When vertebrates ventured onto land, most of their systems (locomotor, respiratory, sensory) and structures (like the skin and axial skeleton) were not optimal for terrestrial life. Life in this new environment must have been fairly difficult for these animals, and the selective pressures leading to adaptation to life on land must have been fairly strong. Adaptation as a process is evolution influenced by selective pressures; the same word also designates the end result of this evolution, namely, a character that improves fitness. Thus, the loss of internal gills early in stegocephalians history can be seen as a terrestrial adaptation.
We must distinguish exaptation (often called preadaptation) from the related concept of adaptation. An exaptation is a structure that acquires a new function. Thus, the presence of a limb with digits may not be an adaptation for terrestrial life, because digits appeared in aquatic vertebrates, perhaps to walk underwater in a cluttered environment, such as a mangrove. In terrestrial stegocephalians, the limb acquired a new function, to walk on dry land. It is thus an exaptation. Another well-known example is the feather. It appeared in flightless dinosaurs, probably to provide thermal insulation, before adopting an exaptive role for flight. This chapter discusses the main adaptations and exaptations for terrestrial life in stegocephalians.
The conquest of land does not necessarily represent an evolutionary trend, which is biased directional evolution. Indeed, although some stegocephalians became terrestrial, some of their descendants reverted to an aquatic lifestyle. Since evolution is not deterministic (it has no purpose), it is usually not directional, at least in the long run. If there was an evolutionary trend towards terrestrial life in some taxa at some times (as in stegocephalians in the Early Carboniferous), there were perhaps trends in the other direction at others (as shown by the frequent returns to an aquatic lifestyle in stegocephalians from the Early Carboniferous to the present). Detecting evolutionary trends requires extensive data and statistical analysis. The conquest of land by vertebrates probably consists of only a few habitat shifts, so it cannot be subjected to such analyses, although transitions between habitats in various taxa have revealed intruiging patterns (Vermeij and Dudley, 2000).
Evolution of the Locomotor System
Since limbs with digits appeared among aquatic vertebrates in the Devonian, they may not be an adaptation to terrestrial life. However, some characters of more recent (Carboniferous or later) stegocephalians are probably adaptations to terrestrial locomotion. To determine which ones, we need to compare the limbs of primitively aquatic stegocephalians with those of the first terrestrial vertebrates. This is not easy to do, because we cannot observe the lifestyle of early stegocephalians, and any inferences that we make on this topic can be wrong. Furthermore, the primitive or secondary status of an aquatic lifestyle is sometimes difficult to assess, especially for Carboniferous taxa. Finally, the limbs are well known in only a small proportion of early stegocephalians. We will compare the limb of Acanthostega (a Devonian stem tetrapod) with that of Cacops (Fig. 6.1; an Early Permian temnospondyl) and assume that the first genus was primitively aquatic and that the second was terrestrial. Among the differences observed, only those that consistently appear between other primitively aquatic and terrestrial stegocephalians are likely to reflect terrestrial adaptations; other differences may reflect only the morphologies of Acanthostega and Cacops and be unrelated to adaptations to terrestrial locomotion.
Figure 6.1. Reconstruction of the temnospondyl Cacops. Cacops is generally thought to have been among the most terrestrial temnospondyls because its limbs were well ossified, and dermal scales on its back may have reinforced its vertebral column. However, this does not imply that it did not occasionally swim, as this reconstruction suggests. Drawing by Douglas Henderson initially published by Lauber (1996). Reproduced with permission.
Limb Evolution
After the transition from an aquatic to a terrestrial lifestyle, the limb became more flexible because well-defined articular surfaces allowing a greater range of flexion appeared. In Acanthostega, the relatively flat articular surfaces suggest that the limbs could not flex much at the wrist, elbow, knee, and ankle (Fig. 4.1). Such a limb must have behaved more like a swim paddle (as seen in extant cetaceans) than a typical tetrapod limb. In most Permo-Carboniferous stegocephalians, the limb extremities bear curved articular surfaces that suggest an increased mobility compatible with walk on emerged land (Fig. 6.2). In such forms, we see genuine elbows and knees that presumably allowed flexion or extension across a range of about 90 degrees. However, the wrists and ankles still involved articulations between several bones, and their flexibility may have resembled that of urodeles. Highly flexible wrists and ankles, as seen in birds and mammals, appeared much later.
