A parabola is a plane curve consisting of all points whose distance from a fixed point, called the focus, equals their perpendicular distance from a fixed line, called the directrix. The focus does not lie on the directrix. It is one of the nondegenerate conic sections, alongside the ellipse and hyperbola. In suitable rectangular coordinates, its equation is , where , or equivalently, after changing the orientation and scale, , where .
The parabola connects classical geometry with algebra, mechanics, optics, and engineering. Its defining distance property gives rise to an exact reflection law: rays parallel to its axis reflect toward its focus. Quadratic functions have parabolic graphs, and ideal projectile trajectories are parabolas in a uniform gravitational field. Its mathematical study extends from ancient Greek constructions and Archimedean area calculations to coordinate geometry, projective geometry, and modern geometric design.
Focal definition and geometric structure
Let be a fixed point and a fixed line in the Euclidean plane, with . The parabola determined by them is the set of points satisfying
Here denotes ordinary Euclidean distance, and the distance to is measured along a perpendicular. The condition specifies an entire unbounded curve, not merely a single branch selected from a larger curve.
The line through perpendicular to is the axis of the parabola. Reflection in this line preserves both defining distances, so it preserves the curve. The point halfway between the focus and the directrix, measured along the axis, is the vertex. If the vertex-to-focus distance is , then the vertex-to-directrix distance is also , and the focus lies from the directrix. The curve opens along the axis toward the side containing the focus.
A chord passing through the focus is a focal chord. The focal chord perpendicular to the axis is the latus rectum; its length is . This gives a direct geometric measure of the parabola’s width relative to its focal distance. Terminology for the parameter varies: some treatments use “focal parameter” for , others for . An equation or an explicit distance definition is therefore needed when comparing conventions.
The focal definition belongs to a wider classification by eccentricity. For a conic with a focus and directrix, eccentricity is the constant ratio
A parabola has , whereas an ellipse has , and a hyperbola has . Unlike an ellipse or hyperbola, a parabola has no finite center of symmetry. It has one axis of reflection symmetry and no straight-line asymptote.
The same curve can be obtained by intersecting a right circular cone with a plane parallel to a generator, meaning a straight line on the cone through its apex. Provided that the plane does not contain the apex, the section is a nondegenerate parabola. A plane through the apex instead produces a degenerate section. The agreement between the cone-section and focus–directrix definitions can be established using spheres tangent to both the cone and the cutting plane, traditionally called Dandelin spheres. The two descriptions emphasize different aspects: one defines the curve intrinsically within a plane, while the other places it within three-dimensional geometry.
Coordinate equations and quadratic functions
Choose a Cartesian coordinate system with vertex at the origin, axis along the positive -axis, focus , and directrix . For a point , the focal condition becomes
Squaring and simplifying gives
Conversely, every real point satisfying this equation has and satisfies the original distance condition. Thus the algebraic equation introduces no extraneous points.
Exchanging the coordinate axes gives . More generally, a parabola with a vertical axis can be written
Its vertex is , its axis is , its focus is
and its directrix is
The sign of determines whether it opens upward or downward. The focal distance is , so increasing makes the graph narrower when the coordinate scales are fixed.
The graph of any real quadratic polynomial
is a parabola. Completing the square gives
Its vertex therefore has coordinates
The discriminant of the corresponding quadratic equation determines whether this graph intersects the horizontal axis twice, touches it once, or does not meet it. This discriminant concerns intersections with a chosen line; it should not be confused with the discriminant used to classify a general second-degree plane curve.
A convenient parametrization of the sideways parabola is
Every finite point occurs exactly once, since . Polynomial parametrizations make intersections, tangent calculations, and geometric constructions especially simple.
There is also a useful equation in polar coordinates. Place the focus at the pole, take the opening direction as polar angle zero, and put the directrix at . Then , so
The denominator vanishes in the forward axial direction, corresponding to the unbounded ends of the curve. Polar equations are particularly useful when the focus, rather than the vertex, has physical significance, as in orbital mechanics.
