This is the accepted version of the article:
Matamales Andreu, Rafel; Mujal, Eudald; Galobart, Àngel; [et al.]. «Insights on
the evolution of synapsid locomotion based on tetrapod tracks from the lower
Permian of Mallorca (Balearic Islands, western Mediterranean)». Palaeogeogra-
phy, Palaeoclimatology, Palaeoecology, Vol. 579 (October 2021), art. 110589.
DOI 10.1016/j.palaeo.2021.110589


This version is available at https://ddd.uab.cat/record/248974

under the terms of the               license
 1       Insights on the evolution of synapsid locomotion based on tetrapod tracks from the lower

 2                   Permian of Mallorca (Balearic Islands, western Mediterranean)

 3            Rafel Matamales-Andreu1,2*, Eudald Mujal1,3, Àngel Galobart1,4 & Josep Fortuny1

 4
     1
 5       Institut Català de Paleontologia Miquel Crusafont, Universitat Autònoma de Barcelona, Edifici
 6   ICTA-ICP, c/ Columnes s/n, Campus de la UAB, 08193 Cerdanyola del Vallès, Barcelona, Spain.
 7
     2
 8       Museu Balear de Ciències Naturals, ctra. Palma-Port de Sóller km 30, 07100 Sóller. Mallorca,
 9   Illes Balears, Spain.
10
     3
11       Staatliches Museum für Naturkunde Stuttgart, Rosenstein 1, 70191 Stuttgart, Germany.
12
     4
13       Museu de la Conca Dellà, c/ del Museu 4, 25650 Isona i Conca Dellà, Lleida, Spain.
14

15   *Corresponding author: rafel.matamales@icp.cat

16

17

18   Abstract

19             “Pelycosaur”-grade synapsids were a successful group of terrestrial tetrapods that lived

20   during the Carboniferous and Permian, utilising a wide diversity of ecological niches. They are

21   considered the trackmakers of the ichnogenera Dimetropus and possibly also Dromopus and

22   Tambachichnium, found in upper Palaeozoic deposits of North America, North Africa and

23   Europe. Here we describe a new morphotype from the lower Permian of Mallorca (Balearic

24   Islands, western Mediterranean), identified as cf. Dimetropus isp., which was purportedly

25   produced by caseid synapsids. The trackmaker identity is inferred based on the digit proportions,

26   as caseids are the only pelycosaurs with species showing mesaxonic autopodia, and a relative

27   depth pattern (corresponding to the functional prevalence of the autopodia of the trackmaker)

28   showing an overall similarity to that of Dimetropus osageorum, which is also attributed to caseids.

29   A detailed study of the expulsion rims and drag traces makes it possible to infer the mode of

30   locomtion of the trackmaker, with a gait alternating the movement of manus and pedes of the two

                                                      1
31   sides of the body and lateral undulation of the spine. Our results provide new information

32   concerning the locomotion of early synapsids, which would undergo important functional

33   modifications latter in their evolutionary history such as a shift from abducted to adducted posture

34   and lateral to sagittal bending of the axial skeleton.

35

36   Key-words: ichnology; photogrammetry; “Pelycosauria”; functional prevalence; equatorial

37   Pangaea

38

39   1. Introduction

40           Synapsids were a key tetrapod clade in late Palaeozoic terrestrial ecosystems, including

41   herbivores with high-fibre diet, small predators and large hypercarnivores, which occupied a

42   variety of ecological niches (Kemp, 2005; Angielczyk & Kammerer, 2018). Carboniferous and

43   Permian synapsids are nowadays included in two major clades: “pelycosaur”-grade synapsids,

44   comprising the most basal groups (Romer & Price, 1940), and non-mammalian therapsids, which

45   diversified over the middle and late Permian (Angielczyk & Kammerer, 2018). Regarding their

46   ichnological record, late Palaeozoic synapsids are represented by nine different ichnogenera, three

47   of which correspond to pelycosaurs (Dromopus, Tambachichnium and Dimetropus) and six to

48   therapsids (Brontopus, Karoopes, Capitosauroides, Dicynodontipus, Dolomitipes and

49   Procolophonichnium; the latter four also present in Triassic deposits) (Marchetti et al., 2019a,

50   2019b, Spindler et al., 2019). Some common traits of these ichnogenera (albeit not always

51   present) are the deeper impression of the median-lateral region of the tracks (median-lateral

52   functiontal prevalence of autopodia), tracks with paw-like shape (with faintly imprinted proximal

53   part of digits and conspicuos metatarso-/metacarpophalangeal pad and digit tip imprints) because

54   of the digital arcade of synapsid phalanges, and well-imprinted palms and soles (Marchetti et al.,

55   2019a; Mujal et al., 2020).

56           The only ichnogenus attributed to synapsids that had been (preliminarily) reported so far

57   from the lower Permian beds of Mallorca (Balearic Islands, western Mediterranean) was

58   Dimetropus (Matamales-Andreu et al., 2019), with forms similar to Dimetropus leisnerianus. The

                                                       2
59   present work analyses newly discovered tetrapod tracks and a trackway that correspond to a

60   second, hitherto unkown morphotype of Dimetropus. Those ichnites, attributable to a synapsid

61   trackmaker, possibly a caseid, enlarge the ichnofaunal diversity of the lower Permian of central

62   equatorial Pangaea, and also provide new insights on the evolution of synapsid locomotion.

63

64   2. Geographical and geological context

65           Mallorca is the largest of the Balearic Islands (western Mediterranean) (Figure 1A), and

66   it is broadly structured in three mountain ranges of SW–NE orientation, with rocks ranging from

67   the Carboniferous to the middle Miocene, and basins filled with sediments ranging from the upper

68   Miocene to the Holocene. The nomenclature of the Permian and Triassic lithostratigraphic units

69   of Mallorca is in need of a thorough revision, as each of the main authors who studied them

70   (Calafat, 1988; Gómez-Gras, 1993; Ramos, 1995) established different names and boundaries for

71   the units, which have never been formally described. However, such a review is beyond the scope

72   of the present work, which uses a nomenclature of correlative letters (from lowermost to

73   uppermost: A, B and C) for the three main lithostratigraphic units (Figure 1B), based on the

74   stratigraphic logs of Gómez-Gras (1993), which are herein considered to be the most exhaustive

75   of those so far published.

76           The Racó de s’Algar tracksite (Calafat et al., 1986, 1986–1987; Calafat, 1988), where the

77   ichnites studied herein were discovered, corresponds to the bases of two fine-grained sandstone

78   beds located in Unit B (metre 32 of the log “E” of Calafat, 1988) (Figure 1B–C). These two beds

79   are mostly tabular, although they wedge laterally, and show climbing ripple lamination,

80   corresponding to two successive point bars of a meandering river deposit separated by a thin bed

81   of mudstones, probably representing seasonal stagnation and desiccation of the channel (Figure

82   1D–E). Their bases preserve, as natural casts, the sedimentary (large mud-cracks) and biogenic

83   (tetrapod ichnites) structures recorded on the surfaces of the underlying sandy mudstone beds,

84   which nowadays are mostly eroded. The irregularity of the ichnite-bearing surfaces, reminiscent

85   of “wrinkle structures”, points to algal mat development, which probably also contributed to the

86   mode of impression and preservation of the tracks. The succession encompassing the tetrapod

                                                    3
 87   tracks is here interpreted as a waterhole environment, with a facies assemblage that is similar to

 88   that of the seasonally-dry fluvial environments from the upper Carboniferous of Nova Scotia

 89   (Falcon-Lang et al., 2004, 2007). In the succession of Mallorca, such a sedimentary setting can

 90   be inferred based on the presence of wave ripples developed on the channel bed, suggesting

 91   stagnation of the water, and subsequent desiccation of the pool, as indicated by the large mud

 92   cracks (Figure 1C, E).

 93           Placed in the general framework of the Permian lithostratigraphic units of Mallorca

 94   (Figure 1B), this tracksite is located in the middle–upper part of what is herein called Unit B

 95   (=“Racó de s’Algar unit” of Calafat, 1988; =“Stretch a” of Gómez-Gras, 1993; =“Port des

 96   Canonge unit” of Ramos, 1995), which was interpreted as a meandering river system (Calafat,

 97   1988; Gómez-Gras, 1993; Ramos, 1995). This unit was dated as Permian based on the discovery

 98   of moradisaurine captorhinid eureptiles, both in the form of bones (Liebrecht et al., 2017) and

 99   tracks (Gand et al., 2010, revising the ichnites figured by Calafat et al., 1986–1987). More

100   recently, Matamales-Andreu et al. (2019) tentatively suggested that the age could be narrowed

101   down to the lower Permian based on the presence of Dimetropus (possibly D. leisnerianus), and

102   further constrained it to the upper part of the Cisuralian (Artinskian–Kungurian) because of the

103   great abundance of tracks attributed to eureptiles (Hyloidichnus) and the rarity of tracks produced

104   by anamniotes (see discussion on the Artinskian–Kungurian reptile radiation by Marchetti et al.,

105   2015, 2019c).

