Floaty: Diverse Support Mechanisms

• Floaty is a term defining systems where sustained suspension arises from multiple mechanisms, including buoyancy, capillary forces, electric fields, and dynamic stabilization.
• The concept encompasses examples from ice floating due to density anomalies to heavy droplets maintained by capillary support and robots leveraging both helium and vertical wind for lift.
• Research on floaty systems integrates experimental, analytical, and numerical modeling to elucidate force balances, orientation dynamics, and control strategies for passive and active buoyancy regulation.

“Floaty” (Editor’s term) denotes systems whose sustained suspension is produced by distinct physical balances rather than by a single universal mechanism. In the cited literature, floaty behavior includes ice remaining buoyant on liquid water because freezing lowers density, heavy droplets supported at a lighter-liquid surface by capillary forces, bodies that float upside down beneath a dynamically stabilized liquid slab, glycerol clusters levitated between electrodes by electric-field gradients, underwater robots that meter trapped gas to reach neutral buoyancy, helium-filled flapping robots for indoor human–robot interaction, and a shape-changing soaring robot named Floaty that hovers by harvesting vertical wind (Sun, 2015, Pototsky et al., 2021, Apffel et al., 2020, Tsai et al., 2023, Kobo et al., 2021, Xu et al., 2 Apr 2025, Elmkaiel et al., 27 Aug 2025).

1. Physical scope and support mechanisms

The common feature across floaty systems is vertical support or sustained suspension, but the underlying mechanism varies sharply with scale, medium, and constitutive physics. In classical floating-body theory, the starting point is Archimedes’ principle: the buoyant force is $F_b=\rho_f g V_{\rm disp}$, and vertical equilibrium requires weight to be balanced by displaced-fluid weight. For long prismatic bodies, orientation stability can then be analyzed by the potential-energy landscape $$U(\theta)=\hat n(\theta)\cdot(\vec G-\vec B(\theta))$$, where $\vec G$ is the center of gravity and $\vec B(\theta)$ the center of buoyancy (Anderson et al., 2022).

A broader classification emerges when interfacial, vibratory, electric, and aerodynamic effects are included.

Mechanism	Representative balance	Representative system
Density anomaly	$\rho = m/V$, with $\rho \propto (r_{OO})^{-3}$	Ice Ih on water
Archimedean buoyancy	$F_b=\rho_f g V_{\rm disp}$	3D-printed floaters, BackBot, Cuddle-Fish
Capillary support	$F_{st}+B\ge W$	Heavy drop on a lighter liquid
Dynamic stabilization	$g_{\rm eff}(t)=g-A\omega^2\cos(\omega t)$	Floaters under a levitated liquid
Electric-field body force	$F_K=\frac{1}{2}\epsilon_0(\epsilon_r-1)\nabla(E^2)$	Levitated glycerol clusters
Drag-supported soaring	$$U(\theta)=\hat n(\theta)\cdot(\vec G-\vec B(\theta))$$0	Floaty soaring robot

This taxonomy shows that “floating” is not reducible to ordinary buoyancy alone. Heavy droplets on oil require a vertical component of surface tension at a triple-phase contact line; glycerol clusters in strong DC fields are supported by Kelvin body forces and plasma-assisted conduction; upside-down floaters beneath a levitated liquid require vibratory Kapitza-like stabilization; and the soaring robot Floaty uses drag-dominated aerodynamics in a vertical updraft rather than stored lift or thrust (Pototsky et al., 2021, Tsai et al., 2023, Apffel et al., 2020, Elmkaiel et al., 27 Aug 2025).

2. Ice as a density-anomaly floater

The mechanism by which ice floats is treated in “Hidden force floating ice” as a cooperative relaxation problem of the hydrogen bond, modeled as an $$U(\theta)=\hat n(\theta)\cdot(\vec G-\vec B(\theta))$$1 composite with two coupled segments: the intramolecular covalent $$U(\theta)=\hat n(\theta)\cdot(\vec G-\vec B(\theta))$$2 segment and the intermolecular $$U(\theta)=\hat n(\theta)\cdot(\vec G-\vec B(\theta))$$3 “nonbond” segment. The paper attributes the density anomaly to segmental specific-heat disparity and Coulomb repulsion between oxygen ions. Quantitatively, $$U(\theta)=\hat n(\theta)\cdot(\vec G-\vec B(\theta))$$4 and $$U(\theta)=\hat n(\theta)\cdot(\vec G-\vec B(\theta))$$5, so the two segments respond differently to cooling. In the quasi-solid freezing regime II, the $$U(\theta)=\hat n(\theta)\cdot(\vec G-\vec B(\theta))$$6 segment becomes the master and contracts slightly, the $$U(\theta)=\hat n(\theta)\cdot(\vec G-\vec B(\theta))$$7 segment elongates more, and the $$U(\theta)=\hat n(\theta)\cdot(\vec G-\vec B(\theta))$$8 containing angle $$U(\theta)=\hat n(\theta)\cdot(\vec G-\vec B(\theta))$$9 widens. The net result is an increase in $\vec G$0 separation, opening of the local tetrahedral network, expansion of molar volume, and a density decrease that renders ice buoyant (Sun, 2015).

