3.4.11 · D4Conic Sections

Exercises — General second-degree equation Ax²+Bxy+Cy²+Dx+Ey+F=0 — discriminant classification

3,087 words14 min readBack to topic

Level 1 — Recognition

(Read off , compute one number, name the family.)

Recall Solution 1.1

What we do: pick the degree-2 coefficients. . Compute the shape number: Negative ⇒ ellipse family. Is it a circle? A circle needs and . Here but , so it is a tilted ellipse, not a circle. Answer: ellipse.

Recall Solution 1.2

. Positive ⇒ hyperbola.

Recall Solution 1.3

. Zero ⇒ parabola. (Sanity check: , a perfect square — the fingerprint of a parabola.)


Level 2 — Application

(One extra factor to handle: unlisted terms, or a missing squared term.)

Recall Solution 2.1

Fit it to the template: (the is a -term, irrelevant to shape). Zero ⇒ parabola — exactly right, is the sideways parabola. Compare with the standard form , where is the focal length (the distance from vertex to focus). Matching gives . See Parabola Standard Equation.

Recall Solution 2.2

Rewrite as . The only degree-2 term is , so . Positive ⇒ hyperbola. (Indeed is the rectangular hyperbola you already know.)

Recall Solution 2.3

. Negative ⇒ ellipse. Because but , it's an axis-aligned (untilted) ellipse, not a circle. See Ellipse Standard Equation.


Level 3 — Analysis

(Now you must also check — the discriminant alone can lie. It can lie in two ways: collapse to lines/points, OR leave no real points at all.)

Recall Solution 3.1

. parabola family. Now the degenerate test. : Expand along the last column: only the entry survives, times the top-left minor: degenerate. Look directly: : two parallel lines, not a parabola. See Degenerate Conics.

Recall Solution 3.2

. hyperbola family. : Expand along the middle row (mostly zeros): the only entry is in position , with sign , giving times the minor . So . degenerate. Directly: : two intersecting lines.

Recall Solution 3.3

. hyperbola family. . The symmetric matrix (halve the off-diagonals): Cofactor expansion along the top row: genuine, non-degenerate hyperbola.

Recall Solution 3.4

. ellipse family. Degenerate test with : So the discriminant says "ellipse" and says "non-degenerate." Yet has no real solutions — a sum of squares can never be negative. The real locus is empty. Applying the real-vs-empty rule (from the intuition box at the top): here and have the same sign, so the ellipse is imaginary (empty) — exactly matching the direct observation . A real ellipse would need and to have opposite signs. Compare with Exercise 2.3, where the ellipse is genuinely real.


Level 4 — Synthesis

(Find the parameter value that switches the shape, or use rotation.)

Recall Solution 4.1

, so This flips sign as crosses :

  • Ellipse when .
  • Parabola when .
  • Hyperbola when or .

The boundary is exactly where the quadratic part becomes a perfect square — the shape "parks" at a parabola while transitioning between loop and split. The figure below plots the three representative curves for so you can see the loop open into a parabola-edge and then split into two branches.

Figure — General second-degree equation Ax²+Bxy+Cy²+Dx+Ey+F=0 — discriminant classification
Recall Solution 4.2

Why ? When we rotate the axes by the new cross-coefficient turns out to be (derived in the parent note). We want so the tilted conic becomes axis-aligned and readable. Setting gives , and dividing by yields . That is the whole origin of the "magic" formula — it is just the condition "kill the cross term." See Rotation of Axes.

Step 1 — angle. Here , so and is infinite ⇒ . (When , the tilt is always .) Step 2 — new quadratic coefficients with : Step 3 — what happens to ? Rotation only mixes the degree-2 and degree-1 terms among themselves; it never touches the constant. Concretely, here (no linear terms), and a rotation of purely degree-2constant terms sends to with and — still no linear part. The constant is a plain number with no or in it, so substituting the rotation formulas cannot change it: . (Rotation moves the axes' directions, not the origin, so the constant term is untouched.) Step 4 — equation: : an ellipse with semi-axes and . Invariance check: original ; rotated . ✓ Identical, as promised.


Level 5 — Mastery

(Build the equation, or combine every tool at once.)

Recall Solution 5.1

Take a positive-definite quadratic part equal to zero: . The only real solution is — a single point. : ellipse family. Degenerate test with : degenerate ellipse = a point. This is the ellipse "shrunk to its centre." See Degenerate Conics and Circle as Special Ellipse.

Recall Solution 5.2

Family: . parabola family. Note . Degenerate test: , matrix uses : Expand along the top row: degenerate parabola. Which kind? Substitute : the equation is . Since , it becomes or : two parallel lines.

Recall Solution 5.3

Both share : parabola family, and . Case (i): . Then : , and : two parallel lines. Case (ii): . Now too: Again , but : a single line, counted twice (a "double line"). The two parallel lines of case (i) have merged into one as the constant shrank to . Lesson: in the parabola family () a degenerate curve is two parallel lines, which coincide into one double line exactly when the perfect square equals (no separating constant).

Recall Solution 5.4

False. is exactly the condition for a genuine, non-degenerate conic. Two parallel lines (and a single double line) are degenerate configurations and always give (as Exercises 3.1, 5.2 and 5.3 all showed). So with forces a genuine parabola — no line-pair possible. Rule summary:

  • → ellipse type (real if opposite signs; imaginary/empty if same signs — Exercise 3.4); → real parabola; → real hyperbola.
  • → degenerate: point, single (double) line, two parallel lines, or two intersecting lines.

Active Recall

Recall Quick self-quiz

Discriminant sign of ? ::: → hyperbola -range making an ellipse type? ::: , : rotation angle to kill ? ::: , , but no real points — what is it? ::: an imaginary (empty) ellipse; here and — what shapes are possible? ::: a single double line or two parallel lines