51 Consequently, extant amphibious fishes, amphibians, and reptiles have been used as functional

52 models to infer the biology of extinct tetrapodomorphs (tetrapods and tetrapod-like

53 sarcopterygian fishes) (Nyakatura et al., 2019, 2014; Pierce et al., 2013), with extant taxa

54 representing alternative strategies for invading land and potentially serving as models for

55 different functional stages as vertebrates became terrestrial. Investigations of extant taxa

56 exhibiting morphological and/or behavioral traits that are consistent with those of extinct

57 tetrapodomorphs offer particularly intriguing opportunities to evaluate how tetrapods were able

58 to leave the water’s edge (Pierce et al., 2013).

59 In considerations of locomotor performance during the invasion of land, salamanders

60 are often used as functional analogues for early tetrapods since they move between water and

61 land (Karakasiliotis et al., 2012), and exhibit a relatively generalized tetrapod Bauplan that has

62 not changed substantially for at least 150 million years (Gao and Shubin, 2001) and possibly as

63 long as 230 million years (Schoch et al., 2020). Previous studies used living salamanders to

64 gain insights into the biomechanics and muscle physiology of walking underwater (Ashley-Ross

65 et al., 2009; Azizi and Horton, 2004; Deban and Schilling, 2009; Frolich and Biewener, 1992)

Page 3 of 34

66 and on land (Ashley-Ross et al., 2009; Brand, 1996; Deban and Schilling, 2009; Delvolvé et al.,

67 1997; Frolich and Biewener, 1992; Kawano and Blob, 2013; Pierce et al., 2020; Sheffield and

68 Blob, 2011), transitioning between water and land (Ashley-Ross and Bechtel, 2004), and

69 assessing how bone histology relates to ecological habits (Canoville and Laurin, 2009; Laurin et

70 al., 2004). Given the greater effect of gravitational loads on the musculoskeletal system in

71 terrestrial environments (Young et al., 2017), one of the most fundamental requirements for

72 moving on land is the ability to support body weight for maintaining posture and physical forces

73 for locomotion. Evaluations of the weight-bearing capabilities in the limbs of early tetrapods

74 have been approached through measurements of ground reaction forces (GRFs) experienced

75 by the tiger salamander, Ambystoma tigrinum, which demonstrated that the hind limbs

76 supported a similar proportion of body weight as the forelimbs but had a greater contribution to

77 acceleration (Kawano and Blob, 2013). To what extent might these results relate to the

78 locomotor function of early tetrapods? Fossil evidence suggests that some of the first tetrapods,

79 such as Acanthostega, were still aquatic (Coates, 1996), and other early tetrapods, such as

80 Ichthyostega, may have had only limited terrestrial capabilities (Pierce et al., 2012). In contrast,

81 A. tigrinum are found in various terrestrial habitats, ranging from conifer forests to deserts, and

82 adults rarely spend extended durations in water for reasons other than reproduction (Petranka,

83 1998). As such, they may not provide an ideal model for the initial invaders of land, in which

84 terrestrial capacity may not have been fully acquired. How might limb function differ for a

85 species that spends most of its time in water?

86 Because salamanders have a diverse range of habitat preferences and life histories

87 (Wake, 2009), they provide an opportunity to model different evolutionary stages in the adoption

Page 4 of 34

88 of terrestrial habits. Examinations of amphibious species with greater aquatic tendencies could

89 yield substantial insights into the limb function of earlier stem tetrapods with digit-bearing limbs.

90 In this study, we compared GRFs of individual limbs produced by semi-aquatic salamanders

91 (Spanish ribbed newts, Pleurodeles waltl Michahelles 1830; see Electronic Supplementary

92 Materials for details), to published data on terrestrial tiger salamanders, Ambystoma tigrinum

93 Green 1825, and semi-aquatic African mudskippers, Periophthalmus barbarus (Linnaeus 1766)

94 (Kawano and Blob, 2013). We also calculated new measures of yank, the rate of force

95 development (sensu Lin et al., 2019), to evaluate how the rate of force production might differ

96 between these species. Quantifying the dynamics of force-time measures contributes valuable

97 information to explain kinematic and muscle activation patterns (Lin et al., 2019). Our aim in

98 these comparisons was to examine extant taxa that model important stages during the transition

99 to land to quantify the functional differences between fins and limbs for terrestrial locomotion.

100 During aquatic locomotion, dominant force production by the rear appendage (“rear-

101 wheel drive”) may have appeared as early as in sarcopterygian fishes (King et al., 2011).

102 However, terrestrial propulsion in stem tetrapods may have been dominated by the forelimb

103 (‘front-wheel drive’), with the transition to hind limb dominance (‘rear-wheel drive’) only occurring

104 among lineages that diverged later, as the hind limbs assumed a more important locomotor role

105 (Boisvert, 2005). The ‘front-wheel drive’ strategy proposed for the sarcopterygian fish

106 Panderichthys (Boisvert, 2005) and the early tetrapod Ichthyostega (Pierce et al., 2012) might

107 be appropriately modeled by the locomotor behaviors of extant mudskipper fishes (McInroe et

108 al., 2016), which use synchronous movements of the pectoral fins to ‘crutch’ over land while

109 keeping the body axis relatively straight (Kawano and Blob, 2013; Pace and Gibb, 2014). On the

Page 5 of 34

110 other hand, the ‘rear-wheel drive’ of more crownward early tetrapods might be better modeled

111 using terrestrial salamanders, such as A. tigrinum (Pierce et al., 2013). Our functional models

112 also show distinct GRF patterns between these groups, with those produced by the pectoral fins

113 of mudskippers directed more medially (~17°) than those produced by the limbs of terrestrial

114 salamanders, or almost any other terrestrial tetrapod (generally less than 11°). Such differences

115 could have substantial implications for the stresses experienced by the limb bones by increasing

116 moment arms due to bending, which could make bones more vulnerable to fracture (Kawano

117 and Blob, 2013).

118 Our data from semi-aquatic P. waltl can give new insight into the functional changes that

119 occurred during the early evolution of terrestrial locomotion. Simply by having limbs, locomotor

120 force production by P. waltl may be similar to that of terrestrial salamanders such as adult A.

121 tigrinum. However, the habitual use of the limbs for aquatic locomotion by adult P. waltl might

122 lead to kinetic similarities with other semi-aquatic taxa, such as mudskipper fishes. These

123 comparisons carry broader implications for generating hypotheses about the emergence of

124 functional disparity between locomotor structures, such as whether changes in functional

125 performance are coupled across major structural changes (e.g., transition from fin to limb) or

126 through gradual steps related to loading regimes that are potentially decoupled from structural

127 changes (e.g., the transition from aquatic to terrestrial habitats, regardless of which locomotor

128 structure is used).

129

130 MATERIALS AND METHODS

131 Animals

Page 6 of 34

132 Five adult P. waltl (body mass: 16.60 ± 0.40 g; snout-vent length: 0.083 ± 0.001 m; total length:

133 0.186 ± 0.003 m) were obtained from a commercial vendor. All values represent means ± 1 S.E.

134 Animals were individually housed in glass aquaria that were aerated with sponge filters, kept on

135 a 12h:12h light:dark cycle, and fed every one - two days on a diet of defrosted bloodworms and

136 krill. Prior to testing, animals were starved for two days to reduce the effects of satiation on

137 locomotor performance (Sass and Motta, 2002).

138

139 Collection of three-dimensional (3D) ground reaction forces

140 Experimental procedures from a previous study on the GRFs of tiger salamanders and

141 mudskipper fish (Kawano and Blob, 2013) were replicated in the present study (see Electronic

142 Supplementary Materials) to obtain forelimb (n = 50) and hind limb (n = 49) GRFs from P. waltl

143 (Fig. S1). Custom code to replicate these analyses is available through GitHub:

144 https://github.com/MorphoFun/kraken. The focal taxa examined in the present study represent

145 models for distinct functional stages during the evolution of terrestrial locomotion: front-wheel

146 drive in a semi-aquatic vertebrate with limited capabilities of the pelvic appendages (semi-

147 aquatic mudskipper fish), a semi-aquatic vertebrate with a generalized tetrapod Bauplan (semi-

148 aquatic P. waltl salamander), and rear-wheel drive in a terrestrial vertebrate with a generalized

149 tetrapod Bauplan (terrestrial A. tigrinum salamander). GRFs in the vertical, mediolateral, and

150 anteroposterior directions were digitally filtered with a custom low-pass, zero phase second

151 order Butterworth filter, and then interpolated to 101 points (0-100% of stance at 1%

152 increments) using a cubic spline with the signal R package (Signal developers, 2013). The

Page 7 of 34

153 stance phase was defined to begin at the first video frame when the entire foot was flat against

154 the ground, and to end at the frame penultimate to the beginning of the swing phase.