The limb also became better ossified in terrestrial stegocephalians. In Acanthostega, the ends of long limb bones must have been covered in a thick layer of cartilage. The carpal and tarsal bones do not articulate closely with each other, which also suggests that a substantial amount of cartilage remained around these bones. In Cacops, the cartilaginous layer around these bones was much thinner, resulting in a stronger limb. Finally, the bones of the palm of the hand and sole of the foot, the metacarpals and metatarsals, became better differentiated. In Acanthostega, they do not differ from phalanges, whereas in Cacops, they are thicker and longer. This differentiation may reflect the appearance of a new flexible joint between the hand and foot proximally and fingers and toes distally. Such flexion is very useful on land, but not in water.
Girdle Evolution
The girdles are skeletal structures to which limbs articulate. The shoulder (or pectoral) girdle is composed of dermoskeletal and endoskeletal parts. The latter is called the scapulocoracoid and may be made of three bones—the scapula, coracoid, and precoracoid. The pelvic (hip) girdle is entirely endoskeletal. As with the limb, several changes are probably exaptations. This includes the loss of contact between the shoulder girdle and the skull (Fig. 6.3) through the loss of the dorsal dermal part of the girdle and of bones from the opercular series that covered the branchial chamber. This increased the mobility of the head through the appearance of a mobile neck; thus, shoulder and neck muscles provided a suspension to stabilize the head while the animal walked on land. This is important because head movements can impede both vision and equilibrium.
Figure 6.2. Skeleton of the temnospondyl Cacops. Each limb segment, as well as dermal and endoskeletal portions of the girdles, are shown in various shades of gray. Modified from Williston (1910).
Figure 6.3. Panderichthys, one of our closest finned relatives. The endoskeletal shoulder girdle is not visible in this view because it was very small and located on the internal surface of the dermal shoulder girdle. The bones that were lost in the first stegocephalians and thus increased the neck flexibility are identified by white dots. Modified from Vorobyeva and Schultze (1991).
Figure 6.4. Skeleton of the stegocephalian Ichthyostega. This aquatic stegocephalian dates from the Late Devonian and was found in Greenland. The same gray shades as in Figure 6.2 are used to identify skull, girdle, and limb segments. Modified from Jarvik (1955).
The limbs acquired a new orientation because sarcopterygian fins extend posteriorly (Fig. 6.3), whereas in the earliest still-aquatic stegocephalians, the limbs extend mostly laterally (Figure 4.1). The new orientation results partly from lateral rotation of the shoulder and hip sockets. Another exaptation is the sacrum, the structure composed of at least one pair of ribs that articulate directly with the pelvic girdle. This contact improves support of the body weight outside the water because it is energetically less costly than the indirect link through muscles and ligaments found in finned vertebrates.
Adaptations to terrestrial life include a reduction in size of the dermal shoulder girdle and an increase in size of the endoskeletal shoulder girdle (Figs. 6.2 and 6.4). The dorsal portion of the latter (the scapula) provides attachment sites for many of the limb muscles; its increase in size presumably reflects an increased development of these muscles.
Primitively aquatic Devonian sarcopterygians such as Eusthenopteron had neural arches (the part of the vertebra that surrounds the spinal chord) fused in the sagittal plane only at the dorsal tip of the neural spines. The arches lacked well-defined articular surfaces for articulation with each other. The vertebral centra (which surrounded the notochord) were composed of a variable number of bony elements—in Eusthenopteron, a median or paired intercentrum and a paired pleurocentrum—that were presumably linked to each other and to the neural arches through cartilage and connective tissues. The vertebral centra formed rings that surrounded a functional notochord retaining an important mechanical role. This was adequate in an aquatic environment, but it was probably not stiff enough to support the body outside water. The vertebrae of the first stegocephalians were barely sturdier. In Acanthostega, the neural arches were slightly more firmly fused in the sagittal plane than in Eusthenopteron. The zygapophyses—articular surfaces between successive neural arches—appeared, but they were poorly developed. The vertebral centrum was a little thicker but retained a configuration reminiscent of Eusthenopteron, and the intercentrum sometimes remained paired. This axial skeleton reflects a primarily aquatic lifestyle.
