This is the accepted version of the article:
Logghe, A.; Mujal, Eudald; Marchetti, Lorenzo; [et al.]. «Hyloidichnus trackways
with digit and tail drag traces from the Permian of Gonfaron (Var, France): New
insights on the locomotion of captorhinomorph eureptiles». Palaeogeography,
palaeoclimatology, palaeoecology, Vol. 573 (July 2021), art. 110436. DOI
10.1016/j.palaeo.2021.110436


This version is available at https://ddd.uab.cat/record/239888

under the terms of the                license
                                                      1


 1       Hyloidichnus trackways with digit and tail drag traces from the Permian of Gonfaron

 2            (Var, France): new insights on the locomotion of captorhinomorph eureptiles

 3


 4   A. Logghe1*, E. Mujal2,3, L. Marchetti4, A. Nel5, J.-M. Pouillon6, S. Giner7, R. Garrouste5, J.-

 5                                                S. Steyer1.

 6   1
         Centre de Recherches en Paléontologie de Paris, UMR 7202 – CNRS, MNHN, SU, EPHE,

 7   Muséum National d’Histoire Naturelle, Sorbonne Universités, 8 rue Buffon, CP38, F-75005

 8   Paris, France.

     2
 9       Staatliches Museum für Naturkunde Stuttgart, Rosenstein 1, D-70191 Stuttgart, Germany.

     3
10       Institut Català de Paleontologica M. Crusafont, ICTA-ICP building, c/ de les columnes, s/n, E-

11   08193 Cerdanyola del Vallès, Catalonia, Spain.

     4
12       Museum für Naturkunde Berlin, Leibniz-Institut für Evolutions- und Biodiversitätsforschung,

13   Invalidenstrasse 43, 10115 Berlin, Germany.

14   5
         Institut de Systématique, Evolution, Biodiversité, ISYEB – UMR 7205 – CNRS, MNHN, SU

15   EPHE, Muséum National d’Histoire Naturelle, Sorbonne Universités, Université des Antilles,

16   57 rue Cuvier, CP 50, Entomologie, F-75005, Paris, France.

     6
17       Nivolas Vermelle, France.

18   7
         Service d’Archéologie, Direction de l’Ingénierie Territoriale, Département du Var, 57 rue

19   Gustave Bret, Fréjus, France.

20   Antoine Logghe:           antoine.logghe@edu.mnhn.fr (corresponding author)

21                             https://orcid.org/0000-0002-2854-77

22   Eudald Mujal:             eudald.mujalgrane@smns-bw.de
                                                    2


23                          https://orcid.org/0000-0002-6310-323X

24   Lorenzo Marchetti:     lorenzo.marchetti85@gmail.com

25                          https://orcid.org/0000-0002-1047-7887

26   André Nel:             andre.nel@mnhn.fr

27                          https://orcid.org/0000-0002-4241-7651

28   Jean-Marc Pouillon:    jmpdb@wanadoo.fr

29                          https://orcid.org/0000-0001-6016-4734

30   Stephen Giner:         sginer@var.fr

31   Romain Garrouste:      romain.garrouste@mnhn.fr

32                          https://orcid.org/0000-0002-0880-5781

33   Jean-Sébastien Steyer: jean-sebastien.steyer@mnhn.fr

34                          https://orcid.org/0000-0003-1835-7852

35


36   Abstract – Newly discovered tetrapod footprints from the middle Permian Pelitic Formation

37   of Gonfaron (Le Luc Basin, Var, France) are described in detail and assigned to the ichnogenus

38   Hyloidichnus. These specimens are very well-preserved, with detailed trackways showing

39   anatomical features, digit drag traces and continuous tail impressions. Together with other

40   Hyloidichnus specimens from the same locality and from the University of Burgundy

41   collections, they allow the identification of the possible trackmakers of Hyloidichnus: small

42   Hyloidichnus footprints are correlated with captorhinomorphs with autopodia similar to

43   Captorhinus whereas large Hyloidichnus footprints might be correlated with larger

44   captorhinomorphs or other ‘basal’ (eventually stem-) moradisaurines. A detailed comparative

45   analysis is proposed to better understand the locomotion style of the trackmakers. Starting from

46   the reconstruction of the stance phase of Captorhinus, we document the swing phase of the

47   fore- and hind-limbs of captorhinomorphs thanks to the continuous digit drag traces associated
                                                    3


48   with Hyloidichnus. A link between gait and track preservation have been highlighted through

49   the analysis of tail impressions and high-resolution 3D models. In all, this study increases our

50   understanding of captorhinomorph locomotion and enhances the integration of both tracks and

51   skeletal remains to highlight the biomechanics of Permian tetrapods.

52


53   Keywords – Biomechanics, Guadalupian-Wordian, tetrapod footprints, track-trackmaker

54   correlation, Catporhinus, ichnology.

55


56   1. Introduction

57          Fossil footprints provide qualitative and quantitative information on anatomy and

58   behaviour of ancient tetrapods. This includes valuable data on, for example, functional

59   morphology and locomotion of the trackmakers (Voigt et al., 2007; Romano et al., 2016;

60   Marchetti et al, 2017; Buchwitz and Voigt, 2018; Mujal and Marchetti, 2020; Mujal and

61   Schoch, 2020; Mujal et al., 2020). Tetrapod footprints are usually compared with the

62   morphology of the appendicular skeletons of contemporary taxa (Gand, 1987; Voigt, 2005;

63   Gand and Durand, 2006; Voigt et al., 2007; Marchetti et al., 2019a, 2020a; Mujal and Marchetti,

64   2020; Mujal and Schoch, 2020). To render these studies possible, well-preserved tracks and

65   trackways are needed: these exhibit morphological features directly related to anatomy, without

66   ‘extramorphologies’ (i.e., morphological features that hide anatomical traits of the producer;

67   Peabody, 1948; Haubold et al., 1995; Marchetti et al., 2019b). Tetrapod footprints arranged in

68   pes-manus couples along trackways provide important information on the trunk length and pes

69   and manus orientation of the trackmaker. Moreover, the possible presence of a continuous tail

70   impression as well as digit drag traces can be key elements to perform a locomotion

71   reconstruction (e.g., Bernardi and Avanzini, 2011; Marchetti et al., 2017; Mujal et al., 2017).
                                                     4


72          One of the most important amniote clades of the Palaeozoic is the Captorhinomorpha

73   (and closely-related protorothyridids). Captorhinomorph eureptiles are relatively well-known

74   thanks to various body fossils from the Permian of Europe, America and Africa, but their

75   footprints were less-intensively studied, although common (e.g., Voigt and Lucas, 2018).

76   Consequently, studies that correlate Captorhinomorpha track and skeletal records are rare (e.g.,

77   Voigt et al., 2010; Marchetti et al., 2020b). Palaeozoic footprints attributable to this group are

78   probably represented by three ichnogenera: Hyloidichnus Gilmore, 1927, Merifontichnus Gand

79   et al., 2000 and Notalacerta Butts, 1891 (e.g., Marchetti et al., 2020b). Although Hyloidichnus

80   is generally attributed to captorhinomorphs (e.g., Voigt et al., 2010), precise studies correlating

81   tracks and trackmakers and analysing locomotion patterns are lacking. Here, we describe in

82   detail four Hyloidichnus trackways from the Permian Pelitic Formation of Gonfaron (Le Luc

83   Basin, Provence, France). This material is exceptionally well-preserved, with continuous tail

84   and digit drag traces, and allows a confident correlation of Hyloidichnus with captorhinomorph

85   eureptiles with autopodia similar to Captorhinus and the performance of a precise analysis of

86   the trackmaker step cycle. This brings new insights on the locomotion of Permian tetrapods and

87   adds new elements on the possible disparities amongst different clades (Mujal et al., 2020).

88


89   2. Geological setting

90          The Permian basins of Provence are from East to West (Fig. 1A) the Estérel and Bas-

91   Argens basins bordered to the north by the Tanneron massif, and the Luc and Toulon-Cuers

92   basins bordered to the South by the Maures massif and the Mediterranean Sea. The Luc and

93   Toulon-Cuers basins are bounded to the North by a thin deposit of Triassic non-marine

94   sediments ('Buntsandstein', Anisian) followed by thick deposits of Mesozoic sediments (see

95   Fig. 1A) (Durand et al., 2011). The topography of these basins is strongly marked by the
                                                     5


 96   Permian depression that extends from the Toulon-Cuers Basin to the Bas-Argens Basin,

 97   corresponding to the erosion of the sedimentary fill of the Toulon-Cuers and Le Luc basins.

 98   The Pelitic Formation is the upper formation of the Luc Basin. It is dated to the Guadalupian,

 99   possibly Wordian, according to radioisotopic ages, palaeomagnetism and correlation of the

100   underlying units with fossil-bearing units of nearby basins, which include macrofloral remains,

101   sporomorphs, ostracods, insects and tetrapod footprints (Gand and Durand, 2006; Durand,

102   2008; Garrouste et al., 2017). The material has been found within a red pelitic level of about 3

103   m thick that ends with a layer of indurated argilits in laminated plates containing mudcracks,

104   ripple marks and raindrops. This red pelitic level also yields abundant ichnites on bio-mats,

105   arthropod trace fossils (Scoyenia, Planolites and Acripes multiformis) and arthropod trackways

106   (Lithographus hieroglyphus), and cone ichnofossils. Several tetrapod trackways such as

107   Hyloidichnus major, Hyloidichnus isp, Microsauripus/Varanopus isp., Dromopus isp. and

108   Batrachichnus salamandroides have also been recovered. This ichno-assemblage characterises

109   a facies close to exondation and a possible ephemeral lake of decametric extension, attracting

110   an abundant fauna.

111


112   3. Material and methods

113          Abbreviations: GONF-A: Gonfaron collection of the MNHN; GONF-G: Gonfaron

114   collection of the University of Burgundy collected by G. Gand.

115          The four trackways described here have been found on the same locality, the ‘Gonfaron

116   A site’ near the eponymous village, Var Département, Southern France. They come from the

117   middle Permian Pelitic Formation of the Le Luc Basin, Provence (Fig. 1). Among the four

118   specimens analysed here, two (GONF-A-04 and GONF-A-11) have been recently discovered

119   by one of us (RG; GPS coordinates available on request from corresponding author) and are

120   stored at the Muséum National d’Histoire Naturelle, Paris (France) (MNHN). The two others,
                                                      6


121   from the University of Burgundy (Dijon, France) collections (GONF-G20 and GONF-G21),

122   were previously described by Gand (1987, 1989):

123           - GONF-A-04 (Fig. 2) is a slab of 266 x 200 mm with a trackway consisting of four

124   consecutive pes-manus couples (Nr. 1-4 Fig. 2), a tail impression (Nr. 5 Fig. 2) and digit drag

125   traces (Nr. 6 and 6’ Fig. 2), preserved in concave epirelief.

126           - GONF-A-11 (Fig. 3) is a slab of 298 × 331 mm with a partial trackway consisting of

127   two pes-manus couples, a right manus and a continuous tail impression, preserved in concave

128   epirelief.

