Equação Dos Pontos Conjugados

Conjugate points in conic geometry: practical guide

equação dos pontos conjugados explained simply

The conjugate points equation comes up whenever you're working with conics and polar relationships. If you have a conic written as Ax² + Bxy + Cy² + Dx + Ey + F = 0 and two points P(x, y) and Q(x, y), those points are conjugate with respect to the conic when A x x + (B/2)(x y + y x) + C y y + (D/2)(x + x) + (E/2)(y + y) + F = 0

This single expression is the core condition. It's symmetric, which matters because conjugacy is a mutual relationship — if P is conjugate to Q, then Q is conjugate to P. The formula is derived from requiring that P lies on the polar line of Q (and vice versa, which follows automatically from symmetry).

How I actually use this in practice

I encounter this most often when solving problems about poles and polars, or when a construction requires finding a point whose polar passes through another specified point. The straightforward approach is to plug the known coordinates into the equation above and solve for whatever unknown coordinate remains. That's it. It reduces to a linear equation in the unknown, which is why people sometimes find it intimidating — there's nothing hard about solving a linear equation, but the setup hides that fact until you write it out. Here's a concrete example. Take the ellipse x²/4 + y²/9 = 1. Rewritten in standard quadratic form: 9x² + 4y² - 36 = 0. So A = 9, C = 4, F = -36, and B = D = E = 0. Let P = (2, 3). I want all points Q = (x, y) conjugate to P. Substituting:

9·2·x + 4·3·y - 36 = 0 18x + 12y = 36 3x + 2y = 6 That's a line. In this case it happens to be the polar of P with respect to the ellipse. Every point on that line is conjugate to P. Check: the point (2, 0) lies on it, and indeed 3(2) + 2(0) = 6. You can verify conjugacy directly: 18·2 + 4·3·0 - 36 = 0. Works.

Edge case that burned me once

I was working on a problem where the conic was degenerate — two intersecting lines, specifically (x - y)(x + y) = 0, which expands to x² - y² = 0. A point on one of the lines has a polar that is the other line. When I applied the conjugate points equation blindly, I got sensible-looking results, but the geometric interpretation was wrong because the "conic" had no interior. The conjugacy relation still holds algebraically, but any problem relying on properties like "the polar of an exterior point intersects the conic at the points of tangency" collapses when the conic degenerates. I caught this when my construction produced a line passing through a point that should have been the vertex of a triangle, but the triangle was flat. The workaround: always check whether the discriminant B² - 4AC tells you the conic is non-degenerate before applying geometric properties that assume one exists. For the degenerate case, stick to the algebraic definition and don't import properties from the non-degenerate theory.

Counter-intuitive thing most people miss

The conjugate points equation does not require the point to be inside or outside the conic. A point on the conic itself is conjugate to every point on its tangent line at that point. This is easy to verify: substitute the conic point into the equation and you get exactly the tangent line equation. Beginners often assume conjugacy is only meaningful for pairs of points both off the curve, but the relationship is universal across all of projective space. Another thing: when the conic is a circle, the conjugate condition doesn't simplify to orthogonality or anything geometrically familiar without projective context. People expect circles to behave nicely and get confused when the formula still involves the full general expression. It doesn't reduce to something shorter just because B = 0 and A = C. It simplifies to A(xx + yy) + D(x+x)/2 + E(y+y)/2 + F = 0, which is about as simple as it gets.

When this method fails outright

If your conic coefficients are numerical approximations rather than exact values, rounding errors can make the conjugacy check unreliable, especially when the resulting linear equation has a near-zero coefficient for the unknown. I've seen this in computer-aided geometry problems where the input comes from measured data. The fix is to work with exact rational arithmetic whenever possible, or to use a tolerance-based check rather than strict equality. Also, if you're dealing with a point at infinity (homogeneous coordinates), the standard Cartesian form of the equation needs to be homogenized first. Skipping that step gives wrong results for problems involving directions rather than finite points.