The Wave Equation and Vibrating Strings

Pluck a guitar string and watch it shiver into a blur. Then it settles back to rest. What equation governs the shape of the string at every instant? This question was first answered rigorously in the eighteenth century. It produced one of the most important partial differential equations in all of mathematics and physics. The one-dimensional wave equation describes not only vibrating strings but sound waves. It also governs electromagnetic waves along transmission lines. Countless other phenomena involve a disturbance propagating through a medium without dissipation. The wave equation matters because it is the simplest model exhibiting genuine wave propagation at finite speed. This feature is absent from the heat equation. The heat equation smooths disturbances instantly and infinitely fast. Understanding how the wave equation arises from Newton's second law builds physical intuition. This intuition applies to an idealized string. It transfers directly to higher-dimensional waves, quantum mechanics, and elasticity theory. The derivation itself is a model of applied mathematics. Physical assumptions are stated explicitly, and a differential element is isolated. Forces are balanced, and a limiting process converts a discrete statement into a continuum PDE. The big picture is this: a taut, flexible, elastic string vibrates with small transverse displacements. It obeys a linear, hyperbolic, second-order PDE with two independent variables. These variables are position and time. This equation belongs to the hyperbolic class discussed in the classification of second-order linear PDEs. Its hyperbolic character permits solutions that travel as waves with finite propagation speed. This stands in sharp contrast to the parabolic heat equation. It also contrasts with the elliptic Laplace equation. This article derives the wave equation from first principles using Newton's second law. It states the precise hypotheses under which the derivation is valid. It proves the central theorem rigorously, and then interprets the equation physically. The constant that appears is a wave speed. The equation supports two independent families of traveling disturbances. Boundary and initial conditions determine a unique solution. The article closes by pointing toward d'Alembert's explicit solution formula. It also points to the separation-of-variables technique. Both of these build directly on the derivation given here.

Key takeaways
  • A perfectly flexible, elastic string under tension TT vibrating with small transverse displacement u(x,t)u(x,t) satisfies utt=c2uxxu_{tt} = c^2 u_{xx}.
  • The constant c2=T/ρc^2 = T/\rho has units of speed squared, where ρ\rho is mass per unit length.
  • The derivation rests on Newton's second law applied to an infinitesimal string element, plus the small-slope (linearization) hypothesis.
  • The wave equation is hyperbolic, matching the general classification of second-order linear PDEs by the sign of B2−4ACB^2 - 4AC.
  • Physically, solutions decompose into disturbances traveling left and right at speed cc, unlike diffusive solutions of the heat equation.
  • Key idea: balancing transverse forces on a string element and linearizing for small slopes converts Newton's second law into the constant-coefficient hyperbolic PDE utt=c2uxxu_{tt} = c^2 u_{xx}.

Where this lives today

The wave equation extends far beyond music. In acoustics, the same derivation technique (balancing forces on an infinitesimal element and linearizing) produces the equation governing pressure waves in air, with cc the speed of sound. In electrical engineering, the telegrapher's equations for voltage and current on a transmission line reduce, in the lossless limit, to the wave equation, with cc determined by the line's inductance and capacitance per unit length. In elasticity theory, the same linearization technique yields wave equations for longitudinal and shear waves in solids, underlying seismology's use of P-waves and S-waves to probe the Earth's interior. In numerical analysis, finite-difference and finite-element schemes for the wave equation (such as the classical explicit leapfrog scheme) must respect the Courant–Friedrichs–Lewy condition, a stability constraint tying the time step to cc and the spatial mesh size, a direct descendant of the finite propagation speed proven here. In quantum field theory, the relativistic Klein–Gordon equation is a wave equation with an added mass term, and the massless case reduces exactly to the equation derived above. The techniques that follow directly from this derivation — d'Alembert's Solution of the Wave Equation and Separation of Variables for the Wave Equation — remain the standard tools for solving it explicitly, whether on the infinite line or on a finite interval with boundary conditions.

Ask the tutor: How does d'Alembert's formula turn the initial shape and velocity of a plucked string into an explicit traveling-wave solution of this equation?