129           - GONF-G20 (Fig. 4) is a slab of 479 × 364 mm and corresponds to a part (GONF-

130   G20A, in concave epirelief) and its partial counterpart (GONF-G20B, in convex hyporelief). It

131   preserves tracks of two ichnotaxa, an incomplete step cycle of Hyloidichnus (see description

132   below) and the holotype of Varanopus rigidus. The step cycle of Hyloidichnus includes a right

133   manus, a right pes-manus couple, a left pes-manus couple and continuous digit drag traces.

134           - GONF-G21A (Fig. 5) is a slab of 463 mm × 227 mm and includes a trackway

135   consisting of six consecutive pes-manus couples, continuous digit drag traces and a continuous

136   tail impression, artificial gypsum cast in convex hyporelief. GONF-G21B is a partial

137   counterpart of the original, from which GONF-G21A was casted, preserved in convex

138   hyporelief.

139

140           The specimens were examined with oblique light and photographed with a Nikon D800

141   and a Canon EOS 70D and then digitally drawn using Krita v.4.2.9 and Adobe Illustrator CC

142   2020. Quantitative and qualitative analyses were carried out following the protocols of Haubold

143   (1971), Leonardi (1987) and Gand (1987) for tetrapod footprints and Hasiotis et al. (2007) for

144   tetrapod tail impressions. Measurements were taken using ImageJ2. 3D models were obtained

145   through scans and photogrammetry. Scans were done with a triangulation-based laser scanner
                                                     7


146   (Next Engine 3D Scanner®). Close-range photogrammetry was performed using photos taken

147   with a Nikon D800 (60 mm macro lens) and a Canon EOS 70D and the software Agisoft

148   Photoscan (standard version 1.1.4) and Meshlab 2020.03 for 3D model generation and edition.

149   Paraview v. 5.8.0 was used for the generation of colour-coded height maps and contours (e.g.,

150   Falkingham, 2012; Mujal et al., 2016, 2020). The track-trackmaker correlation was performed

151   using an anatomy-based correlation (e.g., Voigt et al., 2007; Marchetti et al., 2017, 2019a;

152   Mujal and Marchetti, 2020) between the best-preserved tracks and the manus and pes skeletons

153   of Captorhinus described by Holmes (1977, 2003), including a comparison with other possible

154   producers. The track preservation was evaluated by means of morphological preservation sensu

155   Marchetti et al. (2019b). The measurements of the manus and pes parameters were done on the

156   reconstructions provided by Holmes (1977, fig. 11; 2003, fig. 7). The ratio including the body

157   length was calculated with measurements on the Captorhinus aguti skeleton from the Field

158   Museum of Chicago, specimen FMC-UC 491, Clear Fork Group (Kungurian), Texas.

159

160   4. Systematic palaeontology

161                                    Hyloidichnus Gilmore, 1927

162                                          Hyloidichnus isp.

163                                        Figs. 2-5, Tables 1-4

164   Referred specimens: GONF-A-04; GONF-A-11; GONF-G20; GONF-G21.

165   Description: The manus footprints are semi-plantigrade, the pes footprints are semi-plantigrade

166   to semi-digitigrade. Both manus and pes tracks show a medial-lateral decrease in relief,

167   although this feature is more distinct in the pes. The manus tracks are wider (22 to 76 mm) than

168   long (19.2 to 64 mm). The pes tracks are larger than the manus tracks and almost as long (28 to

169   61 mm) as wide (25 to 75 mm).
                                                     8


170          Manus and pes tracks are pentadactyl and ectaxonic (digit IV imprint is the longest).

171   The relative lengths of the manus digit impressions are I<V≤II<III<IV, the relative lengths of

172   the pes digit impressions are I<V<II<III<IV. The divarication angle of the digits I-V imprints

173   of the manus is around 140°. The divarication angle of the digits I-V of the pes is around 170°.

174   The digit impressions are straight, relatively thin and long. The imprints of digits I and IV are

175   slightly directed inwards, whereas the digit V impressions are often directed outwards. The

176   digits of both manus and pedes were clawed, as indicated by the T-shaped tip impressions (digit

177   tip bifurcation) in most of specimens; otherwise, in GONF-A-04 some pes tracks end in thin tip

178   impressions. In GONF-A-04, expulsion rims associated with both manus and pes tracks are

179   higher around the impressions of digits I, II and III. They are clearer in association with the

180   digit impressions of the manus, where they seem to follow precisely the shape of the digit.


181          The palm impressions are slightly wider (15 to 30.5 mm) than long (12 to 33 mm),

182   whereas the sole impressions are much wider (16 to 22.5 mm) than long (7 to 13 mm). The

183   palm and sole impressions are shallower just beneath the digit IV and V impressions, and the

184   proximal margin of the palm/sole can be slightly convex proximally or rectilinear.

185          The manus tracks are placed in front of the pes, at distances between 19 to 39 mm. The

186   manus tracks are closer to and rotated towards the trackway midline, the pes imprints are almost

187   parallel to or slightly rotated outwards (even though some pes tracks show an inward rotation)

188   relative to the trackway midline. The pace angulations are greater than 90° in the manus tracks

189   and slightly lower than 90° in the pes tracks. The pes imprints are distanced about 10 mm from

190   the tail impression (Nr. 5 in Fig. 2). The tail impression observed in GONF-A-04 is 156 mm

191   long and 5.5 mm wide (maximum value) and is mainly straight, with a low sinusoidal (2.7 mm)

192   shape and a high wavelength (90 mm). A regular pattern is observed with a widening of the tail

193   impression in proximity of the pes-manus couples. In GONF-G21 (Fig. 5), the tail impression

194   is straight (sinuosity of ~1 mm) without identifiable wavelength, whereas in GONF-A-11 (Fig.
                                                      9


195   3) it has a high sinuosity with an amplitude of about 52 mm and a wavelength of 123.5 mm. It

196   also shows a 1 mm deep furrow at its midline.

197          In GONF-A-04, digit drag traces characterised by four curved thin grooves are

198   preserved from the pes of couple Nr. 1 to the pes of couple Nr. 2 (Fig. 2). These impressions

199   are associated with pedal digits II, III, IV and V. Three similar grooves corresponding to digit

200   drag traces are proximal to the pes of couple Nr. 1 (Fig. 2): they are probably departing from

201   the previous pes-manus couple (not preserved on the slab). The digit drag traces do not start

202   directly from the digits and seem to be partially stepped over by the pes and the manus tracks.

203   A second similar trace with three grooves departing from the pes of the couple Nr. 2 (Fig. 2) is

204   also visible but poorly preserved. In GONF-G20 (Fig. 4), three to four digit drag traces are also

205   preserved from one pes to another and associated with the digits II to V. In GONF-G21 (Fig.

206   5), three to four digit drag traces are preserved, either from the pes to the following pes of the

207   next pes-manus couple and associated with digits II to V or three digit drag traces from the

208   manus to the following manus of the next pes-manus couple and associated with digits III to V.

209   The digit drag traces of these two specimens are larger than those on GONF-A-04.

210   Remarks: The straight digit imprints with clawed to bifurcated tip impressions in both pes and

211   manus tracks, the relative length of the digit impressions, increasing from I to IV, with digits

212   III and IV of similar length and a relatively short digit V, the high divarication angle with

213   diverging digit bases, the short palm/sole impressions, the relative size of the manus and pes

214   tracks and the trackway pattern with inwardly-directed manus footprints, are characteristic of

215   the ichnogenus Hyloidichnus Gilmore, 1927 (e.g., Haubold, 1971; Gand, 1987; Haubold et al.,

216   1995; Gand and Durand, 2006; Mujal et al., 2016; Voigt and Lucas, 2018; Marchetti et al.,

217   2020c).

218      Hyloidichnus is a common ichnogenus of the Permian, known from:
                                                 10


219   -   USA, from the Cisuralian Hermit Shale Formation (Gilmore, 1927; Marchetti et al.,

220       2020c), Robledo Mountains Formation (Voigt and Lucas, 2015) and Abo Formation

221       (Lucas et al., 2013; Voigt and Lucas, 2017);

222   -   Argentina, from the Cisuralian-Guadalupian Carapacha Formation of the Carapacha

223       Basin (Melchor and Sarjeant, 2004).

224   -   France, from the Cisuralian (Artinskian) Rabejac Formation of the Lodève Basin (Gand,

225       1987), the Cisuralian (Artinskian-Kungurian) Pélites de Saint Pierre Formation of the

226       Saint-Affrique Basin (Gand, 1993), the Cisuralian of the Blanzy-Le Creusot Basin

227       (Gand, 1981) and the Guadalupian (Wordian-Capitanian) Pelitic Formation of the Le

228       Luc Basin and Pradineaux, Mitan, Muy and La Motte Formation of the Estérel and Bas-

229       Argens basins (Heyler and Montenat, 1980; Demathieu et al., 1992; Gand and Durand,

230       2006);

231   -   Spain, from the Cisuralian (Kungurian) Sotres Formation of the Cantabrian Mountains

232       (Gand et al., 1997), the Cisuralian (Artinskian) Peranera Formation of the Pyrenean

233       Basin (Voigt and Haubold, 2015; Mujal et al., 2016), and the Cisuralian of the Balearic

234       Islands (Matamales-Andreu et al., 2019);

235   -   Italy, from the Cisuralian (Kungurian) Collio, Pizzo del Diavolo and Tregiovo

236       formations (Marchetti et al., 2013; Marchetti et al., 2015a-c; Marchetti, 2016);

237   -   Morocco, from the Cisuralian ‘Upper Formation’ of the Tiddas Basin and Koudiat El-

238       Hamra Formation of the Koudiat El-Hamra Basin (Hmich et al., 2006; Voigt et al., 2011;

239       Zouicha et al., 2020); and the Guadalupian-Lopingian Ikakern Formation of the Argana

240       Basin and Cham-el-Haoua Siltstone Formation of the Marrakech High (Voigt et al.,

241       2010; Hminna et al., 2012; Moreau et al., 2020);

242   -   Niger, from the Guadalupian-Lopingian Moradi Formation (Steyer et al., 2007; Smith

243       et al., 2015);
                                                    11


244      -   Turkey, from the Cisuralian Çakraz Formation of Anatolia (Gand et al., 2011).

245      The type ichnospecies of Hyloidichnus is H. bifurcatus Gilmore, 1927 from the Permian

246   Hermit Formation of the Grand Canyon (Arizona, USA). Gilmore (1928) also erected H. whitei

247   from the same formation. Later, Haubold (1971) proposed the combinations H. major and H.

248   minor based on the ichnospecies ‘Auxipes’ major (Heyler and Lessertisseur, 1963) and

249   ‘Auxipes’ minor (Heyler and Lessertisseur, 1963), from the Permian Rabejac Formation

250   (Lodève Basin, France). Later, two other ichnospecies were erected, H. arnhardti Haubold,

251   1973 and H. tirolensis Ceoloni, Conti, Mariotti and Nicosia, 1986 from Italy. According to

252   Haubold (2000), only H. bifurcatus and H. major are valid, although they may belong to the

253   same ichnospecies (e.g., Gand and Durand, 2006; Marchetti et al., 2020c). Moreover,

254   Gilmoreichnus (Hylopus) hermitanus (Gilmore, 1927) may also be an extramorphological

255   variation of Hyloidichnus (Marchetti et al., 2013). Therefore, the relationships between these

256   ichnotaxa need to be further investigated.