Figure 6.5. Diversity and evolution of vertebral centra. Vertebrae are organized from the most primitive (A) to the most recent (E), but they form two morphoclines. The first goes from Permo-Carboniferous temnospondyls showing the rhachitomous morphology consisting of a large, crescentic, ventral intercentrum and small, dorsal, paired pleurocentra (A) to Triassic stereospondyls (B) with a centrum composed of a cylindrical intercentrum. The second shows the progressive development of the pleurocentrum along the tetrapod stem from the primitive rhachitomous position (A) through embolomeres (D), which had cylindrical pleuro- and intercentra, and more crownward stegocephalians, such as seymouriamorphs and tetrapods (E), which had a large, cylindrical pleurocentrum and a reduced, crescentic intercentrum (the latter disappeared early in amphibian evolution and later in amniotes). The Early Carboniferous form Whatcheeria (C) does not belong to any of these main taxa but fits on the tetrapod stem between Ichthyostega and temnospondyls. It has crescentic intercentra and pleurocentra (one each per centrum).
The axial skeleton of Early Carboniferous stegocephalians was better adapted to terrestrial locomotion, even though many were probably amphibious rather than truly terrestrial. In these forms, the neural arch halves are firmly fused at the sagittal plane and the zygapophyses are well developed. The composition of the vertebral centrum was variable (Fig. 6.5). In temnospondyls, the centrum retains the same composition as in Eusthenopteron and Ichthyostega (Fig. 5.6A). In embolomeres, pleurocentra and intercentra are cylindrical (Fig. 5.6B).
The subsequent evolution of the vertebral centrum depends on the phylogeny. Under the recent alternative favored here (Fig. 5.22), the pleurocentrum becomes cylindrical and the intercentrum becomes very small and crescentic in the smallest clade that includes seymouriamorphs and tetrapods (Batrachomorpha). This morphology has been retained in some extant amniotes. The neural arch of batrachomorphs is firmly fused to the vertebral centrum, which reinforces the axial skeleton. Only in Batrachomorpha is the axial skeleton fully adequate for terrestrial locomotion.
Under the classical hypothesis of a monophyletic Lissamphibia nested within temnospondyls (Fig. 5.21), the increase in size of the pleurocentrum and its fusion with the neural arch along with the dwindling of the intercentrum have to be assumed to occur convergently among reptiliomorphs and in amphibians. The few apparently terrestrial temnospondyls have strengthened their axial skeleton through dermal scales that articulate with each other and with neural arches (DeMar, 1968) or through flanges on the ribs that overlap the next posterior rib and strengthen the rib cage, as in birds. Dermal scales articulating with each other and with neural arches also appeared in the most terrestrial embolomeres, found in the Late Permian and Triassic of Russia (Golubev, 1998). No comparable axial skeleton exists in extant tetrapods.
Lung Evolution
Our aquatic ancestors were already able to breathe air since they had inherited lungs from the first osteichthyans (see “Homology and Analogy: Lungs, Swim Bladders, and Gills” in Chapter One). This suggests that some of our aquatic ancestors lived in poorly oxygenated water and used their lungs to extract oxygen from the air. For such animals, adapting to breathe only air on emerged land was far easier than it was for teleosts that have more recently become amphibious and which had transformed their lungs into a swim bladder (see “Homology and Analogy: Lungs, Swim Bladders, and Gills” in Chapter One).
In terrestrial vertebrates, the lung became more complex. Evolution of this complexification is poorly documented because the lung does not fossilize. Alveoli—small subdivisions of the lung—are present in primitively aquatic osteichthyans (all actinopterygians and dipnoans), but the alveoli are more numerous and smaller in tetrapods, expanding surface/volume ratio and improving lung efficiency. An indirect clue about lung complexity and its increasing importance in breathing is provided by relative rib length. In primitively aquatic vertebrates, ribs are generally very short, whereas in tetrapods that use costal ventilation, as in many amniotes, they are long and curved because they, along with the cartilaginous sternal ribs on the ventral surface, have to encircle the thorax. Relative rib length suggests that pulmonary ventilation was not accomplished primarily through the ribs in Devonian stegocephalians. Predominant costal lung ventilation may have appeared in the smallest clade that includes temnospondyls, embolomeres, seymouriamorphs, and tetrapods. This includes most post-Devonian stegocephalians and may have disappeared in some temnospondyls (which is not surprising, given that many of them returned to an aquatic lifestyle, especially in the Triassic) and in amphibians very early in their history. Thus, the buccopharyngeal breathing of lissamphibians, which does not involve ribs, is probably not a primitive feature; it presumably represents an autapomorphy of this taxon.
