Fact-checked by Grok 2 weeks ago

Curve of constant width

A curve of constant width is a simple closed in the whose width—the minimum distance between any pair of supporting lines—remains constant regardless of the of the lines. This property distinguishes it from other shapes, ensuring uniform "" in all orientations. The is the archetypal example, possessing the maximum possible area for a given width among all such curves. Non-circular curves of constant width abound, with infinitely many variations constructible; notable examples include the , formed by the intersection of three circular disks centered at the vertices of an , and more general Reuleaux polygons based on odd-sided regular polygons. These curves share several remarkable properties: by Barbier's theorem, proved in 1860, all curves of constant width w have the same perimeter \pi w, equivalent to that of a of w. Additionally, the minimizes the enclosed area for a given width. Curves of constant width have practical applications in and , such as in rotary mechanisms where smooth rotation within a square enclosure is required, exemplified by their use in some drill bits to produce square holes or in manhole covers to prevent them from falling through the hole. Historically, the study traces back to Leonhard Euler's work on involutes in 1778, with Émile Barbier formalizing key isoperimetric results in 1860, and later developments by Wilhelm Blaschke in 1915 exploring their analytic properties. Modern research extends these ideas to higher dimensions and generalized widths.

Fundamentals

Definition

A curve of constant width is a simple closed convex curve in the Euclidean plane \mathbb{R}^2 for which the width—the distance between any pair of parallel supporting lines—is the same constant value, say w > 0, regardless of the direction of the lines. This property requires the curve to bound a convex set, as non-convex sets cannot maintain a uniform minimum distance between parallel supporting lines in all directions; convexity ensures that the set lies entirely on one side of any supporting line, which touches the boundary at least at one point. Supporting lines are defined as lines that intersect the curve's boundary while containing the entire bounded convex region on one half-plane. The width in a specific , given by a \mathbf{u} \in S^1, is precisely the between the two supporting lines with outward \mathbf{u} and -\mathbf{u}; for a of constant width, this equals w for every \mathbf{u}. To illustrate, consider two lines touching the at points P and Q; the perpendicular between these lines measures the width in the normal to them, and this measurement remains under of the . An equivalent characterization is that the orthogonal projection of the bounded onto any line in \mathbb{R}^2 has constant length w. This projection length equivalence follows from the fact that the distance between supporting lines equals the extent of the projection in the perpendicular direction. Unlike general curves, where the width varies by direction and the maximum width equals the (the supremum of distances between any two points on the curve), curves of constant width have width equal to the in every direction—a property shared by the circle, which is the unique smooth curve of constant width that is rotationally symmetric in all directions.

History

The concept of curves of constant width originated in the with the work of Leonhard Euler, who in introduced the notion of "orbiform curves" that maintain a constant distance between parallel supporting lines, particularly in connection with shapes capable of rolling without slipping. Euler constructed such curves using involutes of hypocycloids and explored their properties in relation to polygons with an odd number of sides, laying foundational ideas for non-circular examples. In the , significant advancements occurred, beginning with Joseph-Émile Barbier's 1860 , which established that all plane curves of constant width d have the same perimeter \pi d, providing a key characterization independent of the specific shape. Shortly thereafter, in 1875, Franz Reuleaux developed the —a non-circular curve of constant width formed by the of three circular arcs from an —as part of his kinematic mechanisms for applications, popularizing practical constructions beyond the circle. The 20th century brought deeper theoretical insights and generalizations. In 1915, Wilhelm Blaschke demonstrated that any of constant width can be uniformly approximated by Reuleaux polygons, bridging polygonal and smooth cases. Further progress included explorations of smooth variants, with seminal contributions emphasizing their analytic properties and approximations, such as those by Jaglom and Boltyanskii in 1951 on curvature-consistent polygonal limits. These developments solidified the mathematical framework, influencing subsequent studies in .