257      The footprints described here, especially GONF-A-04, are also similar to the ichnospecies

258   Varanopus rigidus Gand, 1989 (Gand and Durand, 2006), whose holotype comes from

259   Gonfaron (Gand, 1989). However, several authors considered this ichnospecies as invalid

260   because material assigned to V. rigidus from the Lodève, Saint-Affrique, Estérel and Bas-

261   Argens basins includes footprints assignable to different ichnotaxa such as other Varanopus

262   ichnospecies, Hyloidichnus and Notalacerta (Voigt et al., 2010; Marchetti et al., 2020b).

263   Pending a comprehensive revision of Hyloidichnus and its possible ichnospecies, we prefer to

264   assign the Gonfaron material studied here to Hyloidichnus isp.

265


266   5. Track-trackmaker correlation
                                                     12


267      Hyloidichnus is generally attributed to captorhinomorph eureptiles such as small

268   captorhinines (Captorhinus) or large moradisaurines (Moradisaurus) depending on the shape

269   and size of the footprints (e.g., Voigt et al., 2010). However, the Hyloidichnus-captorhinomorph

270   correlation is not based on synapomorphic characters as in other valid correlations (e.g., large

271   Ichniotherium cottae-Limnoscelis of Mujal and Marchetti, 2020; small I. cottae-Diadectes and

272   I. sphaerodactylum-Orobates of Voigt et al., 2007; Amphisauropus-Seymouria of Marchetti et

273   al., 2017; Pachypes ollieri-Macroleter of Marchetti et al., 2020a). This is due to the lack of

274   studies on well-preserved Hyloidichnus footprints correlated to detailed skeletal reconstructions

275   in a morpho-functional perspective. The good preservation of the footprints from Gonfaron

276   allows us to investigate and to define this track-trackmaker correlation:

277          Anamniotes such as diadectomorphs (trackmakers of Ichniotherium), seymouriamorphs

278   (Amphisauropus) and amphibians such as eryopoids (Limnopus) have well-ossified, short and

279   distally rounded digits as well as ossified and packed carpus and tarsus, resulting in broad palm

280   and sole impressions (Voigt, 2005; Voigt et al., 2007; Marchetti et al., 2017; Mujal and

281   Marchetti, 2020). This is inconsistent with the clawed and bifurcated digit impressions and the

282   short palm and sole impressions of Hyloidichnus.

283          Parareptiles such as pareiasaurs (trackmakers of large Pachypes dolomiticus) are too

284   large to fit with Hyloidichnus, whereas smaller pareiasauromorphs (trackmakers of Pachypes

285   ollieri) have closely-packed digits and overlapping metatarsals and metacarpals, which are

286   inconsistent with the radiating digit imprints and the large digit divarication angle of

287   Hyloidichnus (Marchetti et al., 2020a). Also, their tails were short, whereas continuous tail

288   impressions are commonly observed in trackways of Hyloidichnus.

289          The Hyloidichnus footprints from Gonfaron show a wide size range in manus (19 to 49

290   mm long) and pes tracks (28 to 61 mm long) as well as in trackway parameters (Tabs. 1-3). The

291   body trunk length, based on the glenoacetabular distance, also greatly varies in size, between
                                                     13


292   50 mm to 184 mm (Tab. 4). This suggests different taxa as trackmakers, or individuals of the

293   same taxon at different ontogenetic stages. Among captorhinids, relatively complete

294   appendicular elements are known from Captorhinus aguti Cope, 1895 (Holmes, 1977),

295   Labidosaurus hamatus Willinston, 1910, Captorhinikos choazensis Olson, 1962, Moradisaurus

296   grandis de Ricqlès and Taquet, 1982 (O’Keefe et al., 2005) and Romeria prima (Clark and

297   Caroll, 1973). Even though the skeletal remains of captorhinids date back to the Artinskian–

298   Kungurian and the Gonfaron footprints to the Wordian (showing a discrepancy of at least 4

299   Myr), the wide temporal range of Hyloidichnus trackways in the fossil record supports this

300   investigation (Fig. 6). Captorhinus aguti fits with the size of the footprints of GONF-A-04. This

301   taxon has been correlated to Varanopus by Fichter (1983) and Voigt (2005), mostly based on

302   overall morphology and proportions. Also, the mesotarsal joint described by Holmes (2003)

303   was considered an explanation for the medial-lateral decrease in relief of the Varanopus pes

304   imprint. Nevertheless, the medial-lateral decrease in relief occurs also in Hyloidichnus, and

305   Varanopus shows parallel and overlapping digit base impressions, a feature that is characteristic

306   of parareptiles (Lee, 1997) and is not in agreement with the radiating and separated metatarsals

307   of articulated skeletons of Captorhinus aguti (Holmes, 2003 figs. 7-8). Therefore, we consider

308   the attribution of C. aguti to Varanopus as currently not well supported.

309          Captorhinus and Hyloidichnus share:

310      -   pentadactyl manus and pes;

311      -   similar relative lengths of the digits (I<V≤II<III<IV for the pes and the manus) (Tabs.

312          1, 2, 5);

313      -   radiating and non-overlapping metatarsals/proximal digit imprints;

314      -   clawed digit tips;

315      -   similar length ratios between: 1) pes/manus length and metatarsus/metacarpus length

316          and 2) pes/manus foot length and sole/palm impression length (Tabs. 1, 2, 5);
                                                     14


317     -   similar length/width ratios of the pes (phalanges + tarsus) and of the manus (phalanges

318         + carpus) (<1 for the manus and ~1 for the pes) (Tabs. 1, 2, 5).

319

320         The carpus of Captorhinus comprises 11 elements: the radiale, the intermedium, the

321   ulnare, the pisiform, two centralia and five distal carpals (Fig. 7). The joint between the radius

322   and the radiale is flat, suggesting no flexion between them: the radiale may act as an extension

323   of the radius. The convex joint between the radiale and the lateral centrale permitted a

324   significant medio-lateral rotation. In addition, the ulnare and the intermedium form a single

325   structural unit with a flat joint oriented like a dorsal process (Holmes, 1977), causing probably

326   a concave proximal margin of the palm: this fits with the Hyloidichnus manus impressions

327   (Fig. 7). The joints are flat between (1) the ulnare and the ulna, (2) the ulnare and the

328   intermedium, and (3) the ulnare and distal carpals 4 and 5. The ulnare and the intermedium

329   contact both the medial centrale, and the latter contacts only marginally the radius-radiale

330   complex. Thus, the lateral centrale is the element of the carpus that experienced the most

331   tension during movement; this element supports both the radius-radiale complex and the ulna-

332   intermedium-ulnare-medial centrale complex (Holmes, 1977; Marchetti et al., 2017). This

333   also fits with the basal part of the Hyloidichnus digit I, II and III imprints of the manus, where

334   the footprint is more deeply-impressed (Fig. 7). The joints between the distal carpals are also

335   flat, allowing very little independent movements; this is consistent with the straight and

336   regularly impressed manus digit bases of Hyloidichnus (Figs. 2, 7). A distal process behind

337   metacarpal 3 prevented too much flexion of digits III and IV (Holmes, 1977), suggesting a

338   more stable position and rigid appearance compared to the imprints of digits I and II of

339   Hyloidichnus. The phalanges of the manus of Captorhinus are proportionally thicker and

340   shorter than those of the pes, with significant ossification at the joints (increasing the distance

341   between the tendon and the joint and thus the flexibility of the digit muscles) (Holmes, 1977).
                                                       15


342    This is also suggested in Hyloidichnus, which has relatively shorter and thicker digit imprints

343    and overall deeper manus imprints (with marked impressions of digits I and II) compared to

344    the pes imprints, as also observed by the higher expulsion rims in GONF-A-04 (Fig. 2).

345          The tarsus of Captorhinus comprises eight elements: the calcaneus, the astragalus, the

346   centrale and five distal tarsals (Fig. 7). The sole impression of Hyloidichnus is smaller than the

347   palm impression, probably due to less distal elements of the tarsus compared to the carpus: this

348   is consistent with Captorhinus. The centrale, calcaneus, and astragalus are closely nested,

349   suggesting little movements between these elements. The calcaneus has flat joints with the

350   distal tarsals 4 and 5, while the centrale has convex joints with the distal tarsals 1, 2 and 3

351   (Holmes, 2003). Thus, the flexion is mainly supported by the centrale and occurs at the level of

352   the distal tarsals 1, 2 and 3, which is consistent with the deeper impressions of the digits I to III

353   of the pes in Hyloidichnus (Fig 7). The phalanges of the Captorhinus pes are long and thin,

354   corresponding to the impressions of the pes digits of Hyloidichnus. The pes joints are more

355   ossified than the manus joints which could explain the straight digit impressions of the pes and

356   the lack of mobility relative to the manus. The unguals of Captorhinus (and captorhinormophs

357   in general) are poorly preserved. However, they seem thinner and shorter in the pes (Holmes,

358   2003) than in the manus (Holmes, 1977). This seems consistent with the thin and shallow pes

359   digit drag traces preserved within the trackways of Hyloidichnus (Figs. 2, 4 and 5). They greatly

360   differ from those of Amphisauropus that leave deeper and larger digit drag traces

361   indistinguishable among each other, probably due to the thicker digits and the absence of

362   unguals. Holmes (2003), contra Schaeffer (1941), Fox and Bowman (1966) and Sumida (1989),

363   hypothesised a ‘mesotarsal’ joint in C. aguti, the dorsiflexion of which suggests a more

364   pronounced medial impression of the sole at the level of the tarsals 1 to 3; this is again consistent

365   with Hyloidichnus.
                                                    16


366          Labidosaurus and Captorhinikos are larger captorhinids than C. aguti. They also show

367   similar manus and pes anatomy with, however, shorter and stubbier distal pedal phalanges,

368   resulting in a relatively shorter and wider pes (Sumida, 1989). Such morphology might be

369   coherent with GONF-G21 and GONF-A-11, where the pes impressions are wider than long.

370   So, Labidosaurus and Captorhinikos may be considered as potential trackmakers for larger

371   Hyloidichnus tracks. Recent phylogenetic studies of captorhinomorphs place Captorhinikos as

372   the basalmost Moradisaurinae (Modesto et al., 2019), but its phylogenetic position has often

373   changed, within or out of the Moradisaurinae (de Ricqlès, 1984; Reisz et al., 2011, 2015;

374   Modesto et al., 2014; Liebrecht et al., 2017; Cisneros et al., 2020). The clade Labidosaurus +

375   Moradisaurinae is well-supported by all these phylogenetic analyses. Inclusive or stem-

376   Moradisaurinae, with autopodials similar to those of Captorhinus, could then be potential

377   trackmakers of relatively large Hyloidichnus tracks.

