9The Life Appearance of Some Fossil Animal Groups
‘ …Mesozoic archosaurs – dinosaurs and pterosaurs especially – are nowadays frequently depicted in high-fidelity skeletal form before, or at the same time as, fleshed out life reconstructions appear. Outside of archosaurs, little similar work is obvious’.
Our discussion of the restoration process has, thus far, been broad and generalized. In this chapter, we will focus specifically on restoration considerations for several groups of popular fossil animals. Entire book chapters could be written about the life appearance of each group covered below and readers should not view them as comprehensive, detailed guides. Rather, they are intended as a basic introduction to the life appearance of certain groups, aids to interpreting their basic anatomy and warnings about common errors, and to demonstrate how the principles outlined in earlier chapters can be applied to specific taxa. They also mention some of the better aids to wider reading and research, though the lack of good review texts for many groups precludes this possibility for all taxa under consideration.
Fossil ‘amphibians’: Tetrapodomorphs, the first tetrapods, and temnospondyls
‘Fossil amphibians’ are a grade of vertebrate animals characterized by their ability to bridge life in water and land (Figs. 9.1–9.2). First appearing in the Devonian Period, they include the tetrapodomorphs (rather fish-like animals with limbs only recently adapted from fins); the first tetrapods (generally heavyset, semi-aquatic species) and temnospondyls (the major ‘amphibian’ radiation, likely including our modern frogs, salamanders and caecilians – the Lissamphibia). The term ‘amphibian’ is a little problematic as fossil examples of this group are often very different from living species, frequently being much larger and living very different lifestyles. Hundreds of millions of years of life at the water-land interface has seen these species shift between predominantly terrestrial or aquatic habits many times – some lineages were truly semi-aquatic, but many were entirely aquatic or more strongly terrestrial. Unlike modern amphibians, Palaeozoic and Mesozoic ‘amphibians’ were not only able to survive in freshwater: some seem to have dwelt in marine habitats.
Fig. 9.1 Jurassic temnospondyl Siderops kehli, demonstrating the stereotyped anatomy of many prehistoric ‘amphibians’. (J. Conway)
Fig. 9.2 Carboniferous/ Permian temnospondyl Platyhystrix rugosus, a terrestrially adapted species with a distinctive sail. This animal is primarily known from its sail, and is reconstructed here with heavy influence from a close relative, Cacops aspidephorus. (M. Witton)
Palaeozoic and Mesozoic ‘amphibians’ are characterized by large heads, generally tubular, low-slung bodies, powerful tails and short limbs. Their closest modern visual analogues are probably large salamanders, crocodylians and non-tetrapod lobe-finned fishes. Their skulls are often flattened but they can bear prominent ridges and depressions. These are not always obvious because literature on these animals prioritizes dorsal and ventral views of skulls, so lateral and/or oblique views of skulls are recommended additional references. As with modern frogs, large shoulder girdles and short necks likely restricted motion of the head compared to later tetrapods. As might be expected for animals switching between terrestrial and aquatic locomotion, much variation is seen in their limb skeleton: they might be paddle-like structures, longer but stocky walking limbs, or wholly reduced in aquatic, tail-propelled species. All species likely sprawled when using their limbs in semi-aquatic or terrestrial locomotion, but not all bear well-defined limb joints. The limbs of these species may have been cartilage-dominant and paddle-like, while those with robust articular surfaces were probably capable of flexing their limbs like other tetrapods. Though small species may have moved quickly, and aquatic species may have been graceful in water, it is unlikely extinct ‘amphibians’ were sprightly and agile as a rule. Non-tetrapod tetrapodomorphs – like the stegocephalians and whatcheeriids – might have dragged, rather than walked, themselves over land. All fossil ‘amphibians’ likely had well-muscled bodies and tails, possibly augmented with soft tissue fins in aquatic species. Trunk flexion was variable, with some species capable of only limited motion because of large ribs and trunk-binding osteoderms.
Several lines of evidence suggest early tetrapods and temnospondyls had tough skin instead of moist, delicate skin like living amphibians. New insights into living amphibian skin suggest a keratinization potential akin to that of fully-terrestrial tetrapods (Maddin et al. 2007). We can imagine that fossil terrestrial ‘amphibians’ would have benefitted from thicker, more keratinized epidermal tissues that reduced water loss, as seen in some modern lungfish and caecilians (Mittal and Whitear 1979). More direct evidence of thickened dermal tissues include skulls covered with conspicuous tubercles, pits, ridges and furrows, as well as with minimized, sharply-defined skull openings. These are thought to reflect a tight, tough facial skin (Witzmann et al. 2010), perhaps somewhat akin to that of living crocodylians. Similar rugosities are seen on their pectoral girdles and (where present) osteoderms, implying similarly tough skin elsewhere on the body (Witzmann et al. 2010), and maybe epidermal scales over the osteoderms. Overlapping bony scales are frequently preserved in Palaeozoic temnospondyls and early tetrapods, implying that fish-like scaly skin with a thinner epidermis may have covered some parts of the body. There is probably no ‘one skin fits all’ rule for fossil ‘amphibians’ however: many Mesozoic and Cenozoic lineages lack evidence of bony scales, and not all species have indications of tightly-bound facial tissues.
Fig. 9.3 Gill structures in early tetrapods. (A) Fossil larva of the temnospondyl Isodectes obtusus showing external gill structures. (B) Distribution of gills throughout tetrapod phylogeny. (C) Reconstruction of Archegosaurus decheni larva. (A) and (B) after Schoch and Witzmann (2011).
