A Taxonomic Review of Ichthyosaurus from Somerset a Thesis Submitted To

Home , Blue lias, Coracoid, Ichthyosaur, Ichthyosaurus

A taxonomic review of Ichthyosaurus from Somerset

A thesis submitted to The University of Manchester for the degree of Master of Philosophy (MPhil) in the Faculty of Engineering and Physical Sciences

2015

Dean R. Lomax

School of Earth, Atmospheric and Environmental Sciences

TABLE OF CONTENTS

TABLE OF CONTENTS ...... 2

LIST OF PRIMARY FIGURES ...... 5

APPENDIX A FIGURES ...... 9

APPENDIX B FIGURES ...... 10

ABSTRACT...... 13

DECLARATION ...... 14

DEDICATIONS ...... 15

INSTITUTIONAL ABBREVIATIONS ...... 16

1. ACKNOWLEDGEMENTS ...... 17

2. INTRODUCTION ...... 18

3. MATERIALS AND METHODS ...... 20

4. GEOLOGICAL SETTING ...... 21

5. A HISTORY OF ICHTHYOSAURUS STUDY WITH EMPHASIS ON SOMERSET

...... 24

6. THE TAXONOMIC IMPLICATIONS OF ‘ICHTHYOSAURUS’ SPECIMENS FROM

SOMERSET AND THEIR PHYLOGENETIC POSITION ...... 31

7. CLADISTIC ANALYSIS ...... 33

7.1. RESULTS ...... 35

7.2.1 SPECIES #1 ...... 38

7.2.1A. Species characteristics: ...... 38

7.2.1B. Species interpretation: ...... 38

7.2.1C. Specimen description(s):...... 39

7.2.2 SPECIES #2 ...... 41

7.2.2A. Species characteristics: ...... 41

7.2.2B. Species interpretation: ...... 41

7.2.2C. Specimen description(s):...... 42

7.2.3 SPECIES #3 (‘CLUSTER 1’) ...... 43

7.2.3A. Species characteristics: ...... 43

7.2.3B. Species interpretation: ...... 44

7.2.3C. Specimen description(s):...... 45

7.2.4 SPECIES #4 (‘CLUSTER 2’) ...... 52

7.2.4A. Species definition: ...... 52

7.2.4B. Species interpretation: ...... 53

7.2.4C. Specimen description(s):...... 54

7.2.5 SPECIES #5? ...... 67

7.2.5A. Species questionable: ...... 67

7.2.5B. Specimen description(s): ...... 68

7.2.6 ADDITIONAL SPECIMENS ...... 69

7.2.6A. Specimen description(s): ...... 69

8. DISCUSSION ...... 75

8.1. Interpretation of results: ...... 75

8.2. Comparison of the neotype with Ichthyosaurus specimens in Somerset:

...... 76

8.3. A cladistic analysis without the neotype: ...... 77

8.4. Ichthyosaurus implications: ...... 79

9. FUTURE RESEARCH ...... 81

10. SUMMARY AND CONCLUSION ...... 84

11. REFERENCES...... 85

12. APPENDICES ...... 94

Appendix A: List of characters for cladistic analysis ...... 94

Appendix B: Interpretation of character maps with comments regarding the

usefulness of characters used ...... 109

Appendix C: Individual character coding for the cladistic analysis ...... 153

Word count, excluding references, appendices and figure explanations: 19,786

List of primary figures Figure 1. Artist illustration of a family of Ichthyosaurus swimming in the Early Jurassic seas. Illustration courtesy of James McKay. Figure 2. Map of Somerset with some of the important ichthyosaur-bearing localities mentioned, plus other major locations. Varied topography: slightly shaded areas show upland (Exmoor, Mendips, Quantocks) and horizontal lines indicate lowland (Somerset Levels). Modified from Haslett (2010; Figure. 1.3). Figure 3. The lithostratigraphy and biostratigraphy of the Blue Lias Formation. Reproduced from Ambrose (2001). Figure 4. Illustration (lithograph) of the original I. communis holotype tooth (Conybeare, 1822; Plate 15, Figure. 8) and paratype forefin (Conybeare, 1822; Plate 20, Figure. 1) modified from Conybeare (1822). Figure 5. An early illustration of an ichthyosaur found in 1818 at Watchet, initially identified as a fish. Unfortunately the specimen’s current whereabouts are unknown, although it was apparently purchased by the Royal College of Surgeons. It appears that an identification of Ichthyosaurus is more than appropriate, given the morphology of the forefin. Modified from Delair (1969; Figure. 3). Figure 6. The variation of the matrix ‘enhancements’/differences of several of the samples specimens from this study; some are potentially suggestive of a composite specimen. A. AGC 11: Two different matrix types which suggests the specimen was placed into a different block, plus several distinct cracks run throughout. B. AGC 12: Painted matrix, covering either original matrix or plaster. C. ANSP 17566: Patched matrix (chisel-like markings) and distinct lines (arrow points to a distinct line in the matrix). D. CAMSM J59575: At least four distinct types of ‘matrix’ are present, including the clear use of plaster and a specific type of patchwork. Figure 7. A selection of forefins belonging to specimens of Ichthyosaurus, exemplifying the high degree of variation found within the genus. Ulnare coloured in blue. Modified from Motani (1999a; Figure. 6, the forefins of Ichthyosaurus). Figure 8. The selected tree (cladogram) from the specimen-level cladistic analysis, representing five potentially distinct species of Ichthyosaurus found in the results. Node 16: Species #1; Node 11: Species #2; Node 29: Species #3 (‘Cluster 1’); Node 21: Species #4 (‘Cluster 2’) and Node 5: Species #5?. Figure 9. A representative photograph of the ‘ichthyosaur wall’ in the Fossil Marine Reptiles gallery at the NHMUK. Some of the most complete, and historically

5 important specimens are on display, many are from Somerset. Note the lighting problems for viewing specimens that are positioned high on the wall. Figure 10. The practically complete skeleton of NMW 93.5G.2, the sole representative of Species #1. Scale bar = 10 cm. Figure 11. Close-up of the pelvic and hindfin area of NMW 93.5G.2. IS = Ischia. IL = Ilium. LF = Left femur. RF = Right femur. Scale bar = 2 cm. Figure 12. The nearly complete skeleton of CAMSM J35183, the sole representative of Species #2. Scale bar = 10 cm. Figure 13. The partial skeleton of CAMSM J59575, a specimen encompassed within Species #3 (Cluster 1). Scale bar = 10 cm. Figure 14. The partial skeleton of BRSMG Cb 3578, a specimen encompassed within Species #3 (Cluster 1). Scale bar = 10 cm. Figure 15. The practically complete skeleton of NHMUK 2013, a specimen encompassed within Species #3 (Cluster 1). Note that this specimen is on display at the NHMUK, behind glass. Scale bar = 20 cm. Figure 16. The relatively complete skeleton of NHMUK R3372, a specimen encompassed within Species #3 (Cluster 1). The arrow points to the position of the embryo. Note that this specimen is on display at the NHMUK, behind glass. Scale bar = 10 cm. Figure 17. Close-up of the embryo positioned in the pelvic area of NHMUK R3372; the arrow points to the skull of the embryo. Figure 18. The partial skeleton of AGC 12, a specimen encompassed within Species #3 (Cluster 1). Scale bar = 10 cm. Figure 19. The practically complete skeleton of ANSP 17429, a specimen encompassed within Species #4 (Cluster 2). Scale bar = 10 cm. Figure 20. The nearly complete skeleton of AGC 11, a specimen encompassed within Species #4 (Cluster 2). Scale bar = 10 cm. Figure 21. The neotype specimen of I. communis (NHMUK R1162), a specimen encompassed within Species #4 (Cluster 2). Note that this specimen is on display at the NHMUK, behind glass. Scale bar = 10 cm. Figure 22. The practically complete skeleton of MOS 8373, a specimen encompassed within Species #4 (Cluster 2). Scale bar = 10 cm. Figure 23. The practically complete skeleton of ANSP 15766, a specimen encompassed within Species #4 (Cluster 2). Scale bar = 10 cm.

Figure 24. The practically complete skeleton of BRSMG Cb 4997, a specimen encompassed within Species #4 (Cluster 2). Note that this specimen is on display at the BRSMG behind glass. Scale bar = 10 cm. Figure 25. The nearly complete skeleton of OUMNH J13799, a specimen encompassed within Species #4 (Cluster 2). Note the arrows pointing out the difference in height/matrix surrounding the head, suggestive of a possible composite. Scale bar = 10 cm. Figure 26. The practically complete skeleton of BRSMG Ce 16611, a specimen encompassed within Species #4 (Cluster 2). The arrow points to the position of an embryo. Note that this specimen is on display at the BRSMG, behind glass, although a cast is readily accessible (and on display) at LEICT. Scale bar = 10 cm. Figure 27. Close-up of the embryo positioned n the pelvic area of BRSMG Ce 16611. Scale bar = 2 cm. Figure 28. The semi-complete skeleton of AGC 17, the only specimen identified as Species #5? although probably better positioned in one of the two clusters of Species #3 or Species #4. Scale bar = 10 cm. Figure 29. The partial skeleton of AGC 14, an additional specimen not used in the primary analysis, but may belong to Species #4 (Cluster 2). Scale bar = 10 cm. Figure 30. The skull and anterior portion of AGC 15, an additional specimen not used in the primary analysis, but may belong to Species #4 (Cluster 2). Scale bar = 10 cm. Figure 31. The fragmentary skull and articulated skeleton of an unnumbered BGS example, an additional specimen not used in the primary analysis and one that could not be reliably referred to any of the species. Scale bar = 10 cm. Figure 32. The practically complete, yet poorly preserved, skeleton of MOS 166/1992. An additional specimen not used in the primary analysis and one that could not be reliably referred to any of the species. Scale bar = 10 cm. Figure 33. The practically complete skeleton of NHMUK R5595, an additional specimen not used in the primary analysis and one that could not be reliably referred to any of the species. Note that this specimen is on display at the NHMUK, behind glass. There is some glare from lighting (positioned in the centre of the specimen). Scale bar = 20 cm.

Figure 34. A tree (cladogram) selected from an additional cladistic analysis undertaken with the sampled specimens, with the removal of the neotype of I. communis. Figure 35. Differences in tooth morphology between some of the studied samples: A. CAMSM J35183. B. AGC 12. C. ANSP 15766. D. BRSMG Cb 4997. A and D are inverted so that the orientation is the same for all specimens featured. Figure 36. An example of the wide (fan-shaped) distal portion of the pubis found in some Somerset specimens. Pictured here is the pubis (note the arrow), ischium and hindfin of ANSP 17429, a specimen found in Species #4 (Cluster 2). Scale bar = 5 cm. Figure 37. An example of one of several oddities found in the analysis, whereby an individual specimen possessed a different character state but which was not used taxonomically (by the analysis). Here is the right forefin (in ventral view) of OUMNH J13799; note the high contact point of the posterior accessory and the irregular shape of both the radius and ulna. Scale bar = 10 cm.

Appendix A Figures Figure A1. Illustration of character 31. Distalmost point of humerus between the radius and ulna: A. Pointed. B. Rounded. Note the arrow. The humeri are depicted as the right in dorsal view. Figure A2. Illustration of character 47. Distal and proximal ends of the femur: A. nearly equal in width. B. Distal end wider than proximal end. The femora are depicted as the right in dorsal view. Figure A3. Illustration of character 48. Anterior portion of distal end of femur: A. Curved (almost right angle). B. Slightly ‘flared’ curve. C. Widely ‘flared’ curve. The femora are depicted as the right in ventral view. Figure A4. Illustration of character 49. Posterior portion of distal end of femur: A. Right angle. B. Gentle ‘flared’ curve. C. Widely ‘flared’ curve. The femora are depicted as the right in dorsal view. Figure A5. Illustration of character 66. Scapula overall robustness: A. Very robust. B. Slightly robust (more delicate) and more elongated. The scapulae are depicted as the right in dorsolateral view. Figure A6. Illustration of character 70. Coracoid anterior notch shape: A. J-shaped. B. Closed C-shaped. C. Wide C-shaped. The coracoids are depicted as the right in ventral view. Note: The open notch is the anterior notch. Figure A7. Illustration of character 71. Coracoid posterior shape: A. J-shaped. B. Closed C-shaped. C. Wide C-shaped. D. Absent. The coracoids are depicted as the right in ventral view. Note: The open notch is the posterior notch. In ‘D’ the notch is absent. Figure A8. Illustration of character 73. Pubis shape, distally: A. Not expanded. B. Expanded ‘spatulate-shaped’. C. Widely expanded and asymmetric (‘fan-shaped’). The pubes are depicted as the right in dorsal view. Figure A9. Illustration of character 75. Ischium shape, distally: A. Not expanded, essentially straight. B. Marginally expanded. C. Widely expanded. The ischia are depicted as the right in dorsal view.

Appendix B Figures Figure B1. The character map for Character 0. Note the autapomorphic differences for both NMW 93.5G.2 and CAMSM J35183 and the synapomorphy between clusters 1 & 2. Figure B2. The character map for Character 1. Note the retained primitive character of CAMSM J35183 with Suevoleviathan. Figure B3. The character map for Character 2. Apart from the autapomorphy of NMW 93.5G.2 note the variation throughout the tree. Figure B4. The character map for Character 3. There is a clear difference with this character from the most derived group (Cluster 2) and all other primitive specimens. Figure B5. The character map for Character 5. Note the retained primitive character of CAMSM J35183 with Temnodontosaurus. Figure B6. The character map for Character 15. Note the sole autapomorphy of NMW 93.5G.2. Figure B7. The character map for Character 16. This is one character that is variable throughout the individual specimens, but may have some usefulness. Figure B8. The character map for Character 23. This is one of only three characters based on dentition. It shows the ‘autapomorphy’ for CAMSM J35183, but the variation in other specimens. Figure B9. The character map for Character 24. This is an example of one of the characters displaying fairly diverse variation. Figure B10. The character map for Character 28. There is a clear difference with this character from the most derived group (Cluster 2) and all primitive specimens. Figure B11. The character map for Character 30. This is a good character map that shows there could be some usefulness of the character (note the grouping in Cluster 1), but the variation is seen throughout Cluster 2. Figure B12. The character map for Character 38. This is another example of how the character state varies throughout the tree, rendering its usefulness as a taxonomic character poor. Figure B13. The character map for Character 39. This is another good character that shows a synapomorphy between NMW 93.5G.2 and Suevoleviathan. Figure B14. The character map for Character 41. Another example of how a character state is spread across the tree, although it may have some potential

10 usefulness in Cluster 1. It also separates the genus (Ichthyosaurus) from the outgroup taxa. Figure B15. The character map for Character 42. One of the characters that did not display any real taxonomic usefulness, apart from separating Temnodontosaurus from the rest. Figure B16. The character map for Character 50. An apparent good genus-level character for Ichthyosaurus, yet the neotype of I. communis is coded differently. Figure B17. The character map for Character 56. One specimen (apart from the outgroup taxa) is coded differently to the others, and may suggest an oddity. Figure B18. The character map for Character 59. A variable character between both cluster 1 & 2. Figure B19. The character map for Character 61. A good genus-level character for Ichthyosaurus. Figure B20. The character map for Character 64. An apparent good genus-level character for Ichthyosaurus, yet the neotype of I. communis is coded differently. Figure B21. The character map for Character 65. A good genus-level character for Ichthyosaurus. Figure B22. The character map for Character 66. An example of a variable character found throughout the tree. Figure B23. The character map for Character 67. A potentially good autapomorphic character to separate the most derived group (Cluster 2) from the others. Figure B24. The character map for Character 69. A good character found in all Ichthyosaurus specimens, apart from CAMSM J35183 which has a character state retained with the outgroup taxa. Figure B25. The character map for Character 70. An example of a character that may have some taxonomic usefulness, but some specimens are coded differently. Figure B26. The character map for Character 73. A character that was independently suggested to show some taxonomic usefulness, but does not, as the character state varies throughout the tree. Figure B27. The character map for Character 74. Note the autapomorphy for NMW 93.5G.2, but the rest of the specimens are variable between the other character states. Figure B28. The character map for Character 75. An example of another character that may be of taxonomic usefulness, although there is an oddity in Cluster 2.

Figure B29. The character map for Character 76. A character retained by the outgroup taxa and NMW 93.5G.2 with both cluster 1& 2 representing a synapomorphy. Figure B30. The character map for Character 78. An autapomorphy for NMW 93.5G.2, but rather variable throughout the rest of the tree. Figure B31. The character map for Character 81. A possible good species-level character, yet the neotype of I. communis is coded differently. Figure B32. The character map for Character 82. A clear difference can be seen between both cluster 1 & 2 although there is the one oddity in Cluster 2. Figure B33. The character map for Character 84. A good genus-level character for Ichthyosaurus. Figure B34. The character map for Character 85. Clearly, Cluster 1 has a retained primitive character with the outgroup taxa and NMW 93.5G.2 and Cluster 2 share a synapomorphy. Figure B35. The character map for Character 86. A good genus-level character for Ichthyosaurus.

ABSTRACT The University of Manchester

Dean R. Lomax The Degree of Master of Philosophy (MPhil) A taxonomic review of Ichthyosaurus from Somerset 2015

Ichthyosaurus specimens from Somerset have not been examined in the light of current views of taxonomy and phylogeny. Presently, all Ichthyosaurus specimens from the Upper Triassic–Lower Jurassic of Somerset are referred to I. communis. Additional species or, perhaps, genera could exist among the Somerset specimens. In order to resolve the taxonomic questions, all Ichthyosaurus specimens from Somerset have been examined as part of this study. A total of 15 complete, or nearly complete, specimens were selected for detailed examination and study. Upon examination of the specimens, in order to separate any potentially distinct species present in the Somerset material, a character list of 87 distinct characters was compiled and each of the specimens was coded. A cladistic analysis was run and a total of 15 equally parsimonious trees were retained in the analysis. A total of 11 trees showed similar groupings of specimens. The results suggest that five species (regarded as: Species #1, #2, #3, #4 and #5) of Ichthyosaurus are present in the Somerset specimens. Three of the species are represented by single specimens (Species #1, #2 and #5) and two others (Species #3 and #4) are each represented by a cluster of specimens. Species #1 is the most primitive Ichthyosaurus species; it possesses autapomorphies of the skull and pelvic girdle morphology. Species #2 has some potential autapomorphies in the skull and tooth morphology; the latter may suggest that tooth morphology may potentially be a reliable character. Both Species #1 and #2 also share an independent combination of features with more primitive ichthyosaur taxa that are not found in any of the other species discovered in this study. Species #3 does not possess any autapomorphies, but it shares characters with more primitive taxa that are not found in Species #4. However, it possesses a combination of characters that may prove to be taxonomically useful for distinction of this species, but it is fairly poorly supported. Species #4 is the most derived species and possesses an autapomorphy in the skull and possibly in the humerus; it appears to be synonymous with I. communis, potentially suggesting its presence among the Somerset specimens. However, the study has uncovered wider implications and issues with this species. Species #5 is very poorly supported and is not considered valid. Coupled with the four currently valid species of Ichthyosaurus from Dorset (I. communis, I. breviceps, I. conybeari, and I. anningae described by Lomax and Massare (2015)), and the recognition of at least three further distinct species (with Species #4 possibly synonymous with I. communis) from Somerset, this study suggests that as many as seven species of Ichthyosaurus are known. The study provides a firm foundation for the much needed revision of the genus, Ichthyosaurus.

DECLARATION

No portion of the work referred to in the thesis has been submitted in support of an application for another degree or qualification of this or any other university or other institute of learning.

COPYRIGHT STATEMENT i. The author of this thesis (including any appendices and/or schedules to this thesis) owns certain copyright or related rights in it (the “Copyright”) and s/he has given The University of Manchester certain rights to use such Copyright, including for administrative purposes. ii. Copies of this thesis, either in full or in extracts and whether in hard or electronic copy, may be made only in accordance with the Copyright, Designs and Patents Act 1988 (as amended) and regulations issued under it or, where appropriate, in accordance with licensing agreements which the University has from time to time. This page must form part of any such copies made. iii. The ownership of certain Copyright, patents, designs, trade marks and other intellectual property (the “Intellectual Property”) and any reproductions of copyright works in the thesis, for example graphs and tables (“Reproductions”), which may be described in this thesis, may not be owned by the author and may be owned by third parties. Such Intellectual Property and Reproductions cannot and must not be made available for use without the prior written permission of the owner(s) of the relevant Intellectual Property and/or Reproductions. iv. Further information on the conditions under which disclosure, publication and commercialisation of this thesis, the Copyright and any Intellectual Property and/or Reproductions described in it may take place is available in the University IP Policy (see http://documents.manchester.ac.uk/DocuInfo.aspx?DocID=487), in any relevant Thesis restriction declarations deposited in the University Library, The University Library’s regulations (see http://www.manchester.ac.uk/library/aboutus/regulations) and in The University’s policy on Presentation of Theses.

Dedications

This study is dedicated to my family, including my mum (Anne Lomax), nan (Joyce Lightfoot), brother (Scott Lomax), sister (Julie Boyles), dad (Ken Lomax), niece (Olivia Boyles) and nephew (Fletcher Boyles) for continuing to believe in me and my career in palaeontology, although they are still unsure exactly what it is that I do! I hope to continue making you all very proud. It is your love and support of this most unusual career that drives me to succeed.

Institutional abbreviations (in alphabetical order): AGC/AGT (Alfred Gillett Collection, cared for by the Alfred Gillett Trust (C & J Clark Ltd), Street, Somerset); BGS (British Geological Survey, Keyworth, Nottingham); ANSP (Academy of Natural Sciences, Philadelphia, Pennsylvania, USA); BRLSI (Bath Royal Literary and Scientific Institute, Bath); BRSMG (Bristol City Museum and Art Gallery, Bristol); CAMSM (Sedgwick Museum, Cambridge University, Cambridge); LEICT (Leicester Arts and Museums Service, New Walk Museum and Art Gallery, New Walk, Leicester); MOS (Museum of Somerset, Taunton, Somerset); NHMUK (formerly BMNH, The Natural History Museum, London); NMW (National Museum of Wales, Cardiff) and OUMNH (Oxford University Museum of Natural History, Oxford).

1. ACKNOWLEDGEMENTS

Firstly, I would like to thank Professor Judy Massare, Nigel Larkin (and Gemma Larkin), Dr David Penney, Dr John Nudds, Dr Paul Barrett and Peter Robinson for encouraging (and having the confidence in) me to undertake this MPhil and for holding my current research and work in palaeontology in high regard. This has helped to inspire me further with my studies and recognize that the time was right for me to undertake an advanced degree. I would also like to extend the acknowledgement to Judy (and indeed Bill Wahl) for sending me down the route of ichthyosaurs (for the past five years); it has certainly been a most enjoyable, yet at times frustrating, ride and I have enjoyed every moment, perhaps not notching, though. John Nudds has been a great friend, colleague and advisor; thanks for making this such a pleasure. Also, I thank Dr Michael A. Taylor for great advice, historical information and general help. I would like to thank a list of museums/institutions (and individuals) for allowing me access to examine some of the specimens in this study; they include: AGC/AGT (Richard Clark, Angela Southern, Charlotte Berry, Judeth Saunders and Julie Mather); ANSP (Dr Edward Daeschler and Edward Gilmore for allowing Judy Massare access to study the specimens); BGS (Louise Neep, Paul Shepherd and Dr Mike Howe); BRLSI (Matt Williams); BRSMG (Roger Vaughan); CAMSM (Matt Riley and Sarah Finney); LEICT (Dr Mark Evans); MOS (Dennis Parson); NHMUK (Sandra Chapman, Dr Paul Barrett and Dr Lorna Steel); NMW (Cindy Howells and Tom Sharpe) and OUMNH (Eliza Howlett). Additionally, thanks to Reece Davies for creating the figures illustrated in the character list. Thanks also to my undergraduate student Ella Goodall (The University of Manchester) for helping with some photographs. The Hitchon Fund (The University of Manchester) is thanked for covering some of my costs to attend the SVPCA (2014) conference and give an oral presentation of preliminary results of this study. Finally, thanks to Dr David Martill (University of Portsmouth), Dr Victoria Egerton (University of Manchester) and Dr Merren Jones (University of Manchester) for their part in the examination of this thesis; your kind comments and suggestions were greatly appreciated.

2. INTRODUCTION

Ichthyosaurs are an extensively researched group of extinct, dolphin-like marine reptiles (Figure 1). They were fully aquatic tetrapods that gave birth to live young and swam using their tails, whilst the fore and hindfins were used for manoeuvering and breaking (Massare, 1988; Buchholtz, 2001). Many species have been described, some measuring only a few feet in length whilst others were in excess of 15 metres, with the most recent review indicating at least 80 valid species (McGowan and Motani, 2003). However, new discoveries and descriptions have since increased this total (e.g. Druckenmiller and Maxwell, 2010; Fischer et al., 2011; Fischer et al., 2012; Maxwell et al., 2012; Stinnesbeck et al., 2014). Numerous features of both the skull and postcranium are used to distinguish genera and species, with many Lower Jurassic forms easily distinguished on the basis of paddle and pectoral girdle morphologies (McGowan and Motani, 2003). To a lesser extent, tooth morphology has been used to distinguish genera and species (Maisch and Matzke, 2000), although separating some species, based solely on tooth morphology, has been considered unreliable (McGowan and Motani, 2003). Tooth form varies among species, with some bearing delicate teeth for piercing, and others displaying more robust teeth for cutting and grasping (smash and crunch) prey (Massare, 1987). Preserved stomach contents have been discovered in several ichthyosaur genera and were first reported over 150 years ago (Buckland, 1836). Contents often comprise the tiny arm hooklets of squid-like cephalopods (Pollard, 1968; Lomax, 2010), although remains of fish and other ichthyosaurs have been recorded (Pollard, 1968; Massare and Young, 2005), suggesting that ichthyosaurs fed upon a variety of animals. Coprolites that may have been produced by ichthyosaurs have also been found with fragments of bone, teeth and scales preserved within (Pollard, 1986; pers. obs.).