378          The pes of Moradisaurus is foreshortened and extremely robust compared to other

379   captorhinids. It has a massive astragalus, larger than that of other captorhinids, and a derived

380   calcaneum (O’Keefe et al., 2005). Nevertheless, the tarsus arrangement is the same as that of

381   other captorhinids, with the only exception being the lack of the distal tarsal 5. The body of

382   Moradisaurus is extremely large (its weight was probably more than 300 kg), being inconsistent

383   with the size of the studied tetrapod footprints. The metatarsals and phalanges of the pes,

384   although similar to those of other captorhinids, suffered a proximo-distal compression: this may

385   reflect a shortening of the pes digit imprints, but this is not observed in the studied Gonfaron

386   specimens. The position of the femur may also restrict the rotation and movement of the

387   hindlimb during the step cycle, suggesting an antero-posterior movement only (O’Keefe et al.,

388   2005). This movement of the hindlimb may not leave digit drag traces as curved as seen in the

389   studied Hyloidichnus trackways. Therefore, Moradisaurus and derived moradisaurines may not

390   be the trackmakers of the Hyloidichnus tracks from Gonfaron.
                                                    17


391


392   6. Locomotor reconstruction

393          The locomotor cycle of the hindlimb of Captorhinus aguti reconstructed by Holmes

394   (2003), in which the outer digits are first to leave the substrate during the kick-off phase, is

395   generally consistent with an increased medial relief of the pes impression. This is regularly

396   observed in Hyloidichnus (e.g., Gand, 1987; Mujal et al., 2020), including small Hyloidichnus

397   tracks as those observed on GONF-A-04. It has to be noted that the pes orientation of

398   Captorhinus proposed by Holmes (2003) is slightly different from that of Hyloidichnus. The

399   absence of a cruro-tarsal joint in Captorhinus is in agreement with the semi-plantigrady of the

400   impressions. The shortened impression of the sole then resulted from the impression of the

401   distal tarsus and proximal metatarsus (Holmes, 2003; Voigt, 2005). Hyloidichnus manus and

402   pes imprints rarely overlap, and the position of the pes-manus couples is consistent with a

403   sprawling locomotion observed in many present-day tetrapods (Schaeffer, 1941; Brinkman,

404   1980, 1981; Pardian and Olsen, 1984) and in trace fossils such as Amphisauropus and its

405   possible trackmaker Seymouria (Marchetti et al., 2017). Holmes (2003) reconstructed the stance

406   phase of the pes of C. aguti based on the hindlimb skeleton. However, the following pes swing

407   phase could not be reconstructed due to the absence of some elements. The locomotor cycle of

408   Captorhinus has been reconstructed mainly according to skeletal elements, while footprints are

409   rarely used (Holmes, 2003 used them only for imposition of constraints). Reconstruction of the

410   pes swing phase was performed for Seymouria thanks to exceptionally preserved

411   Amphisauropus tracks and trackways, including continuous digit drag traces (Marchetti et al.,

412   2017). Tracks give important information on the different steps of the locomotion: how the

413   manus/pes leave, arrive and rest on the sediment (stance phase). While other aspects of the step

414   cycle, when the manus/pes are off the substrate (swing phase), are less certain, the presence of

415   digit drag traces allows a rather precise reconstruction of this part of the locomotion (which
                                                     18


416   cannot be inferred through tracks nor anatomical analyses). The trackways analysed here are

417   consistent with Holmes’ analysis on the stance phase, but not demonstrative of all of it. The

418   well-preserved Gonfaron footprint material including trackways and continuous digit drag

419   traces, coupled with the work of Holmes (2003), allows us to propose a reconstruction of the

420   swing phase of Captorhinus, for both the pes and the manus.

421


422   6.1. Forelimb step cycle

423          Considering the manus tracks, the deep impression of the digits I, II and III suggests an

424   important role of this manus portion during the stance phase to support the weight of the body.

425   However, to accommodate the change in body weight support to the pre-axial side, digits I, II

426   and III moved from a position almost parallel to the trackway midline to a position almost

427   perpendicular (as represented by the inwards rotated manus tracks). This resulted in more

428   excursion of the humerus than that of the femur to accompany this re-orientation of the digits

429   during the midstride. The position of the manus was then nearly perpendicular relative to the

430   substrate with the digits positioned high above it at the end of the stance phase. Digit dragging

431   is generally more frequent in hindlimbs than forelimbs (Willey et al., 2004; Farlow et al., 2017).

432   However, crocodylians and other reptiles such as the Komodo monitors leave manus digit drag

433   traces because of their lateral excursion of the humerus, almost at 90° to the body trunk (Padian

434   and Olsen, 1984). The absence or shallow impression of digit drag traces associated with the

435   Hyloidichnus manus imprints could be due to the more important antero-posterior movement

436   of the elbow compared to crocodylians, which increased the length of the stride with more

437   flexion that prevented digit dragging (Holmes, 1977). This would be in agreement with a more

438   advanced forelimb posture of captorhinids compared to crocodiles. A lifted position of the

439   manus at the end of the stance phase could also be an explanation for the scarcity of manus digit

440   drag traces in Hylodichnus. The stance phase of the manus is consistent with the stance phase
                                                     19


441   reconstructed for the pes by Holmes (2003), with an active role of the digits I to III. The swing

442   phase of the manus can be reconstructed thanks to the well-preserved and continuous digit drag

443   traces of GONF-G21 (Fig. 8-I).

444      -   Beginning of the swing phase (Fig. 8-IA): all digits of the manus have left the substrate.

445          The complex radius/ulna undergoes a rotation of about 90°, so that only the tips of the

446          digits IV and V touch the substrate, leaving continuous digit drag traces. However, the

447          digit drag traces are very thin (and often absent in Hyloidichnus trackways), suggesting

448          a position of the digits well above the substrate. The humerus reaches its maximum

449          height with an excursion over 90°;

450      -   Early swing phase (Fig. 8-IB): almost immediately, the humerus undergoes a rotation

451          of 20° anteriorly, allowing the generation of digit drag traces of digit III (as seen in

452          GONF-G21). The angle with the body axis stays the same;

453      -   Midstride (Fig. 8-IC): during the midstride, the position of the manus relative to the

454          substrate and of the humerus relative to the body axis does not move (or very slightly)

455          resulting in the absence of an arcuate movement of the digit drag traces of the manus

456          compared to those of the pes;

457      -   End of the swing phase (Fig. 8-ID): the humerus reaches its lower height with an

458          excursion below 90°. The manus shifts from an inclined to a parallel orientation towards

459          the substrate. This is permitted by a re-orientation of the carpus. The manus is in its

460          medial-most position, leaving the next imprint.

461


462   6.2. Hindlimb step cycle
463          The specimens GONF-A-04 and GONF-G21 show continuous digit drag traces, which

464   allow to describe the movement of the pes from one step to the following one (Fig. 8-II). Most

465   of the digit drag traces are represented by three to four curved lines that depart from the pedal
                                                      20


466   digit impressions II, III, IV and V. The curved line from digit V impression is the longest. It

467   decreases in length from impressions of digit IV to II. This suggests that the digit I and II left

468   the substrate first, while digits III to V kept their position on the substrate, as shown during the

469   midstride and end of stance phase by Holmes (2003). The swing phase is then similar to that of

470   a living alligator where “early in the hindlimb protraction, the distal end of the femur undergoes

471   slight abduction, relative to its position at the end of the stance phase, as it swings forward, the

472   extent of abduction varying among individual alligators” (Farlow et al., 2017, p. 29). This is

473   consistent with the number of digit drag traces in Hyloidichnus trackways, which ranges

474   between two and four. The digit drag traces in Hyloidichnus and crocodilians trackways (Farlow

475   et al., 2017) consist of similar curved lines, suggesting a similar movement.

476      -   Beginning of the swing phase (Fig. 8-IIA): all digits of the pes have left the substrate.

477          The complex tibia/fibula undergoes a rotation of about 90°. Only the tips of digits IV

478          and V touch the substrate, leaving the continuous digit drag traces. The frequency and

479          sharpness of the pes digit drag traces amongst Hyloidichnus trackways suggest that the

480          femur is less excursed than the humerus, at 90° relative to the body axis;

481      -   Early swing phase (Fig. 8-IIB): the femur undergoes an anterior rotation probably of

482          20°, corresponding to the best-preserved digit drag traces of digits III, IV and V.

483          Simultaneously to this rotation, the angle of the femur with the body axis diminishes.

484          These re-orientations are marked by an arcuate movement of the digit drag traces

485          towards the outside of the trackway midline (as seen in GONF-A-04). This is not seen

486          in the manus digit drag traces, which are more regular;

487      -   Midstride (Fig. 8-IIC): the femur continues its rotation (of about 60° at the end of the

488          midstride) leading to the drag trace of digit II. However, the tip of the digit II pass

489          slightly over the substrate and may not always leave drag traces. During the midstride,

490          digit drag traces of the pes are more rectilinear than those of the manus: this suggests
                                                     21


491          that the femur has a less prominent excursion than that of the humerus. It is already

492          below 90° relative to the body axis;

493      -   End of the swing phase (Fig. 8-IID): the femur reaches its lower height. The pes shifts

494          to a parallel orientation towards the substrate. This is permitted by a re-orientation of

495          the tarsus. The pes is in its medial-most position, leaving the next imprint.

496


497   6.3. Tail impression and gait

498          Deep and wide tail impressions and continuous digit drag traces are rarely observed in

499   Hyloidichnus (Gand, 1987). Similar tail impressions and digit drag traces have been observed

500   in the Permian ichnogenus Erpetopus, the trackmaker of which is interpreted to be a small

501   reptile (Bernardi and Avanzini, 2011). The exceptional preservation of these features on the

502   studied specimens may be due to differences in gait and substrate, which was probably more

503   water-saturated and muddy (Marchetti et al., 2017, 2019b). Contrary to Amphisauropus, where

504   well-impressed and continuous tail traces are associated with digit drag traces and present a

505   significant sinuosity (Marchetti et al., 2017), the Hyloidichnus tail impression shows generally

506   a very low sinuosity, except for GONF-A-11. Variations in the sinuosity of the tail impression

507   are observed in present day tetrapods such as crocodilians (Farlow et al., 2017). The nearly

508   rectilinear tail impression in Hyloidichnus suggests that its captorhinomorph trackmaker had a

509   lower body sway and spine flexibility than those of the seymouriamorph trackmaker of

510   Amphisauropus (Berman et al., 2000; Marchetti et al., 2017). Although, as in Amphisauropus

511   trackways, an arcuate movement of the scratches suggests a rather marked lateral movement of

512   the body (Ashley-Ross, 1994; Marchetti et al., 2017). Captorhinomorph eureptiles, alongside

513   seymouriamorphs, could have had a more derived body trunk flexibility in comparison with

514   earlier tetrapods (Pierce et al., 2013; Marchetti et al., 2017). The body and tail of

515   captorhinomorphs might also have been more lifted than those of seymouriamorphs. A more
                                                    22


516   distal part part of the tail may have been impressed (which was thinner and rectilinear because

517   less subject to the body sway). A subrectilinear tail impression has also been observed in

518   Dimetropus, correlated to pelycosaur-grade synapsids (Voigt, 2005). Variations in the width of

519   the tail impression in correspondence with pes-manus couples suggest a slight change in the

520   orientation of the tail during walking, or a slight lowering-rising of the body (Avanzini and

521   Renesto, 2002). Although such slight variations are seen in GONF-A-04, they are not marked

522   enough to suggest a gait with “brief regular spurts” (Avanzini and Renesto, 2002 p. 57). The

523   tail impression of GONF-A-04 is continuous, suggesting that vertical movement was almost

524   absent (Bernardi and Avanzini, 2011; Marchetti et al., 2017). The tail impression is wide and

525   well-marked in specimens GONF-A-04, GONF-A-11 and GONF-G21. This suggests a

526   relatively large and long tail, touching the ground during progression. The steepness and

527   location of the expulsion rims of the tail impression, opposite to the pes-manus couples and in

528   its inner side, (Figs. 9, 10) suggest a lateral movement in the same direction of the head and

529   opposite to the trunk. Such lateral movement of the body is also seen in seymouriamorphs

530   (Marchetti et al., 2017).

