Why Deep Water Crushes Submarines, Wings Lift Aircraft, and Raindrops Stay Round

Physics uses the word fluid for anything that flows and takes the shape of whatever container it's in -- which covers liquids like water just as much as gases like air. Most of what this chapter covers, pressure, flow, viscosity, applies to both alike, which is exactly why air flowing over an aircraft wing obeys the very same rules as water flowing through a pipe. A few properties, like surface tension, belong to liquids specifically, since a gas has no true free surface of its own to pull taut the way a liquid's does. A sharp needle pressed against skin pierces it easily, while a blunt object pressed with the identical force leaves the skin unharmed -- and a circus performer survives a heavy wooden plank laid across their chest precisely because the plank spreads that weight over a wide area. Both everyday facts point to the same underlying quantity: pressure, the force acting per unit area. When a fluid at rest presses against any submerged surface, that force is always exactly perpendicular (normal) to the surface -- if any part of the force acted sideways, by Newton's third law the surface would push back sideways on the fluid too, setting it flowing, which contradicts the fluid actually being at rest. For a normal force F spread over an area A, the average pressure is Pav=F/A, measured in newtons per square metre, or pascal (Pa), honouring Blaise Pascal's own pioneering studies of fluid pressure. Crucially, pressure itself is a scalar quantity, not a vector -- even though the word 'force' appears in its very definition, that force is always just the size of the push perpendicular to whatever small area is being considered, carrying no independent direction of its own. Examining a small triangular prism-shaped element deep inside a fluid at rest shows this directly: balancing the forces on its three faces (using both equilibrium and simple geometry) proves that the pressure exerted on all three faces must be exactly equal, however each face happens to be tilted -- this is Pascal's Law, that the pressure in a fluid at rest is the same at every point at the same height, and that pressure is transmitted equally in every direction, with no preferred direction of its own at all. Two thighbones, each with a cross-sectional area of 10 cm², supporting a 40 kg body's upper weight, illustrates the calculation directly: with a total supporting area of 20x10⁻⁴ m² and a downward force of 400 N, the average pressure sustained works out to a substantial 2x10⁵ N/m².

Consider a cylindrical element of fluid, area A and height h, sitting between two points at different depths inside a fluid at rest. Since the fluid isn't accelerating, the upward push from below must exceed the downward push from above by exactly enough to support that column's own weight -- working through the balance gives P2-P1=ρgh, where ρ is the fluid's density. Taking the upper point to be the fluid's own open surface (where pressure equals the atmospheric pressure, Pa) gives the general result P=Pa+ρgh: the pressure at depth h below an open liquid surface exceeds atmospheric pressure by exactly ρgh, an amount called the gauge pressure. Notice something genuinely surprising buried in this formula: the container's own cross-sectional area, shape, or total volume of liquid never appears at all -- pressure at a given depth depends only on that depth itself. This produces the striking hydrostatic paradox: three vessels of wildly different shapes (one narrow, one wide, one oddly bulging), connected at their base and filled with water, all settle to precisely the same water level, however different the total amount of water each one actually holds. A swimmer 10 m below a lake's surface experiences this directly: with water density 1000 kg/m³, the pressure works out to P=Pa+ρgh=1.01x10⁵+1000x10x10=2.01x10⁵ Pa, very nearly double the pressure felt at the surface -- and at a full kilometre down, the increase reaches roughly 100 atmospheres, precisely the crushing pressure real submarines must be engineered to withstand.

The atmosphere's own pressure at any point equals the weight of the entire column of air stretching from that point up to the very top of the atmosphere -- at sea level, this works out to 1.013x10⁵ Pa, defined as exactly 1 atmosphere (atm). Evangelista Torricelli devised the first real method for measuring this: inverting a long glass tube, closed at one end and filled with mercury, into an open trough of mercury. The near-vacuum space left at the tube's sealed top exerts essentially zero pressure, so the mercury column's own weight, pressing down, must exactly balance the atmosphere's own push, upward, on the mercury in the open trough below -- giving Pa=ρgh, where h is simply the height of mercury the column settles at. This mercury barometer reliably settles at a height of about 76 cm at sea level, defining a whole family of convenient practical pressure units: 1 torr (named for Torricelli) equals the pressure of exactly 1 mm of mercury, about 133 Pa, while meteorologists commonly use the bar, with 1 bar defined as exactly 10⁵ Pa. If Earth's atmosphere had a genuinely constant density all the way up (its real sea-level value, 1.29 kg/m³), it would need to extend to a height of only about 8 km to produce the observed sea-level pressure -- though the atmosphere's real density actually thins out gradually with altitude, so it genuinely extends, at ever-decreasing pressure, to over 100 km. A submarine's window, 20 cm by 20 cm, at 1000 m ocean depth, illustrates just how much force real gauge pressure can generate: with seawater density 1.03x10³ kg/m³, the gauge pressure alone reaches about 1.03x10⁷ Pa, and multiplied by the window's own 0.04 m² area, the net force pressing inward on that single window works out to a formidable 4.12x10⁵ N.