106

107   3. Material and methods

108           Five slabs with ichnites (natural casts) from the same stratigraphic surface of the Racó de

109   s’Algar tracksite (upper ichnite-bearing bed; Figure 1C) are described in the present work. The

110   material collected in situ for the present study consists of a right manus imprint (DA21/15-05-

111   01), housed at the Museu de Mallorca (Palma, Mallorca, Spain, acronym DA21/15-##-##).

112   Moreover, three more slabs were found in situ but not collected (it was not possible, as they were

113   in the base of an unstable cliff; field acronym NA-26-##): one with a poorly preserved left manus-

114   pes set (NA-26-02), one with a poorly preserved right pes imprint (NA-26-03), and one with a

                                                      4
115   left manus imprint (NA-26-05). Lastly, from the Tomeu Sáez collection (private collection,

116   Binissalem, Mallorca, Spain; acronym TS-#), we have examined one slab with a trackway

117   comprising three manus-pes sets (TS-3), collected in situ in the early 1980s. Ichnites within a

118   same slab have been identified with correlative lowercase letters. In the preservation scale of

119   Marchetti et al. (2019d), most of the tracks studied herein score a 2 (good preservation), although

120   some could be attributed to 1 (intermediate preservation) or even 0 (poor preservation). Therefore,

121   and although they appear to be sensibly different from previously known ichnotaxa of this age,

122   no new ichnospecies will be defined here.

123           Digital 3D models of all the surfaces with ichnites have been produced using the

124   photogrammetry technique, mostly based on the procedure explained by Mujal et al. (2020).

125   Agisoft Photoscan standard version 1.1.4. (http://www.agisoft.com) has been used to align the

126   photographs, to create the meshes and textures, and to crop the model. MeshLab version 2016.12

127   (http://meshlab.sourceforge.net) has been used to clean, align, scale and measure the mesh.

128   ParaView version 4.1.0 64-bit (http://www.paraview.org) has been used to create the false colour-

129   coded depth maps with contours. ImageJ version 1.52d (https://imagej.nih.gov/ij) has been used

130   to measure the main descriptive parameters for each footprint and the trackway (see Leonardi,

131   1987; Hasiotis et al., 2007), herein presented in Tables 1–2. In order to compare the tracks from

132   Mallorca to their possible trackmaker, bones of the caseid Ennatosaurus tecton were measured

133   on the scaled 3D model provided by Romano et al. (2017).

134

135   4. Systematic palaeoichnology

136

137                             Ichnogenus Dimetropus Romer et Price, 1940

138

139                                          cf. Dimetropus isp.

140

141                                       Figures 2–3; Tables 1–2.

142

                                                      5
143           Studied material: DA21/15-05-01, a natural cast of a right manus imprint. NA-26-02, a

144   poorly preserved natural cast of a left manus-pes set (not collected). NA-26-03, a poorly preserved

145   natural cast of a right pes imprint (not collected). NA-26-05, a natural cast of a left manus imprint

146   (not collected). TS-3, a natural cast of a trackway with two left manus-pes sets and a right manus-

147   pes set. All from the lower Permian of the upper ichnite-bearing bed of the Racó de s’Algar

148   tracksite, Banyalbufar, Mallorca, Balearic Islands, western Mediterranean.

149           Description: Pes track semiplantigrade, pentadactyl, almost as long as wide (ca. 7.3–8.3

150   cm). The digit imprints are slender, strongly curved laterally (outwards) and relatively long.

151   Preservation is quite variable: in some cases, the distal parts of digits are not well imprinted

152   (especially digits II and III) (Figure 2A–B), whereas in other cases digit I seems to be imprinted

153   twice (Figures 2C, 3A). No clear claw traces have been observed on any specimen. The relative

154   length of the digit imprints is ordered as follows: V<I≈II≈IV≈III (very weakly mesaxonic) (Table

155   1). However, digit imprints are dragged, deformed and partly incomplete with mudcollapsed

156   distal ends, so it cannot be decided with certainty whether digit III is really longest, because digit

157   IV may not show its true length. The total divergence angle between digits I–V of pes tracks

158   (measured on the proximal part of the digits due to their strong distal curvature) ranges from 77°

159   to 95°. The digit imprints usually have sharp posterolateral margins, whereas their anteromedial

160   margins have gentler slopes. The sole imprint is oval, with an almost straight to weakly convex

161   posteromedial margin and a length of about 1/2 of the whole track. Metatarsophalangeal pads are

162   well impressed, especially at the bases of digits I–III. The most deeply imprinted areas of the pes

163   tracks are the metatarsophalangeal pads of digits I–III (medial functional prevalence). There is an

164   expulsion rim surrounding the pes track posterolaterally.

165           Manus track plantigrade, pentadactyl, usually slightly wider (6.9–7.1 cm) than long (6.7–

166   9.5 cm). Digit I–IV imprints are slender, relatively long and show clawed tips. The proximal parts

167   of the digit imprints are directed medially, whereas the imprints of the tips are curved laterally,

168   giving digit imprints a “hook-shaped” appearance, especially in digit III (Figures 2B–E, 3A).

169   They become thinner and gradually decrease in depth distally. Digit V imprint, which is the

170   shortest, is almost as short as digit I, slender and usually straight, although its tip is often not well

                                                         6
171   impressed (Figures 2B–C, 3A). Some digit imprints, especially I and II, may have drag traces in

172   front of their tips (Figure 2B–E). The relative lengths of the digit imprints are ordered as follows:

173   V≈I<II≈IV<III (weakly mesaxonic) (Table 1). Digits radiate from the palm, usually without

174   significative overlapping, although the bases of digits I and II can be slightly superimposed. The

175   total divergence angle between the imprints of digit I–V is wide, ranging between 115º and 150º.

176   The palm imprint is posterolaterally elongated, with a concave medial margin (“lunate shape”).

177   Metacarpophalangeal pads are rounded and well-impressed, especially at the bases of digits I–IV

178   with slight overlap in I and II. The most deeply imprinted areas of the manus track are the

179   metacarpophalangeal pads of digits II–III (medial-median functional prevalence of the

180   autopodium), followed by those of digits I and IV. The imprints of the phalangeal parts of the

181   digits are shallower, with that of digit III being the deepest, followed by those of digits II and IV,

182   and then the rest of the palm. There are well-marked expulsion rims posteromedially, laterally

183   and between the imprints of digits II and III, and slightly less marked ones between the other digit

184   imprints.

185           Trackway produced by a quadrupedal animal, with pes tracks slightly larger than manus

186   tracks (weak heteropody). The manus-pes sets are alternately arranged, with the manus tracks

187   positioned at the height of the mid-proximal portion of the pes track of the next couple (Figure

188   2A). The pace angulations of the pes and the manus imprints are of about 80º and 90º,

189   respectively, and thus consistent with an abducted (sprawling) limb posture. Respect to the

190   trackway midline, pes imprints are rotated laterally (outwards) by 20–30º, whereas manus

191   imprints are rotated medially (inwards) by 9–26º and are located anteriorly from the pes imprints.

192   Manus tracks can be positioned medially or laterally from the pes tracks; this trait is variable, as

193   in other synapsid tracks (e.g., Marchetti et al., 2019d) The glenoacetabular distance is of about

194   32.5 cm (Table 2). No tail traces have been observed.