531          However, there are slight differences in the relative depth pattern between trackways

532   GONF-A-04 and GONF-A-11 (Fig. 9). This could be linked to slightly different gait as well as

533   different rheological conditions of the substrate. The two trackways are very different in size

534   (pes and manus imprints in GONF-A-11 are nearly two to three times larger than those of

535   GONF-A-04). As discussed above, different captorhinomorph taxa as trackmakers cannot be

536   discarded, but the manus and pes skeletons of captorhinomorphs are extremely conservative

537   and cannot explain such differences. An evolution of the locomotion through ontogenetic stages

538   of the skeleton in captorhinomorph would also be unlikely. A limited skeletal ossification

539   indeed characterizes many eureptilian neonates (age class that express attributes mostly

540   influenced by pre-parturitive development environment such as the egg or the egg-nest,
                                                     23


541   Morafka et al., 2000), but the size of the trackmaker of GONF-A-04 (50 mm) compared to size

542   of the trackmaker of GONF-A-11 (175 mm), would rather indicate a juvenile form.

543   Unfortunately, juvenile captorhinomorph skeletons are not known. The following discussion

544   will therefore concentrate on gait and substrate. High speed is known to affect the trackways of

545   reptiles and amphibians, potentially causing: primary overstepping of the pes to the manus,

546   absence of a tail impression, high pace angulation, high stride length/foot length (SL/FL) ratio,

547   position of the digit V of the pes nearly perpendicular to the trackway midline, high rotation of

548   the manus and pes in respect to the trackway midline and a Y-shaped digit tip impressions of

549   the pes imprints (Peabody, 1956; Irschick and Jayne, 1999; Dietrich and Gardner, 2004;

550   Dietrich, 2008). Both, GONF-A-04 and GONF-A-11 show a tail impression inconsistent with

551   high-speed locomotion but rather a walking-type locomotion (Avanzini and Renesto, 2002;

552   Pierce et al., 2013; Marchetti et al., 2017; Mujal et al., 2017). However, it is not incompatible

553   with a higher speed of one trackmaker relative to the other. The GONF-A-04 trackway has a

554   pace angulation lower than that of GONF-A-11, but a higher SL/FL ratio with a higher inter-

555   pes and inter-manus distance; the rotation of the manus and pes tracks in respect to the trackway

556   midline is higher in GONF-A-04 than in GONF-A-11 with a higher divarication of the digit V

557   (Tables 1, 2). In addition, the morphology of the digit tips of GONF-A-04 present a Y-shaped

558   impression but not in GONF-A-11, where the digit tip impression is directed laterally,

559   suggesting a slower gait (Peabody, 1956). A slightly higher locomotion gait of the trackmaker

560   of GONF-A-04 is then plausible. GONF-A-04 shows a rectilinear tail impression, whereas

561   GONF-A-11 shows a sinuous tail impression. In GONF-G-21, which also presents a nearly

562   rectilinear tail impression, the pace angulations, SL/FL ratio, the inter-manus distance and

563   divarication of the pes are similar to those of GONF-A-04 (Tables 1, 3, 4), although the footprint

564   and supposed trackmaker size are closer to those of GONF-A-11. So, a lower tail sinuosity may

565   be coherent with a slightly higher locomotion gait. Moreover, the substrate of GONF-A-04 and
                                                     24


566   GONF-A-11 was quite different: alongside GONF-A-04 are found traces of bioturbation and

567   burrows and the tracks present high expulsion rims and a more defined shape of the imprints,

568   digit drag traces and tail impressions: this suggests a water-saturated substrate, with sufficient

569   cohesiveness to register small details. In GONF-A-11, the substrate would have been drier, as

570   none of those features are recovered from the slab. Nevertheless, the relative depth pattern of

571   both manus and tracks of both trackways seem to show a medial-median functional prevalence.

572   This is consistent with the fact that rheology does not affect the trackmaker´s functional

573   morphology: the relative depth patterns are the same for each ichnotaxon across different

574   substrates, Mujal et al., 2020). As in the studied material, speed also seems not to have effect

575   on the functional prevalence of the trackmaker. Trackways showing a conclusive high gait

576   should be analysed to eventually confirm this hypothesis.

577


578   7. Conclusions

579          The discovery of new exceptionally well-preserved specimens from Gonfaron (Var,

580   France), together with the revision of those from the Burgundy University collections, allows

581   to precise the ichnological description of Hyloidichnus and to better understand its potential

582   trackmakers, as well as their locomotion style. The correlation between Hyloidichnus and

583   Captorhinus (and captorhinomorphs with similar autopodia) is hence proposed for the first time

584   in a synapomorphy-based approach.

585          The exceptional preservation of continuous digit drag traces and tail impressions

586   allowed a detailed review of the captorhinomorph locomotion. The stance phase of the pes of

587   captorhinomorphs is coherent with Hyloidichnus trackways, suggesting a greater role than

588   previously thought of the manus during the propulsion of the body. The swing phase of both

589   manus and pes for captorhinomorphs has been reconstructed for the first time. A possible
                                                      25


590   marked body trunk flexibility has been highlighted in captorhinomorphs, compared with other

591   Permian tetrapods. However, the tail impression, except for one specimen, suggests moderate

592   body sway. A new hypothesis of a “lifted body” in captorhinomorphs is thus proposed.

593           These trackways finally permitted detailed observations on different locomotion gait.

594   Even though high speed locomotion is not represented, slight changes in speed observed from

595   the Hyloidichnus trackways parameters have been documented. Two interesting outcomes,

596   which need further investigation, are: 1) the correlation between the shape of the tail impression

597   in response to the locomotion gait and 2) the absence of impact of speed and rheology on the

598   functional prevalence of autopodia.

599           Gonfaron is a locality recently re-opened in the Permian of France (Var). Together with

600   new insect specimens recently described, these new discoveries also show that Gonfaron is an

601   important locality to better understand the Permian faunas and their associated

602   palaeoenvironements of France and Europe. This acquires even more importance due to the

603   middle Permian age of this site, since low-palaeolatitude fossil-bearing localities of this age are

604   rare.

605


606   Acknowledgements

607   We thank the Mayor of Gonfaron for authorizations to collect fossils, the Maison de la Nature

608   (Les Mayons) for accommodations, I. van Waveren (Natural History Museum Leiden), J.

609   Fortuny (ICP), Stanislav Stamberg (Eastern Bohemian Museum Czech Republic), G. Lemaitre

610   (Phylogenia Association), G. Gand (Dijon University), M. Durand (Nancy University), T.

611   Arbez (Ottawa University), and the excavation team (M. Raymond, M. Ughetto, X. Pouly, V.

612   Blondel, M. Denis) for their help in the field. We thank E. Fara and J. Thomas for the access to

613   the Dijon University collection. This research was supported by the interdisciplinary programs
                                                    26


614   RedPerm I and II, ATM MNHN, LabEx, and Projet Fédérateur MNHN attributed to J-S.S.,

615   R.G, and A.N. The Alexander von Humboldt Foundation and the Bundesministerium für

616   Bildung und Forschung (BMBF) for the BROMACKER Project 2020-2025 financed LM. This

617   work received support from the CERCA programme (ICP) from the Generalitat de Catalunya.

618   We acknowledge the reviews of Spencer G. Lucas, Sebastian Voigt and Hendrik Klein and of

619   the editors Prof. Howard Falcon-Lang and Prof. Thomas J. Alego, which helped to improve a

620   previous version of the manuscript.

621


622   References

623   Ashley-Ross, M., 1994. Hindlimb kinematics during terrestrial locomotion in a salamander

624   (Dicamptodon tenebrosus). J. Exp. Biol. 193, 255-283.

625   Avanzini, M., Renesto, S., 2002. A review of Rhynchosauroides tirolicus Abel, 1926

626   ichnospecies (Middle Triassic: Anisian-Ladinian) and some inferences on Rhynchosauroides

627   trackmaker. Riv. Ital. Paleontol. S. 108, 51–66.

628   Berman, D.S., Henrici, A.C., Sumida, S.S., Martens, T., 2000. Redescription of Seymouria

629   sanjuanensis (Seymouriamorpha) from the Lower Permian of Germany based on complete,

630   mature specimens with a discussion of paleoecology of the Bromacker locality assemblage. J.

631   Vertebr. Palentol. 20, 253-268.

632   Bernardi, M., Avanzini, M., 2011. Locomotor behavior in early reptiles: insights from an

633   unusual Erpetopus trackway. J. Paleontol. 85, 925-929.

634   Brinkman, D., 1980. The hind limb step cycle of Caiman sclerops and the mechanics of the

635   crocodile tarsus and metatarsus. Can. J. Zool. 58, 2187-2200.
                                                     27


636   Brinkman, D., 1981. The hindlimb step cycle of Iguana and primitive reptiles. J. Zool. 193, 91-

637   103.

638   Buchwitz, M., Voigt, S., 2018. On the morphological variability of Ichniotherium tracks and

639   evolution of locomotion in the sistergroup of amniote. PeerJ 6, e4346.

640   Butts, E., 1891. Recently discovered foot-prints of the Amphibian age, in the Upper Coal

641   Measure group of Kansas City, Missouri. Kansas City Scientist 5, 17-19.

642   Ceoloni, P., Conti, M.A., Mariotti, N., Nicosia, U., 1986. New Late Permian tetrapod footprints

643   from southern Alps. Mem. Soc. Geol. Ital. 34, 45-66.

644   Cisneros, J.C., Angielczyk, K., Kammerer, C.F., Smith, R.M.H., Fröbisch, J., Marsicano, C.A.,

645   Richter, M., 2020. Captorhinid reptiles from the lower Permian Pedra de Fogo Formation, Piauí,

646   Brazil: the earliest herbivorous tetrapods in Gondwana. PeerJ 8: e8719. doi: 10.7717/peerj.8719

647   Clark, J., Caroll, R.L., 1973. Romeriid reptiles from the Lower Permian: Harv. Univ., Mus.

648   Comp. Zool. Bull. 144, 353-407.

649   Cope, E.D., 1895. The reptilian order Cotylosauria. Proc. Am. Philos. Soc. 34, 436-457.

650   Demathieu, G., Gand, G., Toutin-Morin, N., 1992. La palichnofaune des bassins permiens

651   provençaux. Geobios 25, 19-54.

