14The Round Window, Oval Window, and Cochlear Interface

FTwo windows, one incompressible fluid

The cochlea is a fluid-filled tube wound inside the densest bone in the body, and fluid does not compress. That single physical fact dictates the entire architecture of the cochlear interface. When the stapes footplate pushes inward at the oval window, the perilymph it displaces has to go somewhere; because the bony labyrinth is otherwise rigid, the only available outlet is a second, membrane-covered opening — the round window— which bulges outward by an exactly equal volume. Push in at one window, bulge out at the other: the two windows are the inlet and the relief valve of a closed hydraulic system.

This is why an ear cannot hear well through the oval window alone. If the round window were rigidly fixed, the incompressible perilymph would have nowhere to go and the stapes would meet an almost immovable load. The round window’s compliance is therefore not an incidental feature but a functional requirement, and clinical cases of isolated round window atresia— a congenitally bony-sealed round window with an otherwise normal ear — produce a real, predominantly conductive hearing loss of around 30–40 dB [2003]. The two windows are a matched pair; disable either and hearing suffers.

For the reconstructive surgeon this reframes the goal of every operation. Ossiculoplasty is often described as rebuilding the chain from drum to stapes, but its real purpose is to make one window move differently from the other. Everything that follows in this module is about that difference and how easily surgery can squander it.

FAnatomy of the round and oval windows

Both windows sit on the medial (labyrinthine) wall of the middle ear, separated by the rounded bulge of the promontory, the basal turn of the cochlea. The oval window lies above and slightly in front, occupied by the stapes footplate and sealed by the elastic annular ligament; it opens into the scala vestibuli. The round window lies below and behind, set deep in the round window niche of the hypotympanum, and opens into the scala tympani. This staggered geometry — oval window superior and lateral, round window inferior and recessed — is not cosmetic. By tucking the round window into a niche partly hidden from the tympanic cavity, normal anatomy acoustically isolates it from airborne sound, so that the chain can drive the oval window while the round window responds mainly to fluid pushed through the cochlea.

The round window itself is closed by the round window membrane(the secondary tympanic membrane), a thin, three-layered, slightly concave membrane. It is a small and remarkably variable target. Across a systematic review of 808 temporal bones, its surface area ranged from roughly 0.32 to 2.89 mm², with a smallest dimension of about 0.5 to 2.1 mm [2025]. A bony lip, the round window niche, frequently overhangs and partly conceals it, which is why a surgeon can finish an operation without ever truly seeing the membrane — and why an obliterated or covered round window is so easily overlooked at surgery [2003].

The membrane is not a passive lid. Laser-Doppler mapping of human round window membranes shows a complex, non-uniform vibration that moves out of phase with the inflow at the oval window across the auditory range — the membrane is genuinely pistoning fluid out as the stapes drives it in [2004]. That out-of-phase relief is the kinematic expression of the incompressible-fluid principle from the previous section.

TThe pressure difference that drives the cochlea

The cochlear partition — the basilar membrane and the organ of Corti it carries — separates the scala vestibuli from the scala tympani. What deflects it, and therefore what the cochlea actually hears, is not the pressure at the oval window in isolation but the difference in pressure between the two scalae, that is, between the oval and round windows. This is the heart of the two-window view of middle-ear function. In an elegant experiment the two window pressures were controlled independently while cochlear potentials were recorded; the response tracked the differential pressure, and when the two pressures were made equal and in phase the response collapsed by some 40 dB into a sharp minimum [1996]. Equal in-phase pressures cancel; it is the difference that stimulates.

It helps to separate the two parallel routes by which sound can reach the windows. Ossicular coupling is the normal route: the drum and ossicles funnel energy onto the oval window, driving it far harder than the shielded round window and so generating a large differential. Acoustic coupling is the direct airborne drive of the windows by sound pressure in the middle-ear air space. In the unified model of middle-ear transmission, the small acoustic-coupling difference simply adds to the ossicular drive; in the normal ear it is negligible, typically tens of decibels below the ossicular contribution, so it can be ignored [1992]. The geometry of the windows — the round window recessed and isolated — is precisely what keeps acoustic coupling small and lets the ossicular route dominate.