The history of the lissamphibian buccopharyngeal pump is difficult to reconstruct because it leaves even less fossil evidence than costal ventilation. Nevertheless, a few clues suggest that most early stegocephalians lacked such a pump (Gans, 1970), except for temnospondyls, whose large interpterygoid vacuities may have been involved. In such taxa, air enters the mouth when the gular pouch (between the lower jaws) expands. Then, the mouth and nares are closed and air is forced from the mouth into the lungs by contraction of the gular pouch. Expiration is simpler; trunk muscles contract and expel the air. Such a buccal pump also exists in many reptiles, but it has only an olfactory role.
The stegocephalian choana (internal naris) was originally the posterior external naris found in most other osteichthyans. This naris was initially located on the lateral surface of the snout. It provided an outlet for the water that entered the nasal cavity through the anterior naris when the animal swam. The motion of a swimming animal creates a water current in the nasal cavity, which allows the animal to perceive odors, but such animals cannot voluntarily inhale water through the nose because the nasal cavity lacks a connection with the buccal cavity. In the first tetrapodomorphs (the largest clade that includes tetrapods but neither dipnoans nor actinistians) from the Early Devonian, the posterior external naris had migrated ventrally and was located on the lower edge of the snout, next to the mouth. In all other tetrapodomorphs, this opening has migrated onto the palate. This allows direct inhalation of air or water and, thus, the ability to smell without having to move or open the mouth. This may have been especially useful if our ancestors were aquatic ambush predators. We think of the choana as a respiratory structure, but its first function may have been primarily olfactory. The choana was never important as an underwater breathing structure because water is more viscous than air; a much larger opening is needed to efficiently ventilate the gills.
Loss of Gills
One of the main adaptations to life on emerged land was the loss if the internal gills. These structures can breathe in air, but they dehydrate rapidly because the branchial chamber is largely open to the external environment through the gill slits. It is thus not surprising that they were quickly lost towards the end of the Devonian or the Early Carboniferous. And yet, much information (taphonomic, presence of a lateral-line organ, low degree of ossification of the endoskeleton) suggests that most stegocephalians remained mostly aquatic. Thus, the internal gills may have been lost even before stegocephalians became terrestrial. This paradox raises the possibility that stegocephalians lived in oxygen-depleted water, in which case the gills may have been disadvantageous because the oxygen could have flowed from the gills into the water, rather than the other way around.
Figure 6.6. Shoulder girdle of Acanthostega. Postbranchial lamina of the cleithrum (dark gray) of Acanthostega, one of the last stegocephalians that possessed internal gills. The dermal shoulder girdle is shaded a little lighter, and the endoskeletal girdle, lighter still. Scale: 1 cm. Modified from Coates and Clack (1991).
The habitat of Paleozoic stegocephalians remains partly hypothetical, but the loss of internal gills is suggested by the disappearance of the postbranchial lamina of the cleithrum (Fig. 6.6), a structure of the dermal shoulder girdle that normally contributes to the posterior wall of the branchial chamber. This structure was present in the first sarcopterygians and persisted among several Devonian and a few Carboniferous stegocephalians. Another osteological character suggesting an early loss of internal gills is the disappearance of grooves on the ceratobranchials (part of the visceral skeleton; Fig. 6.7). These probably housed arteries that carried blood to the gills, where it was oxygenated. These grooves are present in Acanthostega and Ichthyostega (both from the Devonian), but they are absent from most post-Devonian stegocephalians.
Figure 6.7. Branchial skeleton of Acanthostega. Acanthostega ceratobranchials showing the grooves that must have housed branchial arteries. All visceral skeletal elements are shaded. Scale: 2 mm. Modified from Coates and Clack (1991).
PARTS: A, dorsal view of the skull and associated branchial elements. B, ventral view. C, transversal section of a ceratobranchial the groove that sheltered the branchial artery.
Skin Breathing
Many extant amphibians breathe through the skin. This often provides a substantial amount of gas exchange in addition to the gills and lungs, but in some taxa, such as many plethodontid salamanders, both lungs and gills have disappeared and the skin is solely responsible for gas exchange. Exclusively cutaneous respiration is possible only in organisms of small body size because the surface/volume ratio decreases with body size. Thus, most amphibians with exclusively cutaneous breathing are less than 10 cm in total body length, with the exception of the only known lungless gymnophionan, Atretochoana eiselti, which reaches more than 70 cm in length (Wilkinson and Nussbaum, 1997). Since many Paleozoic stegocephalian species measured much more than 10 cm (Laurin, 2004), this mode of breathing must not have been widespread. Romer (1972) had already concluded that cutaneous breathing was not important in early stegocephalians, because dermal scales, present in many early stegocephalians, prevented cutaneous gas exchange. In fact, these scales are usually located in the dermis, under the skin, and vascular canals may have been present on the external surface of such scales, thus facilitating gas exchange.