Examples

Reuleaux Triangle

The is the archetypal non-circular curve of constant width, formed by the intersection of three circular disks centered at the vertices of an , each with radius equal to the side length of the triangle. Equivalently, it can be constructed by drawing circular arcs of that radius, centered at each vertex and connecting the other two vertices. This shape exhibits of order 3 and possesses a constant width equal to the side length w of the underlying equilateral triangle. Its area is given by A = \frac{1}{2} (\pi - \sqrt{3}) w^2. Visually, the maintains an equilateral triangular outline in all orientations due to its threefold , presenting three distinct vertices regardless of . Kinematically, its constant width enables smooth rolling motion with a fixed height, and it has inspired the design in Wankel rotary engines, where a similar curved triangular shape facilitates within a chamber. Compared to a of the same width, the encloses a smaller area, and in fact achieves the minimal area among all curves of constant width. It serves as the simplest case of a Reuleaux polygon, which generalizes the construction to regular with an odd number of sides.

Smooth and Other Curves

Smooth curves of constant width represent a class of non-circular, infinitely differentiable closed curves that preserve a uniform distance between any pair of parallel supporting lines, distinguishing them from the circle while retaining the core geometric property. These curves avoid the infinite points inherent in polygonal approximations like the Reuleaux triangle, which features sharp vertices that can cause jerky motion in rotational applications; instead, smooth variants ensure finite throughout, facilitating more uniform and efficient in mechanisms such as bearings or cams. The existence of infinitely many such smooth curves follows from the theory of support functions, where a curve has constant width w if its support function p(\theta) satisfies p(\theta) + p(\theta + \pi) = w for all \theta. Any sufficiently smooth periodic function p(\theta) meeting this condition—such as p(\theta) = \frac{w}{2} + \sum_{k=0}^\infty a_k \cos((2k+1)\theta) with coefficients a_k chosen to ensure convexity—yields a distinct smooth curve of constant width w, allowing for arbitrary variation in shape while preserving the width. A seminal construction originates from Leonhard Euler, who demonstrated that the of a with an odd number of cusps produces a smooth curve of constant width; a representative example is the of the deltoid, a three-cusped , resulting in a differentiable boundary with uniform width equal to the deltoid's generating radius. Further examples include algebraic constructions, such as the curve of constant width developed by Rabinowitz, parameterized as \begin{align*} x(t) &= \frac{1}{2}(1 - \cos t) + \frac{1}{4} \sin(2t), \\ y(t) &= \frac{1}{2} \sin t + \frac{1}{4} (1 - \cos(2t)) \sin t \end{align*} for t \in [0, 2\pi], which defines an of degree 4 with width 1, verified through its properties. Another approach generates curves based on a semi-ellipse, where the is the intersection of circular disks of fixed radius centered at points along the semi-ellipse, yielding a constant width equal to a specified ; additional smoothing may be applied for full differentiability. These constructions illustrate the diversity of smooth curves of constant width, which can maintain the defining diameter while exhibiting varied shapes, such as rounded lobes or undulations, to suit specific design needs without compromising the constant width property.

