The ear of stegocephalians is divided into the inner ear (the only part found in primitively aquatic vertebrates such as hagfishes, lampreys, chondrichthyans, and actinopterygians), the middle ear, and the external ear. The external ear is composed of structures (such as the pinna in mammals) that help to channel the sounds to the middle ear. A true external ear does not exist in all stegocephalians; it is unique to mammals, among extant vertebrates. The middle ear consists of a tympanum (ear drum), ossicles that transmit the sounds from the tympanum to the inner ear, and associated structures (nerves, blood vessels, etc.). The tympanum is not always present, but it is required to receive high-frequency, air-borne sounds. However, tetrapods lacking a tympanum (all salamanders, caecilians, some frogs, Sphenodon, and snakes) can still hear seismic vibrations (pressure waves transmitted through the ground). The inner ear consists in a perilymphatic and an endolymphatic system. The vibrations of the stapes are transmitted to the perilymphatic system through the fenestra ovalis (a hole in the otic capsule in which the footplate of the stapes fits), and they are then transmitted to the endolymphatic system, where ciliated cells detect them.
The main middle ear ossicle is the stapes, and it is often the only ossicle involved in transmitting sounds from the tympanum to the inner ear. However, many lissamphibians have an opercular bone between the base of the stapes and the otic capsule, and a cartilaginous extracolumella is present in several groups. In mammals, the malleus and incus (homologous with the articular and the quadrate, respectively) are located between the tympanum and the stapes.
Evolution of the middle ear
The middle ear is an interesting structure because its morphology allows the auditory acuity of many tetrapods to be estimated. Studying the anatomy of the middle ear of extant taxa may not be critical because the auditory acuity of modern tetrapods can be measured more accurately by physiological studies, but the middle ear gives us much-needed clues about hearing in extinct taxa. Unlike the outer and inner ear, the middle ear fossilizes readily and is easily studied (the inner ear is also preserved, but it is a cavity in the braincase and is difficult to study in extinct taxa). Despite this, the evolution of the middle ear has puzzled zoologists for more than a century. Basic questions about the homology of the tympanum of anurans (frogs and toads) and the tympanum of various groups of amniotes can not always be answered confidently (Lombard and Bolt, 1979).
In many tetrapods, the tympanum is supported by a large emargination at the back of the skull. A similar emargination called an otic notch is also found on the skull of most early stegocephalians, and it was believed to have supported a tympanum. This suggested that the tympanum had appeared very early in the evolution of vertebrates (in the Devonian) and that the tympanum present in many groups of tetrapods was a primitive structure inherited from an early ancestor. This theory required that most early stegocephalians have a slender stapes (the bone that transmits sounds from the tympanum to the inner ear) because a massive stapes is not efficient in a tympanic middle ear. Until fairly recently, most known stapes of stegocephalians (temnospondyls and seymouriamorphs) were indeed fairly slender.
Recent discovery of massive stapes in Acanthostega, early temnospondyls (colosteids), and embolomeres suggests that these taxa did not have a tympanum (Clack, 1989). The otic notch of these taxa may instead have had a respiratory function and have housed an open spiracle (the first gill slit in fishes). Therefore, the tympanum probably appeared later than previously believed. Laurin (1998) suggested that the tympanic middle ear probably appeared convergently three to six times in stegocephalians: in anurans (frogs), in synapsids (mammals and their extinct relatives), in diapsids (lizards, crocodiles, and birds), perhaps separately in anapsids (turtles and their extinct relatives), if turtles are not modified diapsids, as suggested by deBraga and Rieppel (1997), probably in seymouriamorphs, and possibly in some temnospondyls. However, if lissamphibians are descendants of temnospondyls, the tympanum appeared only a maximum of five times in vertebrates. There is a general consensus that the tympanum of extant tetrapods appeared independently in all the major groups: anurans, mammals, and reptiles (once or twice in this group, depending on the affinities of turtles). However, the absence of a tympanum in caecilians and salamanders may be primitive, if lissamphibians are derived from lepospondyls, or derived (a reversal), if lissamphibians are descendants of temnospondyls. Definite conclusions on this question must await a consensus on the phylogeny of stegocephalians and on the eventual presence of a tympanum of temnospondyls, a particularly controversial issue. Indeed, a recent description of the Late Carboniferous temnospondyl Iberospondylus schultzei (Laurin and Soler-Gijón, 2006) suggests that the otic notch of that taxon was occluded by a bony lamella. This, along with the robustness of the stapes, suggests that I. schultzei lacked a tympanum, but this conclusion does not necessarily apply to all temnospondyls, and not necessarily to those that are thought to be closely related to lissamphibians.
Lombard and Bolt (1979) studied the homology of the tympanum and the middle ear in living amphibians and in amniotes. They concur that the tympanum of frogs is not homologous to the tympanum of any amniote group. Lombard and Bolt based their conclusions on a detailed anatomical study focusing primarily on extant taxa. The main support for their theory consisted in differences in the spatial relationships between the rami of the mandibular branch of the 7th cranial nerve, the stapes, and the tympanum in anurans, reptiles, and mammals.