Primitively Aquatic Vertebrates
The skin of primitively aquatic vertebrates provides a partial barrier to water exchange because vertebrates living in fresh water are hypertonic compared with their environment, whereas marine vertebrates are generally hypotonic compared with sea water. The integument in most primitively aquatic osteichthyans possesses various features to reduce permeability, such as dermal scales covering most the body and a thick layer of connective tissue (Bond, 1979). Thus, eel skin is fairly waterproof, but it represents 10% of the animal’s mass. In these vertebrates, most water exchange takes place at the gills; these cannot be waterproofed, since this would prevent gas exchange.
Freshwater vertebrates must eliminate the excess water that keeps seeping into the body. This is accomplished through the glomerula in the kidneys. Ions and other substances that the organism must keep are then resorbed by the kidney tubules and the bladder. Osmoregulation has been intensively studied in several actinpterygian species. In marine vertebrates, water tends to move out of the body because sea water is hypertonic relative to the body. These vertebrates drink seawater. This is absorbed by the digestive tract, and such animals excrete monovalent ions (for instance, Na+, Cl-, etc.), especially through the gills and, to a lesser extent, through specialized cells in the skin on the anterior part of the body. Divalent ions (like Ca2+) are eliminated mainly by the kidneys. Glomerula are smaller and less numerous (if present at all) in marine vertebrates because the body does not have to eliminate excess water. They thus produce much less urine than their freshwater relatives, but it has a much higher concentration of divalent ions. This osmoregulatory mode characterizes most marine actinopterygians. However, chondrichthyans and the coelacanth retain urea in the blood, which increases its osmotic pressure to match (and even exceed slightly, in chondrichthyans) that of sea water, which greatly facilitates osmoregulation and helps maintain water balance. The taxonomic distribution of urea retention does not enable us to determine if our marine ancestors ever possessed it.
The habitat of our aquatic ancestors is still controversial, but many paleontological data suggest a marginal marine origin for stegocephalians. This would have facilitated the invasion of land because marine vertebrates produce little urine, which enables water conservation on land.
The First Terrestrial and Amphibious Vertebrates
The skin adaptations of early stegocephalians to face the relative aridity of the terrestrial environments are poorly known because skin does not fossilize, except for the ossified scales. We know that many stegocephalians retained ossified scales on the ventral surface of the body, but fewer species had such scales also on the flanks and the back. By the end of the Devonian, stegocephalians had greatly reduced the dermal scale covering of the body.
To study other aspects of water-balance evolution, we must turn to extant taxa. Lissamphibians are often considered incompletely adapted to a terrestrial lifestyle, but this is an oversimplification. Lissamphibian skin is often highly permeable, but this is not necessarily disadvantageous. Indeed, lissamphibians absorb most of the water they need through specialized surfaces of the skin. Other lissamphibians, such as the tree-dwelling anuran Phyllomedusa sauvagei, produce waxy secretions that waterproof their skin (Pough et al., 2004). Other anurans, such as Cyclorana (Hylidae), Limnodynastes, and Neobatrachus (Myobatrachidae), can secrete a mucous cocoon to prevent water loss in the dry season. These lissamphibains can resist dehydration as well as most reptiles and mammals. Their adaptations differ greatly from those of amniotes, which suggests that they appeared convergently. Consequently, the first tetrapods may have had fairly permeable skin.
The glomerula are generally well developed in extant amphibians, which produce much urine, but this may be linked with their frequently amphibious lifestyle. In some desert-dwelling ampibians, urine is retained in a huge bladder and can be reabsorbed in the dry season to meet the animals’ water requirements.
Amniotes
Amniote skin is generally more waterproof than amphibian skin. The main barrier against dehydration is provided by lipid layers. In this respect, there is little difference between mammalian and reptilian skin. It is often wrongly stated that mammals have a more permeable skin than reptiles because they can lose much water in the form of sweat. However, sweating is an active thermoregulatory mechanism—evaporation of sweat cools off the body—and sweat is produced by specialized glands. Sudation is thus not to be confused with passive water evaporation through the skin; such evaporation also occurs in mammals, of course, but at about the same rate as in reptiles.