195

196   5. Discussion

197   5.1. Comparison of cf. Dimetropus isp. from Mallorca to other Permian ichnogenera


                                                        7
198           The tracks here identified as cf. Dimetropus isp. are pentadactyl, mesaxonic and have

199   anteroposteriorly elongated palm and sole imprints (Figure 4A), which makes it possible to rule

200   out all the Permian ichnogenera attributed to anamniotes, “reptiliomorphs”, eureptiles and

201   parareptiles (compare to Lucas, 2019), thus placing them in the synapsid group. The ichnogenera

202   currently attributed to synapsid trackmakers are: Dromopus, of non-varanopine varanopids and

203   some other groups of diapsids (Spindler et al., 2019; Figure 4B); Tambachichnium, of varanopids

204   (Voigt, 2005; Marchetti et al., 2019a; Figure 4C); Dimetropus, of edaphosaurids,

205   sphenacodontids, ophiacodontids and caseids (Voigt & Ganzelewski, 2010; Sacchi et al., 2014;

206   Marchetti et al., 2019a; Figure 4D–J); Brontopus, of dinocephalians (Marchetti et al., 2019a;

207   Figure 4K–L); Dolomitipes, of dicynodonts (Marchetti et al., 2019a; Figure 4M–N); Karoopes,

208   of gorgonopsians (Marchetti et al., 2019a; Figure 4O); Capitosauroides and at least some

209   ichnospecies of Procolophonichnium, of therocephalians (Marchetti et al., 2019a, 2019b;

210   Buchwitz et al., 2020; Figure 4P–Q); and Dicynodontipus, of cynodonts (Marchetti et al., 2019a;

211   Figure 4R).

212           Although the specimens herein identified as cf. Dimetropus isp. cannot readily be

213   attributed to any of these known ichnogenera (hence the open nomenclature at ichnogenus-level;

214   see Figure 4), their manus imprints are very similar to other tracks attributed to the ichnogenus

215   Dimetropus in the shape of the palm/sole and the digit imprints (compare to Gand, 1988: figure

216   52D; Haubold et al., 1995: fig. 23A; Voigt, 2005: fig. 35D). The strong outwards rotation of the

217   pes imprint, the general proportions, the pace angulation and the stride could be also compared to

218   those of Capitosauroides (see Buchwitz et al., 2020). The latter, however, has hitherto not been

219   found in lower Permian deposits (Buchwitz et al., 2020), and thus here the assignation to

220   Dimetropus (in open nomenclature) is favoured.

221

222   5.2. Comparison of cf. Dimetropus isp. from Mallorca with other ichnospecies of Dimetropus

223           Sacchi et al. (2014) considered five valid ichnospecies within the ichnogenus

224   Dimetropus: D. leisnerianus (Geinitz, 1863), D. osageorum Sacchi, Cifelli, Citton, Nicosia et

225   Romano, 2014, and perhaps D. berea (Tilton, 1931), D. salopensis Haubold et Sarjeant, 1973,

                                                      8
226   and D. nicolasi Gand et Haubold, 1984. However, the last three need revision to rule out that they

227   are undertracks of D. leisnerianus (e.g., Voigt, 2005).

228           cf. Dimetropus isp. from Mallorca differs from D. leisnerianus (Figure 4E–F) by the

229   following features: (1) the manus imprints are weakly mesaxonic in the former, whereas in the

230   latter they are clearly ectaxonic; (2) the digit tips of the manus imprints (especially digit III) are

231   “hook-shaped” in the former, whereas they are mostly straight in the latter; (3) the digits of the

232   manus imprints of the former usually have relatively higher divergence angles than those of the

233   latter; (4) the depth of the digits of the manus imprints decreases distally in the former, whereas

234   in the latter they are deepest proximally and distally, with a shallower region in between; (5) the

235   pes imprints are rotated outwards (laterally) in the former, whereas in the latter they are

236   subparallel to the trackway midline or even slightly rotated inwards (medially); (6) the digits of

237   the pes imprint of the former are all strongly rotated outwards (laterally), whereas they are straight

238   in the former; (7) the sole imprint is anteroposteriorly short and oval in the former, whereas it is

239   usually posteriorly extended in the latter; (8) the autopodial functional prevalence inferred from

240   the relative depth pattern of the manus and pes imprints of the former is medial-median and

241   medial, respectively, whereas in the latter they are both lateral. For comparison, see tracks and

242   trackways of D. leisnerianus in Voigt (2005, 2012), Voigt & Ganzelewski (2010), Voigt et al.

243   (2011, 2012), Lucas et al. (2016), Mujal et al. (2016a, 2020), Marchetti et al. (2019a).

244           The material from Mallorca differs from D. berea, D. salopensis and D. nicolasi

245   essentially in the same characters listed above for D. leisnerianus (see specimens in Haubold &

246   Sarjeant, 1973; Gand & Haubold, 1984; Gand, 1988; Haubold et al., 1995, Lucas et al., 2016).

247   However, these three ichnospecies are of dubious validity (e.g., Voigt, 2005), with definitions

248   based on a few isolated ichnites and manus-pes sets that have slightly different digit lengths, pace

249   angulation angles and distribution of depths than those of D. leisnerianus (see Sacchi et al., 2014).

250   Further research on the type specimens or new materials from the type localities could elucidate

251   this matter, although such a revision is beyond the scope of the present work. Moreover, cf.

252   Dimetropus isp. from Mallorca is also different from the Carboniferous Dimetropus isp. figured

253   by Voigt & Ganzelewski (2010), Dimetropus isp. figured by Lagnaoui et al. (2018) and cf.

                                                        9
254   Dimetropus isp. figured by Marchetti et al. (2019c), because of the reasons listed above for D.

255   leisnerianus.

256           cf. Dimetropus isp. from Mallorca differs from D. osageorum (Figure 4H) by the

257   following features: (1) the digits of the manus imprints of the former are long, slender and digit

258   III is “hook-shaped”, whereas in the latter they are straight, very short and thick; (2) the sole

259   imprints are posteriorly short in the former, whereas in the latter they are elongated, with a marked

260   medial embayment; (3) the pes imprints are strongly rotated outwards (laterally) in the former,

261   whereas in the latter they are subparallel to the trackway midline or even slightly rotated inwards

262   (medially); (4) the pes imprints are only slightly larger than the manus imprints in the former,

263   whereas the latter is notably heteropodic; (5) the functional prevalence of autopodia inferred for

264   the former is medial-median and medial, respectively for the manus and pes imprints, whereas

265   for the latter it is medial-median in both cases (see examples in Sacchi et al., 2014; Romano et

266   al., 2016, 2020).

267           It is at this point important to note that the purported D. leisnerianus figured by Gand

268   (1988) (Figure 4I) show a general size, proportions, heteropody and trackway parameters that are

269   comparable to those reported for D. osageorum by Sacchi et al. (2014) (Figure 4J). Their main

270   difference is that the specimens from Europe have ectaxonic tracks (although relative lengths of

271   the digit imprints appear to be very variable), whereas the North American forms have mesaxonic

272   tracks. Although this aspect is beyond the scope of the present work, future studies should

273   compare these two morphotypes to clarify whether they represent only one ichnospecies (and its

274   range of variability and extramorphological deformation) or they are distinct forms. Moreover,

275   taxonomic status of the footprints illustrated by Gand (1988) (Figure 4I) should probably be

276   reassessed, given the apparent high degree of heretopody of those tracks, atypical of D.

277   leisnerianus (Voigt, 2005; Sacchi et al., 2014) (Figure 4F–H).

278

279   5.3. Trackmaker of cf. Dimetropus isp. from Mallorca

280           Dimetropus is an ichnogenus that has been attributed to various clades of “pelycosaur”-

281   grade synapsids, except for varanopids, which are considered to be the trackmakers of Dromopus

                                                       10
282   and Tambachichnium (see Voigt & Ganzelewski, 2010; Marchetti et al., 2019a; Spindler et al.,

283   2019). The phylogenetic position of varanopids has recently been questioned, suggesting they are

284   actually diapsids (Ford & Benson, 2020), which would also be consistent with some of the

285   parameters of their footprints (Mujal et al., 2020; Buchwitz et al., 2021). However, some other

286   lines of research suggest that they are indeed more closely related to synapsids, as the morphology

287   of their maxillary canals is more similar to ophiacodonts and therapsids than to other reptiles

288   (Benoit et al., 2021). Given their contentious affinities, varanopids will be not further discussed

289   here. In any case and specifically within Dimetropus, D. leisnerianus has been indistinctly

290   attributed to edaphosaurids, sphenacodontids, ophiacodontids and caseids, which have relatively

291   similar manus and pes proportions (see references in Voigt & Ganzelewski, 2010; Sacchi et al.,

292   2014), whereas D. osageorum was proposed to have been produced by large caseids or perhaps

293   edaphosaurids (Sacchi et al., 2014).

