Choosing a Conservativeness Test

Introduction

There is no single conservativeness test that is best for every vector field. A field’s formula may make direct potential construction quick; its domain may make a curl test decisive; and a closed loop may reveal nonconservativeness even when the curl vanishes.

The key is to choose a test whose conclusion is justified by the field’s regularity and the topology of its domain. A quick component check can rule out a potential, but passing that check does not always prove one exists. Similarly, zero curl is decisive only with suitable domain conditions.

Central principle: A failed necessary condition rules out a potential. A successful test establishes a potential only when its additional hypotheses are satisfied.

What Is Being Tested?

Let \(D\) be an open subset of \(\mathbb{R}^2\) or \(\mathbb{R}^3\), and let \(\mathbf F\) be a vector field on \(D\). The field is conservative if there is a differentiable scalar function \(\phi\) on \(D\) such that \(\mathbf F=\nabla\phi\). Such a function \(\phi\) is called a potential.

In the plane, write \(\mathbf F=(P,Q)\), so \(\mathbf F=\nabla\phi\) means \(P=\phi_x\) and \(Q=\phi_y\). In three dimensions, write \(\mathbf F=(P,Q,R)\), so the corresponding conditions are \(P=\phi_x\), \(Q=\phi_y\), and \(R=\phi_z\). The line integral of \(\mathbf F\) along a piecewise smooth curve \(C\) is denoted by \(\int_C\mathbf F\cdot d\mathbf r\).

Two useful consequences of conservativeness are path independence and zero circulation around closed curves. A component or curl test checks a local differential condition; a loop test checks an integral around a particular curve. These tests provide different kinds of information.

Component Compatibility: A Quick Necessary Test

Suppose a planar field has a potential \(\phi\) with continuous second partial derivatives. Then \(P=\phi_x\) and \(Q=\phi_y\), so equality of mixed partial derivatives gives \(P_y=\phi_{xy}=\phi_{yx}=Q_x\). Thus \(P_y=Q_x\) is necessary for a potential under this regularity assumption. In three dimensions, the corresponding necessary condition is \(\nabla\times\mathbf F=\mathbf 0\).

These conditions are useful early checks. If a required equality fails at even one point of the domain, the field cannot be the gradient of a sufficiently smooth potential there. If the equality holds, however, the check alone may not settle the question: the domain matters, and one may need to construct a potential or use another argument.

Decision rule: Use component compatibility first when the component formulas are simple to differentiate. A mismatch is conclusive evidence against a potential; a match is a reason to continue, not automatically a proof of conservativeness.

Worked Example: Ruling Out a Potential by Compatibility

Consider \(\mathbf F(x,y)=(xy,0)\) on \(D=\mathbb R^2\). Here \(P=xy\) and \(Q=0\). Differentiate the components:

P_y=x, \qquad Q_x=0.

At the point \((1,0)\), \(P_y=1\) while \(Q_x=0\). The necessary compatibility condition fails, so this field cannot be the gradient of a potential with continuous second partial derivatives on \(D\).

Check: The mismatch is at a point in the domain, not merely outside it. Since a gradient field with the stated regularity would have equal mixed partials there, the failure rules out a potential.

What the Curl Test Can Establish

For a continuously differentiable planar field \(\mathbf F=(P,Q)\), the scalar curl is \(Q_x-P_y\). Thus the planar curl vanishes exactly when \(P_y=Q_x\). For a continuously differentiable three-dimensional field, the curl is \(\nabla\times\mathbf F\), and the corresponding compatibility condition is \(\nabla\times\mathbf F=\mathbf 0\).

Vanishing curl is necessary for a conservative \(C^1\) field. It is also sufficient when the domain is open and simply connected. Here, simply connected means, informally, that the domain is connected and has no holes that prevent closed curves from being continuously contracted to a point while staying in the domain. The condition matters: on a domain with a hole, zero curl need not imply a global potential.

Use curl carefully: For a \(C^1\) field on an open simply connected domain, zero curl implies the field is conservative. On a domain with holes, zero curl by itself does not settle whether a global potential exists.

When Direct Potential Construction Is Best

Sometimes the field components make it straightforward to build a candidate potential. In the plane, one can integrate \(P\) with respect to \(x\), treating \(y\) as fixed, and include an unknown function of \(y\). Differentiating that candidate with respect to \(y\) then determines whether it can also match \(Q\). In three dimensions, potential recovery similarly requires matching all three components.

A constructed candidate must be checked by differentiating it and recovering every component of the original field. If this succeeds throughout the domain, it directly proves that the field is conservative; no separate curl sufficiency test is needed.

Worked Example: Constructing and Verifying a Potential

Let \(\mathbf F(x,y)=(2xy+3,\ x^2+4y)\) on \(\mathbb R^2\). Set \(P=2xy+3\) and \(Q=x^2+4y\). Integrate \(P\) with respect to \(x\):

\phi(x,y)=\int(2xy+3)\,dx=x^2y+3x+g(y),

where \(g(y)\) is an unknown function because differentiation with respect to \(x\) eliminates any term depending only on \(y\). Differentiate the candidate with respect to \(y\) and match it with \(Q\):

\phi_y=x^2+g'(y)=x^2+4y, \qquad g'(y)=4y, \qquad g(y)=2y^2+C.

We may take \(C=0\), since adding a constant does not change the gradient. The candidate is \(\phi(x,y)=x^2y+3x+2y^2\). Verify both components:

\phi_x=2xy+3=P, \qquad \phi_y=x^2+4y=Q.