Most reptiles produce little urine and have poorly developed glomerula. This is probably linked with their excretion mode of nitrogenous waste products as uric acid, which is fairly insoluble and precipitates in the urine. Excreting nitrogenous wastes thus requires less water in reptiles than in urea-producing mammals. In most mammals and in some birds, glomerula are well developed and produce copious urine, but most of it is reabsorbed in a new portion of the kidney tubules called Henle’s loop. This minimizes water loss linked with nitrogenous waste excretion. Since the kidney does not normally fossilize, and since it evolves quickly in response to habitat shifts, it is difficult to reconstruct its history.
Lateral-line Organ
This organ is composed of neuromasts (ciliated cells), generally located in canals in the dermis, along the flanks and in canals or grooves of some cranial bones. It is present in all primitively aquatic craniates except hagfishes (Janvier, 1996). In lampreys and aquatic lissamphibians, the neuromasts of this organ are located on the skin surface. These cells are used to detect motion in water. They help predators to detect their prey, and the prey to avoid predators. They also facilitate schooling behavior in teleosts; they can stay grouped and avoid collision even if blinded because of the information provided by the lateral-line organ.
A lateral-line system works only in water, and since it dehydrates quickly, it was probably quickly lost in terrestrial and amphibious tetrapods. In lissamphibians, the organ is present in aquatic larvae and in the adults of several aquatic taxa, but it disappears at metamorphosis in species that have terrestrial or amphibious adults. It has never been observed in amniotes, even in early development.
Fossils show that the organ of the first tetrapodomorphs was located in canals in dermal bones and in flank scales, but it migrated to a more superficial position, in grooves at the surface of dermal bones, in early stegocephalians. This migration may be explained by the function of the canals in which the organ is located in most other gnathostomes; they seem to filter the noise generated by the animal as it swims. In teleosts, a similar superficial migration is often associated with neoteny and is generally disadvantageous, but the problem can be minimized by adopting an ambush hunt strategy (Montgomery and Clements, 2000). As long as the animal is immobile, the superficial position of the neuromasts is not disadvantageous. It is thus possible that early stegocephalians were ambush predators.
Since the lateral-line organ does not always leave traces on the skeleton, we can follow only part of its evolution in stegocephalians. We know that it persisted in many temnospondyls (including the last species in the Cretaceous) in various embolomeres, in seymouriamorph larvae, and in the adults of a few Paleozoic amphibians. We have no evidence that it ever existed in reptiliomorphs, even though it was probably present in the first, yet undiscovered, members of this clade.
Ear
ANATOMY AND FUNCTION
The ear is comprised of three distinct parts that do not have the same taxonomic distribution (or geological age). The oldest, found in all craniates, is the inner ear. The middle ear is unique to stegocephalians, and the external ear is found only in mammals.
The inner ear was initially involved mostly in equilibrium, a function that it retains in all craniates; in all except tetrapods, this remains also its main function. In tetrapods (and convergently in ostariophysean teleosts), the inner ear acquired a new auditory function without losing its equilibrium function. But without a middle ear, it is sensitive mostly to sounds traveling in water and in the ground; in air, only relatively low-frequency (less than 1000 Hertz) sounds are easily heard. As for the lateral-line organ, its sensory cells are neuromasts.
The middle ear is composed of one ossicle (the stapes) or more (in mammals, there are three). This structure facilitates reception of airborne sounds to the inner ear. It results from the transformation of the hyomandibular, which is involved in jaw suspension in most gnathostomes, into an ear ossicle (the stapes). To understand its evolution, we must first survey the types of jaw suspensions.
We recognize three main types of jaw suspension (Fig. 6.8): amphistylic, hyostylic, and autostylic. Amphistylic suspension is probably the most primitive because it is found in the first chondrichthyans and osteichthyans, and it persists in gymnophionans and urodeles. It was also found in Paleozoic amphibians and in the first amniotes. In these taxa, the palatoquadrate (the element that composes the dorsal part of the mandibular arch) articulates directly with the neurocranium (braincase) through the basicranial articulation, which is often located close to the orbit, and indirectly with the neurocranium through the hyomandibula (or stapes). Primitively, both articulations may have been slightly mobile. They were involved in, among other things, the buccal pump of actinopterygians and primitively aquatic sarcopterygians.
