The bony labyrinth of late Permian Biarmosuchia: palaeobiology and diversity in non-mammalian Therapsida Julien Benoit1*, Paul R. Manger2, Vincent Fernandez3 & Bruce S. Rubidge1 1Evolutionary Studies Institute (ESI), School of Geosciences, University of the Witwatersrand, PO Wits, Johannesburg, 2050 South Africa 2School of Anatomical Sciences, University of the Witwatersrand, 7 York Road, Parktown, Johannesburg, 2193 South Africa 3European Synchrotron Radiation Facility, 71 rue des Martyrs, 38000 Grenoble, France Received 11 April 2017. Accepted 20 July 2017 INTRODUCTION Understanding the evolution of the different parts of the ear region is central to understanding synapsid and mam- malian palaeobiology (Stokstad 2003; Walsh et al. 2009; Ekdale 2013; Benoit et al. 2013a,b,c; Kemp 2016; Luo et al. 2016). The classic well-established evolutionary series of transformation of the middle ear from the postdentary- bones continues to attract much scientific interest (Allin 1975; Rubidge & Sidor 2001; Kemp 2005, 2016; Luo 2011; Luo et al. 2016; Ramírez-Chaves et al. 2016). Evolution of the inner ear, on the other hand, can be partly recon- structed from casts of the bony labyrinth, i.e. the osseous capsule that housed the inner ear inside the skull (Stokstad 2003; Ekdale 2013; Benoit et al. 2013a–d; Georgi et al. 2013; Walsh et al. 2013). New X-ray computed tomo- graphic scanning techniques (CT scanning) have enabled reconstruction of the internal cast of the bony labyrinth allowing palaeoneurologists to more easily trace the evo- lution of bony labyrinth morphology in the fossil record. As the bony labyrinth houses the organs of both balance and hearing, its morphology facilitates the accurate recon- struction of various palaeobiological traits in extinct tetrapods. These include head posture (Girard 1929; Vidal et al. 1986; Stokstad 2003; Sereno et al. 2007; Benoit et al., in press), locomotion (Spoor et al. 2002, 2007; Stokstad 2003; Silcox et al. 2009), auditory capabilities (West 1985; Meng & Fox 1995; Manoussaki et al. 2008; Walsh et al. 2009; Laaß 2015a,b, 2016), adaptations for life under water (Spoor et al. 2002; Neenan & Scheyer 2012; Georgi et al. 2013; Benoit et al. 2013a), and social group size and communica- tion (Walsh et al. 2009; Benoit et al. 2013c). In addition, the structure of the bony labyrinth has been used to recon- struct the phylogeny of extinct species (Lebrun et al. 2010; Benoit et al. 2013c, 2015). Because of its relevance to mammalian evolution, bony labyrinth morphology has been documented for most groups of non-mammalian Therapsida (NMT) (Fig. 1), particularly that of cynodonts and dicynodonts (e.g. Olson 1944; Cox 1962; Keyser 1965; Sigogneau 1974; Luo 2001; Kielan-Jaworowska et al. 2004; Ivakhnenko 2008; Castanhinha et al. 2013; Rodrigues et al. 2013; Laaß & Schillinger 2015; Laaß 2015a,b, 2016; Benoit et al., in press). Despite this, the structure of the bony labyrinth of the 58 ISSN 2410-4418 Palaeont. afr. (2017) 52: 58–77 *Author for correspondence. E-mail: julien.benoit@wits.ac.za Palaeontologia africana 52: 58–77 — ISSN 2410-4418 [Palaeontol. afr.] Online only Permanently archived on the 1st of August 2017 at the University of the Witwatersrand, Johannesburg, South Africa Both the supplement and the article are permanently archived at: http://wiredspace.wits.ac.za/handle/10539/23023 Biarmosuchia, as the basalmost group of Therapsida (the stem group of mammals), are important for understanding mammalian origins and evolution. Unlike other therapsid groups, the bony labyrinth of biarmosuchians has not yet been studied, despite insightful clues that bony labyrinth morphology can provide to address palaeobiology and phylogeny of extinct animals. Here, using CT scanning, surface reconstruction and a 3D geometric-morphometric protocol of 60 semi-landmarks on the bony labyrinth of 30 therapsids (including three Mammaliaformes), it is demonstrated that bony labyrinth morphology of biarmosuchians is very distinctive compared to that of other therapsids. Despite the primitive nature of their cranial morphology, biarmosuchians display highly derived traits in the structure of the bony labyrinth. The most noticeable are the presence of a long and slender canal linking the vestibule to the fenestra vestibuli, an enlarged and dorsally expanded anterior canal, and the absence of a secondary common crus (except for one specimen), which sets them apart from other non-mammalian therapsids. These characters provide additional support for the monophyly of Biarmosuchia, the most recently recognized major therapsid subclade. Although implications of the derived morphology of the biarmosuchian bony labyrinth are discussed, definitive interpretations are dependent on the discovery of well-preserved postcranial material. It nevertheless sheds light on a previously overlooked diversity of bony labyrinth morphology in non-mammalian therapsids. Keywords: Biarmosuchia, Therapsida, bony labyrinth, inner ear, geometric-morphometric. Palaeontologia africana 2017. ©2017 Julien Benoit, Paul R. Manger, Vincent Fernandez & Bruce S. Rubidge. This is an open-access article published under the Creative Commons Attribution 4.0 Unported License (CC BY4.0). To view a copy of the license, please visit http://creativecommons.org/licenses/by/4.0/. This license permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Both the supplement and the article are permanently archived at: http://wiredspace.wits.ac.za/handle/10539/23023 http://wiredspace.wits.ac.za/handle/10539/23023 http://creativecommons.org/licenses/by/4.0/ http://wiredspace.wits.ac.za/handle/10539/23023 mailto:julien.benoit@wits.ac.za Biarmosuchia, the most basal major therapsid clade (Rubidge & Sidor 2001; Liu et al. 2009), has not yet been described and,yet it is of great importance for understand- ing the most distant evolutionary origin of mammalian traits at the deep root of the therapsid clade. Before Biarmosuchia was recognized as a distinct clade of Therapsida (Hopson & Barghusen 1986; Sigogneau- Russell 1989; Rubidge & Sidor 2001) they were considered to be gorgonopsians (Sigogneau 1970). Few biarmo- suchian genera are represented by more than the holotype material, and many of these specimens are poorly preserved or distorted (e.g. Rubidge & Sidor 2001, 2002; Sidor 2003; Sidor & Welman 2003; Sidor et al. 2004; Smith et al. 2006; Sidor & Rubidge 2006; Rubidge et al. 2006; Sidor & Smith 2007; Ivakhnenko 2008; Kruger et al. 2015; Kammerer et al. 2016; Day et al. 2016). The application of various recently developed imaging techniques could elucidate palaeobiological characteristics of biarmo- suchians, such as locomotion, which have not been addressed. Here, synchrotron and CT scanning, and an advanced protocol of semi-landmark-based geometric morphometrics (Perrier et al. 2016) were applied to quan- tify and compare the bony labyrinth morphology of five biarmosuchian genera to that of 24 other therapsids, shedding new light on the deep evolutionary roots of the therapsid bony labyrinth. MATERIALS AND METHODS Scanning For this study, the skulls of five biarmosuchians were scanned: two genera (Hipposaurus and Herpetoskylax) belonging to the basal paraphyletic non-burnetiamorph biamosuchians (‘Hipposauridae’ and ‘Ictidorhinidae’, ISSN 2410-4418 Palaeont. afr. (2017) 52: 58–77 59 Figure 1. Phylogeny of Therapsida showing the phylogenetic relationships of the Biarmosuchia studied here. Bony labyrinth redrawn after the refer- ences cited directly in the figure. Phylogeny after Rubidge & Sidor (2001). respectively) (Sidor & Rubidge 2001; Sidor & Smith 2007; Kruger et al. 2015; Day et al. 2016); the burnetiamorph Lemurosaurus (two skulls), and an additional specimen of a presumably new burnetiamorph species (SAM-PK- 11112, hereby refer to as indet. burnetiamorph) (see Fig. 1). The scans of the biarmosuchian skulls were compared to scans of 25 other synapsids, including five Anomo- dontia, two Gorgonopsia, four Therocephalia, one Dino- cephalia, ten non-mammaliaform Cynodontia and three Mammaliaforms. A list of the specimens scanned with taxonomic assignment, details of the scanning device, voxel size, and stratigraphic position is provided in Sup- plementary Data 1. Three-dimensional renderings of bony labyrinths were obtained using manual segmenta- tion under Avizo 8 (FEI VSG, Hillsboro OR, USA). Geometric morphometrics of 3D surfaces We used a protocol of 60 three dimensional semi- landmarks, modified from Perier et al. (2016), to character- ize the shape of the three semicircular canals (20 semi- landmarks per canal) of 30 therapsid skulls (Fig. 2A). The semi-landmarks were placed using the software ISE- MeshTools (Lebrun 2008; Lebrun, et al. 2010; http:// morphomuseum.com/meshtools). Using this software, curves were defined in the centre of the lumen of each semicircular canal, from the centre of its ampulla to its opposite limb. The curves were converted into 20 equidis- tant semi-landmarks, which were then analysed using MorphoTools (Specht et al. 2007; Lebrun 2008). Due to the conservative symmetry of the bony labyrinth in mammals (see Welker et al. 2009; Billet et al. 2012), the semi-landmark protocol