155 Interspecific and intraspecific comparisons were made across the GRFs of fins and

156 limbs. First, GRFs from the forelimbs and hind limbs of P. waltl were compared to published

157 data from the pectoral and pelvic appendages of mudskipper fish and tiger salamanders

158 (Kawano and Blob, 2013) to assess whether limb kinetics in semi-aquatic salamanders are

159 more similar to those of mudskipper fins or terrestrial salamander limbs. Second, GRFs were

160 compared between the forelimbs and hind limbs of P. waltl to determine whether a model for a

161 semi-aquatic early tetrapod was forelimb-driven or hind limb-driven. Third, GRF data from these

162 three species were used to conduct new analyses of yank (sensu Lin et al., 2019) to quantify

163 the rate of change in the GRFs produced by fins vs. limbs. Linear mixed effects models (LMMs)

164 were used to compare magnitudes and angles of orientation of GRFs when forces were

165 maximal (“peak net GRF”), and the range of yank values recorded between the species and

166 appendages (see “Statistical analyses”).

167

168 Statistical analyses

169 Linear mixed effects models with random intercepts via the lme4 package (Bates et al.,

170 2015) were used to compare the biomechanical variables of interest while accounting for the

171 non-independence of multiple trials sampled per individual, and fit with Restricted Maximum

172 Likelihood (REML) to produce unbiased estimates. LMMs were conducted to account for the

173 non-independence caused by trials being nested within individuals and the unequal sample

174 sizes across individuals. LMMs also have the advantage that they do not make assumptions

Page 8 of 34

175 about balanced designs, homogeneity of variance, etc., so the model assumptions only include

176 random sampling and normality of the residuals (reviewed in (Smith, 2017). Evaluation of

177 Quantile-Quantile (Q-Q) plots indicated that there were no major deviations from the reference

178 line (Electronic Supplementary Material), suggesting that all variables reasonably met the

179 assumption of normality. However, violations of the model assumptions would not have affected

180 our analyses given our focus on fixed effects, which tend to be unbiased even if assumptions

181 are violated (LeBeau et al., 2018). “Individual” was treated as a random effect, and “appendage

182 type” (pectoral vs. pelvic) or “species” was treated as a fixed effect for intra- and interspecific

183 comparisons, respectively.

184 Large sample sizes are typically required to produce accurate estimates of a random

185 slopes model (reviewed in Harrison et al. 2018) so we assumed that the slopes were similar

186 across the individuals within a species based on our smaller sample size. Assuming common

187 slopes in a LMM can increase the potential for false positives and false negatives when

188 calculating p-values for null hypothesis testing (reviewed in Harrison et al. 2018); however, the

189 present study focused on effect sizes (i.e., the magnitude of the effects or variables) rather than

190 the p-values, so these outcomes do not detract from our analyses. Given that one of the

191 objectives of this study is to quantify the mean differences (i.e., point estimates) and range of

192 variation (i.e., interval estimates) in GRF production between and within salamanders to

193 parameterize computational models of musculoskeletal function, our emphasis on calculating

194 effect sizes is appropriate (Cohen, 1990) due to the limited information that often results from

195 traditional methods of null hypothesis testing (Cumming, 2014).

Page 9 of 34

196 To calculate the point estimates (i.e., mean values) and interval estimates (i.e., standard

197 errors) of the fixed effects, we used the emmeans::emmeans() function (Lenth, 2019). The

198 coefficient of determination (percent of the variation explained by each model) was calculated

2 2 2 199 as Nagakawa’s marginal R (R LMM(m); variation due to the fixed effects only) and conditional R

2 200 (R LMM(c); variation due to the fixed and random effects) using the performance::r2_nagakawa()

201 function (Lüdecke et al., 2019). Dynamic changes in force magnitudes and yank were plotted by

202 pooling individual trials for a given species, but results from the LMMs are reported in the tables

203 to account for repeated trials within individuals. All analyses were conducted in R version 3.6.1

204 (R Core Team, 2019) and RStudio version 1.2.1335 (RStudio Team, 2015).

205

206 RESULTS

207 Force production between the pectoral appendages of mudskippers and salamanders

208 The forelimbs of P. waltl exhibited kinetics that were generally intermediate between the

209 pectoral fins of P. barbarus fish and the forelimbs of A. tigrinum salamanders (Fig. 1). Force

210 production during terrestrial locomotion generally approximated a bell-shape curve that was

211 skewed to the right in P. barbarus and A. tigrinum, but relatively broad and shallow in P. waltl for

212 the net GRF and the vertical component of the GRF (Fig. 1a-b). The mediolateral component of

213 the GRF was directed medially during stance for all three species, yet had a wider range of

214 values in P. barbarus (Fig. 1c). For the anteroposterior component of the GRF, forelimb kinetics

215 in P. waltl were more similar to terrestrial A. tigrinum salamanders than to semi-aquatic P.

216 barbarus fish. The pectoral fin is the primary propulsor during crutching for P. barbarus and its

217 GRF was oriented anteriorly for all of stance (indicating an acceleratory role), whereas the

Page 10 of 34

218 forelimbs of A. tigrinum and P. waltl showed primarily posteriorly directed GRFs (deceleratory

219 role), with only one or two periods in which the GRFs became slightly anterior (Fig. 1d, f). The

220 GRF was angled medially in the horizontal plane throughout stance for P. barbarus and A.

221 tigrinum, and for almost all of stance for P. waltl except for a slight lateral orientation starting

222 around 90% of stance (Fig. 1e). These findings illustrate how different aspects of force

223 production by the forelimbs of P. waltl share similarities, yet also distinct differences (e.g.,

224 mediolateral GRFs), with those of semi-aquatic fins and terrestrial limbs.

225 GRF parameters at the peak net GRF also differed between the pectoral appendages for

226 the three groups examined. The timing of the peak net GRF occurred at 60.4 ± 2.86 % of stance

227 for A. tigrinum, 56.0 ± 2.86 % for P. barbarus, and 46.9 ± 2.93 % for P. waltl (mean ± SE;

2 2 228 R LMM(m) = 0.226, R LMM(c) = 0.462). The magnitude of the peak net GRF was comparable

229 between P. barbarus and A. tigrinum at approximately 44% and 46% of body weight,

230 respectively, but was roughly 40% of body weight and had a wider range of values in P. waltl

231 (Table 1, Fig. 2a). Similarly, the vertical component at the peak net GRF was comparable for A.

232 tigrinum and P. barbarus yet lower and broader for P. barbarus (Fig. 2b). The mediolateral

233 component of the peak net GRF was similar between the two salamanders (Fig. 2c), with values

234 for P. barbarus being 1.5 – 2x more medial than the salamanders. The anteroposterior

235 component was also more similar between the salamanders at the peak net GRF, with values

236 being anterior (acceleratory) for P. barbarus but posterior (deceleratory) for both salamanders.

237 The angle of the GRF was medial for all pectoral appendages at peak net GRF, but values for

238 P. waltl were intermediate (~14°) between P. barbarus (~17°) and A. tigrinum (~9°). These

239 comparisons further demonstrate that kinetics at the time of peak net GRF for P. waltl share

Page 11 of 34

240 similarities to semi-aquatic fins and terrestrial limbs, with some features intermediate between

241 the other two species.

242

243 Force production between salamander forelimbs and hind limbs

244 Kinetic profiles were similar in shape between the hind limbs of semi-aquatic and

245 terrestrial salamanders (Fig. 3), although there were differences in the magnitudes of the

246 curves. Semi-aquatic and terrestrial hind limbs had a peak net GRF occurring around 30% of

247 stance; however, the net GRF and the vertical component at peak GRF for P. waltl were more

248 similar to those of P. waltl forelimbs (at about 30% of body weight) than to the hind limbs of A.

249 tigrinum (~50% of body weight) (Tables 1-2, Fig. 4a - b). The GRFs were directed medially for

250 all limbs (~9-17°) but were 1.5x and 2x more medial in P. waltl than A. tigrinum for the forelimbs

251 and hind limbs, respectively (Table 2, Fig. 4e). The forelimbs of both salamander species had a

252 net deceleratory role (Table 1), whereas the hind limbs had a net acceleratory role (Table 2),

253 with the hind limbs of A. tigrinum showing a mean value ~1.5x higher than those of P. waltl at

254 the time of peak net GRF (Fig. 4d). However, the angle of the GRF in the anteroposterior

255 direction was comparable between salamander hind limbs, inclined 16 – 17° anteriorly (Fig. 4f).