The geometric changes are quantitatively specific. The containing angle $\vec G$1 widens from about $\vec G$2 in liquid water to approximately $\vec G$3 at the lowest-density point near freezing and to about $\vec G$4 deeper into the solid phase. The increase from $\vec G$5 contributes a maximum of about $\vec G$6 to $\vec G$7 bond elongation and about $\vec G$8 to volume expansion in regime II. Over the same interval, $\vec G$9 remains typically $\vec B(\theta)$0–$\vec B(\theta)$1 but shortens by about $\vec B(\theta)$2–$\vec B(\theta)$3, while $\vec B(\theta)$4 is typically $\vec B(\theta)$5–$\vec B(\theta)$6 and elongates by roughly $\vec B(\theta)$7–$\vec B(\theta)$8. X-ray, neutron, and MD trends give a freezing-induced increase in $\vec B(\theta)$9 of roughly $\rho = m/V$0–$\rho = m/V$1, consistent with $\rho = m/V$2.

The macroscopic density anomaly follows directly. Water reaches a density maximum near $\rho = m/V$3, where $\rho = m/V$4. Near $\rho = m/V$5, $\rho = m/V$6, whereas $\rho = m/V$7. The approximately $\rho = m/V$8–$\rho = m/V$9 lower density of ice near $\rho \propto (r_{OO})^{-3}$0 is attributed to the roughly $\rho \propto (r_{OO})^{-3}$1 volume expansion from angle stretching plus additional expansion from $\rho \propto (r_{OO})^{-3}$2 elongation in regime II. A common misconception addressed explicitly by the paper is that the intramolecular $\rho \propto (r_{OO})^{-3}$3 angle drives the anomaly; instead, that angle remains essentially fixed near the familiar $\rho \propto (r_{OO})^{-3}$4, while the decisive variables are the containing angle $\rho \propto (r_{OO})^{-3}$5 and the $\rho \propto (r_{OO})^{-3}$6 distance. Raman and IR spectroscopy support the regime change: near freezing, $\rho \propto (r_{OO})^{-3}$7 stiffens while $\rho \propto (r_{OO})^{-3}$8 softens, indicating $\rho \propto (r_{OO})^{-3}$9 contraction and $F_b=\rho_f g V_{\rm disp}$0 elongation.

3. Capillary support, vibratory floatability, and inverted buoyancy

A heavy liquid drop can float on a lighter liquid because surface tension at the triple-phase contact line supplies an upward force that supplements buoyancy. The general static condition is

$F_b=\rho_f g V_{\rm disp}$1

with

$F_b=\rho_f g V_{\rm disp}$2

where $F_b=\rho_f g V_{\rm disp}$3 is the carrier-liquid–gas interfacial tension and $F_b=\rho_f g V_{\rm disp}$4 the local interface slope along the contact line. In the axisymmetric case, $F_b=\rho_f g V_{\rm disp}$5. The corresponding study shows that sufficiently small heavy drops may possess two distinct stable equilibrium shapes, one with a smaller and one with a larger radius of the triple-phase contact line. For water drops on oil, volumes below about $F_b=\rho_f g V_{\rm disp}$6 display bistability; larger drops retain only the larger-radius branch until static floating equilibria terminate at a saddle-node (Pototsky et al., 2021).

Vertical vibration introduces a second route to floatability. In the same drop-on-oil system, low-frequency vertical forcing excites subharmonic Faraday waves on the drop surface, elongating the drop horizontally and increasing the average contact-line length. At $F_b=\rho_f g V_{\rm disp}$7, the Faraday threshold on the drop is near $F_b=\rho_f g V_{\rm disp}$8, while the bath threshold is near $F_b=\rho_f g V_{\rm disp}$9. In the interval $F_{st}+B\ge W$0, elongated drops with plan-view axis ratio about $F_{st}+B\ge W$1 remain afloat for volumes up to about $F_{st}+B\ge W$2, whereas the static limit is about $F_{st}+B\ge W$3. The decisive point is that the extra support does not come from increased buoyancy: the paper shows that the time-averaged buoyancy is actually reduced relative to the static submerged volume, so the enhanced floatability must arise from a larger time-averaged capillary lifting force.

