Noesis
4.4.25 · D5Multivariable Calculus

Question bank — Curl — definition, physical meaning (rotation)

1,681 words8 min read↑ Back to topic
🇮🇳 Yeh page Hinglish mein padho →

This bank is the conceptual sibling of the parent note Curl. It never asks you to grind a determinant (that lives in the D3/D4 computation banks) — it asks you why the machinery says what it says.

Before we start, three words must be earned so the reveals below never surprise you:

True or false — justify

True or false: If all the arrows of point in the same direction, the curl must be zero.
False — if the speed changes across the wheel (shear), it still spins. has parallel arrows yet curl because the bottom flows faster than the top.
True or false: Curl is a single number, just like divergence.
False — divergence is a scalar (net outflow), but rotation in 3D needs an axis, so curl is a vector. In 2D we keep only the component, which fools people into thinking it's scalar.
True or false: A field can have zero curl but nonzero divergence at the same point.
True — they answer independent questions. The radial field has divergence (spreading out) but curl (nothing twists the wheel).
True or false: A field can have nonzero curl but zero divergence everywhere.
True — pure rotation has curl and divergence . Swirl without spreading.
True or false: Every gradient field is irrotational.
True — because each component is by equality of mixed partials. Downhill flow cannot swirl.
True or false: Every irrotational field is the gradient of some scalar.
Only on a simply-connected domain — with a hole (like the whirlpool field around the -axis), curl can be zero everywhere the field is defined yet no global potential exists. See Gradient and Conservative Fields.
True or false: If curl is zero at a single point, the paddlewheel placed exactly there does not spin.
True at that point, but the value is local — one point being irrotational says nothing about its neighbours, and the wheel has zero size in the limiting definition.
True or false: Curl depends only on how curved the arrows look on a diagram.
False — curl is the circulation per unit area of a shrinking loop, which is set by velocity differences, not by visual curvature. Straight arrows can curl; radial (spreading) straight arrows do not.
True or false: Reversing the whole field () reverses the curl too.
True — curl is linear in , so . The wheel spins the opposite way, so the axis flips.
True or false: A stationary (constant) field has zero curl.
True — every partial derivative of a constant is , so all three curl components vanish. Uniform flow spins nothing.

Spot the error

" has straight arrows, so its curl is zero." — where's the slip?
The reasoner equated "straight arrows" with "no rotation." Curl reads the shear ; the wheel spins because the lower stream outruns the upper one.
"The -component of curl is ." — fix it.
Wrong pairing. It is : derivative of the next axis's component first, in cyclic order .
"Curl of a gradient is zero because gradients are always zero vectors." — fix it.
A gradient need not be zero; the reason is that its curl vanishes, since by Clairaut. The field itself is generally nonzero.
"Divergence of a curl is zero, therefore every divergence-free field is a curl." — spot the overreach.
The identity only says curls are divergence-free; the converse (every divergence-free field is a curl) needs extra topological conditions and is a separate theorem.
"To get the curl's magnitude I take the paddlewheel's angular speed." — fix it.
The magnitude is twice the local angular speed. For the wheel turns at angular speed but curl magnitude is .
"Since the field looks symmetric about the origin, its curl must be zero there." — spot the error.
Visual symmetry is not the test. is rotationally symmetric yet has curl — you must compare speeds across the wheel, not eyeball symmetry.
"I computed for one big loop, so curl is zero inside." — fix it.
One vanishing loop integral does not force pointwise curl to vanish; positive and negative swirl can cancel over a large region. Curl is the limit for shrinking loops. (This is the content of Stokes' Theorem / Green's Theorem.)

Why questions

Why is curl defined as circulation per unit area rather than just circulation?
Raw circulation grows with loop size, so it's not a property of a point; dividing by area and shrinking the loop isolates the local swirl density, which is what a point-sized wheel feels.
Why do we take the limit in the definition ?
A finite loop averages the swirl over a whole region; only shrinking it to a point removes that averaging and gives the curl exactly at the point.
Why does curl need a direction while divergence does not?
Rotation happens in a plane and has a spin axis, so you must say which plane — the picks it. Outflow (divergence) has no preferred direction, so a scalar suffices. See Divergence — definition and physical meaning.
Why does the counterclockwise loop give the curl rather than ?
Counterclockwise is the chosen positive orientation; by the right-hand rule its normal points out of the page along , so positive circulation registers as .
Why do both terms in for add instead of cancel?
Moving right speeds up the upward flow and moving up speeds up the leftward flow — these are the same counterclockwise spin seen two ways, so their contributions reinforce.
Why is interpreted as "a pure swirl has no source"?
A curl field is built entirely from circulation, which loops back on itself; nothing net flows out of any point, so its divergence is zero.
Why can two fields with identical pictures of arrows have different curls?
A diagram usually samples arrows on a coarse grid and hides how speed changes between arrows; curl reads exactly that hidden gradient, so equal-looking arrow plots can encode different shears.

Edge cases

The zero field : what are its curl and divergence, and is it irrotational?
Both curl and divergence are /; it is (trivially) irrotational and source-free — nothing flows, nothing spins.
A field defined only on a plane, : is its curl automatically along ?
Yes — the and components need -dependence or , both absent, so only survives. This is exactly the 2D setting of Green's Theorem.
At a point where flow speed is maximal along a streamline but equal on both sides of the wheel: curl there?
Zero in that plane — curl needs a difference across the wheel, not a change along the flow. Symmetric speed on the two flanks gives balanced torque.
The whirlpool field (undefined at the origin): what is its curl for ?
Exactly everywhere it is defined — it is irrotational off the axis, yet circulation around any loop enclosing the origin is . The origin is a singular point the domain must exclude.
As a paddlewheel shrinks toward a point, why doesn't its measured circulation just go to zero?
Circulation shrinks like the area (order ), so the ratio circulation/area tends to a finite nonzero limit — that finite limit is the curl component.
A field that is irrotational and divergence-free everywhere: what special name and structure does it have?
It is harmonic — locally with , combining zero swirl (gradient) and zero source (Laplace). Such fields model steady, idealized flow.
Recall One-line survival kit

Curl reads speed differences across a tiny wheel, returns a vector (axis = spin direction), is twice the angular speed, kills gradients, and is killed by divergence — none of which you can judge from how curved the arrows look.

Connections