General quadratic equations and affine classification
A general real second-degree equation in the plane has the form
For a nondegenerate real conic, the condition
characterizes the parabolic case. The qualification “nondegenerate” is essential: equations such as , , and also satisfy this condition, but represent two parallel real lines, a repeated line, and no real points, respectively.
The quadratic part is represented by the symmetric matrix
For a genuine parabola, has rank one. A rotation of coordinates, obtained through its eigenvectors, removes the mixed term and leaves only one squared coordinate. In suitable orthogonal coordinates , the equation becomes
The curve is a nondegenerate parabola precisely when . Completing the square in and translating both coordinates then produces a standard equation.
This calculation explains why a parabola has no center. A translation by a vector removes all linear terms from a quadratic equation only if
where . For a parabola, the linear component along the null direction of is nonzero, making this system inconsistent. Ellipses and hyperbolas, by contrast, have an invertible quadratic-part matrix and therefore possess a unique center.
All Euclidean parabolas are similar. For example, uniform scaling of both coordinates by transforms into a parabola with focal distance . A rigid motion can then align any chosen vertex and axis with those of another parabola. Consequently, no dimensionless shape parameter distinguishes one parabola from another in the way eccentricity distinguishes different ellipses.
A nonsingular affine map also sends a parabola to a parabola. It preserves lines, parallelism, tangency, and ratios of lengths along the same line, but generally does not preserve angles or distances. Thus the image is again a Euclidean parabola, although its new focus and directrix are not generally the affine images of the original focus and directrix. This distinction separates properties of the curve that depend on Euclidean measurement from those that depend only on affine structure. Parabolas also have an affine characterization through their single direction at infinity, which accounts for their preservation under nonsingular affine transformations.
Tangents, normals, and the reflection property
For the parametrized parabola
the tangent at parameter has equation
At , this reduces to , the tangent at the vertex. More symmetrically, if lies on , its tangent is
A tangent direction at parameter is , so a perpendicular direction is . The normal line is therefore
The tangent has a geometric relationship with the focus and directrix. Let , and let be the perpendicular projection of onto the directrix. Since , the triangle is isosceles. The tangent at is the perpendicular bisector of , and consequently bisects the angle between the segments and . It is also a supporting line: the entire parabola lies on one side of it.
This angle-bisector property yields the parabola’s exact optical behavior. A ray traveling parallel to the axis toward the concave side strikes the curve and, under specular reflection, passes through the focus. Conversely, a ray emitted from the focus is reflected parallel to the axis. The statement uses the ordinary law of reflection, in which incident and reflected rays make equal angles with the normal.
The reflection can also be checked algebraically. For , take the incoming unit direction and a normal vector . The reflected direction is
Meanwhile,
which is a positive multiple of . Thus the reflected ray points exactly toward the focus at every finite point of the curve.
A related property concerns travel distance. Let an incoming plane wavefront be represented in cross-section by , and consider points of the reflector with . The incident distance from this wavefront to is . The focal definition gives , so the total distance is
independent of the reflection point. In wave applications this equal-path property explains why an ideal parabolic reflector combines an on-axis plane wave coherently at its focus.
Differential geometry and local approximation
A parabola is a smooth curve with a nonzero tangent vector everywhere. For
the velocity is , which never vanishes when . The entire curve is therefore a regular parametrized curve without cusps or self-intersections. Its two unbounded ends belong to one connected curve, and neither end approaches a straight-line asymptote.
In differential geometry, the curvature of a plane curve measures its local rate of turning per unit arc length. For this parametrization,
Curvature is greatest at the vertex, where it is , and decreases toward zero along either end. The corresponding radius of curvature at the vertex is . Thus the circle best matching the parabola there has radius twice the focal distance.
For a graph , the same formula becomes
This shows that the constant second derivative does not imply constant geometric curvature. The second derivative measures change of slope with horizontal displacement; curvature measures change of tangent direction with distance traveled along the curve.
The locus of centers of curvature is called the evolute. For , those centers have coordinates
Eliminating gives
This is a semicubical parabola, a different curve with a cusp at . The name does not make it an ordinary nondegenerate conic. The evolute describes how neighboring normals meet and how the best-fitting circle changes along the original parabola.