A still more nonclassical configuration is “floating under a levitating liquid.” There, a viscous liquid slab is held above an air layer while the container is shaken vertically. The trapped air acts as a gas spring with resonance

$F_{st}+B\ge W$4

and near resonance the liquid-interface motion is amplified by more than an order of magnitude. Experiments fitted $F_{st}+B\ge W$5, $F_{st}+B\ge W$6, $F_{st}+B\ge W$7, and an effective $F_{st}+B\ge W$8. The inverted lower interface is stabilized when

$F_{st}+B\ge W$9

and this enabled stabilization of approximately half a liter of liquid with widths up to about $g_{\rm eff}(t)=g-A\omega^2\cos(\omega t)$0. Under those conditions, immersed bodies can float upside down beneath the slab as if gravity were inverted, obeying a symmetric Archimedes’ principle on the lower interface. The experiments report $g_{\rm eff}(t)=g-A\omega^2\cos(\omega t)$1 spheres with relative densities $g_{\rm eff}(t)=g-A\omega^2\cos(\omega t)$2–$g_{\rm eff}(t)=g-A\omega^2\cos(\omega t)$3 times $g_{\rm eff}(t)=g-A\omega^2\cos(\omega t)$4 floating both above and below the slab, and even boats floating simultaneously on the upper and lower interfaces (Apffel et al., 2020).

4. Electric-field levitation and active buoyancy regulation

Strong electric-field gradients can also generate floaty states. In the glycerol-cluster experiments inspired by the 1893 water-bridge configuration, a single pair of facing electrodes at $g_{\rm eff}(t)=g-A\omega^2\cos(\omega t)$5–$g_{\rm eff}(t)=g-A\omega^2\cos(\omega t)$6 across spout gaps $g_{\rm eff}(t)=g-A\omega^2\cos(\omega t)$7, $g_{\rm eff}(t)=g-A\omega^2\cos(\omega t)$8, or $g_{\rm eff}(t)=g-A\omega^2\cos(\omega t)$9 produced nominal macroscopic fields of about $F_K=\frac{1}{2}\epsilon_0(\epsilon_r-1)\nabla(E^2)$0–$F_K=\frac{1}{2}\epsilon_0(\epsilon_r-1)\nabla(E^2)$1, well above the air-breakdown threshold near $F_K=\frac{1}{2}\epsilon_0(\epsilon_r-1)\nabla(E^2)$2. After the continuous bridge fractured, lump-like glycerol clusters persisted in mid-air with Taylor-cone-like ends pointing toward the electrodes, intermittent plasma clusters in the gaps, and near-periodic oscillations in both vertical and horizontal directions. The paper explicitly states that this levitation does not require a classical continuous bridge form. The force model is Kelvin polarization forcing,

$F_K=\frac{1}{2}\epsilon_0(\epsilon_r-1)\nabla(E^2)$3

with levitation when

$F_K=\frac{1}{2}\epsilon_0(\epsilon_r-1)\nabla(E^2)$4

Observed survival times were $F_K=\frac{1}{2}\epsilon_0(\epsilon_r-1)\nabla(E^2)$5–$F_K=\frac{1}{2}\epsilon_0(\epsilon_r-1)\nabla(E^2)$6; vertical oscillations reached roughly $F_K=\frac{1}{2}\epsilon_0(\epsilon_r-1)\nabla(E^2)$7 in some cases; FFT analysis gave a dominant period near $F_K=\frac{1}{2}\epsilon_0(\epsilon_r-1)\nabla(E^2)$8; and finite-element calculations under the minimal field condition $F_K=\frac{1}{2}\epsilon_0(\epsilon_r-1)\nabla(E^2)$9, $$U(\theta)=\hat n(\theta)\cdot(\vec G-\vec B(\theta))$$00 gave force-to-weight ratio $$U(\theta)=\hat n(\theta)\cdot(\vec G-\vec B(\theta))$$01, compared with an experimental value around $$U(\theta)=\hat n(\theta)\cdot(\vec G-\vec B(\theta))$$02. Three common shapes—Peanut, Dumbbell, and Ellipse—were examined, and analogous detached-cluster behavior was not observed in water, where the detached column broke into droplets through Rayleigh–Plateau instability (Tsai et al., 2023).