was applied only to the left bony labyrinth. When damaged, deformed or not preserved, the left bony labyrinth was substituted by a mirror image of the right. Each specimen’s semi-landmark configuration was represented by its centroid size and by its multidimen- sional shape vector in a linearized Procrustes shape space, using generalized least-squares fitting (Rohlf 1990). These data were subsequently analysed using a principal components analysis (PCA) of shape (Dryden & Mardia 1998). Bony labyrinth measurements Measurements of the bony labyrinth (Table 1) were taken using Avizo 8 (VSG) software and following the protocol illustrated in Fig. 2. The length of the vestibule was measured, including the cochlear recess when present, but excluding the exceedingly long canal joining the fenestra vestibuli to the vestibule in some biarmo- suchians. The radius of curvature of a semicircular canal is half the mean of its width plus its length. The width of a semicircular canal was measured from the centre of the lumen of each opposing limb (Fig. 2C). The height was taken perpendicular to the respective width, and was measured as the greatest distance from the wall of the vestibule to the centre of the lumen of the canal (Fig. 2C). The length corresponds to the length of the canal taken at the centre of the lumen of the canal (Fig. 2C). The ventral expansion of the posterior semicircular canal (Vep) (equiv- alent to the sagittal labyrinthic index of Spoor & Zonneveld [1995]) was measured when the plane of the lateral semicircular canal was parallel to the horizon. According to Spoor & Zonneveld (1995) this measure is the ratio of the distances between the level of the ventralmost point of the lumen of the posterior semicircu- lar canal and the plane of the lateral semicircular canal, and that between the level of the dorsalmost point of the posterior canal and the plane of the lateral semicircular canal (Fig. 2B). The dorsal expansion of the anterior semi- circular canal (Dea) (equivalent to the extension of the anterior semicircular canal projecting to the dorsalmost point of the posterior semicircular canal of Schmelzle et al. [2007]) was measured when the plane of the lateral semi- circular canal was parallel to the horizon and the plane of the posterior semicircular canal was perpendicular to the field of view. This measure is the ratio of the distances between the level of the dorsalmost point of the lumen of the posterior semicircular canal and the level of the dorsalmost point of the lumen of the anterior one, and the length between the dorsalmost point of the lumen of the posterior semicircular canal and the ventralmost point of the anterior semicircular canal and the surface of the vesti- bule (Fig. 2B). The angle between the plane of the lateral semicircular canal and the main axis of the skull was measured as shown in figure 2D. For comparison of relative semicircular canal size (Supplementary Data 2) we used sources from the litera- ture to build a dataset of average semicircular canal radius of curvature and body masses for 280 extant and extinct species, including mammals, reptiles, amphibians, and NMT (see Supplementary Data 2 for details). Body mass was sourced from literature (Tables 1; Supplementary Data 2) but when not available it was estimated based on skull length using the equations of Benoit et al. (in press). The coefficient of agility was calculated using the equation of Spoor et al. (2007) and Silcox et al. (2009) using body mass and the average semicircular canal radius (Table 1). Estimations of the auditory range of detectable airborne sounds (predicted range of audible frequencies and mean best frequency) were calculated using the equations of Walsh et al. (2009). Bony labyrinth description It has long been established that the bony labyrinth morphology does not fully reflect that of the membranous labyrinth in extant species (Gray 1907, 1908). Structures such as the secondary common crus, sacculus and utri- culus on the bony labyrinth look very different from the corresponding structures on the membranous labyrinth, and others such as the osseous semicircular canals only approximate the morphology of their homologues on the membranous organ. For simplicity and readability, most of the literature (e.g. Olson 1944; Ekdale 2013; Ruf et al. 2009; Luo et al. 2011) use the membranous labyrinth termi- nology to describe the bony labyrinth of extinct species. Accordingly, only the bony labyrinth is described and discussed here, and no assumption is made on the morphology of the membranous labyrinth, unless speci- fied. 60 ISSN 2410-4418 Palaeont. afr. (2017) 52: 58–77 http://morphomuseum.com/meshtools http://morphomuseum.com/meshtools DESCRIPTIONS AND RESULTS Unlike most non-mammaliaform therapsids (NMT) in which the medial wall of the vestibule, common crus, ampullae, and anterior and lateral semicircular canals are not ossified (Olson 1944; Sigogneau 1974; Luo 2001; Kielan-Jaworowska et al. 2004; Ivakhnenko 2008; Castanhinha et al. 2013; Rodrigues et al. 2013; Laaß & Schillinger 2015; Laaß 2015a,b, 2016; Benoit et al., in press), the bony labyrinth of biarmosuchians is well-encapsu- lated. Only a small part of the vestibule and the semicircu- lar canal ampullae are medially opened. The vestibule is short in biarmosuchians. In all taxa except Hipposaurus, the fenestra vestibuli is located at the distal end of a long and thin osseous canal that goes through the thickened bone of the basicranium (Fig. 3). In contrast, the vestibule is short and conical in Hipposaurus, and the fenestra vestibuli communicates directly with the exterior surface, which is the most generalized condition among ISSN 2410-4418 Palaeont. afr. (2017) 52: 58–77 61 Figure 2. Methods. A, The protocol of 60 semi-landmarks used to perform the PCA on shape. B and C, measurement protocol of the bony labyrinth (B) and semicircular canals (C). D, Measurement of the angle between the skull and lateral semicircular canal. A, bony labyrinth of Lemurosaurus pricei (BP/1/816); B, bony labyrinth of Iguana sp.; C, drawing of an isolated semicircular canal; D, reconstruction of the bony labyrinth in the transparent skull of Lemurosaurus pricei (BP/1/816). Abbreviations: Amp, ampulla; Angle, angle between the skull and lateral semicircular canal; ASC, anterior semicircular canal; Axis, main axis of the skull; Dea, dorsal expansion of the anterior semicircular canal; Fv, fenestra vestibuli; HSc, semicircular canal height; LabL, length of the bony labyrinth; LSC, lateral semicircular canal; LSC plane, plane of the lateral semicircular canal; PSC, posterior semicircular canal; R, radius of curvature of the semicircular canal; ScC, semicircular canal; Vep, ventral expansion of the posterior canal; Vest, vesti- bule; VesL, length of the vestibule; VesW, width of the vestibule; WSc, semicircular canal width. 62 ISSN 2410-4418 Palaeont. afr. (2017) 52: 58–77 Ta bl e 1. M ea su re m en ts of th e sk ul ls an d bo ny la by ri nt hs ,a nd es tim at io n of he ar in g ra ng e, be st he ar in g fr eq ue nc y an d ag ili ty of th e th er ap si ds st ud ie d. A bb re vi at io ns :D ea ,d or sa le xp an si on of th e an te ri or ca na l; SC ,s em ic ir cu la rc an al ;S C R ,s em ic ir cu la rc an al ra di us ;V ep ,v en tr a ex pa ns io n of th e po st er io rc an al .I n th e ‘le ng th of bo ny la by ri nt h co lu m n’ ,n um be rs in br ac ke ts re pr es en tm ea su rm en ts in cl ud in g th e le ng th of th e sl en de r ca na ll ea di ng to th e fe ne st ra ve st ib ul i. In th e ‘V es tib ul e le ng th co lu m n’ ,n um be rs in br ac ke ts re pr es en tm ea su re m en ts of th e co ch le ar ca na li n m am m al s. Sk ul ll en gt h Ba si cr an ia la xi s Bo dy m as s Le ng th of bo ny Ve st ib ul e le ng th R at io ve st ib ul e/ H ea ri ng ra ng e af te r M ea n be st he ar in g le ng th la by ri nt h ba si cr an ia la xi s W al sh et al .2 00 9 fr eq ue nc y af te r (m m ) (m m ) (g ) (m m ) (m m ) (H z) W al sh et al .2 00 9 (H z) M os ch op s D in oc ep ha lia 34 0. 19 10 6. 27 32 73 67 36 .4 9 21 .8 1 0. 21 82 28 .0 0 46 80 .3 8 H ip po sa ur us Bi ar m os uc hi a 19 3. 92 67 .9 4 23 21 3 25 .4 5 9. 45 0. 14 78 24 .2 7 44 61 .3 8 H er pe to sk yl ax Bi ar m os uc hi a 11 2. 77 31 .0 5 44 90 16 .4 7 (1 7. 86 ) 6. 48 (9 .9 4) 0. 21 82 48 .1 6 46 91 .3 2 Le m ur os au ru s Bi ar m os uc hi a 75 .2 0 17 .8 5 13 27 9. 87 (1 5. 12 ) 2. 19 (7 .0 2) 0. 12 77 24 .1 3 44 07 .0 6 BP /1 /8 16 Le m ur os au ru s Bi ar m os uc hi a 11 8. 53 34 .2 1 52 19 16 .3 6 (1 7. 25 ) 6. 52 (7 .4 6) 0. 19 81 38 .6 0 46 31 .8 9 N M Q R 17 02 In de t. Bi ar m os uc hi a 11 9. 49 32 .2 1 53 48 17 .5 9 (2 3. 62 ) 7. 21 (1 3. 86 ) 0. 22 83 41 .6 1 47 42 .0 1 SA M -P K -1 11 12 Pr is te ro do n D ic yn od on tia 67 .6 5 23 .5 9 96 7 14 .2 2 6. 62 0. 28 86 88 .2 3 49 30 .0 4 In de t. G or go no ps ia 13 4. 69 45 .5 8 76 82 15 .1 2 8. 22 0. 18 80 75 .3 9 45 97 .6 0 BP /1 /1 55 Sc yl ac oc ep ha lu s G or go no ps ia 99 .2 6 34 .1 3 30 56 13 .6 1 7. 63 0. 22 83 39 .8 6 47 41 .0 6 C ho er os au ru s Th er oc ep ha lia 93 .0 7 31 .4 5 25 71 8. 17 3. 90 0. 12 77 32 .1 7 44 11 .4 2 M ic ro go m ph od on Th er oc ep ha lia 86 .9 6 36 .3 1 20 52 13 .7 7 7. 