The load the footplate drives into is the stapes-cochlear input impedance, set by the annular ligament, the inertia and compliance of the cochlear fluids, and the compliance of the round window membrane. Broad-band measurement of this impedance in human temporal bones — made with the round window deliberately insulated — shows it to be largely resistive and consistent across the speech frequencies [1996]. The round window is one term in that load: normally compliant enough to be near-negligible, but capable of dominating when stiffened, obstructed, or atretic.

TWhen coupling fails: the acoustic short-circuit

The clinical power of the two-window idea appears when the ossicular route is broken. Consider a complete ossicular discontinuity behind an intact, normally aerated drum. The drum still vibrates and the middle-ear air space is still driven by sound — but with no chain to focus energy on the oval window, both windows are now driven by the same weak airborne pressure, nearly equally and nearly in phase. The differential that normally stimulates the cochlea almost vanishes. This is the acoustic short-circuit, and it explains the classic finding that an intact-drum discontinuity produces a maximal conductive loss approaching 50–60 dB — as large as that from complete stapes fixation, despite an anatomically normal-looking ear [1992].

The same principle illuminates a counter-intuitive surgical lesson: a technically perfect ossicular bridge can still leave a substantial gap if the windows are no longer differentially driven. The differential can be lost from either end — by weakening the oval-window drive (a poorly coupled or fixed prosthesis) or by strengthening the round-window drive (exposing or covering the round window so airborne sound reaches it directly). Both manoeuvres push the two window pressures toward equality and shrink the useful difference.

CDesigning for the windows in reconstruction

Two practical rules for ossiculoplasty follow directly from the windows. First, keep both niches clear and the round window isolated. Tissue, packing, gelatin sponge, bone cement, or a graft draped across the round window niche all threaten the differential, either by adding load to the outflow or by acoustically connecting the round window to the same sound that drives the oval window. Bone cement near the niche is a particular hazard and is conventionally protected against by shielding the footplate and windows during application. Second, drive the oval window decisively: a well-coupled, correctly tensioned prosthesis on a mobile stapes maintains a strong oval-window pressure and, by contrast, a large differential.

These rules also explain why some technically successful operations disappoint. After canal-wall-down mastoidectomy the protective anatomy of the round window niche is often lost; the round window can be left exposed in the open cavity, free to be driven by airborne sound, so the surgeon must actively re-create the isolation that nature provided. A low cartilage tympanoplasty that drapes over the hypotympanum can do the same harm, turning an excellent-looking repair into a persistent conductive gap. The lesson is that middle-ear pathology and geometry, more than the prosthesis material itself, determine the hearing result— and the windows are central to that geometry.

CRound-window shielding and stimulation

When ossicular coupling cannot be restored, acoustic coupling becomes the only route to the cochlea — and the round window becomes the lever the surgeon must control. The clearest example is type IV/V tympanoplasty, in which sound reaches a naked mobile footplate while the round window is deliberately shielded beneath a graft (the classic “cavum minor”). Experimental analysis of these constructs shows that hearing afterward depends almost entirely on acoustic coupling and on round-window protection: shielding the round window re-creates a pressure difference between the windows and measurably improves transmission, whereas an unshielded round window leaves the windows driven together and the gap large [1997]. Round-window shielding is therefore a true acoustic manoeuvre, not merely a way to stabilise tissue.

The same window can be harnessed deliberately. In round-window stimulationwith an active middle-ear implant, a transducer is coupled to the round window membrane to drive cochlear fluids in reverse when the ossicular route is unusable; the membrane’s small, variable size and the overhanging niche make precise coupling and niche preparation critical to success [2025]. And conditions that stiffen or obliterate the round window — tympanosclerosis, granulation, cholesteatomatous fibrosis, or true atresia — raise the stapes-cochlear input impedance and can produce a conductive loss even with a perfectly reconstructed chain [2003].

In every one of these scenarios the underlying principle is the same one that opened the module: the cochlea is driven by a differencein pressure between two windows. The reconstructive surgeon’s task is not simply to rebuild a chain but to protect and, where necessary, deliberately re-create that difference — keeping the round window isolated, the oval window decisively driven, and both niches clear so that the inner ear receives a clean, differential signal.