256 These findings demonstrate that hind limb kinetics are not necessarily equivalent between

257 salamander taxa with differing degrees of terrestriality, and that locomotor biomechanics differ

258 between the forelimbs and hind limbs during terrestrial walking in both species.

259

260 Rate of change in GRF production

Page 12 of 34

261 Yank values for the pectoral appendages tended to be similar between the three

262 species, but GRFs produced by the forelimbs of P. waltl generally reached their peak more

263 slowly (as indicated by smaller absolute magnitudes of yank) and had a narrower range

264 compared to the semi-aquatic fish and terrestrial salamanders (Table 3). The shapes of the

265 curves for yank differed between P. waltl and the other two species. For P. barbarus and A.

266 tigrinum, yank for the net GRF and vertical component of the GRF (Fig. 5a – b) approximated

267 sine waves that started close to zero at the beginning of stance, increased to positive yank

268 values before decreasing to an inflection point around 60% of stance, and then decreased to

269 negative yank values before returning to the zero baseline. In contrast, net GRF and vertical

270 component of the GRF for P. waltl reached an inflection point at ~50% of stance, and then

271 approached zero again at ~80% of stance, at a time when A. tigrinum and P. barbarus were

272 both reaching their minimum yank values (i.e., peak in negative yank). The anteroposterior

273 component was comparable between the two salamander forelimbs (~0.006 - 0.008 BW for

274 each percent of stance), which increased faster than the pectoral fin of P. barbarus (0.004 ±

275 0.001 BW for each percent of stance). These findings demonstrate that rates of GRF production

276 in the pectoral appendages of the semi-aquatic salamander, P. waltl, were generally similar in

277 magnitude but shifted at different times during stance compared to the other two species.

278 Similarities between A. tigrinum and P. waltl were more evident in the shapes of the

279 yank curves for the hind limbs (Fig. 6). The curves for the net GRF (Fig. 6a) and the vertical

280 component of the GRF (Fig. 6b) approximated sine waves that reached an inflection point at

281 around 25% of stance. The peak for positive yank appears almost 2x higher in A. tigrinum

282 compared to P. waltl (Fig. 6, Table 3), but the confidence intervals overlapped for the fixed

Page 13 of 34

283 effects. The minima for negative yank in the vertical component and net GRF were shallower in

284 P. waltl compared to A. tigrinum. As a result, rates of change in GRF production for the hind

285 limb were slower for the semi-aquatic P. waltl salamander compared to the terrestrial A. tigrinum

286 salamander for the vertical component of the GRF and net GRF, but were generally similar

287 otherwise.

288

289 DISCUSSION

290 Distinctions in limb mechanics for semi-aquatic salamanders

291 Pleurodeles waltl is a semi-aquatic salamander that spends most of its life history

292 underwater. Similar to terrestrial salamanders, the predominant acceleratory forces in this semi-

293 aquatic salamander are produced by the hind limb, signifying rear-wheel drive. The orientation

294 of the GRF in the hind limb was about 17° in the anterior direction for both salamanders, but

295 was more medial in P. waltl (~17°) compared to A. tigrinum (~9°). Similarly, the orientation of the

296 GRF for the pectoral appendages was more medial in the semi-aquatic P. waltl and P. barbarus

297 compared to the primarily terrestrial A. tigrinum (Table 1). These differences in orientation may

298 affect how loads are applied to the appendages, potentially resulting in the lower vertical

299 component and net magnitude of the GRF in P. waltl forelimbs and hind limbs.

300 Although the forelimbs of both salamanders had a net deceleratory role as evidenced by

301 the negative value for the anteroposterior component of the GRF, this value was larger in

302 magnitude throughout almost all of stance in P. waltl, whereas A. tigrinum exhibited two periods

303 of slightly net positive values (Fig. 1, Table 1). Comparisons between the limbs of P. waltl

304 demonstrate that the hind limbs are the primary propulsors, whereas the forelimbs may be

Page 14 of 34

305 operating as ‘struts’ that act as pivot points to rotate the body. This suggests that semi-aquatic

306 and terrestrial salamanders may be using different mechanisms for propulsion between the

307 forelimbs and hind limbs.

308 McElroy and colleagues (2014) found that differential limb function occurred in diverse

309 ways across sprawling quadrupeds (e.g., diverse lizard taxa), and that ‘rear-wheel drive’ was

310 one of the only unifying themes among the seven species compared. Instead, differences in the

311 locomotor biomechanics of limbs in running quadrupedal lizards were based on interspecific

312 differences in limb morphology, with a more medial orientation of the GRFs produced by the

313 hind limb in lizard species that had disproportionately longer hind limbs than the forelimbs

314 (McElroy et al., 2014). Differential limb function in sprawling quadrupeds is proposed to increase

315 maneuverability while maintaining stability (McElroy et al., 2014). Although this idea has been

316 proposed in the context of running lizards that may experience instability during rapid turns,

317 similar arguments apply to the locomotion of crabs, enabling them to avoid over-turning

318 moments from hydrodynamic forces (Chen et al., 2006). Future studies could factor

319 environmental variation into comparative analyses of differential limb function to examine how it

320 might evolve in the context of selection pressures presented by different environments.

321 Disparities in function between fore- and hind appendages, as well as other differences

322 between biomechanical profiles for aquatic and terrestrial species, could relate to the different

323 demands imposed by the primary environments in which the appendages of these taxa function.

324 For example, point estimates for the medial orientation of the peak GRF in semi-aquatic P. waltl

325 limbs (14-16°) falls between that of P. barbarus fins (17°) and most previously evaluated

326 tetrapod limbs (<11°), including primarily terrestrial A. tigrinum salamanders (Fig. 7, Tables 1 -

Page 15 of 34

327 2). A shift to a GRF directed less medially could reduce joint moments and, thus, the stresses

328 experienced by the appendicular bones during terrestrial locomotion (Kawano and Blob, 2013).

329 However, the greater medial orientation of the GRF in semi-aquatic P. waltl likely relates to the

330 greater lateral spread of their distal limb segments compared to terrestrial taxa, so that the feet

331 are placed lateral to the elbow or knee joint during stance (Fig. S1A), rather than directly below

332 these joints (as in terrestrial salamanders: Fig. 1A, B in Kawano and Blob, 2013). A recent

333 comparison of terrestrial locomotion in salamanders found that walking mechanics and muscle

334 activation patterns were generally similar among species, but that some interspecific differences

335 emerged (Pierce et al., 2020). For instance, maximum extensions of the wrist (160°) and ankle

336 (159°) were higher in P. waltl than A. tigrinum (150° and 150°, respectively) despite its elbow

337 and knee having greater (55° vs. 30°) or equal (55°) total maximum extensions, respectively

338 (Pierce et al., 2020). Given that this more pronounced sprawling limb posture is also found in

339 the mudskipper fish, a more medial orientation of the GRF also might be found in other semi-

340 aquatic taxa. For example, Ashley-Ross and Bechtel (Ashley-Ross and Bechtel, 2004) found

341 that Taricha torosa newts also had greater lateral spread of the limbs, caused by more flattened

342 angles of the limb elements, when walking in water than on land.

343 The observation that P. waltl maintains a more extended sprawling posture on land may

344 mean that some species have greater stereotypy (Wainwright et al., 2008) in their locomotor

345 biomechanics compared to others, with the degree of flexibility influenced by how often habitat

346 transitions occur during a season. The broadening of the gait pattern that would result from such

347 lateral foot placement might convey additional stability against currents or other flows in aquatic

348 habitats (Martinez et al., 1998) by reducing pitching and rolling (Chen et al., 2006). However,

Page 16 of 34

349 when on land, species that tend to be more aquatic may not be able to adjust to the more

350 upright orientations of distal limb segments that are seen in terrestrial taxa (Kawano and Blob,

351 2013). Producing more acute limb angles could facilitate elevating the body off the ground, and

352 shift the bone loading regime to reduce bending and increase compression (Ashley-Ross and

353 Bechtel, 2004; Kawano and Blob, 2013). Lateral spread of the distal appendage may also

354 contribute to the high medial orientation of the GRF in mudskippers (Fig. 1C in Kawano and

355 Blob, 2013), but furthermore could prevent rolling of the body axis during terrestrial crutching,

356 given the lack of extended posterior appendages in this taxon. Thus, shifts to less sprawling

357 limb posture could have major biomechanical consequences that could facilitate terrestrial

358 locomotion.