Constructions

Geometric Constructions

Curves of constant width can be constructed geometrically using basic tools such as a compass and straightedge, often by replacing straight sides of polygons with circular arcs or by employing intersection and rotation techniques. The serves as a foundational example, built from an . To construct it, first draw a horizontal line of length 10 cm using a . Set the to a 10 cm and place it at one of the base to draw an above the line; repeat from the other endpoint so the arcs intersect at a point, forming the . Connect this apex to the base endpoints with the ruler to complete the equilateral triangle, labeling the vertices A (left base), B (right base), and C (apex). Keeping the compass at 10 cm, place it at A and draw a semicircular arc from C to B; repeat with the compass at C (arc from B to A) and at B (arc from A to C). The intersections of these arcs form the boundary of the Reuleaux triangle, which has constant width equal to the side length of the original triangle. This method extends to Reuleaux polygons based on regular polygons with an odd number of sides, such as a . Begin by constructing a regular using standard and ruler techniques, ensuring all sides are equal. For each side, identify the opposite (skipping two vertices in a pentagon) and use the set to the side length as to draw a centered at that opposite , connecting the adjacent vertices of the side. Repeating this for all five sides yields the Reuleaux pentagon, where the arcs bulge outward to maintain constant width equivalent to the polygon's side length. This construction works only for odd-sided polygons because even-sided ones would result in overlapping or non-constant width boundaries due to symmetric vertex pairings. Another approach, known as the crossed-lines method, generates curves of constant width through successive intersections and rotations of lines. Start with three lines intersecting pairwise at points, forming a triangular configuration, and select an additional point on one line. Rotate this point around the nearest intersection vertex to trace an arc until it reaches the next line, using the compass for the rotation with a fixed radius. The new intersection becomes the center for the subsequent rotation, continuing this process for six steps until the curve closes. This method allows flexibility in line angles and radii, producing various constant-width shapes, including non-convex ones if lines are arranged accordingly. Involute constructions offer a way to create smooth curves of constant width from base curves with an odd number of cusps, such as . First, geometrically construct the —for instance, a deltoid (three-cusped ) by drawing a fixed and rolling a smaller (one-third ) inside it, marking the path of a point on the rolling 's circumference using string or parametric intersection points. The is then the locus of a point at fixed distance along the tangent lines unwrapping from this , achievable by drawing perpendiculars from the cusp curve at equal arc lengths and connecting the endpoints. Leonhard Euler demonstrated that such of odd-cusped curves yield constant-width boundaries, with the width determined by the unwrapping string length. These methods can produce smoother variants beyond polygonal arcs.

Algebraic Constructions

Algebraic constructions of curves of constant width rely on mathematical formulations that enable precise definition and computational generation, primarily through s and their parametric representations. The h(\theta) of a describes the signed distance from the to the line at \theta, and for a curve of constant width w, it satisfies h(\theta) + h(\theta + \pi) = w for all \theta. This condition ensures the width—the distance between parallel supporting lines—is under . A general parametric representation derives from the support function as x(\theta) = h(\theta) \cos \theta - h'(\theta) \sin \theta, \quad y(\theta) = h(\theta) \sin \theta + h'(\theta) \cos \theta, where h'(\theta) is the derivative with respect to \theta, and \theta ranges from 0 to $2\pi. This form parameterizes the curve as the envelope of its tangent lines. For the simplest non-circular case, an example support function is h(\theta) = \frac{w}{2} + a \cos(3\theta) with a > 0 small enough to ensure convexity, yielding a smooth curve via the parametric equations above. Extensions to higher orders involve adding terms like b \cos(5\theta) or c \sin(7\theta), maintaining only odd harmonics to preserve the constant width condition. Fourier series provide a flexible framework for constructing smooth curves of constant width by representing the support function as h(\theta) = \frac{w}{2} + \sum_{k=1,3,5,\dots}^{\infty} \left( a_k \cos(k\theta) + b_k \sin(k\theta) \right), where even harmonics are absent to satisfy h(\theta + \pi) = w - h(\theta). This series ensures the curve is C^\infty-smooth if sufficiently many terms are included. Polynomial approximations arise by truncating the series and eliminating the parameter \theta, resulting in algebraic equations of even degree. A seminal example is an 8th-degree polynomial curve given by (x^2 + y^2)^4 - 45(x^2 + y^2)^3 + \cdots + (x^2 - 3y^2)x[16(x^2 + y^2)^2 - 5544(x^2 + y^2) + 266382] = 720^3, derived from the support function h(\theta) = 8 + 2 \cos^2\left(\frac{3\theta}{2}\right), with a corresponding parametric form x(\theta) = 9 \cos \theta + 2 \cos 2\theta - \cos 4\theta, \quad y(\theta) = 9 \sin \theta - 2 \sin 2\theta - \sin 4\theta. Higher-degree polynomials follow similarly from support functions with more harmonics. Computational generation often employs algorithms based on rotating an initial curve and intersecting with its offsets to enforce constant width. Starting from a convex set, involves computing the 180-degree rotation around the and taking the with an version scaled to width w, iteratively refining until closure. More directly, numerical evaluation of the form uses the : discretize \theta, compute h(\theta) and h'(\theta) via series summation, then plot the points. Basic for plotting in a language like is: 
import numpy as np
import [matplotlib](/page/Matplotlib).pyplot as plt