Internal or external gills were ubiquitous among early tetrapods and temnospondyls, though few linages had both (Witzmann et al. 2004; Schoch and Witzmann, 2011). The fact that extinct temnospondyls had external gills during their larval phase is evidenced by fossils showing three soft tissue lobes projecting from the neck region, structures which look very comparable to those of axolotls and larval salamanders (Fig. 9.3; Schoch and Witzmann, 2011). External gills also have osteological correlates: smooth ceratobranchials (bones that support each gill arch) and a reduced operculum (a large bone at the back of the skull), whereas the opposite is true in animals with internal gills (Schoch and Witzmann, 2011). Be warned that gill apparatus change with growth, so lack of external gill features in adult animals has little bearing on their larval state – fossils of larval specimens are the only means to detect or refute the presence of external gills in early life. Some temnospondyls seem to have kept their external gills even as adults, such as the branchiosaurs, but this does not generally seem common.
Carrol’s textbook The Rise of Amphibians (2009) gives an excellent, extremely well-illustrated overview of Palaeozoic and Mesozoic ‘amphibians’. Clack’s Gaining Ground: The Origin and Evolution of Tetrapods (2012, 2nd Edition) provides a good introduction to the complexities of interpreting tetrapod evolution and anatomy.
Synapsids: the mammal-line
The synapsid line has received mixed attention from artists. Crown group mammals are some of the most restorable of all fossil animals, while their stem relatives – particularly those of the Palaeozoic – are much more enigmatic. All synapsid reconstructions benefit from having our well-studied and diverse mammalian fauna as reference material, but artists are advised to resist undue ‘mammalification’ of stem-mammal species. The skeletons of many Carboniferous and Permian synapsids are only slightly modified from those of early tetrapods, and there is no indication that features of true mammals – upright poses, hair and whiskers, fleshy noses, pinnae and so on – emerged on the synapsid line the moment we split from other amniotes. Though characteristics of crown mammals likely had origins in earlier synapsid relatives, most evidence suggests a gradual acquisition of mammalian traits throughout the Palaeozoic and Mesozoic, and it might only be bone fide mammals that bear the full complement of features.
Early synapsids
The first synapsids are the ‘pelycosaurs’, an evolutionary ‘grade’ containing the eothyridids, caseids, varanopids, ophiacodontids, edaphosaurids and sphenacodontids. They are best known by the sail-backed species Dimetrodon and Edaphosaurus (Figs. 8.1, 9.4). They retain many characteristics of early tetrapods including stout, sprawling limbs with large limb girdles; long, slow-slung bodies; dragging tails that anchored hindlimb muscles, and short necks. Their torsos are generally taller in lateral view however, and some of this increased height is facilitated by longer neural spines on the dorsal vertebrae. These likely reflected large muscles running along the spine. Presumably, this was also true of sail-backed pelycosaurs too – the ‘sails’ probably had as much muscle at their base as did other synapsids, with the ‘sail’ region emerging above this (Huttenlocker et al. 2010). Recent work suggests the large sails of Dimetrodon were not entirely covered in webbed-skin: the tips seem to have protruded beyond any significant soft tissue ‘sail’ structure (Rega et al. 2012).
Fig. 9.4 Permian scene starring the sail-backed synapsid Edaphosaurus boanerges. Early synapsids were many anatomical miles from their mammalian descendants, and probably bore greater superficial similarity to reptiles and early tetrapods. Meganeuropsis, Ophiacodon and Dimetrodon also feature. (J. Csotonyi)
Pelycosaur skulls are often rugose and bear small openings, suggesting relatively tight skin over much of the face, though sheathed teeth are predicted through phylogenetic bracketing and low numbers of labial foramina. They have not yet developed large openings for the tympanic membrane, and likely lacked externally visible ear anatomy. Detailed insight into pelycosaur skin is wanting, but at least some species had transversely broad, bony belly scales which, presumably, were covered in a thin epidermis – sediment imprints from a resting pelycosaur show the scale bands were prominent enough to leave traces in soft mud (Niedźwiedzki and Bojanowski 2012). Further evidence of scales has recently been found though soft-tissue preservation of a Permian varanopid, the entire body of which was covered in lizard-like epidermal scales (Spindler et al. 2018). For the time being, it is unclear how long scaly bodies were retained by early synapsids: some lineages may have lost them, and simply had tough, naked skin, well before the evolution of hair.
Later synapsids, such as the dinocephalians, dicynodonts and therapsids (a group that includes the famous gorgonopsians) have somewhat more mammal-like anatomy, but are superficially reptile-like in many respects (Fig. 7.13). A gradual reduction in tail development and expanded pelves suggest a move toward mammal-like rounded haunches and away from the integrated hindlimb/tail muscle configuration of pelycosaurs, though sprawling or semi-sprawling postures still typified synapsid limbs of this grade (Kemp 2009). We also see retention of ribs along the entire dorsal series, which may have limited trunk motion somewhat. The development of longer necks means these synapsids can be restored with a greater variety of head and neck poses than their pelycosaur-grade ancestors. Very little is known about the skin of these animals, though pelycosaur-like belly bone scales seem absent. Dicynodonts bear correlates for cornified sheaths on their beaks, and rugose skull textures are common throughout the grade: tight, tough facial skin may have occurred widely in these animals. The housing for tympanic membranes is expanded in some therapsids, and may reflect a visible ear opening. Many synapsids of this grade have upper caniform teeth which are sufficiently large to extend below the mandible when their mouths were closed, and they may be large enough to protrude from any lips that were present.