Ichthyosaurs achieved a worldwide distribution. The first ichthyosaurs appeared in the late Early Triassic (Olenekian), with the last specimens having been documented from the early Late Cretaceous (Cenomanian; McGowan and Motani, 2003; Motani, 2009). They were most diverse during the Triassic and Jurassic, although the majority of specimens in the UK are known from the historically important Lower Jurassic sites along the Dorset coast, around Lyme Regis (Milner and Walsh, 2010);

18 historic quarries around Street, Somerset (Delair, 1969); from the Lower Jurassic exposures along the Whitby coast (Benton and Taylor, 1984); from the Middle Jurassic Oxford Clay Formation, especially in Peterborough (Kirton, 1983; Martill, 1991); and from the Upper Jurassic Kimmeridge Clay Formation of Dorset (McGowan, 1976). In comparison, Cretaceous ichthyosaurs are much less abundant, perhaps due to preservation and collecting bias. However, recent work on Cretaceous ichthyosaurs has yielded new discoveries, suggesting further diversification (e.g. Motani, 2005a; Fischer et al., 2011; Fischer et al., 2012; Zammit, 2012; Fischer et al., 2013; Fischer et al., 2014a; Fischer et al., 2014b; Fischer et al., 2014c; Stinnesbeck et al., 2014; Green and Lomax, 2014).

Figure 1. Artist illustration of a family of Ichthyosaurus swimming in the Early Jurassic seas. Illustration courtesy of James McKay.

Ichthyosaur specimens had been documented as early as the late 1600s and 1700s, although their true identity was not recognized and most were regarded as ‘crocodiles’ or ‘alligators’ at the time (Delair, 1969; McGowan and Motani, 2003). The first genus of ichthyosaur to be recognized scientifically was the common Upper

Triassic/Lower Jurassic form Ichthyosaurus, first scientifically reported in 1821 from remains collected at Lyme Regis, Dorset (De la Beche and Conybeare, 1821). Specimens of ichthyosaurs collected during the 19th century were largely referred to the genus Ichthyosaurus, with over 50 new species described before 1900 (see list of taxonomic names in McGowan and Motani, 2003). Subsequent studies, and further discoveries of specimens and exploration of new localities, resulted in the separation and definition of new ichthyosaurian genera and species. Today, all specimens of Lower Jurassic ichthyosaurs with a wide forefin of five or more digits are referred to Ichthyosaurus (McGowan and Motani, 2003; Lomax and Massare, 2015). Despite being the first ichthyosaur described, the definition of the genus is rather poor. Many of the morphometric characteristics overlap between species (pers. obs.) and the diagnostic features and individual species within Ichthyosaurus require thorough examination and revision (see Lomax and Massare, 2015).

The primary intention of this study is to determine how many species of Ichthyosaurus from the Late Triassic/Early Jurassic rocks of Somerset are valid. Many specimens have been recorded from inland quarries around the town of Street, and only a few specimens have been collected from the coastline. The intention of this study is not to record each and every ichthyosaur element collected in Somerset, but instead to focus on the more complete to near-complete specimens, along with anything else that can be determined diagnostic or useful. A cladistic analysis incorporating the more complete specimens shows the presence (and possible species diversity) of Ichthyosaurus within the Somerset specimens. For specimens that are partial skeletons or comprise fragmentary skeletons and are not used in the cladistic analysis, an attempt is made to place them within species/groups that may arise from the study. A preliminary examination suggests that at least two or three distinct ‘species’ or ‘groups’ will be found.

3. MATERIALS AND METHODS

For each specimen deemed to be sufficiently complete, well preserved and regarded of potential significance, morphologic data and measurements were recorded (where necessary) and photographs taken. This included the comparison and documentation of all of the characters/features of the skull and postcranium which

20 could be of importance for this study. Isolated elements (e.g. isolated forefins, vertebral centra, etc.) were not recorded and are not discussed. A variety of institutional collections were examined and the data were collected. The majority of ichthyosaur specimens from Somerset are stored in the collections of the NHMUK, OUMNH, AGC/AGT and CAMSM. Several other institutions have a few specimens from Somerset, and some of their more complete remains are also used in this study.

For features of specimens with dimensions less than 15 cm, measurements were taken using digital callipers and recorded to the nearest 0.01 cm. For larger features (or total length of specimen), a tape measure was used, and measurements were recorded to the nearest 0.1 cm. Photographs were taken of each of the specimens examined (camera model: Canon PowerShot SX40 HS). Additional notes regarding preservation, conservation, etc., were also recorded where necessary. The package used for the cladistic analysis was the Willi Hennig Society edition of the program TNT (Goloboff et al., 2008). A total of 87 characters were compiled for the character list; 44 of which were based on observations of differences in morphology among the studied specimens (Appendix A).

4. GEOLOGICAL SETTING

Somerset is a county in southwestern England (Figure 2), bounded by the Bristol Channel in the northwest, by Bristol and Gloucestershire in the north, by Wiltshire in the east, by Dorset in the southeast and by Devon in the southwest and west. The geology of the county has been well studied, and represents some of the most diverse in the British Isles. Ranging from Palaeozoic (Silurian) to the more recent Quaternary sedimentary deposits, a host of fossils have been discovered and described (Hardy, 1999; Haslett, 2010). However, it is the sediments of the uppermost Triassic and lowermost Jurassic (part of the Lower Lias Group) that comprise the focus of this study. The Lower Lias Group has yielded many fossils, including several exceptionally well-preserved marine reptiles (Hawkins, 1834; Knight and Dutton, 1909; Hardy, 1999) discussed later. The Lower Lias Group includes the Blue Lias Formation, which possibly comprises the uppermost Triassic

(Rhaetian), the entire Hettangian, and the lowermost Sinemurian (Ambrose, 2001). The formation is further divided into three members: the Wilmcote Limestone Member (upper Rhaetian-lower Hettangian), the Saltford Shale Member (lower-mid Hettangian) and the Rugby Limestone Member (upper Hettangian-lower Sinemurian), each of which are sub-divided further by distinct ammonite zones and subzones (Figure 3; Ambrose, 2001). The Wilmcote Limestone Member comprises interbedded limestones and mudstones, the former being predominantly shelly and argillaceous (Ambrose, 2001). The Saltford Shale Member comprises mostly grey mudstones with some fine-grained argillaceous limestone layers and nodule beds, and is generally poorly fossiliferous (Ambrose, 2001). The Rugby Limestone Member is composed of alternating limestones and mudstones, with the former characterized as grey, fine-grained and argillaceous; nodular limestones are common (Ambrose, 2001). It is the Blue Lias Formation that has yielded the ichthyosaur specimens that form the basis of this study. At inland localities, the Blue Lias Formation was extensively quarried for building stone in the 19th century, largely around the town of Street, with quarrying not restrained to just a few pits (Storrs and Taylor, 1996; Benson et al., 2012). Coastal exposures of the Blue Lias are found at, for example, Watchet (McGowan, 2003; Benson et al., 2012) and Kilve (Deeming et al., 1993). The often complete and well-articulated marine reptile remains found in this formation (collected from Somerset) appear to indicate a low-energy, shallow marine environment with little current action (Storrs and Taylor, 1996; Benson et al., 2011). It has also been suggested that during this time interval, life may have been undergoing, or perhaps recovering from, a major mass extinction (Benson et al., 2011). Finally, Upper Lias rocks (Toarcian), which have also yielded well-preserved remains of ichthyosaurs, have been found in Somerset around the village of Ilminster (Caine and Benton, 2011).

Unfortunately, many of the historic ichthyosaur specimens have poorly recorded geological data, but almost certainly derive from the Blue Lias Formation. Thus their age spans the Triassic-Jurassic boundary, with most specimens probably having derived frp, the Rhaetian and Hettangian. However, Benson et al. (2012) determined that most of the Street plesiosaur specimens likely originated from the Pre-planorbis beds, which occur below the first documented occurrence of the zonal ammonite Psiloceras planorbis. They further suggested that most probably they derived from

22 the earliest Hettangian P. tilmanni Chronozone, immediately following the Triassic– Jurassic boundary. However, some specimens may be from slightly younger horizons (Storrs and Taylor, 1996). It was previously suggested that the Street material from the Pre-planorbis beds was largely Rhaetian in age (Arkell, 1933; McGowan and Motani, 2003). Recent discoveries of ichthyosaur specimens at coastal localities have better geological and stratigraphic information and have been collected in situ from both the Hettangian and Sinemurian strata of the Blue Lias (Deeming et al., 1993; McGowan, 2003 and some specimens examined in this study). Coupled with the age of the plesiosaur remains (Benson et al., 2012), this suggests that the majority of ichthyosaur specimens in Somerset probably derive from the Lower Jurassic (Hettangian), rather than the Upper Triassic (Rhaetian).

Figure 2. Map of Somerset with some of the important ichthyosaur-bearing localities mentioned, plus other major locations. Varied topography: slightly shaded areas show upland (Exmoor, Mendips, Quantocks) and horizontal lines indicate lowland (Somerset Levels). Modified from Haslett (2010; Figure. 1.3).

Figure 3. The lithostratigraphy and biostratigraphy of the Blue Lias Formation in the UK. Reproduced from Ambrose (2001).

5. A HISTORY OF ICHTHYOSAURUS STUDY WITH EMPHASIS ON SOMERSET

The coastal town of Lyme Regis in Dorset is generally considered the historical birthplace of ichthyosaur study (McGowan and Motani, 2003) due to its association with the local fossil collector, and one of the greatest fossil collectors of all time, Mary Anning (1799-1847). Mary tirelessly searched for and discovered many fossils, including the first ichthyosaurs, plesiosaurs and the first British pterosaur (Torrens, 1995). In fact, Mary, along with her brother, discovered the first ichthyosaur (1810/11) that was brought to the attention of the scientific community (Home, 1814). Home (1814) initially thought the specimen was a crocodile or some sort of fish, but later (Home, 1819) named the remains (a large skull and some postcranial elements) as Proteosaurus, now referred to Temnodontosaurus platyodon (NHMUK R1158). However, the name Ichthyosaurus had been proposed a year earlier by König (1818), although not formally described by him. De la Beche and Conybeare (1821,

24 pp. 563-564), whilst describing a different specimen, showed that König’s Ichthyosaurus had precedence over Proteosaurus, and their formal description was based upon the former name (McGowan and Motani, 2003). Ichthyosaurus has since prevailed in the literature. The type species (I. communis) was mentioned by De la Beche and Conybeare (1821, p. 594), with a brief description of a partial skeleton from Lyme Regis. Some comments regarding the tooth morphology of I. communis were discussed in Conybeare (1822), and the original descriptions and distinctions of the first Ichthyosaurus species (I. communis, ‘I. intermedius’, I. platyodon [now Temnodontosaurus platyodon] and I. tenuirostris [now Leptonectes tenuirostris]) were based largely on tooth morphology with some comments on the postcrania (Conybeare, 1822; Owen, 1840; Lydekker, 1889). Unfortunately, the lack of both a detailed description and a figure of the entire holotype of I. communis (as inferred, only a single tooth was figured for I. communis specifically, with perhaps a forefin also [Figure 4]; both considered syntypes [Conybeare, 1822. Pl. XV. Fig. 1 and Pl. XX. Fig. 1]), and the lack of information regarding the repository of the type specimen caused some problems. Additional specimens of Ichthyosaurus had already been described and discussed by De la Beche and Conybeare (1821. p. 565) and Conybeare (1822. p. 112). As mentioned above, the holotype has since been deemed lost and a neotype (NHMUK R1162) designated by McGowan (1974). He proposed that the neotype was a good replacement for the holotype as it was consistent with both the location of discovery (Lyme Regis), and also De la Beche and Conybeares’ (1821) description, which rendered the holotype too large to be referred to either I. breviceps or I. conybeari (the two contemporary species). Moreover, the jaw size of the original description falls within the size range of the taxon (including the proposed neotype), and the figured tooth of I. communis in Conybeare (1822) is also consistent with the neotype. Thus, I. communis is considered to be everything else that is not I. breviceps or I. conybeari. As the neotype is behind glass (on public display) it was not possible to examine it in detail, but this must be done for a thorough review of the genus.

Figure 4. Illustration (lithograph) of the original I. communis holotype tooth (Conybeare, 1822; Plate 15, Figure. 8) and paratype forefin (Conybeare, 1822; Plate 20, Figure. 1) modified from Conybeare (1822).

Despite the fame of Lyme Regis, marine reptile remains in Somerset have been documented since the 1700s, and were first scientifically reported in the early 19th century; although quarrying prior to the 1700s (even within Roman times) probably yielded finds (McGarvie, 1987). Some of the earliest discoveries were in Bath, many from various working quarries around the city. Some of these remains (including ichthyosaurs, then considered crocodiles) were reported in 1779 (Delair, 1969). The majority of marine reptile specimens from Somerset were collected from around the historic town of Street from quarries that are no longer accessible (Hawkins, 1834; Delair, 1969; Storrs and Taylor, 1996; McGowan and Motani, 2003; Benson et al., 2012). One of the earliest records of a nearly complete, articulated ichthyosaur skeleton, although again considered a crocodile at the time, was discovered in the Blue Lias of a ‘Somerset quarry’ in 1803 (reported in 1807), but unfortunately, the specimen has since been lost (Delair, 1969). Indeed, several historical specimens have been ‘lost’ (Figure 5), whilst others were probably never recorded or documented. Numerous other discoveries during the early 1800s, including skulls

26 and partial skeletons of ichthyosaurs, were collected across several quarries in Somerset, including at Twerton, Keynsham, Kingsdon and Weston, and also from the coastal site of Watchet (Delair, 1969; Howe et al., 1981).

Figure 5. An early illustration of an ichthyosaur found in 1818 at Watchet, initially identified as a fish. Unfortunately the specimen’s current whereabouts are unknown, although it was apparently purchased by the Royal College of Surgeons. It appears that an identification of Ichthyosaurus is more than appropriate, given the morphology of the forefin. Modified from Delair (1969; Figure. 3).

The celebrated anatomist and surgeon, John Hunter (1728-1793), had several ichthyosaur specimens in his personal collection, collected from both Dorset and Somerset, and of course were collected prior to 1793 (Delair, 1969; McGowan and Motani, 2003). These remains, along with hundreds of other fossils, formed the foundations of the palaeontology collections of the Museum of the Royal College of Surgeons of England. Unfortunately, of the 400 or so reptilian fossils, almost all were destroyed by enemy bombs in 1941 during World War II (Delair, 1969).

Another notable individual was the great fossil collector and dealer, Thomas Hawkins (1810-1889). ‘The Hawkins Collection’ comprised many ichthyosaurs and plesiosaurs, most of which were collected from working quarries around Street. This

27 collection would later become one of the most valuable private British collections of Liassic reptilian specimens ever made (Delair, 1969; Howe et al., 1981; Taylor, 1989). His collection of marine reptiles is held in notable museums in the UK, including the NHMUK, CAMSM and OUMNH, among others. Hawkins (1834) loosely described several new ichthyosaur taxa from specimens that he had obtained across Somerset and Dorset. The species were poorly defined and all of them were found to be synonymous with previously described species (e.g., I. communis). These specimens represent some of the most complete ichthyosaurs collected from the Lias of Somerset and Dorset (Delair, 1969; McGowan and Motani, 2003). Interestingly, several specimens from Hawkins’ collection were found to represent composites, pieced together from separate individuals, so one must be cautious with specimens from this collection (McGowan, 1989; Taylor, 1989; McGowan, 1990). In fact, caution should be applied to any historic specimen; differences in the matrix, matrix covered by a plaster veneer, painted or patterned matrix (e.g. chisel-like marks), and an unusual positioning/orientation of the bones can further suggest a composite (Figure 6) even though some of these ‘enhancements’ were done solely for aesthetics (McGowan, 1990; Lomax and Massare, 2012; Massare and Lomax, 2014a; Massare and Lomax, 2014b).

Joseph Chaning Pearce (1811-1847), medic and fossil collector, was another individual who amassed an important collection of fossils, including ichthyosaurs from Somerset. One of his most important specimens from Somerset represented the first occurrence of an ichthyosaur embryo found with its mother, suggesting for the first time that ichthyosaurs gave birth to live young (Pearce, 1846).

Another notable individual was Alfred Gillett (1814-1904), who amassed a collection of at least 18 ichthyosaurs, one plesiosaur and numerous other fossils, most of which were personally collected by him (Lydekker, 1891; Delair, 1978; Howe et al., 1981; Michael A. Taylor, pers. comm. 2014). Several of the ichthyosaurs are nearly complete. All of the remains, except for one, are held in the care of the Alfred Gillett Trust and C. & J. Clark Ltd (now Clarks Ltd), but Gillett’s most complete ichthyosaur specimen, a Leptonectes tenuirostris, was donated to the NHMUK in 1883/4, where it was once on display (NHMUK R498) (Lydekker, 1891. pg. 290, Fig. 1). This collection became quite famous due to its association with the shoes and

28 leatherworking company, Clarks. The Clark family had a keen and curious interest in the natural world, especially the geology of Street, and some members of the family had also collected ichthyosaur specimens (McGarvie 1987; Michael A. Taylor, pers. comm. 2014). Gillett subsequently founded a museum in Crispin Hall (Street) in 1887, which displayed local fossils. The museum closed in 1948 and the local fossils were put into the care of the Clarks Archive and more recently the Alfred Gillett Trust (Anon [Wallis, F. S.], 1948; Martill et al., 2009; Michael A. Taylor, pers. comm. 2014). The decline of quarrying had begun in the early 20th century, and this resulted in many of the small quarries around Street being abandoned during World War I. Thus Gillett’s collection was probably one of the last of its type amassed from Street (Arkell, 1933; McGarvie, 1987; Storrs and Taylor, 1996; Michael A. Taylor, pers. comm. 2014). As some of the early, historic ichthyosaurs collected from across Somerset are either lost or destroyed, and as the inland quarries no longer exist, the Gillett Collection represents a scientifically and historically important collection (Delair, 1978; Michael A. Taylor, pers. comm. 2014).

Given that Ichthyosaurus was first described and named in 1821 (De la Beche and Conybeare), at about the same time that many ichthyosaurs were being discovered around Street, naturally the first named ichthyosaurs to be recorded from Somerset were referred to Ichthyosaurus (Delair, 1969). Since then, other species from the Lias of Street have been recorded, including: Ichthyosaurus communis, I. ‘intermedius’, Leptonectes tenuirostris (previously I. tenuirostris), Excalibosaurus costini and possibly Temnodontosaurus sp. (McGowan, 1974; McGowan, 1986; McGowan, 1989; McGowan and Motani, 2003; pers. obs.). Other important ichthyosaur discoveries from Somerset include three ichthyosaurs (two Ichthyosaurus and one Leptonectes tenuirostris) with associated embryos (Pearce, 1846; Deeming et al., 1993; Lomax and Massare, 2012). Interestingly, these are the only ichthyosaur embryos recorded from the UK.

Despite interest in the ichthyosaurs of Somerset during the 19th century, when compared to the extensive research on ichthyosaurs described from the area around Lyme Regis, Dorset, there have been very few recent detailed descriptions. More recently, ichthyosaurs from Somerset have been somewhat overlooked, although more recent discoveries at coastal locations such as Watchet, Kilve and Lilstock

29 have been reported (McGowan, 1986; Deeming et al., 1993; McGowan, 2003). Many of the historic ichthyosaur specimens from Street and surrounding areas have not been studied recently in terms of taxonomy (McGowan, 1974; Appleby, 1979; McGowan, 1989) nor have cladistic analyses been used. The plesiosaurs, on the other hand, have received recent attention (Storrs and Taylor, 1996; Benson et al., 2011; Benson et al., 2012). Applying new techniques and recording a number of morphologically distinct features, Benson et al. (2012) described several new plesiosaur taxa, largely based on the historic collections of Street; it is intended that this work will do likewise for the ichthyosaurs.

Figure 6. The variation of the matrix ‘enhancements’/differences of several of the samples specimens from this study; some are potentially suggestive of a composite specimen. A. AGC 11: Two different matrix types which suggests the specimen was placed into a different block, plus several distinct cracks run throughout. B. AGC 12: Painted matrix, covering either original matrix or plaster. C. ANSP 17566: Patched matrix (chisel-like markings) and distinct lines (arrow points to a distinct line in the matrix). D. CAMSM J59575: At least four distinct types of ‘matrix’ are present, including the clear use of plaster and a specific type of patchwork.

6. THE TAXONOMIC IMPLICATIONS OF ‘ICHTHYOSAURUS’ SPECIMENS FROM SOMERSET AND THEIR PHYLOGENETIC POSITION

The genus Ichthyosaurus currently includes all Lower Jurassic forms with a wide forefin (five or more digits) in which a digital bifurcation occurs anteriorly (Figure 7; Motani, 1999a; McGowan and Motani, 2003). The diagnosis includes: forefin with no fewer than five digits; ulnare larger than intermedium; anterior digital bifurcation; phalanges numerous and close-packed; distal end of humerus wider than proximal end; a tripartite pelvic girdle without fusion between pubis and ischium; and a preflexural vertebral count <80 (McGowan and Motani, 2003). The basioccipital with an extensive extracondylar area and well developed basioccipital peg has also been used but is rarely seen (Motani, 1999a; McGowan and Motani, 2003). The arrangement of dermal skull elements has also been considered diagnostic (Motani, 2005b), but sutures are not always evident or are poorly preserved even on complete skulls.

Figure 7. A selection of forefins belonging to specimens of Ichthyosaurus, exemplifying the variation found within the genus. Ulnare coloured in blue. Modified from Motani (1999a; Figure. 6, the forefins of Ichthyosaurus).

Regarding the above, generic assignment is relatively simple, but individual species determination is more difficult. Currently, species of Ichthyosaurus are distinguished by overall size and skull proportions (snout ratio, orbit ratio), the presacral and preflexural vertebral count, the number of digits in the forefin, the number of phalanges in the longest digit and the presence of notching on the forefin (McGowan, 1974; McGowan and Motani, 2003), all of which are used inconsistently (Lomax and Massare, 2015). Despite the numerous features currently used to distinguish species of Ichthyosaurus, tooth morphology is not one, yet the original definition of both I. communis and I. intermedius (see below) was based on tooth morphology (Conybeare, 1822. p. 108). The last general overview was undertaken by McGowan and Motani (2003). The key features used to distinguish species of Ichthyosaurus is the relative rostrum length (snout ratio = preorbital length of rostrum divided by mandible length [McGowan, 1974; McGowan and Motani, 2003]). With the exception of I. breviceps, which has a distinctly shorter snout than the three other species (McGowan and Motani, 2003; Massare and Lomax, 2014a), the snout ratio is variable. The supposedly longer snout of I. conybeari is within the range of I. communis (McGowan, 1974; pers. obs.), as is I. anningae (Lomax and Massare, 2015). The orbit ratio has also been suggested to have some diagnostic value. However, with the exception of I. breviceps, where the larger orbit ratio is presumably related to the shorter snout, the orbit ratio does not appear to display much variation among species (and individual specimens) of Ichthyosaurus (see discussion in Lomax and Massare, 2015). The orbit, when compared to the relative size of the sclerotic ring, has also been used for estimating the age of some ichthyosaur species (Fernández et al., 2005). The problem with both ratios is that some specimens may not have a complete snout, or the orbit may have been crushed or broken potentially resulting in unreliable measurements (pers. obs.), although a method for estimating missing portions of the snout has been used (McGowan, 1996a). Other species level characters also show overlap among species, but a complete review of the genus is beyond the scope of this study.

All of the Somerset specimens identified as Ichthyosaurus are currently referred to I. communis (McGowan, 1974; McGowan and Motani, 2003). I. breviceps and I. conybeari have not been recorded from the Lias of Somerset and are, apparently, restricted to Lyme Regis, Dorset (McGowan, 1974; McGowan and Motani, 2003).

Historically, another species of Ichthyosaurus, I. intermedius, was suggested to have been present at Somerset, based on tooth morphology (Owen, 1840. pp. 110-112; Maisch and Matzke, 2000). However, the recognition of I. intermedius as a distinct species has been challenged and it was synonymised with I. communis on the grounds that the supposed features of I. intermedius were found in I. communis specimens (McGowan, 1974; McGowan and Motani, 2003). Maisch (1997) argued for the retention of I. intermedius on the basis of tooth morphology (“waisted teeth”, where the crown is narrow and the root inflated), but this was disregarded by McGowan and Motani (2003). Interestingly, Lydekker (1899, p. 42) noted that it was very difficult to distinguish I. communis from I. intermedius in “young skeletons and imperfect skulls”. In fact, based largely on the historic identifications, perhaps half of the Ichthyosaurus specimens from Somerset were initially referred to I. intermedius, with the rest referred to I. communis (pers. obs.). To complicate the taxonomic issues further, the holotype of both I. communis and I. intermedius are lost, and perhaps even destroyed (Maisch and Matzke, 2000; McGowan and Motani, 2003; Michael A. Taylor, pers. comm. 2014). Finally, together with the I. communis specimens from Lyme Regis, Dorset, McGowan (1974. p. 11) suggested that they represented ‘two populations’ of I. communis, which varied greatly in forefin morphology (phalange shape; number of phalanges; number of digits; presence/absence of notching; spatial configuration of elements; bifurcation row, etc. pers. obs.); yet, the variation could not be used taxonomically as it was continuous between specimens.

7. CLADISTIC ANALYSIS

The data used in this analysis were personally collected (unless otherwise stated) from specimens held in the collections listed on page 16. A cladistic analysis based on 15 ichthyosaur specimens from Somerset was undertaken. The examined specimens comprise complete to partial skeletons and are at least 50% complete. Admittedly there were additional specimens that could have been included in this analysis, but because they are fragmentary and represent isolated forefins, hindfins, etc., or the primary features (e.g. of the skull, forefin etc.) were missing, they would not benefit the study and could potentially hinder the results (see comments below). Unfortunately, of the specimens used in the analysis, four are behind glass and

33 could not be examined in detail. However, as the characters in the character list do not require specific measurements, it was decided to include them within this analysis as they represent some of the most complete examples and, despite the lack of measurements offer useful details on the morphology.