w = 1.0  # width
a3 = 0.1  # [coefficient](/page/Coefficient) for [cos](/page/Cos)(3θ)
theta = np.linspace(0, 2*np.pi, 1000)
h = w/2 + a3 * np.[cos](/page/Cos)(3*theta)
dh = -3*a3 * np.sin(3*theta)
x = h * np.[cos](/page/Cos)(theta) - dh * np.sin(theta)
y = h * np.sin(theta) + dh * np.[cos](/page/Cos)(theta)
plt.plot(x, y)
plt.axis('equal')
plt.show()
This generates a approximation, with additional odd terms extensible for complexity. Software like Mathematica or facilitates elimination to algebraic forms via symbolic computation.

Properties

Geometric Properties

Curves of constant width w enclose areas bounded above by that of the circle, \pi (w/2)^2, achieved uniquely by the circle itself via the applied to the fixed perimeter \pi w from Barbier's theorem. The lower bound is attained by the , with area \frac{1}{2} (\pi - \sqrt{3}) w^2, which minimizes the area among all such curves by the Blaschke-Lebesgue theorem. These curves possess distinctive curvature properties, including the relation that the \rho(\theta) at any point satisfies \rho(\theta) + \rho(\theta + \pi) = w. Non-circular examples feature at least six vertices—points where the attains a local extremum—often manifesting as cusps or extrema in representations. Their projections onto any line are line segments of fixed length w, exhibiting rotational invariance in width measurements. The defining geometric trait involves supporting lines: the distance between any pair of parallel supporting lines equals the constant width w, establishing a uniform caliper diameter across all orientations. As convex sets, these curves coincide with the boundaries of their own convex hulls, ensuring no interior points lie outside the spanned region. In kinematic contexts, a curve of constant width rolls without slipping along a straight line while maintaining a constant distance w/2 from the line to its centroid, mimicking the motion of a circle despite deviations in shape.

Analytic Properties

Curves of constant width exhibit several key analytic properties that distinguish them from general curves. A fundamental result is Barbier's theorem, which asserts that the perimeter L of any of constant width w is L = \pi w. This holds regardless of the specific shape, as long as the width—the distance between parallel supporting lines—is constant in every direction. The proof relies on the Cauchy-Crofton formula from integral geometry, which relates the length of a to the measure of lines intersecting it; for constant width sets, this integral simplifies to yield the perimeter as \pi times the width. The h(\theta) provides a powerful analytic for such curves, defined as the from the to the supporting line in direction \theta. For a curve of constant width w, it satisfies h(\theta) + h(\theta + \pi) = w for all \theta. This condition implies that h(\theta) can be expressed as h(\theta) = \frac{w}{2} + f(\theta), where f(\theta + \pi) = -f(\theta), reflecting the odd symmetry relative to the constant term. The implications for follow directly: the is \rho(\theta) = h(\theta) + h''(\theta), and since \rho(\theta) + \rho(\theta + \pi) = w, the total integrates to L = \int_0^{2\pi} \rho(\theta) \, d\theta = \pi w, recovering Barbier's theorem via this parameterization. Regarding curvature, the total curvature of a convex closed plane curve, including those of constant width, is exactly $2\pi, as established by the fact that the tangent turning angle over the full loop sums to $2\pi. In terms of the support function, the curvature \kappa(\theta) = 1/\rho(\theta), so the integral \int \kappa \, ds = \int_0^{2\pi} d\theta = 2\pi. A notable feature is the Blaschke theorem, which guarantees at least three pairs of points on a C^2 closed plane curve with parallel tangents and equal radii of curvature; for constant width curves, these pairs occur at antipodal points where \rho(\theta) = \rho(\theta + \pi) = w/2. Finally, constant width implies constant diameter for convex sets: the diameter d, defined as the supremum of distances between any two points, equals the constant width w. This follows because the distance between any two points is at most the width in the direction perpendicular to the line joining them, and equality is achieved in some directions, making d = w.