Fig. 9.5 Jurassic mammaliaform Morganucodon watsoni pursues a kidnapping Clevosaurus bairdi. Mammaliaforms and other derived cynodonts likely resembled small mammals in appearance, but might have differed in some aspects, such as limb carriage, the location of the ear, and the development of ear pinnae. (M. Witton)
Cynodonts and Mammaliaformes
It is probably among cynodont-grade synapsids where classically mammalian features became apparent, though the earliest members of this clade were probably still quite therapsid-like in their overall form. Progressively more mammalian features are seen in cynodont groups like the epicynodonts and, later, the eucynodonts. Their widened cheek regions, narrowed braincases and infraorbital foramina imply the development of mammal-like jaw musculature and muscular cheeks; their nasal openings become increasingly large and open, presumably reflecting development of cartilaginous nasal tissues; their posterior ribs are reduced, suggesting the beginnings of mammal-like spine flexibility; their limb girdles indicate semi-sprawling or erect postures; and their pelves elongate to accommodate large hindlimb muscles. Their tympanic apparatus was large and probably housed conspicuous ear tissues, but remained attached to the lower jaw. Studies on the internal cranial nerves of eucynodonts imply that whiskers and hair were present in Triassic members of this clade (specifically, in the prozostrodontians) (Benoit et al. 2016), but hair found in a Permian coprolite (Bajdek et al. 2016), which is probably from a cynodont, may indicate older eucynodonts (perhaps early probainognathians) had hair of some kind too. Overall, the appearance of derived cynodonts was probably not too dissimilar to that of Mammaliaformes, the group that includes the famous morganucodontids, the docodonts, and several other lineages closely related to the true mammals. Perhaps the only obvious differences for artists to note are the development of a true lumbar spine region – and thus greater trunk flexibility – and a capacity to place the feet and hands beneath the body, albeit with slightly bowed limbs. (Fig. 9.5)
Crown-group mammals
Bowed limbs seem to characterize the earliest grades of the true mammal group, including extant monotremes, and it may have only been among the eutherian mammals – the line that includes marsupials and placentals – that synapsids achieved fully upright limb postures. As discussed in detail in Chapter 7, current evidence indicates that eutherians may be the only synapsids to have conspicuous ear pinnae (Fig. 6.16), though our insight into this topic must be regarded as tentative at best.
Fig. 9.6 Fossils of crown-mammals – such as these Pliocene arctic camels – are among the most restorable of all extinct animals because our modern dataset of mammal anatomy is so varied and comparatively well-understood. (J. Csotonyi)
The glut of artistic reference material concerning living mammals makes extinct members of the crown group among the most readily restored of extinct species. When executed well, these palaeoartworks are some of our most credible insights into ancient life (Fig. 9.6). Though the well-understood anatomy of living mammals leaves relatively little ambiguity about the probable appearance of their ancestors, it is an oversimplification to suggest that all crown group mammals are easy to restore. Indeed, our relatively swollen dataset of extant forms demands a keen eye for detail and thorough research to hit the highest reconstruction quality. An understanding of living mammal musculature is essential for any reasonable attempt at restoring any fossil crown-mammal and avoids a common pitfall of prehistoric mammal palaeoart: undue monsterization by minimization of musculature, fleshy components and fur. Living mammals are, of course, poster-children for non-shrink-wrapped soft tissues, and this was almost certainly true of ancient mammals too.
Accessible artistic references for early synapsids are thin on the ground. Kemp’s (2009) The Origin and Evolution of Mammals is a terrific, well-illustrated and accessible overview of mammal evolution, with an emphasis on Palaeozoic and Mesozoic species. Prothero’s (2016) The Princeton Field Guide to Prehistoric Mammals is a useful introduction to modern ideas of mammal evolution and palaeobiology, but the illustrations – basic life restorations and photos of museum displays – make it a less useful artistic reference. Goldfinger‘s (2005) Animal Anatomy for Artists is an essential purchase for anyone wanting to understand mammal form and musculoskeletal anatomy.
Diapsids: reptiles, birds and their relatives
The diapsid lineage is the most popular palaeoart subject, containing a multitude of fantastic fossil reptiles including – of course – dinosaurs and birds. It is useful to know that modern diapsids are divided into two large groups, the lepidosaurs (lizards and allies) and the archosaurs (birds and crocodiles) (Fig. 3.7). Though similar, these groups bear distinctions in anatomy and function which need consideration from artists; an animal allied closer to lizards will look quite different to one allied to Archosauria. We will consider some of the most popular diapsid groups in more detail below, after outlining some general properties of the lepidosaur-line and stem-archosaur species.
Fig. 9.7 The spectacular Tanystropheus longobardicus. Elaborate neck tissues aside, much of the visible anatomy of this animal is typical of stem-archosaur reptiles. (M. Witton)
Ancestral diapsid gaits are variations on sprawling (Fig. 9.7). Many species – particularly those with large limbs and girdles – are capable of a semi-sprawled stance which lifts the body clear from the ground. Sprawling stances are often assumed to reflect slow, lumbering gaits but are actually capable of producing tremendous turns of acceleration and speed – there is no reason why palaeoart of early diapsids needs a low speed limit. Virtually all diapsids retain a heavy, hindlimb-linked tail, but it is not a given that the tail dragged. Modern reptiles which walk with their bellies clear from the ground often carry the tail base horizontally, after which only the distal portion falls to the ground, or is otherwise held elevated. Trackways are capable of recording tail drags, making them an ideal means to assess tail posture in these extinct reptiles (Fig. 4.17).