294           Mujal et al. (2020) showed that most ichnogenera currently attributed to Carboniferous–

295   Triassic synapsid trackmakers share a median-lateral functional prevalence of the autopodia.

296   Interestingly enough, the ichnogenus Tambachichnium, purportedly produced by varanopids,

297   showed a different functional prevalence (median), typical of diapsids. Mujal et al. (2020)

298   hypothesised that the basal position of varanopids within synapsids could explain their having

299   this particular type of functional prevalence, or that it could be regarded as a trait evolved

300   convergently in a similar functionality.

301           Interestingly, the ichnospecies D. osageorum also appears to be at odds with the

302   hypothesis that the median-lateral functional prevalence is a trait typical of all synapsids. In the

303   touch-down phase (posterior part of the ichnite), the functional prevalence of D. osageorum was

304   considered to be median-lateral, but during the kick-off phase (anterior part of the ichnite) it

305   appeared to be medial-median (Romano et al., 2016, 2020). D. osageorum was attributed to large

306   caseid trackmakers (Sacchi et al., 2014), which are also a synapsid clade that diverged before the

307   eupelycosaurs, the latter containing all ichnogenera that display a median-lateral functional

308   prevalence according to Mujal et al. (2020). Moreover, both manus and pes tracks of D.

309   osageorum are mesaxonic (Sacchi et al., 2014), whereas the autopodia of all non-therapsid

                                                      11
310   eupelycosaurs have a digit IV longer than III, being caseids the only pelycosaurs with species

311   known to have subequal digit II–IV lengths (e.g., Romer & Price, 1940; Olson, 1968).

312           In this sense, cf. Dimetropus isp. from Mallorca is similar to D. osageorum in that they

313   both share a medial-median functional prevalence (with only slight differences), especially in the

314   anterior portion of the tracks, and that they are both mesaxonic, being their digit III imprint

315   slightly longer than that of digit IV. Following the reasoning explained above, it is here tentatively

316   suggested that caseids could also be considered the trackmakers of these cf. Dimetropus isp.

317   tracks. Nevertheless, as noted by Sacchi et al. (2014), the complete structure of edaphosaurid

318   autopodia is unknown, so they cannot be fully excluded from being the trackmakers of these two

319   morphotypes of Dimetropus. In any case, the size of manus and pes tracks of cf. Dimetropus isp.

320   from Mallorca is consistent with the general proportions and dimensions of the autopodia of the

321   middle Permian genus Ennatosaurus, which is the only caseid of relatively small size with digits

322   III and IV being subequal in length (e.g., Olson, 1968; Romano et al., 2017); all other small-sized

323   genera have a digit IV that is clearly longer than III (e.g., Spindler et al., 2016; Berman et al.,

324   2020). Moreover, the manus/pes proportions (1.02 in Ennatosaurus and 0.83–0.95 in cf.

325   Dimetropus isp. from Mallorca) and the glenoacetabular distance in relation to the pes length

326   (4.97 in Ennatosaurus and 3.93–4.41 in cf. Dimetropus isp. from Mallorca) are also similar

327   (measurements of Ennatosaurus based on the complete skeleton figured by Romano et al., 2017).

328   Interestingly, a partial vertebra tentatively attributed to cf. Ennatosaurus tecton, of notably larger

329   size than the Russian specimens, was reported from the middle Permian of the Pyrenees (Mujal

330   et al., 2016b), which were palaeogeographically close to the Mallorcan basin.

331           Taking all of the above into account, a caseid identity for the trackmaker of cf.

332   Dimetropus isp. from Mallorca becomes feasible; specifically, a relatively small-sized form with

333   digits III and IV subequal in length, similar to Ennatosaurus. Nevertheless, further osteological

334   discoveries are needed to confirm this attribution, especially of autopodial material of more small

335   forms of caseids and also edaphosaurs.

336

337   5.4. Locomotion of the trackmaker of cf. Dimetropus isp. from Mallorca

                                                       12
338           Despite the abundant record of upper Palaeozoic ichnites (e.g., Lucas, 2019), detailed

339   studies of functionality and locomotion are all relatively recent (Voigt et al., 2007; Nyakatura et

340   al., 2015, 2019; Romano et al., 2016, 2020; Marchetti et al., 2017; Buchwitz & Voigt, 2018;

341   Mujal & Marchetti, 2020; Buchwitz et al., 2021; Logghe et al., 2021). Therefore, although the

342   general patterns of locomotion are well understood, work remains to be done on the evolutionary

343   timing of the locomotion style of different clades. In this sense, a detailed study of the distribution

344   of trackway parameters and the depth patterns of the ichnites and associated expulsion rims makes

345   it possible to infer different phases in the step cycle and the mode of locomotion of cf. Dimetropus

346   isp. from Mallorca, thus adding new insight on the locomotion patterns of primitive “pelycosaur”-

347   grade synapsids.

348           The general functional interpretation for the manus track of cf. Dimetropus isp. from

349   Mallorca differs little from that offered by Romano et al. (2016) for D. osageorum. In fact,

350   expulsion rims of different height arranged around the tracks further complement this

351   interpretation. In the touch-down phase, the weight was mostly applied on the pad below the

352   ulnare, proximally to digit V (lateral functional prevalence). In the manus tracks of cf. Dimetropus

353   isp. from Mallorca, this is not only represented by a relatively deep imprint of the posterolaterally

354   part of the palm, but also by a high expulsion rim located around it (Figures 2B–E, 5A). After

355   that, the weight shifted towards the pad below the radiale and the metacarpophalangeal pads of

356   digits I, II and III (medial-median functional prevalence) during the kick-off phase, making them

357   the deepest-imprinted areas of the track (Figure 5B). Digit drag traces (Figures 2B–C, E, 5C) were

358   also formed during this phase evidencing an initial movement of the manus towards the medial-

359   anterior part.

360           The functional interpretation of the pes track of cf. Dimetropus isp. from Mallorca is, on

361   the other hand, quite different to that of D. osageorum by Romano et al. (2016). Possibly, the

362   most noteworthy trait of the pes tracks of cf. Dimetropus isp. from Mallorca is their strong lateral

363   rotation, coupled with an additional lateral curvature of the imprints of digits III–V. It is likely

364   that, during locomotion, the trackmaker used an alternate lateral bending of the trunk, as inferred

365   for other pelycosaurs (Hopson, 2015). It is herein interpreted that in the touch-down phase, there

                                                        13
366   was an initial dragging of the tips of digits IV and V (and, to a lesser extent, III), and then the

367   weight was applied on the pad below the astragalus and the metatarsophalangeal pads of digits

368   IV and V (lateral functional prevalence), developing a moderately high expulsion rim between

369   digits IV and V, and around the posterior region of the ichnite (Figures 2A–C, 3A–B, 5D). Shortly

370   after, and as a readjustment of the body weight when the trunk was bent towards the opposite

371   side, the pes slightly rotated medially and the imprints of digits IV, V and, to a lesser extent, III,

372   became curved laterally (Figure 5G). This movement obscured the original imprints of digits III–

373   V and the posterolateral part of the sole. In the kick-off phase, the weight shifted towards the

374   anterior part of the autopodium, specifically, on the metatarsophalangeal pads of digits I and II

375   (medial functional prevalence), where the deepest impression of the track was left (Figures 2A–

376   C, 3A–B, 5E). The fact that the anteromedial margins of the pes digit imprints have a much gentler

377   slope than the posterolateral margins (Figure 2B–C) further supports that the pes was risen off the

378   ground with an anteromedial movement (Figure 5G). Drag traces of digits I (in some cases

379   imprinted twice) and II (Figures 2C, 3A, 5F) reveal an initial anterolateral movement of the pes

380   when it was taken off the ground (Figure 5H).

381           Therefore, in both the manus and the pes imprints, the functional prevalence of the

382   autopodia changed from being lateral in the touch-down phase (posterolateral region of the

383   ichnites) to being medial-median in the kick-off phase (anterior region of the ichnites, being the

384   deepest imprinted part overall), similarly to what Romano et al. (2016) observed in D. osageorum.

385   As interpreted herein, the gait would have had lateral sequence (Hildebrand, 1980), that is, the

386   sequence of movement of manus and pedes would have alternated the left and right side of the

387   body of the trackmaker (Figure 5G–I).