A total of 43 characters were selected, compiled from a variety of character matrices, including: Callaway (1989); Motani (1999b); Sander (2000); Maisch and Matzke (2000); Fernández (2007); Druckenmiller and Maxwell (2010); Fischer et al. (2011); Maxwell et al. (2012); and Fischer et al. (2013). In addition, 44 characters, bringing the total to 87, have been added which represent some of the differences in morphology that were noted among examined specimens (Appendices A and B).

As each of the studied specimens varies with regards to completeness, preservation and orientation, some of the data is missing or difficult to discern. This, in turn, may not give a full representation of the taxonomic differences between specimens, in that certain features determined to be of high importance by the analysis may be missing or poorly preserved in another specimen. Nevertheless, I envisage that differences will be noted. As this study is focused upon the species I. communis, the neotype specimen (NHMUK R1162) was also included in the data matrix, to discover which of the studied specimens, if any, plotted closely with that individual. As the specimen is behind glass at the NHMUK, measurements are not possible, but the specimen is displayed at eye-level and the features that are required for the character list do not require specific measurements. Thus, observations were recorded directly from the specimen. Temnodontosaurus and Suevoleviathan were included as the outgroup taxa for the analysis as the two genera are well- represented, are Jurassic in age and appear in many cladistic analyses (e.g. McGowan and Motani, 2003; Maxwell et al., 2012; Fischer et al., 2013). Information for both genera was compiled from descriptions, detailed illustrations and previous phylogenetic studies (McGowan, 1994; McGowan, 1996b; Maisch, 1998; Motani, 1999b; Maisch and Matzke, 2000; Maisch and Hungerbȕhler, 2001; McGowan and Motani, 2003; Maxwell et al., 2012).

Each of the 15 Somerset specimens selected, as well as Temnodontosaurus, Suevoleviathan and the neotype of I. communis, were coded for the analysis using

34 the data matrix of 87 characters (Appendices A and C). A total of 15 equally parsimonious trees were retained in the analysis with a best score of 204 steps. Eleven of the trees showed similar groupings of specimens.

7.1. RESULTS

Having examined the individual trees of the cladistic analysis and compared the specimens at each branch (that grouped together), one of these trees was selected. A strict interpretation of the tree results suggests that there are at least five species of Ichthyosaurus present within the Somerset remains: three are represented by single specimens and two others are each represented by a cluster of specimens (Figure 8). The clusters (potential species) of specimens appear in other trees found by the analysis, which suggests that they may indeed represent distinct species. However, in the same analysis some of the trees displayed continual branching, with little or no grouping. Thus, despite the separation of potentially distinguishable species (detailed below), the analysis has shown that among the Somerset specimens some issues are unresolved and require further work beyond the scope of this study. Nevertheless, the chosen tree showed that as many as five species of Ichthyosaurus are present.

Detailed below are descriptions of the proposed species found by the analysis and a basic overview description of the specimens included in each species. This study focused entirely on the more complete remains, although a few less complete specimens (30-49% complete) are discussed in Section 7.2.6. Several other ichthyosaur specimens are held in the collections of the NMW. These specimens are on long-term loan from the BRLSI and form part of the famous Charles Moore Collection. It is possible, based on the preservation of the matrix and bone (pers. obs.), that some of these specimens were collected from Somerset and probably from Street (see Appleby, 1979). However, the lack of reliable information currently prohibits detailed work on these individuals and they are not discussed further. Additionally, there are several complete and nearly complete ichthyosaur specimens from Somerset on display at the NHMUK, but unfortunately access to these specimens (they are on display in a public gallery, behind glass and positioned high on the wall) is not possible. However, some of the most complete Somerset

35 specimens on display (at eye level) have been used in this analysis, as mentioned previously (Figure 9).

Figure 8. The selected tree (cladogram) from the specimen-level cladistic analysis, representing five potentially distinct species of Ichthyosaurus found in the results. Node 16: Species #1; Node 11: Species #2; Node 29: Species #3 (‘Cluster 1’); Node 21: Species #4 (‘Cluster 2’) and Node 5: Species #5?.

The following descriptions of the potential species found within the analysis are determined to belong to the same genus, defined at Node 25, which unite all specimens including NMW 93.5G.2, CAMSM J35183, AGC 17 and both clusters 1 and 2 (see position of Node 25 on Figure 8). Some of the primary features found at Node 25 include: nasal/external naris contact present, digital bifurcation present, scapula with large acromion process absent, but anterior expansion of proximal scapula present, and preflexural vertebral count less than 80. Features that were not used for the analysis, but are shared between each specimen, include phalanges numerous and close-packed, forefin with no fewer than five digits, and a tripartite pelvic girdle without fusion between pubis and ischium. All of these features were found to be present (co-occurring) in the studied specimens (where available) and appear to represent good genus-level (Ichthyosaurus) characters. Other features in common within the specimens include the lack of an overbite, the lunate, or J-

36 shaped jugal and ilium morphology. Based solely on a strict interpretation of the results, a total of five possible species of Ichthyosaurus have been recognized in the Somerset remains. (For specific character-level interpretations of each of the 87 characters used, see Appendix B.)

Figure 9. A representative photograph of the ‘ichthyosaur wall’ in the Fossil Marine Reptiles gallery at the NHMUK. Some of the most complete, and historically important specimens are on display, many are from Somerset. Note the lighting problems for viewing specimens that are positioned high on the wall.

7.2.1 SPECIES #1

7.2.1A. Species characteristics: This species is represented by a single specimen at Node 16, NMW 93.5G.2. The specimen displays a variety of unusual features not seen in any of the other specimens examined. It is the most plesiomorphic of the proposed species, positioned after the outgroup taxa Temnodontosaurus and Suevoleviathan. Based upon the analysis of the sampled specimens, the autapomorphies in this species include a long, delicate and slender anterior process of the maxilla (in lateral aspect) (Char. 2: 3); a narrow postorbital (Char. 15: 1); a short, robust and wide ischium (Char. 74: 2); and the shortness of the supranarial process relative to the subnarial process (Char. 0: 2). The longer ilium than ischium (Char. 78: 1) may also represent an important feature for this species. The species has a short jugal process of the maxilla hidden from lateral view by the jugal (Char. 4: 1), a character also seen in Suevoleviathan (this was not coded for Temnodontosaurus); and the rounded metacarpals not forming a tightly interlocking mosaic (Char. 39: 0), which is shared only with Suevoleviathan (synapomorphy). The ischium with greatly expanded medial edge (Char. 76: 0) is a synplesiomorphic character shared between this species and both outgroup taxa. The primarily rounded phalanges of the hindfin (Char. 60: 0) may also represent a retained primitive character of this species, shared with both the outgroup taxa and one oddity, MOS 8373 (Species #4). The glenoid contribution of the scapula is also extensive (Char. 67: 0) and is a retained primitive character, shared with Temnodontosaurus and Species #3. Other characters that may represent a unique character combination include a slight ‘flared’ curve to the anterior portion of the distal end of the femur (Char. 48: 1); a humerus/femur ratio of 1.3 or less (Char. 46: 0); and narrow, high and straight anterior-middle dorsal neural spines (Char. 85: 1).

7.2.1B. Species interpretation: There are several strong diagnostic characters for this species. The proposed species also displays features that are quite different to the currently accepted species of Ichthyosaurus, I. communis, I. breviceps, I. conybeari (McGowan and Motani, 2003) and I. anningae (Lomax and Massare, 2015). The exceedingly slender anteriormost portion of the maxilla and narrow postorbital appear to represent skull

38 autapomorphies not found in any of the other species (McGowan and Motani, 2003; Lomax and Massare, 2015). The shortness of the supranarial process relative to the subnarial process may represent a diagnostic character, but it is only coded for eight of the 15 specimens studied, so it is difficult to use with confidence. The ischium morphology is unusual and has not been found in any other species of Ichthyosaurus. A longer ilium than ischium is also unusual and is not found in any specimens from this study, but also occurs in I. breviceps (Massare and Lomax, 2014a). Forefin notching is present on this specimen, but it also occurs in other specimens that represent other species, and cannot be reliably used to distinguish this species from any other. Some of the other features are poorly supported. One example is the glenoid contribution of the scapula. This character was unknown in several of the studied specimens and may not be reliable for differentiating the species. Based on the well-defined characters (slender anteriormost portion of maxilla; narrow postorbital and ischium morphology), this represents a well supported new species of Ichthyosaurus. The skeleton of the specimen, on which the species is based, was 71% complete for this analysis.

7.2.1C. Specimen description(s): 1). NMW 93.5G.2 (Figures 10, 11). This specimen is the sole representative of this species, and thus provided information for all characters. Specifically, characters of the skull, forefin and pelvic girdle were significant for understanding this species. The specimen comprises a small, practically complete and fully articulated skeleton lying on its right side. It is one of few of the study specimens to have been collected since the 19th century and has the best stratigraphic information. The specimen was discovered in the 1980s, collected during a low tide from the foreshore at Doniford Bay, Watchet from the Blue Lias Formation, possibly the liasicus Zone (middle Hettangian). A significant portion (12 cm) of the anterior rostrum is missing and has been reconstructed, although the posterior is well-preserved with several teeth present. The left side of the skull is dorsolaterally flattened, displaying a small portion of the skull roof. Some features of the skull, including the maxilla, are well defined and display no crushing or minimal deformation (e.g. orbit). The pectoral girdle is present. The complete left scapula is preserved, but the right is missing. Only the left portion of the clavicle is exposed. Both coracoids are preserved but are overlain by the scapula and ribs such that the morphology is obscured. A single left forefin is

39 displayed in dorsal view and comprises the humerus and the most proximo-medial portion of the forefin. The carpal is notched anteriorly. A partial pelvis is preserved with ischia, one ilium and no pubes present, although it is possible that the ischia may represent an ischiopubis. One nearly complete hindfin, probably representing the left in dorsal view, is also associated with this section. The femur is disarticulated from the rest of the fin, which is largely articulated. Some of the elements are notched. The right femur is isolated and preserved near the ischia (Figure 11). An almost entire, fully articulated vertebral column is preserved. Underneath the ribs is a large mass of gastric contents, comprising fish scales and cephalopod hooklets. Based upon size (total length of 87.5 cm), this specimen may represent a juvenile.

Figure 10. The practically complete skeleton of NMW 93.5G.2, the sole representative of Species #1. Scale bar = 10 cm.

Figure 11. Close-up of the pelvic and hindfin area of NMW 93.5G.2. IS = Ischia. IL = Ilium. LF = Left femur. RF = Right femur. Scale bar = 2 cm.

7.2.2 SPECIES #2

7.2.2A. Species characteristics: This species is represented by a single specimen at Node 11, CAMSM J35183. The species displays features that are distinct from others, but also unusual features that are shared with basal plesiomorphic taxa. After the outgroup taxa and Species #1, it is the most plesiomorphic species. The autapomorphies of this species include very slender, small and delicate teeth (Char. 23: 2) and short supranarial and subnarial processes, relative to the external narial length (Char. 0: 3). Several retained primitive characters are shared with the outgroup taxa and Species #1. The absence of premaxilla/lacrimal contact below the external nares (Char. 1: 0) is shared with Suevoleviathan and the presence of contact between the maxilla and external naris (Char. 5: 1) is shared with Temnodontosaurus only. The anterior notch of the coracoid is extremely small (Char. 69: 1) and is shared with both outgroup taxa. The shape of the anterior notch (closed C-shaped) (Char. 70: 1) is also shared with Temnodontosaurus and one other specimen, MOS 8373 (Species #4). A widely ‘flared’ curve at the posterior portion of the distal end of the femur (Char. 49: 2) is shared with Temnodontosaurus, Species #1 and OUMNH J13799 (Species #4). The widely expanded shape of the ischium (distally) (Char. 75: 2) is also a retained primitive character, shared with Temnodontosaurus, Species #2 and another specimen, BRSMG Ce 16611 (Species #4). Finally, other features that may represent unusual combinations (or synapomorphies) include the absent or subtle root striations of the teeth (Char. 22: 1); narrow contact between the radius and ulna (Char. 35: 1); and approximately equal length/width of the intermedium (Char. 36: 2).

7.2.2B. Species interpretation: Some of the key characters that differentiate this species from those within the studied sample (and currently accepted species of the genus) represent a good basis for a distinct taxon. However, the autapomorphies of this species are not sufficiently reliable. Tooth morphology in Ichthyosaurus has been disregarded as a useful taxonomic character (e.g. McGowan and Motani, 2003), although seen to be of taxonomic value by others (e.g. Maisch, 1997; Maisch and Matzke, 2000), yet the very slender, small and delicate teeth of CAMSM J35183 appear to represent a unique feature. This suggests that tooth morphology may represent a useful

41 taxonomic character for differentiating at least one species, but this requires further study. The other autapomorphy, the short supranarial and subnarial processes relative to the external narial length, is difficult to accept confidently as only eight specimens of the 15 in this study could be coded for this character. What really helps to distinguish this species is the occurrence of retained primitive characters, shared with the outgroup taxa or Species #1, but not found in more derived species. The key features here include the absence of premaxilla/lacrimal contact below the external nares, presence of contact between the maxilla and external naris, and the extremely small anterior notch of the coracoid. The three features represent an unusual combination of characters not found in any of the other specimens in this sample. The combination of these characters, coupled with the proposed autapomorphies listed above, are unique in Ichthyosaurus. Pending further research, this can be tentatively considered to represent a potentially distinct species of Ichthyosaurus. The other features (mentioned above) may also be distinctive for this species, but as they are shared with some species/groups across the tree it is difficult to make a confident assessment of those characters. It should also be noted that the skeleton of the specimen, for which the species is based, was only 56% complete for this analysis.

7.2.2C. Specimen description(s): 1). CAMSM J35183 (Figure 12). As this specimen is the sole representative of this species, it provided information for all characters. Specifically, characters of the skull, tooth morphology, pectoral girdle and pelvic girdle were informative for this species. The specimen is a relatively small, but practically complete and articulated skeleton lying on its left side. It was collected from Street and formed part of the Thomas Hawkins Collection. It was previously a referred specimen of I. communis (McGowan, 1974. p. 9). A complete skull is present, but the posterior is partially disarticulated. The rostrum is well-preserved with numerous small and delicate teeth in the dental groove. There appears to be a repaired crack or break located across the orbit and lower portion of the jaw, but the skull bones match-up on either side. A practically complete left forefin (in ventral view) is present, although the deltopectoral crest of the humerus is broken and missing. Notching is absent. A large portion of the pectoral girdle is present with both coracoids articulated, well-preserved and articulated with a large portion of the right clavicle and interclavicle; the left side is

42 missing. Only the right scapula is present, but is partially buried by the distal portion of the clavicle. Both femora are present, one is located directly beneath a portion of the caudal vertebrae whereas the other (left?) is associated with a few phalanges. A fragmentary pelvis is exposed, comprising a small portion of the shaft and entire distal end of an articulated pubis and ischium (both are expanded). It has an almost entire, fully articulated vertebral column with the very distal end of the tail missing. However, there are some distal caudal vertebrae preserved posterior to the forefin which probably correspond to the end of the tail.

Figure 12. The nearly complete skeleton of CAMSM J35183, the sole representative of Species #2. Scale bar = 10 cm.

7.2.3 SPECIES #3 (‘CLUSTER 1’)

7.2.3A. Species characteristics: This is the first of two possible species that comprise several specimens (referred to as ‘Cluster 1’). The species is represented by five specimens at Node 29, CAMSM J59575, BRSMG Cb 3578, NHMUK 2013, NHMUK R3372 and AGC 12. There are no autapomorphies for this species, but some features that are shared with more basal taxa suggest an unusual combination of characters (metataxon). Several characters are shared with basal taxa including the absence of a circular depression in the articular surface of the humerus, anterior to the dorsal process (Char. 28: 1); extensive glenoid contribution of the scapula at least as large as the coracoid facet (Char. 67: 0); and narrow, relatively short and straight anterior-middle

43 dorsal neural spines (Char. 85: 0). Other characters of the species that may potentially represent an unusual combination or are shared synapomorphies with more basal species and Species #4 include: the tapering of the anterior margin of the jugal, extending partly between the lacrimal and maxilla (Char. 6: 0); a lunate, large and robust postorbital (Char. 16: 2); tibia length approximately twice as proximodistally long as more distal elements (Char. 53: 0); the nearly equal size of the pubis and femur (Char. 82: 0); and the nearly equal size of the ischium and femur (Char. 83: 0).

7.2.3B. Species interpretation: This possible species does not display any autapomorphies, although the nearly equal size of the ischium and femur was only found in this and MOS 8373 (within Species #4). It is possible that this character may prove to be autapomorphic, but it is currently not considered as such. A retained character, shared only with more basal specimens, comprises a fairly weak argument for this as a distinct species. The absence of a depression in the articular surface of the humerus is one of the strongest characters, separating this species from Species #4; although, four specimens in that cluster do not preserve that part of the humerus. The extensive glenoid contribution of the scapula is another character that may potentially represent something distinct, yet only a single specimen in this species was coded and renders the argument for the taxonomic usefulness of this character questionable. It is possible that the shape and height of the anterior-middle dorsal neural spines could be a useful character. This is shared with the current species and both outgroup taxa only. The other features of this species are difficult to use with confidence, as they are found to vary among the studied specimens and may reflect variation in the genus. All specimens in this group are united by a combination of features shared with more basal taxa, with one or two potentially distinct features. This may represent a distinct, but somewhat poorly supported species. Nevertheless, it is not associated with the neotype of I. communis and thus probably represents something different from that species. The skeletal completeness of specimens for this species ranges from 63–74%.

7.2.3C. Specimen description(s): 1). CAMSM J59575 (Figure 13). For the analysis, this specimen provided informative details on the forefin, pectoral girdle and pelvic girdle. The specimen is a fairly large incomplete individual lying on its right side, largely exposed in ventral view. It is nearly complete with a disarticulated assemblage of ribs extending anteroposteriorly; only the distal-most portion of the skeleton is missing. It was collected from Street and formed part of the Thomas Hawkins Collection. It was referred to I. communis by McGowan (1974: p. 9). The skull is practically complete and well-preserved, but is slightly displaced and upturned with the lowermost portion of the right side of the mandible exposed. A portion of the distal end of the skull is slightly distorted, including the orbit. Numerous well-preserved teeth are present. Both forefins are preserved and are practically complete, although the humerus of the right forefin is broken and at least half is missing. The left forefin is partially buried by ribs. Notching is absent. All of the pectoral elements are associated, although both scapulae are covered slightly by the coracoids. The coracoids are well-articulated. It appears that the right coracoid may be pathological (or damaged) as both the anterior and posterior notch are different from those of the left, more complete coracoid. A single partial hindfin is preserved and associated with pelvic elements. The hindfin comprises a femur and a largely articulated portion of the fin. Notching is present. One side of the pelvis is complete with all three bones associated; the pubis is spatulate-shaped distally. Portions of the matrix appear to vary with regards to the preparation, something that McGowan (1990) had flagged as a sign of composite specimens. However, the specimen appears to be genuine.

Figure 13. The partial skeleton of CAMSM J59575, a specimen encompassed within Species #3 (Cluster 1). Scale bar = 10 cm.

2). BRSMG Cb 3578 (Figure 14). This specimen has provided informative details of the skull, pectoral girdle, forefin and pelvic girdle. It is a relatively small individual lying on its right side. The majority of the ribs are broken and disarticulated, although the skull and vertebral column are fully articulated. It was found in Street and collected and prepared by Joseph Chaning Pearce. The skull is well-preserved, with minimal deformation, although the posteriormost portion is slightly damaged. There is a slight overbite although the upper portion of the skull may be pushed forwards. Some of the individual bones of the skull are well-defined, including the premaxilla (especially the supranarial and subnarial processes). The orbit is entire or crushing. Many teeth are preserved. The right forefin is fully exposed in ventral view and it is nearly complete associated with the pectoral girdle. The left forefin is overlain by a mass of ribs, but the humerus is well exposed in dorsal view. The majority of the mid-posterior portion of the fin is disarticulated. Almost an entire pectoral girdle is present, although the distal portions of the clavicle are broken and missing. Both coracoids are articulated and in ventral view; they are partially articulated with the right humerus and left scapula. Both scapulae are positioned close to the coracoids. The distal portions of the scapulae are covered over by other bones, but the proximal portion is exposed. Only a single femur (left) is preserved, positioned with a mass of broken ribs near the left forefin. At least some portions of

46 the pelvic girdle are present, including an isolated ilium next to the femur and another positioned proximal to the caudal vertebrae. An articulated pubis and ischium are surrounded by a large mass of ribs. The vertebral column with short neural spines is present with only the very distal end of the tail is missing.

Figure 14. The partial skeleton of BRSMG Cb 3578, a specimen encompassed within Species #3 (Cluster 1). Scale bar = 10 cm.

3). NHMUK 2013 (Figure 15). This specimen has provided detailed information of the skull, limbs and vertebral column. It is a nearly complete, large individual lying on its ventral side with almost all features of the skull and postcranium present. The specimen was collected in Street and formed part of the Thomas Hawkins Collection. The skull is very well preserved, with many of the sutures well defined. It is displayed in dorsolateral view with a large portion of the skull roof exposed. This is one of few specimens where the skull roof remains undamaged, visible and with sutures and features, such as the pineal foramen, very well-defined. The nasals are also well defined. The orbit appears to be slightly deformed but this is minimal and has not distorted any other portion of the skull. Many interlocking and heavily striated teeth are present. Both forefins are well- preserved. The right is exposed in dorsal view and is probably complete. The left is seen in dorsal view, but partially obscured by the vertebral column, although the exposed portion is largely articulated. The humeri are rather robust, especially the proximal-most portion. Notching is present. Unfortunately, only some of the pectoral girdle is exposed, although all of the elements are probably preserved. The vertebral column obscures much of the pectoral girdle, but portions of both scapulae are visible, the left coracoid is partly exposed and a portion of the clavicle is also visible. Both hindfins are present. The right is seen in dorsal view and is complete. The left is also exposed in dorsal view but is partially covered by the vertebral column. Notching of the tibiale, distal tarsal and metatarsal is present. At least one ilium is present and positioned next to the right femur. Other pelvic bones may be present but are preserved under the ribs. A closer examination of the specimen with the glass removed may allow for the identification of further pelvic elements. The entire vertebral column is preserved including the characteristic bend in the vertebral column defining the fluke.

Figure 15. The practically complete skeleton of NHMUK 2013, a specimen encompassed within Species #3 (Cluster 1). Note that this specimen is on display at the NHMUK, behind glass. Scale bar = 20 cm.

4). NHMUK R3372 (Figures 16, 17). This specimen has provided good information regarding the pelvic girdle and hindfin. It is a relatively complete and large-bodied specimen lying on its left side but largely exposed in ventral view. The specimen was collected from an unspecified quarry in Somerset and first described by Pearece (1846) who identified the small, embryonic ichthyosaur bones positioned near the pelvis (Figure 17). This represents the earliest documented account of ichthyosaur live birth in the literature (Lomax and Massare, 2012). The skull is fairly well-preserved, but damage to the posterior and proximalmost portions is evident. Certain elements of the skull are difficult to define and are positioned out of place. The posteriormost portion of the left side of the mandible is disarticulated and also exposed. Only a few teeth are preserved but are fairly robust. The right forefin is exposed in ventral view and the humerus is buried beneath the coracoid; it has been reconstructed at the distal end. The left forefin is exposed in ventral view and appears to be complete, but has been heavily reconstructed. The pectoral girdle is present with both coracoids associated. Only the right scapula is complete and a portion of the left is preserved. Only a small portion of the clavicle is present. A well- detailed and probably complete left hindfin is seen in ventral view. The right femur is only partially exposed. The femur is exceptionally long and slender. The left hindfin is positioned next to an articulated pubis and ischium with associated ilium. The pubis is widely expanded distally. Positioned next to these elements is the small embryo described by Pearce (1846). Despite a few breaks in the mounted specimen, the majority of the vertebral column is preserved including the tip of the tail.

Figure 16 (Above). The relatively complete skeleton of NHMUK R3372, a specimen encompassed within Species #3 (Cluster 1). The arrow points to the position of the embryo. Note that this specimen is on display at the NHMUK, behind glass. Scale bar = 10 cm.

Figure 17 (Below). Close-up of the embryo positioned in the pelvic area of NHMUK R3372; the arrow points to the skull of the embryo.

5). AGC 12 (Figure 18). This specimen has provided detailed information on the tooth morphology, skull elements, pectoral girdle and forefin. It comprises a largely articulated specimen lying on its right side exposed in ventrolateral view. The skull is displayed in left lateral. This specimen is part of the Gillett Collection and was therefore almost certainly found within one of the working quarries around Street. The skull is well-preserved with many of the bones individually defined; the maxilla is a good example. The anteriormost portion of the snout is missing. The bones surrounding the orbit (i.e. jugal, lacrimal and postorbital) are well-defined and display only a small portion of crushing (prefrontal). Some of the sclerotic ring is also present. The right side of the mandible is dislodged and exposed below the left. The teeth are exceptional; they are robust and have heavily striated roots. Both forefins are present. The right forefin (anteriormost) is exposed in dorsal view and is articulated with the coracoid. It is largely complete and is only missing the distalmost portion. The left forefin is also exposed in dorsal view, positioned above the right lying on top of the vertebral column and articulated with the left coracoid. It is largely complete and is missing only a few phalanges. Notching is absent. The pectoral girdle is nearly complete and fully articulated. Both coracoids are articulated with the clavicle and interclavicle. The left scapula is complete but buried slightly by the clavicle. The right scapula is absent. A single nearly complete hindfin is present, although a portion of the femur is broken and some of the digits are disarticulated. Some of the disarticulated phalanges are notched. Despite the unusual orientation, the vertebral column is almost complete. The posterior dorsal and anterior caudal vertebrae are coiled around the uppermost portion of the specimen and the distal caudal vertebrae are intertwined with the mass of ribs.

