The Object: A Circle That Turns
In the marshes of the Paraná Delta, near the mouth of the river in Argentina, there is a circle of land roughly 120 metres across. It floats in a channel of water that is itself an almost perfect circle, separated from the surrounding wetland by a thin moat. The island is not anchored. It rotates — slowly, continuously, grinding its own circular shoreline against the channel that contains it. Locals and the filmmakers who surveyed the site in 2016 named it El Ojo: the Eye.
Two facts make it remarkable. First, the geometry is too clean: a circular island inside a circular channel, concentric to within metres, is not what erosion usually produces. Second, it moves. A patch of vegetation and peat, neutrally buoyant, set turning by the slow circulation of water beneath and around it — a state that, once established, sustains itself. Nothing pushes it on a schedule. It simply keeps turning.
This chapter takes El Ojo seriously, not as folklore but as evidence. The claim is modest and precise: El Ojo is what the attractor of Chapter 3 looks like when nature, rather than an engineer, sets the initial conditions.
Why It Is Not a Coincidence
A rotating circular island is a limit cycle made of mud. In the language of Chapter 3, it is a trajectory that has reached the unit circle r = 1 and rotates at near-constant angular speed θ̇ ≈ 1, while the slow vertical exchange of water and sediment plays the role of the height coordinate z. The shape persists because it is stable: small disturbances — a flood pulse, a dry season, a boat wake — decay rather than grow. That is the signature of a negative transverse Lyapunov exponent, the μ < 0 of the dm³ system.
The reason such a thing can be self-sustaining rather than merely transient is the contact structure. A planar circulation that did not exchange anything vertically would spin down by friction. The vertical coupling — water welling up through the peat, sediment cycling — is exactly the third coordinate that turns a decaying planar eddy into a persistent helix. Chapter 2 showed that promoting that third direction to a geometric axis produces a contact form; Chapter 3 showed that on the resulting 3-manifold the attractor is a helix, not a circle.
El Ojo realises, in sediment and water, the three ingredients the dm³ theorem requires: a radial restoring tendency toward a preferred circle (r → 1), a steady rotation (θ̇ ≈ 1), and a vertical exchange that keeps the rotation from dissipating (the z-coupling). The island is the projection to the (r, θ) plane of a helical Reeb orbit on a contact 3-manifold.
El Ojo as a dm³ Attractor
Write the island's state in cylindrical coordinates (r, θ, z): r the distance from the channel centre, θ the angular position of a marked point on the shoreline, z a measure of the vertical water/sediment exchange. The dm³ system of Chapter 3 is:
The radial term r(1 − r²) pulls every radius toward the unit circle. The coupling ε(r − 1)e−z vanishes on the attractor and decays as the exchange builds. The result, proved as a global statement for the outer basin in Chapter 10, is that every trajectory starting outside the unit circle converges exponentially to the helix at rate μ → −2, with period T* = 2π for one full turn.
The basin matters for a natural object. An island that started too small — too far inside the unit circle — would not lock into rotation; it would spin inward and dissolve. The honest, numerically-corrected boundary from Chapter 10 is asymmetric:
El Ojo exists because its initial radius fell in the convergent basin (r(0) > r*). That is the whole content of "why is it so round, and why does it keep turning": roundness is the attractor, turning is the Reeb flow, persistence is μ < 0, and existence-at-all is the basin condition.
This is a model, not a field measurement. We claim El Ojo is consistent with the dm³ attractor and that the qualitative features (circularity, rotation, persistence, sensitivity to size) match the theorem's predictions. A decisive test would be bathymetric and flow data fitting r*, μ, and T* — proposed as field work in Chapter 9. The interlude states the conjecture; it does not claim the measurement.