388

389   5.5. Notes on the evolution of synapsid locomotion

390           The particular functional prevalence inferred for these purported caseid tracks may

391   provide new insights on the evolution of the synapsid mode of locomotion. Anamniotes and basal

392   amniotes (i.e., seymouriamorphs, diadectomorphs, captorhinomorphs) produced icnites that

393   showed a medial-median functional prevalence of their autopodia (Mujal et al., 2020). On the

                                                       14
394   other hand, eupelycosaurs, which include all the more derived synapsids, seem to homogeneously

395   feature a median-lateral functional prevalence of their autopodia (Mujal et al., 2020). Therefore,

396   a transition between these two states of functional prevalence probably occurred between those

397   basal amniotes and the eupelycosaurs. Caseids, the purported trackmakers of the ichnites studied

398   here, correspond to a clade that diverged quite early in the synapsid tree (Ford & Benson, 2020).

399   Their tracks, as interpreted herein, show a mixture of lateral (posterior part) and medial-median

400   (anterior part) functional prevalence of their autopodia, which could be referred to the transitional

401   state referred above.

402           Caseids with mesaxonic autopodia, such as Ennatosaurus and Cotylorhynchus, nest

403   highly within the phylogeny of caseasaurs (Maddin et al., 2008). This could appear to conflict

404   with the foregoing hypotheses, as the most derived caseids would have the “intermediate” state

405   of functional prevalence argued above, whereas small, primitive forms would not. Therefore, it

406   must be stressed that the present work does not necessarily consider mesaxony as a primitive state

407   of caseids, and the fact that two mesaxonic track morphotypes (cf. Dimetropus isp. from Mallorca

408   and D. osageorum) show a combination of posterolateral and anteromedial-median deeper areas

409   does not exclude ectaxonic tracks having a similar pattern. It must be noted, however, that the

410   relative depth patterns of small tracks attributed to Dimetropus have not been exhaustively

411   described in the literature so far, and in fact, they are known only from five specimens from the

412   West Midlands, the Thüringer Wald, the Massif Central and the Pyrenees (Meade et al., 2016;

413   Mujal et al., 2016a, 2020; Marchetti et al., 2019a). Since Dimetropus is thought to have been

414   produced by a wide variety of “pelycosaur”-grade synapsids, it is likely that there are forms with

415   different types of relative depth patterns among the pool of specimens from different ages and

416   formations. Further research is needed to confirm this aspect, but it would be expected for a part

417   of the typical D. leisnerianus to show a relative depth pattern similar to that of cf. Dimetropus

418   isp. from Mallorca and D. osageorum. Especially interesting would be to assess this aspect in the

419   Carboniferous Dimetropus specimens (e.g., Van Allen et al., 2005; Voigt & Ganzelewski, 2010;

420   Meade et al., 2016; Lagnaoui et al., 2018; Marchetti et al., 2019c), as they may show other traits



                                                       15
421   closer to the primitive state of synapsids, some of them already pointed out by Buchwitz et al.

422   (2021).

423

424   6. Conclusions

425             The present work describes a trackway and several isolated ichnites from the lower

426   Permian of Mallorca (Balearic Islands, western Mediterranean) identified as cf. Dimetropus isp.

427   This form differs from Dimetropus leisnerianus and allied forms because cf. Dimetropus isp. from

428   Mallorca has mesaxonic manus and pes imprints, the pes imprints are strongly rotated laterally

429   (outwards), and the deepest regions of both manus and pes imprints are located laterally in their

430   posterior half and medial-median in their anterior half. It differs from Dimetropus osageorum

431   because cf. Dimetropus isp. from Mallorca has long and slender manus and pes digit imprints,

432   and pes imprints that are strongly rotated laterally (outwards). It has been here suggested that the

433   trackmaker of cf. Dimetropus isp. from Mallorca can probably be attributed to caseid synapsids,

434   based on the proportions of the autopodia and its relative depth pattern being similar to that of D.

435   osageorum, which was also purportedly produced by caseids. The general relative depth pattern

436   (which corresponds to the functional prevalence of the autopodia of the trackmaker) of cf.

437   Dimetropus isp. of Mallorca, which is deeper in the lateral region of the posterior half and in the

438   medial-median region of the anterior half, appears to be transitional from the primitive state of

439   anamniotes and basal amniotes (medial-median functional prevalence), to the derived state of

440   eupelycosaur synapsids (lateral-median functional prevalence). This has important implications

441   for the evolution of the mode of locomotion throughout the synapsid line, as caseids, occupying

442   a basal position in the synapsid tree, probably produced footprints with a relative depth pattern

443   (and thus functional prevalence of the autopodia) that was intermediate between that of

444   anamniotes/basal amniotes and eupelycosaurs.

445

446   Acknowledgements

447             To Tomeu Sáez, who kindly let us study and make a 3D model of the trackway specimen

448   in his collection, as well as the precise indications on its discovery and exact stratigraphic position.

                                                        16
449   To Sebastià Matamalas for aid in fieldwork. To the reviewers, Michael Buchwitz and Hendrik

450   Klein, for their critical review of the manuscript that substantially improved it. To Prof. Howard

451   Falcon-Lang, editor, for reviewing and handling the manuscript. We acknowledge support from

452   the CERCA program (Generalitat de Catalunya, Spain). R.M.A. was supported by the predoctoral

453   grant FPU17/01922 (Ministerio de Ciencia, Innovación y Universidades, Spain). À.G. is member

454   of the consolidated research group 2017 SGR 1666 GRC (Generalitat de Catalunya, Spain). J.F.

455   is member of the consolidated research group 2017 SGR 086 GRC (Generalitat de Catalunya,

456   Spain). We acknowledge support from the project “Mallorca abans dels dinosaures: estudi dels

457   ecosistemes continentals del Permià i Triàsic amb especial èmfasi en les restes de vertebrats” (ref.

458   15 - 619/2020), based at the ICP and funded by the Departament de Cultura, Patrimoni i Política

459   Lingüística (Consell Insular de Mallorca).

460

461   References

462   Angielczyk, K.D. & Kammerer, C.F. (2018). Non-Mammalian synapsids: the deep roots of the

463           mammalian family tree. In: Zachos, F. & Asher, R. (eds.). Mammalian Evolution,

464           Diversity and Systematics. De Gruyter, 117–198.

465   Benoit, J.; Ford, D.P.; Miyamae, J.A. & Ruf, I. (2021). Can maxillary canal morphology inform

466           varanopid phylogenetic affinities? Acta Palaeontologica Polonica, 66(2): 389–393.

467   Berman, D.S.; Maddin, H.C.; Henrici, A.C.; Sumida, S.S.; Scott, D. & Reisz, R.R. (2020). New

468           primitive caseid (Synapsida, Caseasauria) from the early Permian of Germany. Annals of

469           Carnegie Museum, 86(1): 43–75.

470   Buchwitz, M. & Voigt, S. (2018). On the morphological variability of Ichniotherium tracks and

471           evolution of locomotion in the sistergroup of amniotes. PeerJ, 6:e4346.

472   Buchwitz, M.; Marchetti, L.; Jansen, M.; Falk, D.; Trostheide, F. & Schneider, J.W. (2020).

473           Ichnotaxonomy and trackmaker assignment of tetrapod tracks and swimming traces from

474           the middle Permian Hornburg Formation of Saxony-Anhalt (Germany). Annales

475           Societatis Geologorum Poloniae, 90: 291–320.



                                                      17
476   Buchwitz, M.; Jansen, M.; Renaudie, J.; Marchetti, L. & Voigt, S. (2021). Evolutionary change

477          in locomotion close to the origin of amniotes inferred from trackway data in an ancestral

478          state reconstruction approach. Frontiers in Earth Science, 9:674779.

479   Calafat, F. (1988). Estratigrafía y sedimentología de la litofacies Buntsandstein de Mallorca.

480          Bachelor’s Thesis. Universitat de Barcelona. 127 pp. + 31 annexed figs.

481   Calafat, F.; Fornós, J.J.; Marzo, M.; Ramos, E. & Rodríguez-Perea, A. (1986). Icnología de

482          vertebrados de las facies Buntsandstein de Mallorca. In: Cabrera, L. (ed.). XI Congreso

483          Español de Sedimentología. Libro de Resúmenes: 40.

484   Calafat, F.; Fornós, J.J.; Marzo, M.; Ramos-Guerrero, E. & Rodríguez-Perea, A. (1986–1987).

485          Icnología de vertebrados de las facies Buntsandstein de Mallorca. Acta Geològica

486          Hispànica, 21–22: 515–520.

487   Falcon-Lang, H.J.; Rygel, M.C.; Calder, J.H. & Gibling, M.R. (2004). An early Pennsylvanian

488          waterhole deposit and its fossil biota in a dryland alluvial plain setting, Joggins, Nova

489          Scotia. Journal of the Geological Society, London, 161: 209–222.

490   Falcon-Lang, H.J.; Benton, M.J. & Stimson, M. (2007). Ecology of earliest reptiles inferred from

491          basal Pennsylvanian trackways. Journal of the Geological Society, London, 164: 1113–

492          1118.