The hyostylic suspension appeared through the disappearance of the basicranial articulation, resulting in a greater mobility of the palatoquadrate and, hence, of the upper jaw. It is found in most elasmobranchs (sharks and rays). In this type of mandibular suspension, as in amphistyly, the hyomandibular plays an essential role in mandibular suspension, which imposes mechanical constraints preventing this structure from evolving in response to evolutionary pressures to improve hearing.
In the third kind of mandibular suspension, autostylic, the palatoquadrate articulates with (and often fuses to) the neurocranium only through the basicranial articulation. This frees the stapes from its support function, which was probably crucial to allow the appearance of middle ear. To efficiently transmit high-frequency airborne sounds, the stapes must be very light, which is incompatible with a role in mandibular suspension. All stegocephalians with a tympanum or ear drum (a tympanic ear), such as extant reptiles, mammals, and anurans, possess an autostylic suspension. Mammals are a partial exception because the stapes retains an articulation with the posterior part of the palatoquadrate, the incus, but these minute elements form part of the middle ear and not the functional jaw. A tympanic ear enables all these animals to hear high-frequency (over 1000 Hertz) airborne sounds well.
Figure 6.8. Mandibular suspension. The three types of jaw suspension (amphistylic, hyostylic, and autostylic) are defined by the number and position of the articulations (not necessarily mobile) between the palatoquadrate (element of the upper part of the mandibular arch) and the neurocranium.
TYMPANIC EAR
The tympanic ear is necessary only in the air because sounds transmitted in water can easily reach the water-filled inner ear when the animal is submerged. However, the bulk of the sound waves transmitted in air would be reflected at the air/water interface (the vertebrate inner ear is always filled with water) without a force-amplification mechanism, because water is about 1000 times more dense than air. This means that the animal would be nearly deaf. For sound waves to be transmitted efficiently into the inner ear, their force is amplified through two mechanisms (Fig. 6.9). The first is a great size ratio (between 10 × and 40 ×) between the surface of the tympanum and the oval window, into which the footplate of the stapes rests. Vibrations of the stapes induced by airborne sounds are transmitted to the fluid-filled inner ear through the oval window. The second is a lever system, which decreases movement amplitude and increases its strength by the same factor. The lever is different (not homologous) in several taxa, which is not surprising given that the tympanum appeared several times (Fig. 6.10). In squamates, the extracolumella (which is generally cartilaginous) plays this lever role. The distal end of its pars inferior (ventral part) is located in the center of the tympanum and moves with it when sound waves reach it. The extracolumella is fixed to the skull at the tip of its pars superior (upper part), which is immobile. The pars inferior is usually three times longer than the pars superior (Werner and Igic, 2002, which means that the motion amplitude is decreased by a factor of about 4. Between this lever effect and the amplification provided by the ratio between the surface of the tympanum and of the oval window (10 × to 40 ×), the total amplification factor is from 40 to 160. This allows adequate transmission of sounds from air to the aqueous solution that fills the inner ear.
Figure 6.9. Middle ear. Middle ear of a gecko (squamate) Ptyodactylus guttatus. Modified from Werner and Igic (2002).
It is fairly advantageous for terrestrial vertebrates to hear high-frequency airborne sounds—these include sounds emitted by predators, potential prey, and the vocalizations of conspecific individuals, such as calls emitted by males to call females or to delimit their territory. It is thus not surprising that the tympanum appeared fairly quickly in terrestrial taxa (Fig 6.10). These include anurans and mammals, which acquired the tympanum convergently, probably in the Triassic, or perhaps as late as the Jurassic in anurans. Among reptiles, the tympanum appeared at least once, but more likely twice, if turtles are not diapsids (their phylogenetic position is controversial). The appearance of the tympanum in reptiles probably occurred in the Late Permian (Müller and Tsuji, 2007; Senter, 2008). There was probably another appearance in seymouriamorphs in the Late Carboniferous, and perhaps one in some terrestrial temnospondyls at about the same time. The tympanum may thus have appeared up to six times in stegocephalians.
Figure 6.10. Middle ear evolution. Evolution of the middle ear in stegocephalians according to an old phylogeny (to the left) and to a recent one (to the right). Modified from Laurin (1998b).
The tympanum does not fossilize, but middle ear evolution is fairly well known because the stapes, despite its slenderness, is often preserved. When it is preserved, the stapes gives indirect clues about the presence of a tympanum because the latter can vibrate efficiently only if the stapes is light and has low inertia. Thus, in most extant vertebrates, a tympanum is associated with a stapes with a diameter less than 1 mm. On the other hand, tetrapods lacking a tympanum usually have a massive stapes involved in mandibular suspension (Fig. 6.11).
