A polar coordinate system is a coordinate system for the Euclidean plane in which a point is specified by a distance from a fixed point and an angle measured from a fixed direction. The fixed point is called the pole, the reference ray is the polar axis, and the coordinates are usually written . With the standard convention , is the distance from the pole and gives the direction of the ray containing the point.
Polar coordinates are especially useful when a problem involves rotation, distance from a center, or circular boundaries. A circle centered at the pole has the simple equation , while a ray from the pole has constant . These features make the system fundamental in analytic geometry, calculus, complex analysis, and mathematical physics. The same geometry underlies range-and-bearing measurements, descriptions of orbital motion, and the analysis of fields on disks.
Unlike Cartesian coordinates, polar coordinates do not assign a unique unrestricted numerical pair to every point. Angles differing by a full revolution describe the same direction, and every angle represents the same point when . These properties are not merely notational inconveniences: they determine how polar transformations, differentiation, integration, and numerical calculations must be interpreted.
Coordinate conventions and the geometry of the grid
To construct a polar coordinate system, choose a pole , a ray beginning at , and an orientation for measuring angles. In the usual mathematical convention, the reference ray points horizontally to the right and positive angles are measured counterclockwise. Angles in calculus are normally expressed in radians, so one complete revolution is . Degrees remain common in surveying and navigation. The choice of angular unit matters in derivative formulas: the familiar derivatives of sine and cosine take their simplest form only when the argument is measured in radians.
For a point other than the pole, its radial coordinate is its distance from , and its angular coordinate specifies its direction. Thus denotes a point three units above the pole in the standard orientation. For every integer ,
represent the same point. A unique angle can be selected by restricting it to a half-open interval such as or . This choice introduces a discontinuity along a reference ray: directions on opposite sides of that ray have nearby positions but numerically very different selected angles.
Some treatments permit negative radial coordinates. Under this convention, means travel a signed distance along the direction . Consequently,
as descriptions of a point. Negative radii are particularly convenient when graphing equations such as or . They do not represent negative physical distances. When polar coordinates are used for integration over a region, the convention is usually adopted so that the geometry and area factor remain straightforward.
The coordinate grid consists of concentric circles and radial rays . These families meet at right angles wherever , making polar coordinates an orthogonal curvilinear coordinate system. Equal changes in angle, however, do not correspond to equal distances everywhere. At radius , an angular increment corresponds to tangential displacement . Angular grid lines therefore spread farther apart as their distance from the pole increases.
The pole is exceptional. All pairs represent it, so its angle is undefined rather than intrinsically zero. A program or table may assign zero by convention, but no direction is geometrically distinguished there. This distinction becomes essential when interpreting formulas containing , constructing coordinate grids, or deciding whether an apparent singularity belongs to the underlying object or only to its coordinates.
Transformation to and from Cartesian coordinates
Choose a Cartesian coordinate system with origin at the pole and positive -axis along the polar axis. Elementary trigonometry gives the forward transformation
For , the inverse radial transformation follows from the Pythagorean theorem:
Recovering the angle requires more care. The expression alone cannot distinguish points in opposite quadrants, and it is undefined when . Computational systems therefore commonly provide a two-argument function, usually written , which uses the signs of both coordinates to select the appropriate quadrant.
For example, has radius and angle , whereas has the same radius and angle . Both points have , so a one-argument arctangent cannot identify them correctly without additional information. At , no inverse angular value is geometrically defined. A numerical return value at that input is an implementation convention, not an extension of the mathematical angle function.
The transformation also gives useful formulas for geometric measurements. If two points have polar coordinates and , their Euclidean distance satisfies
This is the law of cosines applied to the triangle formed with the pole. It shows that ordinary distance cannot be computed by treating and as perpendicular Cartesian components. The angular difference enters through a cosine, and its effect depends on both radial distances.
Coordinate changes are particularly simple for rotations about the pole. An active counterclockwise rotation through angle sends
Reflection across the polar axis sends to , while uniform scaling about the pole multiplies by the scale factor. Translation generally has no comparably simple polar formula because it changes the center from which both distance and direction are measured. Rotating the reference axis, rather than the point, instead subtracts the rotation angle from the coordinate.
