Tangents at Self-Intersections of Polar Curves

Why a Repeated Point Needs Two Checks

A polar curve can pass through the same point at two different parameter values. At that point, the curve may cross itself with distinct tangent lines, touch itself with a shared tangent line, or retrace part of a previously traced path. To tell these situations apart, first find the parameter values that represent the same Cartesian point, then compare the tangent directions there.

The first check matters because polar coordinates allow a negative radius. The second matters because two different parameter values can differ by more than one half-turn, especially when the parameter domain includes several turns or the radius function is not periodic.

Definition: For a polar curve \(r=r(\theta)\) on a domain \(D\), a self-intersection point is a point represented by two distinct parameters \(\theta_1,\theta_2\in D\). If the curve is regular at both parameters, the intersection is transverse when its two tangent lines are distinct, and tangent when the tangent lines coincide.

When Do Two Polar Parameters Represent the Same Point?

The Cartesian point represented by parameter \(\theta\) is

\boldsymbol{\gamma}(\theta) = \bigl(r(\theta)\cos\theta,\;r(\theta)\sin\theta\bigr).

For a nonzero point, equality of two such Cartesian representations has two possibilities: the angles agree modulo \(2\pi\) and the signed radii agree, or the angles differ by an odd multiple of \(\pi\) and the signed radii have opposite signs. The pole is a separate case: every parameter with radius zero represents the origin, regardless of its angle.

Criterion: Distinct parameters \(\theta_1,\theta_2\in D\) represent the same point precisely when either \(r(\theta_1)=r(\theta_2)=0\), or one of these conditions holds:

\theta_2\equiv\theta_1\pmod{2\pi} \quad\text{and}\quad r(\theta_2)=r(\theta_1),

\theta_2\equiv\theta_1+\pi\pmod{2\pi} \quad\text{and}\quad r(\theta_2)=-r(\theta_1).

In every case, require \(\theta_1\ne\theta_2\). The parameters must also lie in the domain being considered.

The congruences in this criterion are important. For example, two angles can differ by \(3\pi\), not just by \(\pi\), and still differ by an odd multiple of \(\pi\). Replacing the congruence condition with a single equation involving \(\theta+\pi\) can miss repeated points when the domain includes more than one turn or \(r\) is not \(2\pi\)-periodic.

Example: A Repeated Point Separated by Three Half-Turns

Worked Example: Checking a Pair on a Multi-Turn Domain

Let \(r(\theta)=1-\frac{2\theta}{3\pi}\) on \(D=[0,4\pi]\). Check whether \(\theta_1=0\) and \(\theta_2=3\pi\) represent the same point.

At the first parameter, \(r(0)=1\), so

\boldsymbol{\gamma}(0) = (1\cos 0,\;1\sin 0) = (1,0).

At the second parameter,

r(3\pi) = 1-\frac{2(3\pi)}{3\pi} = 1-2 = -1.

Since \(3\pi\equiv 0+\pi\pmod{2\pi}\) and \(r(3\pi)=-r(0)\), the criterion applies. Direct substitution confirms the point:

\boldsymbol{\gamma}(3\pi) = \bigl((-1)\cos 3\pi,\;(-1)\sin 3\pi\bigr) = (1,0).

Thus these two distinct parameters represent the same nonzero point. Checking only \(r(\theta+\pi)=-r(\theta)\) would not test this pair, since the parameter difference is \(3\pi\), not exactly \(\pi\).

A Complete Search for Candidate Parameters

To find self-intersections, work on the stated parameter domain rather than assuming the curve is traced over just one turn. First identify multiple parameters at the pole by solving \(r(\theta)=0\). Then examine distinct parameter pairs satisfying the two angle-and-radius alternatives in the criterion. If the domain contains multiple turns, also check for repeated parameters whose angles agree modulo \(2\pi\). A periodic radius function can make such repetitions common, but periodicity should not be assumed unless it is given or established.

Check the pole.
Find every parameter in \(D\) for which \(r(\theta)=0\). Any two distinct such parameters represent the origin.

Check angles equal modulo \(2\pi\).
For distinct parameters with \(\theta_2\equiv\theta_1\pmod{2\pi}\), test whether their signed radii agree.