Applications

Mechanical Engineering

In mechanical engineering, curves of constant width have been employed since the late 19th century to enable precise, uniform motion in mechanisms where traditional circular components would introduce variability or instability. Franz Reuleaux, a pioneering figure in , explored these curves—particularly the —in his seminal work The Kinematics of Machinery (1875), demonstrating their utility in constrained motion systems such as regulators and rollers that maintain consistent spacing between parallel planes without gaps or wobbling during rotation. This foundational analysis laid the groundwork for practical applications, emphasizing how the constant breadth ensures predictable kinematic behavior under load. Reuleaux's kinematic models, preserved in collections like those at , illustrated these principles through physical demonstrations of linkages and cams, influencing machine design well into the 20th century. One key application lies in cams and linkages, where constant-width profiles provide positive drive without reliance on springs or weights, ensuring reliable intermittent motion. For instance, box beater cams in textile machinery utilize shapes to deliver consistent reciprocating action for beating warp threads, maintaining uniform contact and eliminating backlash during operation. Similarly, early rotary engines employed Reuleaux rotors meshed with internal gears to produce smooth , serving as conceptual precursors to the Wankel engine's eccentric rotary design, which adapts a curved Reuleaux-like triangle for internal combustion. These configurations exploit the curve's invariant width to achieve stable rotation and load distribution in high-speed linkages. In precision structures, constant-width curves enhance tools and components requiring unwavering diameter for effective performance. Drilling bits based on the , developed by engineer Harry James Watts in the early , enable the of near-square holes by rotating against a square socket, with the constant breadth ensuring even cutting edges and minimal deviation—ideal for filleted square holes in mechanical fittings. Reuleaux rollers, cylindrical components with constant-width cross-sections, facilitate stable material transport in conveyor systems or bearings by rolling between parallel surfaces without axial wobble, preserving uniform support under varying loads. This property also benefits seals in rotary applications, where the invariant diameter maintains consistent contact pressure to prevent leakage in dynamic environments like pumps. Advancing into the , computer (CNC) has revitalized these curves for fabricating complex precision parts, allowing lathes and mills to turn solids of constant width with high accuracy for custom mechanisms. Such modern techniques enable the production of Reuleaux-based components for advanced and actuators, building on Reuleaux's to achieve sub-millimeter tolerances in non-circular profiles.

Numismatics and Design

Curves of constant width have found notable applications in numismatics, particularly in coin design, where their unique properties enhance functionality and security. The British 50 pence coin, introduced in 1969, features a seven-sided equilateral curve heptagon shape, a Reuleaux polygon variant that maintains constant width across all orientations. This design allows the coin to roll smoothly like a circle in vending machines while being easily distinguishable by sight and touch from round coins, facilitating quick recognition in transactions. Similarly, the British 20 pence coin, first circulated on June 9, 1982, adopts the same seven-sided constant-width profile to reduce the overall weight of pocket change post-decimalisation, while preserving compatibility with automated coin-handling systems. The Canadian one-dollar "loonie," launched in 1987, employs an 11-sided Reuleaux polygon outline, ensuring constant width for reliable vending machine acceptance and aiding anti-counterfeiting through its distinctive non-circular form that mimics the diameter of existing round coins like the U.S. Susan B. Anthony dollar. These shapes deter counterfeits by complicating replication of the precise curvature, which machines can verify alongside weight and composition for authentication. In and , constant-width curves inspire creations that exploit their paradoxical appearance—resembling circles in motion yet revealing angularity when stationary—for visual illusions of circularity. Sculptors like Dario Santacroce have incorporated Reuleaux triangles into 3D-printed works, such as interlocking spherical forms, to evoke dynamic, rounded motion in static pieces. Mathematical models blending and science, including Reuleaux-based volumes, appear in installations that generalize the triangle's form, often evoking the curved tracery in Gothic cathedral rose windows for an illusion of seamless . Tiles employing constant-width profiles, such as Reuleaux-inspired mosaics, create flooring or wall patterns that simulate rolling circular paths, enhancing perceptual depth without traditional round elements. Industrial design leverages rounded constant-width profiles for ergonomic benefits in everyday products, promoting comfortable handling through uniform grip dimensions. The adoption of constant-width coins has had measurable economic impacts, including cost savings in and machinery compatibility. The British 20 pence's 1982 introduction, following trials in 1981, lightened average pocket change compared to equivalent round denominations, easing daily carry for consumers. Vending machine operators benefited from reduced jamming incidents, as the constant width ensured smooth insertion akin to circular coins. Overall, these designs balanced aesthetic innovation with practical efficiency, influencing global numismatic trends toward non-circular security features.