DefinitionThe one-dimensional wave equation

Let u(x,t)u(x,t) denote the transverse displacement, at horizontal position x∈[0,L]x \in [0,L] and time t≥0t \geq 0, of a vibrating string from its equilibrium straight-line configuration. The one-dimensional wave equation is the second-order linear partial differential equation

∂2u∂t2(x,t)=c2∂2u∂x2(x,t),0<x<L, t>0,\frac{\partial^2 u}{\partial t^2}(x,t) = c^2 \frac{\partial^2 u}{\partial x^2}(x,t), \qquad 0 < x < L,\ t>0,

where c>0c > 0 is a constant called the wave speed. When the equation models a physical string of constant mass density ρ\rho (mass per unit length) under constant tension TT, the wave speed is given by

c=Tρ.c = \sqrt{\frac{T}{\rho}}.

A well-posed problem for the wave equation on a finite string additionally specifies two boundary conditions (for instance u(0,t)=u(L,t)=0u(0,t) = u(L,t) = 0 for a string fixed at both ends) and two initial conditions,

u(x,0)=f(x),ut(x,0)=g(x),0≤x≤L,u(x,0) = f(x), \qquad u_t(x,0) = g(x), \qquad 0 \le x \le L,

giving the initial shape ff and initial velocity gg of the string.

Intuition

Picture the string as a continuum of infinitesimal masses connected by tension, each free to move only vertically. Tension is the internal force that pulls neighboring pieces of string toward straightness; when the string is bent, the tension on the two ends of a small piece no longer point in exactly opposite directions, and the mismatch produces a net transverse force. Newton's second law then says that this net force equals mass times transverse acceleration. The wave equation is nothing more than the continuum limit of this force balance, together with the physical fact that for small displacements the tension's horizontal components stay essentially constant while the vertical components are proportional to the local slope of the string. The constant c=T/ρc = \sqrt{T/\rho} has a transparent physical meaning: tightening the string (increasing TT) increases the restoring force and speeds up vibrations, while a heavier string (larger ρ\rho) is more sluggish and slows them down. This matches everyday experience tuning a guitar: tightening a string raises its pitch, corresponding to a higher effective wave speed and hence higher frequencies of vibration, while thicker, heavier strings (larger ρ\rho, used for bass notes) vibrate more slowly. The picture below shows a snapshot of a string element and the tension forces acting on its two ends, tilted at slightly different angles because the string is curved.

Figure
A string element of length Δx with tension forces at slightly different angles at its two ends
T(x, t) T(x+Δx, t) θ(x) θ(x+Δx) Δx u(x,t): transverse displacement

The second figure illustrates why the equation is called hyperbolic: characteristic lines x−ct=constx - ct = \text{const} and x+ct=constx+ct = \text{const} in the (x,t)(x,t)-plane carry information at finite speed cc, foreshadowing d'Alembert's traveling-wave solution.

Figure
Characteristic lines in the (x,t) plane along which disturbances propagate at speed c
x t x−ct=const x+ct=const x₀
TheoremDerivation of the wave equation

Let a string occupy, at rest, the segment of the xx-axis from x=0x=0 to x=Lx=L. Assume: 1. (Perfect flexibility) The string offers no resistance to bending; the only internal force it can exert is tension directed tangent to the string. 2. (Constant linear density) The string has constant mass per unit length ρ>0\rho > 0, and no stretching occurs during vibration (the motion is purely transverse). 3. (No external transverse forces) Gravity and damping are neglected; the only forces acting are the internal tensions at the two ends of any string element. 4. (Planar, transverse motion) Each point of the string moves only vertically; u(x,t)u(x,t) denotes this vertical displacement, and horizontal motion of material points is negligible. 5. (Small slope) The displacement is small enough that ux(x,t)u_x(x,t) is small for all x,tx,t under consideration, so that sin⁡θ≈tan⁡θ=ux\sin\theta \approx \tan\theta = u_x and cos⁡θ≈1\cos\theta \approx 1 to the order of approximation used, where θ(x,t)\theta(x,t) is the angle the string makes with the horizontal at (x,t)(x,t). Under hypotheses 1–5, uu satisfies

utt=c2uxx,c2=T0ρ,u_{tt} = c^2 u_{xx}, \qquad c^2 = \frac{T_0}{\rho},

where T0T_0 is the (constant, to leading order) magnitude of the tension in the string.