Check angles opposite modulo \(2\pi\).
For distinct parameters with \(\theta_2\equiv\theta_1+\pi\pmod{2\pi}\), test whether their signed radii are opposites.

Compare tangents at each repeated point.
Compute the tangent vectors at both parameters. If either vector is zero, the regular tangent-line comparison requires further analysis.

The Tangent Vector and the Classification Test

Writing the Cartesian coordinates as \(x(\theta)=r(\theta)\cos\theta\) and \(y(\theta)=r(\theta)\sin\theta\), differentiate each coordinate. The resulting vector gives the direction of motion along the curve. This computation assumes \(r\) is differentiable at the parameter in question.

\boldsymbol{\gamma}'(\theta) = \bigl(r'(\theta)\cos\theta-r(\theta)\sin\theta,\; r'(\theta)\sin\theta+r(\theta)\cos\theta\bigr).

At a regular parameter, \(\boldsymbol{\gamma}'(\theta)\ne\mathbf{0}\), and the tangent line has this vector as a direction. Two nonzero vectors in the plane determine the same line exactly when they are scalar multiples of one another. Equivalently, their determinant is zero.

Theorem: Suppose \(\theta_1\ne\theta_2\) represent the same point and the curve is regular at both parameters. Write \(\mathbf{v}_1=\boldsymbol{\gamma}'(\theta_1)\) and \(\mathbf{v}_2=\boldsymbol{\gamma}'(\theta_2)\). The intersection is transverse if

\det(\mathbf{v}_1,\mathbf{v}_2)\ne 0,

and tangent if \(\det(\mathbf{v}_1,\mathbf{v}_2)=0\). Here, for \(\mathbf{v}_1=(a,b)\) and \(\mathbf{v}_2=(c,d)\), \(\det(\mathbf{v}_1,\mathbf{v}_2)=ad-bc\).

Why the Determinant Test Works

The determinant test follows directly from the condition for two direction vectors to be parallel. Suppose both vectors are nonzero. If \(\mathbf{v}_2=k\mathbf{v}_1\) for some nonzero real number \(k\), then

\det(\mathbf{v}_1,\mathbf{v}_2) = \det(\mathbf{v}_1,k\mathbf{v}_1) = k\det(\mathbf{v}_1,\mathbf{v}_1) = 0.

Conversely, if \(\mathbf{v}_1=(a,b)\ne\mathbf{0}\) and \(\det(\mathbf{v}_1,\mathbf{v}_2)=ad-bc=0\), where \(\mathbf{v}_2=(c,d)\), then \(\mathbf{v}_1\) and \(\mathbf{v}_2\) are linearly dependent. Since \(\mathbf{v}_2\ne\mathbf{0}\), one is a nonzero scalar multiple of the other. They therefore determine the same tangent line. A nonzero determinant means the two direction vectors are not parallel, so the tangent lines are distinct and the intersection is transverse.

Example: Two Transverse Intersections

Worked Example: Comparing Tangents at an Isolated Crossing

Consider \(r(\theta)=\sin\theta+\cos 2\theta\) on the domain \(D=[\frac{\pi}{6},\frac{\pi}{3}]\cup[\frac{6\pi}{5},\frac{4\pi}{3}]\). This restricted domain isolates a single repeated point. For a pair from the first and second intervals, the angle difference can represent opposite directions only when \(\theta_2=\theta_1+\pi\). The signed-radius condition becomes

r(\theta_1+\pi)=-r(\theta_1) \quad\Longleftrightarrow\quad -\sin\theta_1+\cos 2\theta_1 = -\sin\theta_1-\cos 2\theta_1,

so \(\cos 2\theta_1=0\). In the first interval, the only solution is \(\theta_1=\frac{\pi}{4}\), and its partner is \(\theta_2=\frac{5\pi}{4}\). Neither parameter is a zero of \(r\), so this is not a pole intersection. The common point is

r\left(\frac{\pi}{4}\right)=\frac{\sqrt{2}}{2}, \qquad \boldsymbol{\gamma}\left(\frac{\pi}{4}\right) = \left(\frac12,\frac12\right),

and the opposite signed radius at \(5\pi/4\) gives that same point.