Numerical conversion requires attention to numerical stability. Direct evaluation of can overflow or underflow even when the final answer is representable; a scaled hypotenuse routine avoids many such failures. Angle differences also require periodic normalization. For instance, directions of and differ by across the branch boundary, not by in the shorter geometric sense.
Equations of curves and the interpretation of polar graphs
A polar curve can be defined explicitly by , parametrically by and , or implicitly by an equation . Its Cartesian parametrization is
This formulation is important because a polar graph is a parametrized set of points, not an ordinary Cartesian graph with angle and radius serving as perpendicular axes. Different parameter values may trace the same point or even the same entire arc.
Basic geometric objects often have compact polar equations. A circle centered at the pole is , with . A ray is , ; a complete line through the pole requires both opposite rays or a signed-radius convention. A line whose unit normal makes angle with the polar axis has equation
where is its signed displacement in that normal direction. The equation , for , describes the circle
The pole lies on this circle rather than at its center.
A conic section with one focus at the pole can be represented, with a suitable orientation, by
where is the semilatus rectum and is the eccentricity. The cases , , and correspond respectively to an ellipse, a parabola, and a hyperbola. This form follows from the focus–directrix definition and is central to orbital geometry. Where is required to be nonnegative, only angular intervals producing a positive denominator belong to that radial representation.
Spirals display another strength of the system. An Archimedean spiral has equation ; successive turns are separated radially by when . A logarithmic spiral has equation , with , and meets each radial ray at a constant angle. Rose curves of the form , with positive integer , illustrate the significance of negative radii: under the standard signed-radius interpretation, there are petals for odd and petals for even .
Plotting and intersection calculations must respect nonuniqueness. Solving finds intersections represented at the same angle, but it can miss intersections represented by opposite angles and opposite signed radii. The pole requires separate attention whenever either radius becomes zero. Likewise, a full interval of length may trace a curve more than once: traces its circle twice over such an interval. Correct tracing intervals are therefore necessary before calculating enclosed area or arc length.
Differential geometry, tangents, and local distance
The local geometry of polar coordinates is encoded by the displacement formula. Differentiating the Cartesian transformation gives
Adding their squares eliminates the cross terms:
This expression is the Euclidean metric written in polar coordinates. It states that a radial increment has length , whereas a small angular increment at fixed radius has length . The two coordinate directions are orthogonal, but their scale factors differ.
For a differentiable curve , the arc-length element becomes
Thus, over a parameter interval ,
For a circle , integration over gives . If a chosen interval traces an arc repeatedly, this formula counts the repeated traversal; it is a length formula for the parametrization, not automatically for the distinct point set.
The tangent slope follows by differentiating the Cartesian parametrization:
provided the denominator is nonzero. A zero numerator with a nonzero denominator indicates a horizontal tangent; a zero denominator with a nonzero numerator indicates a vertical tangent. If both vanish, the parametrization has zero velocity at that parameter value, and higher-order analysis may be required. The derivative is therefore not the Cartesian slope: it describes how radial distance changes with angle.
For a regular polar parametrization, the magnitude of curvature is
where the denominator must be nonzero. A circle gives . At a passage through the pole, the geometry must be distinguished from the coordinate failure. For example, the circle is smooth at the pole even though its angle there is not uniquely defined.
Polar coordinates also provide a useful distinction between coordinate and physical components. The coordinate basis vector associated with changing has length , not one. A vector written as therefore has squared length . In the corresponding orthonormal radial and tangential directions, its components are and . Confusing these conventions produces missing or extra factors of in geometric and physical calculations.
Area, integration, and probability
The area element is determined by the Jacobian matrix of the transformation from to :
Its determinant is . Under the nonnegative-radius convention,
Geometrically, a small polar cell has radial width and tangential width approximately . The factor expresses the increasing area occupied by the same angular interval farther from the pole.
For a suitable integrand , the change-of-variables formula is
The parameter region must cover the physical region without unintended repeated coverage, apart from boundary sets that do not affect the integral. For a disk of radius , the usual bounds are , . For an annular sector, radial bounds and angular bounds give area , assuming the sector is represented once.