Proof

We use the method of balancing forces on an infinitesimal element, then pass to the limit as the element's length tends to zero; this is the standard continuum-mechanics derivation technique.

Step 1: Isolate an element and write Newton's second law. Fix tt and consider the piece of string lying between xx and x+Δxx + \Delta x. By hypothesis 1, the string exerts only tangential tension forces, so the piece is acted on by tension T(x,t)T(x,t) pulling to the left at xx (directed into the piece from the left neighbor) and tension T(x+Δx,t)T(x+\Delta x, t) pulling to the right at x+Δxx + \Delta x, each tangent to the string at its respective point. Let θ(x,t)\theta(x,t) denote the angle the tangent to the string makes with the positive xx-axis at the point (x,u(x,t))(x, u(x,t)). By hypothesis 4, only vertical motion occurs, so we apply Newton's second law separately in the horizontal and vertical directions to the transverse displacement.

Step 2: Horizontal force balance. By hypothesis 4, the horizontal acceleration of every material point is zero (no horizontal motion), so the net horizontal force on the element is zero:

T(x+Δx,t)cos⁡θ(x+Δx,t)−T(x,t)cos⁡θ(x,t)=0.T(x+\Delta x, t)\cos\theta(x+\Delta x, t) - T(x,t)\cos\theta(x,t) = 0.

Since xx and Δx\Delta x are arbitrary, this says T(x,t)cos⁡θ(x,t)T(x,t)\cos\theta(x,t) is independent of xx; call this quantity T0(t)T_0(t). Under hypothesis 5, cos⁡θ≈1\cos\theta \approx 1 to leading order, so T(x,t)≈T0(t)T(x,t) \approx T_0(t) for all xx: the tension magnitude is spatially uniform to the order of approximation being used. We will further assume (consistent with small-amplitude, non-stretching vibration) that T0(t)=T0T_0(t) = T_0 is also constant in time; this is the standard idealization for a string vibrating about a fixed equilibrium tension.

Step 3: Vertical force balance and Newton's second law. Let Δm=ρ Δx\Delta m = \rho\, \Delta x be the mass of the element (hypothesis 2: mass per unit length constant, no stretching). The vertical component of Newton's second law for this element is

Δm ∂2u∂t2(xˉ,t)=T(x+Δx,t)sin⁡θ(x+Δx,t)−T(x,t)sin⁡θ(x,t),\Delta m\, \frac{\partial^2 u}{\partial t^2}(\bar{x}, t) = T(x+\Delta x,t)\sin\theta(x+\Delta x,t) - T(x,t)\sin\theta(x,t),

for some xˉ∈[x,x+Δx]\bar x \in [x, x+\Delta x], where we have used that the vertical acceleration of the element is well-approximated by uttu_{tt} evaluated at an interior point (this uses continuity of uttu_{tt} and the mean value theorem for integrals, applied to ∫xx+Δxρ utt ds\int_x^{x+\Delta x}\rho\,u_{tt}\,ds). By hypothesis 3, there are no other vertical forces.

Step 4: Linearize using the small-slope hypothesis. By hypothesis 5, sin⁡θ≈tan⁡θ=ux\sin\theta \approx \tan\theta = u_x to leading order (since tan⁡θ\tan\theta is exactly the slope of the string, this is exact by definition of θ\theta, and the approximation sin⁡θ≈tan⁡θ\sin\theta\approx\tan\theta is valid whenever θ\theta is small). Combining with Step 2 (T≈T0T \approx T_0 constant), the right side becomes

T0[ux(x+Δx,t)−ux(x,t)].T_0\big[u_x(x+\Delta x, t) - u_x(x,t)\big].