BackBot addresses an opposite problem: not levitation in air, but underwater neutral floatation through active gas management. Inspired by the backswimmer insect, it uses low-voltage electrolysis on interlocking $$U(\theta)=\hat n(\theta)\cdot(\vec G-\vec B(\theta))$$03 stainless-steel comb electrodes separated by $$U(\theta)=\hat n(\theta)\cdot(\vec G-\vec B(\theta))$$04 to nucleate microbubbles, a 3D-printed curved cellular canopy to trap them as an external gas reservoir, and a mechanically decoupled linear vibration mechanism with $$U(\theta)=\hat n(\theta)\cdot(\vec G-\vec B(\theta))$$05 stroke at about $$U(\theta)=\hat n(\theta)\cdot(\vec G-\vec B(\theta))$$06 to release excess bubbles. Neutral buoyancy follows from

$$U(\theta)=\hat n(\theta)\cdot(\vec G-\vec B(\theta))$$07

or, with bubble storage,

$$U(\theta)=\hat n(\theta)\cdot(\vec G-\vec B(\theta))$$08

The robot’s trimmed mean density was approximately $$U(\theta)=\hat n(\theta)\cdot(\vec G-\vec B(\theta))$$09, so it was slightly negatively buoyant without trapped gas. The measured buoyancy-force generation rate was $$U(\theta)=\hat n(\theta)\cdot(\vec G-\vec B(\theta))$$10, corresponding to about $$U(\theta)=\hat n(\theta)\cdot(\vec G-\vec B(\theta))$$11 of effective captured gas. Bubble release exhibited a rapid first phase of $$U(\theta)=\hat n(\theta)\cdot(\vec G-\vec B(\theta))$$12 followed by a slower second phase of $$U(\theta)=\hat n(\theta)\cdot(\vec G-\vec B(\theta))$$13. With PID gains $$U(\theta)=\hat n(\theta)\cdot(\vec G-\vec B(\theta))$$14, $$U(\theta)=\hat n(\theta)\cdot(\vec G-\vec B(\theta))$$15, and $$U(\theta)=\hat n(\theta)\cdot(\vec G-\vec B(\theta))$$16, the system reached a new depth setpoint with $$U(\theta)=\hat n(\theta)\cdot(\vec G-\vec B(\theta))$$17 rise time, no overshoot, and $$U(\theta)=\hat n(\theta)\cdot(\vec G-\vec B(\theta))$$18 settling to within $$U(\theta)=\hat n(\theta)\cdot(\vec G-\vec B(\theta))$$19. The paper reports that neutral floatation can be reached within seconds once the canopy is preloaded, and maintained for extended periods subject to gas diffusion losses (Kobo et al., 2021).

5. Robotic embodiments of floaty design

One robotic interpretation of floaty behavior is near-neutral helium buoyancy combined with compliant propulsion. Cuddle-Fish is a soft floating robot with an aluminum-coated nylon envelope approximately $$U(\theta)=\hat n(\theta)\cdot(\vec G-\vec B(\theta))$$20, two soft flapping wings supported by $$U(\theta)=\hat n(\theta)\cdot(\vec G-\vec B(\theta))$$21 carbon-fiber rods, a passive tail, an internal center-of-mass shifter for pitch control, an ESP32 controller, a $$U(\theta)=\hat n(\theta)\cdot(\vec G-\vec B(\theta))$$22 $$U(\theta)=\hat n(\theta)\cdot(\vec G-\vec B(\theta))$$23 LiPo battery, and total mass around $$U(\theta)=\hat n(\theta)\cdot(\vec G-\vec B(\theta))$$24. Wingspan varies from $$U(\theta)=\hat n(\theta)\cdot(\vec G-\vec B(\theta))$$25 at the highest position to $$U(\theta)=\hat n(\theta)\cdot(\vec G-\vec B(\theta))$$26 at the lowest position, enabling passage through standard doorframes. The design exploits helium’s practical lift constraint of approximately $$U(\theta)=\hat n(\theta)\cdot(\vec G-\vec B(\theta))$$27 per liter of displaced air, so aggressive mass reduction and slight ballast tuning are necessary. Maneuvers include Fly Towards, Fly Overhead, Ascend and Descend, Circle Around, Spin Overhead, and Wave. In a study with $$U(\theta)=\hat n(\theta)\cdot(\vec G-\vec B(\theta))$$28, $$U(\theta)=\hat n(\theta)\cdot(\vec G-\vec B(\theta))$$29 participants ($$U(\theta)=\hat n(\theta)\cdot(\vec G-\vec B(\theta))$$30) engaged in touch-based interaction, and valence increased significantly with Wilcoxon statistic $$U(\theta)=\hat n(\theta)\cdot(\vec G-\vec B(\theta))$$31, $$U(\theta)=\hat n(\theta)\cdot(\vec G-\vec B(\theta))$$32; spontaneous behaviors included patting, stroking, hugging, cheek touch, high-five, and head bump. The reported interpretation is that the soft, quiet, low-kinetic-energy form constitutes a socially acceptable alternative to rigid quadrotors for companionship, play, and gentle indoor interaction (Xu et al., 2 Apr 2025).