493   Ford, D.P. & Benson, R.B.J. (2020). The phylogeny of early amniotes and the affinities of

494          Parareptilia and Varanopidae. Nature Ecology & Evolution, 4: 57–65.

495   Gand, G. (1988). Les traces de vertébrés tétrapodes du Permien français: paléontologie,

496          stratigraphie, paléoenvironnements. Ph.D. Thesis. Université de Bourgogne. 368 pp.

497   Gand, G. & Haubold, H. (1984). Traces de Vertèbrés tetrapodes du Permien du Bassin de Saint-

498          Affrique (Description, datation, comparaison avec celle du bassin de Lodève). Géologie

499          Méditerranéenne, 11(4): 321–348.

500   Gand, G.; De La Horra, R.; Galán-Abellán, B.; López-Gómez, J.; Barrenechea, J.F.; Arche, A. &

501          Benito, M.I. (2010). New ichnites from the Middle Triassic of the Iberian Ranges (Spain):

502          palaeoenvironmental and paleogeographical implications. Historical Biology, 22(1–3):

503          40–56.

                                                    18
504   Geinitz, H.B. (1863). Beiträge zur Kenntnis der organischen Überreste in der Dyas. Neues

505          Jahrbuch für Mineralogie, Geologie und Paläontologie, 1863: 385–398 + pl. III–IV.

506   Gómez-Gras, D. (1993). El Permotrías de las Baleares y de la vertiente mediterránea de la

507          Cordillera Ibérica y del Maestrat: Facies y Petrología Sedimentaria (Parte II). Boletín

508          Geológico y Minero, 104: 467–515.

509   Hasiotis, S.T.; Platt, B.F.; Hembree, D.I. & Everhart, M.J. (2007). The Trace-Fossil Record of

510          Vertebrates. In: William Miller, III (ed.). Trace Fossils: Concepts, Problems, Prospects.

511          Elsevier: 196–218.

512   Haubold, H. (1971). Ichnia Amphibiourum et Reptiliorum fossilium. In: Kuhn, O. (ed.).

513          Handbuch der Paläoherpetologie. Gustav Fischer Verlag, Stuttgart. 124 pp.

514   Haubold, H. & Sarjeant, W.A.S. (1973). Tetrapodenfährten aus den Keele und Enville Groups

515          (Permokarbon: Stefan und Autun) von Shropshire und South Staffordshire,

516          Großbritannien. Zeitschrift für Geologische Wissenschaften, 1: 895–933.

517   Haubold, H.; Hunt, A.P.; Lucas, S.G. & Lockley, M.G. (1995).Wolfcampian (Early Permian)

518          vertebrate tracks from Arizona and New Mexico. New Mexico Museum of Natural History

519          and Science Bulletin, 6: 135–165.

520   Hildebrand, M. (1980). The adaptive significance of tetrapod gait selection. American Zoologist,

521          20(1): 255–267.

522   Hopson, J.A. (2015). Fossils, trackways, and transitions in locomotion: a case study of

523          Dimetrodon. In: Dial, K.P.; Shubin, N. & Brainerd, E.L. (eds.). Great Transformations

524          in Vertebrate Evolution. University of Chicago Press, Chicago & London. 424 pp.

525   Kemp, T.S. (2005). The Origin and Evolution of Mammals. Oxford University Press, New York.

526          331 pp.

527   Klein, H.; Lucas, S.G. & Voigt, S. (2015). Revision of the ?Permian-Triassic tetrapod ichnogenus

528          Procolophonichnium Nopcsa 1923 with description of the new ichnospecies P. lockleyi.

529          Ichnos, 22: 155–176.




                                                    19
530   Lagnaoui, A.; Voigt, S.; Belahmira, A.; Saber, H.; Klein, H.; Hminna, A. & Schneider, J.W.

531           (2018). Late Carboniferous Tetrapod Footprints from the Souss Basin, Western High

532           Atlas Mountains, Morocco. Ichnos, 25(2–3), 81–93.

533   Leonardi, G. (1987). Glossary and Manual of Tetrapod Footprint Palaeoichnology.

534           Departamento Nacional de Produção Mineral, Brasília. 117 pp.

535   Liebrecht, T.; Fortuny, J.; Galobart, A.; Müller, J. & Sander, P.M. (2017). A large, multiple-tooth-

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

537           (Balearic Islands, Western Mediterranean). Journal of Vertebrate Paleontology, 37(1):

538           e1251936, 1–6.

539   Logghe, A.; Mujal, E.; Marchetti, L.; Nel, A.; Pouillon, J.-M.; Giner, S.; Garrouste, R. & Steyer,

540           J.-S. (2021). Hyloidichnus trackways with digit and tail drag traces from the Permian of

541           Gonfaron (Var, France): New insights on the locomotion of captorhinomorph eureptiles.

542           Palaeogeography, Palaeoclimatology, Palaeoecology, 573: 110436.

543   Lucas, S.G. (2019). An ichnological perspective on some major events of Paleozoic tetrapod

544           evolution. Bollettino della Società Paleontologica Italiana, 58(3): 223–266.

545   Lucas, S.G.; Kollar, A.D.; Berman, D.S. & Henrici, A.C. (2016). Pelycosaurian-grade (Amniota:

546           Synapsida) footprints from the lower Permian Dunkard Group of Pennsylvania and West

547           Virginia. Annals of Carnegie Museum, 83(4): 287–294.

548   Maddin, H.C.; Sidor, C.A. & Reisz, R.R. (2008). Cranial anatomy of Ennatosaurus tecton

549           (Synapsida: Caseidae) from the Middle Permian of Russia and the evolutionary

550           relationships of Caseidae. Journal of Vertebrate Paleontology, 28: 160–180.

551   Marchetti, L.; Ronchi, A.; Santi, G.; Schirolli, P. & Conti, M.A. (2015). Revision of a classic site

552           for Permian tetrapod ichnology (Collio Formation, Trompia and Caffaro valleys, N.

553           Italy), new evidences for the radiation of captorhinomorph footprints. Palaeogeography,

554           Palaeoclimatology, Palaeoecology, 433: 140–155.

555   Marchetti, L.; Mujal, E. & Bernardi, M. (2017). An unusual Amphisauropus trackway and its

556           implication for understanding seymouriamorph locomotion. Lethaia, 50: 162–174.



                                                      20
557   Marchetti, L.; Belvedere, M.; Voigt, S.; Klein, H.; Castanera, D.; Díaz-Martínez, I.; Marty, D.;

558           Xing, L.; Feola, S.; Melchor, R.N. & Farlow, J.O. (2019d). Defining the morphological

559           quality of fossil footprints. Problems and principles of preservation in tetrapod ichnology

560           with examples from the Palaeozoic to the present. Earth-Science Reviews, 193: 109–145.

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

562           & Groenewald, G.H. (2019a). Permian-Triassic vertebrate footprints from South Africa:

563           Ichnotaxonomy, producers and biostratigraphy through two major faunal crises.

564           Gondwana Research, 72: 139–168.

565   Marchetti, L.; van der Donck, H.; van Hylckama Vlieg, M. & During, M.A.D. (2019b). Leaving

566           only trace fossils-the unknown visitors of Winterswijk. Staringia, 16: 250–257.

567   Marchetti, L.; Voigt, S.; Lucas, S.G.; Francischini, H.; Dentzien-Dias, P.; Sacchi, R.; Mangiacotti,

568           M.; Scali, S.; Gazzola, A.; Ronchi, A. & Millhouse, A. (2019c). Tetrapod ichnotaxonomy

569           in eolian paleoenvironments (Coconino and De Chelly formations, Arizona) and late

570           Cisuralian (Permian) sauropsid radiation. Earth-Science Reviews, 190: 148–170.

571   Marchetti, L.; Voigt, S. & Klein, H. (2019e). Revision of Late Permian tetrapod tracks from the

572           Dolomites (Trentino-Alto Adige, Italy). Historical Biology, 31(6): 748–783.

573   Matamales-Andreu, R.; Fortuny, J.; Mujal, E. & Galobart, À. (2019). Tetrapod tracks from the

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

575           biogeographic inferences. In: The Palaeontological Association, 63rd annual meeting,

576           15th–21st December 2019: 107.