98 0. 22 83 16 .7 7 47 28 .5 4 Eu ch am be rs ia Th er oc ep ha lia 12 1. 04 ? 55 60 ? ? ? ? ? N H M U K 56 96 Eu ch am be rs ia Th er oc ep ha lia 83 .7 4 25 .7 1 18 32 11 .6 4 4. 50 0. 18 80 43 .6 3 45 80 .3 7 BP /1 /4 00 9 O liv ie ro su ch us Th er oc ep ha lia 96 .0 8 34 .7 2 27 70 ? ? 69 75 .2 0 40 00 .8 0 C yn os au ru s C yn od on tia 49 .7 0 17 .1 8 38 7 6. 22 2. 22 0. 13 77 64 .0 0 44 28 .6 9 BP /1 /1 56 3 C yn os au ru s C yn od on tia 11 5. 83 46 .1 8 48 68 12 .5 8 6. 43 0. 14 78 25 .1 5 44 61 .8 6 BP /1 /3 92 6 Th ri na xo do n C yn od on tia 70 .8 8 20 .2 0 11 12 8. 82 3. 41 0. 17 80 05 .6 8 45 59 .7 9 BP /1 /4 26 3 Th ri na xo do n C yn od on tia 72 .5 4 22 .7 9 11 92 5. 11 5. 21 0. 23 83 70 .7 47 57 .7 9 BP /1 /7 19 9 G al es au ru s C yn od on tia 93 .4 2 34 .3 4 26 29 10 .7 0 5. 34 0. 16 79 24 .4 4 45 15 .7 2 Tr ir ac ho do n C yn od on tia 95 .6 2 28 .2 4 27 30 10 .2 9 1. 57 0. 06 73 14 .5 7 41 84 .8 9 A M 46 1 Tr ir ac ho do n C yn od on tia 10 1. 45 33 .8 5 32 63 9. 82 6. 22 0. 18 80 95 .9 7 46 08 .7 7 BP /1 /4 65 8 Tr ir ac ho do nt id C yn od on tia 45 .5 4 12 .9 6 29 9 6. 74 2. 40 0. 19 81 05 .6 3 46 14 .0 0 in de t. Lu m ku ia C yn od on tia 59 .0 5 17 .3 7 64 5 6. 91 2. 25 0. 13 77 65 .9 1 44 29 .7 2 C on tin ue d on p. 63 ISSN 2410-4418 Palaeont. afr. (2017) 52: 58–77 63 Ta bl e 1 (c on tin ue d) Sk ul ll en gt h Ba si cr an ia la xi s Bo dy m as s Le ng th of bo ny Ve st ib ul e le ng th R at io ve st ib ul e/ H ea ri ng ra ng e af te r M ea n be st he ar in g le ng th la by ri nt h ba si cr an ia la xi s W al sh et al .2 00 9 fr eq ue nc y af te r (m m ) (m m ) (g ) (m m ) (m m ) (H z) W al sh et al .2 00 9 (H z) Ta ch yg lo ss us M am m al ia fo rm 10 8. 50 25 .7 1 39 96 9. 89 (9 .9 8) – – – O rn ith or hy nc hu s M am m al ia fo rm 85 .2 5 10 .7 0 19 33 8. 36 (7 .3 4) – – - M eg az os tr od on M am m al ia fo rm 29 .8 4 10 .4 4 87 4. 66 3. 10 0. 30 87 87 .7 8 49 84 .0 4 Ly st ro sa ur us D ic yn od on tia 10 1. 63 33 .6 7 32 81 30 .3 2 12 .2 1 0. 36 91 88 .8 5 52 01 .6 0 N M Q R 35 93 Ly st ro sa ur us D ic yn od on tia 10 1. 79 32 .2 6 32 96 28 .4 8 16 .8 0 0. 52 10 15 4. 13 57 25 .2 2 N M Q R 81 5 Pa tr an om od on A no m od on tia 63 .6 4 16 .3 5 80 6 10 .5 6 4. 91 0. 30 88 08 .3 6 49 95 .2 0 M as se to gn at hu s C yn od on tia 89 .1 8 21 .5 5 22 14 9. 33 5. 20 0. 24 84 48 .1 6 47 99 .8 1 Eo di cy no do n D ic yn od on tia 99 .7 7 35 .5 9 31 03 17 .4 1 7. 56 0. 21 82 71 .8 7 47 04 .1 8 Ve st ib ul e w id th A nt SC ra di us A nt SC le ng th La tS C ra di us La tS C le ng th Po st SC ra di us Po st SC le ng th Av er ag e SC R (m m ) (m m ) (m m ) (m m ) (m m ) (m m ) (m m ) (m m ) M os ch op s D in oc ep ha lia 14 .7 3 5. 73 20 .2 8 4. 27 12 .0 8 4. 06 12 .3 0 4. 69 H ip po sa ur us Bi ar m os uc hi a 9. 42 6. 07 22 .1 3 3. 52 9. 26 4. 42 18 .5 8 4. 67 H er pe to sk yl ax Bi ar m os uc hi a 5. 18 3. 76 14 .8 3 3. 03 9. 41 2. 25 10 .6 2 3. 01 Le m ur os au ru s Bi ar m os uc hi a 4. 41 3. 43 12 .0 0 2. 19 6. 10 2. 22 9. 17 2. 61 BP /1 /8 16 Le m ur os au ru s Bi ar m os uc hi a 4. 93 3. 87 13 .8 5 2. 34 6. 70 3. 06 12 .6 0 3. 09 N M Q R 17 02 In de t. Bi ar m os uc hi a 5. 08 4. 58 16 .1 6 3. 71 12 .7 7 3. 02 8. 66 3. 77 SA M -P K -1 11 12 Pr is te ro do n D ic yn od on tia 3. 67 2. 87 9. 29 2. 43 6. 58 2. 38 7. 54 2. 56 In de t. G or go no ps ia 4. 36 3. 15 10 .3 0 2. 22 8. 13 2. 34 7. 28 2. 57 BP /1 /1 55 Sc yl ac oc ep ha lu s G or go no ps ia 3. 86 2. 91 9. 88 ? ? 2. 08 6. 70 2. 50 C ho er os au ru s Th er oc ep ha lia 2. 73 2. 22 6. 89 1. 58 4. 76 1. 73 4. 77 1. 84 M ic ro go m ph od on Th er oc ep ha lia 3. 71 2. 46 7. 85 ? ? ? ? 2. 46 Eu ch am be rs ia Th er oc ep ha lia ? 3. 91 13 .3 1 2. 84 9. 36 2. 61 8. 57 3. 12 N H M U K 56 96 Eu ch am be rs ia Th er oc ep ha lia 4. 50 3. 13 9. 07 2. 75 8. 23 3. 07 9. 09 2. 98 BP /1 /4 00 9 O liv ie ro su ch us Th er oc ep ha lia ? ? ? 2. 94 ? ? ? 2. 94 C yn os au ru s C yn od on tia 2. 56 1. 74 5. 98 1. 18 3. 09 1. 24 4. 10 1. 39 BP /1 /1 56 3 C yn os au ru s C yn od on tia 4. 51 2. 65 9. 02 1. 63 4. 56 2. 16 7. 68 2. 14 BP /1 /3 92 6 C on tin ue d on p. 64 64 ISSN 2410-4418 Palaeont. afr. (2017) 52: 58–77 Ta bl e 1 (c on tin ue d) Ve st ib ul e w id th A nt SC ra di us A nt SC le ng th La tS C ra di us La tS C le ng th Po st SC ra di us Po st SC le ng th Av er ag e SC R (m m ) (m m ) (m m ) (m m ) (m m ) (m m ) (m m ) (m m ) Th ri na xo do n C yn od on tia 3. 23 1. 94 7. 78 1. 93 4. 81 1. 56 3. 88 1. 81 BP /1 /4 26 3 Th ri na xo do n C yn od on tia 3. 56 1. 96 6. 43 1. 50 4. 42 1. 79 5. 60 1. 75 BP /1 /7 19 9 G al es au ru s C yn od on tia 3. 81 2. 15 6. 98 1. 54 4. 06 1. 71 6. 03 1. 80 Tr ir ac ho do n C yn od on tia 1. 28 2. 82 9. 40 2. 17 6. 27 2. 46 7. 05 2. 48 A M 46 1 Tr ir ac ho do n C yn od on tia 5. 65 2. 73 9. 40 2. 20 6. 41 2. 13 6. 80 2. 35 BP /1 /4 65 8 Tr ir ac ho do nt id C yn od on tia 1. 43 1. 75 5. 59 1. 40 3. 85 1. 18 4. 04 1. 45 in de t. Lu m ku ia C yn od on tia 2. 35 1. 61 5. 98 1. 39 4. 31 1. 50 4. 96 1. 50 Ta ch yg lo ss us M am m al ia fo rm 2. 08 2. 31 7. 51 1. 45 5. 08 1. 81 6. 96 1. 86 O rn ith or hy nc hu s M am m al ia fo rm 1. 52 1. 77 5. 07 1. 31 4. 03 1. 75 4. 74 1. 61 M eg az os tr od on M am m al ia fo rm 1. 18 1. 93 2. 91 – – – – 1. 93 Ly st ro sa ur us D ic yn od on tia 5. 69 3. 78 14 .3 6 2. 96 10 .4 1 3. 82 13 .2 5 3. 52 N M Q R 35 93 Ly st ro sa ur us D ic yn od on tia 5. 95 4. 17 13 .9 1 3. 08 9. 54 3. 49 12 .5 7 3. 58 N M Q R 81 5 Pa tr an om od on A no m od on tia 3. 13 2. 99 10 .6 9 2. 26 7. 94 1. 68 7. 83 2. 31 M as se to gn at hu s C yn od on tia 4. 28 2. 42 8. 58 1. 80 6. 39 2. 08 6. 29 2. 10 Eo di cy no do n D ic yn od on tia 9. 08 4. 39 14 .7 8 2. 97 9. 06 3. 50 11 .3 9 3. 62 C oe ff ic ie nt Ve p D ea A ng le be tw ee n A ng le be tw ee n A ng le be tw ee n A ng le be tw ee n of ag ili ty an te ri or an d po st er io r an d an te ri or an d th e cr an ia la xi s la te ra lc an al s la te ra lc an al s po st er io r ca na ls an d th e pl an of th e th e ho ri zo nt al ca na l (° ) (° ) (° ) (° ) M os ch op s D in oc ep ha lia 4 – 0. 51 90 .1 0 73 .9 3 75 .6 9 76 H ip po sa ur us Bi ar m os uc hi a 6 1. 43 0. 78 86 .2 1 81 .0 3 93 .3 1 11 H er pe to sk yl ax Bi ar m os uc hi a 5 2. 15 0. 77 78 .6 7 72 .7 7 12 8. 47 26 Le m ur os au ru s Bi ar m os uc hi a 6 1. 51 0. 95 82 .6 2 88 .4 1 91 .9 7 37 BP /1 /8 16 Le m ur os au ru s N M Q R 17 02 Bi ar m os uc hi a 5 1. 30 1. 42 73 .1 9 75 .2 4 13 3. 54 25 In de t. Bi ar m os uc hi a 6 – 0. 54 77 .0 7 10 1. 12 82 .0 8 20 SA M -P K -1 11 12 Pr is te ro do n D ic yn od on tia 6 – 0. 24 80 .6 1 76 .7 2 77 .8 0 18 C on tin ue d on p. 65 ISSN 2410-4418 Palaeont. afr. (2017) 52: 58–77 65 Ta bl e 1 (c on tin ue d) C oe ff ic ie nt Ve p D ea A ng le be tw ee n A ng le be tw ee n A ng le be tw ee n A ng le be tw ee n of ag ili ty an te ri or an d po st er io r an d an te ri or an d th e cr an ia la xi s la te ra lc an al s la te ra lc an al s po st er io r ca na ls an d th e pl an of th e th e ho ri zo nt al ca na l (° ) (° ) (° ) (° ) In de t. G or go no ps ia 4 – 0. 19 76 .2 7 83 .2 4 86 .0 2 18 BP /1 /1 55 Sc yl ac oc ep ha lu s G or go no ps ia 5 – 0. 25 77 .5 4 74 .0 1 75 .8 4 33 C ho er os au ru s Th er oc ep ha lia 4 – 0. 28 84 .0 2 80 .0 9 97 .8 4 52 M ic ro go m ph od on Th er oc ep ha lia 5 ? ? 86 .3 7 87 .4 5 ? 32 Eu ch am be rs ia Th er oc ep ha lia 5 – 0. 16 78 .9 1 87 .7 4 76 .0 2 19 N H M U K 56 96 Eu ch am be rs ia Th er oc ep ha lia 6 – 0. 09 44 .9 4 10 2. 96 80 .3 6 30 BP /1 /4 00 9 O liv ie ro su ch us Th er oc ep ha lia 6 ? ? 78 .2 5 76 .8 0 ? 31 C yn os au ru s C yn od on tia 4 – 0. 25 76 .4 3 73 .4 8 81 .6 7 12 BP /1 /1 56 3 C yn os au ru s C yn od on tia 4 – 0. 46 77 .5 4 80 .5 7 95 .7 6 21 BP /1 /3 92 6 Th ri na xo do n C yn od on tia 4 – 0. 28 76 .5 4 93 .6 6 11 3. 94 29 BP /1 /4 26 3 Th ri na xo do n C yn od on tia 4 – 0. 07 80 .6 4 95 .5 1 94 .0 0 33 BP /1 /7 19 9 G al es au ru s C yn od on tia 4 – 0. 24 78 .4 0 81 .0 6 82 .1 2 60 Tr ir ac ho do n C yn od on tia 5 – 0. 00 88 .5 6 74 .8 0 73 .8 4 18 A M 46 1 Tr ir ac ho do n C yn od on tia 4 – 0. 07 78 .6 3 82 .1 3 82 .9 1 24 BP /1 /4 65 8 Tr ir ac ho do nt id C yn od on tia 4 – 0. 32 78 .1 1 86 .0 8 89 .6 1 22 in de t. Lu m ku ia C yn od on tia 4 – 0. 00 70 .2 3 75 .9 2 88 .6 9 23 Ta ch yg lo ss us M am m al ia fo rm 3 – 0. 16 65 .0 3 97 .4 0 10 0. 09 24 O rn ith or hy nc hu s M am m al ia fo rm 3 – 0. 24 70 .9 4 92 .0 2 91 .4 1 9 M eg az os tr od on M am m al ia fo rm 7 – – – – – 44 Ly st ro sa ur us D ic yn od on tia 7 – 0. 30 82 .8 9 75 .0 6 80 .0 8 –2 3 N M Q R 35 93 Ly st ro sa ur us D ic yn od on tia 7 – 0. 27 75 .9 9 76 .5 1 78 .8 7 –1 9 N M Q R 81 5 Pa tr an om od on A no m od on tia 6 – 0. 56 80 .9 1 90 .2 5 73 .4 9 21 M as se to gn at hu s C yn od on tia 4 – 0. 34 80 .7 7 67 .5 5 79 .0 5 24 Eo di cy no do n D ic yn od on tia 7 – 0. 