Figure 18. The partial skeleton of AGC 12, a specimen encompassed within Species #3 (Cluster 1). Scale bar = 10 cm.

7.2.4 SPECIES #4 (‘CLUSTER 2’)

7.2.4A. Species characteristics: This is the second of two suggested species that comprise several specimens (referred to as ‘Cluster 2’). The species is represented by eight specimens at Node 21, ANSP 17429, AGC 11, I. communis (neotype), MOS 8373, ANSP 15766, BRSMG Cb 4997, OUMNH J13799 and BRSMG Ce 16611. Note the inclusion of the neotype of I. communis. This species does not display any retained characters with the outgroup taxa, unless they are also shared with all other species. This is different from the three previously mentioned species. Thus this group of specimens represents the most derived species. The proposed autapomorphies in this species include the highest point (dorsoventrally) of the maxilla positioned posterior to the external nares (Char. 3: 0); the presence of a circular depression on the articular surface of the humerus anterior to the dorsal process (Char. 28: 0); and reduced glenoid contribution of scapula markedly smaller than the coracoid facet (Char. 67: 1). Other possible autapomorphies include the presence of a much larger robust proximal portion of the humerus compared to the distal (Char. 30: 0) and longer pubis than femur (Char. 82: 1). This species shares synapomorphies with Species #1 and Species #3, such as the presence of four digits in the hindfin (Char. 50: 1, although see below); presence of notching in the hindfin but not tibia (Char. 64: 2, although see below); the gradual decrease in mid-caudal centra height (Char. 84: 0);

52 and preflexural vertebral count of less than 80 (Char. 86: 0). Other features that may be useful include the narrow, high and straight neural spines on the anterior-middle dorsal vertebrae (Char. 85: 1) shared solely with Species #1; presence of a well- developed anterior notch of the coracoid (Char. 69: 0); ischium with lack of medial expansion (Char. 76: 1) both shared with Species #3; and a J-shaped posterior coracoid notch (Char. 71: 0), which is a derived character shared between this species and Species #2.

7.2.4B. Species interpretation: This possible species displays several autapomorphies that may represent useful taxonomic characters. The high point of the maxilla positioned posterior to the external nares is a good character as 11 specimens provide information on this character. For this character, there is a clear distinction between this species and all other specimens, which (where coded) displayed the highest point of the maxilla centred in the middle of the external nares. The same can be suggested for the presence of a circular depression in the humerus, although only ten specimens were preserved adequately to provide information for this character. However, the reduced glenoid contribution of the scapula (markedly smaller than the coracoid facet), one of the suggested autapomorphies, is a difficult character to use confidently to define this species. Despite five of the eight specimens in the current cluster providing information, only two of the more primitive species (Temnodontosaurus and Species #1) and only a single specimen in Species #3, displayed a different character state. This demonstrates that differences are evident with this character, but the lack of complete remains renders its taxonomic usefulness low. The large, robust proximal portion of the humerus might be an autapomorphy, but two specimens (ANSP 15766 and BRSMG Ce 16611) did not share this character, suggesting that this may represent variation within a species. The longer pubis than femur may also be an autapomorphy, yet one specimen (MOS 8373) displayed a different character state. This may raise the argument for the usefulness of the character, although this difference is represented by only one of the eight specimens. Several characters of this species are shared with Species #1 and Species #3. However, some of these features, such as the presence of four digits in the hindfin and presence of notching in the hindfin (not in tibia) are problematic as one specimen (I. communis neotype) did not display either. As the neotype specimen of I. communis is incorporated within

53 this group of specimens, it would be correct to assume that this species would be defined as I. communis. However, as a distinct species, it is poorly supported with only few potentially reliable characters. Of most concern are the several unusual features that are present in the neotype of I. communis, but are not found in any of the other specimens in the group. In addition, several characters found at the node defining the species vary within the cluster (e.g. distal and proximal ends of the humerus, humerus/femur ratio, intermedium length/width ratio, ilium/pubis relative length, pubis/ischium relative length and ilium/femur size). This all suggests a poorly supported species and perhaps further separation of these specimens could be made. It also raises the question of whether I. communis is actually present in the Somerset material at all (see Section 8). The completeness of specimens for this species ranges from 57–92%.

7.2.4C. Specimen description(s): 1). ANSP 17429 (Figure 19). This specimen provided important details on the pelvic girdle and hindfin. It is a practically complete specimen lying on its left side, but with portions of the skeleton displaced or partially disarticulated. There is little information present with this specimen, but the information and preservation of the skeleton suggests it was collected from a quarry in Somerset. Like many of the historic specimens, it is displayed and set into a large wooden frame. The posterior portion of the skull is almost completely crushed and displaced with many ribs and vertebrae lying on top. However, this has allowed several features of the skull, such as the basioccipital with extensive extracondylar area, to be seen. The rostrum and all elements anterior to the orbit are unaffected by this crushing and are well- preserved. Some of the skull sutures are very clear. Numerous well-detailed and striated teeth are present. Both forefins are preserved. The right is positioned posterior to the skull, in dorsal view and is practically complete. The small depression in the proximal head of the humerus is evident. The left forefin is preserved under a mass of ribs and is partially complete. The humerus is damaged. Notching is absent. Much of the pectoral girdle is missing with only one nearly complete but poorly preserved coracoid (left). The other is damaged and largely missing. Only the ?right scapula is present but buried under the posterior portion of the skull with just the distal end showing. A practically complete hindfin (the right in dorsal view) is present. There appears to be a depression or ?break positioned on

54 the proximal portion of the shaft. The femur is a fairly robust bone; it is rather wide at the distal end. The anterior elements of this fin are damaged and missing. The left femur is also present. Presence (or absence) of notching cannot be confirmed. The femora are positioned alongside a virtually complete pelvis. The right hindfin is in contact with an articulated ischium and pubis and associated ilium. The left femur is associated with the articulated left pubis and ischium; the other ilium cannot be located. The pubis is widely expanded distally. The majority of the vertebral column is present and articulated, but the anterior is largely distorted and intertwined with the skull. The tail of this individual appears to become abruptly shorter and may represent a composite (pers. obs.). This is also suggested by the matrix of the specimen which changes from ‘smooth’ to ‘patchy’ (see McGowan, 1990). The specimen requires conservation as many areas of matrix are broken.

Figure 19. The practically complete skeleton of ANSP 17429, a specimen encompassed within Species #4 (Cluster 2). Scale bar = 10 cm.

2). AGC 11 (Figure 20). This specimen has provided good information regarding the forefin and pectoral girdle. It is a medium-sized and moderately well- preserved individual lying on its left side. The specimen is nearly complete and fully articulated but with some of the elements displaced or missing. The preservation is not as good as some of the other sampled specimens. As an Alfred Gillett specimen, this was almost certainly collected from one of the working quarries around Street. The skull is nicely preserved, but the posteriormost portion is damaged and some of the bones are missing. The snout is surprisingly quite short with the orbit large; perhaps the large orbit is exaggerated by the shortness of the snout and the missing portion of the skull. The lowermost portion of the mandible is exposed from both

55 sides and is dislodged. Few small and striated teeth are present. Some of the teeth have fine striations that do not extend all the way to the top; some of the roots have good fluting and some are bulbous. This shows that some of the teeth are waisted, but others are not and supports the poor taxonomic usefulness of teeth. Both forefins are present and are probably nearly complete. The humerus of the right forefin, seen in dorsal view, is positioned on top of the middle portion of the left; there is a slight depression evident in the humerus. This forefin is displaced, positioned posterior to the left. The left forefin is in ventral view and is associated with the pectoral girdle. The left forefin displays a bifurcation in the metacarpal row, while the right forefin displays a bifurcation in the first row of phalanges. This may be pathological or simply unusual. The intermedium also appears to be a different shape on either fin; this may be due to the orientation of the fins (ventral v. dorsal) or may simply represent shape difference in the intermedium. It may suggest that the taxonomic usefulness of these characters is poor. Notching is absent. A good portion of the pectoral girdle is present with both coracoids associated and practically complete, although the left is broken. The left scapula is complete but damaged in the centre; the proximal portion is wide and rounded. The right scapula is also preserved but is awkwardly oriented above the coracoids and partly buried by the distal end of the skull. Portions of the clavicle are covered by the right coracoid. A single hindfin is present and is probably the right in dorsal view; the femur is damaged slightly. Some of the middle and posterior phalanges are missing. The left femur is positioned anterior to the right hindfin next to the vertebral column and covered by ribs. Notching is present. A single ?ilium is the only representative of the pelvic girdle. The vertebral column is complete, fully-articulated and well defined, but partially overlain by ribs. A nicely defined bend in the tail is present. Stomach contents are preserved.

Figure 20. The nearly complete skeleton of AGC 11, a specimen encompassed within Species #4 (Cluster 2). Scale bar = 10 cm.

3). I. communis (neotype) NHMUK R1162 (Figure 21). This specimen provided detailed information regarding the skull, forefin, pelvic girdle, hindfin and vertebral column. It was designated the neotype of I. communis by McGowan (1974) and is currently on display behind glass. The specimen was collected from Lyme Regis, Dorset and is the only Ichthyosaurus specimen used in this study that is from outside Somerset. It is an almost complete skeleton of a medium-sized ichthyosaur lying on its right side, missing only the distalmost portion of the skeleton. The skull is complete with just some deformation of the posterior portion, although many of the sutures are difficult to determine. The orbit shape has not been affected by the deformation. The mandible is well preserved and comprises many well-striated teeth. A single complete forefin (left in dorsal view) is preserved and associated with the pectoral girdle. The humerus is orientated at an unusual angle so correct interpretation is difficult. The right forefin is a jumble of phalanges, positioned underneath the left. Notching is absent. Only a small portion of the pectoral girdle is exposed. Fragments of both coracoids are present, but are largely overlain by the forefin and by one another; some of the clavicle is exposed. A single scapula (right) is associated with the right coracoid and is complete; it is buried slightly at the anterior end and is rather slender. The left hindfin is well-preserved and in dorsal view. Both the anteriormost and posteriormost portions are flared. There appears to be a significant number of bones positioned at the distal end of the hindfin, which suggests either an exceedingly long fin or perhaps the presence of the second hindfin (examination without the glass is required for confirmation). Notching is absent. The femur is associated with the complete left side of the pelvis. The pubis, ischium and ilium are all associated together. The pubis is spatulate-shaped distally.

The vertebral column is fully articulated missing only the distalmost portion. Immediately posterior to the skull, the neural spines are exceptionally high and stay high until the mid-posterior dorsals. As well as the ribs, a mass of gastralia is present. A further description of this specimen is in Section 8.

Figure 21. The neotype specimen of I. communis (NHMUK R1162), a specimen encompassed within Species #4 (Cluster 2). Note that this specimen is on display at the NHMUK, behind glass. Scale bar = 10 cm.

4). MOS 8373 (Figure 22). This specimen has provided important information regarding the skull, forefin, hindfin and vertebral column. It represents a practically complete medium-to-large-sized individual lying on its right side and has little deformation or displacement. It is recorded as ‘collected from Somerset’. The specimen is in one main block of matrix with portions fitted together backed by resin (for safety). Of interest, from a conservation point of view, the entire specimen is curved in concave aspect. The skull is finely preserved with only the minimal amount (if any) of deformation, which appears to have slightly compressed the orbit, but this is not reflected elsewhere in the skull. Although some of the posterior skull bones are present, there is matrix surrounding them and other bones may still be preserved. Sutures are nicely defined, especially the subnarial and supranarial processes, the lacrimal, jugal and postorbital. The rostrum is complete with many slender, striated and well-defined teeth. Both forefins are present, although the right is largely covered by the dorsal vertebrae and neural spines. The left is in dorsal view and practically complete with only a portion of the middle section missing. The humerus is robust, especially towards the proximal end. A depression is evident in the

58 proximal portion of the humerus, behind the dorsal process. Noteworthy are the rather spacious gaps between the phalanges, which are not as interlocked as in other specimens. Notching is absent. A portion of the pectoral girdle is present, although the clavicle is not seen and is perhaps covered over by matrix and ribs. Both coracoids are preserved and articulated, although only the left is nearly fully exposed, associated with the left humerus. The complete left scapula is positioned next to the humerus and coracoid; it is fairly slender posteriorly. The left hindfin (in dorsal view) is preserved and is probably complete. The femur is quite slender with a pronounced flare to the posterior. Phalanges are quite well spaced with no real interlocking. Notching is present. The right femur is present, but only the proximal portion is exposed, with the rest buried between ribs and beneath vertebrae. The left side of the pelvis is present and associated with the left hindfin. Here, the pubis and ischium (which are damaged) are associated together, positioned next to the ilium. The distal portion of the pubis is spatulate-shaped. The other ischium is preserved near the isolated right femur. The vertebral column is fully articulated, displays the characteristic tail bend and is practically complete, missing just the last few vertebral centra of the tail. Aside from ribs, some gastralia are also preserved.

Figure 22. The practically complete skeleton of MOS 8373, a specimen encompassed within Species #4 (Cluster 2). Scale bar = 10 cm.

5). ANSP 15766 (Figure 23). This specimen has provided important information regarding the skull, forefin, hindfin and elements of the vertebral column. It comprises the skeleton of a large, practically complete individual lying on its right side. The specimen is recorded as ‘collected from Somerset’. The skull is exceptionally preserved with no distortion, although a small portion of the skull roof is exposed. Skull suture definition is among one of the best in the sampled specimens and other Ichthyosaurus remains (pers. obs.). Only the very anteriormost portion of the snout (the tip) is missing, but it was certainly present (based on the space left behind). Many interlocking and nicely defined teeth are in the rostrum. Both forefins are preserved with the left the more complete (and exposed) of the two. This forefin is complete and is in dorsal view; the humerus is rather robust. The right forefin is most probably in dorsal view, but is largely covered by dorsal vertebrae and neural spines. Notching is absent. Only portions of the pectoral girdle are exposed, including the left side of the clavicle, the left scapula (although only the shaft and distal portion is present) and the ?left coracoid, although portions of the coracoid are covered over. It is possible that the rest of the pectoral girdle is covered by the ribs and/or matrix. Both hindfins, in dorsal view, are present and are nearly complete, perhaps missing a few distal phalanges. The left hindfin is positioned posterior to the right; the right is nestled next to the vertebral column and partially overlain by caudal ribs. The posteriormost portion of the distal end of the femur has a noticeably flared edge. The femur is rather slender. Notching is present. Pelvic bones are not present, although a curved bone positioned above the right femur, under the ribs, is probably an isolated ilium. The vertebral column is well-preserved and fully articulated. A fused atlas-axis is positioned posterior to the skull. The neural spines are consistently fairly low. Only the very distal end of the tail is missing.

Figure 23. The practically complete skeleton of ANSP 15766, a specimen encompassed within Species #4 (Cluster 2). Scale bar = 10 cm.

6). BRSMG Cb 4997 (Figure 24). This specimen has provided good, informative details of the pectoral girdle, hindfin and pelvic girdle. It is a practically complete, fully articulated large individual lying on its left side with only some of the remains slightly displaced. This is one of the specimens that was behind glass and could not be examined in more detail. To complicate this further the specimen is mounted on a slant tilted away from the glass (not mounted parallel to glass). It was collected from Street. The skull is complete and is well-preserved; only the distalmost portion is slightly distorted and the underside of the mandible (right side) has become dislodged and displaced. The orbit has been unaffected by the distortion and retains its original shape. The maxillary/narial section is slightly compressed and difficult to interpret properly. There is a bend from the maxillary segment and the remainder of the premaxilla and rostrum; this is probably due to preservation. Numerous robust, large and heavily striated teeth are densely packed in the snout. Both forefins are present; the right partly overlaying the distal portion of the skull. Noticeably, the humerus is robust and massive, especially when compared to the size of the coracoid and scapula. This forefin is practically complete, but many of the elements, despite being in contact, are displaced or oddly oriented. It is associated with the right coracoid and scapula. The left forefin is slightly posterior to the right, but is also practically complete, though the humerus is broken proximally and partially buried by a mass of ribs. A large crack splits the forefin in half. The anteriormost phalanges are large polygons. Notching appears to be present although it is difficult to determine completely and may actually represent distinct cracks in the

61 bone, although that is unlikely (confirmation is required with the removal of the glass). The right coracoid and scapula are preserved and articulated; they are both rather large and robust bones. The left coracoid is associated with the left forefin, but is only partially preserved. Both hindfins are preserved and are practically complete. The most anterior of the two is the left which is seen in ventral view; unfortunately many of the distal phalanges have fallen away (?lost) with several spaces present in the matrix showing where they once sat. The right hindfin, seen in dorsal view, is positioned marginally posterior to the left and is more complete. The femora have a distinctive flare to the posterior edge and are fairly robust. Due to damage of some digits, the presence/absence of notching cannot be confirmed. The right hindfin is next to a single ilium, presumably the right. Positioned anterior to this is a partially associated complete ischium and fragmentary pubis (probably the right); the rest of the pubis has become detached and is preserved slightly posterior to it. The left hindfin has both the left ilium and ischium preserved next to it with a small fragment of what is presumably the pubis. The pubis is spatulate-shaped distally. The entire vertebral column is preserved and fully articulated with exceptional preservation. The very distal end of the tail is also present. Many of the anterior vertebrae and neural spines are covered over by ribs.

7). OUMNH J13799 (Figure 25). This specimen has provided informative details on the pectoral girdle, forefin, pelvic girdle and hindfin. It is an exceptional, almost complete, large-bodied ichthyosaur lying on its dorsal side positioned in ventral view. It was collected from the Blue Lias of Street and said to have been excavated in 1856. It was formerly in the possession of the seventh Duke of Marlborough and was presented to Oxford University in 1865. However, there is some indication that the specimen may be a composite. Firstly, the skull is set on what appears to be a different block of matrix, raised slightly above the rest of the skeleton. The skull is also in dorsal view and is opposite to the rest of the skeleton; although, it is not impossible that the skull had twisted round during preservation, due to taphonomic processes. Secondly, the tail appears to be complete, but appears suspiciously short, though this may be due to the scattering of many anterior vertebrae. The matrix also varies in colour and composition throughout, something McGowan (1990) highlighted when looking for potential composites. Disregarding these comments, the specimen is described as if it represents a single individual (it requires further examination). The skull is well-preserved and displays the skull roof including the supratemporal fenestrae and pineal foramen. Some of the posterior portion is damaged. The rostrum is complete and few teeth are displaced to the left side of the skull; they are large and striated. Both forefins are preserved in ventral view and are exceptional. They are both complete and bear minimal damage. The oddly shaped radius and ulna and the posterior accessory digit contact with the ulna (a definitive edge to the ulna is present) appears unusual. The humerus is quite short and is thin at the centre of the shaft. Both forefins are articulated with a complete and almost fully articulated pectoral girdle. Notching is absent. The coracoids are nicely articulated and associated with all bones of the pectoral girdle, including the clavicle and interclavicle; the latter is on top of the coracoid articulation. Both scapulae are present although the right is marginally displaced from articulation with the coracoid. The shaft and distal portion is quite delicate, but the anterior is wide and large. Both hindfins are present although both are disarticulated. The left is largely articulated, however, with just the femur displaced (positioned to the posterior). The right is a jumble of phalanges with the femur associated. Proximally, the femora are quite slender and thin. Notching is present. Both hindfins are associated with pelvic elements. Next to the right femur is the right ilium; positioned slightly anterior is an associated pubis and ischium. The left femur is next to the left

63 ilium and ischium and possibly the pubis. The pubis is spatulate shaped-distally. The majority of the vertebral column is present, but almost all of the anterior elements from the pelvic area forward are dislodged and scattered throughout the column.

Figure 25. The nearly complete skeleton of OUMNH J13799, a specimen encompassed within Species #4 (Cluster 2). Note the arrows pointing out the difference in height/matrix surrounding the head, suggestive of a possible composite. Scale bar = 10 cm.

8). BRSMG Ce 16611 (Figures 26, 27). This specimen has provided informative details of the pectoral girdle, forefin, pelvic girdle and hindfin. It is represented by an almost complete and articulated, large-bodied individual lying on its dorsal side, exposed in ventral view. It is a rather fine example and was first described by Deeming et al. (1993) as an Ichthyosaurus specimen with an embryo (Figure 27). In comparison to the majority of specimens examined in this study (apart from NMW 93.5G.2) this is a relatively recent discovery, found in 1985. The remains were found in beds of the Psiloceras johnstoni subzone (Hettangian) and were collected at low water mark at Kilve (Deeming et al., 1993). It should be noted that the tip of the mandible and the tail bend have been restored and are not original (Deeming et al., 1993). The original specimen is behind glass at the BRSMG and could not be examined. However, a cast of the same specimen (unnumbered cast) is on display at the LEICT and was examined. The preserved skull is positioned in ventral view and largely crushed, although the posterior portion of the mandible remains undamaged. Some teeth can be seen in the dental groove but are only marginally exposed. Both forefins are present with the right almost articulated with the coracoid. The right is exceptionally well-preserved and complete, and the

64 humerus is quite robust. The left is slightly displaced from the pectoral elements and is pathological. The individual digits and phalanges are pushed to the side, which generates a rather odd-shaped forefin. The left humerus morphology is also quite different to the right humerus in that it is much more slender and delicate. Notching is present. Despite the crushing and displacement of the skull the majority of the pectoral girdle is intact and is positioned directly next to (and partly on top of) the skull. Both coracoids are articulated and have a fairly wide ‘neck’; the clavicle and interclavicle are articulated with both coracoids. Both coracoid notches are rather wide. The right scapula is articulated with the coracoid. The left is displaced and broken distally. The shaft and distalmost portion are quite delicate, but the anterior is fairly wide. Both hindfins are preserved, displayed in ventral view. The right hindfin is complete and the left is only represented by the femur and a few paddle elements. The femora are rather long, but quite robust. Notching is present. The right hindfin is associated directly with the right pubis and ischium which are almost articulated. The right ilium cannot be located but may be within a mass of ribs. The left hindfin is also closely associated with an articulated pubis and ischium. The pubis is widely expanded distally. The left ilium is positioned posterior to these elements. It is in this area that the complete and exceptionally preserved embryo is located. Aside from the restored distal portion of the vertebral column, the majority is present and largely articulated. It should be noted that the very distal ends of the ribs are ‘flattened’ and exceedingly wide.

Figure 26. The practically complete skeleton of BRSMG Ce 16611, a specimen encompassed within Species #4 (Cluster 2). The arrow points to the position of an embryo. Note that this specimen is on display at the BRSMG, behind glass, although a cast is readily accessible (and on display) at LEICT. Scale bar = 10 cm.

Figure 27. Close-up of the embryo positioned n the pelvic area of BRSMG Ce 16611. Scale bar = 2 cm.

7.2.5 SPECIES #5?

7.2.5A. Species questionable: As this possible species is very poorly supported at the defining node and may be better placed in one of the other species, it is not described with a ‘species characteristics’ and ‘species interpretation’ as was done with the other species.

Species #5? is found at Node 5 and consists of a single specimen, AGC 17. For this analysis, the skeleton of the specimen is 52% complete and is the least complete of the study specimens. Many potentially important characters of the other species are not preserved on this specimen and, for example, none of the skull characters could be coded. Much of the pelvic girdle characters were also uncoded. The specimen displays no autapomorphies. It appears to be distinct within this analysis based on some poorly supported characters, shared with other specimens across the tree and with individual specimens in other species. It does not share any synapomorphies with one distinct species or solely share any specific features with more primitive species or the outgroup taxa. All of the features (where coded) are either present in all sampled specimens (e.g. ulna wider than long (Char. 33: 1); posterior accessory digit present (Char. 44: 1); and tibia wider than long (Char. 51: 1)) or shared with Ichthyosaurus specimens (e.g. digital bifurcation present (Char. 61: 1) and scapula with large acromion process absent, but anterior expansion of proximal portion present (Char. 65: 1)). Many of the other features of this specimen which may represent something distinct are found to fluctuate throughout both Species #3 and Species #4 (e.g. the contact between the radius and ulna (Char. 35); a wider intermedium than radiale (Char. 38: 1); circular shape of the tibia (Char. 52: 2); and intermedium and tibiale equal in size). Given that these features are found on this specimen, and across both cluster 1 and 2, they are probably not good taxonomic characters and probably represent individual variation. The only potential apomorphy is the nearly equal width of the proximal ends of the femur (Char. 47: 0); although a single specimen from Species #4 also shows this, so it may be an artefact of preservation/orientation. This suggests that if the specimen was more complete it would probably group with one of those species. Moreover, this specimen could not be coded for character state 28 (the presence or absence of a circular depression in the humerus) as the proximal portion of the humerus is worn. However, the

67 possibility of the presence of a depression may be evident. If indeed this is a depression, then this specimen may therefore be tentatively placed within Species #4. However, as mentioned above, that group of specimens requires further assessment. As a final comment, examining the other trees retained in this analysis, this specimen was also found to group with specimens that are currently in Species #4. This may further suggest its placement in that group.

7.2.5B. Specimen description(s): 1). AGC 17 (Figure 28). As this specimen is the sole representative of this ‘species’ it provided all of the characters, although the forefin, pelvic girdle and hindfin were informative. This is a partial skeleton of a relatively large individual, disarticulated into two large portions, lying on its ventral side and positioned in dorsal view. As an Alfred Gillett specimen it was almost certainly collected from one of the working quarries around Street. The skull is fairly poorly-preserved, yet practically intact; it is crushed posteriorly and many of the bones are unrecognisable. The premaxilla is complete and numerous well preserved teeth are present. The dentary cannot be seen. An isolated basioccipital with wide extracondylar area is positioned slightly anteroproximally to the skull on a small separate block. The colour and morphology suggests that it belongs with this specimen. A natural ‘divet’ is located in the centre of the condyle and the exoccipitals surround two-thirds of the foramen magnum. Both forefins are present and are in dorsal view. The uppermost forefin is the right and is partially covered by ribs. It is articulated with the coracoid. The left forefin is probably complete and fully articulated. The humeri are robust. Notching is absent, but some damage to some proximal elements resembles notching. It is probable that a complete pectoral girdle is preserved, although as this specimen is in dorsal view the majority of the coracoids and interclavicle are covered over. Nevertheless, the left scapula is well exposed and is robust throughout. The right scapula cannot be located. A single nearly complete hindfin, representing the left in dorsal view, is present. The femur is robust anteriorly with a fairly slender shaft, while the posterior is quite wide. Notching is present. The femur is associated with an articulated pubis and ischium. Other bones which surround the femur (and are broken or covered over) probably represent the rest of the pelvis. The distal end of the preserved ischium is covered, but the rest of the ischium is fairly wide. The pubis is widely expanded distally. The anterior and middle portions of the vertebral column

68 are largely fully articulated, but there is a significant split from the pectoral girdle, whereby the entire vertebral column is separated (yet still remains largely articulated). Beyond the pelvis few vertebral centra are preserved; some are scattered across the matrix. Some of the ribs are ‘flattened’ and widely spaced at their distalmost end.