Computational Uses

Software tools for generating and plotting curves of constant width have been developed in various programming environments, facilitating digital experimentation and visualization. In , interactive programs enable the creation of generalizations by parameterizing curves based on odd-sided regular polygons and circular arcs, allowing users to adjust parameters like the number of sides and arc radii to maintain constant width. Similarly, file exchanges provide scripts for rendering Reuleaux polygons and related shapes, supporting parametric equations for arcs centered at vertices opposite each side. In , libraries such as and are used to construct Reuleaux polygons through vertex calculations and output, with scripts handling odd numbers of sides (e.g., 3 for the triangle, 7 for the ) to ensure constant width via equal arc radii. Simulations of curves of constant width leverage parametric representations to analyze properties like projections and shadows in computer graphics applications. Shadow functions, defined as t_1(\theta) = x(\theta)\cos\theta + y(\theta)\sin\theta for a parametric curve \mathbf{r}(\theta) = (x(\theta), y(\theta)), are computed piecewise for shapes like the , revealing variations in projected width under different light angles while confirming overall constancy. In (CAD) software such as , optimization routines model solids of constant width by revolving 2D curves, targeting minimal area enclosures; for instance, theoretical characterizations identify as optimal for least area given fixed width and inradius constraints. Recent advances in the 2020s include algorithmic tools for constructing constant-width curves suitable for . C++-based generators produce Reuleaux polyhedra by embedding self-dual planar graphs and optimizing arc intersections via , outputting STL files for additive manufacturing while preserving constant width. A 2020 study on the graphs behind Reuleaux polyhedra used computational to construct hundreds of new 3D constant-width bodies, extending beyond traditional tetrahedra for printable prototypes. In 2024, mathematicians developed algorithms for building n-dimensional shapes of constant width with volumes at most 0.9^n times that of the ball, with potential applications in optimization and . In , curves of constant width model uniform-width objects for path planning, particularly in locomotion systems requiring stable trajectories. Bipedal robots employ cam-follower mechanisms to generate foot paths with constant body height, where the triangle's rotation parameter \alpha (from \pi/3 to $2\pi/3) ensures a straight-line support phase and quasi-static stability during stepping. This approach simplifies kinematic planning by leveraging the curve's inherent width constancy for predictable contact and velocity (e.g., 21.9 mm/s step speed).

Generalizations

Three-Dimensional Bodies

A three-dimensional body of constant width is a compact in \mathbb{R}^3 such that the between every pair of parallel supporting planes—known as the width—is the same constant value w regardless of direction. This property requires the body to be , as non-strict convexity would lead to varying widths in certain directions. The width w equals the of the body, the maximum between any two points on its . The sphere of diameter w provides the trivial example of such a body, possessing and uniform properties in all directions. A prominent non-spherical example is the Meissner tetrahedron, introduced by Ernst Meissner in through a modification of the . The , formed by the intersection of four balls of radius w centered at the vertices of a with edge length w, approximates constant width but fails strictly due to greater separation between opposite edges than between a vertex and opposite face; the Meissner construction remedies this by replacing three edges with suitably curved spherical portions centered at the opposite vertices, yielding a surface of exact constant width. The Meissner tetrahedron is significant for minimizing volume among bodies of constant width w, with its volume serving as a in this class. Key properties include bounds on volume and surface area established by Wilhelm Blaschke in : among all such bodies, maximizes both, achieving \frac{\pi}{6} w^3 and surface area \pi w^2, the latter serving as the three-dimensional analog to Barbier's theorem for two-dimensional perimeters. Tetrahedral bodies like the Meissner tetrahedron realize the minimal , approximately $0.1276 w^3, illustrating the variability possible within the constant width constraint while underscoring 's extremal role. All such bodies are also and possess a of w. Constructions typically rely on intersections of balls of radius w; a body of constant width coincides with the intersection of all balls of radius w centered at points on its own boundary, providing a self-consistent geometric . Alternative methods involve offset surfaces parallel to a core at distance w/2 or parametric formulations using the h(\mathbf{u}), where h(\mathbf{u}) + h(-\mathbf{u}) = w for all unit vectors \mathbf{u} on S^2. These approaches extend two-dimensional techniques, such as rotating curves of constant width, to generate three-dimensional examples like solids of revolution.