359 Similarities in the overall shapes of the GRF and yank produced by individual hind limbs

360 suggest that rates of GRF production are consistent across hind limbs of semi-aquatic and

361 terrestrial salamander species, but the different peaks observed in these curves indicate that the

362 maxima differ between them. The similarities in force production between semi-aquatic and

363 terrestrial hind limbs could indicate that the use of the hind limbs as a primary propulsor

364 imposes strong selection on kinetics for acceleration, but that other aspects, such as

365 maneuverability and body support, can be altered by changing the GRFs and yank. For

366 example, the smaller absolute value for negative yank of the vertical component of the GRF and

367 net GRF for the hind limbs in P. waltl, compared to A. tigrinum, suggest that there may be

368 steadier loading regimes in the former. Loading the appendages in more predictable ways can

369 have numerous advantages (Granatosky et al., 2020). For instance, McNeil Alexander proposed

370 that higher safety factors (margins of protection against failure) may be found in structures that

Page 17 of 34

371 experience more variable loading regimes (Alexander, 1997). Consequently, steadier loading

372 regimes could reduce the safety factors needed in the limb bones and, therefore, reduce the

373 energetic costs to produce the higher safety factors. Such factors might be advantageous if

374 loads were elevated due to the highly sprawling posture of P. waltl (Fig. 7). The limb bones of A.

375 tigrinum have been found to possess generally high safety factors against failure (Blob et al.,

376 2014; Kawano et al., 2016; Sheffield and Blob, 2011) and studies are ongoing to determine

377 whether a similar pattern is found in semi-aquatic salamanders.

378 The rate of load application can also affect the load magnitudes experienced by bones.

379 Strain magnitude is positively correlated with strain rate for both cyclic and discrete loading

380 regimes in vertebrates (Aiello et al., 2015), so smaller yank values in P. waltl should be

381 associated with slower strain rates and lower peak strain magnitudes. Lower peak strains would

382 keep bones further from failure, whereas slower strain rates could affect bone mass (Rubin and

383 Lanyon, 1982). Thus, peak strains could have a direct impact on the safety factors of the limb

384 bones, and strain rates could affect the remodelling of bones in response to different loading

385 regimes. Strain rate has been found to be a main factor driving bone remodelling and was

386 positively correlated with higher Young’s modulus values (Aiello et al., 2015), which would

387 confer greater stiffness and greater resistance to bending in bones. Such properties would be

388 advantageous for terrestrial taxa that must habitually withstand the downward effects of gravity

389 on land. Semi-aquatic taxa that spend most of their time underwater might not need higher

390 stiffnesses in their limb bones while their bodies are primarily being supported by buoyancy;

391 however, such traits could be beneficial during the terrestrial eft phase of many newt species

392 (including P. waltl). Future studies that examine changes in the mechanical properties and

Page 18 of 34

393 loading regimes of limb bones across ontogenetic stages within salamanders and between

394 species with different life histories (e.g., direct vs. indirect developers) would provide further

395 insights into the factors that shape limb function.

396

397 Paleontological implications

398 Evidence from the fossil record suggests that terrestrial adaptations first appeared in the

399 anterior regions of the body (Lebedev, 1997), but how rear-wheel drive evolved within stem

400 tetrapods is still debated. Anatomical evaluations of some of the earliest stem tetrapods, such

401 as the elpistostegalid fish Panderichthys (Boisvert, 2005) and the Devonian tetrapod

402 Ichthyostega (Pierce et al., 2012), indicate that the pelvic appendages were likely not effective

403 propulsors on land. As a result, front-wheel drive has been proposed in the early stages of

404 terrestrial locomotion for at least some early tetrapods (Boisvert, 2005; Nyakatura et al., 2014).

405 However, empirical work on the African lungfish (Protopterus annectens) suggests that rear-

406 wheel drive likely evolved for aquatic locomotion when vertebrates were still fully aquatic (King

407 et al., 2011). Further, recent paleontological examinations of the pelvic girdle of the

408 elpistostegalid tetrapodomorph fish Tiktaalik (a relative of Panderichthys) indicate that it

409 exhibited a mosaic of tetrapod-like and fish-like characteristics, including precursors for

410 achieving rear-wheel drive (Shubin et al., 2014). Given the morphological diversity across

411 tetrapodomorphs, it is appropriate to evaluate a range of extant taxa as functional models that

412 represent different locomotor strategies. Our GRF data from semi-aquatic P. waltl limbs build

413 upon previous work on the kinetics of mudskipper pectoral fins and terrestrial salamander limbs

414 (Kawano and Blob, 2013) to offer additional insights for interpreting evolutionary patterns in the

Page 19 of 34

415 incipient stages of terrestrial locomotion, providing a functional model for semi-aquatic tetrapods

416 that exhibit locomotor biomechanics intermediate between those of finned taxa and crownward

417 tetrapods.

418 Taxa with intermediate morphologies and locomotor biomechanics may be particularly

419 relevant for stem tetrapods. Morphological comparisons across the humeri of 40

420 tetrapodomorphs spanning the water-to-land transition demonstrated that stem tetrapods had

421 intermediate morphologies that placed them in a valley between two adaptive peaks (one for

422 aquatic fishes and second for terrestrial crown tetrapods) (Dickson et al., 2021). This was a

423 likely transitional state before natural selection was strong enough to favor the movement of

424 stem tetrapods to the adaptive peak representing the terrestrial characteristics now associated

425 with crown tetrapods (Dickson et al., 2021). Work by Nyakatura and colleagues (Nyakatura et

426 al., 2014) suggests that tetrapods may have gone through an intermediate stage during the

427 transition from front-wheel drive to rear-wheel drive. Specifically, their work evaluated the limb

428 mechanics of a sprawling, belly-dragging lizard, and proposed that belly dragging could have

429 allowed early tetrapods to move on land using appendicular muscles that were less developed

430 (Nyakatura et al., 2014). This intermediate belly-dragging stage of front-wheel drive among early

431 tetrapods would have allowed initial capacities for terrestrial locomotion, after which the role of

432 rear-wheel drive gradually increased. Our findings from the semi-aquatic salamander, P. waltl,

433 may provide a model for a subsequent stage after belly-dragging with front-wheel drive, in which

434 rear-wheel drive has been adopted on land but the forelimbs have not yet acquired fully

435 terrestrial limb mechanics.

Page 20 of 34

436 Kinetic data from the semi-aquatic salamander may serve as a foundation for following up

437 with two hypotheses regarding how terrestrial locomotion evolved (discussed in Pierce et al.,

438 2013). The first hypothesis suggested that the first locomotor mode on land was a trot with

439 lateral bending of the axial system, producing a traveling wave and limbs that acted as ‘struts’.

440 The second hypothesis proposed a lateral sequence walk involving a standing wave, with the

441 limbs generating propulsion. Given that the semi-aquatic forelimb was deceleratory while the

442 hind limb was acceleratory (Tables 1-2, Figs. 1 - 4), P. waltl may be using a modified standing

443 wave in which the hind limbs are generating forward propulsion while the forelimbs are being

444 used as ‘struts’. Such disparity in the propulsive roles of the limbs is not as pronounced in the

445 terrestrial salamander (Table 1 in Kawano and Blob, 2013) and may be due to the different

446 habitats and ecologies that are associated with these species. Similarly, Pierce and colleagues

447 (2020) found interspecific differences in both kinematic and muscle activation patterns during

448 terrestrial walking and proposed that the degree of terrestriality, or other factors related to

449 ecology, may be driving such variation. A gait that was similar to the one used by P. waltl may

450 have allowed the earliest limbed tetrapods to traverse the terrestrial environment with a

451 musculoskeletal system that still primarily functioned for underwater behaviors (McInroe et al.,

452 2016), potentially representing an intermediate stage between sarcopytergian fishes that could

453 accomplish rear-wheel drive underwater (King et al., 2011) and crownward tetrapods that used

454 rear-wheel drive on land.

455 How functional changes evolve has been considered in a variety of systems. Historically, the

456 evolution of locomotor posture has been viewed as a sequential series of gradual evolutionary

457 steps, leading from sprawling to upright (Charig, 1972). However, more recent work highlighted

Page 21 of 34

458 the potential for intermediate taxa to exhibit a broader range of capabilities between the ends of

459 this functional continuum, rather than a graded series of incremental changes between them

460 (Blob, 2001; Kemp, 1978). Hind limb function in the tetrapodomorph fish Tiktaalik has been

461 described with a wide range of capacities that may have been possible due to the relaxed

462 constraints placed upon musculoskeletal systems to overcome gravitational loads (Shubin et al.,

463 2014), potentially indicating intermediate flexibility in locomotor function for an early stage of the

464 fin-to-limb transition that allowed for greater rotations about the hip than tetrapods (and greater

465 stability in the shoulder compared to the hip), but reduced ability to withstand locomotor

466 stresses. Such disparity in limb function is reflected in our data from P. waltl, which suggest that

467 the forelimbs and hind limbs have different functional roles for generating forward movement.