20 80 .3 5 85 .7 6 79 .6 0 35 therapsids (Olson 1944; Cox 1962; Keyser 1965; Sigogneau 1974; Luo 2001; Kielan-Jaworowska et al. 2004; Ivakh- nenko 2008; Castanhinha et al. 2013; Rodrigues et al. 2013; Laaß & Schillinger 2015; Laaß 2015a,b, 2016; Benoit et al., in press). There is no cochlear recess, except in one specimen of Lemurosaurus (NMQR 1702) which shows a distinct papilla medial to the vestibule (Fig. 3E) comparable to those observed in some cynodonts, dicynodonts and therocephalians (Luo 2001; Kielan-Jaworowska et al. 2004; Castanhinha et al. 2013; Rodrigues et al. 2013; Laaß 2015a,b; Fig. 4D, G, I–K). There is no evidence for the separation of the sacculus and utriculus of the vestibule in any of our specimens. As in other NMT, the three ampullae are opened medially to the brain cavity in biarmosuchians, except in one Lemurosaurus (BP/1/816) in which the posterior ampulla only is not ossified medially. All three ampullae are promi- nent except in Herpetoskylax, in which the lateral ampulla is distinctly smaller (Fig. 3C). All NMT have a secondary common crus, which corresponds to the fusion of the ventral arm of the posterior canal with the posterior arm of the lateral canal (Olson 1944; Cox 1962; Keyser 1965; Sigogneau 1974; Luo 2001; Kielan-Jaworowska et al. 2004; Ivakhnenko 2008; Castanhinha et al. 2013; Rodrigues et al. 2013; Laaß & Schillinger 2015; Laaß 2015a,b, 2016; Benoit et al., in press; Fig. 4A–K). As a result, the entrance into the posterior ampulla is shared by both osseous canals. A short secondary common crus is present in the indet. burnetiamorph, with the two canals being separated until they fuse at the entry of the posterior ampulla (Fig. 3F), but not in the other biarmosuchians studied (Fig. 3). Instead, Lemurosaurus, Herpetoskylax, and Hipposaurus display an apomorphic condition in which there is no secondary common crus (Fig. 3). In these taxa, the lateral and posterior canals are fused for a short section, distal to the vestibule. However, instead of entering the posterior ampulla together as in other therapsids, the posterior canal projects ventrally and separates from the lateral canal (Fig. 3). Therefore, the lateral canal enters the vesti- bule dorsal to the posterior ampulla, independent of the posterior canal. In Lemurosaurus, the posterior canal enters the vestibule through the posterior ampulla (Fig. 3D, E), but in the basal biarmosuchians Hipposaurus and Herpetoskylax the posterior canal enters the vestibule directly, medial to the posterior ampulla (Fig. 3A, B). As a result, neither the posterior nor the lateral canals enter the vestibule through the posterior ampulla in these two taxa (Fig. 3A, B), which is a unique condition among therapsids and mammals (Olson 1944; Cox 1962; Keyser 1965; Sigogneau 1974; Luo 2001; Stokstad 2003; Kielan- Jaworowska et al. 2004; Ivakhnenko 2008; Castanhinha et al. 2013; Rodrigues et al. 2013; Ekdale 2013; Benoit et al. 2013a–d, 2015, in press; Georgi et al. 2013; Walsh et al. 2013; Laaß & Schillinger 2015; Laaß 2015a,b, 2016; Fig. 4). This ventral expansion of the posterior canal (Vep) is more pro- nounced in Herpetoskylax than in other biarmosuchians (Table 1). As in mammals, the anterior canal is the largest of the semicircular canals in NMT in terms of length and radius (Table 1) and encircles the floccular fossa of the endocast. In Lemurosaurus, Herpetoskylax, and the indet. burnetia- morph, the anterior canal is more than 50% larger than the posterior one, which is very large for a therapsid (Table 1). Moschops is the only other NMT in our sample to approach this condition, with the anterior canal about 40% larger than the posterior one (Table 1). The dorsal expansion of the anterior canal (Dea) also reaches more than 50% of the height of the posterior canal in biarmosuchians and Moschops, which indicates the great enlargement of this canal (Table 1). The radius of the posterior canal is also proportionately larger in Lemurosaurus, Herpetoskylax, and Hipposaurus because of the ventral expansion of the ventral posterior canal (Table 1). In the PCA of shape, most of the variation is explained by axis PC1 (34.03%). Axes PC2 and PC3 explain 27.79% and 10.89% of the variation, respectively. Shape differences on PC1 mostly affects the lateral canal and its relative position with respect to the ampullary arm of the poste- rior canal (presence or absence of a secondary common crus) (Fig. 5). Shape differences on PC2 mostly affect the relative size of the anterior canal and its dorsal projection (Fig. 5). All NMT are mixed in the scatter plots, but the Biarmosuchia is the only therapsid group that is isolated by the geometry of the semicircular canals (Fig. 5). Both axes distinguish biarmosuchians from other NMT, with the greatest degree of differenciation being on axis PC2. PC3 captures more minor differences in the angle between the common crus and the lateral canal, and the width of the lateral semicircular canal (Fig. 5). Four of the biarmosuchians fall outside the range of morphological variation of semicircular canal of other NMT (Fig. 5), which reflects the unique configuration of their semicircu- lar canals (e.g. the absence of secondary common crus, the Dea). One noticeable exception is the indet. burnetia- morph (SAM-PK-K11112) which appears close to a gorgonopsian (Scylacocephalus) but still stands apart from all other NMT on both scatter plots (Fig. 5). This position of SAM-PK-K11112 in the geometric morphometric analysis may reflect the presence of a secondary common crus in this specimen, which strongly affects the 3D conforma- tion of semicircular canals, as described above. The average radius of curvature of the semicircular canals is relatively larger in biarmosuchians, the dino- cephalian Moschops, anomodonts, the gorgonopsian Scylacocephalus, the cynodont Trirachodon (AM 461) and the therocephalians Euchamberia, Olivierosuchus and Micro- gomphodon, than in the indet. gorgonopsian, the thero- cephalian Choerosaurus and most cynodonts (Fig. 6). Like most non-cynodont therapsids (apart from one specimen of Trirachodon), the biarmosuchians have a quotient of agility above 5 (Table 1). Non-mammalian cynodonts and Moschops have a quotient value of 4 while the monotremes have a value of 3 (Table 1). The anomodonts have the larg- est semicircular canal radius compared to body size, with values reaching up to 7 (Table 1). DISCUSSION The morphology of the bony labyrinth in NMT is often assumed to be conservative with respect to that of mam- mals (Sigogneau 1974; Luo 2001; Laaß 2015; Angielczyk 66 ISSN 2410-4418 Palaeont. afr. (2017) 52: 58–77 ISSN 2410-4418 Palaeont. afr. (2017) 52: 58–77 67 Figure 3. The bony labyrinth of Biarmosuchia. A, an indeterminate gorgonopsia (BP/1/155) for outgroup comparison; B, Hipposaurus (CG-WB123); C, Herpetoskylax hopsoni (BP/1/3924). Abbreviations: Ant amp, anterior ampulla; ASC, anterior semicircular canal; CC, common crus; CCII, secondary common crus; Dea, dorsal expansion of the anterior semicircular canal; Floc., floccular (subarcuate) fossa; Fv, fenestra vestibuli; Lat amp, lateral ampulla; LSC, lateral semicircular canal; Post amp, posterior ampulla; PSC, posterior semicircular canal; Vep, ventral expansion of the posterior canal; Vest, vestibule; Vest tb, vestibular tube. The brain endocast is in transparent. Scale bar, 5 mm. Continued on p. 68 68 ISSN 2410-4418 Palaeont. afr. (2017) 52: 58–77 Figure 3 (continued) D, Lemurosaurus pricei (BP/1/816); E, Lemurosaurus pricei (NMQR 1702); F, indet. burnetiamorph (SAM-PK-K11112). ISSN 2410-4418 Palaeont. afr. (2017) 52: 58–77 69 Figure 4. The bony labyrinth in a variety of therapsids examined for this study. A, Scylacocephalus watermeyeri (BP/1/216); B, an indeterminate gorgonopsia (BP/1/155); C, Euchambersia mirabilis (BP/1/4009); D, Microgomphodon oligocynus (SAM-PK-10160). Abbreviations: ASC, anterior semicircu- lar canal; LSC, lateral semicircular canal; PSC, posterior semicircular canal. Continued on p. 70 70 ISSN 2410-4418 Palaeont. afr. (2017) 52: 58–77 Figure 4 (continued). E, Patranomodon nyaphulii (NMQR 3000); F, Eodicynodon oosthuizeni (NMQR 2978); G, Pristerodon sp. (BP/1/2642); H, Lystrosaurus declivis (NMQR 815). Continued on p. 71 ISSN 2410-4418 Palaeont. afr. (2017) 52: 58–77 71 Figure 4 (continued) I, Thrinaxodon liorhinus (BP/1/7199); J, Cynosaurus suppostus (BP/1/3926); K, Massetognathus pascuali (BP/1/4245); L, Tachyglossus aculeatus (MS86); M, Ornithorhynchus anatinus (BP/4/908). From left to right: lateral, posterior and dorsal view of the bony labyrinth. et al. 2016), similar to that of early ‘pelycosaur’ grade synapsids (Case 1914). Because there is no cochlear canal, the bony labyrinth in NMT shows little variation of vesti- bule morphology compared to that of mammals. In addi- tion, the radius of curvature of the semicircular canals are usually subequal, and the posterior and lateral canals usually share a secondary common crus (Olson 1944; Cox 1962; Keyser 1965; Sigogneau 1974; Luo 2001; Kielan- Jaworowska et al. 2004; Ivakhnenko 2008; Castanhinha et al. 2013; Rodrigues et al. 2013; Laaß & Schillinger 2015; Laaß 2015a,b, 2016; Benoit et al., in press; Fig. 4). The unique and autapomorphic bony labyrinth morphology of biarmosuchians demonstrate that bony labyrinth morphology in therapsids is in fact more diverse than previously recognized. The large dorsal expansion of the anterior semicircular canal (Dea), absence of a secondary common crus in all but one of our specimens (SAM-PK- 11112), and the presence of a long and slender canal join- ing the vestibule to the distant fenestra vestibuli (except in Hipposaurus) clearly set biarmosuchians apart as a distinct group of NMT. The status of the Biarmosuchia as a higher taxon was recognized only recently (Sigogneau-Russell 72 ISSN 2410-4418 Palaeont. afr. (2017) 52: 58–77 Figure 5. Results of the PCA on shape. Figure 6. Comparison of average semicircular canal radius of curvature and body mass in tetrapods (Supplementary Data 2). 