Parabolas also arise as local approximations to more general smooth graphs. If is twice continuously differentiable near , then
When , the second-order approximation is a parabola. This is a local statement, not a claim that the full graph is parabolic. It underlies quadratic approximation in calculus and optimization, where local slope and curvature are used to model a function near a point. In several variables, the corresponding second-order model involves a quadratic form rather than a single plane parabola.
Areas, arc lengths, and Archimedean quadrature
A parabolic segment is the bounded region between a parabola and one of its chords. Its area has a remarkable affine-invariant description: it equals four-thirds of the area of the triangle whose base is the chord and whose third vertex is the point of the parabola where the tangent is parallel to that chord. Archimedes proved this theorem in Quadrature of the Parabola, long before the development of integral calculus.
For a simple example, consider , with , and the horizontal chord joining the points with and , where . The chord lies on . The segment area is
The triangle formed by the chord’s endpoints and the vertex has area
Hence .
Writing the chord length as and the perpendicular height of the segment as gives
This last formula applies directly when the chord is perpendicular to the axis. For an arbitrarily oriented chord, the general four-thirds theorem remains valid, but the distinguished point of the parabola is determined by parallel tangency rather than necessarily being the curve’s vertex.
Archimedes’ geometric proof repeatedly inserts triangles into the remaining smaller parabolic segments. At each stage, the total area of the newly added triangles is one-quarter of the previous stage’s total. The resulting geometric series is
His argument also used exhaustion reasoning to establish that the residual area can be made arbitrarily small. The result is a major example of exact area determination before the modern integral.
Arc length behaves differently from area. For the parametrization , the length from the vertex to a point with is
Equivalently, the logarithmic term is . An arbitrary finite arc can be measured by subtracting the corresponding values of this increasing arc-length function. Unlike the arc lengths of general ellipses, parabolic arc lengths can therefore be expressed using elementary functions. Nevertheless, equal horizontal or parameter increments do not correspond to equal arc lengths, an important distinction in geometric sampling and curve-based motion.
Projective structure and rational parametrization
In projective geometry, the distinction between ellipse, parabola, and hyperbola depends on which line is chosen as the line at infinity. All nondegenerate real projective conics with real points are projectively equivalent. A parabola is distinguished in an affine chart because its projective completion meets the line at infinity at one real point with intersection multiplicity two.
Using homogeneous coordinates , the affine equation becomes
The line at infinity is . Substitution gives , so the only point there is
The tangent to the conic at this point is precisely . Thus the two ends of the ordinary affine parabola meet at a single point in its projective completion. In comparison, a hyperbola has two distinct real points at infinity, while an ellipse has no real points at infinity.
This provides a unified geometric interpretation of the three real affine conic types. If the chosen line at infinity cuts a projective conic in two real points, the affine part is a hyperbola; if it is tangent, the affine part is a parabola; if there are no real intersections, the affine part is an ellipse. Projective transformations preserve incidence and tangency, but do not preserve ordinary lengths, angles, or the focus–directrix construction.
The parabola is also a basic example of a rational algebraic curve. Its polynomial parametrization extends projectively to
For , setting recovers
The remaining parameter value gives the point at infinity. A single projective parameter line therefore accounts for the complete conic.
A related construction intersects the parabola with lines through a fixed point on it. One intersection is already known; solving for the second gives rational coordinate expressions in the slope. This method generalizes to any nonsingular conic with a suitable known point and is important in algebraic geometry and computational geometry.
Finite parabolic arcs can also be represented exactly as noncollinear quadratic Bézier curves,
Here is the intersection of the endpoint tangents. Such a curve is an affine image of a standard quadratic parametrization. This exact polynomial representation is useful in drawing systems and computer-aided design; circular and elliptical arcs generally require rational, rather than ordinary polynomial, quadratic Bézier representations.
Physical models and engineering applications
In classical mechanics, a projectile moving under constant downward gravitational acceleration , with air resistance neglected, satisfies
If , eliminating time gives
The trajectory is therefore a parabola. If the horizontal velocity vanishes, the motion remains on a vertical line instead. The parabolic model assumes a uniform gravitational field and neglects drag, lift, and rotational effects. For motion over large distances around Earth, an inverse-square central-force model replaces this approximation.