Step 5: Divide and take the limit Δx→0\Delta x \to 0. Substituting into the equation from Step 3 and dividing by Δx\Delta x:

ρ utt(xˉ,t)=T0 ux(x+Δx,t)−ux(x,t)Δx.\rho\, u_{tt}(\bar x, t) = T_0\,\frac{u_x(x+\Delta x,t) - u_x(x,t)}{\Delta x}.

Assume uxu_x is differentiable in xx (this is a smoothness hypothesis implicit in seeking a classical solution). As Δx→0\Delta x \to 0, xˉ→x\bar x \to x by the squeeze theorem, and the difference quotient on the right converges to uxx(x,t)u_{xx}(x,t) by the definition of the partial derivative. The left side converges to ρ utt(x,t)\rho\, u_{tt}(x,t) by continuity of uttu_{tt}. Hence

ρ utt(x,t)=T0 uxx(x,t),\rho\, u_{tt}(x,t) = T_0\, u_{xx}(x,t),

for every x∈(0,L)x \in (0,L) and t>0t > 0. Dividing by ρ>0\rho > 0 and setting c2:=T0/ρ>0c^2 := T_0/\rho > 0 gives

utt=c2uxx,u_{tt} = c^2 u_{xx},

which is the claimed equation. ■\blacksquare

∎

Worked examples

Example 1 — Computing the wave speed and a boundary/initial value problem for a cello string. A cello C-string has linear mass density ρ=3.2×10−3 kg/m\rho = 3.2 \times 10^{-3}\ \text{kg/m} and is stretched to tension T=55 NT = 55\ \text{N}, with length L=0.69 mL = 0.69\ \text{m} between the nut and the bridge. We are asked: (a) find the wave speed cc; (b) write the full initial-boundary value problem if the string is plucked at t=0t=0 into the (asymmetric) triangular shape

f(x)={ha x,0≤x≤a,hL−a(L−x),a≤x≤L,f(x) = \begin{cases} \dfrac{h}{a}\,x, & 0 \le x \le a,\\[4pt] \dfrac{h}{L-a}(L - x), & a \le x \le L,\end{cases}

with pluck point a=0.15 La = 0.15\ L (deliberately off-center) and released from rest. Solution. (a) By the theorem,

c=Tρ=553.2×10−3=17187.5≈131.1 m/s.c = \sqrt{\frac{T}{\rho}} = \sqrt{\frac{55}{3.2\times 10^{-3}}} = \sqrt{17187.5} \approx 131.1\ \text{m/s}.

(b) The string is fixed at both ends (clamped to the cello body), giving Dirichlet boundary conditions u(0,t)=u(L,t)=0u(0,t) = u(L,t) = 0. Released from rest means zero initial velocity, ut(x,0)=0u_t(x,0) = 0. The full problem is

utt=c2uxx,0<x<L, t>0,u_{tt} = c^2 u_{xx}, \quad 0<x<L,\ t>0,
u(0,t)=0,u(L,t)=0,t≥0,u(0,t) = 0,\quad u(L,t)=0, \quad t \ge 0,
u(x,0)=f(x),ut(x,0)=0,0≤x≤L,u(x,0) = f(x), \quad u_t(x,0)=0, \quad 0\le x\le L,

with c≈131.1c \approx 131.1 m/s, L=0.69L = 0.69 m, a=0.15(0.69)=0.1035a = 0.15(0.69) = 0.1035 m, and ff the asymmetric triangular pluck shape given above. Note the peak of ff occurs at x=a=0.1035x=a=0.1035 m, close to one end — this asymmetry is essential to producing the rich harmonic content characteristic of a string plucked near its end, and the equation derived above applies unchanged regardless of where the string is plucked, since the derivation used no assumption about the initial shape. This problem is precisely the type solved by separation of variables, using the wave speed and boundary data computed here as raw ingredients.