Figure 28. The semi-complete skeleton of AGC 17, the only specimen identified as Species #5? although probably better positioned in one of the two clusters of Species #3 or Species #4. Scale bar = 10 cm.

7.2.6 ADDITIONAL SPECIMENS

Based upon the species interpreted above, the less complete individuals, described below, have been assessed and have temporarily been placed, where possible, within one of those species. The following comprises a brief description of each specimen and the reason for its tentative referral to one of the species.

7.2.6A. Specimen description(s): 1). AGC 14 (Figure 29). This is the least complete individual that was considered for the analysis, being only 30% complete. As an Alfred Gillett specimen, it was almost certainly collected from one of the working quarries around Street. The specimen comprises a practically fully-articulated individual lying on its ventral side, exposed in dorsal view. Despite the appearance of a nearly complete individual, many of the characters used in the analysis were either obscured, not preserved or

69 difficult to determine. The skull is fairly well preserved with much of the skull roof present, although the very distal portion is damaged. The rostrum appears to be complete and relatively short, with some of the dentary exposed. The right forefin is almost fully exposed, although the humerus is buried underneath ribs and matrix. The left forefin is missing, but an isolated humerus is present, with a depression in the head of the humerus. Notching is probably absent. The entire pectoral girdle is probably buried under a mass of ribs and vertebrae. Both femora are preserved and the left is fully exposed. The right femur is associated with an articulated pubis and ischium. The ischium is quite wide distally and the pubis is widely expanded distally. The vertebral column is probably all present, including the very distal end of the tail. Some of the anterior caudal vertebrae are disarticulated and dislodged. Given the low percentage of characters seen on the specimen it is exceedingly difficult to assign it with confidence to one of the described species. The presence of a depression in the humerus (Char. 28: 0) suggests that this could be placed within Species #4 (Cluster 2), but the referral is poorly supported. The specimen also has stomach contents preserved, comprising numerous cephalopod hooklets.

Figure 29. The partial skeleton of AGC 14, an additional specimen not used in the primary analysis, but may belong to Species #4 (Cluster 2). Scale bar = 10 cm.

2). AGC 15 (Figure 30). For this analysis, the specimen was 41% complete, but none of the pectoral, pelvic, hindfin or vertebral column characters could be coded. As an Alfred Gillett specimen, it was almost certainly collected from one of the working quarries around Street. Despite the low percentage of characters for the matrix, the specimen represents a well preserved anterior portion of a skeleton, comprising the skull, forefins, some elements of the pectoral girdle and few vertebrae

(and neural spines). The skull is complete, but the posteriormost portion is damaged. The orbit is distorted also. The rostrum is well preserved with many fairly robust teeth with heavily striated crowns (but with blunt tips). The lowermost forefin is the right in dorsal view and is practically complete; the humerus is rather robust. The left forefin is unusually preserved as it is positioned between the distal portions of the mandibular ramus, an unusual taphonomic phenomenon. Notching is absent. Nothing much can be gleaned from the pectoral or vertebral elements, although two notches are present in the coracoid. As the specimen displays none of the defining characters of the previously described species, and a significant portion of the specimen is missing, it is not possible to assign it to one of the species.

Figure 30. The skull and anterior portion of AGC 15, an additional specimen not used in the primary analysis, but may belong to Species #4 (Cluster 2). Scale bar = 10 cm.

3). Unnumbered BGS specimen (Figure 31). For this analysis, the skeleton of the specimen is 49% complete, but none of the skull or humerus characters could be coded. It is a fairly complete large-bodied individual lying on its ventral side. The specimen was collected from the Lias of Walton, near Street. The skull comprises just the mandible and a few scattered fragments; in contrast to the postcrania it is flipped and lying on its dorsal surface (perhaps a suggestion of a composite, similar to OUMNH J13799). Apart from the skull, the rest of the skeleton is fairly well preserved and fully articulated. It appears that both forefins were once present (there is an outline in the matrix where the right forefin would have been). The left (in dorsal view) is fairly complete, but the majority of the humerus is covered by ribs. Notching is absent. A long and fairly robust scapula (the left) is the sole representative of the pectoral girdle, although a portion of the coracoid is present, but covered over by ribs and vertebrae. Both hindfins are present (in dorsal view) and practically complete. The femora are fairly slender with a relatively robust proximal portion. The fibula is noticeably larger than the tibia. Notching is present. The right side of the pelvis is complete with a fully associated pubis, ischium and ilium, although a fragment of the ilium is missing. The pubis is widely expanded distally. As no characters can be gleaned from the skull or the humerus, it is difficult to assign this to any of the discovered species. However, it can be ruled out from referral to Species #1 on the basis of the pelvic girdle morphology. It is most probable that this specimen could be placed within Species #3 or Species #4, although it is better regarded as Ichthyosaurus sp.

Figure 31. The fragmentary skull and articulated skeleton of an unnumbered BGS example, an additional specimen not used in the primary analysis and one that could not be reliably referred to any of the species. Scale bar = 10 cm.

4). MOS 166/1992 (Figure 32). For this analysis, the skeleton of the specimen is only 34% complete. However, the specimen actually represents a small-bodied and probably complete individual lying on its ventral surface, exposed in dorsal view. The preservation is quite poor. In comparison to all other sampled specimens this is the most recent discovery. It was collected in November, 1992 from the Kilve Shales, Coroniceras bucklandi Zone (lower Sinemurian) on the beach at Kilve. The size of the individual suggests it is probably a juvenile. The skull is complete, but poorly- preserved with sutures poorly-defined. Both forefins are present and are displayed in dorsal view; the humeri are largely covered over. Notching is absent. No pectoral girdle elements are exposed and are presumably buried beneath the skeleton. The right hindfin is nearly complete, but the femur has become detached and is positioned slightly anterior to the rest of the fin intertwined with ribs. Notching appears to be absent. An isolated ilium is preserved near the femur. A possible ischium is also present. The vertebral column is associated, but almost all of it is disarticulated or dislodged. Given the poor preservation of the specimen, it is difficult to assign it with confidence to one of the defined species. It is better regarded as Ichthyosaurus sp. Stomach contents are preserved and include the bones of small fish.

Figure 32. The practically complete, yet poorly preserved, skeleton of MOS 166/1992. An additional specimen not used in the primary analysis and one that could not be reliably referred to any of the species. Scale bar = 10 cm.

5). NHMUK R5595 (Figure 33). The skeleton of the specimen is 48% complete for this analysis. This is a large-bodied individual, practically complete and lying on its ventral surface exposed in dorsal view. The specimen was collected from Street in 1847 by Dr W. Cunnington; it was subsequently donated to the NHMUK in 1931. Of the preserved skeleton, the skull is most probably the least well-preserved. The posterior appears to be damaged and largely disarticulated with a portion of the skull roof pushed forward. The rostrum is probably complete, but the dentary is not present. Numerous teeth are preserved. Features of the skull could not be coded. Both forefins are present, practically complete and are displayed in dorsal view. The lowermost forefin is the right. A distinct crack runs across the anterior portion of this fin whereby there is a colour difference between the previous portion and the latter. Parts of the fin do appear to match up, but it is dubious and perhaps a composite. The left is well preserved and fairly slender. The humerus is quite short and robust. Notching is absent. Almost the entire pectoral girdle is buried beneath the skeleton, although both scapulae are protruding from the mass. Only the distal portion of the scapulae can be seen and are fairly robust but have a somewhat slender shaft. Both hindfins are present and exposed in dorsal view and are practically complete. Notching is present. Both hindfins are associated with some elements of the pelvic girdle including the ilia. The vertebral column is fully articulated and exceptionally preserved, including the very distal end of the tail. As well as some of the key elements being missing (e.g. skull characters and pectoral elements), the specimen was one of the highest mounted on display at the NHMUK (behind glass) and good quality photographs and notes were difficult to acquire. Given the fairly high percentage of coding for the specimen, it would suggest that this example would be easier to assign to one of the previously mentioned groups, although this is not the case. Some of the noticeable features of this specimen include the short robustness of the humerus, the fairly slender shaft of the scapula (Char. 66: 1) and the roughly equal size of the femur and ilium (Char. 81: 0). However, these features are found to vary between both Species #3 and Species #4 and this suggests they are probably not good taxonomic characters. For this reason, the specimen is better regarded as Ichthyosaurus sp.

Figure 33. The practically complete skeleton of NHMUK R5595, an additional specimen not used in the primary analysis and one that could not be reliably referred to any of the species. Note that this specimen is on display at the NHMUK, behind glass. There is some glare from lighting (positioned in the centre of the specimen). Scale bar = 20 cm.

8. DISCUSSION

8.1. Interpretation of results: The data matrix comprised 43 distinct characters derived from numerous data matrices supplemented with 44 additional characters identified in this study (see Appendix A). The majority of characters have been useful for distinguishing differences among specimens and individual species. Based on a strict interpretation of the results of this study, a total of four possible species of Ichthyosaurus has been identified (disregarding Species #5? as a distinct species on the grounds explained in Section 7.2.5). Of the putative Ichthyosaurus species, Species #1 is the most primitive and is represented by a single individual which displays a strongly defined set of autapomorphies. This species shares features with the more primitive outgroup taxa (Temnodontosaurus and Suevoleviathan) which are not shared with derived species. Species #2 is represented by a single individual and has two autapomorphies. It also possesses features which are shared solely with either, or both, of the outgroup taxa. Species #3 is represented by a cluster of five specimens. Despite the lack of any autapomorphies, this species does share features with more primitive taxa which are not found in the most derived species (Species #4). Species #4 is represented by a total of eight specimens including the neotype of I. communis; it has at least two autapomorphies, perhaps three. This species does not share any

75 characters (solely) with either of the outgroup taxa, but does share some features with Species #1 and Species #3.

Based upon the four possible species, Species #1 is the best defined. Together, the strong autapomorphic characters which are unique to the species and the primitive characters shared with outgroup taxa represent a well supported species. Species #2 is another fairly well-supported species; although, one of the autapomorphies is based on tooth morphology, which has been considered unreliable (McGowan and Motani, 2003). The lack of autapomorphies in Species #3, but unusual combinations of characters, may suggest that this is a valid species. Similar interpretation can be stated for Species #4. Despite the proposed autapomorphies for this species, some are poorly supported. Also, considering the issues surrounding the neotype of I. communis (see below) this species is poorly supported. As detailed in this study, the findings suggest that only two of these species (Species #1 and Species #2) can be confidently separated with the potential of, at the very least, another distinct species represented between Species #3 and Species #4, or both. However, as both clusters of specimens are poorly supported, it is difficult to determine which of these groups (and specimens within) could represent a distinct species.

8.2. Comparison of the neotype with Ichthyosaurus specimens in Somerset: To assist with the recognition of additional species of Ichthyosaurus in Somerset, as well as I. communis itself, the neotype was included in the cladistic analysis of this study. As discussed above (Section 7), four possible species (disregarding Species #5?) of Ichthyosaurus were found, with the neotype placed in the most derived group, Species #4 (comprising eight individuals). The three autapomorphies of the group include the highest dorsoventral point of the maxilla positioned posterior to the external nares (Char. 3: 0); the presence of a circular depression on the articular surface of the humerus, anterior to the dorsal process (Char. 28: 0); and reduced glenoid contribution of the scapula markedly smaller than the coracoid facet (Char. 67: 1); although the latter could not be used with confidence (see Section 7.2.4B). The neotype displayed the first and last character but could not be assessed for Character 28 because of the way it is preserved.

Perhaps most alarming for the neotype was the identification of several features found to be unique and not found in any of the studied Ichthyosaurus specimens. First, the neotype possesses five digits in the hindfin, whereas as all other specimens possess four digits (where coded in ten specimens). Secondly, the neotype does not have notching in the hindfin, yet all eleven Somerset specimens which preserved the hindfin had hindfin notching. Lastly, a longer ilium than femur was found in the neotype, whereas almost all Somerset specimens, where the character state could be recorded, possessed an ilium and femur that were either equal in size (eight specimens) or the femur was longer than the ilium (three specimens). It should be noted that I. breviceps has a longer ilium than femur, but this is a lot longer (and wider) than seen in the neotype of I. communis (Massare and Lomax, 2014a). Some of the trees generated from the analysis positioned the neotype towards the primitive end of the tree, isolated.

8.3. A cladistic analysis without the neotype: In order to assess the importance of unique features of the neotype, a second analysis was run without the inclusion of the neotype (Figure 34). A total of 31 equally parsimonious trees were retained. As the majority of trees displayed similar clustering, and few trees showed one continual branch, the best representative tree was selected. The tree displayed four species, Species #1 (NMW 93.5G.2), Species #2 (CAMSM J35183) and two clusters of specimens. The problematic specimen referred to Species #5? from the primary cladistic analysis was included in one of the clusters. Compared with the results found in the primary analysis, some specimens remained in the same cluster, but others switched position. Nevertheless, the key feature to separate the two clusters of specimens was Character 46 (humerus/femur size ratio). The presence of a depression in the humerus (Char. 28: 0) was found in all specimens in Cluster 1 (where coded), but a single specimen in Cluster 2 also had a depression. There were also some distinctions between the ilium/pubis relative length (Char. 77) and the pubis/femur size (Char. 82). The results of this analysis further demonstrate that both Species #3 (Cluster 1) and Species #4 (Cluster 2) from the primary analysis are poorly supported. It also raises additional questions regarding the presence or absence of I. communis at Somerset.

Figure 34. A tree (cladogram) selected from an additional cladistic analysis undertaken with the sampled specimens, with the removal of the neotype of I. communis.

8.4. Ichthyosaurus implications: There are two primary outcomes of this study, which have wider implications for the genus. Firstly, as demonstrated above, a problem surrounds the inclusion of the neotype of I. communis in Species #4 (Cluster 2) and whether this group is a good representative of that species. The combination of unique features in the neotype not found in other specimens and a lack of features missing in the neotype, but present in other specimens in the same cluster, introduces some doubt regarding the presence of I. communis at Somerset. This may suggest either that the characters are not taxonomically useful, which would render Species #4 even more poorly supported; both Species #3 and Species #4 may be the same species, but display sexual dimorphism or intraspecific variation; or that I. communis is either exceptionally rare or absent at Somerset. Secondly, as McGowan (1974) and McGowan and Motani (2003) only recognized I. communis in Somerset, they disregarded the apparent contemporary species, I. intermedius (a taxonomic history regarding Ichthyosaurus is discussed in Section 6). As the holotype of I. intermedius is missing, was never fully described, and only brief descriptions were ever made of supposed subsequent specimens (e.g. Conybeare, 1822; Owen, 1840 and Lydekker, 1889), it is difficult to justify I. intermedius as a distinct species in this study. Based on the analysis, Species #2 has an autapomorphy that was defined upon tooth morphology, but the morphology is different to that described for I. intermedius. Nevertheless, this does suggest that tooth morphology may yet perhaps prove its taxonomic usefulness (Figure 35). The findings of this study suggest that I. intermedius should still be considered invalid, but also that perhaps I. communis may not be present at Somerset.

Figure 35. Differences in tooth morphology between some of the studied samples: A. CAMSM J35183. B. AGC 12. C. ANSP 15766. D. BRSMG Cb 4997. A and D are inverted so that the orientation is the same for all specimens featured.

9. FUTURE RESEARCH

In order to understand better and to confirm the discovered species in this study, further research is required. Although both Species #1 and Species #2 (especially the former) appear to represent distinct species, Species #3 and Species #4 require further detailed study. With the addition of extra specimens from historical collections, such as those currently inaccessible at the NHMUK (displayed behind glass and/or displayed high) and those at the NMW (which require detailed examination and conservation in order to confirm Somerset origin), it should be possible to gain a better understanding of both Species #3 and Species #4. Examining the neotype of I. communis (without the glass) is vital and will no doubt assist in answering some of the questions regarding important characters in this study. The addition of more complete specimens should help to group them better and to unite them with stronger taxonomic diagnoses.

Some of the characters that may have implications and be of taxonomic importance in the analysis (e.g. pubis distal shape, forefin notching etc.) were variable throughout the results. One interesting point regarding the exceptionally wide (fan- shaped) distal portion of the pubis (Figure 36) is that this morphology has not been seen in any Dorset Ichthyosaurus remains and is only present in the Somerset specimens (pers. obs.). At Dorset, forefin notching has been used as a taxonomic character and is used to distinguish I. conybeari (McGowan and Motani, 2003), but in the Somerset specimens forefin notching has no taxonomic usefulness. It should also be noted that some individual specimens in both Species #3 and Species #4 displayed unusual features (Figure 37). Further study of such features may hold additional information for those individuals (perhaps individual variation, ontogeny, sexual dimorphism, pathology etc.) and their respective species.

Figure 36. An example of the wide (fan-shaped) distal portion of the pubis found in some Somerset specimens. Pictured here is the pubis (note the arrow), ischium and hindfin of ANSP 17429, a specimen found in Species #4 (Cluster 2). Scale bar = 5 cm.

Figure 37. An example of one of several oddities found in the analysis, whereby an individual specimen possessed a different character state but which was not used taxonomically (by the analysis). Here is the right forefin (in ventral view) of OUMNH J13799; note the high contact point of the posterior accessory and the irregular shape of both the radius and ulna. Scale bar = 10 cm.

In addition to the more complete examples, lots of fragmentary specimens from Somerset were recognized, but not included in this study (e.g. isolated skulls, paddles and other postcrania). It is possible that a better understanding and definition of the species discovered in this study may allow for the referral of fragmentary (and isolated) remains to one or other of the species. Significantly, almost all of the sampled specimens are from historic collections (19th century) and were collected from quarries that are no longer accessible or simply do not exist. One issue regarding many of the historic specimens was the lack of specific horizon, age and detailed locality information. These specimens were not as well documented as those collected more recently. For this reason, the collection of new material from coastal exposures in Somerset is extremely important and could possibly help to enable further interpretation of the historical specimens.

Another potential option to help distinguish the species may be to decipher and recognize other characters. Of the 87 characters used, 44 of these were observed to vary between the studied specimens and were included in the matrix. Some of the 87 characters did not separate any specimens or only did so at the genus-level, such that a slightly revised character list with the addition of new characters may help further distinguish the species in this study. Another potential problem that affected the results was the preservation and orientation of specimens. The majority were preserved lying on their side largely exposed in ventral view. If all specimens were displayed in the same view and with the same level of preservation, this would argue for a more confident and reliable interpretation. For this reason, some of the characters used in the analysis were slightly biased based upon the completeness and or orientation of certain elements.

10. SUMMARY AND CONCLUSION

The main objective of this study is to review ‘Ichthyosaurus’ specimens from Somerset that have not been examined in light of phylogenetic analyses and to decipher their taxonomic diversity. Reviewing the study specimens and performing a cladistic analysis on the more complete examples has resulted in the discovery of additional species of Ichthyosaurus. It has been found that the taxonomic diversity of Ichthyosaurus species at Somerset is significantly greater than was previously suggested (McGowan, 1974; McGowan and Motani, 2003). This research recognizes that, at the very least, two of the species (Species #1 and Species #2) are well-supported, but that three of them are not (Species #3, Species #4 and Species #5?). Species #1 possesses autapomorphies of the skull and pelvic girdle; the latter includes features that are unique and have not been seen in Ichthyosaurus before. Species #2 has an autapomorphy in the skull and possibly the tooth morphology; the latter may perhaps help determine that tooth morphology is distinct in at least one species. Species #3 possesses a combination of characters that are not found in any of the other species. Species #4 has one autapomorphy in the skull and possibly one in the humerus. Considering that Species #5? is very poorly supported (and disregarded as something distinct), and that Species #4 may or may not be I. communis, it would appear that four species may be present in Somerset. With the recognition of possibly three new Somerset species (Species #1, #2 and #3), added to those found in Dorset, it suggests that at least seven species of Ichthyosaurus are known, three from Somerset and four from Dorset.

11. REFERENCES

Ambrose, K. 2001. The lithostratigraphy of the Blue Lias Formation (Late Rhaetian– Early Sinemurian) in the southern part of the English Midlands. Proceedings of the Geologists’ Association, 112, 97–110.

Anon [Wallis, F. S.], 1948. ‘Street Museum: temporary storage and closure’, Museums Journal, 48, 39.

Appleby, R. M. 1979. The affinities of Liassic and later ichthyosaurs. Palaeontology, 22, 921–946.

Arkell, W. J. 1933. The Jurassic System in Great Britain. Clarendon Press, Oxford. p. 681.

Benson, R. B. J., Bates, K. T., Johnson, M. R. and Withers, P. J. 2011. Cranial anatomy of Thalassiodracon hawkinsii (Reptilia, Plesiosauria) from the Early Jurassic of Somerset, United Kingdom. Journal of Vertebrate Paleontology, 31, 562– 574.

Benson, R. J. B., Evans, M. and Druckenmiller, P. S. 2012. High Diversity, Low Disparity and Small Body Size in Plesiosaurs (Reptilia, Sauropterygia) from the Triassic–Jurassic Boundary. PLoS ONE, 7, e31838. doi:10.1371/journal.pone.0031838

Benton, M. J. and Taylor, M. A. 1984. Marine reptiles from the Upper Lias (Lower Toarcian, Lower Jurassic) of the Yorkshire coast. Proceedings of the Yorkshire Geological Society, 44, 399–429.

Buchholtz, E. A. 2001 Swimming styles in Jurassic ichthyosaurs. Journal of Vertebrate Paleontology, 21, 61-73.

Caine, H. and Benton, M. J. 2011. Ichthyosauria from the Upper Lias of Strawberry Bank, England. Palaeontology, 54, 1069–1093.

Callaway, J. M. 1989. Systematics, phylogeny, and ancestry of Triassic ichthyosaurs. Unpublished Ph.D. dissertation, University of Rochester, Rochester, pp. 204.

Conybeare, W. D. 1822. Additional notices on the fossil genera Ichthyosaurus and Plesiosaurus. Transactions of the Geological Society of London, 1, 103–123.

Deeming, D. C., Halstead, L. B., Manabe, M. and Unwin. D. M. 1993. An ichthyosaur embryo from the Lower Lias (Jurassic: Hettangian) of Somerset, England, with comments on the reproductive biology of ichthyosaurs. Modern Geology, 18, 423- 442.

De la Beche, H. T. and Conybeare, W. D. 1821. Notice of the discovery of a new fossil animal, forming a link between the Ichthyosaurus and crocodile, together with general remarks on the osteology of the Ichthyosaurus. Transactions of the Geological Society of London, 5, 559–594.

Delair, J. B. 1969. A history of the early discoveries of Liassic ichthyosaurs in Dorset and Somerset (1779-1835). Proceedings of the Dorset Natural History and Archaeological Society, 90, 1–9.

Delair, 1978. The Clark (Street) Exhibition. Unpublished, detailed notes regarding the Alfred Gillett collection of ichthyosaurs. Stored at the AGC, Gillet Collection, Street, Somerset.

Druckenmiller, P. S. and Maxwell, E. E. 2010. A new Lower Cretaceous (lower Albian) ichthyosaur from the Clearwater Formation, Alberta, Canada. Canadian Journal of Earth Sciences, 47, 1037–1053.

Fernández, M. S., Archuby, F., Talevi, M. and Ebner, R. 2005. Ichthyosaurian eyes: paleobiological information content in the sclerotic ring of Caypullisaurus (Ichthyosauria, Ophthalmosauria). Journal of Vertebrate Paleontology, 25, 330–337.

Fernández, M. 2007. Redescription and phylogenetic position of Caypullisaurus (Ichthyosauria: Ophthalmosauridae). Journal of Paleontology, 81, 368–375.

Fischer, V., Masure, E., Arkhangelsky, M.S. and Godefroit, P. 2011. A new Barremian (Early Cretaceous) ichthyosaur from western Russia. Journal of Vertebrate Paleontology, 31, 1010–1025.

Fischer, V., Maisch, M. W., Naish, D., Liston, J., Kosma, R., Joger, U., Krüger, F. J., Pardo-Pérez, J., Tainsh, J., and Appleby, R. M. 2012. New ophthalmosaurid ichthyosaurs from the European Lower Cretaceous demonstrate extensive ichthyosaur survival across the Jurassic–Cretaceous boundary. PLoS ONE, 7, 1, e29234.

Fischer. V., Appleby. R. M., Naish, D., Liston. J., Riding. J. B., Brindley. S., and Godefroit. P. 2013. A basal thunnosaurian from Iraq reveals disparate phylogenetic origins for Cretaceous ichthyosaurs. Biology Letters, 9, 20130021. http://dx.doi.org/10.1098/rsbl.2013.0021.

Fischer, V., Arkhangelsky, M. S., Uspensky, G. N., Stenshin, I. M. and Godefroit, P. 2014a. A new Lower Cretaceous ichthyosaur from Russia reveals skull shape conservatism within Ophthalmosaurinae. Geological Magazine, 151, 60–70.

Fischer. V., Bardet. N., Guiomar. M., and Godefroit. P. 2014b. High diversity in Cretaceous ichthyosaurs from Europe prior to their extinction. Plos One, 9, e84709. http://doi:10.1371/journal.pone.0084709.