468 Moreover, these changes may not have been strictly coupled to evolutionary changes in

469 appendicular structure since such stark differences between the forelimbs and hind limbs were

470 not found in the GRFs for terrestrial A. tigrinum salamanders. Synthesis of empirical data from

471 biomechanical experiments and morphological analyses of fossils from palaeontology,

472 therefore, holds promise for developing more comprehensive reconstructions of the

473 transformations in the vertebrate musculoskeletal system that led to digit-bearing tetrapods

474 becoming terrestrial (Nyakatura et al., 2019) and the causes and effects of functional evolution

475 more broadly.

476

477 CONCLUSION

478 Our results provide novel insights into the functional changes associated with tetrapods

479 becoming terrestrial, with broader implications for whether functional capabilities could be

Page 22 of 34

480 closely associated with major structural changes (e.g., transition from fin to limb) or could occur

481 through gradual changes that included intermediate stages. The kinetics of terrestrial walking in

482 semi-aquatic P. waltl salamanders differed between the limbs, with the hind limbs sharing

483 numerous similarities to the hind limbs of primarily terrestrial salamanders but the forelimbs

484 exhibiting kinetics that were distinct from the pectoral appendages of semi-aquatic fish and

485 primarily terrestrial salamanders. These data provide additional context to account for

486 differential limb function in models of musculoskeletal function, including those that use

487 salamanders as a blueprint for estimating the locomotor function of early tetrapods. Together,

488 our results present new context for the functional consequences of different locomotor

489 strategies that are possible in functional analogs of tetrapodomorphs that spanned the water-to-

490 land transition.

491

492 ACKNOWLEDGMENTS

493 We thank Zach Marion, Jim Fordyce, and Yu Xia for helpful discussions on statistical analyses

494 that were considered for this study. We are also grateful to Margaret Ptacek, Miriam Ashley-

495 Ross, and anonymous reviewers for feedback on earlier drafts of this manuscript. An earlier

496 draft of this manuscript was submitted by S.M.K. in partial fulfilment of a doctoral dissertation at

497 Clemson University.

498

499 FUNDING

500 Funding to conduct this work was provided to S.M.K. by the American Society for Ichthyologists

501 and Herpetologists Gaige and Raney Awards, Sigma Xi Grants-in-Aid of Research, and

Page 23 of 34

502 Clemson University Stackhouse Fellowship, and to R.W.B. through the National Science

503 Foundation (IOS 0517240 and IOS 0817794).

504

505 COMPETING INTERESTS

506 The authors declare no competing financial, professional or personal interests.

507

508 DATA ACCESSIBILITY

509 All R code are open access via GitHub (https://github.com/MorphoFun/kraken), with the specific

510 scripts for this manuscript available in the “compareGRFs” folder. The raw files for the original

511 force data were too large to include on GitHub but will be made available on Open Science

512 Framework (https://osf.io/) upon publication of the results.

513

514 AUTHOR CONTRIBUTIONS

515 S.M.K and R.W.B. designed the experiment. S.M.K. ran the experiments, wrote the R scripts,

516 and analyzed the data. S.M.K. and R.W.B. contributed to interpreting the results, writing the

517 manuscript, and approving all drafts of the manuscript.

Page 24 of 34 bioRxiv preprint doi: https://doi.org/10.1101/2021.05.01.442256; this version posted May 2, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

REFERENCES Aiello, B.R., Iriarte-Diaz, J., Blob, R.W., Butcher, M.T., Carrano, M.T., Espinoza, N.R., Main, R.P., Ross, C.F., 2015. Bone strain magnitude is correlated with bone strain rate in tetrapods: implications for models of mechanotransduction. Proc. R. Soc. B Biol. Sci. 282, 20150321. https://doi.org/10.1098/rspb.2015.0321 Alexander, R.McN., 1997. A theory of mixed chains applied to safety factors in biological systems. J. Theor. Biol. 184, 247–252. https://doi.org/10.1006/jtbi.1996.0270 Ashley-Ross, M.A., Bechtel, B.F., 2004. Kinematics of the transition between aquatic and terrestrial locomotion in the newt Taricha torosa. J. Exp. Biol. 207, 461–474. https://doi.org/10.1242/jeb.00769 Ashley-Ross, M.A., Lundin, R., Johnson, K.L., 2009. Kinematics of level terrestrial and underwater walking in the California newt, Taricha torosa. J. Exp. Zool. 311A, 240–257. https://doi.org/10.1002/jez.522 Azizi, E., Horton, J.M., 2004. Patterns of axial and appendicular movements during aquatic walking in the salamander Siren lacertina. Zoology 107, 111–120. https://doi.org/10.1016/j.zool.2004.03.002 Bates, D., Maechler, M., Bolker, B., Walker, S., 2015. Fitting linear mixed-effects models using lme4. J. Stat. Softw. 67, 1–48. https://doi.org/doi:10.18637/jss.v067.i01 Blob, R.W., 2001. Evolution of hindlimb posture in nonmammalian therapsids: biomechanical tests of paleontological hypotheses. Paleobiology 27, 14–38. https://doi.org/10.1666/0094-8373(2001)027<0014:EOHPIN>2.0.CO;2 Blob, R.W., Espinoza, N.R., Butcher, M.T., Lee, A.H., D’Amico, A.R., Baig, F., Sheffield, K.M., 2014. Diversity of limb-bone safety factors for locomotion in terrestrial vertebrates: evolution and mixed chains. Integr. Comp. Biol. 54, 1058–1071. https://doi.org/10.1093/icb/icu032 Boisvert, C.A., 2005. The pelvic fin and girdle of Panderichthys and the origin of tetrapod locomotion. Nature 438, 1145–7. https://doi.org/10.1038/nature04119 Brand, L.R., 1996. Variations in salamander trackways resulting from substrate differences. J. Paleontol. 70, 1004–1010. Canoville, A., Laurin, M., 2009. Microanatomical diversity of the humerus and lifestyle in lissamphibians. Acta Zool. 90, 110–122. https://doi.org/10.1111/j.1463- 6395.2008.00328.x

Page 25 of 34 bioRxiv preprint doi: https://doi.org/10.1101/2021.05.01.442256; this version posted May 2, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

Charig, A.J., 1972. The evolution of the archosaur pelvis and hind-limb: an explanation in functional terms, in: Joysey, K.A., Kemp, T.S. (Eds.), Studies in Vertebrate Evolution. Oliver and Boyd, Edinburgh, pp. 121–155. Chen, J.J., Peattie, A.M., Autumn, K., Full, R.J., 2006. Differential leg function in a sprawled- posture quadrupedal trotter. J. Exp. Biol. 209, 249–59. https://doi.org/10.1242/jeb.01979 Coates, M.I., 1996. The Devonian tetrapod Acanthostega gunnari Jarvik: postcranial anatomy, basal tetrapod interrelationships and patterns of skeletal evolution. Trans. R. Soc. Edinb. Earth Sci. 87, 363–421. Cohen, J., 1990. Things I have learned (so far). Am. Psychol. 45, 1304–1312. Cumming, G., 2014. The new statistics: why and how. Psychol. Sci. 25, 7–29. https://doi.org/10.1177/0956797613504966 Deban, S.M., Schilling, N., 2009. Activity of trunk muscles during aquatic and terrestrial locomotion in Ambystoma maculatum. J. Exp. Biol. 212, 2949–59. https://doi.org/10.1242/jeb.032961 Delvolvé, I., Bem, T., Cabelguen, J.M., 1997. Epaxial and limb muscle activity during swimming and terrestrial stepping in the adult newt, Pleurodeles waltl. J. Neurophysiol. 78, 638–50. Dickson, B.V., Clack, J.A., Smithson, T.R., Pierce, S.E., 2021. Functional adaptive landscapes predict terrestrial capacity at the origin of limbs. Nature 589, 242–245. https://doi.org/10.1038/s41586-020-2974-5 Frolich, L.M., Biewener, A.A., 1992. Kinematic and electromyographic analysis of the functional role of the body axis during terrestrial and aquatic locomotion in the salamander Ambystoma tigrinum . J. Exp. Biol. 162, 107–130. Gao, K., Shubin, N.H., 2001. Late Jurassic salamanders from northern China. Nature 410, 574– 577. Gatesy, S.M., Middleton, K.M., Jenkins Jr., F.A., Shubin, N.H., 1999. Three-dimensional preservation of foot movements in Triassic theropod dinosaurs. Nature 399, 141–144. https://doi.org/10.1038/20167 Granatosky, M.C., McElroy, E.J., Lemelin, P., Reilly, S.M., Nyakatura, J.A., Andrada, E., Kilbourne, B.M., Allen, V.R., Butcher, M.T., Blob, R.W., Ross, C.F., 2020. Variation in limb loading magnitude and timing in tetrapods. J. Exp. Biol. 223, jeb201525. https://doi.org/10.1242/jeb.201525 Karakasiliotis, K., Schilling, N., Cabelguen, J.-M., Ijspeert, A.J., 2012. Where are we in understanding salamander locomotion: biological and robotic perspectives on kinematics. Biol. Cybern. 107, 529–544. https://doi.org/10.1007/s00422-012-0540-4