Fig. 9.8 Garjainia madiba, a stem-archosaur from South Africa. The large bosses projecting from the head of this species are based on epidermal correlates seen in fossil skulls of this species. (M. Witton)
Diapsids were ancestrally scaly, though the lepidosaur-line is characterized by overlapping scales rather than, in archosaurs, non-overlapping tubercles. Skin impressions are rare outside of certain groups, and we often must infer skin types instead of basing them on direct data. There is great variation in scale morphology across modern diapsids, ranging from huge tubercles and scutes (sometimes underlain by osteoderms) to soft, velvet-like integument composed of tiny scales: perhaps the same was true in the past. Many species have epidermal correlates on their skulls which allow for precise allocation of facial scales, bosses and hornlets. (Fig. 6.4, 9.8) Midline frills, dewlaps, ornament scales and enlarged fatty display structures are common features of living diapsids, and can be safely assumed for fossil species too. There is no evidence of fibres or filaments in diapsids outside of the pterosaur and dinosaur clade. Claws are present on most digits, but are absent on the fourth and fifth fingers of most archosaurs. Tooth-covering lips are predicted as ancestral for all diapsids by phylogenetic bracketing.
Mosasaurs
These Cretaceous marine reptiles have a long history in palaeoart, much of it portraying mosasaurs as crocodile-like serpents. Crocodiles are poor analogues because mosasaurs are, in fact, true lizards, most often (though not incontrovertibly) allied with monitor lizards and snakes. Like snakes, they have prominent palatal teeth and somewhat flexible lower jaws – both these features are conspicuous enough for artists to consider including them in the right artworks. Several lines of evidence indicate that their tongues were forked (Schulp et al. 2005), and – like all their modern relatives – their teeth were probably covered with simple lips. Their dorsally-open nares recall those of monitor lizards, and may have housed similarly substantial cartilaginous nasal tissues in life. Fossils show that their nostrils were positioned at the front of the narial opening.
Fig. 9.9 The large, globally-distributed mosasaur Prognathodon demonstrates many soft tissue characteristics of its group: a bilobed tail, expanded paddle tissues and essentially smooth skin. Its teeth are reconstructed as fully sheathed, as is the condition in all living lizards. (M. Witton)
Fig. 9.10 Soft tissue preservation in the mosasaur Prognathodon sp. (A) Extended paddle tissues of the left forefin. (B) Preserved soft tissue of the caudal fin. Redrawn from Lindgren et al. (2013)
Mosasaur life appearance has benefitted from several recent fossil discoveries and investigations (Fig. 9.9; Lindgren et al. 2009, 2013). Multiple records show that their bodies were covered by overlapping scales less than 10 mm across, though larger scales (up to 10–20mm) existed on the faces of some species. There is no evidence for frills, scutes or other elaborations of their skin despite their persistence in art. These are either based on mistakes (historic ideas about a midline frill were derived on displaced throat cartilage) or elaborate speculation. Mosasaurs were supremely adapted to aquatic life and, like comparably adapted modern species, their bodies were probably devoid of structures which would induce unnecessary drag, such as oversize scales or frills. The posterior margin of their paddles was extended by soft tissue, creating a superficially triangular profile, and their tails were also augmented with considerable fin tissues (Fig. 9.10). Classically, mosasaurs have been given eel- or crocodylian-like tail fins with long, low fins and a straight tail skeleton. However, at least some species had reversed heterocercal tails (Fig. 9.10), with downturned tail tips and large soft tissue dorsal lobes tissue. It is conceivable that the tails of different mosasaurs may have varied depending on their ecology, as discussed in Chapter 6. A further distinction from classic mosasaur reconstructions concerns restriction of vertebral motion to the tail. Rather than having serpent-like bodies, mosasaur torsos seem to have been stiff, like those of whales or fast-swimming fish. Melanosome data for the giant mosasaur Tylosaurus suggests it was at least partly covered with a dark colour (brown or black), and microscopic differences in scale anatomy indicate their dorsal regions were less reflective than the ventral, a condition which might be linked to a form of countershading (Lindgren et al. 2014).
Plesiosauria
These four-flippered Mesozoic marine reptiles were some of the first creatures to feature in palaeoartworks but our standard reconstruction approach needs modernization. Many familiar elements of plesiosaur palaeoart are at odds with fossil soft tissue data, modern muscle studies and flipper arthrology, as well as the generalities of vertebrate anatomy.
Fig. 9.11 Soft tissue outlines and musculature of plesiosaur forelimbs. Skeletal form and musculature based on Araújo and Correia (2015).
Fig. 9.12 Reconstruction of Jurassic plesiosaur Attenborosaurus conybeari. Note the tail fin, well-muscled torso and upper limbs, thick neck and flipper mobility. (M. Witton)
The ‘classic’ approach to plesiosaur flipper reconstruction – a tight, ‘oar-shaped’ profile which hugs the contours of the flipper skeleton – is erroneous on three aspects (Fig. 9.11). Reconstructions of plesiosaur forelimb musculature suggest large, powerful muscles around the shoulders, especially ventrally, implying that their flipper tissues were bulky and thick as they approached the body, and not like the shaft of an oar (Araújo and Correia 2015). In this respect their limb anatomy seems comparable to that of whales, seals and turtles. Equivalent muscle studies have not been done on the hindlimb, but the size of the pelvic girdle indicates similar muscular development. Secondly, rare examples of plesiosaur soft tissue indicate that the posterior edge of the flippers were extended from the skeleton in life (Figs. 8.9; 9.11). The exact flipper shape remains unknown, but it’s clear that most of the length was deepened by a posteriorly-situated wedge of soft tissue that tapered towards t he flipper tip. A third likely error is our restriction of flipper movement to only very tight arcs. Recent studies on their shoulder and pelvic limb joints (which were formed primarily of cartilage in many species) suggest they may have been capable of much wider arcs than previously realized, allowing for a much wider range of flipper poses and more dynamic depictions of movement (Fig. 9.12; Liu et al. 2014; Muscutt et al. 2017).