Therefore \(\nabla\phi=\mathbf F\) throughout \(\mathbb R^2\), so the field is conservative.

Check: Differentiating the proposed potential reproduces the first and second components exactly, including their constant and \(y\)-dependent terms.

Closed Loops: A Direct Way to Disprove Conservativeness

If \(\mathbf F=\nabla\phi\) and \(C\) is a piecewise smooth closed curve, the chain rule gives \(\mathbf F(\mathbf r(t))\cdot\mathbf r'(t)=\frac{d}{dt}\phi(\mathbf r(t))\) on each smooth piece. Integrating along the pieces gives the total change in \(\phi\); because the curve ends where it began, that change is zero. Consequently, a single closed curve with nonzero line integral proves that the field is not conservative.

The converse requires care. One zero integral, or even zero integrals around a few selected loops, does not prove that all closed-loop integrals vanish. And an assertion that all closed-loop integrals vanish implies a potential only with appropriate hypotheses.

Worked Example: Curl Zero but a Loop Integral Is Nonzero

Consider \(\mathbf F(x,y)=\left(-\frac{y}{x^2+y^2},\frac{x}{x^2+y^2}\right)\) on the punctured plane \(D=\mathbb R^2\setminus\{(0,0)\}\). Write \(P=-y/(x^2+y^2)\) and \(Q=x/(x^2+y^2)\). Their relevant partial derivatives are

P_y=\frac{y^2-x^2}{(x^2+y^2)^2}, \qquad Q_x=\frac{y^2-x^2}{(x^2+y^2)^2}.

Thus \(P_y=Q_x\), so the planar curl vanishes everywhere in the domain. But the domain has a hole, so the simply connected sufficiency condition does not apply. Test the unit circle, traversed counterclockwise: \(\mathbf r(t)=(\cos t,\sin t)\), \(0\le t\le2\pi\). Then \(\mathbf r'(t)=(-\sin t,\cos t)\), and the field along the curve is \(\mathbf F(\mathbf r(t))=(-\sin t,\cos t)\). Therefore

\int_C\mathbf F\cdot d\mathbf r =\int_0^{2\pi}(-\sin t,\cos t)\cdot(-\sin t,\cos t)\,dt =\int_0^{2\pi}1\,dt =2\pi.

The nonzero loop integral proves that \(\mathbf F\) is not conservative on the punctured plane, despite its zero curl there.

Check: The parametrization stays in the domain, and the dot product is \(\sin^2t+\cos^2t=1\). Its integral is \(2\pi\), not zero.

Why Vanishing Integrals Around All Loops Need Regularity

On an open path-connected domain \(D\), path independence means that the line integral between two fixed endpoints has the same value along every piecewise smooth path in \(D\). If all closed piecewise smooth loops have zero integral, then path independence follows: follow one path from \(A\) to \(B\), then the reverse of another path from \(A\) to \(B\). This makes a closed loop, whose integral is the difference between the two path integrals.

To obtain a potential from path independence, assume also that \(\mathbf F\) is continuous on \(D\). Fix a base point \(A\) and define \(\phi(X)\) as the line integral from \(A\) to \(X\). Path independence makes this definition unambiguous. Since \(D\) is open, sufficiently short line segments from \(X\) in each coordinate direction remain in \(D\). Continuity of \(\mathbf F\) then gives the corresponding partial derivatives of \(\phi\) as the components of \(\mathbf F\), so \(\nabla\phi=\mathbf F\). Thus continuity is part of this implication, not an optional detail.

Precise conclusion: If \(\mathbf F\) is continuous on an open path-connected domain and its integral around every closed piecewise smooth curve is zero, then its line integral is path independent and \(\mathbf F\) has a potential.

A Practical Selection Process

Use the formula and domain to decide what to try, then verify the conclusion the chosen test actually supports.

Check the components.
For a planar field, compare \(P_y\) and \(Q_x\); in three dimensions, check the curl. A mismatch rules out a potential under the relevant differentiability assumptions.

Inspect the domain.
If the field is \(C^1\), the domain is open and simply connected, and the curl vanishes, the curl test establishes conservativeness.

Try constructing a potential.
When the component formulas are easy to integrate, build a candidate and differentiate it to check every component. A successful global verification proves conservativeness directly.

Test a closed loop when useful.
A nonzero integral around one valid closed curve disproves conservativeness. Zero on a selected loop is not enough to prove it.

Match the conclusion to the hypotheses.
For the all-loops criterion to yield a potential, use an open path-connected domain and a continuous field.

Common Pitfall: Treating Zero Curl as a Universal Test

A tempting but incorrect argument is: “The curl is zero, so the field is conservative.” For the punctured-plane example, this calculation establishes only that the curl vanishes on the domain. It does not establish that the domain is simply connected. The unit-circle integral \(2\pi\) supplies a concrete contradiction to conservativeness.

The correction is to state the domain condition before using the sufficiency direction: for a \(C^1\) field on an open simply connected domain, zero curl implies a potential. On a domain with a hole, investigate further—for example, attempt potential construction or compute a closed-loop integral. A failed necessary test is decisive; a passed test must be interpreted with its hypotheses in view.

Check Your Understanding

Use the hypotheses and conclusions of the tests, rather than relying on a remembered slogan.

For a planar field with a potential having continuous second partial derivatives, what component equality must hold?
What additional domain condition makes zero curl sufficient for a \(C^1\) field to be conservative?
Why does one nonzero closed-loop integral rule out a potential?
Does a zero integral around one selected loop prove path independence? Explain.
What regularity and domain assumptions allow vanishing integrals around all closed piecewise smooth curves to imply a potential?