1989; Hopson & Barghusen 1986; Rubidge & Sidor 2001) and, given the highly autapomorphic morphology of their bony labyrinth, this study of their bony labyrinth provides further support to their clade status. These highly distinctive characters provide an opportunity to consider the palaeobiological implications of these new data. Cochlear recess and auditory capabilities We here describe the presence of a recess medial to the vestibule in the biarmosuchian Lemurosaurus (NMQR 1702) which could correspond to a cochlear recess. In biarmosuchians more derived than Hipposaurus (Fig. 1), the vestibule elongates ventrally into a slender canal leading to the fenestra vestibuli (Fig. 3), but this is probably not homologous to the cochlear recess since the cochlear recess does not open distally into the fenestra vestibuli (Luo 2001). The cochlear recess is the antecedent of the coiled cochlear canal of mammals and was previously recog- nized only in cynodonts (i.e. Cynosaurus, Thrinaxodon, Massetognathus, Probelesodon, Probainognathus, Yunnanodon and Brasilitherium) for which it constitutes a synapo- morphy (Luo 2001; Kielan-Jaworowska et al. 2004; Rodrigues et al. 2013; Fig. 4I–K). In contrast, despite substantial ossification in this area, a cochlear recess is absent in most Biarmosuchia and Anomodontia, and all Dinocephalia and Gorgonopsia studied so far (Olson 1944; Sigogneau 1974; Araujo et al. 2017; Benoit et al., in press). The probable presence of the cochlear recess in the biarmosuchian Lemurosaurus broadens the phylogen- etic occurrence of this character among therapsids. Among our comparative sample, a cochlear recess is also present in the therocephalian Microgomphodon but not in other therocephalians in which the vestibule is weakly ossified medially (Olson 1944; Fig. 4D). Recently, the presence of a cochlear recess was also hypothesized in two dicynodonts, Niassodon mfumukasi and Pristerodon sp. (Castanhinha et al. 2013; Laaß 2015a, 2016). In the cistecephalid Kawingasaurus fossilis, Laaß (2015b) identi- fied a dramatically enlarged vestibule, a feature strongly correlated with fossoriality, but found no cochlear recess. A cochlear recess is absent in the basal anomodonts Patranomodon and Eodicynodon (Fig. 4E, F). Therefore, apart from Cynodontia, a cochlear recess is present in iso- lated taxa only among Anomodontia (two occurences), Therocephalia (one occurence) and Biarmosuchia (one occurence). In addition to its absence in more basal repre- sentatives of these clades, this strongly suggests that the presence of a distinct and ossified cochlear recess evolved multiple times among therapsids. Just as coiling of the membranous cochlea is not reflected by the bony cochlear canal in monotremes (Gray 1908), the absence of an ossi- fied cochlear recess does not imply that the correspond- ing membranous structure, the basilar papilla, was absent. For example, a small membranous basilar papilla is present in sarcopterygian fishes but does not reflect on their bony labyrinth (Fritzsch 1987; Manley 2012). More data are required to address the evolution of this structure in NMT, but given the presence of a possible cochlear recess in a basal species such as Lemurosaurus, it may be hypothesized that an enlarged basilar papilla, which is ancestral to the membranous cochlea, was already present in the last common ancestor of therapsids. Estimations of auditory capability based on vestibular length (Table 1) show that the presence or absence of a cochlear recess does not greatly influence the predicted range of detectable airborne frequencies or the mean best hearing frequency in therapsids (Table 1). According to Walsh et al.’s equation (Walsh et al. 2009), the predicted hearing capabilities of air-borne sound of therapsids averages 4–5 kHz ± 4–5 kHz (Table 1), which overlaps the highest values of those measured in extant reptiles (Walsh et al. 2009). These values are less than 20 kHz, which corresponds to the lower limit of frequencies that only mammals can detect (Manley 2012). Althought there are some uncertainties that may affect the accuracy of this prediction (e.g. the nature and position of the tympanum in NMT and the effect of the morphology of the stapes [Gaetano & Abdala 2015; Kemp 2016]), it remains consis- tent with current consensus that mammals evolved high frequency hearing capabilities only later in the Mesozoic (Meng & Fox 1995; Laaß 2015a, 2016; Kemp 2016; Luo et al. 2016). Semicircular canals and locomotion Compared to other therapsids, the Biarmosuchia have a disproportionately large anterior semicircular canal (Fig. 3; Table 1). An enlarged anterior canal is not uncom- mon among tetrapods, including mammals (Gray 1907, 1908; Walsh et al. 2009; Ekdale 2013), but it is rare among NMT and elsewhere has only been documented in the dinocephalian Moschops and Brasilitherium (Fig. 3; Table 1; Rodrigues et al. 2013; Benoit et al., in press). In biarmo- suchians and Moschops a prominent dorsal expansion of the anterior canal (Dea) is also present (Table 1). The morphology of the anterior canal could be linked to that of the floccular fossa and the corresponding paraflocular lobe of the cerebellum (Sanchez-Villagra 2002). A larger flocular fossa may result in a larger anterior canal (Sanchez-Villagra 2002), but in dinocephalians the floccular fossa is shallow, unlike the condition in biarmosuchians and most other NMT (Fig. 3; Laaß & Kaestner 2017; Benoit et al., in press). Actually, there seems to be no direct link between the presence of a large and deep flocular fossa and the size of the anterior canal in NMT (Olson 1944; Cox 1962; Keyser 1965; Sigogneau 1974; Luo 2001; Kielan-Jaworowska et al. 2004; Ivakhnenko 2008; Castanhinha et al. 2013; Rodrigues et al. 2013; Laaß & Schillinger 2015; Laaß 2015a,b, 2016; Araujo et al. 2017; Fig. 4) though, this would need to be quantitatively addressed. Semicircular canals record the rotation of the head in three dimensions in tetrapods (Graf & Klam 2006; Malinzak et al. 2012). It is thus hypothesized that their size is correlated to head mobility and agility in tetrapods (Spoor et al. 2002; Stokstad 2003; Clarke 2005; Georgi et al. 2013) and enlargement of the anterior canal would reflect greater vertical mobility of the head (Clarke 2005). How- ever, the body masses of dinocephalians and biarmo- suchians differ greatly (Table 1), and the orientation of the ISSN 2410-4418 Palaeont. afr. (2017) 52: 58–77 73 anterior semicircular canal with respect to the skull in Moschops and biarmosuchians is also very different. Indeed, the angle between the cranial axis and the plane of the horizontal canal approximates 80° in Moschops whereas in Biarmosuchians it smaller (Table 1). This may reflect a difference in neutral head posture (Benoit et al., in press). Lystrosaurus is remarkable in this respect as it is the only NMT to display an angle value below zero, which means the head was likely held high (Table 1). The differ- ences in both body mass and bony labyrinth orientation makes it difficult to imagine that a common pattern of head motion are the reason for the similarity in anterior semicircular canal enlargement in Moschops and biarmo- suchians. Herpetoskylax, Hipposaurus and Lemurosaurus are the only NMT currently known that do not have a complete secondary common crus (Fig. 3B, C; Olson 1944; Cox 1962; Keyser 1965; Sigogneau 1974; Luo 2001; Kielan- Jaworowska et al. 2004; Ivakhnenko 2008; Castanhinha et al. 2013; Rodrigues et al. 2013; Laaß & Schillinger 2015; Laaß 2015a,b, 2016; Benoit et al., in press; Fig. 4A–K). Fusion of the osseous lateral and posterior semicircular canals into a secondary common crus, and fact that both canals share the same ampulla (the posterior ampulla) to enter the vestibule, is a primitive feature for therapsids and mammals (Ekdale 2013; Luo et al. 2011) and is present in all other therapsids and Mesozoic mammals studied to date (e.g. Henkelotherium, Haldanodon and Dryolestes) (Ruf et al. 2009, 2013; Luo et al. 2011). It is also present in the extant monotreme Ornithorhyncus (but not Tachyglossus, Fig. 4L, M), a variety of marsupials (e.g. Didelphys), as well as some placental mammals (e.g. Orycteropus, Equus) (Gray 1907 1908; Ekdale 2013; Benoit et al. 2015). Lemurosaurus, Herpetoskylax, and Hipposaurus depart from this ancestral anatomy since the posterior and lateral semicircular canals are not completely fused. In these taxa, the posterior canal projects ventrally and only merges with the lateral canal at the level where they cross each other (Fig. 3). This condition, though unique among NMT, is sporadically represented in some extant placental mammals (e.g. the Hyracoidea, Dasypus, some Macro- scelididae; Ekdale 2013; Benoit et