Non-Convex and Non-Euclidean Cases

While the classical definition of a curve of constant width applies to sets in the , where the distance between parallel supporting lines remains constant regardless of orientation, generalizations to non- curves introduce significant definitional challenges. In particular, for non- shapes such as star-shaped or self-intersecting curves, the notion of "width" must be carefully adapted using supporting lines that may touch the at multiple points, potentially leading to ambiguities in since inward dents could violate the constant separation property. One approach to overcome this is through hedgehogs of constant width, which are non- generalizations constructed as the of the union of line segments connecting points on a closed to a fixed interior point, maintaining a constant width via these supports despite concavities. These hedgehogs allow for self-intersecting or star-like forms while preserving the directional width invariance, though their construction often requires ensuring the remains well-defined across non-smooth regions. Another generalization employs envelopes of families of lines defined by periodic functions, such as a(t) = 16 \sin(3t), which yield non-convex curves of constant width by generating astroid-like shapes with cusps, though these may not be simple closed curves. Such constructions highlight the tension between maintaining constant width and avoiding self-intersections or non-simplicity, often resulting in curves that are only piecewise smooth. In non- geometries, the concept of constant width adapts to the underlying metric, with supporting "lines" replaced by s and widths measured along geodesic parallels. In , curves of constant width are represented using a track function x(\theta) relative to a l, where the curve \gamma satisfies a ensuring constant geodesic distance between parallel horocycles. These curves, such as analogs of the , exhibit constant width B but vary in length depending on the h(\theta), as given by the formula L = \frac{e^{kB} - e^{-kB}}{2k} \int_0^{2\pi} e^{kh(\theta)} \, d\theta for K = -k^2. An adaptation of Barbier's in the Lobachevski plane states that for a horocycle-convex curve of constant width B, the length differs from the Euclidean \pi B, with the geodesic circle achieving L = 2\pi \frac{\sinh(kB/2)}{k} and the Reuleaux analog L = 6 \frac{\sinh(kB)}{k} \arcsin(4 \sech(kB/2)). results further characterize the hyperbolic disk as the only body of constant width with constant projection lengths onto normal geodesics. On the sphere S^d, spherical bodies of constant width w are defined such that the width across any supporting G—measured as the thickness of the narrowest lune containing the body—is constantly w. A key property is that the spherical equals w, and for w < \pi/2, the body is . Moreover, on the 2-sphere, a body has constant \delta it has constant width \delta, providing a direct analog to the case without additional conditions. Approximations of such bodies by spherical polygons or reductions in surface area have been studied, with Blaschke's theorem extended to show that any constant width body with w < \pi/2 can be approximated arbitrarily well by inscribed polytopes. Barbier's theorem adapts here as well, though the perimeter ( length) scales with the spherical excess rather than \pi w. Advanced generalizations extend constant width to higher dimensions via bodies of constant girth, where girth denotes the constant length of projections onto great hyperspheres or the invariant separation by parallel hyperplanes in curved spaces. In discrete settings, analogs appear as polytopes of constant lattice width, defined by the minimal number of parallel lattice lines separating the polytope, with minimal volume examples studied in \mathbb{Z}^n.