Differentiate \(r\): \(r'(\theta)=\cos\theta-2\sin 2\theta\). At \(\theta_1=\frac{\pi}{4}\), \(r'(\theta_1)=\frac{\sqrt{2}}{2}-2\). Substitution into the Cartesian tangent-vector formula gives

\mathbf{v}_1 = \left(-\sqrt{2},\;1-\sqrt{2}\right).

At \(\theta_2=\frac{5\pi}{4}\), \(r'(\theta_2)=-\frac{\sqrt{2}}{2}-2\), and substitution gives

\mathbf{v}_2 = \left(\sqrt{2},\;1+\sqrt{2}\right).

Both vectors are nonzero. Their determinant is

\det(\mathbf{v}_1,\mathbf{v}_2) = (-\sqrt{2})(1+\sqrt{2}) - (1-\sqrt{2})(\sqrt{2}) = (-\sqrt{2}-2)-(\sqrt{2}-2) = -2\sqrt{2}\ne 0.

The tangent lines are distinct, so the repeated point is a transverse crossing.

Example: Tangency Caused by Retracing

Worked Example: A Curve Traced Twice with the Same Tangent

Let \(r(\theta)=\sin\theta\) on \(D=[0,2\pi)\). For every \(\theta\in[0,\pi)\), the distinct parameter \(\theta+\pi\) also lies in \(D\), and

r(\theta+\pi)=\sin(\theta+\pi)=-\sin\theta=-r(\theta).

The two parameters differ by \(\pi\), so they represent the same point. In fact, the Cartesian coordinates simplify to

x(\theta)=\sin\theta\cos\theta, \qquad y(\theta)=\sin^2\theta.

Both coordinates are unchanged when \(\theta\) is replaced by \(\theta+\pi\). Differentiating gives

\boldsymbol{\gamma}'(\theta) = \bigl(\cos 2\theta,\;\sin 2\theta\bigr), \qquad \boldsymbol{\gamma}'(\theta+\pi) = \bigl(\cos 2\theta,\;\sin 2\theta\bigr).

The vector has length \(1\), since \(\cos^2 2\theta+\sin^2 2\theta=1\), so the curve is regular at both parameters. The vectors are equal and nonzero; their determinant is zero. The tangent lines coincide. This parametrization retraces the same curve, rather than making isolated transverse crossings. At the pole, the pair \(\theta=0,\pi\) also represents the origin, and the same vector calculation gives the same tangent line there.

Common Pitfall: Confusing a Zero Derivative with Tangency

A zero tangent vector does not establish that the curve has a tangent line shared with another branch. The determinant test assumes two nonzero tangent vectors. For a parameter where \(\boldsymbol{\gamma}'(\theta)=\mathbf{0}\), the first-order direction is unavailable; the parameter is singular, and its local tangent behavior must be analyzed separately. In particular, treating a zero determinant as proof of tangency without first checking that both vectors are nonzero is invalid.

Practical check: Before classifying a repeated point, verify that the parameters are distinct and that both tangent vectors are nonzero. Then compare their lines, not merely the orientations in which the parametrization travels along them. Opposite vectors still determine the same tangent line.

Why This Classification Matters

Identifying a repeated point is a geometric question, while classifying its tangents is a local question. The same point can be reached with different directions, producing a crossing, or with the same direction line, producing tangency or retracing. Keeping these questions separate helps prevent two common errors: overlooking a repeated point because the parameters differ by more than \(\pi\), and calling two branches transverse merely because their tangent vectors point in opposite directions.

When working with a polar curve, state the parameter domain and keep track of signed radii. Use the full angle congruences to find candidate pairs, handle the pole explicitly, and apply the determinant only after confirming regularity. These checks make the classification precise even when the curve is traced more than once.

Check Your Understanding

Use the same-point criterion and tangent-line test to answer these questions.

What radius condition accompanies \(\theta_2\equiv\theta_1+\pi\pmod{2\pi}\) for two non-pole polar representations to agree?
Why can checking only parameter pairs separated by exactly \(\pi\) miss a repeated point on a multi-turn domain?
What must be checked before using a determinant to classify a self-intersection as transverse or tangent?
If two nonzero tangent vectors have determinant zero, what does that tell you about their tangent lines?
For \(r(\theta)=\sin\theta\) on \([0,2\pi)\), why do \(\theta\) and \(\theta+\pi\) represent the same point whenever \(0\le\theta<\pi\)?