al. 2013d) and reptiles (Fig. 2; Jones & Spells 1963 ; Ramprashad et al. 1986; Olori 2010). Herpetoskylax and Hipposaurus share a unique condition with no equivalent in any known extant or extinct species. Their posterior canal enters directly into the vestibule without passing through the corresponding posterior ampulla (Fig. 3). As a result, the ampulla appears isolated from both the lateral and posterior semicircular canals (Fig. 3). Another interpretation for this morphology would be to consider that the posterior ampulla is extremely small or absent and that the structure here identified as the posterior ampulla is neomorphic. Herpetoskylax and Hipposaurus are two of the most basal biarmosuchians (e.g. Sidor & Rubidge 2006; Kruger et al. 2015; Day et al. 2016). Given that this very particular condi- tion is not present in the basalmost burnetiamorph Lemurosaurus (Fig. 3D, E) and that a secondary common crus is present in the unidentified burnetiamorph SAM-PK-11112 (Fig. 3F), it appears possible that the secondary common crus displayed by the latter was re-acquired secondarily after it was lost in more basal species. In this case, the absence of a secondary common crus was only transitory in the evolution of biarmo- suchians. Another possibility would be that Herpetoskylax and Hipposaurus comprise a distinct clade, which is not supported by any phylogenetic analysis to date (Sidor & Welman 2003; Smith et al. 2006; Sidor & Smith 2007; Kruger et al. 2015; Kammerer et al. 2016; Day et al. 2016). Because of the lack of fossil material, the postcranial anatomy of biarmosuchians is poorly documented, limited to few limb bones and vertebrae in Biarmosuchus and Hipposaurus, and little is known about their locomo- tion (Olson 1962; Sigogneau 1970; Sues 1986; Signogneau- Russell 1989). It appears that their postcranial skeleton was similar to that of gorgonopsians and other basal therapsids, which were able to adopt both a sprawling posture or a more parasagittal gait, as opposed to epicynodonts, which were more permanently para- sagittal (Olson 1962; Jenkins 1971; Signogneau-Russell 1989; Kemp 2005). The average radius of semicircular canals indicates that biarmosuchians, along with Moschops, anomodonts, Scylacocephalus and some thero- cephalians did have a larger semicircular canal radius than other therapsids, including most cynodonts (Fig. 6). Indeed, all cynodonts have semicircular canal radii that fall within the range of variation observed in extant mammals, whereas biarmosuchians, anomodonts and many other therapsids are different (Fig. 6). Based on the correlation demonstrated between semi- circular canal radius, agility and locomotion in mammals (Spoor et al. 2002, 2007; Silcox et al. 2009), it is tempting to interpret this difference as reflecting a more mammal- like posture in the derived cynodonts (Jenkins 1971). In contrast, the more basal forms, such as biarmosuchians, anomodonts and dinocephalians, would have had a more sprawling posture. The fact that radius of curvature of the semicircular canals of basal theriodonts (gorgonopsian and therocephalians) are spread between these two groups possibly reflects an intermediate postural behav- iour (Fig. 6). However, the calculation of agility scores based on semi- circular radius (Spoor et al. 2007; Silcox et al. 2009) gives a different signal. It predicts that biarmosuchians, gorgo- nopsians and anomodonts were more agile than most cynodonts and Moschops (Table 1). On the one hand, these higher scores may reflect a more diverse range of locomotory possibilities permitted by the use of both a ‘reptile-like gait’ and ‘mammal-like gait’ in biarmo- suchians, gorgonopsians and anomodonts, whereas cynodonts and Moschops were more constrained to a single type of posture, perhaps more parasagittal. But on the other hand, this is inconsistent with the fossil record of footprints attributed to dinocephalians, which supports a partial sprawling posture (Gand et al. 2000; Surkov et al. 2007) and the hypothesis that some dicynodonts evolved a more upright posture independently because of the biomechanical constraints imposed by their large body size (Blob 2001; Fröbisch 2006). Finally, the range of semi- 74 ISSN 2410-4418 Palaeont. afr. (2017) 52: 58–77 ISSN 2410-4418 Palaeont. afr. (2017) 52: 58–77 75 circular canal radii of sprawling tetrapods overlaps that of mammals and NMT (Fig. 6). This means that, based on the average radius of the semicircular canals alone (and the so calculated coefficient of agility), it is impossible to discrim- inate between sprawling and parasagittal forms. Thus the question of whether the very derived semicircular canal anatomy of the biarmosuchians reflects a different type of locomotion will have to be addressed when well- preserved postcranial material becomes available. CONCLUSION With their large pachyostosed, ornamented cranial structures and intermeshing incisors, biarmosuchians constitute a fascinating, yet poorly understood radiation of early NMT (Rubidge & Sidor 2001; Kemp 2005; Day et al. 2016). Along with these traits, the study performed here has identified additional potential synapomorphies that support the clade. It also highlights an unexpected diver- sity of morphology of the bony labyrinth in NMT which, and along with other recent studies (Castanhinha et al. 2013; Rodrigues et al. 2013; Laaß & Schillinger 2015; Laaß 2015a,b, 2016; Benoit et al., in press), strongly suggests that the auditory system of different NMT taxonomic groups was not as homogenous as previously postulated (Luo 2001; Kielen Jaworowska et al. 2004; Kemp 2016; Luo et al. 2016). Finally, the unique morphology of the semicircular canals in biarmosuchians, with no extant equivalents, suggests a highly derived type of locomotion. Future studies and reappraisal of the scarce postcranial material of biarmosuchians will certainly shed some light on this very autapom INSTITUTIONAL ABBREVIATIONS AM Albany Museum, Grahamstown, South Africa BP/1 Evolutionary Studies Institute, Johannesburg, South Africa CG Council for Geosciences, Pretoria, South Africa MS School of Anatomical Sciences, Johannesburg, South Africa NHMUK Natural History Museum, London, United Kingdom NMQR National Museum, Bloemfontein, South Africa SAM-PK Iziko Museum, Cape Town, South Africa Thanks to F. Ahmed, S. Chapman (NHMUK, London), K. Jakata, L. Norton, S. Jirah, B. Zipfel (ESI, Johannesburg), H. Fourie (Ditsong National Museum of Natural His- tory, Pretoria), E. de Kock (Council for Geosciences, Pretoria), E. Butler, J. Botha-Brink (National Museum, Bloemfontein), R. Smith, Zeituna Erasmus (Iziko Museum, Cape Town), A. Du plessis (University of Stellenbosch), B. de Klerk, R. Prevec (Albany Museum, Grahamstown). We acknowledge the European Syn- chrotron Radiation Facility for provision of synchrotron radiation facilities and we would like to thank P. Tafforeau for assistance in using beamline ID17. This research was conducted with financial support from PAST and its Scatterlings of Africa Programmes; the NRF; and the DST-NRF Centre of Excellence in Palaeosciences (CoE in Palaeosciences). Opinions expressed and conclusions arrived at, are those of the authors and are not necessarily to be attributed to the CoE in Palaeosciences. The authors declare no competing interest. REFERENCES ANGIELCZYK, K.D., RUBIDGE, B.S., DAY, M.O. & LIN, F. 2016. A reevaluation of Brachyprosopus broomi and Chelydontops altidentalis, Dicynodonts (Therapsida, Anomodontia) from the Middle Permian Tapinocephalus Assemblage Zone of the Karoo Basin, South Africa. Journal of Vertebrate Paleontology 36(2), e1078342. ALLIN, E.F. 1975. Evolution of the mammalian middle ear. Journal of Mor- phology 147, 403–438. ARAUJO, R.M., FERNANDEZ, V., POLCYN, M.J., FRÖBISCH, J. & MARTINS & R.M.S. 2017. Aspects of the gorgonopsian paleobiology: insights from the basicranium, occiput, osseous labyrinth and neuroanatomy of the immature gorgonopsian skull of Aloposaurus gracilis (Therapsida: Theriodontia: Gorgonopsia). PeerJ Preprints 4, e2313v1. BENOIT, J., BEN HAJ ALI, M., ADNET, S., EL MABROUK, E., HAYET, K., MARIVAUX, L., MERZERAUD, G., MERIGEAUD, S., VIANEY-LIAUD & M., TABUCE, R. 2013a. Cranial remain from Tunisia provides new clues for the origin and evolution of Sirenia (Mammalia, Afrotheria) in Africa. PLOS ONE 8, e54307. BENOIT, J., MABROUK, E.E., LEBRUN, R., TABUCE, R. & MARIVAUX, L. 2013b. New insights into the ear region anatomy and cranial blood supply of stem Strepsirhini: evidence from three primate petrosals from the Eocene of Chambi, Tunisia. Journal of Human Evolution 65, 551–572. BENOIT, J., MERIGEAUD, S. & TABUCE, R. 2013c. Homoplasy in the ear region of Tethytheria and the systematic position of Embrithopoda (Mammalia, Afrotheria). Geobios 46, 357–370. BENOIT, J., CRUMPTON, N., MERIGEAUD, S. & TABUCE, R. 2013d. Petrosal and bony labyrinth morphology supports paraphyly of Elephantulus within Macroscelididae (Mammalia, Afrotheria). Journal of Mammalian Evolution 21, 173–193. BENOIT, J., LEHMANN, T., VATTER, M., LEBRUN, R., MERIGEAUD, S., COSTEUR, L. & TABUCE, R. 2015. Comparative anatomy and three dimensional geometric-morphometric study of the bony labyrinth of