A second robotic interpretation is the soaring robot Floaty, which is not lighter-than-air but passively propelled by vertical wind. Floaty has a footprint of about $$U(\theta)=\hat n(\theta)\cdot(\vec G-\vec B(\theta))$$33, total mass $$U(\theta)=\hat n(\theta)\cdot(\vec G-\vec B(\theta))$$34, a 3D-printed fiber-reinforced Onyx body, four PLA Tough flaps of about $$U(\theta)=\hat n(\theta)\cdot(\vec G-\vec B(\theta))$$35 each, KST X08H Plus $$U(\theta)=\hat n(\theta)\cdot(\vec G-\vec B(\theta))$$36 servos, and power from two single-cell $$U(\theta)=\hat n(\theta)\cdot(\vec G-\vec B(\theta))$$37 LiPo batteries in series. Four independently actuated flaps are combined through a control-allocation matrix to yield largely decoupled roll, pitch, yaw, and vertical-force control near hover. Passive stability is built into the morphology: the center of mass lies about $$U(\theta)=\hat n(\theta)\cdot(\vec G-\vec B(\theta))$$38 below the flap plane, and each flap has a $$U(\theta)=\hat n(\theta)\cdot(\vec G-\vec B(\theta))$$39 kink that changes the roll cross-section from unstable to stable. Wind-tunnel experiments in vertical updrafts from about $$U(\theta)=\hat n(\theta)\cdot(\vec G-\vec B(\theta))$$40 to $$U(\theta)=\hat n(\theta)\cdot(\vec G-\vec B(\theta))$$41 showed hovering for $$U(\theta)=\hat n(\theta)\cdot(\vec G-\vec B(\theta))$$42 within a $$U(\theta)=\hat n(\theta)\cdot(\vec G-\vec B(\theta))$$43 box, position error below $$U(\theta)=\hat n(\theta)\cdot(\vec G-\vec B(\theta))$$44 on each axis during most of the experiment, attitude error below $$U(\theta)=\hat n(\theta)\cdot(\vec G-\vec B(\theta))$$45, yaw steps up to $$U(\theta)=\hat n(\theta)\cdot(\vec G-\vec B(\theta))$$46 with settling time under $$U(\theta)=\hat n(\theta)\cdot(\vec G-\vec B(\theta))$$47, and disturbance rejection under crosswind up to about $$U(\theta)=\hat n(\theta)\cdot(\vec G-\vec B(\theta))$$48. Average electrical power was about $$U(\theta)=\hat n(\theta)\cdot(\vec G-\vec B(\theta))$$49, giving specific power near $$U(\theta)=\hat n(\theta)\cdot(\vec G-\vec B(\theta))$$50, compared with typical thruster-powered hover in the range $$U(\theta)=\hat n(\theta)\cdot(\vec G-\vec B(\theta))$$51–$$U(\theta)=\hat n(\theta)\cdot(\vec G-\vec B(\theta))$$52 (Elmkaiel et al., 27 Aug 2025).

Taken together, these two platforms show that “floaty robot” can denote either buoyancy-assisted, touch-safe indoor locomotion or passive soaring with active morphological control. The common engineering pattern is that environmental support—helium displacement in one case, vertical wind in the other—carries most of the weight-support burden, while actuation is reserved for modulation, steering, and interaction.