Page 26 of 34 bioRxiv preprint doi: https://doi.org/10.1101/2021.05.01.442256; this version posted May 2, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

Kawano, S.M., Blob, R.W., 2013. Propulsive forces of mudskipper fins and salamander limbs during terrestrial locomotion: implications for the invasion of land. Integr. Comp. Biol. 53, 283–294. https://doi.org/10.1093/icb/ict051 Kawano, S.M., Economy, D.R., Kennedy, M.S., Dean, D., Blob, R.W., 2016. Comparative limb bone loading in the humerus and femur of the tiger salamander: testing the ‘mixed-chain’ hypothesis for skeletal safety factors. J. Exp. Biol. 219, 341–353. https://doi.org/10.1242/jeb.125799 Kemp, T.S., 1978. Stance and gait in the hindlimb of a therocephalian mammal-like reptile. J. Zool. 186, 143–161. King, H.M., Shubin, N.H., Coates, M.I., Hale, M.E., 2011. Behavioral evidence for the evolution of walking and bounding before terrestriality in sarcopterygian fishes. Proc. Natl. Acad. Sci. U. S. A. 108, 21146–21151. https://doi.org/10.1073/pnas.1118669109/- /DCSupplemental.www.pnas.org/cgi/doi/10.1073/pnas.1118669109 Laurin, M., Girondot, M., Loth, M.-M., 2004. The evolution of long bone microstructure and lifestyle in lissamphibians. Paleobiology 30, 589–613. LeBeau, B., Song, Y.A., Liu, W.C., 2018. Model misspecification and assumption violations with the linear mixed model: a meta-analysis: SAGE Open 8, 1–16. https://doi.org/10.1177/2158244018820380 Lebedev, O.A., 1997. Fins made for walking. Nature 390, 21–22. https://doi.org/10.1038/36215 Lenth, R., 2019. emmeans: estimated marginal means, aka least-squares means. R package version 1.4.3. https://CRAN.R-project.org/package=emmeans. Lin, D.C., McGowan, C.P., Blum, K.P., Ting, L.H., 2019. Yank: the time derivative of force is an important biomechanical variable in sensorimotor systems. J. Exp. Biol. 222, jeb180414. https://doi.org/10.1242/jeb.180414 Lüdecke, D., Makowski, D., Waggoner, P., 2019. performance: assessment of regression models performance. R package version 0.4.0. https://CRAN.R- project.org/package=performance. Maidment, S.C.R., Bates, K.T., Falkingham, P.L., VanBuren, C., Arbour, V., Barrett, P.M., 2014. Locomotion in ornithischian dinosaurs: an assessment using three-dimensional computational modelling. Biol. Rev. 89, 588–617. https://doi.org/10.1111/brv.12071 Martinez, M.M., Full, R.J., Koehl, M.A.R., 1998. Underwater punting by an intertidal crab: a novel gait revealed by the kinematics of pedestrian locomotion in air versus water. J. Exp. Biol. 201, 2609–23.

Page 27 of 34 bioRxiv preprint doi: https://doi.org/10.1101/2021.05.01.442256; this version posted May 2, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

McElroy, E.J., Wilson, R., Biknevicius, A.R., Reilly, S.M., 2014. A comparative study of single leg ground reaction forces in running lizards. J. Exp. Biol. 217, 735–742. https://doi.org/10.1242/jeb.095620 McInroe, B., Astley, H.C., Gong, C., Kawano, S.M., Schiebel, P.E., Rieser, J.M., Choset, H., Blob, R.W., Goldman, D.I., 2016. Tail use improves performance on soft substrates in models of early vertebrate land locomotors. Science 353, 154–158. https://doi.org/10.1126/science.aaf0984 Nyakatura, J.A., Andrada, E., Curth, S., Fischer, M.S., 2014. Bridging “Romer’s Gap”: limb mechanics of an extant belly-dragging lizard inform debate on tetrapod locomotion during the early Carboniferous. Evol. Biol. 41, 175–190. https://doi.org/10.1007/s11692- 013-9266-z Nyakatura, J.A., Melo, K., Horvat, T., Karakasiliotis, K., Allen, V.R., Andikfar, A., Andrada, E., Arnold, P., Lauströer, J., Hutchinson, J.R., Fischer, M.S., Ijspeert, A.J., 2019. Reverse- engineering the locomotion of a stem amniote. Nature 565, 351–355. https://doi.org/10.1038/s41586-018-0851-2 Pace, C.M., Gibb, A.C., 2014. Sustained periodic terrestrial locomotion in air-breathing fishes. J. Fish Biol. 84, 639–660. https://doi.org/10.1111/jfb.12318 Petranka, J.W., 1998. Salamanders of the United States and Canada. Smithsonian Institution Press, Washington, D.C. Pierce, S.E., Clack, J.A., Hutchinson, J.R., 2012. Three-dimensional limb joint mobility in the early tetrapod Ichthyostega. Nature 486, 523–526. https://doi.org/10.1038/nature11124 Pierce, S.E., Hutchinson, J.R., Clack, J.A., 2013. Historical perspectives on the evolution of tetrapodomorph movement. Integr. Comp. Biol. 53, 209–23. https://doi.org/10.1093/icb/ict022 Pierce, S.E., Lamas, L.P., Pelligand, L., Schilling, N., Hutchinson, J.R., 2020. Patterns of limb and epaxial muscle activity during walking in the fire salamander, Salamandra salamandra. Integr. Org. Biol. 2. https://doi.org/10.1093/iob/obaa015 R Core Team, 2019. R: a language and environment for statistical computing. R Foundation for Statistical Computing. https://www.R-project.org/, Vienna, Austria. RStudio Team, 2015. RStudio: integrated development for R. R Studio, Inc., Boston, MA. Rubin, C.T., Lanyon, L.E., 1982. Limb mechanics as a function of speed and gait: a study of functional strains in the radius and tibia of horse and dog. J. Exp. Biol. 101, 187–211.

Page 28 of 34 bioRxiv preprint doi: https://doi.org/10.1101/2021.05.01.442256; this version posted May 2, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

Sass, G.G., Motta, P.J., 2002. The effects of satiation on strike mode and prey capture kinematics in the largemouth bass, Micropterus salmoides. Environ. Biol. Fishes 65, 441–454. https://doi.org/10.1023/A:1021108519634 Schoch, R.R., Werneburg, R., Voigt, S., 2020. A Triassic stem-salamander from Kyrgyzstan and the origin of salamanders. Proc. Natl. Acad. Sci. https://doi.org/10.1073/pnas.2001424117 Sheffield, K.M., Blob, R.W., 2011. Loading mechanics of the femur in tiger salamanders (Ambystoma tigrinum) during terrestrial locomotion. J. Exp. Biol. 214, 2603–2615. https://doi.org/10.1242/jeb.048736 Shubin, N.H., Daeschler, E.B., Jenkins, F.A., 2014. Pelvic girdle and fin of Tiktaalik roseae. Proc. Natl. Acad. Sci. U. S. A. 111, 893–9. https://doi.org/10.1073/pnas.1322559111 Signal developers, 2013. signal: signal processing. Smith, P.F., 2017. A guerilla guide to common problems in ‘neurostatistics’: essential statistical topics in neuroscience. J. Undergrad. Neurosci. Educ. 16, R1–R12. Wainwright, P.C., Mehta, R.S., Higham, T.E., 2008. Stereotypy, flexibility and coordination: key concepts in behavioral functional morphology. J. Exp. Biol. 211, 3523–3528. https://doi.org/10.1242/jeb.007187 Wake, D.B., 2009. What salamanders have taught us about evolution. Annu. Rev. Ecol. Evol. Syst. 40, 333–352. https://doi.org/10.1146/annurev.ecolsys.39.110707.173552 Young, V.K.H., Wienands, C.E., Wilburn, B.P., Blob, R.W., 2017. Humeral loads during swimming and walking in turtles: implications for morphological change during aquatic reinvasions. J. Exp. Biol. 220, 3873–3877. https://doi.org/10.1242/jeb.156836