Bibymalagasia (Mammalia, Afrotheria). Journal of Vertebrate Paleontol- ogy 35(3), e930043. BENOIT, J., MANGER, P.R., NORTON, L., FERNANDEZ, V. & RUBIDGE, B.S. Synchrotron scanning reveals the palaeoneurology of the head-butting Moschops capensis (Therapsida, Dinocephalia). PeerJ, in press. BILLET, G., HAUTIER, L., ASHER, R., SCHWARZ, C., CRUMPTON, N., MARTIN, T. & RUF, I. 2012. High morphological variation of vestibular system accompanies slow and infrequent locomotion in three-toed sloths. Proceedings of the Royal Society B 279, 3932–3939. CASE, E.C. 1914. On the structure of the inner ear in two primitive rep- tiles. Biological Bulletin 27, 213–216. BLOB, R.W. 2001. Evolution of hindlimb posture in nonmammalian therapsids: biomechanical tests of paleontological hypotheses. Paleo- biology 27, 14–38. CASTANHINHA, R., ARAÚJO, R., JÚNIOR, L. C., ANGIELCZYK, K. D., MARTINS, G.G., MARTINS, R.M.S., CHAOUIYA, C., BECKMANN, F. & WILDE, F. 2013. Bringing dicynodonts back to life: paleobiology and anatomy of a new emydopoid genus from the Upper Permian of Mozambique. PLOS ONE 8(12), e80974. CLARKE, A.H. 2005. On the vestibular labyrinth of Brachiosaurus brancai. Journal of Vestibular Research 15, 65–71. COX, C.B. 1962. A natural cast of the inner ear of a dicynodont. American Museum Novitates 2116, 1–6. DAY, M.O., RUBIDGE, B.S. & ABDALA, F. 2016. A new mid-Permian burnetiamorph therapsid from the Main Karoo Basin of South Africa and a phylogenetic review of Burnetiamorpha. Acta Palaeontologica Polonica 61, 701–719. DRYDEN, I.L. & MARDIA, K.V. 1998. Statistical Shape Analysis. Chichester, John Wiley and Sons. EKDALE, E.G. 2013. Comparative anatomy of the bony labyrinth (inner ear) of placental mammals. PLOS ONE 8(6), e66624. FRÖBISCH, J. 2006. Locomotion in derived dicynodonts (Synapsida, Anomodontia): a functional analysis of the pelvic girdle and hind limb of Tetragonias njalilus. Canadian Journal of Earth Sciences 43, 1297–1308. FRITZSCH, B. 1987. The inner ear of the coelacanth fish Latimeria has tetrapod affinities. Nature 327, 153–154. GAETANO, L.C. & ABDALA, F. 2015. The stapes of gomphodont cyno- donts: insights into the middle ear structure of non-mammaliaform cynodonts. PLOS ONE 10, e0131174. GAND, G., GARRIC, J., DEMATHIEU, G. & ELLENBERGER, P. 2000. La Palichnofaune de vertébrés tétrapodes du Permien supérieur du Bassin de Lodève (Languedoc-France). Palaeovertebrata 29, 1–82. GEORGI, J.A., SIPLA, J.S. & FORSTER, C.A. 2013. Turning semicircular canal function on its head: dinosaurs and a novel vestibular analysis. PLOS ONE 8(3), e58517. GIRARD, L. 1929. L’attitude normale de la tête déterminée par le labyrinthe de l’oreille. Bulletins et Mémoires de la Société d’anthropologie de Paris, VII° Série 10, 79–99. GRAF, W.M. & KLAM, F. 2006. Le système vestibulaire: anatomie fonctionnelle et comparée, évolution et développement. Comptes Rendus Palevol 5, 637–655. GRAY, A.A. 1907. The Labyrinth of Animals, Vol. 1. London, Churchill. GRAY, A.A. 1908. The Labyrinth of Animals, Vol. 2. London, Churchill. HOPSON, J.A. & BARGHUSEN, H. 1986. An analysis of therapsid rela- tionships. In: Hotton, N. III, MacLean, P.D., Roth, J.J. & Roth, E.C. (eds), The Ecology and Biology of Mammal-like Reptiles, 83–106. Washington, Smithsonian Institution Press. IVAKHNENKO, M. F. 2008. Cranial morphology and evolution of Perm- ian Dinomorpha (Eotherapsida) of eastern Europe. Paleontological Journal 42, 859–995. JENKINS, F.A.Jr. 1971. The postcranial skeleton of African cynodonts. Bulletin of the Peabody Museum of Natural History, Yale University 36, 1–216. JONES, M.G. & SPELLS, K.E. 1963. A theoretical and comparative study of the functional dependence of the semicircular canal upon its physi- cal dimensions. Proceedings of the Royal Society B 157, 403–419. KEMP, T.S. 2005. The Origin and Evolution of Mammals. Oxford, Oxford University Press. KEMP, T.S. 2016. Non-mammalian synapsids: the beginning of the mam- mal line. In: Clack, J.A. et al. (eds), Evolution of the Vertebrate Ear – Evi- dence from the Fossil Record, 107–137. Springer Handbook of Auditory Research 59. Springer International Publishing AG. KIELAN-JAWOROWSKA, Z., CIFELLI, R. L. & LUO, Z-X. 2004. Mammals from the Age of Dinosaurs, Origins, Evolution, and Structure. New York, Columbia University Press. KAMMERER, C. F. 2016. Two unrecognised burnetiamorph specimens from historic Karoo collections. Palaeontologia africana 50, 64–75. KEYSER, A.W. 1965. The morphology of the anomodont genus Cistecephalus Owen (1876). M.Sc. thesis, Universiteit van Pretoria. KRUGER, A., RUBIDGE, B.S., ABDALA, F., CHINDEBVU, E.G. & JACOBS, L.L. 2015. Lende chiweta, a new therapsid from Malawi, and its influence on burnetiamorph phylogeny and biogeography. Journal of Vertebrate Paleontology 35(6), e1008698. LAAß, M. 2015a. Virtual reconstruction and description of the cranial endocast of Pristerodon mackayi (Therapsida, Anomodontia). Journal of Morphology 276, 1089–1099. LAAß, M. 2015b. Bone-conduction hearing and seismic sensitivity of the late Permian anomodont Kawingasaurus fossilis. Journal of Morphology 276, 121–143. LAAß, M. 2016. The origins of the cochlea and impedance matching hear- ing in synapsids. Acta Palaeontologica Polonica 61, 267–280. LAAß, M. & KAESTNER, A. 2017. Evidence for convergent evolution of a structure comparable to the mammalian neocortex in a Late Permian therapsid. Journal of Morphology 278, 1033–1057. LAAß, M. & SCHILLINGER, B. 2015. Reconstructing the auditory appa- ratus of therapsids by means of neutron tomography. Physics Procedia 69, 628–635. LEBRUN, R. 2008. Evolution and development of the strepsirrhine primate skull. Ph.D. thesis, Université de Montpellier II, Montpellier, France, University of Zürich, Zürich, Switzerland. LEBRUN, R., DE LEÓN, M.P., TAFFOREAU, P. & ZOLLIKOFER, C. 2010. Deep evolutionary roots of strepsirrhine primate labyrinthine mor- phology. Journal of Anatomy 216, 368–380. LIU, J., RUBIDGE, B. & LI, J. 2009. New basal synapsid supports Laurasian origin for therapsids. Acta Palaeontologica Polonica 54, 393–400. LUO, Z-X. 2001. Inner ear and its bony housing in tritylodonts and impli- cations for evolution of mammaliaform ear. Bulletin of the Museum of Comparative Zoology 156, 81–97. LUO, Z-X. 2011. Developmental patterns in Mesozoic evolution of mammal ears. Annual Review of Ecology, Evolution and Systematics 42, 355–380. LUO, Z-X., RUF, I., SCHULTZ, J.A. & MARTIN, T. 2011. Fossil evidence on evolution of inner ear cochlea in Jurassic mammals. Proceedings of the Royal Society B 278, 28–34. LUO, Z-X., SCHULTZ, J.A. & EKDALE, E.G. 2016. Evolution of the mid- dle and inner ears of mammaliaforms: the approach to mammals. Clack J.A. et al. (eds), Evolution of the Vertebrate Ear – Evidence from the Fossil Record, 107–137. Springer Handbook of Auditory Research 59. Springer International Publishing AG. MALINZAK, M., KAY, R. & HULLAR, T. 2012. Locomotor head move- ments and semicircular canal morphology in primates. Proceedings of the National Academy of Sciences USA 109, 17914–17919. MANLEY, G.A. 2012. Evolutionary paths to mammalian cochleae. Journal of the Association for Research in Otolaryngology 13, 733–743. MANOUSSAKI, D., CHADWICK, R.S., KETTEN, D.R., ARRUDA, J., DIMITRIADIS, D. & O’MALLEY, J.T. 2008. The influence of cochlear shape on low-frequency hearing. Proceedings of the National Academy of Sciences USA 105, 6162–6166. MENG, J. & FOX, R.C. 1995. Osseous inner ear structures and hearing in early marsupials and placentals. Zoological Journal of the Linnean Society 115, 47–71. NEENAN, J.M. & SCHEYER, T.M. 2012. The braincase and inner ear of Placodus gigas (Sauropterygia, Placodontia) – a new reconstruction based on micro-computed tomographic data. Journal of Vertebrate Pale- ontology 32, 1350–1357. OLORI, J.C. 2010. Digital endocasts of the cranial cavity and osseous labyrinth of the burrowing snake Uropeltis woodmasoni (Alethinophi- dia: Uropeltidae). Copeia 2010, 14–26. OLSON, E.C. 1944. Origin of mammals based upon cranial morphology of the therapsid suborders. Geological Society of America, Special Papers 55, 1–136. OLSON, E.C. 1962. Late Permian terrestrial vertebrates, USA and USSR. Transactions of the American Philosophical Society, new series 52, 1–224. PERIER, A., LEBRUN, R. & MARIVAUX, L. 2016. Different level of intraspecific variation of the bony labyrinth morphology in slow- ver- sus fast-moving primates. Journal of Mammalian Evolution 23, 353–368 RAMÍREZ-CHAVES, H-E., WEISBECKER, V., WROE, S. & PHILLIPS, M.J. 2016 Resolving the evolution of the mammalian middle ear using Bayesian inference. Frontiers in Zoology 13:39. RAMPRASHAD, F., LANDOLT, J.P., MONEY, K.E. & LAUFER, J. 1986. Comparative morphometric study of the vestibular system of the vertebrata: Reptilia, Aves, Amphibia, and Pisces. Acta Oto-laryngolo- gica, Supplement 427, 1–42. RODRIGUES, P. G., RUF, I. & SCHULTZ, C.L. 2013. Digital reconstruction of the otic region and inner ear of the non-mammalian cynodont Brasili- therium riograndensis (Late Triassic, Brazil) and its relevance to the evo- lution of the mammalian ear. Journal of Mammalian Evolution 20, 291–307. ROHLF, F.J. 1990. Rotational fit (Procrustes) Method. In: