Modern mammals have three tiny bones in their middle ear that aid hearing. Oddly, these bones evolved from remnants of jaw bones, and how they migrated to form the ear has fascinated biologists for 200 years1. The middle ear and lower jaw of mammals today differ strikingly from those of other vertebrates. In addition to the three bones of the middle ear (termed the malleus, incus and stapes), mammalian ears have an ectotympanic bone, from which hangs the tympanic membrane (also known as the eardrum) that is needed for hearing. But they have just one lower jaw bone (the dentary bone). By contrast, other vertebrates have only one bone in the middle ear (the stapes), but more than eight lower jaw bones.
Nineteenth-century biologists were the first to recognize the similarities in development between some of these extra elements of the lower jaw of non-mammalian vertebrates and the middle ear of mammals, and in doing so chronicled one of the most fundamental transitions in mammalian evolution: the transformation of lower jaw bones to form the middle ear. On the basis of new fossil evidence, Wang et al.2 now crucially revise this classic story of mammalian evolution.
Once the basic similarities in relative position and structure (homologies) between bones of the lower jaw and middle ear had been established, the question of how such a major transformation could have occurred baffled scientists. One way to gain insight is through fossil evidence, which provides the only direct evidence that can capture key evolutionary moments in deep time. Discoveries of early fossil mammals and their closest relatives (called mammaliaforms) originally indicated a gradual transition as the lower jaw bones formed the middle ear3. This transition aligned with major groups of our evolutionary tree. On the basis of this initial fossil evidence, biologists coined the terms transitional mammalian middle ear (describing a partial detachment of the bones from the lower jaw) and definitive mammalian middle ear (describing the full detachment of the bones from the lower jaw).
However, subsequent fossil discoveries clouded this picture of a gradual transition4–6. Fossils discovered more recently show that the seemingly complex evolution that led to the detachment of the middle ear occurred independently at least three times in mammals7. Genetic evidence helps to explain these multiple origins of the middle ear, and what looks like a complex morphological (shape) transition is explained by relatively simple genetic mechanisms8,9.
Given the current fossil and developmental evidence, it is clear that the established terminology is outdated. A definitive mammalian middle ear is neither a defining feature of mammals nor is it found in all mammals. Evidence from 200 years of research in various fields of biology has resulted in contradictory and inconsistent terminology and interpretations. The classification system used for evolutionary studies of the mammalian middle ear has therefore been in urgent need of revision.
Wang et al. have now provided exactly that, in the form of a detailed assessment of the fossil and developmental evidence. Their article combines the description of a new fossil, a re-evaluation of previously reported fossil middle ears, and a discussion of the development of the middle ear in a wide range of modern mammals. On the basis of this much-needed, comprehensive comparison of the fossil and developmental data, Wang and colleagues have established a more-cohesive terminology system that is anchored in bone morphology (Fig. 1). This terminology identifies three types of middle ear: detached (the middle ear is completely separated from the lower jaw); Meckelian-attached (the middle ear is connected to the lower jaw through a structure called Meckel’s cartilage); and postdentary-attached (the middle ear is not separated from the jaw).
At the core of Wang and colleagues’ work is a fossil specimen from the Middle Jurassic epoch (160 million years ago), newly discovered in China. At first glance, the fossil itself is not spectacular; indeed, the species was described10 in 2017. But, crucially, the middle ear of this new specimen is much better preserved than in the previously discovered fossil, and the insights it provides, even though based on only a few tiny bones of the middle ear, have broad implications.
The fossil, Vilevolodon diplomylos, belongs to a group called Haramiyida, or Euharamiyida, whose classification has been a source of contention, being placed either in or outside the class Mammalia (mammals)10–12. Vilevolodon was described originally as having its middle ear attached to the lower jaw10. This interpretation placed the species outside Mammalia in some previous studies. The new fossil suggests that Vilevolodon actually had a middle ear with all elements detached from the lower jaw (Fig. 1). Wang and colleagues’ work thus places haramiyidans firmly within Mammalia in the evolutionary tree, and so favours an ancient origin for mammals, at least 215 million years ago (during the Late Triassic epoch). This is some 30 million years earlier than the timing of mammalian origins suggested by phylogenetic analysis in the original study of Vilevolodon10.
In addition, as part of their broad evaluation of the morphology of the middle ear, Wang et al. determined that a morphology that was widely held to be a specialization of monotremes (egg-laying mammals such as platypuses) is actually more widespread than previously considered. Monotremes have a flat, overlapping joint (with a broad contact) between the malleus and incus (Fig. 1), whereas adult mammals in the therian group, comprising placentals (which includes humans) and marsupials, have a saddle-shaped joint, with the incus lying behind the malleus. Wang and colleagues indicate that the overlapping joint arrangement of monotremes seems to be present in all major early mammalian clades, as well as in the early developmental growth stages of marsupials and placentals.
The authors speculate that this joint shape might have balanced the needs for both increased auditory and load-bearing functions during ear evolution. It is unclear whether the overlapping morphology in different mammalian groups is a shared innovation in the ancestral animals that gave rise to Mammalia, or whether it evolved independently (by convergent evolution) in different groups as their middle ears became detached from their jaws.
Although their research is a major contribution to the understanding of Haramiyida, Wang and colleagues’ study will not settle the debate about the placement and composition of this group. The striking increase in fossil discoveries, in particular new specimens from China, has fuelled routine revision of the mammalian tree of life during the past few years10–12. In the end, the lasting impact of Wang and colleagues’ research will probably lie in their detailed evaluation of fossil and developmental morphologies of the middle ear, and the establishment of new terminology that is consistent with current evidence. Their work is a foundational reference for future studies that provides a framework on which to evaluate the evolution of the middle ear and new fossil discoveries.