Fischer, V., Arkhangelsky, M. S., Naish, D., Stenshin, I. M., Uspensky, G. N. and Godefroit, P. 2014c. Simbirskiasaurus and Pervushovisaurus reassessed: implications for the taxonomy and cranial osteology of Cretaceous platypterygiine ichthyosaurs. Zoological Journal of the Linnean Society, 171, 822–841.

Goloboff, P. A., Farris, J. S. and Nixon, K. C. 2008. TNT, a free program for phylogenetic analysis. Cladistics, 24, 774–786.

Green, J. P. and Lomax, D. R. 2014. An ichthyosaur (Reptilia: Ichthyosauria) specimen from the Lower Cretaceous (Berriasian) Spilsby Sandstone Formation of Nettleton, Lincolnshire, UK. Proceedings of the Geologists’ Association, 125, 432– 436.

Hardy, P. 1999. The Geology of Somerset. Ex Libris Press, Bradford-on-Avon. pp. 223.

Haslett, S. K. 2010. Somerset Landscapes: Geology and Landforms. Blackbarn Books. p. 184.

Hawkins, T. H. 1834, Memoirs on Ichthyosauri and Plesiosauri; Extinct monsters of the ancient Earth. Relfe and Fletcher, London. p. 58.

Home, E. 1814. Some account of the fossil remains of an animal more nearly allied to fishes than any other classes of animals. Philosophical Transactions of the Royal Society of London, 101, 571–577.

Home, E. 1819. An account of the fossil skeleton of the Proteo-Saurus. Philosophical Transactions of the Royal Society of London, 109, 209–211.

Howe, S. R., Sharpe, T. and Torrens, H. S. 1981. Ichthyosaurs: a history of fossil 'sea dragons'. National Museum of Wales, Cardiff. p. 32.

Knight, F. A. and Dutton, L. M. 1909. Somerset. Cambridge University Press, Cambridge. p. 206.

König, C. D. E. 1818. Synopsis of the contents of the British Museum. London.

Kirton, A. M. 1983. A review of British Upper Jurassic ichthyosaurs. Unpublished Ph.D. Thesis, University of Newcastle-upon-Tyne, Newcastle-upon-Tyne. pp. 239.

Lomax, D. R. 2010. An Ichthyosaurus (Reptilia, Ichthyosauria) with gastric contents from Charmouth, England: first report of the genus from the Pliensbachian. Paludicola, 8, 22–36.

Lomax, D. R. and Massare, J. A. 2012. The first Leptonectes (Reptilia: Ichthyosauria) with associated embryos, from Somerset, England. Paludicola, 8, 263–276.

Lomax, D. R. and Massare, J. A. 2015. A new species of Ichthyosaurus from the Lower Jurassic of West Dorset, England. Journal of Vertebrate Paleontology. DOI: 10.1080/02724634.2014. 903260.

Lydekker, R. 1888. Note on the classification of the Ichthyopterygia with a notice of two new species. Geological Magazine, 5, 309–314.

Lydekker, R. 1889. Catalogue of the fossil reptilia and amphibia in the British Museum (Natural History). Part II. The orders Ichthyopterygia and Sauropterygia. Printed by order of the trustees, Taylor and Francis, London. p. 307.

Lydekker, R. 1891. Note on a nearly perfect skeleton of Ichthyosaurus tenuirostris from the Lower Lias of Street, Somerset. Geological Magazine, 290–292. Note, there is an error in the original numbering of this article.

Maisch, M. W. 1997. The cranial osteology of Ichthyosaurus intermedius CONYBEARE, 1833 from the Lias of Great Britain. Stuttgarter Beiträge zur Naturkunde, Serie B (Geologie und Paläontologie), 258, 1–26.

Maisch, M. W. 1998. Short review of the ichthyosaurs of the Posidonienschiefer with remarks on the taxonomy of the Stenopterygidae and Temnodontosauridae. Neues Jahrbuch für Geologie und Paläontologie, Abhandlungen, 209, 401–431.

Maisch, M. W. and Matzke, A. T. 2000. The Ichthyosauria. Stuttgarter Beiträge zur Naturkunde Serie B (Geologie und Paläontologie), 298, 1–159.

Maisch, M. W. and Hungerbȕhler, A. 2001. New evidence for a discrete supratemporal bone in the Jurassic ichthyosaur Temnodontosaurus. Historical Biology, 11, 335–345.

Martill, D. M. 1991. Marine Reptiles. In Martill D. M. and Hudson A. D. (eds.), Field Guide to Fossils Series: Fossils of the Oxford Clay. The Palaeontological Association. 226–244.

Martill, D. M., Motani, R. and Taylor, M. A. 2009. Alfred Gillett and fossils from Street. Unpublished handout sheet for the Palaeontological Association Review Seminar.

Massare, J. A. 1987. Tooth morphology and prey preference of Mesozoic marine reptiles. Journal of Vertebrate Paleontology, 7, 121–137.

Massare, J. A. 1988. Swimming speeds in Mesozoic marine reptiles: implications for predation. Paleobiology, 14, 187–205.

Massare, J. A. and Young, H. A. 2005. Gastric contents of an ichthyosaur from the Sundance Formation (Jurassic) of Central Wyoming. Paludicola, 5, 20–27.

Massare, J. A. and Lomax. D. R. 2014a. An Ichthyosaurus breviceps collected by Mary Anning: New information on the species. Geological Magazine, 151, 21–28. (First published on line April, 2013).

Massare, J. A. and Lomax, D. R. 2014b. Recognising composite specimens of Jurassic ichthyosaurs in historical collections. The Geological Curator, 10, 9–15.

Maxwell, E. 2010. Generic reassignment of an ichthyosaur from the Queen Elizabeth Islands, Northwest Territories, Canada. Journal of Vertebrate Paleontology, 30, 403– 415.

Maxwell, E., Fernández, M. S. and R. R. Schoch. 2012. First diagnostic marine reptile remains from the Aalenian (Middle Jurassic): A new ichthyosaur from southwestern Germany. PLoS One, 7, 1–13.

McGarvie, M. 1987. The book of Street. A history from the earliest times to 1925. Barracuda Books, Buckingham.

McGowan, C. 1974. A revision of the latipinnate ichthyosaurs of the Lower Jurassic of England (Reptilia: Ichthyosauria). Life Science Contributions of the Royal Ontario Museum, 100, 1–30.

McGowan, C. 1976. The description and phonetic relationships of a new ichthyosaur genus from the Upper Jurassic of England. Canadian Journal of Earth Sciences, 13, 668–683.

McGowan, C. 1986. A putative ancestor for the swordfish-like ichthyosaur Eurhinosaurus. Nature, 322, 454–456.

McGowan, C. 1989. Leptopterygius tenuirostris and other long-snouted ichthyosaurs from the English Lower Lias. Palaeontology, 32, 409–427.

McGowan, C. 1990. Problematic ichthyosaurs from southwest England: a question of authenticity. Journal of Vertebrate Paleontology, 10, 72–79.

McGowan, C. 1994. Temnodontosaurus risor is a juvenile of T. platyodon (Reptilia: Ichthyosauria). Journal of Vertebrate Paleontology, 14, 472–479.

McGowan, C. 1996a. A new and typically Jurassic ichthyosaur from the Upper Triassic of British Columbia. Canadian Journal of Earth Science, 33, 24–32.

McGowan, C. 1996b. Giant ichthyosaurs of the Early Jurassic. Canadian Journal of Earth Sciences, 33, 1011–1021.

McGowan, C. 2003. A new specimen of Excalibosaurus from the English Lower Jurassic. Journal of Vertebrate Paleontology, 23, 950–956.

McGowan, C. and Motani, R. 2003. Handbook of Paleoherpetology, Part 8 Ichthyopterygia. Verlag Dr. Friedrich Pfeil, Munich. p. 175.

Milner, A. C. and Walsh, S. 2010. Reptiles. In Lord A. R. and Davis P. G. (eds.), Field Guide to Fossils Series: Fossils from the Lower Lias of the Dorset Coast. The Palaeontological Association. 372–394.

Motani, R. 1999a. On the evolution and homologies of ichthyopterygian forefins. Journal of Vertebrate Paleontology, 19, 28–41.

Motani, R. 1999b. Phylogeny of the Ichthyopterygia. Journal of Vertebrate Paleontology, 19, 473–496.

Motani, R. 2005a. Evolution of fish-shaped reptiles (Reptilia: Ichthyopterygia) in their physical environments and constraints. Annual Review of Earth and Planetary Sciences, 33, 395–420.

Motani, R. 2005b. True skull roof configuration of Ichthyosaurus and Stenopterygius and its implications. Journal of Vertebrate Paleontology, 35, 338–342.

Motani, R. 2009. The evolution of marine reptiles. Evolution: Education and Outreach, 2, 224–235.

Owen, R. 1840. Report on British fossil reptiles. Report of the 9th Meeting of the British Association for the Advancement of Science, 8, 43–126.

Owen, R. 1881. A monograph of the fossil reptilia of the Liassic Formations. Part Ill. Palaeontographical Society, London. (Ichthyopterygia, pp. 83-134).

Pearce, J. C. 1846. Notice of what appears to be the embryo of an Ichthyosaurus in the pelvic cavity of Ichthyosaurus (communis?). Annals and Magazine of Natural History (First Series), 17, 44–46.

Pollard, J. E. 1968. The gastric contents of an ichthyosaur from the Lower Lias of Lyme Regis. Palaeontology, 11, 376–388.

Sander, P. M. 2000. Ichthyosauria: their diversity, distribution, and phylogeny. Paläontologische Zeitschrift, 74, 1–35.

Stinnesbeck, W., Frey, E., Rivas, L., Pérez, J. P., Cartes, M. L., Soto, C. S. and Lobos, P. Z. 2014. A Lower Cretaceous ichthyosaur graveyard in deep marine slope channel deposits at Torres del Paine National Park, southern Chile. Geological Society of America Bulletin, 126, B30964.1.

Storrs, G. W. and Taylor, M. A. 1996. Cranial anatomy of a plesiosaur genus from the Lowermost Lias (Rhaetian/Hettangian) of Street, Somerset, England. Journal of Vertebrate Paleontology, 16, 403–420.

Taylor, M. A. 1989. Thomas Hawkins FGS. The Geological Curator, 5, 112-114.

Torrens, H. S. 1995. Mary Anning (1799–1847) of Lyme; ‘the greatest fossilist the world ever knew’. British Journal for the History of Science, 28, 257–84.

Zammit, M. 2012. Cretaceous ichthyosaurs: dwindling diversity, or the Empire strikes back? Geosciences, 2, 11–24.

12. APPENDICES

Appendix A: List of characters for cladistic analysis

The following list of characters comprises a combination of those used by Motani (1999b); Maxwell et al. (2012); Fernández (2007); Druckenmiller and Maxwell (2010); Maxwell (2010); Sander (2000); Maisch and Matzke (2000); Callaway (1989), Fischer et al. (2011; 2012; 2013). Forty four characters that varied among the study samples are highlighted in italics and were identified as part of this study. There is a corresponding figure legend to help interpret some of the included characters they appear below as Figure A1, A2 and so forth. Comments regarding the usability (or issue) of a certain character (where required) are in bold and italics.

Skull 0. Premaxilla: (0) supra- and sub-narial processes absent; (1) well-developed supranarial process overlying approximately 1/3–1/2 of external nares; (2) supranarial process short relative to subnarial process; (3) both processes short relative to external narial length; (4) both processes long, extending half way into the external naris (modified from Maxwell et al., 2012: character 1).

1. Premaxilla/lacrimal contact below the external nares: (0) absent, (1) narrow, (2) broad (Fernández, 2007: character 0).

2. Anterior process of maxilla in lateral aspect: (0) long and broad, (1) long and narrow, (2) relatively short and narrow, (3) long, delicate and slender. (modified from Fernández, 2007: character 1).

3. Highest point (dorsoventrally) of maxilla: (0) posterior to the external nares, (1) anterior to the external nares, (2) centred in the middle of the external nares (modified from Druckenmiller and Maxwell, 2010: character 4).

4. Maxilla, jugal process: (0) long, extending as far under orbit as lacrimal in lateral view, (1) short, hidden from lateral view by jugal (Maxwell et al., 2012: character 3).

5. Maxilla/external naris contact: (0) absent, (1) present (Motani, 1999b: character 3).

6. Anterior margin of jugal: (0) tapering, running partly between the lacrimal and maxilla, (1) broad and fan-like, covering large area of maxilla ventrolaterally, (2) robust and short, not separating maxilla and lacrimal (modified from Druckenmiller and Maxwell, 2010: character 6).

7. Jugal shape: (0) triradiate, (1) lunate, or J-shaped (Motani, 1999b: character 24).

8. Jugal, dorsal ramus angle: (0) well-developed and strongly curved dorsally (almost 90 degree bend), (1) poorly developed, jugal essentially straight (modified from Maxwell et al., 2012: character 7).

9. Lacrimal: (0) excluded from contact with external nares by dorsal process of maxilla, (1) participates in external narial opening (Sander, 2000: character 10).

10. Overbite: (0) absent or slight (accounting for preservation), (1) present (Motani, 1999b: character 33).

11. External naris orientation: (0) dorso-lateral, (1) lateral, scarcely visible in dorsal view (Motani, 1999b: character 4). *Note this may be dependent on preservation.

12. Nasal/external naris contact: (0) absent, (1) present (Motani, 1999b: character 5).

13. Nasals, lateral view: (0) extend as far forward as the anterior process of the maxilla, (1) do not extend as far forward as the anterior process of maxilla.

14. Prefrontal–external narial contact: (0) absent, (1) present (Druckenmiller and Maxwell, 2010: character 7).

15. Postorbital: (0) broad, (1) narrow (Fernández, 2007: character 7).

16. Postorbital shape: (0) triradiate, (1) lunate and small, (2) lunate, large and robust (modified from Motani, 1999b: character 11).

17. Supratemporal-postorbital contact: (0) absent, (1) present (Sander, 2000: character 27).

18. Parietal, dorsal view: (0) with no contribution to anterior edge of UTF, postfrontal excluded by frontal, (1) with minor contribution to anterior edge of UTF, (2) large contribution, almost completely excluding postfrontal from contact with anterior edge of UTF (Maxwell et al., 2012: character 14).

19. Surangular, lateral view: (0) anteriormost portion of surangular seen to extend as far forward as maxilla, (1) anteriormost portion does not extend as far forward as the maxilla.

20. Angular lateral exposure: (0) much smaller than surangular exposure, (1) extensive, almost completely concealing surangular, (2) equal to surangular exposure (modified from Fischer et al., 2013, modified from Motani, 1999b: character 32).

Dentition 21. Crown striations: (0) striations present, (1) crown enamel subtly ridged or smooth, (2) striations absent or subtle (Druckenmiller and Maxwell, 2010: character 25).

22. Root striations: (0) present, (1) absent or subtle (Fischer et al., 2013: character 4).

23. Tooth robusticity: (0) teeth robust (especially roots) and moderately large, (1) teeth moderately slender, (2) teeth mostly very slender, small and delicate.

Forefin

24. Distal and proximal ends of the humerus: (0) nearly equal in width, (1) distal end noticeably wider than proximal end, (2) proximal end noticeably wider than distal end (modified from Fernández, 2007: character 12).

25. Deltopectoral crest of the humerus: (0) present and large, (1) present and small, (2) extremely reduced (modified from Maxwell, 2010: character 20).

26. Humerus size: (0) much longer than wide, (1) nearly squarish (modified from Callaway, 1989: character 28).

27. Humerus distal articular facets: (0) radius facet larger than ulna facet, (1) ulna facet larger than radius facet, (2) two facets nearly equal (modified from Motani, 1999b: character 52). *Note that rotation of the humerus will potentially affect the outcome.

28. Humerus circular depression on the articular surface anterior to the dorsal process of humerus: (0) present, (1) absent.

29. Humerus dorso-ventral constriction in humerus head: (0) present, (1) absent.

30. Humerus shape: (0) proximal portion larger and more robust compared to distal (more delicate), (1) distal portion wider and more robust compared to proximal, (2) overall robustness of proximal and distal portion equal. *Note the orientation of the humerus, and how it is preserved, effects the decision.

31. Humerus radius-ulna distal contact (distalmost point of humerus (in dorsal/ventral view) between the radius and ulna): (0) pointed, (1) rounded.

Figure A1. Illustration of character 31. Distalmost point of humerus between the radius and ulna: A. Pointed. B. Rounded. Note the arrow. The humeri are depicted as the right in dorsal view.

32. Radius L/W ratio: (0) longer than wide, (1) wider than long, (2) roughly equal (modified from Motani, 1999b: character 61).

33. Ulna L/W ratio: (0) longer than wide, (1) wider than long, (2) roughly equal.

34. Radius/ulna relative overall size: (0) practically equal, (1) radius larger than ulna, (2) ulna larger than radius (modified from Motani, 1999b: character 64).

35. Contact between radius and ulna: (0) broad, (1) narrow.

36. Intermedium L/W ratio: (0) longer than wide (rounded hexagon), (1) wider than long (elongated hexagon), (2) roughly equal.

37. Ulnare/intermedium overall relative size: (0) ulnare noticeably larger than intermedium, (1) intermedium larger than ulnare, (2) roughly equal (modified from Motani, 1999b: character 66).

38. Radiale/intermedium width: (0) radiale noticeably wider than intermedium, (1) intermedium wider than radiale, (2) roughly equal.

39. Metacarpals: (0) rounded, not forming tightly interlocking mosaic, (1) polygonal, forming a tightly interlocking mosaic (Maxwell et al., 2012: character 46).

40. Phalanges shape: (0) mostly rounded, (1) mostly polygonal, (2) combination of both rounded and rectangular (from Fischer et al., 2011, modified from Maisch and Matzke, 2000: character 102). *Note that in general the proximal elements are almost always polygonal and the distal elements become more rounded. However, there are some differences. Though, this character also depends on the completeness of the fin.

41. Digital bifurcation row: (0) no bifurcation, (1) occurs in metacarpal row, (2) occurs in p1 or p2 row, (3) occurs in more distal row.

42. Extent of notching: (0) absent, (1) radius and other elements notched, (2) radius not notched, but other elements are.

43. Anterior accessory digit: (0) absent, (1) present.

44. Posterior accessory digit: (0) absent, (1) present.

45. Posterior accessory digit initial point of contact: (0) ulna, (1) ulnare, (2) metacarpal, (3) phalangeal level.

46. Humerus/femur ratio (humerus length divided by femur length): (0) humerus 1.3 times or less as large as femur, (1) humerus 1.4 times as large as femur, (2) humerus 1.5 times as large as femur.

Hindfin

47. Distal and proximal ends of the femur: (0) nearly equal in width, (1) distal end wider than proximal end.

Figure A2. Illustration of character 47. Distal and proximal ends of the femur: A. nearly equal in width. B. Distal end wider than proximal end. The femora are depicted as the right in dorsal view.

100

48. Anterior portion of distal end of femur: (0) curved (almost right angle), (1) slight ‘flared’ curve, (2) widely (‘flared’) curve.

Figure A3. Illustration of character 48. Anterior portion of distal end of femur: A. Curved (almost right angle). B. Slightly ‘flared’ curve. C. Widely ‘flared’ curve. The femora are depicted as the right in ventral view.

49. Posteriormost portion of distal end of femur: (0) right angle, (1) gentle ‘flared’ curve, (2) widely ‘flared’ curve.

Figure A4. Illustration of character 49. Posterior portion of distal end of femur: A. Right angle. B. Gentle ‘flared’ curve. C. Widely ‘flared’ curve. The femora are depicted as the right in dorsal view.

101

50. Number of digits in hindfin: (0) three, (1) four, (2) five, (3) six.

51. Tibia L/W ratio: (0) longer than wide, (1) wider than long (Motani, 1999b: character 90).

52. Tibia shape: (0) pentagonal, (1) polygonal, (2) circular. *Note the orientation of the femur makes this difficult to assess on some specimens, especially if only a single hindfin is present, i.e. it is near-impossible to identify left/right.

53. Tibia length: (0) approximately twice as proximodistally long than more distal elements, (1) tibia only slightly longer than successive elements (from Maxwell et al., 2012, modified from Motani, 1999b: character 90).

54. Fibula L/W ratio: (0) longer than wide, (1) wider than long, (2) roughly equal.

55. Fibula shape: (0) pentagonal, (1) polygonal/rectangular, (2) circular.

56. Tibia and fibula – relative size: (0) tibia larger than fibula, (1) approximately equal in size, (2) tibia anterioposteriorly and proximodistally smaller than fibula, (3) fibula proximodistally larger than tibia, but anteroposteriorly roughly equal, (4) fibula both as proximodistally longer and anteroposteriorly wider (modified from Maxwell et al., 2012: character 56).

57. Intermedium L/W ratio: (0) longer than wide, (1) wider than long, (2) roughly equal.

58. Intermedium shape: (0) pentagonal, (1) hexagonal, (2) rectangular, (3) rounded.

59. Intermedium and tibiale – relative size: (0) intermedium larger than tibiale, (1) approximately equal in size, (2) tibiale larger than intermedium.

102

60. Shape of phalanges: (0) mostly rounded, (1) mostly polygonal, (2) combination of rounded and rectangular phalanges. *Note that in general the proximal elements are almost always polygonal and the distal elements become more rounded. However, there are some differences. Although this character also depends on the completeness of the fin.

61. Digital bifurcation: (0) absent, (1) present.

62. Anterior accessory digit: (0) absent, (1) present.

63. Posterior accessory digit: (0) absent, (1) present.

64. Notching: (0) absent, (1) tibia and other elements notched, (2) tibia not notched, but other elements notched.

Pectoral girdle

65. Scapula: (0) with prominent acromion process, (1) large acromion process absent, but anterior expansion of proximal scapula present, (2) proximal expansion absent, scapula strap-shaped (Maxwell et al., 2012: character 31).

66. Scapula overall robustness: (0) very robust, (1) slightly robust (more delicate) and more elongated.

Figure A5. Illustration of character 66. Scapula overall robustness: A. Very robust. B. Slightly robust (more delicate) and more elongated. The scapulae are depicted as the right in dorsolateral view.

103

67. Glenoid contribution of the scapula: (0) extensive, being at least as large as the coracoid facet, (1) reduced, being markedly smaller than the coracoid facet (Fischer et al., 2012: character 27).

68. Coracoid: (0) mediolaterally wider than anteroposteriorly long, (1) approximately equidimensional, (2) anteroposteriorly longer than mediolaterally wide (Maxwell et al., 2012: character 32).

69. Coracoid: (0) with well-developed anterior notch, (1) anterior notch reduced (small), (2) anterior notch absent (modified from Maxwell et al., 2012: character 33).

70. Coracoid anterior notch shape: (0) J-shaped, (1) closed C-shaped, (2) wide C- shaped.

Figure A6. Illustration of character 70. Coracoid anterior notch shape: A. J-shaped. B. Closed C-shaped. C. Wide C-shaped. The coracoids are depicted as the right in ventral view. Note: The open notch is the anterior notch.

104

71. Coracoid posterior shape: (0) J-shaped, (1) closed C-shaped, (2) wide C- shaped, (3) absent.

Figure A7. Illustration of character 71. Coracoid posterior shape: A. J-shaped. B. Closed C-shaped. C. Wide C-shaped. D. Absent. The coracoids are depicted as the right in ventral view. Note: The open notch is the posterior notch. In ‘D’ the notch is absent.

105

Pelvic girdle

72. Ilium: (0) with expanded proximal region, (1) narrow proximally and distally, rib-like (from Maxwell et al., 2012, modified from Sander 2000: character 106).

73. Pubis shape, distally: (0) not expanded, (1) expanded ‘spatulate shape’, (2) widely expanded and asymmetric (‘fan-shaped’).

Figure A8. Illustration of character 73. Pubis shape, distally: A. Not expanded. B. Expanded ‘spatulate-shaped’. C. Widely expanded and asymmetric (‘fan-shaped’). The pubes are depicted as the right in dorsal view.

74. Ischium overall shape: (0) long and slender (especially in centre), (1) long, fairly robust and oblong-shaped, (2) short, robust and wide (modified from Motani, 1999b: character 87).

106

75. Ischium shape, distally: (0) not expanded, essentially straight, (1) marginally expanded, (2) widely expanded.

Figure A9. Illustration of character 75. Ischium shape, distally: A. Not expanded, essentially straight. B. Marginally expanded. C. Widely expanded. The ischia are depicted as the right in dorsal view.

76. Ischium: (0) with greatly expanded medial edge, (1) ischium lacking medial expansion (Maxwell et al., 2012: character 53).

77. Ilium/pubis relative length: (0) nearly equal, (1) ilium longer than pubis, (2) pubis longer than ilium.

78. Ilium/ischium relative length: (0) nearly equal, (1) ilium longer than ischium, (2) ischium longer than ilium.

79. Pubis/ischium relative length: (0) nearly equal, (1) pubis longer than ischium, (2) ischium longer than pubis (modified from Motani, 1999b: character 86).

107

80. Ischium/pubis overall robustness: (0) pubis more robust, (1) ischium more robust, (2) roughly equal.

81. Ilium/femur size: (0) nearly equal, (1) ilium longer than femur, (2) femur longer than ilium.

82. Pubis/femur size: (0) nearly equal, (1) pubis longer than femur, (2) femur longer than pubis.

83. Ischium/femur size: (0) nearly equal, (1) ischium longer than femur, (2) femur longer than ischium.

Vertebral column

84. Mid-caudal centra height change: (0) gradual decrease, (1) sudden decrease, (2) increase, (3) sudden increase (modified from Motani, 1999b: character 98).

85. Anterior-middle dorsal neural spine: (0) narrow, relatively short and straight, (1) narrow, high, and straight, (2) abruptly high (modified from Callaway, 1989: character 16). Note this can be difficult to confirm as displaced ribs/vertebrae or matrix are often obscuring the neural spines.