If a region consists of all points between the pole and a nonnegative curve , its area is
For a region lying between inner and outer radial boundaries, the integrand becomes . Signed-radius graphs and self-intersecting curves require more careful interval selection. Squaring a radius does not by itself resolve whether an interval traces a region once, several times, or with overlapping parts.
A standard application evaluates the Gaussian integral. If
then, using positivity to justify the product integral,
Hence . Polar coordinates work here because the integrand depends only on distance from the origin and the domain includes every direction.
The same area factor governs probability densities. A point uniformly distributed over a disk of radius does not have a uniformly distributed radius. Its radial density is for , because larger radii contain more area per unit radial width. A uniform disk sample can therefore be constructed from independent uniform variables on by taking and . Taking instead produces excessive concentration near the center.
Vector calculus and equations with circular symmetry
At a point away from the pole, define the radial and tangential unit vectors
They form an orthonormal basis, but unlike Cartesian unit vectors they change with position. Their angular derivatives are
Differentiation of vector quantities must account for these changes as well as changes in their components.
For a scalar field , its gradient is
The factor converts change per unit angle into change per unit tangential distance. If a vector field is written in physical components as , its divergence is
The out-of-plane component of its curl is
Combining gradient and divergence yields the planar Laplace operator:
This formula is especially useful for the heat equation, the wave equation, and Laplace’s equation on disks or annuli. If is rotationally symmetric, its angular derivatives vanish, reducing a two-variable spatial operator to . A radial harmonic function on an annulus therefore has the form . On a disk containing the origin, boundedness excludes the logarithmic term.
For nonradial problems, angular periodicity makes Fourier series natural. Substitution of into Laplace’s equation gives
For positive integer , the solutions are and ; the zero mode gives a constant and a logarithm. In related wave and diffusion eigenvalue problems, separation produces Bessel equations instead. The resulting functions describe radial oscillation, while sine and cosine modes describe angular variation.
The coefficients and appear singular at the origin, but this does not mean every corresponding field is singular there. For example, is smooth and has Laplacian . A boundary-value problem on a full disk must impose regularity consistent with a smooth Cartesian field at the center. On an annulus, where the center is excluded, logarithmic or negative-power terms may instead be admissible.
Complex numbers, periodicity, and the topology of angle
Identify the Cartesian plane with the complex plane by writing . Euler’s formula gives the polar representation
Here is the modulus, and is an argument of . This representation connects plane geometry directly with complex arithmetic. Multiplication multiplies moduli and adds arguments:
Division by a nonzero complex number divides moduli and subtracts arguments. Thus multiplication by a fixed nonzero complex number combines scaling with rotation.
Powers and roots also become geometrically transparent. For an integer ,
If and is a positive integer, its distinct complex th roots are
They lie at equally spaced angles on a circle of radius . The ambiguity of the original argument is essential: adding multiples of before division by generates the different roots.
The angle is intrinsically periodic rather than an unrestricted real-valued coordinate. No globally continuous choice of real argument exists on the entire punctured plane. To see the obstruction, follow a point once around a circle centered at the origin. A continuously tracked angle increases by , even though the point returns to its starting location. A single-valued angle function would have to return to its original real value. Cutting the plane along a ray removes this obstruction on the remaining domain and permits a continuous branch of argument.
The complex logarithm consequently has the multivalued expression
Choosing a branch of argument selects one analytic logarithm on an appropriate domain. The origin must remain excluded, and a domain encircling it cannot support a single-valued analytic logarithm throughout. This relationship explains why angular branch cuts appear in complex analysis rather than being merely artifacts of plotting software.
In topological terms, the punctured plane corresponds to positive radius together with a point on a circle, not positive radius together with one unrestricted real angle. It is therefore naturally represented by . Allowing instead produces a covering description that keeps track of accumulated turns. This distinction is useful in phase unwrapping: a measured principal phase jumps at a branch boundary, while a continuous lifted phase can record many revolutions. Such lifting depends on continuity and sampling assumptions; periodic angle values alone do not reveal how many full turns occurred between observations.