6. Modeling, control, and transport of floaters

Floaty systems are unusually sensitive to coupled fluid–body dynamics, so analytical and numerical modeling is central. For rigid floating-body CFD, FloatStepper was introduced as a non-iterative, added-mass–aware coupling algorithm for incompressible single- and two-phase flow. The key decomposition writes hydrodynamic reactions as an added-mass term proportional to instantaneous body acceleration plus a remainder:

$$U(\theta)=\hat n(\theta)\cdot(\vec G-\vec B(\theta))$$53

The method measures $$U(\theta)=\hat n(\theta)\cdot(\vec G-\vec B(\theta))$$54 at each time step by prescribed probe accelerations along active degrees of freedom, and measures $$U(\theta)=\hat n(\theta)\cdot(\vec G-\vec B(\theta))$$55 by a zero-acceleration probe. It was implemented as an open-source OpenFOAM extension. Benchmarks include a rising light disc in ideal fluid, where the relative error in acceleration was about $$U(\theta)=\hat n(\theta)\cdot(\vec G-\vec B(\theta))$$56, a two-phase disc entering water with density ratio about $$U(\theta)=\hat n(\theta)\cdot(\vec G-\vec B(\theta))$$57, a massless wiggling ellipse showing correct translation–rotation cross-coupling, and free and moored wave-structure cases. The explicit goal is to remove the classical added-mass instability that appears when $$U(\theta)=\hat n(\theta)\cdot(\vec G-\vec B(\theta))$$58 (Roenby et al., 2023).

For static orientation stability, “Mathematics of Floating 3D Printed Objects” formulates floating as an energy-landscape problem over waterline angle $$U(\theta)=\hat n(\theta)\cdot(\vec G-\vec B(\theta))$$59. Stable orientations are minima of

$$U(\theta)=\hat n(\theta)\cdot(\vec G-\vec B(\theta))$$60

with $$U(\theta)=\hat n(\theta)\cdot(\vec G-\vec B(\theta))$$61. The paper provides closed-form square results, polygon-based centroid formulas via Green’s theorem, MATLAB code, and 3D-printing workflows. Representative results include a symmetric square with eight stable orientations at $$U(\theta)=\hat n(\theta)\cdot(\vec G-\vec B(\theta))$$62, four corner-up equilibria at $$U(\theta)=\hat n(\theta)\cdot(\vec G-\vec B(\theta))$$63, and four flat-side-up equilibria at $$U(\theta)=\hat n(\theta)\cdot(\vec G-\vec B(\theta))$$64. Off-center ballast shifts $$U(\theta)=\hat n(\theta)\cdot(\vec G-\vec B(\theta))$$65 and reduces or bifurcates the number of stable minima. A nonconvex “Mason M” cross-section with $$U(\theta)=\hat n(\theta)\cdot(\vec G-\vec B(\theta))$$66 exhibited four stable orientations, and the framework is explicitly linked to Ulam’s floating body problem (Anderson et al., 2022).

Once afloat, bodies are transported, dispersed, and reoriented by waves. In laboratory surface-wave turbulence, buoyant $$U(\theta)=\hat n(\theta)\cdot(\vec G-\vec B(\theta))$$67 PVC spheres tracked by 3D PTV exhibited three temporal regimes: early-time ballistic motion dominated by wave drag up to about $$U(\theta)=\hat n(\theta)\cdot(\vec G-\vec B(\theta))$$68, intermediate-time saturation associated with trapping by short-lived horizontal eddies up to about $$U(\theta)=\hat n(\theta)\cdot(\vec G-\vec B(\theta))$$69, and late-time superdiffusive growth with exponent $$U(\theta)=\hat n(\theta)\cdot(\vec G-\vec B(\theta))$$70 driven by mean circulation. Delaunay tessellation of floater positions gave Gamma-distributed normalized triangle areas with shape parameter $$U(\theta)=\hat n(\theta)\cdot(\vec G-\vec B(\theta))$$71, indicating preferential concentration (Grosso et al., 2019). For thin flexible strips in monochromatic gravity waves, a diffractionless Froude–Krylov model predicts a mean yaw moment favoring longitudinal alignment with the wave direction; experiments confirmed systematic rotation toward $$U(\theta)=\hat n(\theta)\cdot(\vec G-\vec B(\theta))$$72, and found empirical scaling

$$U(\theta)=\hat n(\theta)\cdot(\vec G-\vec B(\theta))$$73

with the mean angular velocity decreasing as strip length increases. The same study predicts a small reduction in Stokes drift relative to a material point (Dhote et al., 2024).

A plausible implication is that floaty behavior is governed as much by orientation dynamics, added mass, and wave-mediated transport as by the primary support mechanism. In that sense, floaty systems form a coherent research domain not because they share one force law, but because they repeatedly couple weight support to geometry, fluid response, and slow manifold stability.