Plesiosaur neural spines are relatively tall compared to those of other tetrapods, indicating well developed vertebral musculature, including large muscles supporting the limb girdles (Figs. 4.13, 5.6). Their cervical vertebrae bear stout ribs and the posterior surface of their skulls have broad sites for muscle attachment. Combined with their tall cervical neural spines, these features suggest their necks should be restored with generous amounts of musculature. Fossils show that plesiosaur tails were deeply buried in soft tissue, being broad-based structures that tapered from the body to a point (Fig. 8.9; Frey et al. 2017) and/or bearing vertical caudal fins (Fig. 6.10). Caudal fins are predictable for many species based on osteological hallmarks – see Chapter 6 for more details.
The cervical and cranial anatomy of plesiosaurs is highly variable and close attention should be paid to body proportions and skull geometry. Some large-toothed species probably had teeth emerging from their sealed mouths, though many have dentitions of size and orientations that could be accommodated within oral sheaths. Long-necked species were seemingly capable of bending their necks into simple curves, and species with greater cervical counts could achieve tighter degrees of curvature. ‘S’ shapes and more advanced cervical contortion are unlikely, however. Surface-floating plesiosaurs should not be restored in swan-like poses; this violates understanding of neck arthrology as well as the physics of buoyancy.
Plesiosaurs are often restored with a ‘barrel-shaped’ trunk. This is appropriate for some taxa but not all (Fig. 4.10). The construction of their torso skeleton remains an area of investigation (and a challenging one because of the degree of cartilage likely apparent in this region) but detailed reconstruction attempts show that some species had dorsoventrally compressed bodies (O’Keefe et al. 2011). This is augmented in certain taxa, such as Tatanectes, by reduced rib curvature. Neural spine height is not always consistent along the vertebral column. Genera like Attenborosaurus have taller vertebrae towards the anterior end of the torso, making for a proportionally deeper shoulder region and a ‘tear-drop’ profile in lateral aspect. This was likely further augmented by deepened muscle tissue anchoring to this region, especially that stabilizing the pectoral girdle and supporting the base of the neck (Fig. 9.11).
Ichthyosauria
Ichthyosaurs have been restored as amphibious creatures, angry sea-serpents and – in their modern guise – as fish- or dolphin-like animals supremely adapted to life at sea (Fig. 9.13). Specimens of Stenopterygius with soft tissue outlines from Jurassic rocks of Holzmaden, Germany, are critical references for ichthyosaur artists as soft tissue preservation is rare in this group, despite the relatively ubiquitous nature of their fossils (Fig. 8.9; McGowan and Motani 2003). It is from these specimens that we know some ichthyosaurs had dorsal fins as well as reversed heterocercal, crescentic tail fins. Less commonly discussed is that they also reveal posteriorly-expanded flipper soft tissues akin to those of plesiosaurs and mosasaurs. Allegations that these body outlines were manufactured by preparators or happy accidents of bacterial decay have not been upheld.
Fig. 9.13 Reconstruction of Ophthalmosaurus icenicus, a large ichthyosaur with some of the largest eyes of any animal. As indicated by fossil remains, much of the underlying skeletal anatomy is obscured by the soft tissues and the visible eye area is much less than the full eyeball size. The dorsal and caudal fins are based on ichthyosaur fossils, but some liberty has been taken with their shape. (M. Witton)
A Stenopterygius-like body plan may have typified most Jurassic and Cretaceous ichthyosaurs, though the tail tips of these species are dipped to varying extents – their tail fins were probably similar to, but not exact copies of the morphology seen in Stenopterygius. Triassic ichthyosaur tails lack pronounced downturns however, and had tail fins composed of heightened neural spines and chevrons. This condition is quite varied across different Triassic species (McGowan and Motani 2003) and artists should pay close attention to the proportions and vertebral posture of the tails of their subject species. Tail morphology seems to correlate with swimming styles: long, low tails reflect a more flexible, eel-like tail motion, while crescentic tail fins occurred in species with largely stiffened bodies, where only the distal part of the tail provided thrust (Motani et al. 1996).
Reports of ichthyosaur skin impressions lack mention of epidermal structures, so they may have been smooth-skinned in life. Body outlines show that the skeletons of these animals were buried deeply under soft tissue, particularly in the dorsal region where about one fifth of the body height reflects soft tissue overlying the vertebral column (Fig. 8.9). This deep soft tissue probably continued over the back of the skull. Their skulls lack osteological correlates for epidermal structures and, although some species do have elevated numbers of pits towards the ends of their snouts, they lack sufficient numbers of foramina on their jaws to imply liplessness. Many ichthyosaurs likely had enormous eyes but, as we saw in Chapter 7, their sclerotic ring morphology reveals a relatively small area of visible eye tissue, located entirely within the centre of the sclerotic ring (Fig. 7.2).
Like all marine reptiles, ichthyosaurs lack a modern, accessible overview that can serve as a useful source for artistic reference. McGowan and Motani’s Handbook of Paleoherpetology: Ichthyopterygia (2003) is a very useful and well-illustrated overview of the group, but is written for ichthyosaur experts and sports a hefty price tag.
Pseudosuchia: crocodile-line reptiles
Pseudosuchians are those archosaurs more closely related to crocodylians than other archosaurs. Many, particularly members of the Crocodylomorpha, resembled crown-group crocodylians to greater or lesser extents but other pseudosuchians were radically different to their living relatives. Do not assume that fossil pseudosuchians were simply crocodiles stretched into different shapes.
Several gaits have been employed by pseudosuchians over time. Most seem to have been sprawlers but – like modern crocodylians – they may have been capable of ‘high walking’, where the limbs are positioned nearly vertically. Many Triassic species, including the aetosaurs and poposauroids, experimented with fully erect hindlimbs and even bipedality, and most seem to have retained plantigrade foot postures (Fig. 4.17). All pseudosuchians seem to have borne heavy, hindlimb-propelling tails typical of most other diapsids.
