Vector Calculus & the Green–Stokes–Gauss Theorems

Vector calculus connects three levels of description:

• a field assigns a scalar or vector to each point;
• derivatives such as gradient, divergence, and curl describe local change;
• integrals over curves, surfaces, and volumes measure accumulated effects.

Green’s, Stokes’, and the divergence theorem say that an integral of a derivative over a region equals an integral of the original field over the oriented boundary. Most mistakes are therefore not algebra mistakes but boundary, topology, or orientation mistakes.

The physical questions behind the notation #

The operators describe the same field at different scales: derivatives inspect one tiny neighborhood, while integrals add the effect along a path, across a surface, or through a volume.

Curves and line integrals #

Let a piecewise $C^1$ curve be parameterized by

$$r(t)=(x(t),y(t),z(t)),\qquad a\le t\le b.$$

For a scalar field $f$,

$$\int_C f\,ds =\int_a^b f(r(t))\|r'(t)\|\,dt.$$

This integral is independent of orientation because reversing the parameter changes $dt$ and $r'$ together, leaving $ds=\|r'\|dt$ positive.

For a vector field $F$,

$$\int_C F\cdot dr =\int_a^b F(r(t))\cdot r'(t)\,dt.$$

This integral changes sign when the curve orientation reverses. For a force field, it is work; around a closed curve, it is circulation.

The parameterization must trace the intended curve with the intended orientation. A nonregular point where $r'(t)=0$ can be acceptable in a piecewise-regular representation, but formulas that divide by speed cannot be used there.

Conservative fields and potentials #

A field $F$ is conservative on a domain $D$ if there is a scalar potential $\phi$ with

$$F=\nabla\phi.$$

Then the chain rule gives

$$\frac{d}{dt}\phi(r(t)) =\nabla\phi(r(t))\cdot r'(t),$$

so

$$\int_C F\cdot dr =\phi(r(b))-\phi(r(a)).$$

Consequences:

• the integral depends only on endpoints;
• every closed-loop integral is zero;
• work can be computed without following the path.

If $F=(P,Q)$ is $C^1$ and conservative, then $P_y=Q_x$. The converse needs topology. On an open simply connected domain, the equality implies conservativity. Without the no-hole condition it can fail.

A standard counterexample is

$$F(x,y)=\left(-\frac{y}{x^2+y^2},\frac{x}{x^2+y^2}\right)$$

on $\mathbb R^2\setminus\{0\}$. Its scalar curl is zero wherever defined, but its circulation around the unit circle is $2\pi$. The hole contains the obstruction.

Divergence and curl #

For $F=(P,Q,R)$,

$$\nabla\cdot F=P_x+Q_y+R_z$$

measures local net outflow per unit volume.

The curl

$$\nabla\times F = \begin{vmatrix} \mathbf i&\mathbf j&\mathbf k\\ \partial_x&\partial_y&\partial_z\\ P&Q&R \end{vmatrix}$$

measures local circulation density and axis.

These interpretations are local and differential. A pointwise zero divergence does not by itself say that flux through an arbitrary open surface is zero, and zero curl implies a global potential only with suitable domain hypotheses.

For sufficiently smooth fields,

$$\nabla\cdot(\nabla\times F)=0, \qquad \nabla\times(\nabla\phi)=0.$$

The smoothness matters because the identities use equality of mixed partial derivatives.

Surface area and flux #

Let an oriented parametric surface be

$$r(u,v),\qquad (u,v)\in D.$$

The tangent vectors $r_u,r_v$ span the tangent plane. Their cross product supplies both a normal direction and the area scale:

$$dS=\|r_u\times r_v\|\,du\,dv.$$

A scalar surface integral is

$$\iint_S f\,dS =\iint_D f(r(u,v))\|r_u\times r_v\|\,du\,dv.$$

The flux of $F$ through an oriented surface is

$$\iint_S F\cdot n\,dS =\iint_D F(r(u,v))\cdot(r_u\times r_v)\,du\,dv.$$

Swapping $u$ and $v$ reverses the cross product and hence the flux sign.

A surface must be orientable to choose a continuous unit normal globally. A Möbius strip has no global two-sided orientation, so an ordinary global flux integral is not defined without additional treatment.