Pascal's law carries a genuinely powerful practical consequence: pushing on any part of a fluid enclosed in a vessel transmits that extra pressure completely undiminished to every other part of the fluid, and to every wall of its container, whatever the vessel's own shape. A hydraulic lift exploits this directly: a small piston of area A1 pushes down on an enclosed liquid with force F1, creating a pressure P=F1/A1 that spreads, entirely unchanged, throughout the connected fluid -- reaching a second, much larger piston of area A2, where that same pressure now produces a correspondingly much larger upward force, F2=P x A2=F1 x (A2/A1). The ratio A2/A1 is the device's mechanical advantage: a small applied force at the narrow piston genuinely lifts a vastly heavier load at the wide one, exactly how a car lift or a hydraulic jack works, and hydraulic brakes use the identical principle, a gentle push on a car's brake pedal transmitting a far larger force to each wheel's own brake cylinder. Two connected syringes, with piston diameters 1.0 cm and 3.0 cm, show the multiplication directly: a modest 10 N force on the smaller piston produces 90 N at the larger one, exactly the ratio of their cross-sectional areas (9:1) -- though water's own near-total incompressibility means the trade-off runs both ways, since the smaller piston must be pushed in a full 9 times farther than the larger piston moves out.

A gently-opened water tap flows smoothly at first, but loses this smoothness once the flow speeds up -- the study of fluids actually in motion, fluid dynamics, begins by tracking the path each individual fluid particle actually follows. Flow is called steady if, at any one fixed point in space, every fluid particle passing through has exactly the same velocity as the particle before it -- the velocity can genuinely differ from one location to another, but at any single chosen location, it stays constant over time. The path traced by a fluid particle under steady flow is a streamline, defined so its own tangent, at any point, always points along the fluid's actual velocity there -- and since two streamlines crossing would mean a fluid particle arriving at that exact crossing point could have two different velocities at once, genuine streamlines in steady flow can never cross. Since the total mass of fluid flowing in must equal the total mass flowing out (mass itself cannot simply appear or disappear), the product of a pipe's cross-sectional area and the fluid's own speed through it, Av, must stay exactly constant all along the pipe, wherever the pipe's own width happens to change -- the equation of continuity. This single relation explains something genuinely counterintuitive from everyday experience: fluid speeds up precisely where a pipe (or a river channel) narrows, and slows down again wherever it widens, since the same volume must pass through a smaller cross-section in the same time. Steady, orderly streamline flow, called laminar flow, only persists below a certain critical speed -- push a fluid faster than this, and the flow abruptly loses its own steadiness altogether, breaking into the chaotic, swirling eddies of turbulent flow, exactly the foamy whirlpools visible where a fast-flowing stream crashes over rocks.

Since a steady-flowing incompressible fluid's speed must genuinely change wherever a pipe's cross-section changes (by the equation of continuity), some force has to actually cause that change in speed -- and that force comes directly from a difference in pressure between different regions of the pipe. Applying the work-energy theorem to a small element of fluid moving through a pipe of varying cross-section and height, and carefully tracking the work done on it by pressure at each end, alongside its changing kinetic and gravitational potential energy, leads to Bernoulli's equation: P+½ρv²+ρgh=constant, valid all along one single streamline. In words: moving along a streamline, the sum of pressure, kinetic energy per unit volume, and gravitational potential energy per unit volume never changes -- a direct statement of energy conservation applied to flowing fluid, valid so long as the fluid genuinely has no internal friction (non-viscous) and stays incompressible. One particularly clean special case is Torricelli's law: a fluid escaping from a small hole low in the side of a wide, open tank (where the surface above barely moves) exits with speed v=√(2gh), where h is the depth of the hole below the open surface -- exactly the same formula as an object simply falling freely under gravity through that same height h, since the escaping fluid, in a genuine sense, has effectively been in gravitational free-fall the whole way down to the hole.