Example 2 — Checking a candidate solution and reading off a constraint on the wave speed. Consider a string of length L=2L=2 fixed at both ends. Suppose it is claimed that

u(x,t)=sin⁡(3πx)cos⁡(6πt)u(x,t) = \sin(3\pi x)\cos(6\pi t)

is a solution of utt=c2uxxu_{tt} = c^2 u_{xx} on this string for a particular value of cc, and satisfies the boundary conditions. Verify the boundary conditions, determine the unique value of cc for which uu solves the PDE, and check the initial velocity. Solution. Boundary conditions. At x=0x=0: u(0,t)=sin⁡(0)cos⁡(6πt)=0u(0,t) = \sin(0)\cos(6\pi t) = 0 for all tt, so the left condition holds. At x=L=2x=L=2: u(2,t)=sin⁡(6π)cos⁡(6πt)=0⋅cos⁡(6πt)=0u(2,t) = \sin(6\pi)\cos(6\pi t) = 0 \cdot \cos(6\pi t) = 0, since sin⁡(6π)=0\sin(6\pi) = 0. Both boundary conditions u(0,t)=u(2,t)=0u(0,t)=u(2,t)=0 are satisfied.

Substitute into the PDE. Compute the needed partial derivatives:

ut=−6πsin⁡(3πx)sin⁡(6πt),utt=−36π2sin⁡(3πx)cos⁡(6πt),u_t = -6\pi \sin(3\pi x)\sin(6\pi t), \qquad u_{tt} = -36\pi^2 \sin(3\pi x)\cos(6\pi t),
ux=3πcos⁡(3πx)cos⁡(6πt),uxx=−9π2sin⁡(3πx)cos⁡(6πt).u_x = 3\pi\cos(3\pi x)\cos(6\pi t), \qquad u_{xx} = -9\pi^2 \sin(3\pi x)\cos(6\pi t).

For utt=c2uxxu_{tt} = c^2 u_{xx} to hold identically (for all x∈(0,2)x\in(0,2), t>0t>0, not just at isolated points), we need, wherever sin⁡(3πx)cos⁡(6πt)≠0\sin(3\pi x)\cos(6\pi t) \neq 0,

−36π2=c2(−9π2)⟹c2=4⟹c=2-36\pi^2 = c^2(-9\pi^2) \quad\Longrightarrow\quad c^2 = 4 \quad\Longrightarrow\quad c = 2

(taking c>0c>0 as required by the definition). Since sin⁡(3πx)cos⁡(6πt)\sin(3\pi x)\cos(6\pi t) is not identically zero on (0,2)×(0,∞)(0,2)\times(0,\infty), this single value c=2c=2 is forced, and one checks directly that with c=2c=2 the equation utt=4uxxu_{tt} = 4u_{xx} holds identically: −36π2sin⁡(3πx)cos⁡(6πt)=4⋅(−9π2)sin⁡(3πx)cos⁡(6πt)-36\pi^2\sin(3\pi x)\cos(6\pi t) = 4\cdot(-9\pi^2)\sin(3\pi x)\cos(6\pi t), which is true since −36π2=−36π2-36\pi^2 = -36\pi^2.

Initial velocity. ut(x,0)=−6πsin⁡(3πx)sin⁡(0)=0u_t(x,0) = -6\pi\sin(3\pi x)\sin(0) = 0, so this solution corresponds to a string released from rest, consistent with g(x)≡0g(x) \equiv 0 in the general initial-value formulation. The initial shape is f(x)=u(x,0)=sin⁡(3πx)f(x) = u(x,0) = \sin(3\pi x), an oscillatory (three-half-wavelength) shape on [0,2][0,2] — this is one of the normal modes that separation of variables produces for this string, and it satisfies the wave equation only for the specific speed c=2c=2 computed here, illustrating how the PDE constrains the temporal frequency once the spatial mode and wave speed are fixed.

CounterexampleDropping the small-slope hypothesis breaks linearity of the equation

Suppose we retain hypotheses 1–4 of the theorem but drop hypothesis 5 (small slope), allowing θ(x,t)\theta(x,t) to be an arbitrary angle rather than a small one. Returning to Step 4 of the proof, the vertical force balance becomes, without linearization,

ρ utt=∂∂x[T0sin⁡θ(x,t)],tan⁡θ=ux,\rho\, u_{tt} = \frac{\partial}{\partial x}\Big[T_0\sin\theta(x,t)\Big], \qquad \tan\theta = u_x,

so that sin⁡θ=ux1+ux2\sin\theta = \dfrac{u_x}{\sqrt{1+u_x^2}} exactly (from the right-triangle relation between the tangent and sine of the slope angle). The vertical balance equation is then the genuinely nonlinear PDE

ρ utt=T0 ∂∂x(ux1+ux2)=T0 uxx(1+ux2)3/2.\rho\, u_{tt} = T_0\, \frac{\partial}{\partial x}\left(\frac{u_x}{\sqrt{1+u_x^2}}\right) = \frac{T_0\, u_{xx}}{(1+u_x^2)^{3/2}}.