652   Dietrich, C., 2008. Millions of reptile tracks – Early to Middle Triassic carbonate tidal flat

653   migration bridges of Central Europe – reptile immigration into the Germanic Basin.

654   Palaeogeogr. Palaeoclimatol. Palaeoecol. 259, 410-423.

655   Dietrich C.G., Gardner, A.S., 2004. Lacertilian trackway experiments in the carbonate tidal flats

656   of All Dabb’iya, western Abu Dhabi, U.A.E. Triubulus 13, 23-28.
                                                    28


657   Durand, M., 2008. Permian to Triassic continental successions in southern Provence (France):

658   an overview. Boll. Soc. Geol. Ital. 127, 697-716.

659   Falkingham, P.L., 2012. Acquisition of high resolution three-dimensional models using free,

660   open-source, photogrammetric software. Palaeontol. Electronica 15, 15.

661   Farlow, J.O., Robinson, N.J., Kumagai, C.J., Paladino, F.V., Falkingham, P.L., Elsey, R.M.,

662   Martin, A.J., 2017. Trackways of the American crocodile (Crocodylus acutus) in Northwestern

663   Costa      Rica:     implications     for     crocodylian        ichnology.   Ichnos.   doi:

664   10.1080/10420940.2017.1350856.

665   Fichter, J., 1983. Tetrapodenfährten aus dem saarpfälzischen Rotliegenden (Ober-Karbon-

666   Unter-Perm; SW-Deutschland), Teil II: Die Fährten der Gattungen Foliipes, Varanopus,

667   Ichniotherium, Dimetropus, Palmichnus, Phalangichnus, cf. Chelichnus, cf. Laoporus und

668   Anhomoiichnium. Mainz. Naturwiss. Arch. 21, 125-186.

669   Fox, R.C., Bowman, M.C., 1966. Osteology and relationships of Captorhinus aguti (Cope)

670   (Reptilia: Captorhinomorpha). Univ. Kansas Paleontol. Contrib. Vertebrata. 11, 1-79.

671   Gand, G., 1981. Découverte de traces de reptiles cotylosauriens dans le Permien du bassin de

672   Blanzy-Le Creusot (Saône-et-Loire, France). Une étape dans la succession des palichnofaune

673   de vertébrés tétrapodes. CR. Acad. Sci. Ser III. 292, 163-167.

674   Gand, G., 1987. Les traces de vertébrés tétrapodes du Permien français, paléontologie,

675   stratigraphie, paléoenvironnements. (PhD thesis). Université de Bourgogne, Dijon (341 pp.)

676   Gand, G., 1989. Varanopus rigidus : une nouvelle ichnoespèce de vertébré tétrapode du

677   Permien français attribuable à des captorhinomorpha ou à des procolophonoidea. Geobios 22,

678   277-291.
                                                     29


679   Gand, G., 1993. La palichnofaune de vertébrés tétrapodes du bassin permien de Saint-Affrique

680   (Aveyron) : comparaisons et conséquences stratigraphiques. Géol. France 1, 41-56.

681   Gand, G., Garric, J., Demathieu, G., Ellenberger, P., 2000. La palichnofaune du Permien

682   Supérieur du Bassin de Lodève (Languedoc – France). Palaeovertebr. 29, 1-81.

683   Gand, G., Durand, M., 2006. Tetrapod footprint ichno-associations from French Permian

684   basins. Comparisons with other Euramerican ichnofaunas. In: Lucas, S.G., Cassinis, G.,

685   Schneider, J.W. (Eds.), Non-Marine Permian Biostratigraphy and Biochronology. Geological

686   Society of London, Special Publications 265, pp. 157-177.

687   Gand, G., Kerp, H., Parsons, C., Martinez-García, E., 1997. Palaeoenvironmental and

688   stratigraphic aspects of animal traces and plant remains in Spanish Permian red beds (Peña

689   Sagra, Cantabrian Moutains, Spain). Geobios 30, 295-318.

690   Gand, G., Tüysüz, O., Steyer, J.-S., Allain, R., Sakinç, M., Sanchez, S., Şengor, A.M.C., Sen,

691   S., 2011. New Permian tetrapod footprints and macroflora from Turkey (Çakraz Formation,

692   northwestern Anatolia): biostratigraphic and palaeoenvironmental implications. CR Palevol 10,

693   617-625.

694   Garrouste, R., Lapeyrie, J., Steyer, J.-S., Giner, S., Nel, A., 2017. Insects in the Red Middle

695   Permian of Southern France: first Protanisoptera (Odonatoptera) and new Caloneurodea

696   (Panorthoptera), with biostratigraphical implications. Hist. Biol. 30, 546-553.

697   Gilmore, G.W., 1927. Fossil footprints from the Grand Canyon: Second contribution. Smithson.

698   Misc. Collect. 80, 1-78.

699   Gilmore, G.W., 1928. Fossil footprints from the Grand Canyon: Third contribution. Smithson.

700   Misc. Collect. 80, 1-21.
                                                    30


701   Hasiotis, S.T., Platt, B.F., Hembree, D.I., Everheart, M.J., 2007. The trace-fossil record of

702   vertebrates. In: III Miller (Ed.), Trace Fossils: Concepts, Problems, Prospects. Elsevier, pp.

703   196-218.

704   Haubold, H., 1971. Ichnia amphibiorum et Reptiliorum fossilium. In : Kuhn, O. (Ed.),

705   Handbuch der Paläoherpetologie. 18 Gustav Fischer Verlag, Stuttgart and Portland (124 pp.).

706   Haubold, H., 1973. Die Tetrapodenfährten aus dem Perm Europas. Freiberger Forsch.-Hefte C.

707   285, 5-55.

708   Haubold, H., 2000. Tetrapodenfährten aus dem Perm – Kenntnisstand und progress 2000.

709   Hallesches Jahrb. Geowiss. B. 22, 1-16.

710   Haubold, H., Lockley, M.G., Hunt, A.P., Lucas, S.G., 1995. Lacertoid footprints from Permian

711   dune sandstones, Cornberg and De Chelly Sandstones. N. M. Mus. Nat. Hist. Sci. Bull. 6, 235-

712   244.

713   Heyler, D., Montenat, C., 1980. Traces de pas de vertébrés du Permien du Var. Intérêt

714   biostratigraphique. Bull. Mus. Natl. Hist. Nat. 4C. 4, 407-451.

715   Heyler, D., Lessertisseur, J., 1963. Pistes de tétrapodes permiens dans la région de Lodève,

716   Hérault. Mém. Mus. Natl. Hist. Nat. S. C 11, 125-220.

717   Hmich, D., Schneider, J.W., Saber, H., Voigt, S., El Wartiti, M., 2006. New continental

718   Carboniferous and Permian faunas of Morocco: implications for biostratigraphy,

719   palaeobiogeography and palaeoclimate. In: Lucas, S.G., Cassinis, G., Schneider, J.W. (Eds.),

720   Non-Marine Permian Biostratigraphy and Biochronology. Geological Society of London,

721   Special Publications 265, pp. 297-324.
                                                       31


722   Hminna, A., Voigt, S., Saber, H., Schneider, J.W., Hmich, D., 2012. On a moderately diverse

723   continental ichnofauna from the Permian Ikakern Formation (Argana Basin, Western High

724   Atlas, Morocco). J. Afr. Earth Sci. 68, 15-23.

725   Holmes, R.B., 1977. The osteology and musculature of the pectoral limb of small captorhinids.

726   J. Morphol. 152, 101-140.

727   Holmes, R.B., 2003. The hind limb of Captorhinus aguti and the step cycle of basal amniotes.

728   Can. J. Earth Sci. 40, 515-526.

729   Irschick, D.J., Jayne, B.C., 1999. Comparative three-dimensional kinematics of the hindlimb

730   for high-speed bipedal and quadrupedal locomotion of lizards. J. Exp. Biol. 202, 1047-1065.

731   Lee, M.S.Y., 1997. The evolution of the reptilian hindfoot and the homology of the hooked fifth

732   metatarsal. J. Evolution Biol. 10, 253-263.

733   Leonardi, G., 1987. Glossary and Manual of Tetrapod Footprint Paleoichnology. Departamento

734   Nacional de Produção Mineral, Brasilia (117 pp.).

735   Liebrecht, T., Fortuny, J., Galobart, À, Müller, J., Sander, P.M., 2017. A large multiple-tooth-

736   rowed captorhinid reptile (Amniota: Eureptilia) from the Upper Permian of Mallorca (Balearic

737   Islands western Mediterranean). J. Vertebr. Paleontol. 37, e1251936.

738   Lucas, S.G., Krainer, K., Chaney, D.D., Dimichele, W.A., Voigt, S., Berman, D.S., Henrici,

739   A.C. 2013. The Lower Permian Abo Formation in Central New Mexico. N. M. Mus. Nat. Hist.

740   Sci. Bull. 59, 161-180.

741   Marchetti, L., Avanzini, M., Conti, M.A., 2013. Hyloidichnus bifurcatus Gilmore, 1927 and

742   Limnopus heterodactylus (King, 1845) from the Early Permian of Southern Alps (N Italy): A

743   new equilibrium in the Ichnofauna. Ichnos 20, 202-213.
                                                      32


744   Marchetti, L., Ronchi, A., Santi, G., Schirolli, P., Conti, M.A., 2015a. Revision of a classic site

745   for the Permian tetrapod ichnology (Collio Formation, Rompia and Caffaro valleys, N. Italy),

746   new evidences for the radiation of captorhinomorph footprints. Palaeogeogr. Palaeoclimatol.

747   Palaeoecol. 433, 140-155.

748   Marchetti, L., Ronchi, A., Santi, G., Voigt, S., 2015b. The Gerola Valley site (Orobic Basin,

749   Northern Italy): a key for understanding late Early Permian tetrapod ichnofaunas. Palaeogeogr.

750   Palaeoclimatol. Palaeoecol. 439, 97-116.


751   Marchetti, L., Forte, G., Bernardi, M., Wappler, T., Hartkpf-Fröder, C., Krainer, K.,

752   Kustatscher, E., 2015c. Reconstruction of a Late Cisuralian (Early Permian) floodplain lake

753   environment: Palaeontology and sedimentology of the Tregiovo Basin (Trentino-Alto Adige,

754   Northern Italy). Palaeogeogr. Palaeoclimatol. Palaeoecol. 440, 180-200.


755   Marchetti, L., 2016. New occurrences of tetrapod ichnotaxa from the Permian Orobic Basin

756   (Northern Italy) and critical discussion of the age of the ichnoassociation. Pap. Palaeontol. 2,

757   363-386.


758   Marchetti, L., Mujal, E., Bernardi, M., 2017. An unusual Amphisauropus trackway and its

759   implication for understanding seymouriamorph locomotion. Lethaia 50, 162-174.