It is safe to assume that pseudosuchians were scaly, an idea bolstered by the presence of a paired series of osteoderms along the dorsal midline in most species. Some taxa, like the aetosaurs, elaborated this series into extensive armour plating and spikes across the body, neck, belly and tail (Figs. 4.6, 7.8). Pseudosuchian osteoderms are uniquely shaped and distributed in each lineage and often differ markedly from those of living crocodylians. The assumption that modern crocodile scutes can be dropped onto fossil pseudosuchians is a common mistake when restoring these reptiles. One function of these rows of osteoderms is stiffening the body (Salisbury and Frey 2000), so their presence is a cue not to overly contort body regions overlain by these structures.
Pseudosuchian skulls often contrast markedly from those of living crocodylians, being taller, bearing large openings, and lacking rugose, pitted surfaces. This almost certainly reflects a different facial integument to the tight, cracked skin of modern crocodylians: many pseudosuchians may have had more ‘conventional’ tetrapod facial tissues, including extra-oral tissues.
Some thalattosuchians, a radiation of marine pseudosuchians, bore several conspicuous soft tissue adaptations to aquatic habitats that are of note to artists. The most aquatically adapted, the metriorhynchids, lost their osteoderms (and, indeed, we might speculate that their scales were generally reduced, as seen in other marine reptiles) and developed reversed heterocercal tails with dorsal fin lobes supported by enlarged vertebral processes. Soft tissue preservation suggests they had tail tissues not dissimilar to those of mosasaurs (Fig. 9.10). The teleosaurids, being seemingly less adapted for a fully aquatic life, do not seem to have had this tail morphology, and probably had caudal tissues more akin to those of living crocodylians.
Truly artist-friendly overviews of pseudosuchians are lacking. Anatomy, Phylogeny and Palaeobiology of Early Archosaurs and their Kin (Nesbitt et al. 2013) contains excellent summaries of pseudosuchian anatomy and functional morphology, but is expensive and aimed at academics – lay readers may struggle with some text. Grigg and Kirshner’s (2015) Biology and Evolution of Crocodylians is an accessible and comprehensive overview of crocodylian biology, but is also expensive.
Pterosaurs
Flying reptiles can be the subject of truly amazing palaeoartworks, but their unusual, unfamiliar anatomy and strange proportions renders them challenging subjects for artists. Issues with proportions are perhaps most egregious problems of pterosaurs restorations, often because artists assume their proportions were like those of birds – they were not. At their most extreme (but not uncommon) proportions, pterosaur bodies are only thirty or fifty per cent longer than their humeri, while their heads can be at least two to three times as long. Wing bone proportions are another aspect to carefully restore as these are not only highly conspicuous, but also characteristic to different groups.
Flying reptiles have a long tradition of being restored with minimal soft tissue, to the extent of many restorations appearing highly skeletal. Muscle reconstructions suggest contrarily however (Fig. 5.6), and we should assume that – like living flying animals – pterosaur bodies and the proximal regions of their limbs were strongly muscled. Large muscle attachment regions on the posterior skull, well-developed vertebrae, and large pectoral girdles hint at muscular necks.
Fig. 9.14 An overview of pterosaur wing tissue distribution and internal anatomy. After Witton (2013).
The shape and extent of pterosaur wing membranes remains a matter of controversy, but enough specimens have been recovered with the main flight membrane (brachiopatagium) attaching below the knee to assume this was standard for most or all pterosaurs (Fig. 9.14; Elgin et al. 2011). The curvature of the trailing membrane edge and depth of the wing (the chord) remains uncertain. Further membranes were present between the wrist and shoulder (propatagium), and behind the legs (uropatagium). The shape of the latter varies depending on the shape of the fifth toe. Where the toe is short or missing, the membrane is small, narrow, and trails behind the leg between the ankle and hip. A large fifth toe correlates with an expansive membrane that occupied much of the space between the hindlimbs, spanning from toe to toe (Unwin and Bakhurina 1994). Pterosaur tails do not seem to have been incorporated into this membrane, but the number of specimens with bearing on this issue is small. Exceptionally preserved specimens show that pterosaur membranes were not simple sheets of skin but were reinforced in places and padded – via wedges of tissue and airsacs – to minimize the contours of the limb bones and musculature. We can assume that, in life, pterosaur wing anatomy was not sharply prominent beneath the membranes.
Fig. 9.15 Reconstruction of Tupandactylus imperator, one of the most visually impressive pterosaurs, demonstrating the rounded crest tissues known from pterosaur fossils. Also note the small body, the enormous skull, and the proportions of the limb bones. (J. Conway)
Fossils show that pterosaur bodies, necks, heads and the proximal parts of their limbs were covered in simple fibres which may, or may not, be homologous to dinosaur filaments and fuzz (Barrett et al. 2015). All pterosaur fibres – technically called pycnofibres – known to date are relatively short, show evidence of having been soft, and are sufficiently dense to suggest they obscured the underlying skin and smooth body contours. Epidermal correlates on pterosaur faces suggest many species bore beak-like keratinous sheaths, even if teeth were present, and such tissues also seemed to cover bony crests (Chapters 6 and 7). In some species, keratinous extensions of the jaw tips enhanced the jaw length, but these are only known from soft tissue preservation and do not seem to have obvious bony correlates. Soft tissue crests made of cornified skin seem to have been common across many, perhaps most, pterosaur species, and their presence can be inferred via the presence of low, striated bony crests over the rostrum (Fig. 6.7). Soft tissue remains show that these crests were vast, tall structures with rounded lateral profiles (Fig. 9.15). Foot pads and long claw sheaths are known from exceptional pterosaur fossils.