86. Preflexural vertebral count: (0) less than 80, (1) greater than 80.

108

Appendix B: Interpretation of character maps with comments regarding the usefulness of characters used

Below are interpretations of each character map of the selected tree from the cladistic analysis. The potential significance of the character in the tree and its taxonomic usefulness are discussed. This is a strict interpretation of this tree and analysis; some of the features that are here deemed ‘useful’, ‘not useful’ or ‘rejected’ may result in a different outcome if combined with other characters, or if other characters are removed or added, or if more complete analyses is undertaken. Thus, the following should only be used as a preliminary analysis of the Somerset ichthyosaurs, based solely on the data matrix and tree. The outgroup taxa (Temnodontosaurus and Suevoleviathan) are only discussed when/if necessary. Please note, to help interpret some of the important (or unusual) character maps some have a corresponding figure legend and are figured below as Figure B1, B2 and so forth. The figures are colour coded to character states, i.e. state 0, 1, etc. They correspond directly to the character list, i.e. State 0 on the map is State 0 in the character.

As this information is taken directly from the character maps the corresponding specimen numbers are used in the following interpretations (e.g. NMW 93.5G.2) instead of the individual species allocation (i.e. Species #1). To ensure there is no confusion, the corresponding terminology relates directly as: NMW 93.5G.2 (Species #1); CAMSM J35183 (Species #2); Cluster 1 (Species #3); Cluster 2 (Species #4) and AGC 17 (Species #5?).

Skull Char 0: Premaxilla: (0) supra- and sub-narial processes absent, (1) well-developed supranarial process overlying approximately 1/3–1/2 of external nares, (2) supranarial process short relative to subnarial process, (3) both processes short relative to external narial length, (4) both processes long, extending half way into the external naris (modified from Maxwell et al., 2012: character 1). - NMW 93.5G.2 (2) = Autapomorphy - CAMSM J35183 (3) = Autapomorphy

109

- Clusters 1 & 2 (4) = Synapomorphy (although this is ambiguous as 9 specimens have a ‘?’) - This is a difficult character to define as several specimens from cluster 1 & 2 are missing. However, based upon the separation of the outgroup taxa and both NMW 93.5G.2 and CAMSM J35183, this is a good character to distinguish species.

Figure B1. The character map for Character 0. Note the autapomorphic differences for both NMW 93.5G.2 and CAMSM J35183 and the synapomorphy between clusters 1 & 2.

110

Char 1: Premaxilla/lacrimal contact below the external nares: (0) absent, (1) narrow, (2) broad (Fernández, 2007: character 0). - CAMSM J35183 (0) = Retained primitive character - NMW, cluster 1 & 2 (1) = Synapomorphy - This is a variable character (appearing in both cluster 1 & 2), but appears to display an unusual combination of characters in NMW 93.5G.2 (shared with cluster 1 & 2).

Figure B2. The character map for Character 1. Note the retained primitive character of CAMSM J35183 with Suevoleviathan.

111

Char 2: Anterior process of maxilla in lateral aspect: (0) long and broad, (1) long and narrow, (2) relatively short and narrow, (3) long, delicate and slender. (modified from Fernández, 2007: character 1). - NMW 93.5G.2 (3) = Autapomorphy - This character appears to be rather variable, found in cluster 1 & 2, and shared with the outgroup taxa. However, it is a distinguishable character for NMW 93.5G.2.

Figure B3. The character map for Character 2. Apart from the autapomorphy of NMW 93.5G.2 note the variation throughout the tree.

112

Char 3: Highest point (dorsoventrally) of maxilla: (0) posterior to the external nares, (1) anterior to the external nares, (2) centred in the middle of the external nares (modified from Druckenmiller and Maxwell, 2010: character 4). - All specimens where present bar Cluster 2 (2) = Retained primitive character - Cluster 2 (0) = Autapomorphy - There is a clear-cut distinction between specimens with this character and therefore it is determined to represent a good, taxonomic character.

Figure B4. The character map for Character 3. There is a clear difference with this character from the most derived group (Cluster 2) and all other primitive specimens.

Char 4: Maxilla, jugal process: (0) long, extending as far under orbit as lacrimal in lateral view, (1) short, hidden from lateral view by jugal (Maxwell et al., 2012: character 3). - NMW 93.5G.2 (1) = Retained primitive character - Shared with Suevoleviathan - Other specimens where present (0) = Synapomorphy - This character state suggests that it is variable among CAMSM J35183 and both cluster 1 & 2. However, the difference found in Suevoleviathan and NMW 93.5G.2 suggests its usefulness to distinguish at least one species.

113

Char 5: Maxilla/external naris contact: (0) absent, (1) present (Motani, 1999b: character 3). - NMW 93.5G.2 (0) = Synapomorphy - Unusual combination of character – shared with cluster 1 & 2 - CAMSM J35183 (1) = Retained primitive character - This character suggests that it is variable between cluster 1 & 2, but shared with NMW 93.5G.2 (unusual combination). Yet, in CAMSM J35183 it is shared with Temnodontosaurus, and is thus a potentially reliable character.

Figure B5. The character map for Character 5. Note the retained primitive character of CAMSM J35183 with Temnodontosaurus.

Char 6: Anterior margin of jugal: (0) tapering, running partly between the lacrimal and maxilla, (1) broad and fan-like, covering large area of maxilla ventrolaterally, (2) robust and short, not separating maxilla and lacrimal (modified from Druckenmiller and Maxwell, 2010: character 6). - NMW 93.5G.2 (0) = Retained primitive character - Clusters 1 & 2 (0) = Retained primitive character - This character is not considered to represent a useful character to differentiate between species, as only the single specimen (AGC 12) and Suevoleviathan share this. Thus, it is probably variable and is not a good character to separate potential species.

114

Char 7: Jugal shape: (0) triradiate, (1) lunate, or J-shaped (Motani, 1999b: character 24). - All specimens where present (1) - This is not a reliable character to distinguish among species, as all specimens (where coded) are the same.

Char 8: Jugal, dorsal ramus angle: (0) well-developed and strongly curved dorsally (almost 90 degree bend), (1) poorly developed, jugal essentially straight (modified from Maxwell et al., 2012: character 7). - All specimens where present bar two (0) = Synapomorphy - Temnodontosaurus & ANSP 15766 (1) = Retained primitive character - This is unusual as ANSP 15766 is the only specimen to retain this character (aside from Temnodontosaurus). Thus, it is not considered to be a reliable character.

Char 9: Lacrimal: (0) excluded from contact with external nares by dorsal process of maxilla, (1) participates in external narial opening (Sander, 2000: character 10). - All specimens where present bar one (1) = Retained primitive character - A good character to separate specimens from Suevoleviathan, but nothing else.

Char 10: Overbite: (0) absent or slight (accounting for preservation), (1) present (Motani, 1999b: character 33). - This is not a reliable character to distinguish among species as all specimens (where coded) are the same.

Char 11: External naris orientation: (0) dorso-lateral, (1) lateral, scarcely visible in dorsal view (Motani, 1999b: character 4). - All specimens where present (1) - This is not a reliable character to distinguish among species as all specimens (where coded) are the same.

Char 12: Nasal/external naris contact: (0) absent, (1) present (Motani, 1999b: character 5).

115

- Outgroup taxa & BRSMG Cb 3578 (0) = Retained primitive character - All other specimens where present (1) = Synapomorphy - The occurrence of this character in BRSMG Cb 3578 is unusual, although over half of the sampled specimens could not be coded for this character. Thus, it is possible that this character would be more prevalent. It is not considered to be a strong character.

Char 13: Nasals, lateral view: (0) extend as far forward as the anterior process of the maxilla, (1) do not extend as far forward as the anterior process of maxilla. - Temnodontosaurus (1) - All other specimens where present (0) = Synapomorphy - This is a good character for separating the distinct groups from Temnodontosaurus, but does not assist with any further grouping. The character is variable.

Char 14: Prefrontal–external narial contact: (0) absent, (1) present (Druckenmiller and Maxwell, 2010: character 7). - All specimens where present (0) - This is not a reliable character to distinguish among species as all specimens (where coded) are the same.

116

Char 15: Postorbital: (0) broad, (1) narrow (Fernández, 2007: character 7). - NMW 93.5G.2 (1) = Autapomorphy - This represents a good, strong character for NMW 93.5G.2 as a distinct species.

Figure B6. The character map for Character 15. Note the sole autapomorphy of NMW 93.5G.2.

117

Char 16: Postorbital shape: (0) triradiate, (1) lunate and small, (2) lunate, large and robust (modified from Motani, 1999b: character 11). - All specimens where present (1) = Retained primitive character - NHMUK R3372 (cluster 1) & AGC 11, MOS 8373, ANSP 15766 (cluster 2) (1) = Retained primitive character - CAMSM J59575, NHMUK 2013, AGC 12 (cluster 1) & BRSMG Cb 4997, CRSMG Ce 16611 (Cluster 2) (2) = Synapomorphy - This is an unusual character. Both characters (1 & 2) are present between the two clusters of specimens, which suggest this character is not good for a solid species distinction.

Figure B7. The character map for Character 16. This is one character that is variable throughout the individual specimens, but may have some usefulness.

Char 17: Supratemporal-postorbital contact: (0) absent, (1) present (Sander, 2000: character 27). - Temnodontosaurus, NMW 93.5G.2, NHMUK 2013 & AGC 12 (1) = Retained primitive character - Only a total of five specimens are coded for this character, thus it is not very reliable.

118

Char 18: Parietal, dorsal view: (0) with no contribution to anterior edge of UTF, postfrontal excluded by frontal, (1) with minor contribution to anterior edge of UTF, (2) large contribution, almost completely excluding postfrontal from contact with anterior edge of UTF (Maxwell et al., 2012: character 14). - All specimens where present (1) - This is another character that was not reliable for this study, as most specimens (only four of the Somerset samples) displayed this.

Char 19: Surangular, lateral view: (0) anteriormost portion of surangular seen to extend as far forward as maxilla, (1) anteriormost portion does not extend as far forward as the maxilla. - All specimens where present (1) - CAMSM J59575 (0) = ‘Autapomorphy’ - As only one specimen has been documented with this feature, it is probably an unusual occurrence and may be due to preservation. Thus, it is not considered to represent an autapomorphy of the specimen.

Char 20: Angular lateral exposure: (0) much smaller than surangular exposure, (1) extensive, almost completely concealing surangular, (2) equal to surangular exposure (modified from Fischer et al., 2013, modified from Motani, 1999b: character 32). - All specimens where present (0) = Retained primitive character - CAMSM J59575, AGC 12 (cluster 1) & BRSMG Cb 4997 (2) = Synapomorphy - Due to three of the sample specimens representing something different, and spread between the two clusters, this character is variable, and would have probably been seen in the other specimens that could not be coded.

Dentition Char 21: Crown striations: (0) striations present, (1) crown enamel subtly ridged or smooth, (2) striations absent or subtle (Druckenmiller and Maxwell, 2010: character 25). - All specimens where present (0) = Retained primitive character - BRSMG Cb 3578 (2) = ‘Autapomorphy’

119

- As this occurrence is based on a single specimen, and as the striations are very subtle, it is possible that the crown striations were previously much stronger and were perhaps abraded over subsequent years. Thus, this character is rejected as an ‘autapomorphy’ for this individual.

Char 22: Root striations: (0) present, (1) absent or subtle (Fischer et al., 2013: character 4). - All specimens where present (0) = Retained primitive character - CAMSM J35183 (1) = Unusual combination of characters Synapomorphy - BRSMG Cb 3578 (cluster 1) & MOS 8373, ANSP 15766 (Cluster 2) (1) = Synapomorphy - The variation in the root striations suggest that this is not a strong character, although the unusual combination in CAMSM J35183 may suggest otherwise.

120

Char 23: Tooth robusticity: (0) teeth robust (especially roots) and moderately large, (1) teeth moderately slender, (2) teeth mostly very slender, small and delicate. - Temnodontosaurus and part of Cluster 1 & 2 (0) = Retained primitive character - NMW 93.5G.2 and part of cluster 1 & 2 (1) = Synapomorphy - CAMSM J35183 (2) = Autapomorphy - The variation between the tooth robusticity suggests that this is not a reliable character. However, the very slender and delicate nature of the teeth in CAMSM J35183 appears to represent a good, strong character.

Figure B8. The character map for Character 23. This is one of only three characters based on dentition. It shows the ‘autapomorphy’ for CAMSM J35183, but the variation in other specimens.

121

Forefin Char 24: Distal and proximal ends of the humerus: (0) nearly equal in width, (1) distal end noticeably wider than proximal end, (2) proximal end noticeably wider than distal end (modified from Fernández, 2007: character 12). - Outgroup taxa, NMW 93.5G.2, AGC 12 (Cluster 1) & OUMNH J13799 (cluster 2) = Retained primitive character - CAMSM J35183, AGC 17 and portion of cluster 1 & 2 = Synapomorphy - BRSMG Cb 4997 & BRSMG Ce 16611 (Cluster 2) = Synapomorphy - The wide dispersal of results for this character strongly suggests that there is a wide variation in this character. Thus, it is rejected as a reliable character.

Figure B9. The character map for Character 24. This is an example of one of the characters displaying fairly diverse variation.

Char 25: Deltopectoral crest of the humerus: (0) present and large, (1) present and small, (2) extremely reduced (modified from Maxwell, 2010: character 20). - As less than half of the sample displayed this character, it is rendered as not useful for the analysis. Of the specimens whereby it was used, there were no clear differences.

122

Char 26: Humerus size: (0) much longer than wide, (1) nearly squarish (modified from Callaway, 1989: character 28). - All specimens (0) = Retained primitive character - This character was found in all specimens and thus it was not useful for separating individual groups or species.

Char 27: Humerus distal articular facets: (0) radius facet larger than ulna facet, (1) ulna facet larger than radius facet, (2) two facets nearly equal (modified from Motani, 1999b: character 52). - All specimens (2) = Retained primitive character - This character was found in all specimens and was not useful for separating individual groups or species. The only specimen that this could not be coded was the neotype of I. communis.

123

Char 28: Humerus circular depression on the articular surface anterior to the dorsal process of humerus: (0) present, (1) absent. - Outgroup taxa, NMW 93.5G.2 and Cluster 1 (1) = Retained primitive character - Cluster 2 (0) = Autapomorphy - There is a clear distinction between specimens based on this character, therefore it is considered to be a useful, taxonomic character; although the neotype of I. communis could not be coded for this character.

Figure B10. The character map for Character 28. There is a clear difference with this character from the most derived group (Cluster 2) and all primitive specimens.

Char 29: Humerus dorso-ventral constriction in humerus head: (0) present, (1) absent. - As the majority of specimens could not be coded for this character, this is rendered not useful for this data matrix.

124

Char 30: Humerus shape: (0) proximal portion larger and more robust compared to distal (more delicate), (1) distal portion wider and more robust compared to proximal, (2) overall robustness of proximal and distal portion equal. - Outgroup taxa, NMW 93.5G.2 and one specimen in cluster 2 (1) = Retained primitive character - CAMSM J35183, AGC 17, Cluster 1 and one specimen in cluster 2 (2) = Synapomorphy - Cluster 2 (0) = Potential ‘Autapomorphy’ - This character appears to be a good, useful character in order to separate the two clusters of specimens and NMW 93.5G.2 (plus the outgroup taxa). However, two oddities in Cluster 2 display a different character state. This could be due to preservation/orientation of the humerus, thus giving an odd character state. The character is considered to be good.

Figure B11. The character map for Character 30. This is a good character map that shows there could be some usefulness of the character (note the grouping in Cluster 1), but the variation is seen throughout Cluster 2.

125

Char 31: Humerus radius-ulna distal contact (distalmost point of humerus (in dorsal/ventral view) between the radius and ulna): (0) pointed, (1) rounded. - All specimens, majority of cluster 1 and two specimens in cluster 2 (1) = Retained primitive character - Two specimens in cluster 1 and majority of specimens in cluster 2 (0) = Synapomorphy - This character appears to be rather variable with the character appearing in both states in cluster 1 & 2. This is probably not a strong character for separating the specimens. Also, a single specimen (AGC 11) had a different character state for either fin, although this may have been due to the orientation of the fins (ventral v. dorsal view).

Char 32: Radius L/W ratio: (0) longer than wide, (1) wider than long, (2) roughly equal (modified from Motani, 1999b: character 61). - All specimens (1) = Retained primitive character - This character was found in all specimens and thus was not useful for separating individual groups or species.

Char 33: Ulna L/W ratio: (0) longer than wide, (1) wider than long, (2) roughly equal. - All specimens (1) = Retained primitive character - This character was found in all specimens and was not useful for separating individual groups or species. The only specimen that remained uncoded was Suevoleviathan.

Char 34: Radius/ulna relative overall size: (0) practically equal, (1) radius larger than ulna, (2) ulna larger than radius (modified from Motani, 1999b: character 64). - All specimens aside from three within Cluster 1 and two in Cluster 2 (0) = Retained primitive character - Cluster 1 (CAMSM J59575, NHMUK 2013 & AGC 12) and Cluster 2 (AGC 11 & MOS 8373) = Synapomorphy - This character is not reliable for separation of individual species as it is present in the two clusters of specimens.

Char 35: Contact between radius and ulna: (0) broad, (1) narrow.

126

- This character is seen in CAMSM J35183, a specimen in Cluster 1 and several in Cluster 2 (0) = Synapomorphy - All other specimens (aside from Suevoleviathan as ‘?’) (1) = Retained primitive character - As the character state is found in several specimens and in the two clusters, this is not a good character to separate species.

Char 36: Intermedium L/W ratio: (0) longer than wide (rounded hexagon), (1) wider than long (elongated hexagon), (2) roughly equal. - This character is split between several specimens at the primitive end of the tree (NMW 93.5G.2 and CAMSM J35183) and is inconsistent throughout both clusters. It is, however, an autapomorphy for Suevoleviathan. For this reason, it is not considered a reliable character for separating the species.

Char 37: Ulnare/intermedium overall relative size: (0) ulnare noticeably larger than intermedium, (1) intermedium larger than ulnare, (2) roughly equal (modified from Motani, 1999b: character 66). - The character state is not consistent with any specimen, aside from the outgroup taxa, as it is variable between all other specimens. Therefore, it is not considered to be a reliable taxonomic character.

127

Char 38: Radiale/intermedium width: (0) radiale noticeably wider than intermedium, (1) intermedium wider than radiale, (2) roughly equal. - This character is extremely variable and is found across the tree. However, the majority of Cluster 2 shares the same state (2). But, the same character is found in two specimens of Cluster 1. It is not considered to be a reliable character.

Figure B12. The character map for Character 38. This is another example of how the character state varies throughout the tree, rendering its usefulness as a taxonomic character poor.

128

Char 39: Metacarpals: (0) rounded, not forming tightly interlocking mosaic, (1) polygonal, forming a tightly interlocking mosaic (Maxwell et al., 2012: character 46). - Suevoleviathan and NMW 93.5G.2 (0) = Synapomorphy - Temnodontosaurus and all other specimens(1) = Retained primitive character - This is considered to be a reliable character demonstrating that NMW 93.5G.2 is a distinct species to the rest of the sampled specimens.

Figure B13. The character map for Character 39. This is another good character that shows a synapomorphy between NMW 93.5G.2 and Suevoleviathan.

Char 40: Phalanges shape: (0) mostly rounded, (1) mostly polygonal, (2) combination of both rounded and rectangular (from Fischer et al., 2011, modified from Maisch and Matzke, 2000: character 102). - The character state is interchangeable between specimens at single branches and throughout both clusters. Thus, it is not considered to be a distinct character.

129

Char 41: Digital bifurcation row: (0) no bifurcation, (1) occurs in metacarpal row, (2) occurs in p1 or p2 row, (3) occurs in more distal row. - This character is extremely variable. It demonstrates the variability of the positioning of the bifurcation in Ichthyosaurus specimens. However, the outgroup taxa are separate as they do not have a bifurcation.

Figure B14. The character map for Character 41. Another example of how a character state is spread across the tree, although it may have some potential usefulness in Cluster 1. It also separates the genus (Ichthyosaurus) from the outgroup taxa.

130

Char 42: Extent of notching: (0) absent, (1) radius and other elements notched, (2) radius not notched, but other elements are. - The presence of this character is seen across the tree. It is therefore determined that notching cannot be used to distinguish between species of Ichthyosaurus, at least in the given sample.

Figure B15. The character map for Character 42. One of the characters that did not display any real taxonomic usefulness, apart from separating Temnodontosaurus from the rest.

Char 43: Anterior accessory digit: (0) absent, (1) present. - As all other specimens (aside from NHMUK 2013) do not have the presence of this state this cannot be used as a good, reliable character. It is possible that an anterior accessory digit may have been present in other specimens, but perhaps removed during preparation or eroded away.

Char 44: Posterior accessory digit: (0) absent, (1) present. - All specimens (where coded) (1) = Retained primitive character - This is not considered a reliable character for species distinction as it is found in all specimens (where coded).

131

Char 45: Posterior accessory digit initial point of contact: (0) ulna, (1) ulnare, (2) metacarpal, (3) phalangeal level. - Where preserved this is an unusual character, but it is found to vary among the clusters and thus may therefore be considered as intraspecific variation (if preserved/present) in the genus Ichthyosaurus.

Char 46: Humerus/femur ratio (humerus length divided by femur length): (0) humerus 1.3 times or less as large as femur, (1) humerus 1.4 times as large as femur, (2) humerus 1.5 times as large as femur. - NMW 93.5G.2, CAMSM J35183, Cluster 2 (and one specimen from cluster 1 AGC 12) (0) = Retained primitive character - AGC 17, Cluster 1 (two specimens) and cluster (two specimens) (1) = Synapomorphy - It appears that this character is not useful for Cluster 1, however, it may be of some use for Cluster 2 (disregarding AGC 12). But, the character state does differentiate between the two clusters.

Hindfin Char 47: Distal and proximal ends of the femur: (0) nearly equal in width, (1) distal end wider than proximal end. - All specimens bar two (1) = Retained primitive character - AGC 17 and BRSMG Cb 4997 (in cluster 2) (0) = Synapomorphy - This is quite an unusual character. It may prove to be a potentially useful taxonomic character.

Char 48: Anterior portion of distal end of femur: (0) curved (almost right angle), (1) slight ‘flared’ curve, (2) widely (‘flared’) curve. - NMW 93.5G.2 and three specimens in cluster 1 (1) = Shared derived character - AGC 17, Cluster 1 and three specimens in cluster 2 (0) = Synapomorphy - It is possible that this character may have some use. However, as several of the specimens in Cluster 2 vary with regards to the character state, it cannot currently be determined to represent a diagnostic species character.

132

Char 49: Posteriormost portion of distal end of femur: (0) right angle, (1) gentle ‘flared’ curve, (2) widely ‘flared’ curve. - Temnodontosaurus, NMW 93.5G.2, CAMSM J35183 and OUMNH J13799 (cluster 2) = Retained primitive character - AGC 17, Cluster 1 and Cluster 2 = Synapomorphy - Aside from the oddity of OUMNH J13799 this appears to be a viable character and demonstrates that both clusters 1 and 2 share this feature, but more primitive specimens do not.

Char 50: Number of digits in hindfin: (0) three, (1) four, (2) five, (3) six. - NMW 93.5G.2 and all other specimens (bar those coded with a ‘?’, outgroup taxa and one specimen from cluster 2 (1) = Synapomorphy - I. communis (2) = ‘Autapomorphy’ - This character appears to be useful for determining specimens at genus level. However, the real oddity is that the neotype of I. communis is the only specimen (aside from outgroup taxa) to have a different character state. Not useful for separating species, but requires further assessment regarding generic assignment.

Figure B16. The character map for Character 50. An apparent good genus-level character for Ichthyosaurus, yet the neotype of I. communis is coded differently.

133

Char 51: Tibia L/W ratio: (0) longer than wide, (1) wider than long (Motani, 1999b: character 90). - All specimens (where coded) (1) = Retained primitive character - This is not considered to represent a reliable character for species distinction as it is found in all specimens.

Char 52: Tibia shape: (0) pentagonal, (1) polygonal, (2) circular. - This character is found to vary between both clusters of specimens and is not useful for separating species.

Char 53: Tibia length: (0) approximately twice as proximodistally long than more distal elements, (1) tibia only slightly longer than successive elements (from Maxwell et al., 2012, modified from Motani, 1999b: character 90). - This character is found to vary between both clusters of specimens and is thus not useful for determining species. - Outgroup taxa and AGC 17 along with a specimen in cluster 1 (AGC 12) and three specimens from cluster 2 (1) = Retained primitive character - Majority of Cluster 1 and Cluster 2 (0) = Synapomorphy - As both character states are found to vary between both clusters, this is not considered to represent a good taxonomic character.

Char 54: Fibula L/W ratio: (0) longer than wide, (1) wider than long, (2) roughly equal. - This character is found in both clusters and is variable. It is not considered to be a distinctive character between species.

Char 55: Fibula shape: (0) pentagonal, (1) polygonal/rectangular, (2) circular. - AGC 17, majority of cluster 1 (bar CAMSM J59575) and majority of cluster 2 (bar MOS 8373) (1) = Synapomorphy - This may suggest intraspecific variation in both clusters, and perhaps an ontogenetic/sexual dimorphic character.

134

Char 56: Tibia and fibula – relative size: (0) tibia larger than fibula, (1) approximately equal in size, (2) tibia anterioposteriorly and proximodistally smaller than fibula, (3) fibula proximodistally larger than tibia, but anteroposteriorly roughly equal, (4) fibula both as proximodistally longer and anteroposteriorly wider (modified from Maxwell et al., 2012: character 56). - All specimens (aside from Temnodontosaurus and those coded ‘?’, plus OUMNH J13799) (3) = Synapomorphy - The only oddity with this character is OUMNH J13799, which is part of Cluster 2. This may suggest that the aforementioned specimen may represent an oddity.

Figure B17. The character map for Character 56. One specimen (apart from the outgroup taxa) is coded differently to the others, and may suggest an oddity.