Air is a fluid, exactly like water, and everything already established for fluids in this chapter, streamlines, Bernoulli's principle, pressure differences from changing speed, applies to it in exactly the same way -- which is precisely what makes a spinning ball curve and an aircraft wing generate lift. A ball moving through air without spinning leaves the streamlines above and below it genuinely symmetric, so the air speed (and hence pressure, by Bernoulli's principle) above and below stays identical, and the ball feels no sideways force from the air at all. A spinning ball, though, drags a thin layer of air along with its own spin -- on the side where the ball's own surface spin moves in the same direction as the oncoming air, the two motions add, crowding the streamlines closer together and speeding the local airflow up; on the opposite side, where surface spin opposes the oncoming air, the streamlines spread apart and the local airflow slows. By Bernoulli's principle, the faster-moving side develops lower pressure, and the slower side higher pressure, producing a genuine net sideways force -- the Magnus effect, exactly why a spinning cricket, tennis, baseball or golf ball visibly curves away from the simple parabolic path an ordinary, non-spinning projectile would follow. An aircraft wing's own cross-section, an aerofoil, achieves an equivalent effect through its shape rather than spin: oriented against the oncoming airflow, an aerofoil's particular curve crowds streamlines closer together above the wing than below it, so air moves measurably faster over the top surface than underneath -- creating a genuine, sustained upward pressure difference, dynamic lift, strong enough to support the aircraft's entire weight. A fully loaded Boeing aircraft, mass 3.3x10⁵ kg, with a total wing area of 500 m², flying level at 960 km/h (267 m/s), needs its own weight exactly balanced by this pressure difference: ΔP x A = mg gives ΔP≈6.5x10³ N/m², and working through Bernoulli's equation with air density 1.2 kg/m³ shows the air above the wing needs to move only about 8% faster than the air below it -- a genuinely modest speed difference, quietly keeping hundreds of tonnes of aircraft airborne.

Real fluids, unlike Bernoulli's own idealised non-viscous fluid, genuinely resist flowing -- an internal friction called viscosity, arising whenever adjacent layers of a fluid move at different speeds relative to one another and drag on each other as a result. Sandwiching oil between two glass plates, one fixed and one dragged sideways at a steady velocity v, shows this directly: the oil layer touching the moving plate moves right along with it at v, the layer touching the fixed plate stays at rest, and the layers in between settle into a smooth, uniform velocity gradient connecting the two -- honey, needing noticeably more force to drag the same plate at the identical speed, is simply more viscous than oil. Unlike a solid, where stress depends on strain itself, a fluid's internal stress depends instead on the rate at which strain is growing over time -- the coefficient of viscosity, η, is defined as the ratio of shearing stress to this strain rate, with SI unit the poiseuille (Pa·s). Real measured values span an enormous range: thin fluids like water and air are far less viscous than thick ones like honey or glycerine, and, strikingly, a liquid's viscosity generally falls as it warms up, while a gas's viscosity actually rises with temperature instead. A body falling through a viscous fluid drags along the fluid layer touching it, setting up relative motion between fluid layers that produces a genuine retarding drag force, given by Stokes' law: F=6πηav, proportional to the fluid's viscosity η, the sphere's radius a, and its own speed v. As a falling sphere (a raindrop, for instance) speeds up, this drag grows too, until it exactly balances gravity's own pull (adjusted for buoyancy) -- at that exact point, acceleration stops entirely, and the sphere descends at a fixed terminal velocity, vt=2a²(ρ-σ)g/(9η), where ρ and σ are the sphere's and fluid's own densities. A 2.0 mm copper ball, falling through oil at a measured terminal velocity of 6.5 cm/s, lets the oil's own viscosity be calculated directly from this formula, working out to about 0.99 Pa·s -- turning a simple falling-ball timing experiment into a genuine, practical viscosity measurement.

A molecule deep inside a liquid is pulled equally by neighbours in every direction, but a molecule right at the liquid's own surface has neighbours only below and to the sides, never above -- leaving it with less negative potential energy (roughly half) than a molecule fully surrounded, meaning surface molecules genuinely carry extra energy compared to those in the interior. Since creating more surface area means putting more molecules into this higher-energy surface state, a liquid naturally tends toward the smallest possible surface area its circumstances allow -- and among all shapes enclosing a given volume, a sphere has the least surface area of all, exactly why a free liquid drop, unaffected by gravity or other outside forces, is spherical. This same extra surface energy defines surface tension directly: stretching a liquid film by a small distance d, using a movable bar, increases the film's own total surface area, and the work done against this internal resisting force, divided by the extra area created (counting both sides of the film), gives the surface tension, S=F/2l, the extra energy per unit area, equally expressible as a force per unit length acting along the surface. Since a liquid touching a solid surface will genuinely wet that surface (spread out) only if doing so lowers the overall surface energy, some liquids wet certain solids readily (water on clean glass) while others do not (mercury on virtually any surface, water on a waxy lotus leaf) -- described by the angle of contact between the liquid's own surface and the solid, an acute angle for wetting liquids and an obtuse angle for non-wetting ones. Because a curved liquid surface always has higher pressure on its concave (inward-curving) side than its convex side, the pressure inside a spherical drop genuinely exceeds the pressure outside by 2S/r, and inside a soap bubble (which has two liquid surfaces, not just one) by fully 4S/r -- exactly why blowing a soap bubble takes a small but real extra push of air pressure. This same curved-surface pressure difference drives capillary rise: water climbing up a fine, narrow tube against gravity itself, rising to a height h given by hρg=2Scosθ/a, where a is the tube's own radius -- for a genuinely thin capillary (radius 0.05 cm), ordinary water climbs to a real, measurable height of about 2.98 cm, exactly the same physics letting sap rise through the narrow vessels inside a tall tree.