Grounded pterosaurs can be challenging to render to novices (Fig. 9.16). They are quadrupedal, with plantigrade, forward-facing feet and erect hindlimbs. The forelimb has complicated articulation. In some early species the forelimb sprawls, but in most pterosaurs it was erect or nearly so (Witton 2015). The forelimb joints are mechanically linked so that as the elbow opens and closes, the wrist follows suit, but the joints are not uniaxial and the forelimb does not fold nearly. As the elbow closes, the forearm rotates inwards while the wrist rotates out. Thus, unless the wing is completely extended (as in flight), the forelimb would not assume a completely straight profile in anterior view. Trackways show that pterosaurs walked on their three small fingers, which splayed sideways, and this demands that the forelimb is rotated somewhat away from the body during terrestrial locomotion (Fig. 4.1). The wing finger is stowed alongside the wing when walking, and fossils suggest it could be tightly pressed against the wing metacarpal. Tracks show that pterosaurs ‘paced’ when walking, moving the limbs on one side of the body, then the other, like giraffes or greyhounds. Trackways show that they were also capable of running.
Pterosaurs have been well-served with accessible and well-illustrated reviews in recent decades: Wellnhofer’s The Illustrated Encyclopaedia of Pterosaurs(1991), Unwin’s Pterosaurs from Deep Time (2005) and my own Pterosaurs: Natural History, Evolution, Anatomy (2013) are useful artistic references for this group.
Dinosaurs (including birds)
More time, energy and resources have been spent restoring dinosaurs than any other group of fossil animals. Despite their popularity, many artworks contain errors that are easily avoided with some basic knowledge of their fundamental biology.
Virtually all dinosaurs stood with erect limbs, with bipeds walking and running with their feet more medially placed than those of quadrupeds. The horned dinosaurs, or ceratopsids, are an exception in that their forelimbs were bowed somewhat outwards, though their hands were still placed beneath their bodies. Dinosaur backs and tails were orientated perpendicular to the ground, not arcing or sloping down as commonly depicted in the nineteenth and twentieth century. The angle of the tail base, though generally horizontal, is variable because of differences in the orientation of the pelvic region and shape of the tail vertebrae. In some groups, like the brachiosaurids, therizinosaurids and ceratopsids, the tail base slopes down somewhat, but in some sauropods, like the mamenchisaurids, the tail base arcs slightly upwards. Most dinosaur tails are expanded to accommodate large caudofemoralis muscles, and should have thick bases which grade into the back of the thigh. Paravian dinosaurs – the birds and ‘bird-like dinosaurs’ – reduced the size of this muscle, however.
Dinosaurs experimented liberally with gaits. The armoured dinosaurs (ankylosaurs and stegosaurs), some larger sauropodomorphs (including the true sauropods) and neoceratopsians (ceratopsids and close allies) were quadrupeds, while theropods, lither sauropodomorphs and small ornithischians were bipeds. The ornithopods and seemed unwilling to commit to either strategy, seemingly walking on four limbs but running on two. Bipedal dinosaurs are always digitigrade and would always have bent ankle and knee joints. In birds, and some paravians, the femur is almost horizontal and would move little during the walk cycle. In all other dinosaurs, the femur was more vertically inclined.
Dinosaur hand and foot morphology is highly variable with different numbers of digits, as well as much diversity in claws and hoofs, being present across the group. Most species were digitigrade, but sauropod feet were plantigrade. Footprints show the presence of large pads on dinosaur walking limbs, but sauropod hands are famous for lacking these. Rather, their hand prints are horseshoe- or crescent-shaped impressions, indicating a hollow, pad-less posterior surface. It is very common to see dinosaur hands with the palms facing backwards. This is incorrect: in most species the palms do not face downwards or backwards, but face one another (or nearly so). Dinosaur wrists and elbows did not allow the hands to rotate from this posture so, in short, bipedal dinosaurs should always look like they’re ready to burst into a round of applause. Quadrupedal dinosaurs have hands which faced somewhat more rearwards, but no species had palms that were fully rotated to face backwards (Senter 2012; Hutson 2015).
Fig. 9.17 The variable integument types of Kulindadromeus zabaikalicus, including multiple types of fibres, fibre/scale structures, and true epidermal scales. Integuments of this kind may have been common across Dinosauria, and confuse an already complex picture of dinosaur skin evolution. (E. Willoughby)
Fig. 9.18 Pterosaur and dinosaur integument data plotted to the traditional dinosaur tree (blocks), and predictions of integument types in different clades (circles). The shaded proportion of each circle shows the likelihood of scales, fibres or feathers being the ancestral condition for that group. This tree is one result from Barrett et al. (2015) – other models are available.
Dinosaur skin is among the most complex of any vertebrate. Scales, fibres and feathers of varying complexity are present across the group and even sometimes on single animals (Fig. 9.17). The ancestral state of dinosaur skin is controversial, in part because of continued uncertainty over the relationships of the major dinosaur groups and a lack of skin data from early dinosaur fossils. Their shared ancestry with pterosaurs (which were covered in epidermal fibres) gives a reasonable chance of dinosaurs being ancestrally fuzzy, though the fact that some dinosaur clades seem mostly or wholly scaly precludes any confident prediction of their ancestral skin type (Fig. 9.18, Barrett et al. 2015). Currently, it is difficult to say whether dinosaur skin was either ancestrally fibrous with multiple groups independently becoming predominantly or entirely scaly; or if dinosaur skin was ancestrally scaly, with multiple groups then independently developing fibres and feathers.