Mechanics, measurement, and computational applications
For a moving particle, write its position as . Differentiating the position and the moving unit vectors gives
and
The radial acceleration contains the centripetal term . The tangential acceleration contains both the effect of angular acceleration and the term associated with simultaneous radial and angular motion. These formulas are kinematic identities for a particle described in a fixed plane; they do not require the observer’s reference frame itself to rotate.
In classical mechanics, a particle of mass has kinetic energy
For a central potential , the angular coordinate does not appear explicitly in the Lagrangian . The Euler–Lagrange equations then imply conservation of
the angular momentum perpendicular to the plane. The swept-area rate is , giving equal areas in equal times. For an attractive inverse-square force, the orbit equation becomes a conic in polar form, with the force center at a focus.
Range-and-bearing measurements are direct practical realizations of polar coordinates. Radar and surveying instruments determine distance and direction relative to an instrument location. Their angular conventions may differ from those in mathematics: a navigation bearing commonly increases clockwise from north. Converting such data requires identifying the reference direction, orientation, and units before applying trigonometric formulas. A bearing is not interchangeable with a standard mathematical angle merely because both are angular measurements.
Measurement uncertainty also transforms geometrically. A small angular error at range produces transverse error approximately , with in radians. An error of milliradian therefore corresponds to approximately meter transversely at a range of kilometer. For sufficiently small errors, a polar covariance matrix can be propagated to Cartesian coordinates by , using the local transformation Jacobian. Large angular uncertainties or distributions crossing a branch boundary require treatment beyond a single linear approximation.
Digital images and numerical models use polar grids when rotational structure is more important than rectangular alignment. In log-polar coordinates, , positive scaling becomes translation in , while rotation becomes translation in angle. These properties support methods for comparing scaled or rotated patterns. Polar meshes nevertheless have uneven physical resolution: cells widen tangentially with increasing radius, and all angular lines converge at the center. Numerical methods must therefore account for cell areas, central regularity, and potentially restrictive time steps near the pole rather than treating the grid as an ordinary rectangular array.
Historical development and higher-dimensional extensions
The geometry underlying polar coordinates predates their modern notation. Ancient astronomy and geography described directions through angles, while work on circles and spirals naturally related angular displacement to distance from a center. Archimedes, in the third century BCE, studied the curve now called the Archimedean spiral. Its modern equation expresses the relation between uniform rotation and uniform outward motion used in its construction. This is an important antecedent, but it should not be equated with the later systematic use of ordered coordinate pairs for arbitrary plane points.
In the seventeenth century, the development of analytic geometry created a setting in which different methods of assigning numbers to points could be treated systematically. Grégoire de Saint-Vincent and Bonaventura Cavalieri used constructions relating radial distance and angle in their investigations of curves. Isaac Newton subsequently treated coordinate arrangements beyond perpendicular Cartesian axes. The historical development was therefore distributed across geometric construction, curve analysis, and the emerging calculus rather than confined to one universally decisive invention.
By the eighteenth century, the work of Leonhard Euler helped establish polar methods within analysis. Their usefulness became especially clear in problems where the geometry itself selected a center: planetary motion, rotationally symmetric fields, and circular boundaries. The notation and conventions used today developed through this broader mathematical practice. Historical discussions consequently distinguish early uses of radius-and-angle descriptions from later explicit treatment of polar coordinates as a general coordinate system.
The most direct three-dimensional extension is the cylindrical coordinate system, written . It retains polar coordinates in each horizontal plane and adds a perpendicular Cartesian height:
Its distance element is
and its volume element is . It is adapted to cylinders, pipes, and fields with axial symmetry. The angular degeneracy occurs along the entire axis , rather than at a single point.
A different extension is the spherical coordinate system. With one common convention, denotes distance from the origin, is the angle from the positive vertical axis, and is the azimuthal angle:
The volume element is . Angle symbols vary between disciplines, so their geometric definitions are more reliable than their names alone. In -dimensional Euclidean space, the analogous radial decomposition is , where is the surface element on the unit -sphere. The planar area factor is the two-dimensional instance of this general dependence of volume on distance from a center.