Figure 6.11. Stapes. Massive stapes of the temnospondyl Iberospondylus schultzei (Late Carboniferous), which probably lacked a tympanum (A to E), and of Rhinella marina (formerly known as Bufo marinus), one of the largest anurans (F to J). Both are left stapes, in anterior (A, F), dorsal (B, G), posterior (C, H), ventral (D, I), and medial (E, J) views. Note that the stapes is much more gracile in R. marina, which has a tympanum, than in I. schultzei. Modified from Laurin and Soler-Gijón (2006).
The skull usually has an otic notch that supports the tympanum, whenever the latter is present (Fig. 6.12). However, the presence of such a notch is insufficient to infer the presence of a tympanum because, in some cases, the notch is associated with a massive stapes. By combining both criteria (size of stapes and presence of the notch), the presence of a tympanic middle ear can be inferred.
Eye
The eye probably evolved less than the lateral-line organ when vertebrates invaded land because it did not require drastic modifications to work in air. The main innovations that improved its performance on land are the lachrymal glands that keep the eye wet and a change in shape. The latter is required because the refraction index of air is 1, but that of water is 1.33. Without changes, an eye adapted to underwater vision gives a fuzzy image on land, and vice versa (as we experience when we dive without goggles). Another change, which probably occurred later, is the appearance of an eyelid to protect the eye from dehydration and dust. In amniotes, the main eyelid (the most mobile one) is the upper eyelid, whereas in lissamphibians, it is the lower eyelid. This suggests that eyelids appeared convergently in these taxa (or at least its mobility did). Its evolution is difficult to reconstruct because it generally does not fossilize. Its presence seems to be strongly linked with habitat (at least in lissamphibians) because gymnophionans, which are aquatic or fossorial, and pipids, which are aquatic, lack eyelids.
Figure 6.12. Otic notch and tympanum. Seymouria baylorensis skull showing the otic notch, which probably supported a tympanum. Modified from Laurin (1996).
Olfactory Organ
Migration of the posterior external naris into the mouth, thus becoming the choana, facilitated olfaction as well as breathing. Another noteworthy change is the disappearance of many taste buds, which cover much of the skin in aquatic vertebrates. In tetrapods, taste buds have become restricted to the mouth because they work only in water. Various primitively aquatic vertebrates can detect electrical fields. This is useful to detect prey and predators alike. Thus, sharks can detect immobile prey, even if buried under sediment. This perception rests on special cells, called the ampullae of Lorenzini, which can also measure ambient temperature. They resemble neuromasts, but lack large cilia. Similar structures persist in some aquatic urodeles and gymnophionans, but are absent in anurans, in amphibious or terrestral lissamphibians, and in amniotes. The taxonomic distribution of these ampullae in early stegocephalians is poorly known because they leave few traces, but small solated pores on the skull of the seymouriamorph Discosauriscus have been interpreted as traces of these ampullae (Klembara, 1994).
The transition from an aquatic to a terrestrial lifestyle was facilitated by several exaptations, as early as the Late Devonian. The shoulder girdle became detached from the skull through the loss of opercular bones, and this gave a new flexibility to the neck, which could thus stabilize the head while the animal walked. The sacrum transmitted the weight of the animal on land at a lower energetic cost and with less muscle fatigue than would have been possible without it. The axial skeleton was reinforced by zygapophyses, new articular surfaces between successive neural arches. The choana appeared by migration of the posterior external naris into the mouth.
This habitat shift also triggered adaptations that appeared in the Carboniferous or, in some cases, later still. Limbs became more flexible, with better-defined articular surfaces, especially in the knee and elbow and later in the wrist and ankle. The axial skeleton was reinforced through consolidation of the vertebral centrum and its fusion with the neural arches. Internal gills, which dehydrate quickly on land, were lost. Costal pulmonary ventilation appeared in stem tetrapods and was subsequently lost in lissamphibians. The lateral-line and electrosensory organs were lost and the tympanum appeared, in both cases several times. The tympanum seems to have appeared well after the terrestrial lifestyle. Lachrymal glands and eyelids appeared (for the latter, probably convergently in lissamphibians and amniotes). Several of these characters leave little or no fossil evidence, which hampers evolutionary studies.