Presently, there is no evidence for fibres in ankylosaurs, hadrosaurs and ceratopsids despite extensive skin samples from these clades, and less exemplary skin data from sauropods, stegosaurs, abelisaurids, tyrannosaurids and a few non-coelurosaurian clades indicate scaly hides too (Fig. 9.18). Osteological correlates for scaly features – including hornlets, and rows or sheets of tubercles – are present on the skulls of ankylosaurs, ceratopsids and many theropods. By and large dinosaur body scales were very small, often being less than a centimetre across (Fig. 6.3). They are larger in sauropods, but still only a few centimetres in width. Larger, sparsely distributed ornamental scales are a common dinosaur feature, and are particularly common across their dorsal regions. Many dinosaurs adorned their dorsal midlines with ornaments of various kinds, such as frills, spikes or small osteoderms.
Fibres are currently known from several small-bodied ornithischian dinosaurs and extensively among the coelurosaurian theropods. These structures vary enormously in fine anatomy and not all variants on fibres and feathers known from fossil dinosaurs exist in living species. They include simple filaments and bristles, filaments erupting from scale-like plates, central filaments with radiating smaller filaments, and variants on pennaceous feathers (e.g. Prum et al. 1999; Godefroit et al. 2014). Soft tissue outlines of fuzzy dinosaurs show that, as with birds, their fibres were thick enough to significantly alter their body outlines. Some ornithischians had fibres on only parts of the body, such as the top of the tail in Psittacosaurus (Fig. 6.19), and the head, neck torso and proximal limbs in Kulindadromeus (Fig. 9.17). The rest of their bodies were covered in scales. Coelurosaurs seem to have been largely covered in fibres but many seem to have retained scales on their hands and feet, sometimes directly under fibrous regions. Tyrannosaurid skin is unusual in being – so far as it is known – exclusively scaly despite their ancestors, the tyrannosauroids, being covered in fibres (Fig. 6.2, Bell et al. 2017). Known skin impressions from other large dinosaurs also show scaly hides and this may, as discussed in Chapter 6, reflect a need to lose fibrous insulation for thermoregulatory requirements at larger body sizes (the largest fibrous dinosaur known is Yutyrannus, with an estimated mass of 1.5 tonnes). It is possible that other large coelurosaurs, like the giant therizinosaurids and Deinocheirus, had scale-dominated skin too.
Fig. 9.19 Restoration of Archaeopteryx lithographica showing the orientation of the dinosaur forelimb and distribution of feathers over the hands and fingers – the hand is not ‘free’ of the wing, but part of it. New information suggests the hindlimb feathers of this animal may be more extensive than shown here. (R. Groom)
Fig. 9.20 Structure of the paravian wing, and evolution of modern wing anatomy. (A) Schematic of feather attachment in paravian dinosaurs, showing how the large primary and secondary feathers are anchored to bones of the forearm and second digit. (B) Development of wing feathers across grades of bird evolution, from homogenous, multilayered complexes of narrow feathers in early paravians to differentiated feathers with reduced layering in true birds. (B) After Longrich et al. (2012).
Fig. 9.21 Reconstruction of the aquatic Cretaceous bird Hesperornis regalis. Other than some minor differences in feathering and the presence of teeth, it seems aspects of modern bird appearance were well established in the Mesozoic Era. Note the absence of cornified sheath tissues adjacent to the teeth. (J. Conway)
True feathers, or structures very like them, are seen in paravians, including the development of wing-like feather arrangements on their forelimbs (and sometimes hindlimbs). A common mistake is to show dinosaur fingers free from the wing apparatus, with their long primary feathers emerging from the wrist. Rather, as with living dinosaurs, the entire hand was buried in feathers, with the largest primary feathers emerging from the second finger (Fig. 9.19). The avian wing developed in stages from the first paravians, with some of the earliest variants being composed of numerous small, overlapping feathers instead of the larger, stronger wing feathers seen in modern birds (Fig. 9.20). Other than the presence of teeth and some differences in facial tissue, there may have been few major visual differences between the earliest fossil birds and their living relatives (Fig. 9.21).
Dinosaur skulls suggest a lot of variation in facial features. Sauropod skull architecture is extremely open and likely housed a lot of tissue now lost to time. Their nasal apparatus was especially enormous (and extended to the front of their snouts – only an internal nasal opening is found at the back of the skull – see Witmer 2001) and their nasal tissues may have added significant tissue to their faces (Fig. 7.7). The faces of horned dinosaurs, pachycephalosaurs and ankylosaurs are covered with epidermal correlates that allow for the location and size of sales and keratin sheaths to be predicted, though most art ignores this data (Figs. 6.6, 6.13, 6.15 and 7.1). Several theropods have rugose skull textures indicative of tough, maybe crocodile-like facial skin, though they lack the extensive foramina associated with beaks or liplessness. Lips seem likely for dinosaurs generally, and ‘pseudocheeks’ may well have been present in some species, though their presence is speculative at this time. Armour-like skin was present on the face of the abelisaurid Majungasaurus (Fig. 6.7).
Dinosaurs are well covered by accessible, artist-friendly literature. The Complete Dinosaur (Brett-Surman et al. 2012) is perhaps the most comprehensive single volume published on the group. A new edition is underway at time of writing. Gregory S. Paul’s The Princeton Field Guide to Dinosaurs (2016)is an extensive library of high-quality skeletal reconstructions only matched in breadth and quantity by Scott Hartman’s Skeletal Drawing website (www.skeletaldrawing.com). Van Grouw’s The Unfeathered Bird (2013) offers a unique and fascinating insight into the skeletons of living birds and their functional morphology, much of which aids understanding of fossil forms. Modern birds are also well served by dedicated illustration guides, and the advice therein has direct relevance to the reconstruction of extinct birds and feathered non-avian dinosaurs. Garner’s Wildlife Artists Handbook (2013) and Wootton’s Drawing and Painting Birds (2010) are recommended.