577   Meade, L.E.; Jones, A.S. & Butler, R.J. (2016). A revision of tetrapod footprints from the late

578           Carboniferous of the West Midlands, UK. PeerJ, 4: e2718.

579   Mujal, E. & Marchetti, L. (2020). Ichniotherium tracks from the Permian of France, and their

580           implications for understanding the locomotion and palaeobiogeography of large

581           diadectomorphs. Palaeogeography, Palaeoclimatology, Palaeoecology, 547:109698.

582   Mujal, E.; Fortuny, J.; Oms, O.; Bolet, A.; Galobart, À. & Anadón, P. (2016a).

583           Palaeoenvironmental reconstruction and early Permian ichnoassemblage from the NE

584           Iberian Basin (Pyrenean Basin). Geological Magazine, 153(4): 578–600.

                                                      21
585   Mujal, E.; Gretter, N.; Ronchi, A.; López-Gómez, J.; Falconnet, J.; Diez, J.B.; De la Horra, R.;

586           Bolet, A.; Oms, O.; Arche, A.; Barrenechea, J.F.; Steyer, J.-S. & Fortuny, J. (2016b).

587           Constraining the Permian/Triassic transition in continental environments: Stratigraphic

588           and palaeontological record from the Catalan Pyrenees (NE Iberian Peninsula).

589           Palaeogeography, Palaeoclimatology, Palaeoecology, 445: 18–37.

590   Mujal, E.; Marchetti, L.; Schoch, R.R. & Fortuny, J. (2020). Upper Paleozoic to lower Mesozoic

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

592           functional prevalence of autopodia. Frontiers in Earth Science, 8:248, 1–23.

593   Nyakatura, J.A.; Allen, V.R.; Lauströer, J.; Andikfar, A.; Danczak, M.; Ullrich, H.-J.; Hufenbach,

594           W.; Martens, T. & Fischer, M.S. (2015). A three-dimensional skeletal reconstruction of

595           the stem amniote Orobates pabsti (Diadectidae): analyses of body mass, centre of mass

596           position, and joint mobility. PLoS ONE, 10(9): e0137284.

597   Nyakatura, J.A.; Melo, K.; Horvat, T.; Karakasiliotis, K.; Allen, V.R.; Andikfar, A.; Andrada, E.;

598           Arnold, P.; Lauströer, J.; Hutchinson, J.R.; Fischer, M.S. & Ijspeert, A.J. (2019). Reverse-

599           engineering the locomotion of a stem amniote. Nature, 565: 351–355.

600   Ramos, A. (1995). Transition from alluvial to coastal deposits (Permian–Triassic) on the Island

601           of Mallorca, western Mediterranean. Geological Magazine, 132(4): 435–447.

602   Romano, M.; Citton, P. & Nicosia, U. (2016). Corroborating trackmaker identification through

603           footprint funcional analysis: the case study of Ichniotherium and Dimetropus. Lethaia,

604           49: 102–116.

605   Romano, M.; Brocklehurst, N. & Fröbisch, J. (2017). The postcranial skeleton of Ennatosaurus

606           tecton (Synapsida, Caseidae). Journal of Systematic Palaeontology, 16(13): 1097–1122.

607   Romano, M.; Citton, P. & Nicosia, U. (2020). A review of the concepts of ‘axony’ and their

608           bearing on tetrapod ichnology. Historical Biology, 32(5): 611–619.

609   Romer, A.S. & Price, L.I. (1940). Review of the Pelycosauria. Geological Society of America

610           Special Papers, 28: 1–538.




                                                      22
611   Sacchi, E.; Cifelli, R.; Citton, P.; Nicosia, U. & Romano, M. (2014). Dimetropus osageorum n.

612           isp. from the Early Permian of Oklahoma (USA): A Trace and its Trackmaker. Ichnos,

613           21: 175–192.

614   Sevillano, A. & Barnolas, A. (2019). Mapa Geológico Digital Continuo E. 1:50.000, Zona

615           Mallorca (Zona-2210). In: GEODE. Mapa Geológico Digital continuo de España.

616           Available online at: http://info.igme.es/visorweb/ (last accessed on 26/07/2019).

617   Spindler, F.; Falconnet, J. & Fröbisch, J. (2016). Callibrachion and Datheosaurus, two historical

618           and previously mis-taken basal caseasaurian synapsids from Europe. Acta

619           Palaeontologica Polonica, 61(3): 597–616.

620   Spindler, F.; Werneburg, R. & Schneider, J.W. (2019). A new mesenosaurine from the lower

621           Permian of Germany and the postcrania of Mesenosaurus: implications for early amniote

622           comparative osteology. Paläontologische Zeitschrift, 93: 303–344.

623   Tilton, J.L. (1931). Permian vertebrate tracks in West Virginia. Bulletin of the Geological Society

624           of America, 42: 547–556.

625   Van Allen, H.E.K.; Calder, J.H. & Hunt, A.P. (2005). The trackway record of a tetrapod

626           community in a walchian conifer forest from the Permo-Carboniferous of Nova Scotia.

627           New Mexico Museum of Natural History and Science Bulletin, 30: 322–332.

628   Voigt, S. (2005). Die Tetrapodenichnofauna des kontinentalen Oberkarbon und Perm im

629           Thüringer Wald - Ichnotaxonomie, Paläoökologie und Biostratigraphie. Cuvillier Verlag,

630           Göttingen. 308 pp.

631   Voigt, S. (2012). Tetrapodenfährten im Rotliegend. In: Lützner, H. & Kowalczyk, G. (eds.).

632           Stratigraphie von Deutschland X. Rotliegend. Teil I: Innervariscische Becken.

633           Schriftenreihe der Deutschen Gesellschaft für Geowissenschaften, 61: 161–175.

634   Voigt, S. & Ganzelewski, M. (2010). Toward the origin of amniotes: Diadectomorph and synapsid

635           footprints from the early Late Carboniferous of Germany. Acta Palaeontologica

636           Polonica, 55(1): 57–72.

637   Voigt, S.; Berman, D.S. & Henrici, A.C. (2007). First well-established track-trackmaker

638           association of paleozoic tetrapods based on Ichniotherium trackways and diadectid

                                                      23
639          skeletons from the Lower Permian of Germany. Journal of Vertebrate Paleontology,

640          27(3): 553–570.

641   Voigt, S.; Lagnaoui, A.; Hminna, A.; Saber, H. & Schneider, J.W. (2011). Revisional notes on

642          the Permian tetrapod ichnofauna from the Tiddas Basin, central Morocco.

643          Palaeogeography, Palaeoclimatology, Palaeoecology, 302: 474–483.

644   Voigt, S.; Niedźwiedzki, G.; Raczyński, P.; Mastalerz, K & Ptaszyński, T. (2012). Early Permian

645          tetrapod ichnofauna from the Intra-Sudetic Basin, SW Poland. Palaeogeography,

646          Palaeoclimatology, Palaeoecology, 313–314: 173–180.




                                                    24
647   Figures

648   Figure 1. Geographical and geological context for the Racó de s’Algar tracksite, Permian of

649   Mallorca (Balearic Islands, western Mediterranean). A: Simplified geological map of Mallorca

650   (based on Sevillano & Barnolas, 2019), with indication of the location of the studied site. B:

651   Simplified stratigraphic log of the Permian of Mallorca (thicknesses based on Gómez-Gras,

652   1993), distinguishing the three main lithostratigraphic units and indicating the location of the

653   studied site. C: Detailed stratigraphic log of the Racó de s’Algar tracksite, indicating the position

654   of the two track-bearing beds and the dominant sedimentary structures and fossil content. D: Field

655   photograph of the Racó de s’Algar tracksite with indication of the main sedimentary environments

656   and the track-bearing beds. See person for scale. E: Close-up photograph of the Racó de s’Algar

657   tracksite showing the two beds with ichnites. See hammer for scale (30 cm), pointed by an

658   arrowhead.

659

660   Figure 2. Selection of the best preserved specimens of cf. Dimetropus isp. studied herein, all as

661   textured 3D models, false colour-coded 3D models with contours and interpretative drawings. A:

662   TS-3, natural cast of a trackway with two left manus-pes sets and a right manus-pes set. B: Close-

663   up of the natural cast of the third manus-pes set of TS-3. C: Close-up of the natural cast of the

664   second manus-pes set of TS-3. D: NA-26-05, natural cast of a left manus imprint. E: DA21/15-

665   05-01, natural cast of the right manus imprint. All from the lower Permian Racó de s’Algar

666   tracksite (Mallorca, Balearic Islands, western Mediterranean), Permian.