760   Marchetti, L., Klein, H., Buchwitz, M., Ronchi, A., Smith, R.M., De Klerk, W.J., Sciscio, L.,

761   Groenewald, G.H., 2019a. Permian-Triassic vertebrate footprints from South Africa:

762   Ichnotaxonomy, producers and biostratigraphy through two major faunal crises. Gondwana

763   Res. 72, 139-168.


764   Marchetti, L., Belvedere, M., Voigt, S., Klein, H., Castanera, D., Díaz-Martínez, I., Marty, D.,

765   Xing, L., Feola, S., Melchor, R.N., Farlow, J.O., 2019b. Defining the morphological quality of
                                                    33


766   fossil footprints. Problems and principles of preservation in tetrapod ichnology with examples

767   from the Palaeozoic to the present. Earth Sci. Rev. 193, 109-145.


768   Marchetti, L., Voigt, S., Mujal, E., Lucas, S.G., Francischini, H., Fortuny, J., Santucci, V.L.,

769   2020a. Extending the footprint record of Pareiasauromorpha to the Cisuralian: earlier

770   appearance    and    wider   palaeobiogeography     of   the   group.   Pap.   Paleontol.   doi:

771   10.1002/spp2.1342.


772   Marchetti, L, Voigt, S., Lucas, S. G., Stimson, M. R., King O. A., Calder, J. H., 2020b.

773   Footprints of the earliest reptiles: Notalacerta missouriensis – Ichnotaxonomy, potential

774   trackmakers, biostratigraphy, palaeobiogeography and palaeoecology. Ann. Soc. Geol. Pol., 90

775   doi: 10.14241/asgp.2020.13.


776   Marchetti, L., Francischini, H., Lucas, S.G., Voigt, S., Hunt, A.P., Santucci, V.L., 2020c.

777   Paleozoic vertebrate ichnology of Grand Canyon National Park. In: Santucci, V.L., Tweet,

778   J.S.(Eds.), Grand Canyon National Park: Centennial paleontological resource inventory.

779   National Park Service, Fort Collins, Colorado, pp. 333-379.


780   Matamales-Andreu, R., Fortuny, J., Mujal, E., Galobart, À., 2019. Tetrapod tracks from the

781   Permian of Mallorca (western Mediterranean): preliminary data, biostratigraphic and

782   biogeographic inferences. In: The Palaeontological Association, 63rd annual meeting, 15th–

783   21st December 2019: 107.


784   Melchor, R.N., Sarjeant, W.A.S., 2004. Small amphibian and reptile footprints from the

785   Permian Carapacha basin, Argentina. Ichnos 11, 57-78.

786   Modesto, S., Smith, R.M., 2001. A new Late Permian captorhinid reptile: a first record from

787   the South African Karoo. J. Vertebr. Paleontol. 21, 405-409.
                                                     34


788   Modesto, S.P., Lamb, A.J., Reisz, R.R., 2014. The captorhinid reptile Captorhinikos valensis

789   from the lower Permian Vale Formation of Texas and the evolution of herbivory in eureptiles.

790   J. Vertebr. Paleontol. 34, 291-302.

791   Modesto, S.P., Richards, C.D., Ide, O., Sidor, C.A., 2019. The vertebrate fauna of the Upper

792   Permian of Niger – X. The mandible of the captorhinid reptile Moradisaurus grandis. J.

793   Vertebr. Paleontol. 38, e1531877.

794   Morafka, D.J., Spangenberg, E.K., Lance, V.A., 2000. Neonatalogy of the reptiles. Herpetol.

795   Monogr. 14, 353-370.

796   Moreau, J.-D., Benaouiss, N., Tourani, A., Steyer, J.-S., Laurin, M., Peyer, K., Béthoux, O.,

797   Aouda, A., Jalil, N.-E., 2020. A new ichnofauna from the Permian of the Zat Valley in the

798   Marrakech High Atlas of Morocco. J. Afr. Earth. Sci. 172, 103973.

799   Mujal, E., Marchetti, L., 2020. Ichniotherium tracks from the Permian of France, and their

800   implications for understanding the locomotion and palaeobiogeography of large

801   diadectomorph.     Palaeogeogr.       Palaeoclimatol.   Palaeoecol.   547,    109698.     doi :

802   10.1016/j.palaeo.2020.109698


803   Mujal, E., Schoch, R. R., 2020. Middle Triassic (Ladinian) amphibian tracks from the Lower

804   Keuper succession of southern Germany: implications for temnospondyl locomotion and track

805   preservation.    Palaeogeogr.       Palaeoclimatol.     Palaeoecol.   543,    109625.      doi:

806   10.1016/j.palaeo.2020.109625


807   Mujal, E., Fortuny, J., Oms, O., Bolet, A., Galobart, À., Anadón, P., 2016. Palaeoenvironmental

808   reconstruction and early Permian ichnoassemblage from the NE Iberian Peninsula (Pyrenean

809   Basin). Geol. Mag. 153, 578-600.
                                                     35


810   Mujal, E., Fortuny, J., Bolet, A., Oms, O., López, J.Á., 2017. An archosauromorph dominated

811   ichnoassemblage in fluvial settings from the late Early Triassic of the Catalan Pyrenees (NE

812   Iberian Peninsula). PLoS ONE 12, e0174693. https://doi.org/10.1371/journal.pone.0174693.

813   Mujal, E., Marchetti, L., Schoch, R.R., Fortuny, J., 2020. Upper Paleozoic to Lower Mesozoic

814   tetrapod ichnology revisited: photogrammetry and relative depth pattern inferences on

815   functional prevalence of autopodia. Front. Earth Sci. 8, 248.

816   O’Keefe, F.R., Sidor, C.A., Larsson, H.C., Maga, A., Ide, O., 2005. The vertebrate fauna of the

817   Upper Permian of Niger-III, morphology and ontogeny of the hindlimb of Moradisaurus

818   grandis (Reptilia, Captorhinidae). J. Vertebr. Paleontol. 25, 309-319.

819   Olson, E.C., 1962. Permian vertebrates from Oklahoma and Texas, Part II. The osteology of

820   Captorhinikos chozaensis Olson. Okla. Geol. Surv. Circ. 59, 49-68.

821   Olson, E.C., Barghusen, H., 1962. Permian vertebrates from Oklahoma and Texas (Vol. 2).

822   University of Oklahoma (68 pp.).

823   Padian, K., Olsen, P.E. Footprints of the Komodo monitor and the trackways of fossil reptiles.

824   Copeia 3, 662-671.

825   Peabody, F.E., 1948. Reptile and amphibian trackways from the Lower Triassic Moenkopi

826   Formation of Arizona and Utah. University of California Publications. Bull. Depart. Geol. Sci.

827   27, 295-468.

828   Peabody, F.E., 1956. Gilmore’s split-toed footprint from the Grand Canyon Hermit Shale

829   Formation.: Plateau, Univ. California 29, 41-43.

830   Pierce, S.E., Ahlberg, P.E., Hutchinson, J.R., Molnar, J.L., Sanchez, S., Tafforeau, P., Clack,

831   J.A., 2013. Vertebral architecture in the earliest stem tetrapods. Nature 494, 226.
                                                    36


832   Reisz, R.R., Liu, J., Li, J., Müller, J., 2001. A new captorhinid reptiles Gansurhinus

833   qingtoushanensis gen. et sp. nov., from the Permian of China. Naturwiss. 98, 435-441.

834   Reisz, R.R., LeBlanc, A.R.H., Sidor, C.A., Scott, D., May, W. 2015. A new captorhinid reptile

835   from the lower Permian of Oklahoma showing remarkable dental and mandibular convergence

836   with microsaurian tetrapods. Sci. Nature 102, 50. doi: 10.1007/s00114-015-1299-y.

837   Ricqlès, A.de., 1984. Remarques systématiques et méthodologiques pour server à l’étude de la

838   famille des Captorhinides (Reptilia, Cotylosauria, Captorhinimorpha). Ann. Paléontol. 70, 18-

839   21.

840   Ricqlès, A.de., Taquet, P., 1982. La faune de vertébrés du Permien supérieur du Niger. I. Le

841   captorhinomorphe Moradisaurus grandis (Reptilia, Cotylosauria). Le crâne. Ann. Paléontol.

842   68, 33-106.


843   Romano, M., Citton, P., Nicosia, U., 2016. Corroborating trackmaker identification through

844   footprint functional analysis: the case study of Ichniotherium and Dimetropus. Lethaia 49, 102-

845   116.


846   Schaeffer, B., 1941. The morphological and functional evolution of the tarsus in amphibians

847   and reptiles. Bull. Am. Mus. Nat. Hist. 78, 398-472.

848   Smith, R.M., Sidor, C.A., Tabor, N.J., Steyer, J.-S., 2015. Sedimentology and vertebrate

849   taphonomy of the Moradi Formation of northern Niger: A Permian wet desert in the tropics of

850   Pangeae. Palaeogeogr. Palaeoclimatol. Palaeoecol. 440, 128-141.

851   Steyer, J.-S., Gand, G., Smith R.M., Sidor, C.A., Tabor, N., 2007. The first tetrapod trackways

852   from the Paleozoic of West Africa: one step towards reconstructing the landscape of Central

853   Pangea. J. Vertebr. Paleontol. 27, 152A.
                                                     37


854   Sumida, S.S., 1989. The appendicular skeleton of the Early Permian genus Labidosaurus

855   (Reptilia, Captorhinomorpha, Captorhinidae) and the hind limb musculature of captorhinid

856   reptiles. J. Vertebr. Paleontol. 9, 295-313.

857   Voigt, S., 2005. Die Tetrapodenichnofauna des kontinentalen Oberkarbon und Perm im

858   Thüringer Wald: Ichnotaxonomie, Paläoökologie und Biostratigraphie (PhD thesis). Martin-

859   Luther Universität Halle-Wittenberg, Göttingen (179 pp).

860   Voigt, S., Lucas, S.G., 2015. Permian tetrapod ichnodiversity of the Prehistoric Trackways

861   National Monument (south-central New Mexico, U.S.A.). N. M. Mus. Nat. Hist. Sci. Bull. 65,

862   153-167.

863   Voigt, S., Lucas, S.G., 2017. Early Permian tetrapod footprints from central New Mexico. N.

864   M. Mus. Nat. Hist. Sci. Bull. 77, 333-352.

865   Voigt, S., Lucas, S.G., 2018. Outline of a Permian tetrapod footprint ichnostratigraphy. In:

866   Lucas, S.G. and Zhen, Z. (Eds.), The Permian Timescale. Geological Society of London,

867   Special Publication 450 pp. 387-404.

868   Voigt, S., Berman, D.S., Henrici, A.C., 2007. First well-established track-trackmaker

869   association of Paleozoic tetrapods based on Ichniotherium trackway and diadectid skeletons

870   from the Lower Permian of Germany. J. Vertebr. Paleontol. 27, 553-570.