This is not the wave equation: it is nonlinear because the coefficient of uxxu_{xx} depends on uxu_x itself, so superposition of solutions fails (if u1,u2u_1, u_2 solve this equation, u1+u2u_1+u_2 generally does not). Only when ∣ux∣≪1|u_x| \ll 1 can we Taylor-expand (1+ux2)−3/2≈1−32ux2+⋯≈1(1+u_x^2)^{-3/2} \approx 1 - \tfrac{3}{2}u_x^2 + \cdots \approx 1 and recover the linear equation ρutt≈T0uxx\rho u_{tt} \approx T_0 u_{xx} of the theorem. Physically, a string plucked so violently that its slope is not small (for example, pulled into a sharp near-vertical kink) genuinely vibrates according to this nonlinear equation, and predictions from the linear wave equation — such as the exact superposition of normal modes used in separation of variables — become inaccurate. This shows hypothesis 5 is not a technical nicety but is load-bearing: without it the derivation yields a different, harder equation with fundamentally different mathematical structure (no simple d'Alembert formula, no straightforward superposition).

Scope, edge cases & common traps

SituationWhat happensCorrect handling
Non-uniform density ρ(x)\rho(x)Equation becomes ρ(x)utt=T0uxx\rho(x) u_{tt} = T_0 u_{xx}, no longer constant-coefficientRe-derive; cc becomes position-dependent, c(x)2=T0/ρ(x)c(x)^2 = T_0/\rho(x)
Damping present (air resistance)An extra term +γut+\gamma u_t appears from a resistive force proportional to velocityUse the damped wave equation utt+γut=c2uxxu_{tt}+\gamma u_t = c^2 u_{xx}, not the undamped one
External transverse forcing (e.g. gravity, magnetic pluck)Hypothesis 3 fails; a forcing term F(x,t)/ρF(x,t)/\rho appears on the rightUse utt=c2uxx+F(x,t)/ρu_{tt} = c^2 u_{xx} + F(x,t)/\rho
Large-amplitude pluckingHypothesis 5 fails; equation becomes genuinely nonlinear (see counterexample)Use the nonlinear vibrating-string equation, or restrict to small amplitude
Confusing cc with the physical speed of any material pointMaterial points move only vertically; cc is the propagation speed of the waveform, not a particle speedDistinguish transverse particle velocity utu_t from wave speed cc

The mathematician behind it

Origins · 1 minJean le Rond d'Alembert and how this was discovered
JL
Jean le Rond d'Alembert
1717–1783 · Paris, Age of Enlightenment

D'Alembert was a French mathematician, physicist, and philosopher, a leading figure of the Enlightenment and co-editor of the Encyclopédie. In 1747 he derived the one-dimensional wave equation for a vibrating string and produced the first general solution to it, expressing the displacement as a sum of two traveling waveforms moving in opposite directions. His work ignited a famous and productive controversy with Euler, Daniel Bernoulli, and later Lagrange over what kinds of functions could legitimately represent the string's initial shape, a dispute that helped clarify the modern concept of a function and anticipated the theory of Fourier series.

History
- 1727: Johann Bernoulli studies a discrete chain of masses connected by springs, an early precursor to the continuous string.
- 1747: D'Alembert derives the wave equation for a vibrating string and gives its general traveling-wave solution.
- 1748–1749: Euler extends d'Alembert's analysis, insisting more general (non-analytic) initial shapes should be admissible.
- 1753: Daniel Bernoulli proposes that every vibration is a superposition of sinusoidal normal modes, seeding separation of variables.
- 1822: Fourier's treatise on heat systematizes the trigonometric series methods that resolve the Bernoulli–d'Alembert–Euler dispute.
Why is only the transverse (vertical) displacement modeled, when a real string can also stretch?