In the ideal Newtonian two-body problem, a parabolic orbit has eccentricity and zero specific orbital energy. It lies between bound elliptical trajectories with negative energy and hyperbolic escape trajectories with positive energy. With gravitational parameter , speed , and distance from the attracting center, this boundary is
A parabolic orbit is therefore a distinct use of the same geometric curve, derived from an inverse-square force rather than a uniform force.
Optical and radio systems use the reflection property. Rotating around its axis produces a paraboloid of revolution. Ideally, rays parallel to its axis are reflected to one focus. Satellite dishes, radio telescopes, reflecting telescope mirrors, and some solar concentrators use this geometry. A parabolic cylinder, produced by translating a parabola perpendicular to its plane, concentrates parallel incident rays onto a line rather than a point. Finite apertures, diffraction, surface errors, and off-axis aberrations limit real systems; the exact focal property concerns ideal on-axis rays.
Parabolas also occur in structural mechanics under specific loading assumptions. A perfectly flexible cable carrying a uniform vertical load per unit horizontal distance assumes a parabolic shape. This approximates the main cable of a suspension bridge when the suspended deck supplies the dominant load. A cable supporting only its own uniformly distributed weight instead forms a catenary. The distinction arises because uniform loading per horizontal distance is not the same as uniform loading per length of cable.
In road and railway alignment, parabolic vertical curves provide a controlled transition between two gradients. Because the derivative of a quadratic function is linear, grade changes uniformly with horizontal distance. Quadratic fitting also appears in least-squares analysis and numerical interpolation. In those applications, however, a fitted parabola is a chosen mathematical model, not evidence that the underlying process necessarily follows a focal or mechanical law.
Historical development
The study of parabolas arose within ancient Greek investigations of conic sections. Menaechmus, active in the fourth century BCE, is traditionally credited with early work on these curves and with using intersections of conics in the problem of doubling a cube. Early approaches associated different conic curves with sections of different types of cone. The later unified theory showed that all three principal types could be obtained from a single suitable cone by varying the cutting plane.
Apollonius of Perga, active around the late third and early second centuries BCE, developed an extensive systematic treatment in his Conics. The names parabola, ellipse, and hyperbola are associated with his terminology and the Greek geometric language of applying areas. In that setting, the terms described an exact application, a deficiency, or an excess in certain area constructions. Their historical meanings concern relationships expressed geometrically, rather than merely the modern visual appearance of the curves.
Archimedes’ Quadrature of the Parabola, from the third century BCE, established the four-thirds area theorem by geometric reasoning. The work is notable both for its exact result and for its method of successively approximating a curved region with rectilinear figures. It belongs to the tradition of the method of exhaustion, in which inequalities and repeated subdivision establish an area without relying on modern infinitesimal notation.
The development of analytic geometry in the seventeenth century changed the treatment of conics. René Descartes and Pierre de Fermat connected geometric loci with algebraic equations, allowing parabolas to be studied through quadratic relations, coordinate transformations, and intersections obtained by equation solving. This did not replace synthetic geometry, but supplied a powerful complementary language in which a whole family of curves could be handled systematically.
Galileo’s Two New Sciences, published in 1638, gave a mathematical account of ideal projectile motion as a combination of uniform horizontal motion and uniformly accelerated vertical motion. The resulting parabolic trajectory became a central example of geometry applied to mechanics. The subsequent development of differential and integral calculus made tangent slopes, curvature, areas, and arc lengths accessible through general methods rather than curve-specific constructions.
Projective geometry later clarified why the three affine conic types belong to one larger family. In the nineteenth century, the use of points at infinity and projective transformations placed the parabola within a theory based on incidence and tangency rather than measurement. Modern treatments retain these several viewpoints side by side: the focus–directrix locus for Euclidean geometry, the quadratic equation for algebra and analysis, the projective conic for transformation theory, and the parametrized curve for computation and design.