Rohlf, F.J. & Bookstein, F.L. (eds), Proceedings of the Michigan Morphometrics Work- shop, 227–236. Ann Arbor, University of Michigan Museum of Zoology Press. RUBIDGE, B.S. & SIDOR, C.A. 2001. Evolutionary patterns among Permo-Triassic therapsids. Annual Review of Ecology and Systematics 32, 449–480. RUBIDGE, B.S. & SIDOR, C.A. 2002. On the cranial morphology of the basal therapsids Burnetia and Proburnetia (Therapsida: Burnetiidae). Journal of Vertebrate Paleontology 22, 257–267. RUBIDGE, B.S., SIDOR, C.A. & MODESTO, S.P. 2006. A new burnetia- morph (Therapsida: Biarmosuchia) from the Middle Permian of South Africa. Journal of Paleontology 80, 740–749. RUF, I., LUO, Z-X. & MARTIN, T. 2013. Reinvestigation of the basicra- nium of Haldanodon exspectatus (Mammaliaformes, Docodonta). Jour- nal of Vertebrate Paleontology 33, 382–400. RUF, I., LUO, Z-X., WIBLE, J.R. & MARTIN, T. 2009. Petrosal anatomy and inner ear structures of the Late Jurassic Henkelotherium (Mammalia, Cladotheria, Dryolestoidea): insight into the early evolution of the ear region in cladotherian mammals. Journal of Anatomy 214, 679–693. SANCHEZ-VILLAGRA, M.R. 2002. The cerebellar paraflocculus and the subarcuate fossa in Monodelphis domestica and other marsupial mam- mals: the ontogeny and phylogeny of a brain–skull interaction. Acta Theriologica 47, 1–14. SCHMELZLE, T., SANCHEZ-VILLAGRA, M.R. & MAIER, W. 2007. Vestibular labyrinth diversity in diprotodontian marsupial mammals. Mammal Study 32, 83–98. SERENO, P.C., WILSON, J.A., WITMER, L.M., WHITLOCK, J.A., MAGA, A., IDE, O. & ROWE, T.A. 2007. Structural extremes in a Cretaceous dinosaur. PLOS ONE 2, e1230. SIDOR, C.A. 2003. The naris and palate of Lycaenodon longiceps (Therapsida: Biarmosuchia), by comments on their early evolution in the Therapsida. Journal of Paleontology 77, 977–984. SIDOR, C.A. & RUBIDGE, B.S. 2006. Herpetoskylax hopsoni, a new biarmo- suchian (Therapsida: Biarmosuchia) from the Beaufort Group of South Africa. In: Carrano, M.T., Gaudin, T., Blob, R. & Wible, J. (eds), Amniote Paleobiology: Perspectives on the Evolution of Mammals, Birds, and Reptiles, 76–113. Chicago, University of Chicago Press. SIDOR, C.A. & SMITH, R.M. 2007. A second burnetiamorph therapsid from the Permian Teekloof Formation of South Africa and its associ- ated fauna. Journal of Vertebrate Paleontology 27, 420–430. SIDOR, C.A. & WELMAN, J. 2003. A second specimen of Lemurosaurus pricei (Therapsida: Burnetiamorpha). Journal of Vertebrate Paleontology 23, 631–642. SIDOR, C.A., HOPSON, J.A. & KEYSER, A.W. 2004. A new burnetia- morph therapsid from the Teekloof Formation, Permian, of South Africa. Journal of Vertebrate Paleontology 24, 938–950. SIGOGNEAU, D. 1970. Revision systematique des gorgonopsiens Sud-Afri- cains. Paris, Cahiers de Paleontologie. SIGOGNEAU, D. 1974. The inner ear of Gorgonops (Reptilia, Therapsida, Gorgonopsia). Annals of the South African Museum 64, 53–69. SIGOGNEAU-RUSSELL, D.S. 1989. Theriodontia I: Phthinosuchia, Biarmosuchia, Eotitanosuchia, Gorgonopsia. In: Wellnhofer, P. (ed.), Encyclopedia of Paleoherpetology Part 17C, 1–127. Stuttgart, Gustav Fischer. SILCOX, M.T., BLOCH, J.I., BOYER, D.M., GODINOT, M., RYAN, T.M., SPOOR, F. & WALKER, A. 2009. Semicircular canal system in early primates. Journal of Human Evolution 56, 315–327. 76 ISSN 2410-4418 Palaeont. afr. (2017) 52: 58–77 SMITH, R.M., RUBIDGE, B.S. & SIDOR, C.A. 2006. A new burnetiid (Therapsida: Biarmosuchia) from the Upper Permian of South Africa and its biogeographic implications. Journal of Vertebrate Paleontology 26, 331–343. SPECHT, M., LEBRUN, R. & ZOLLIKOFER, C.P.E. 2007. Visualizing shape transformation between chimpanzee and human braincases. The Visual Computer 23, 743–751. SPOOR, F. & ZONNEVELD, F. 1995. Morphometry of the primate bony labyrinth: a new method based on high resolution computed tomog- raphy. Journal of Anatomy 186, 271–286. SPOOR, F., BAJPAL, S., HUSSAIM, S.T., KUMAR, K. & THEWISSEN, J.G.M. 2002. Vestibular evidence for the evolution of aquatic behaviour in early cetaceans. Nature 417, 163–166. SPOOR, F., GARLAND, T., KROVITZ, G., RYAN, M., SILCOX, M.T. & WALKER, A. 2007. The primate semicircular canal system and locomo- tion. Proceedings of the National Academy of Science USA 104, 10808–10812. STOKSTAD, E. 2003. Peering into ancient ears. Science 302(5646), 770–771. SUES, H-D. 1986. Locomotion and body form in early Therapsida (Dinocephalia, Gorgonopsia, and Therocephalia). In: Hotton III, N., MacLean, P.D., Roth, J.J. & Roth, E.C. (eds), The Ecology and Biology of Mammal-like Reptiles, 61–70. Washington, D.C, Smithsonian Institution Press. SURKOV, M.V., BENTON, M.J., TWITCHETT, R.J., TVERDOKHLEBOV, V.P. & NEWELL, A.J. 2007. First occurence of footprints of large therapsids from the Upper Permian of European Russia. Palaeontology 50, 641–652. VIDAL, P.P., GRAF, W. & BERTHOZ, A. 1986. The orientation of the cervi- cal vertebral column in unrestrained awake animals. I. Resting posi- tion. Experimental Brain Research 61, 549–559. WALSH, S.A., LUO, Z-X. & BARRETT, P.M. 2013. Modern imaging tech- niques as a window to prehistoric auditory worlds. In: Köppl, C. et al. (eds), Insights from Comparative Hearing Research, Springer Handbook of Auditory Research 49, 227–261. New York, Springer Science, Business Media. WALSH, S.A., BARRETT, P.M., MILNER, A.C., MANLEY, G. & WITMER, L.M. 2009. Inner ear anatomy is a proxy for deducing auditory capabil- ity and behaviour in reptiles and birds. Proceedings of the Royal Society B 276, 1355–1360. WELKER, K.L., ORKIN, J.D. & RYAN, T.M. 2009. Analysis of intra- individual and intraspecific variation in semicircular canal dimen- sions using high-resolution x-ray computed tomography. Journal of Anatomy 215, 444–451. WEST, C.D. 1985. The relationship of the spiral turns of the cochlea and the length of the basilar membrane to the range of audible frequencies in ground dwelling mammals. Journal of the Acoustical Society of America 77, 1091–1101. ISSN 2410-4418 Palaeont. afr. (2017) 52: 58–77 77 << /ASCII85EncodePages false /AllowTransparency false /AutoPositionEPSFiles true /AutoRotatePages /All /Binding /Left /CalGrayProfile (Dot Gain 20%) /CalRGBProfile (sRGB IEC61966-2.1) /CalCMYKProfile (U.S. Web Coated \050SWOP\051 v2) /sRGBProfile (sRGB IEC61966-2.1) /CannotEmbedFontPolicy /Warning /CompatibilityLevel 1.3 /CompressObjects /Tags /CompressPages true /ConvertImagesToIndexed true /PassThroughJPEGImages true /CreateJDFFile false /CreateJobTicket false /DefaultRenderingIntent /Default /DetectBlends true /ColorConversionStrategy /LeaveColorUnchanged /DoThumbnails false /EmbedAllFonts true /EmbedJobOptions true /DSCReportingLevel 0 /EmitDSCWarnings false /EndPage -1 /ImageMemory 1048576 /LockDistillerParams false /MaxSubsetPct 100 /Optimize true /OPM 1 /ParseDSCComments true /ParseDSCCommentsForDocInfo true /PreserveCopyPage true /PreserveEPSInfo true /PreserveHalftoneInfo false /PreserveOPIComments false /PreserveOverprintSettings true /StartPage 1 /SubsetFonts false /TransferFunctionInfo /Apply /UCRandBGInfo /Preserve /UsePrologue false /ColorSettingsFile () /AlwaysEmbed [ true /AmazoneBT-Regular ] /NeverEmbed [ true ] /AntiAliasColorImages false /DownsampleColorImages true /ColorImageDownsampleType /Bicubic /ColorImageResolution 350 /ColorImageDepth -1 /ColorImageDownsampleThreshold 1.14286 /EncodeColorImages true /ColorImageFilter /DCTEncode /AutoFilterColorImages false /ColorImageAutoFilterStrategy /JPEG /ColorACSImageDict << /QFactor 0.40 /HSamples [1 1 1 1] /VSamples [1 1 1 1] >> /ColorImageDict << /QFactor 0.15 /HSamples [1 1 1 1] /VSamples [1 1 1 1] >> /JPEG2000ColorACSImageDict << /TileWidth 256 /TileHeight 256 /Quality 30 >> /JPEG2000ColorImageDict << /TileWidth 256 /TileHeight 256 /Quality 30 >> /AntiAliasGrayImages false /DownsampleGrayImages true /GrayImageDownsampleType /Bicubic /GrayImageResolution 350 /GrayImageDepth -1 /GrayImageDownsampleThreshold 1.14286 /EncodeGrayImages true /GrayImageFilter /DCTEncode /AutoFilterGrayImages false /GrayImageAutoFilterStrategy /JPEG /GrayACSImageDict << /QFactor 0.15 /HSamples [1 1 1 1] /VSamples [1 1 1 1] >> /GrayImageDict << /QFactor 0.15 /HSamples [1 1 1 1] /VSamples [1 1 1 1] >> /JPEG2000GrayACSImageDict << /TileWidth 256 /TileHeight 256 /Quality 30 >> /JPEG2000GrayImageDict << /TileWidth 256 /TileHeight 256 /Quality 30 >> /AntiAliasMonoImages false /DownsampleMonoImages true /MonoImageDownsampleType /Bicubic /MonoImageResolution 1200 /MonoImageDepth -1 /MonoImageDownsampleThreshold 2.00583 /EncodeMonoImages true /MonoImageFilter /CCITTFaxEncode /MonoImageDict << /K 0 >> /AllowPSXObjects false /PDFX1aCheck false /PDFX3Check false /PDFXCompliantPDFOnly false /PDFXNoTrimBoxError true /PDFXTrimBoxToMediaBoxOffset [ 0.00000 0.00000 0.00000 0.00000 ] /PDFXSetBleedBoxToMediaBox true /PDFXBleedBoxToTrimBoxOffset [ 0.00000 0.00000 0.00000 0.00000 ] /PDFXOutputIntentProfile (None) /PDFXOutputCondition () /PDFXRegistryName (http://www.color.org) /PDFXTrapped /Unknown /SyntheticBoldness 1.000000 /Description << /FRA /JPN /DEU /PTB /DAN /NLD /ESP /SUO /ITA /NOR /SVE /ENU >> >> setdistillerparams << /HWResolution [1200 1200] /PageSize [5952.756 8418.897] >> setpagedevice