667

668   Figure 3. Other specimens attributed to cf. Dimetropus isp., both as textured 3D models, false

669   colour-coded 3D models with contours and interpretative drawings. A: NA-26-02, natural cast of

670   a left manus-pes set. B: NA-26-03, natural cast of a right pes imprint. Both from the lower

671   Permian Racó de s’Algar tracksite (Mallorca, Balearic Islands, western Mediterranean), Permian.

672

673   Figure 4. Schematic drawings of morphotypes of Permian and Triassic tetrapod tracks attributed

674   to non-mammalian synapsids. They are idealised and based on the best-preserved tracks, arranged

                                                       25
675   symmetrically in the usual trackway disposition of each form. A: cf. Dimetropus isp. from

676   Mallorca. B: Dromopus lacertoides, based on Voigt (2005) and Buchwitz et al. (2021). C:

677   Tambachichnium schmidti, based on Voigt (2005). D: Dimetropus isp., based on Voigt &

678   Ganzelewski (2010). E: cf. Dimetropus isp., based on Marchetti et al. (2019d). F: Dimetropus

679   leisnerianus, based on Voigt (2005). G: Dimetropus leisnerianus, based on Voigt (2005). H:

680   Dimetropus leisnerianus, based on Voigt (2005). I: Dimetropus isp., based on Gand (1988). J:

681   Dimetropus osageorum, based on Sacchi et al. (2014). K: Brontopus antecursor, based on Gand

682   et al. (2000). L: Brontopus giganteus, based on Marchetti et al. (2019a). M: Dolomitipes accordii,

683   based on Marchetti et al. (2019e). N: Dolomitipes icelsi, based on Marchetti et al. (2019a). O:

684   Karoopes gansfonteinensis, based on Marchetti et al. (2019a). P: Capitosauroides bernburgensis,

685   based on Marchetti et al. (2019a). Q: Capitosauroides isp., based on Buchwitz et al. (2020). R:

686   Procolophonichnium haarmuehlensis, based on Klein et al. (2015). S: Dicynodontipus geinitzi,

687   based on Marchetti et al. (2019a).

688

689   Figure 5. Functional interpretation of cf. Dimetropus isp. from Mallorca. A: Manus track in

690   touch-down phase. B: Manus track in kick-off phase. C: Manus track after the forefoot was lifted.

691   D: Pes track in touch-down phase. E: Pes track in kick-off phase. F: Pes track after the hindfoot

692   was lifted. G–I: Reconstruction of the step cycle of the trackmaker (based on Ennatosaurus tecton

693   illustrated by Romano et al., 2017), with indication of the main movements occurring. G: Initial

694   phase of the step cycle, rotating the autopodia slightly medially and lining the pectoral and pelvic

695   girdles perpendicular to the body axis to gain a forward thrust. H: Swing phase moving the limbs

696   to the new position simultaneously to bending the trunk and finishing the movement of the girdles.

697   I: Final phase of the step cycle.




                                                      26
698   Figure 1




699




                 27
700   Figure 2




701

702   Figure 3




703




                 28
704   Figure 4




705




                 29
706   Figure 5




707




                 30
708   Table 1. Measurements taken on the studied ichnite specimens of cf. Dimetropus isp. from
709   Mallorca. Numbers in brackets correspond to parameters that have not been confidently measured
710   on poorly preserved ichnites. Hyphens (-) mark measurements impossible to obtain. In the
711   divarication angles, positive values correspond to ichnites rotated medially (inwards), whereas
712   negative values (−) correspond to ichnites rotated laterally (outwards).
                                                Parameters of the TRACKS
                                                                                                           NA-
                     DA21/1                                                                                         NA-
       Slab                                               TS-3                              NA-26-02       26-
                     5-05-01                                                                                       26-05
                                                                                                            03
       Figure          2D                              2A–C                                      3A        3B       2E
       Ichnite          -        a         b          c           d         e        f      a      b         -       -
                                                                 Left              Right   Lef    Left             Left
                     Right     Right     Right       Left                 Right                            Righ
       Anatomy                                                   manu              manu     t     manu             manu
                     manus      pes      manus       pes                   pes                             t pes
                                                                  s                  s     pes     s                 s
       Length
                      7.14     (7.78)    (8.21)      8.27        6.86     (7.37)   7.02     -     6.66       -     6.49
       (cm)
       Width
                      6.69     (13.46)   (13.15)    (8.35)       9.50     7.70     8.22     -     (7.11)     -       -
       (cm)
       Length
                      4.03     (5.25)       -        5.57        3.83     4.17     3.77     -     3.57       -     3.93
       palm (cm)
       Width
                      3.40     (8.64)       -        6.00        3.63     6.47     4.67     -     4.10       -     3.07
       palm (cm)
       Length I
                      1.57     (5.18)       -       (4.97)       2.53     2.66     2.90     -     2.43       -     3.20
       (cm)
       Length II
                      2.96     (5.68)       -         -          3.01     3.22     3.71     -     2.77       -     3.30
       (cm)
       Length III
                      3.30     (2.98)       -         -          3.59     3.61     3.92     -     3.49       -     3.83
       (cm)
       Length IV
                      2.57      3.79        -        4.17        (5.22)   (2.77)   3.26     -     3.00       -       -
       (cm)
       Length V
                     (2.04)     2.21      2.15       2.08        (1.62)     -      2.69     -     (1.40)     -       -
       (cm)
       Divergenc
       e angle I–     45.83    (4.18)    (63.57)     4.80        15.70    8.52     12.93    -     23.37      -     36.72
       II (º)
       Divergenc
       e angle II–    23.42    (76.54)      -        6.32        31.38    14.60    22.58    -     14.93      -     18.68
       III (º)
       Divergenc
       e angle        40.40    (3.27)    43.91      36.01        42.44    28.24    39.31    -     35.41      -     24.26
       III–IV (º)
       Divergenc
       e angle        37.18    (10.76)   45.08      41.51        35.98    25.86    44.45    -     68.96      -     48.03
       IV–V (º)
       Divergenc
                                                                 125.3             118.9          143.3            117.3
       e angle I–    147.41    (95.12)   152.45     89.19                 77.65             -                -
                                                                   9                 2              6                0
       V (º)
       Divergenc
       e angle II–    63.79    (80.12)   (42.06)    43.22        73.67    42.72    62.57    -     50.67      -     42.72
       IV (º)
       Divergenc
                                                                 110.3             106.1          120.0
       e angle II–   101.07    (90.40)   (88.97)    84.76                 68.73             -                -     91.38
                                                                   4                 6              0
       V (º)
       Divaricati
                               (−90.7    (−23.0     −31.2                 −22.2
       on from          -                                        9.01              26.10    -         -      -       -
                                 0)        3)         6                     5
       midline (º)
713

714




                                                             31
715   Table 2. Mean trackway measurements taken on the studied specimens of cf. Dimetropus isp.
716   from Mallorca. Brackets mean that that parameter has been partially inferred. Hyphens (-) mark
717   measurements impossible to obtain.
                                       Parameters of the TRACKWAY
       Slab                              DA21/15-05-01 TS-3 NA-26-02      NA-26-03   NA-26-05
       Figure                                 2D          2A–C    3A        3B         2E
       Pace angulation pes (º)                 -          83.15    -         -          -
       Pace angulation manus (º)               -          93.90    -         -          -
       Stride pes (cm)                         -          34.63    -         -          -
       Stride manus (cm)                       -          38.64    -         -          -
       Pace pes (cm)                           -          26.33    -         -          -
       Pace manus (cm)                         -          27.31    -         -          -
       Length pace pes (cm)                    -          17.22    -         -          -
       Length pace manus (cm)                  -          21.05    -         -          -
       Width pace pes (cm)                     -          19.57    -         -          -
       Width pace manus (cm)                   -          14.17    -         -          -
       Interpes distance (cm)                  -          12.07    -         -          -
       Intermanus distance (cm)                -           6.00    -         -          -
       Pes-manus distance (cm)                 -          14.32 (13.12)      -          -
       Width external (cm)                     -          26.91    -         -          -
       Width internal (cm)                     -           6.00    -         -          -
       Glenoacetabular distance (cm)           -          32.53    -         -          -
718




                                                         32