Page 29 of 34 bioRxiv preprint doi: https://doi.org/10.1101/2021.05.01.442256; this version posted May 2, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

FIGURE LEGENDS Figure 1. Profiles of GRF production in the pectoral appendages during stance. Individual trials were pooled within each species (N = 50 trials for each) and profiles are plotted as the mean ± s.e.m. for each percent of stance. Mediolateral angles were set relative to vertical (0°), so negative values indicate a medial direction of the GRF. Anteroposterior angles were set relative to vertical (0°), so negative values represent a posterior direction of the GRF. Although the general shapes of the curves were comparable across the species for a given parameter, pectoral appendages for P. waltl (dotted green lines) tended to support lower magnitudes of body weight, and have a more deceleratory role and intermediate mediolateral angle, compared to the semi-aquatic P. barbarus (dashed blue lines) and the primarily terrestrial A. tigrinum (solid brown lines).

Figure 2. Parameters at the peak net GRF compared across the pectoral appendages. P. waltl (green) generally has lower magnitudes for the vertical component and net GRF as demonstrated by the histograms, yet also had a greater variability compared to the semi-aquatic fish P. barbarus (blue) and the primarily terrestrial salamander A. tigrinum (brown) as illustrated by the whisker-box plots. Sample sizes for data from the pectoral appendages: P. barbarus (n = 50), P. waltl (n = 50), and A. tigrinum (n = 50).

Figure 3. Profiles of GRF production in the pelvic appendages during stance. Individual trials were pooled within each species (N = 49 trials for each) and profiles are plotted as the mean ± s.e.m. for each percent of stance. Mediolateral angles were set relative to vertical (0°), so negative values indicate a medial direction of the GRF. Anteroposterior angles were set relative to vertical (0°), so negative values represent a posterior direction of the GRF. Although the general shapes of the curves were comparable across the species for a given parameter, P. waltl hind limbs (dashed green lines) tended to support lower magnitudes of body weight and had less of an acceleratory role compared to the primarily terrestrial salamander A. tigrinum (solid brown lines).

Figure 4. Parameters at the peak net GRF compared across the pelvic appendages. P. waltl hind limbs (green) generally had lower magnitudes for the vertical component and net GRF as demonstrated by the histograms, yet the angle of GRF orientation could become almost 2x more medial than for the primarily terrestrial salamander A. tigrinum (brown) as illustrated by the whisker-box plots. Sample sizes for the pelvic appendages: P. waltl (n = 49), and A. tigrinum (n

Page 30 of 34 bioRxiv preprint doi: https://doi.org/10.1101/2021.05.01.442256; this version posted May 2, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

= 49).

Figure 5. Profiles of the rate of GRF production (‘yank’) in the pectoral appendages during stance. Individual trials were pooled within each species (N = 50 trials for each) and profiles are plotted as the mean ± s.e.m. for each percent of stance. Mediolateral components were set relative to vertical (0°), so negative values indicate a medial direction of the GRF. Anteroposterior components were set relative to vertical (0°), so negative values represent a posterior direction of the GRF. Yank values for P. waltl (dotted green lines) tended to span a narrower range of values and did not generally follow the same shape dynamics as the semi- aquatic fish P. barbarus (dashed blue lines) and the primarily terrestrial salamander A. tigrinum (solid brown lines), which tended to yield comparable data.

Figure 6. Profiles of the rate of GRF production (‘yank’) in the pelvic appendages during stance. Individual trials were pooled within each species (N = 49 trials for each) and profiles are plotted as the mean ± s.e.m. for each percent of stance. Mediolateral components were set relative to vertical (0°), so negative values indicate a medial direction of the GRF. Anteroposterior components were set relative to vertical (0°), so negative values represent a posterior direction of the GRF. Yank values for P. waltl hind limbs (dashed green lines) tended to span a narrower range of values for the vertical component and net GRF, but were otherwise relatively comparable to those from the primarily terrestrial salamander A. tigrinum (solid brown lines).

Figure 7. Data on GRF production in semi-aquatic P. waltl salamanders contribute towards a broader perspective on the evolution of the locomotor biomechanics across the fish-tetrapod transition. The orientation of the GRF (represented as a black arrow) in the medial direction was historically thought to be evolutionarily conserved across tetrapods, despite variation in limb postures. It was unclear whether the same pattern was found in fishes, but our previous work found that GRFs produced by pectoral fins in mudskipper fishes were more medial, likely resulting from their more sprawled posture compared to tetrapods. GRF data from P. waltl in this study demonstrate that the more medial orientation of GRFs is not exclusive to fishes, suggesting that ecology may have an important role in shaping the kinetics during terrestrial locomotion.

Page 31 of 34

bioRxiv preprint

Table 1. Comparison of locomotor variables between the pectoral appendages of mudskippers and salamanders at the peak net GRF

was notcertifiedbypeerreview)istheauthor/funder.Allrightsreserved.Noreuseallowedwithoutpermission. doi:

Periophthalmus barbarus Pleurodeles waltl Ambystoma tigrinum Goodness-of-fit https://doi.org/10.1101/2021.05.01.442256 n = 50 n = 50 n = 50

2 2 FE ± s.e. CI FE ± s.e. CI FE ± s.e. CI R LMM(m) R LMM(c) Net GRF (BW) 0.443 ± 0.015 0.410, 0.476 0.397 ± 0.016 0.363, 0.431 0.458 ± 0.015 0.424, 0.491 0.110 0.225

Vertical GRF (BW) 0.417 ± 0.013 0.389, 0.445 0.374 ± 0.013 0.346, 0.403 0.447 ± 0.013 0.420, 0.475 0.154 0.214

Mediolateral GRF (BW) -0.129 ± 0.012 -0.155, -0.103 -0.087 ± 0.012 -0.114, -0.061 -0.068 ± 0.012 -0.094, -0.041 0.260 0.491

Anteroposterior GRF (BW) 0.048 ± 0.014 0.018, 0.078 -0.079 ± 0.014 -0.110, -0.049 -0.029 ± 0.014 -0.059, 0.0004 0.515 0.651

Mediolateral angle (°) -17.12 ± 1.46 -20.3, -13.94 -14.16 ± 1.50 -17.4, -10.89 -8.64 ± 1.46 -11.8, -5.46 0.258 0.421 ; this versionpostedMay2,2021. Anteroposterior angle (°) 6.69 ± 1.64 3.09, 10.28 -11.11 ± 1.69 -14.79, -7.43 -3.32 ± 1.64 -6.92, 0.27 0.543 0.645

Statistical analyses were based on the model: lmer(y ∼ Species + (1|Individual), REML=True). Fixed effect (FE) values, standard errors, and 2 confidence interval (CI) were estimated from the linear mixed effects model using the emmeans R package. R LMM(m) represents the variance 2 explained by the fixed effects whereas R LMM(c) represents the variance explained by the combination of fixed and random effects, and were calculated using the performance::r2_nakagawa() function in R.

The copyrightholderforthispreprint(which

Page 32 of 34

bioRxiv preprint

Table 2. Comparison of locomotor variables between the pelvic appendages of salamanders at the peak net GRF

was notcertifiedbypeerreview)istheauthor/funder.Allrightsreserved.Noreuseallowedwithoutpermission. Pleurodeles waltl Ambystoma tigrinum Goodness-of-fit doi: n = 49 n = 49 https://doi.org/10.1101/2021.05.01.442256

2 2 FE ± s.e. CI FE ± s.e. CI R LMM(m) R LMM(c) Net GRF (BW) 0.349 ± 0.017 0.309, 0.389 0.497 ± 0.017 0.457, 0.537 0.585 0.715

Vertical GRF (BW) 0.314 ± 0.021 0.266, 0.362 0.466 ± 0.021 0.419, 0.514 0.558 0.734

Mediolateral GRF (BW) -0.096 ± 0.014 -0.129, -0.063 -0.071 ± 0.014 -0.104, -0.037 0.056 0.350

Anteroposterior GRF (BW) 0.092 ± 0.017 0.053. 0.132 0.132 ± 0.017 0.092, 0.172 0.102 0.418

Mediolateral angle (°) -16.97 ± 2.27 -22.2, -11.7 -8.73 ± 2.27 -14.0, -3.5 0.215 0.483 ;

Anteroposterior angle (°) 17.0 ± 2.71 10.7, 23.2 16.3 ± 2.70 10.0, 22.5 0.001 0.315 this versionpostedMay2,2021.