Green’s theorem #

Let $D$ be a planar region with positively oriented boundary $\partial D$, meaning the region stays on the left as the boundary is traversed. Under the usual piecewise-smooth boundary and $C^1$ field hypotheses,

$$\oint_{\partial D}P\,dx+Q\,dy = \iint_D(Q_x-P_y)\,dA.$$

The flux form is

$$\oint_{\partial D}F\cdot n\,ds = \iint_D\nabla\cdot F\,dA.$$

For regions with holes, each inner boundary is oriented clockwise so that the region remains on the left. Simply writing “counterclockwise” for every component gives the wrong sign.

Stokes’ theorem #

Let $S$ be an oriented piecewise-smooth surface with positively induced boundary orientation $\partial S$, and let $F$ have continuous first partial derivatives on a neighborhood of $S$. Then

$$\oint_{\partial S}F\cdot dr = \iint_S(\nabla\times F)\cdot n\,dS.$$

The right-hand rule couples the normal to the boundary direction. If the fingers curl along $\partial S$, the thumb points along the chosen normal.

Stokes’ theorem depends only on the boundary curve and the field, not on which spanning surface is selected, provided the replacement surface stays in a region where $F$ is smooth and the orientation agrees. This often turns a difficult curved-surface integral into a simple planar one.

Worked Stokes example #

Take

$$F(x,y,z)=\left(-\frac y2,\frac x2,0\right)$$

and let $C$ be the unit circle in the $xy$-plane, counterclockwise from $+z$. Since

$$\nabla\times F=(0,0,1),$$

choose the unit disk $S$ with upward normal:

$$\oint_C F\cdot dr =\iint_S1\,dA =\pi.$$

Directly, $r(t)=(\cos t,\sin t,0)$, so

$$F(r(t))\cdot r'(t) =\left(-\frac{\sin t}{2},\frac{\cos t}{2},0\right) \cdot(-\sin t,\cos t,0) =\frac12,$$

and $\int_0^{2\pi}\frac12\,dt=\pi$, confirming the orientation and scale.

Divergence theorem #

Let $E$ be a solid with piecewise-smooth closed boundary $S=\partial E$, oriented outward. If $F$ has continuous first partial derivatives on an open neighborhood containing $E$, then

$$\iint_{\partial E}F\cdot n\,dS = \iiint_E\nabla\cdot F\,dV.$$

For $F=(x,y,z)$ on the ball of radius $R$,

$$\nabla\cdot F=3,$$

so

$$\iint_{\partial E}F\cdot n\,dS =3\mathop{\mathrm{Vol}}(E) =4\pi R^3.$$

On the sphere, $F=Rn$, so the direct calculation also gives $R\cdot4\pi R^2=4\pi R^3$.

The theorem applies to the complete closed boundary. If a side surface is open, cap it, apply the theorem to the closed surface, and subtract the cap flux with its induced outward orientation.

One boundary principle #

The theorems can be remembered structurally:

$$\int_{\partial\Omega}\text{field} = \int_\Omega\text{derivative of field}.$$

• Fundamental theorem of calculus: interval $\to$ endpoints.
• Fundamental theorem for line integrals: curve $\to$ endpoints.
• Green: planar region $\to$ boundary curve.
• Stokes: oriented surface $\to$ boundary curve.
• Divergence: volume $\to$ closed boundary surface.

This mnemonic does not replace hypotheses; it tells you which object and orientation to audit.

Failure modes #

• Omitting $\|r'(t)\|$ from a scalar line integral.
• Adding an absolute value to a vector line integral and destroying orientation.
• Inferring “curl-free implies conservative” on a domain with a hole.
• Using a potential across disconnected domain components without fixing constants separately.
• Treating $r_u\times r_v$ and $r_v\times r_u$ as interchangeable.
• Applying Green’s theorem with every boundary component counterclockwise.
• Applying Stokes without matching the curve direction to the surface normal.
• Applying the divergence theorem to an open surface without adding and subtracting caps.
• Ignoring a field singularity inside the region, even when the field is smooth on the boundary itself.
• Computing one symmetric half of a flux and reporting it as the whole-surface answer.