135

Char 57: Intermedium L/W ratio: (0) longer than wide, (1) wider than long, (2) roughly equal. - Temnodontosaurus, cluster 1 (all specimens) and some specimens in cluster 2 (2) = Retained primitive character - AGC 17 and majority of cluster 2 (1) = Synapomorphy - Cluster 1 displays this character in all specimens. Although, the character state is the same in Cluster 2 (three specimens). A potentially helpful character.

Char 58: Intermedium shape: (0) pentagonal, (1) hexagonal, (2) rectangular, (3) rounded. - Temnodontosaurus and MOS 8873 (cluster 2) = Retained primitive character - All other specimens (where coded) (0) = Synapomorphy - It appears that MOS 8873 is an oddity here. This is not considered a character for reliable species distinction.

Char 59: Intermedium and tibiale – relative size: (0) intermedium larger than tibiale, (1) approximately equal in size, (2) tibiale larger than intermedium. - This character is variable between both clusters of specimens and thus not considered to be useful in differentiating species.

Figure B18. The character map for Character 59. A variable character between both cluster 1 & 2.

136

Char 60: Shape of phalanges: (0) mostly rounded, (1) mostly polygonal, (2) combination of rounded and rectangular phalanges. - NMW 93.5G.2, outgroup taxa and MOS 8373 (cluster 1) (0) = Retained primitive character - AGC 17, majority of Cluster 1 and Cluster 2 (2) = Synapomorphy - There is one specimen in cluster 1 that is different and two specimens in Cluster 2. Thus, it is not considered a strong character.

Char 61: Digital bifurcation: (0) absent, (1) present. - All specimens (aside from outgroup taxa) (1) = Synapomorphy - This is a good character to determine the specimens at generic level, i.e. Ichthyosaurus.

Figure B19. The character map for Character 61. A good genus-level character for Ichthyosaurus.

Char 62: Anterior accessory digit: (0) absent, (1) present. - As only two specimens have this character present, and they are within different clusters, it is not considered a reliable character.

Char 63: Posterior accessory digit: (0) absent, (1) present.

137

- All specimens (where coded) aside from Cluster 2 (1) = Retained primitive character - Cluster 2 (aside from one specimen, OUMNH J13799) (0) = Synapomorphy - The oddity here is OUMNH J13799 as it appears with a different character state to the rest in this cluster. Nevertheless, 11 specimens from the total of 18 were not coded and therefore this character cannot be used reliably.

Char 64: Notching: (0) absent, (1) tibia and other elements notched, (2) tibia not notched, but other elements notched. - All specimens aside from outgroup taxa and one specimen from Cluster 2 (2) = Synapomorphy - I. communis (Cluster 2) (0) = ‘Autapomorphy’ - The only specimen not to share this character state is the I. communis neotype. This character would have otherwise been a strong genus-level character. Further assessment of the neotype is required (once made accessible).

Figure B20. The character map for Character 64. An apparent good genus-level character for Ichthyosaurus, yet the neotype of I. communis is coded differently.

138

Pectoral girdle Char 65: Scapula: (0) with prominent acromion process, (1) large acromion process absent, but anterior expansion of proximal scapula present, (2) proximal expansion absent, scapula strap-shaped (Maxwell et al., 2012: character 31). - All specimens (aside from outgroup taxa) (1) = Synapomorphy - This is a good genus-level character, i.e. Ichthyosaurus.

Figure B21. The character map for Character 65. A good genus-level character for Ichthyosaurus.

139

Char 66: Scapula overall robustness: (0) very robust, (1) slightly robust (more delicate) and more elongated. - This character is seen to vary throughout the tree and therefore is considered to not be a reliable character with regards to differentiating species.

Figure B22. The character map for Character 66. An example of a variable character found throughout the tree.

140

Char 67: Glenoid contribution of the scapula: (0) extensive, being at least as large as the coracoid facet, (1) reduced, being markedly smaller than the coracoid facet (Fischer et al., 2012: character 27). - Temnodontosaurus, NMW 93.5G.2 and BRSMG Cb 3578 (Cluster 1) (0) = Retained primitive character - Cluster 2 (1) = Autapomorphy - There is a clear distinction with this character from Cluster 2 and the rest. However, several specimens in Cluster 1 (total of four) were not coded. Nevertheless, this may be a potentially useful character for separating species.

Figure B23. The character map for Character 67. A potentially good autapomorphic character to separate the most derived group (Cluster 2) from the others.

Char 68: Coracoid: (0) mediolaterally wider than anteroposteriorly long, (1) approximately equidimensional, (2) anteroposteriorly longer than mediolaterally wide (Maxwell et al., 2012: character 32). - This character is unusual in that both the outgroup taxa share the same character state, but one specimen from either cluster also. The rest display a single character state. Thus, this demonstrates considerable variation in the coracoid width/length and is thus variable among specimens.

141

Char 69: Coracoid: (0) with well-developed anterior notch, (1) anterior notch reduced (small), (2) anterior notch absent (modified from Maxwell et al., 2012: character 33). - Outgroup taxa and CAMSM J35183 (1) = Retained primitive character - AGC 17, Cluster 1 and Cluster 2 (0) = Synapomorphy - This is a good character for separating CAMSM J35183 from the other groups of specimens. It is a potentially strong character for this individual.

Figure B24. The character map for Character 69. A good character found in all Ichthyosaurus specimens, apart from CAMSM J35183 which has a character state retained with the outgroup taxa.

142

Char 70: Coracoid anterior notch shape: (0) J-shaped, (1) closed C-shaped, (2) wide C-shaped. - The shape of the coracoid anterior notch varies among the two clusters (primarily) and in more primitive specimens. It is not considered a reliable character.

Figure B25. The character map for Character 70. An example of a character that may have some taxonomic usefulness, but some specimens are coded differently.

Char 71: Coracoid posterior shape: (0) J-shaped, (1) closed C-shaped, (2) wide C- shaped, (3) absent. - This character is largely the same (0) throughout the tree, but three specimens display a different character state. Nevertheless, this suggests that at least the posterior coracoid notch is present in Ichthyosaurus (although variable), and absent in other genera i.e. Temnodontosaurus.

Pelvic girdle Char 72: Ilium: (0) with expanded proximal region, (1) narrow proximally and distally, rib-like (from Maxwell et al., 2012, modified from Sander 2000: character 106). - All specimens (aside from Suevoleviathan) (1) = Retained primitive character - Not a good character for separating species.

143

Char 73: Pubis shape, distally: (0) not expanded, (1) expanded ‘spatulate shape’, (2) widely expanded and asymmetric (‘fan-shaped’). - This character is seen to vary among both Cluster 1 and Cluster 2 and is therefore not considered to currently represent a character of taxonomic use. Perhaps it is sexual dimorphic.

Figure B26. The character map for Character 73. A character that was independently suggested to show some taxonomic usefulness, but does not, as the character state varies throughout the tree.

144

Char 74: Ischium overall shape: (0) long and slender (especially in centre), (1) long, fairly robust and oblong-shaped, (2) short, robust and wide (modified from Motani, 1999b: character 87). - NMW 93.5G.2 (2) = Autapomorphy - It appears that the character state is variable among the other species (groups) and the outgroup taxa; yet, is a good, taxonomic character for NMW 93.5G.2.

Figure B27. The character map for Character 74. Note the autapomorphy for NMW 93.5G.2, but the rest of the specimens are variable between the other character states.

145

Char 75: Ischium shape, distally: (0) not expanded, essentially straight, (1) marginally expanded, (2) widely expanded. - Temnodontosaurus, NMW 93.5G.2, CAMSM J35183 and BRSMG Ce 16611 (Cluster 2) (2) = Retained primitive character - Cluster 1 and Cluster 2) (1) = Synapomorphy - Aside from the oddity of the specimen in Cluster 2, which may be due to orientation, this appears to be a reliable character. It is useful for both NMW 93.5G.2 and CAMSM J35183 which may further suggest their identification as distinct species.

Figure B28. The character map for Character 75. An example of another character that may be of taxonomic usefulness, although there is an oddity in Cluster 2.

146

Char 76: Ischium: (0) with greatly expanded medial edge, (1) ischium lacking medial expansion (Maxwell et al., 2012: character 53). - Outgroup taxa and NMW 93.5G.2 (0) = Retained primitive character - Cluster 1 and Cluster 2 (1) = Synapomorphy - A good character that demonstrates separation of NMW 93.5G.2 from the two clusters, but a character that is shared between both clusters.

Figure B29. The character map for Character 76. A character retained by the outgroup taxa and NMW 93.5G.2 with both cluster 1& 2 representing a synapomorphy.

Char 77: Ilium/pubis relative length: (0) nearly equal, (1) ilium longer than pubis, (2) pubis longer than ilium. - Temnodontosaurus, cluster 1 and two specimens from cluster 2 (0) = Retained primitive character - Cluster 1 (three specimens) (2) = ‘Autapomorphy’ - As Cluster 2 has both character states present in the specimens, it is not considered to be a good character. However, coupled with other characters, it is useful for potentially distinguishing those from Cluster 1.

147

Char 78: Ilium/ischium relative length: (0) nearly equal, (1) ilium longer than ischium, (2) ischium longer than ilium. - NMW 93.5G.2 (1) = Autapomorphy - Cluster 1 and Cluster 2 (aside from two specimens) (0) = Synapomorphy - This character is extremely useful for NMW 93.5G.2 as a distinct species. The oddity of two specimens in Cluster 2 does, however, remove the usefulness of this character in that group of specimens.

Figure B30. The character map for Character 78. An autapomorphy for NMW 93.5G.2, but rather variable throughout the rest of the tree.

Char 79: Pubis/ischium relative length: (0) nearly equal, (1) pubis longer than ischium, (2) ischium longer than pubis (modified from Motani, 1999b: character 86). - All specimens aside from ANSP 17429 (Cluster 2) (0) = Retained primitive character - As only a single specimen displays something different (although eight specimens could not be coded), this character is not useful for differentiating species.

Char 80: Ischium/pubis overall robustness: (0) pubis more robust, (1) ischium more robust, (2) roughly equal.

148

- All primitive specimens could not be coded for this character, and the two clusters are interchangeable. Due to the lack of primitive specimens coded, this character is not useful.

Char 81: Ilium/femur size: (0) nearly equal, (1) ilium longer than femur, (2) femur longer than ilium. - NMW 93.5G.2, ANSP 17429 & BRSMG Ce 16611 (cluster 2) (2) = Retained primitive character - CAMSM J35183, cluster 1 and cluster 2 (majority) (0) = Synapomorphy - I. communis (1) = ‘Autapomorphy’ - This is an interesting character. It appears that NMW 93.5G.3 can be distinguished from other specimens (aside from two dubious specimens in Cluster 2) on the basis of this character state. Cluster 1 comprises a single character state that may be taxonomically useful, in combination with other characters. The neotype of I. communis has a completely different state to the others and further suggests that this specimen must be examined more closely.

Figure B31. The character map for Character 81. A possible good species-level character, yet the neotype of I. communis is coded differently.

149

Char 82: Pubis/femur size: (0) nearly equal, (1) pubis longer than femur, (2) femur longer than pubis. - AGC 17, cluster 1 and MOS 8373 (Cluster 2) (0) = Synapomorphy - Cluster 2 (aside from above specimen) (1) = Autapomorphy - This is a good character for separating the individual clusters of specimens, although MOS 8373 is an oddity.

Figure B32. The character map for Character 82. A clear difference can be seen between both cluster 1 & 2 although there is the one oddity in Cluster 2.

Char 83: Ischium/femur size: (0) nearly equal, (1) ischium longer than femur, (2) femur longer than ischium. - NMW 93.5G.2 and ANSP 17429 (Cluster 2) (2) = ‘Synapomorphy’ - Cluster 1 and MOS 8373 (Cluster 2) (0) = ‘Synapomorphy’ - Cluster 2 (aside from above specimens) (1) = Autapomorphy - This character state may be dependent on how this is interpreted, in that character state (2) may be much longer (NMW 93.5G.2) or only marginally (ANSP 17429). With regards to the latter, this may suggest preservation or orientation of the individual element has produced an ‘oddity’. The same can be said for MOS 8373.

150

Vertebral column Char 84: Mid-caudal centra height change: (0) gradual decrease, (1) sudden decrease, (2) increase, (3) sudden increase (modified from Motani, 1999b: character 98). - NMW 93.5G.2 (and all other specimens where coded, aside from outgroup taxa) (0) = Synapomorphy - This is a good character to determine that all specimens (aside from outgroup) belong to Ichthyosaurus, i.e. a good genus-level character.

Figure B33. The character map for Character 84. A good genus-level character for Ichthyosaurus.

Char 85: Anterior-middle dorsal neural spine: (0) narrow, relatively short and straight, (1) narrow, high, and straight, (2) abruptly high (modified from Callaway, 1989: character 16). - Outgroup taxa and Cluster 1 (0) = Retained primitive character - NMW 93.5G.2 and Cluster 2 (1) = Synapomorphy - This appears to represent a good taxonomic character, although three specimens in Cluster 1 and five specimens in Cluster 2 were not coded.

Char 86: Preflexural vertebral count: (0) less than 80, (1) greater than 80.

151

- NMW 93.5G.2 (and all other specimens where coded, aside from outgroup taxa) (0) = Synapomorphy - This is a good character to determine that all specimens (aside from outgroup taxa) belong to Ichthyosaurus, i.e. a good genus-level character.

Figure B34. The character map for Character 85. Clearly, Cluster 1 has a retained primitive character with the outgroup taxa and NMW 93.5G.2 and Cluster 2 share a synapomorphy.

Figure B35. The character map for Character 86. A good genus-level character for Ichthyosaurus.

152

Appendix C: Individual character coding for the cladistic analysis

The specific coding for all of the individual specimen character states for the cladistic analysis is listed below. A ‘?’ represents a character that is unknown in that individual due either to poor preservation, or that the feature could not be reliably confirmed, i.e. that it was missing. A character coded as (01) or (12) indicates that both character states are present in that specimen. Please see Appendix A for the characters and their descriptions. See Appendix B for interpretations of the individual usefulness of the characters (and character maps) used during the analysis. *A = Number of characters uncoded. *B = Percentage of completeness for that individual.

Specimen/Taxa 0 1 2 3 4 5 6 7 8 9 10 11 Temnodontosaurus 1 1 1 2 [01] 1 0 1 1 1 0 1 Suevoleviathan 1 0 2 ? 1 ? 2 1 0 0 0 1 I. communis (Neotype) ? 1 1 0 0 0 0 1 0 1 0 1 AGC 11 ? 1 ? ? 0 ? ? ? ? 1 0 1 AGC 12 ? 1 2 2 0 0 2 1 0 ? ? 1 AGC 17 ? ? ? ? ? ? ? ? ? ? ? ? ANSP 15766 4 1 2 0 0 0 0 1 1 1 0 1 ANSP 17429 4 1 2 0 0 0 ? ? ? ? 0 1 BRSMG Cb 3578 4 ? 1 2 ? 0 0 1 0 1 0 1 BRSMG Cb 4997 ? 1 2 0 ? ? 0 1 0 1 0 1 BRSMG Ce 16611 ? ? ? ? ? ? ? ? ? ? ? ? CAMSM J35183 3 0 ? 2 0 1 ? 1 0 1 0 1 CAMSM J59575 ? ? 2 ? ? ? ? ? ? ? ? ? MOS 8373 4 1 1 0 0 0 0 1 0 1 0 1 NHMUK 2013 ? 1 2 2 0 0 0 1 0 1 0 1 NHMUK R3372 ? ? ? ? ? ? ? ? ? ? 0 ? NMW 93.5G.2 2 1 3 2 1 0 0 1 0 1 ? 1 OUMNH J13799 ? ? ? ? ? ? ? ? ? ? ? ?

153

Specimen/Taxa 12 13 14 15 16 17 18 19 20 21 22 23 Temnodontosaurus 0 1 0 ? 1 1 1 1 0 0 0 0 Suevoleviathan 0 0 0 0 1 0 1 ? 0 ? ? ? I. communis (Neotype) ? 0 0 ? ? ? ? 1 ? 0 0 1 AGC 11 1 ? 0 0 1 ? ? 1 ? 0 0 1 AGC 12 ? ? ? 0 2 1 ? 1 2 0 0 0 AGC 17 ? ? ? ? ? ? ? ? ? 0 0 1 ANSP 15766 1 0 0 0 1 ? ? 1 0 0 1 1 ANSP 17429 ? 0 ? ? ? ? ? ? ? 0 0 1 BRSMG Cb 3578 0 0 ? ? ? ? ? 1 ? 2 1 1 BRSMG Cb 4997 ? ? 0 0 2 ? ? 1 2 0 0 0 BRSMG Ce 16611 ? ? ? 0 2 ? ? ? ? ? ? ? CAMSM J35183 1 ? 0 0 1 ? ? 1 0 0 1 2 CAMSM J59575 ? ? ? 0 2 ? ? 0 2 0 0 1 MOS 8373 1 0 0 0 1 ? 1 ? 0 0 1 1 NHMUK 2013 1 0 ? 0 2 1 1 ? ? 0 0 0 NHMUK R3372 ? ? ? 0 1 ? ? ? 0 0 0 0 NMW 93.5G.2 1 0 0 1 1 1 1 1 0 0 0 1 OUMNH J13799 ? ? ? ? ? ? 1 ? ? 0 ? 0

154

Specimen/Taxa 24 25 26 27 28 29 30 31 32 33 34 35 Temnodontosaurus 1 2 0 2 1 1 1 1 1 1 0 0 Suevoleviathan 1 ? 0 2 1 1 1 ? 1 ? 0 ? I. communis (Neotype) 0 ? 0 ? ? ? ? 0 1 1 0 0 AGC 11 0 0 0 2 0 ? 0 [01] 1 1 2 1 AGC 12 1 1 0 2 1 1 2 1 1 1 2 0 AGC 17 0 ? 0 2 ? ? 2 1 1 1 0 0 ANSP 15766 0 ? 0 2 0 ? 2 0 1 1 0 0 ANSP 17429 0 ? 0 2 0 ? 0 1 1 1 0 1 BRSMG Cb 3578 0 1 0 2 ? ? 2 1 1 1 0 1 BRSMG Cb 4997 2 1 0 2 ? ? 0 0 1 1 0 1 BRSMG Ce 16611 2 1 0 2 ? ? 1 0 1 1 0 0 CAMSM J35183 0 ? 0 2 ? ? 2 1 1 1 0 1 CAMSM J59575 ? ? 0 2 ? ? ? 1 1 1 2 0 MOS 8373 0 ? 0 2 0 ? 0 0 1 1 2 0 NHMUK 2013 0 ? 0 2 1 ? 2 0 1 1 2 0 NHMUK R3372 0 ? 0 2 1 ? 2 0 1 1 0 0 NMW 93.5G.2 1 ? 0 2 1 1 1 1 1 1 0 0 OUMNH J13799 1 1 0 2 ? ? 0 1 1 1 0 0

155

Specimen/Taxa 36 37 38 39 40 41 42 43 44 45 46 47 Temnodontosaurus [12] 1 0 1 2 0 1 0 1 ? ? 1 Suevoleviathan 0 1 ? 0 0 0 0 0 1 ? ? 1 I. communis (Neotype) 1 2 0 1 2 2 0 0 1 ? ? 1 AGC 11 1 2 0 1 2 [12] 0 0 ? ? ? ? AGC 12 2 2 0 1 2 3 0 0 1 1 0 ? AGC 17 1 2 1 1 2 2 0 ? ? ? 1 0 ANSP 15766 2 0 2 1 1 2 0 0 1 2 1 1 ANSP 17429 1 2 1 1 1 3 0 0 ? ? 0 1 BRSMG Cb 3578 2 2 2 1 2 3 0 0 1 ? 1 ? BRSMG Cb 4997 2 2 2 1 1 1 2 0 1 2 ? 0 BRSMG Ce 16611 2 0 2 1 1 2 2 0 1 0 0 1 CAMSM J35183 2 0 0 1 2 2 0 0 1 ? 0 1 CAMSM J59575 2 0 2 1 1 3 0 ? 1 2 1 ? MOS 8373 2 0 2 1 2 2 0 ? 1 ? 1 1 NHMUK 2013 1 2 1 1 1 2 2 1 1 0 ? 1 NHMUK R3372 1 0 0 1 2 3 0 0 ? 0 ? 1 NMW 93.5G.2 1 0 0 0 2 2 2 ? ? ? 0 1 OUMNH J13799 1 2 2 1 1 3 0 0 1 0 0 1

156

Specimen/Taxa 48 49 50 51 52 53 54 55 56 57 58 59 Temnodontosaurus 2 2 0 1 2 1 2 2 1 2 3 ? Suevoleviathan ? 0 0 1 ? 1 ? ? [12] ? ? ? I. communis (Neotype) 1 1 2 1 2 1 2 1 3 2 0 0 AGC 11 0 1 1 1 2 1 ? ? 3 2 0 1 AGC 12 0 1 ? 1 1 1 1 1 3 2 0 0 AGC 17 0 1 1 1 2 1 2 1 3 1 0 1 ANSP 15766 0 1 1 1 2 1 2 1 3 1 0 0 ANSP 17429 0 1 ? ? ? ? ? ? ? ? ? ? BRSMG Cb 3578 ? ? ? ? ? ? ? ? ? ? ? ? BRSMG Cb 4997 0 1 1 1 2 0 1 1 3 1 0 1 BRSMG Ce 16611 1 1 1 1 1 0 1 1 3 2 0 1 CAMSM J35183 ? 2 ? ? ? ? ? ? ? ? ? ? CAMSM J59575 ? ? 1 1 1 0 2 0 3 2 0 1 MOS 8373 0 1 1 1 2 0 1 0 3 1 3 0 NHMUK 2013 0 1 1 1 2 0 2 1 3 2 0 1 NHMUK R3372 0 1 1 1 2 0 1 1 3 2 0 0 NMW 93.5G.2 1 2 1 ? ? ? ? ? ? ? ? ? OUMNH J13799 1 2 1 1 2 0 1 1 4 1 0 0

157

Specimen/Taxa 60 61 62 63 64 65 66 67 68 69 70 71 Temnodontosaurus 0 0 0 ? 1 2 0 0 1 1 1 3 Suevoleviathan 0 0 0 1 1 2 ? ? 1 1 ? ? I. communis (Neotype) 2 1 0 ? 0 1 1 1 ? ? ? ? AGC 11 2 1 0 ? 2 1 1 1 2 0 2 0 AGC 12 2 ? ? ? 2 ? 0 ? 2 0 2 1 AGC 17 2 1 0 ? 2 1 0 ? ? 0 2 ? ANSP 15766 1 1 0 0 2 ? 0 ? ? ? ? ? ANSP 17429 2 1 ? ? ? ? ? ? ? ? ? 0 BRSMG Cb 3578 ? ? ? ? ? 1 1 0 2 0 0 2 BRSMG Cb 4997 2 1 0 0 ? 1 0 1 2 0 2 0 BRSMG Ce 16611 2 ? ? ? 2 1 1 1 1 0 2 0 CAMSM J35183 ? ? ? ? ? 1 1 ? 2 1 1 0 CAMSM J59575 2 1 0 ? 2 ? ? ? 1 0 2 2 MOS 8373 0 ? 0 0 2 1 1 1 2 0 1 0 NHMUK 2013 2 1 0 1 2 1 0 ? ? ? ? ? NHMUK R3372 1 1 1 1 2 1 0 ? 2 0 2 0 NMW 93.5G.2 0 1 ? ? 2 1 1 0 ? ? ? ? OUMNH J13799 2 1 1 1 2 1 1 ? 2 0 2 0

158

Specimen/Taxa 72 73 74 75 76 77 78 79 80 81 82 83 Temnodontosaurus 1 1 1 2 0 0 2 0 ? ? ? ? Suevoleviathan 0 2 1 ? 0 ? ? 0 ? ? ? ? I. communis (Neotype) 1 1 0 1 1 0 0 0 2 1 1 1 AGC 11 ? ? ? ? ? ? ? ? ? ? ? ? AGC 12 ? ? ? ? ? ? ? ? ? ? ? ? AGC 17 ? 2 1 ? ? ? ? ? ? ? 0 ? ANSP 15766 ? ? ? ? ? ? ? ? ? ? ? ? ANSP 17429 1 2 0 1 1 2 0 1 2 2 1 2 BRSMG Cb 3578 1 1 0 1 1 0 0 0 2 0 0 0 BRSMG Cb 4997 1 1 0 1 1 ? 0 ? 1 0 ? 1 BRSMG Ce 16611 1 2 0 2 1 2 2 0 2 2 1 1 CAMSM J35183 1 2 ? 2 ? ? ? ? ? 0 ? ? CAMSM J59575 1 1 1 1 1 0 0 0 1 0 0 0 MOS 8373 1 1 1 ? 1 0 0 0 1 0 0 0 NHMUK 2013 1 ? ? ? ? ? ? ? ? 0 ? ? NHMUK R3372 1 2 0 1 1 0 0 0 2 0 0 0 NMW 93.5G.2 1 ? 2 2 0 ? 1 ? ? 2 ? 2 OUMNH J13799 1 1 1 1 1 2 2 0 1 0 1 1

159

Specimen/Taxa 84 85 86 *A *B Temnodontosaurus 2 0 1 9 90 Suevoleviathan 2 0 1 31 64 I. communis (Neotype) ? 1 0 20 77 AGC 11 0 ? 0 32 63 AGC 12 ? ? 0 28 68 AGC 17 ? ? ? 42 52 ANSP 15766 0 1 0 22 75 ANSP 17429 0 ? ? 37 57 BRSMG Cb 3578 0 0 0 29 67 BRSMG Cb 4997 ? ? ? 17 80 BRSMG Ce 16611 ? ? ? 29 67 CAMSM J35183 ? ? ? 38 56 CAMSM J59575 ? ? ? 32 63 MOS 8373 0 1 0 7 92 NHMUK 2013 0 0 ? 23 74 NHMUK R3372 ? ? ? 24 72 NMW 93.5G.2 0 1 0 25 71 OUMNH J13799 ? ? ? 26 70

160