871   Voigt, S., Hminna, A., Saber, H., Schneider, J.W., Klein, H., 2010. Tetrapod footprints from

872   the uppermost level of the Permian Ikakern Formation (Argana Basin, Western High Atlas,

873   Morocco). J. Afr. Earth Sci. 57, 470-478.

874   Voigt, S., Lagnaoui, A., Hminna, A., Saber, H., Schneider, J.W., 2011. Revisional notes on the

875   Permian tetrapod ichnofauna from the Tiddas Basin, central Morocco. Palaeogeogr.

876   Palaeoclimatol. Palaeoecol. 302, 474-483.
                                                     38


877   Willey, J.S., Biknevicius, A.R., Reilly, S.M., Earls, K.D. 2004. The tale of the tail: limb

878   function and locomotor mechanics in Alligator mississippiensis. J. Exp. Biol. 207, 553-563.

879   Willinston, S.W., 1910. The skull of Labidosaurus. Am. J; Anat. 10, 69-84.

880   Zouicha, A., Voigt, S., Saber, H., Marchetti, L., Hminna, A., El Attari, A., Ronchi, A.,

881   Schneider, J.W., 2020. First record of permian continental trace fossils in the Jebilet Massif,

882   Morocco. J. Afr. Earth. Sci. 173, 104015.

883
884   Table 1. Parameters of the pes imprints of Hyloidichnus. FL: foot length (mm), FW: foot width
885   (mm), I-V: digit number, L: digits length (mm), div: digits divarication (degrees).
       Specimen       FL      FW      I-      II-    III-    IV-    V-      div    div      div    div    div

                                      L       L      L       L      L       I/II   II/III III/IV   IV/V

       GONF-A-04      28      25      3.6     4.95   7.2     9      4.5     32     30       35     75     172

       GONF-G20       53.3    ?46.7   18.7    22.7   29.3    32     18.7    33     40       45     39     154


       GONF-G21       44.1    51.2    17.6    21.2   28.2    33.5   19.4    45     28       32     35     140


       GONF-A-11      61      75      15      24     30      36     25      50     35       40     53     178

886
887   Table 2. Parameters of the manus imprints of Hyloidichnus. FL: foot length (mm), FW: foot
888   width (mm), I-V: digit number, L: digits length (mm), div: digits divarication (degrees).
       Specimen        FL      FW     I-      II-    III-    IV-L   V-      div    div      div    div    div

                                      L       L      L              L       I/II   II/III III/IV   IV/V

       GONF-A-04       19.2    22     2.25    3.6    4.5     5.4    3.6     30     55       19     50     154

       GONF-G20        30.4    48     12.8    19.2   22.4    25.6   9.6     36     43       60     38     177


       GONF-G21        35.3    42.4   14.1    15.9   19.4    26.4   12.4    50     45       23     34     152


       GONF-A-11       61      48     17      19     27      29     15      50     38       32     53     173

889
                                                                 39


    890     Table 3. Trackway parameters. SLp: pes stride length (mm), PLp: pes pace length (mm), Pap:
    891     pes pace angulation (degrees), PWp: pes width of pace (mm), Lp-p: pes length of pace (mm),
    892     dist ip: distance interpes (mm), divp : pes divarication (degrees), SLm: manus stride length
    893     (mm), PLm: manus pace length (mm), Pam: manus pace angulation (degrees), PWm: manus
    894     width of pace (mm), Lp-m: manus length of pace (mm), dist im: distance intermanus (mm),
    895     divm : manus divarication (degrees).
Specimen         SLp         PLp     Pap PWp         Lp-    dist      div-       SLm         PLm      Pam     PWm       Lp-       dist     div-

                                                     p      ip        p                                                 m         im       m

GONF-A-04        108         78.4    85      49.5    45     38.7      80         101.7       78.75    95      45        56.7      49.5     120

GONF-G20                     80              104     88     52.8      50         240         136      120     72        352       24       130


GONF-G21         158.8       123.5   80      112.9   79.4   60        80         155.3       114.7    85      97        70.6      49.4     80

GONF-A-11                    21              98      18     20        65         130         19.5     130     85        17.5      16       53

    896
    897
    898     Table 4. Tail impression and digit drag marks parameters. dist m-p: distance pes-manus (mm),
    899     FLp/FLm: pes length/manus length (mm), BL: body length (mm). BL/FL: body length/pes
    900     length, SL/FL: stride length/pes length, SL/BL: stride length/body length, bsl:tail baseline
    901     (mm), mdl: tail midline (mm), WL: wave length (mm), amp: amplitude (mm), TW: tail width
    902     (mm).

 Specimen       dist     FLp/FLm sole/palm BL               BL/FL         SL/FL SL/BL bsl                   mdl WL          amp TW

                m-p

 GONF-A-04      39       1.45         1.37           50     1.78          4.29         2.4          156     160    90       2.7    4-5.5

 GONF G 20      21.3     1.75         4.4            184    3.45          11.6 ?       1.9


 GONF G21       19.4     1.25         1.19           141.2 3,18           3.9          1.1          273.5 300               1      8-9

 GONF-A-11      20       1            0.8            175    2,86          2.5          0.74         222     255    123.5 52        10

    903
    904
                                                      40


905   Table 5. Parameters of the manus and pes of Captorhinus aguti and comparisons with the
906   avergage ratios of manus and pes tracks.

       Captorhinus aguti   pes      manus    ratio                  C. aguti   Track averages



       Length              50       42.5     pes/manus length       1.18       1.36

       Width               55       50       pes length/width       0.9        0.98

       digit I/IV          0.5      0.43     manus length/width     0.85       0.9

       digit II/IV         0.63     0.71     metatarse/metacarpus   1.65       1.94

       digit III/IV        0.81     0.93     length

       digit V/IV          0.63     0.57

       digit IV/length     0.9      0.82

       BL/FL               3,7

907
908
909
910
                                                      41




911
912   Figure 1 | A. Map of the Permian basins of Provence (after Durand, 2011, modified) with
913   position of the study area and Gonfaron A site with position of footprints levels. B. Stratigraphic
914   log of the Gonfaron A site. C. Photograph of the Gonfaron A site (Romain Garrouste, MNHN).
915
                                                     42




916
917   Figure 2 | A. Photograph of Hyloidichnus trackways GONF-A-04 with digit drag traces and
918   tail impression. B. Interpretative drawing with Nr. 1 to 4: pes-manus couples, Nr. 5: tail
919   impression, Nr. 6 and 6’: digit drag traces. C. Photograph of best preserved right manus imprint.
920   D. Photograph of best preserved left pes imprint. D. Photograph showing details of digit drag
921   traces. Scale bars: 20 mm (A, B), 10 mm (C–E). Photographs by Philippe Loubry, CNRS Paris.
922
                                                 43




923
924   Figure 3 | A. Photograph of Hyloidichnus incomplete step cycle GONF-A-11 with sinuous tail
925   impression, concave epirelief. B. Interpretive drawing. Scale bar: 20 mm. Photograph by
926   Phillipe Loubry, CNRS Paris.
927
                                                   44




928
929   Figure 4 | A. Photograph of Hyloidichnus incomplete step cycle GONF-G20 with digit drag
930   marks, convex hyporelief. Note the holotype trackway of Varanopus rigidus on the same
931   surface. B. Close-up of left pes-manus couple and digit drag traces. C. Interpretive drawing.
932   Scale bars: 50 mm (A), 40 mm (B, C). Photograph by Lorenzo Marchetti, MB.
933
                                                   45




934
935   Figure 5 | A. Photograph of Hyloidichnus trackway with digit drag traces GONF-G21 with
936   straight and shallow tail impression, convex hyporelief, artificial cast. B. Close up on best
937   preserved right pes-manus couple. C. Interpretive drawing. Scale bars: 10 cm (A), 50 mm (B),
938   30 mm (C). Photograph by Lorenzo Marchetti, MB.
939
                                                   46




940
941   Figure 6 | Stratigraphic distribution of Hyloidichnus trackways compared with the consensual
942   stratocladogram of the captorhinomorphs (compiled from The Paleobiology Database
943   https://paleobiodb.org; Modesto and Smith, 2001; Modesto et al., 2019; Voigt and Lucas 2018).
944
                                                       47




945
946   Figure 7 | A. Photograph of skeletal reconstitution of Captorhinus aguti (Field Museum of
947   Chicago), the red lines indicate the pes length and the glenoacetabular distance, scale bar: 50
948   mm. B. Anatomy of the manus redrawn after Holmes (1977). ul: ulna, ue: ulnare, in:
949   intermedium, ra: radius, re: radiulare, pc: proximal centrale, dc: distal centrale, 1-5: distal
950   carpals, scale: 10 mm. C. Anatomy of the pes redrawn after Holmes (2003). ti: tibia, fi: fibula,
951   cal : calcaneus, ast: astragalus, c: centrale, 1-5: distal tarsals, scale bar: 10 mm. D. Superposition
952   of the best preserved manus track of specimen GONF-A-04 (in orange) with the skeletal
953   reconstruction of the manus redrawn after Holmes (1977); the red circle indicates the zone with
954   most power flexion during the locomotor cycle, scale bar: 10 mm. E. Superposition of the best
955   preserved pes track of GONF-A-04 (in orange) with the skeletal reconstruction of the pes
956   redrawn after Holmes (2003); the red circle indicates the zone with most power flexion during
957   the locomotor cycle, scale bar: 10 mm. Photograph in A by Lorenzo Marchetti, MB.
958
                                                    48




959
960   Figure 8 | I. Swing phase of the manus. A. Beginning. B. Early. C. Midstride. D. End. II. Swing
961   phase of the pes. In the square: re-orientation phase between the end of the stance phase and
962   the beginning of the swing phase. A. Reconstruction of captorhinomorph locomotion based on
963   GONF-A-04 trackway. B. Beginning. C. Early. D. Mistride. E. End.
964
                                                    49




965
966   Figure 9 | A. 3D model in false-colour depth and with contours of GONF-A-04. B. Interpretive
967   drawings. C. Photograph of the best preserved right manus track. D. 3D model of the track in
968   C with superimposition of Captorhinus manus skeleton. E. Photograph of the best-preserved
969   left pes. F. 3D model of the track in E with superimposition of Captorhinus pes skeleton. G. 3D
970   model of GONF-A-11. H. Interpretive drawing. I. Photograph of the left pes-manus couple. J.
971   3D model of the track in I with superimposition of Captorhinus autopodians skeletons. The red
972   zones represent the most deeply impressed part of the imprints. The arrows point to the steepest
973   parts of the tail impressions. Captorhinus skeletons after Holmes (2003), modified. Scale bars:
974   50 mm.
                                                  50




975
976   Figure 10 | A. 3D model in false-colour depth and with contours of GONF-G20. B. Interpretive
977   drawing. C. Photograph of the best-preserved left pes-manus couple. D. 3D model of tracks in
978   C with superimposition of Captorhinus autopodial skeletons. E. 3D model of GONF-G21. F.
                                                   51


979   Interpretive drawing. G. Photograph of the best-preserved right pes-manus couple. H. 3D model
980   of the tracks in G with superimposition of Captorhinus autopodial skeletons. The red zones
981   represent the most deeply impressed part of the imprints. Captorhinus skeletons after Holmes
982   (2003), modified. Scale bars: 10 cm (A, B, E, F), 50 mm (C, D, G, H).