The derivation assumes purely transverse motion and no stretching (hypotheses 2 and 4) because, for small-amplitude vibrations of an already-taut string, longitudinal displacement is a much smaller effect than transverse displacement — it is a second-order correction. Physically, tension in a slightly stretched elastic string responds strongly to small length changes, so any incipient stretching is suppressed almost instantly compared to the timescale of transverse oscillation. Keeping only transverse displacement is therefore an excellent approximation for the musically and physically relevant regime of small-amplitude vibration, and it is exactly what makes the resulting equation linear and analytically tractable, in contrast to the fully nonlinear elastic string equation.

Does the wave equation apply to a string with variable tension or density, like a tapered instrument string?

Not in the constant-coefficient form given here. If density ρ(x)\rho(x) or tension T0(x)T_0(x) vary with position, the same force-balance argument (Steps 1–5 of the proof) still applies locally, but Step 2 no longer gives a spatially uniform tension, and the final equation becomes ρ(x)utt=∂x(T0(x)ux)\rho(x) u_{tt} = \partial_x(T_0(x) u_x), a variable-coefficient hyperbolic PDE. This equation is still linear and still supports finite-speed wave propagation, but it generally cannot be solved by elementary traveling waves; techniques like Sturm–Liouville theory replace simple separation of variables.

What happens physically if the small-slope hypothesis fails, such as when a string is plucked extremely hard?

As shown in the counterexample, dropping the small-slope assumption produces a genuinely nonlinear PDE, ρutt=T0uxx/(1+ux2)3/2\rho u_{tt} = T_0 u_{xx}/(1+u_x^2)^{3/2}, rather than the linear wave equation. Physically this means the restoring force per unit displacement depends on how steep the string already is, so the principle of superposition fails: you cannot add two known vibration patterns to get a new one. Large-amplitude plucking of a real string does exhibit measurable nonlinear effects, such as amplitude-dependent pitch shifts, which the linear wave equation cannot predict at all.

How does the wave equation relate to the heat equation covered under the classification of second-order linear PDEs?

Both are constant-coefficient second-order linear PDEs in one space and one time variable, but they belong to different classes under the discriminant-based classification: the wave equation utt−c2uxx=0u_{tt} - c^2 u_{xx} = 0 has B2−4AC=4c2>0B^2-4AC = 4c^2 > 0, making it hyperbolic, while the heat equation ut−kuxx=0u_t - k u_{xx}=0 is first order in time and parabolic. The practical consequence is stark: the wave equation propagates information at exactly finite speed cc along characteristics, while the heat equation instantaneously (if very weakly) affects every point, however far away, the moment a disturbance is introduced.

Why do boundary conditions and initial conditions each need to be specified twice (two boundary, two initial) for a well-posed problem?

The wave equation is second order in both xx and tt. Just as a second-order ordinary differential equation in time needs two initial conditions (position and velocity) to pin down a unique solution, the wave equation needs an initial position f(x)f(x) and an initial velocity g(x)g(x) to determine the future evolution uniquely; this matches the two arbitrary functions that appear in d'Alembert's general solution. Likewise, being second order in xx requires two boundary conditions, one at each end of the string, to select a unique spatial solution consistent with how the string is physically held or free.

Can the same derivation be adapted to justify the wave equation for sound or electromagnetic transmission lines?

Yes, and this is exactly why the equation is so important. For sound in a gas, one balances forces on an infinitesimal slab of gas using Newton's second law together with the linearized equation of state, playing the role that tension plays for the string; the result is again utt=c2uxxu_{tt}=c^2u_{xx} with cc the speed of sound. For a lossless transmission line, Kirchhoff's laws applied to an infinitesimal segment of the line, combined with linearized inductance and capacitance relations, produce the same equation with c=1/LCc = 1/\sqrt{LC} per unit length. The common thread is always a linear restoring mechanism combined with inertia, linearized over infinitesimal elements.

Before this

Where this leads