The copyrightholderforthispreprint(which

Page 33 of 34

bioRxiv preprint

Table 3. Range of average yank values compared between appendages and species

was notcertifiedbypeerreview)istheauthor/funder.Allrightsreserved.Noreuseallowedwithoutpermission. Vertical Mediolateral Anteroposterior Net doi: https://doi.org/10.1101/2021.05.01.442256 FE ± s.e.m. CI FE ± s.e.m. CI FE ± s.e.m. CI FE ± s.e.m. CI Pectoral - Max P. barbarus 0.010 ± 0.001 0.008, 0.012 0.004 ± 0.001 0.003, 0.006 0.004 ± 0.001 0.002, 0.005 0.011 ± 0.001 0.008, 0.013

P. waltl 0.009 ± 0.001 0.007, 0.011 0.006 ± 0.001 0.004, 0.008 0.008 ± 0.001 0.007, 0.010 0.009 ± 0.001 0.007, 0.012

A. tigrinum 0.012 ± 0.001 0.010, 0.014 0.004 ± 0.001 0.002, 0.005 0.006 ± 0.001 0.004, 0.008 0.012 ± 0.001 0.010, 0.014

Pectoral - Min

P. barbarus -0.012 ± 0.001 -0.014, -0.010 -0.005 ± 0.001 -0.006, -0.003 -0.004 ± 0.001 -0.005, -0.003 -0.013 ± 0.001 -0.015, -0.011 ; this versionpostedMay2,2021.

P. waltl -0.009 ± 0.001 -0.011, -0.007 -0.006 ± 0.001 -0.008, -0.005 -0.005 ± 0.001 -0.006, -0.003 -0.010 ± 0.001 -0.012, -0.007

A. tigrinum -0.013 ± 0.001 -0.016, -0.011 -0.004 ± 0.001 -0.006, -0.003 -0.007 ± 0.001 -0.008, -0.005 -0.014 ± 0.001 -0.016, -0.011

Pelvic - Max P. waltl 0.005 ± 0.001 0.002, 0.008 0.006 ± 0.001 0.005, 0.008 0.006 ± 0.001 0.004, 0.008 0.006 ± 0.001 0.003, 0.009

A. tigrinum 0.010 ± 0.001 0.007, 0.013 0.006 ± 0.001 0.005, 0.008 0.009 ± 0.001 0.007, 0.011 0.010 ± 0.001 0.007, 0.014 The copyrightholderforthispreprint(which Pelvic - Min P. waltl -0.009 ± 0.001 -0.011, -0.007 -0.004 ± 0.0004 -0.005, -0.003 -0.006 ± 0.001 -0.007, -0.005 -0.009 ± 0.001 -0.011, -0.007

A. tigrinum -0.014 ± 0.001 -0.016, -0.012 -0.004 ± 0.0004 -0.005, -0.003 -0.007 ± 0.001 -0.008, -0.006 -0.015 ± 0.001 -0.016, -0.013

Sample sizes for the pectoral yank LMM: P. barbarus (n = 50), P. waltl (n = 50), and A. tigrinum (n = 50). Sample sizes for the pelvic yank LMM: P. waltl (n = 49), and A. tigrinum (n = 49). FE = estimates of the coefficients for the fixed effects based on estimated marginal means with the emmeans::emmeans() function. Statistical analyses were based on the model: lme4::lmer(y ∼ species + (1 | individual), REML = True). The confidence interval (CI) is listed as the lower and upper bounds, respectively, that were reported from the LMMs.

Page 34 of 34

bioRxiv preprint doi: https://doi.org/10.1101/2021.05.01.442256; this version posted May 2, 2021. The copyright holder for this preprint (which was not certified Figureby peer review) is1: the author/funder. Pector All alrights appendagesreserved. No reuse allowed without permission.

a 0.5 b 0.5

0.4 0.4

0.3 0.3 tical (BW) er 0.2 0.2 GRF − Net (BW)

0.1 GRF − v 0.1

0.0 0.0 0 25 50 75 100 0 25 50 75 100 c 0.2 d 0.2

0.1 0.1 ior (BW) al (BW)

0.0 0.0

−0.1 −0.1 GRF − mediolater GRF − anteroposter −0.2 −0.2 0 25 50 75 100 0 25 50 75 100 e 50 f 50

25 25 ior angle (deg) al angle (deg) 0 0

−25 −25

GRF − mediolater −50 −50 GRF − anteroposter 0 25 50 75 100 0 25 50 75 100

Percent Stance

species A. tigrinum P. barbarus P. waltl bioRxiv preprint doi: https://doi.org/10.1101/2021.05.01.442256; this version posted May 2, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. Figure 2

A 0.6 B C

0.0 0.5 0.5 al

tical 0.4 −0.1 0.4 er GRF − net

GRF − v 0.3 0.3 GRF − mediolater −0.2

0.2 0.2 −0.3

um altl us um altl us um altl us in in in P. w P. w P. w A. tigr P. barbar A. tigr P. barbar A. tigr P. barbar

D E F 20 0.1 0 10 ior 0.0 −10 0 ior angle al angle

−0.1 −20 −10 mediolater anteroposter GRF − anteroposter

−30 −20 −0.2

um altl us um altl us um altl us in in in P. w P. w P. w A. tigr P. barbar A. tigr P. barbar A. tigr P. barbar

bioRxiv preprint doi: https://doi.org/10.1101/2021.05.01.442256; this version posted May 2, 2021. The copyright holder for this preprint (which was not certified by Figurepeer review) is the 3: author/funder. Pelvic All rights appendages reserved. No reuse allowed without permission.

a 0.5 b 0.5

0.4 0.4

0.3 0.3 tical (BW) er 0.2 0.2 GRF − Net (BW)

0.1 GRF − v 0.1

0.0 0.0 0 25 50 75 100 0 25 50 75 100 c 0.2 d 0.2

0.1 0.1 ior (BW) al (BW)

0.0 0.0

−0.1 −0.1 GRF − mediolater GRF − anteroposter −0.2 −0.2 0 25 50 75 100 0 25 50 75 100 e 50 f 50

25 25 ior angle (deg) al angle (deg) 0 0

−25 −25

GRF − mediolater −50 −50 GRF − anteroposter 0 25 50 75 100 0 25 50 75 100

Percent Stance

species A. tigrinum P. waltl bioRxiv preprint doi: https://doi.org/10.1101/2021.05.01.442256; this version posted May 2, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. Figure 4

A B C 0.6 0.00 0.6

0.5 −0.05 0.5 al tical

er 0.4 0.4 −0.10 GRF − net

GRF − v 0.3 0.3 GRF − mediolater −0.15

0.2 0.2 −0.20

um altl um altl um altl in in in P. w P. w P. w A. tigr A. tigr A. tigr

D E F 0.3 0

−10 0.2 40 ior ior angle al angle −20 0.1 20

mediolater −30 anteroposter GRF − anteroposter

0.0 0 −40

um altl um altl um altl in in in P. w P. w P. w A. tigr A. tigr A. tigr

Figure 5: Pectoral appendages

a 0.02 b 0.02

0.01 0.01 tical er 0.00 0.00 ank − Net Y ank − v Y

−0.01 −0.01

−0.02 −0.02

0 25 50 75 100 0 25 50 75 100 c 0.02 d 0.02

0.01 0.01 ior al

0.00 0.00 ank − mediolater Y ank − anteroposter

−0.01 Y −0.01

−0.02 −0.02

0 25 50 75 100 0 25 50 75 100

Percent Stance

Figure 6: Pelvic appendages

a 0.02 b 0.02

0.01 0.01 tical er 0.00 0.00 ank − Net Y ank − v Y

−0.01 −0.01

−0.02 −0.02

0 25 50 75 100 0 25 50 75 100 c 0.02 d 0.02

0.01 0.01 ior al

0.00 0.00 ank − mediolater Y ank − anteroposter

−0.01 Y −0.01

−0.02 −0.02

0 25 50 75 100 0 25 50 75 100

Percent Stance

Figure 7