Why a Stone Can Bounce Off Water
The Flatness Happens at Impact
Imagine picking through stones at the water’s edge. The heavy, round one goes back. The thin, broad one stays in your hand. Before you throw anything, you have already made a little engineering judgment: this shape has a chance.
It is a good judgment. It is also easy to explain badly.
“Water gets hard if you hit it fast enough” sounds plausible, especially if you have ever regretted a belly flop. But a skipping stone does not meet a temporary floor. It meets liquid that it must accelerate and redirect. The reaction can turn the stone upward—provided the encounter produces the right force before the stone loses the geometry that makes a rebound possible.
The surprise is that an object need not arrive with that geometry. A soft, round ball can make it on contact.
That is the part I wanted to understand: not just why a flat stone skips, but why a ball can get away with ignoring the stone-picker’s first rule.
The water never becomes a floor
Lydéric Bocquet’s deliberately simplified account of stone skipping starts with a fast object, a partly immersed face and a force that depends on how much of that face encounters water. In the inertial regime he considers, the force scale grows with water density, wetted area and the square of speed. The details of orientation and fluid flow supply the coefficients and directions. [1]
The force does two jobs that pull against the thrower’s ambitions. Its upward component can reverse the stone’s descent. It must act long enough to overcome the incoming downward momentum as well as gravity during contact. Its backward component slows the stone down. A successful skip buys another flight at a cost.
This is not the same as supporting a motionless stone. For these fast, macroscopic ricochets, the relevant loading is dominated by inertia, with negligible surface-tension contribution in the regime discussed by the later experimental literature. There is no need to imagine the surface as a stretched trampoline holding up the stone. [2]
The speed dependence is worth making tangible. Choose a five-centimeter object traveling at ten meters per second. Using ordinary water properties, its Reynolds number is about 500,000: a very different situation from something creeping through a viscous liquid. The standard dynamic-pressure scale, one-half density times speed squared, is 50,000 pascals. Multiply that by an assumed wetted area of ten square centimeters and you get a force scale of fifty newtons, before accounting for the force coefficient or its vertical direction.
That is an illustrative calculation, not a prediction that a particular stone will skip. The area changes during impact, and so does the encounter with the water. But it gives speed some teeth: with area and geometry held fixed, halving speed quarters that force scale. The water has not changed its state. The moving object has changed the loading it can generate.
There are two angles in your hand
A stone can have its face almost parallel to the surface while its path points steeply downward. Those are different angles: the tilt of the face and the direction of travel.
Confusing them makes a throwing tip sound simpler than the physics. The stone needs a useful face orientation during contact, not just an attractive silhouette before release. Because water acts on an immersed region away from the center, the force can also twist the stone out of that orientation.
In Bocquet’s account, this is why spin matters. Spin helps stabilize attitude against the torque of impact; it is not the principal source of the upward hydrodynamic force. The familiar finger-flick helps the stone keep presenting the face you chose. [1, Section V]
The model is not a universal instruction sheet. It assumes shallow entry, simplified force coefficients and approximately fixed tilt during a collision. Real stones are irregular. Real water has waves. Its estimated skip counts are not promises, and its assumptions should not be smuggled into a magic launch angle.
Still, the broad lesson survives: a rigid stone has to bring a useful geometry and preserve it.
The ball that makes its own face
A 2016 study by Jesse Belden and colleagues gives that requirement a wonderfully physical counterexample. The team launched rigid and compliant spheres at water and used high-speed imaging alongside numerical and analytical models to examine what happened. [2]
The opening figure describes a rigid nylon sphere arriving at 24.3 meters per second and about thirty degrees to the surface. It dives. A soft silicone sphere, of almost the same size and density, arrives at twenty-two meters per second and thirty-two degrees. It skips.
These are not perfectly matched launches: the speeds, angles and material properties differ. The comparison is a vivid example, not proof that one changed variable explains everything. The broader series of stiffness and speed tests supplies the stronger evidence.
What distinguishes the soft sphere is what happens after contact. Water loading deforms it into a broad, disk-like shape. Its water-contacting face also changes inclination as elastic deformation travels through the material. The object that started as a ball briefly acquires a geometry that looks much more useful to a stone skipper.
The authors identify two ways this can help: a larger wetted area and a favorable orientation of that area. Their experiments and models connect the material response to enhanced upward forcing. In the elastic regime they studied, reducing shear modulus—a measure of resistance to shear deformation—reduced the minimum speed required to skip. [2, Figures 3–4]
“Soft” by itself is not enough of an explanation. The important comparison is material stiffness against the hydrodynamic pressure associated with impact. A ball that hardly changes shape under one launch may deform dramatically under another. The material also has a clock: elastic disturbances take time to travel through it, and their return can alter an encounter still in progress. For the traveling-wave mode examined in the model, softer material slows the changing angle of the contact face, helping it maintain favorable loading during contact.
There is a limitation worth keeping beside that explanation. The reduced analytical model does not independently calculate the entire pressure distribution or every change of attitude. It borrows the attack-angle evolution from the more detailed numerical model. The laboratory observations, simulations and simpler equations illuminate different parts of the event; they are not three independent replications. [2, Methods]
Nor does the result mean that making every ball softer indefinitely must improve it. The paper proposes an extremely soft limit in which surface tension could interfere with recovery of the ball’s shape, but explicitly leaves that prediction unvalidated. A mechanism demonstrated in one regime is not permission to erase its boundaries.
Flat is something the impact can do
What changed for me was the status of the shape.
For the rigid stone, flatness is mostly a property selected before the throw. For the compliant sphere, a broad contact face is something the collision creates. The water pushes, the material changes form, and that new form changes how the water pushes back.
The round ball has not abolished geometry. It has made geometry part of the motion.
I like that this does not make the old shore-side judgment foolish. A thin stone really does arrive with an advantage. It just reveals what the judgment was reaching for: not membership in the category “flat objects,” but a favorable interaction with moving water.
The next question is what happens when the water brings a changing shape of its own. A wave can alter the local surface and the next contact. How much of the ball’s advantage survives when the landing place is no longer still?
Sources and limits
[1] Lydéric Bocquet, “The physics of stone skipping,” American Journal of Physics 71, 150–155 (2003). Read as the complete 2002 author preprint, especially Sections II–V; this is a simplified theoretical treatment, not a new experiment conducted here.
[2] Jesse Belden et al., “Elastic spheres can walk on water,” Nature Communications 7, 10551 (2016). Main text, methods and captions support the experimental and modeling account. Its June 2016 erratum corrects the scaling of Figure 5’s colored background; no numerical regime boundaries here were taken from that background. The experimental team is independent of Bocquet’s, but its force model extends earlier rigid-disk work rather than independently deriving every part of it.
The opening scene is imagined. I did not throw stones, watch the supplementary movies or inspect the figure images; the laboratory sequences above follow the authors’ descriptions and captions. The numerical example is checked arithmetic using chosen dimensions, not a measured force, a trajectory simulation or a guaranteed recipe for a skip.