Fact-checked by Grok 2 weeks ago

Real projective plane

The real projective plane, denoted \mathbb{RP}^2, is a two-dimensional manifold in projective geometry defined as the set of all straight lines through the origin in three-dimensional Euclidean space \mathbb{R}^3, or equivalently, the quotient space of nonzero vectors in \mathbb{R}^3 by scalar multiplication.^[1] This construction identifies points that represent the same direction, resulting in a space where every pair of distinct lines intersects exactly once, embodying the core axiom of projective planes.^[2] As a topological space, \mathbb{RP}^2 is compact and connected, with the standard topology induced from the quotient of the unit sphere S^2 under antipodal identification, where each point on the sphere is paired with its opposite.^[3] A key characteristic of \mathbb{RP}^2 is its non-orientability, meaning it lacks a consistent choice of "left" and "right" across its surface, distinguishing it from orientable surfaces like the sphere or torus.^[4] Topologically, it can be visualized as a closed disk with antipodal points on the boundary identified, or as a Möbius strip with an additional disk attached along its boundary, yielding an Euler characteristic of 1.^[5] Its fundamental group is \mathbb{Z}/2\mathbb{Z}, reflecting a single non-contractible loop that corresponds to traversing the antipodal identification twice to close.^[3] Unlike the Euclidean plane, \mathbb{RP}^2 includes a "line at infinity," compactifying the space and enabling homogeneous coordinates [x:y:z] where not all coordinates vanish.^[1] The real projective plane plays a central role in algebraic geometry, topology, and computer vision, serving as a model for projective transformations and non-Euclidean geometries.^[6] It cannot be embedded in \mathbb{R}^3 without self-intersections but admits immersions such as the Boy's surface, a famous parametrization that realizes \mathbb{RP}^2 in three dimensions.^[3] Historically, \mathbb{RP}^2 emerged from efforts to unify affine and ideal points in geometry, as developed by mathematicians like Jean-Victor Poncelet and August Ferdinand Möbius in the 19th century, influencing modern applications in robotics and image processing.^[2]

Construction

As a quotient space

The real projective plane, denoted \mathbb{RP}^2, can be constructed topologically as a quotient space. In general, a quotient space is formed by partitioning a topological space into equivalence classes and equipping the set of these classes with the quotient topology, where a set is open if its preimage under the quotient map is open in the original space.^[7] This construction identifies points that are considered equivalent, yielding a new space that captures the desired geometric structure. Specifically, \mathbb{RP}^2 is the quotient space S^2 / \sim, where S^2 is the 2-sphere, defined as the set of points (x, y, z) \in \mathbb{R}^3 with x^2 + y^2 + z^2 = 1, and the equivalence relation \sim identifies each point x \in S^2 with its antipodal point -x.^[7] The projection map \pi: S^2 \to \mathbb{RP}^2 sends each point to its equivalence class = \{x, -x\}, which is continuous and surjective by definition. The quotient topology on \mathbb{RP}^2 is the finest topology making \pi continuous, ensuring that open sets in \mathbb{RP}^2 are precisely those whose preimages under \pi are open in S^2.^[7] This quotient map establishes \mathbb{RP}^2 as a covering space, with S^2 serving as a double cover of \mathbb{RP}^2. The map \pi is a 2-to-1 local homeomorphism, meaning each point in \mathbb{RP}^2 has exactly two preimages in S^2, and it is a principal \mathbb{Z}/2\mathbb{Z}-bundle where the deck transformation is the antipodal map x \mapsto -x.^[7] As a quotient of the compact, connected, and Hausdorff space S^2 under a closed equivalence relation (the antipodal identification), \mathbb{RP}^2 inherits these properties: it is compact, connected, and Hausdorff.^[7] This topological model complements the algebraic construction via homogeneous coordinates in \mathbb{R}^3.^[7]

Via homogeneous coordinates

The real projective plane, denoted \mathbb{RP}^2, can be defined algebraically using homogeneous coordinates as the set of lines through the origin in \mathbb{R}^3. A point in \mathbb{RP}^2 is represented by a triple [x : y : z], where (x, y, z) \in \mathbb{R}^3 \setminus \{(0,0,0)\}, and two triples (x, y, z) and (x', y', z') represent the same point if there exists a scalar \lambda \in \mathbb{R} \setminus \{0\} such that (x', y', z') = \lambda (x, y, z).^[8]^[9] This equivalence relation identifies each point with the one-dimensional subspace (line) it spans in \mathbb{R}^3.^[10] To work with these coordinates computationally, a representative vector for each equivalence class can be chosen by normalization, such as requiring \|(x, y, z)\| = 1, or equivalently, x^2 + y^2 + z^2 = 1. This selects a unique point on the unit sphere S^2 for each line through the origin, though antipodal points on S^2 correspond to the same projective point. Every point in \mathbb{RP}^2 has homogeneous coordinates that are unique up to this scalar multiple, ensuring a well-defined algebraic structure.^[9]^[10] This coordinate system extends the affine plane \mathbb{R}^2 by incorporating points at infinity. Points in the affine plane (u, v) \in \mathbb{R}^2 correspond to [u : v : 1] in \mathbb{RP}^2, forming an affine patch where the third coordinate z \neq 0; dehomogenization yields u = x/z and v = y/z. The remaining points, where z = 0 (i.e., [x : y : 0]), represent directions at infinity, allowing parallel lines in the affine plane to intersect at these ideal points, thus unifying projective geometry.^[8]^[9] Subsets of \mathbb{RP}^2, such as algebraic curves, are defined using homogeneous polynomials. A projective variety is the zero set of a homogeneous polynomial P(x, y, z) = [0](/page/0) in the homogeneous coordinates, where P is invariant under scaling. For example, conics in \mathbb{RP}^2 arise as quadratic forms, satisfying equations like ax^2 + by^2 + cxy + dxz + eyz + fz^2 = [0](/page/0), which include both finite and infinite components of classical conic sections.^[2]^[8]

Geometric Elements

Points and lines

In the real projective plane \mathbb{RP}^2, points are defined as the equivalence classes of directions in \mathbb{R}^3, corresponding to lines through the origin excluding the origin itself. These points are represented using homogeneous coordinates [x : y : z], where (x, y, z) \in \mathbb{R}^3 \setminus \{(0,0,0)\} and two triples (x,y,z) and (x',y',z') represent the same point if there exists a nonzero scalar \lambda \in \mathbb{R} such that (x',y',z') = \lambda (x,y,z).^[11] This construction identifies antipodal points on the unit sphere S^2, yielding \mathbb{RP}^2 as the quotient space S^2 / \sim where p \sim -p.^[10] Lines in \mathbb{RP}^2 are the 1-dimensional projective subspaces, which correspond to planes through the origin in \mathbb{R}^3. Each line is represented by homogeneous coordinates [a : b : c] for the normal vector to the plane, satisfying the equation a x + b y + c z = 0 for points [x : y : z] on the line.^[11] Unlike Euclidean geometry, lines in \mathbb{RP}^2 do not extend infinitely in two directions but form closed loops topologically.^[2] The incidence relation between points and lines is given by the linear equation: a point [x : y : z] lies on a line [a : b : c] if and only if a x + b y + c z = 0.^[10] This relation is bilinear and symmetric in the projective sense, allowing points and lines to be treated dually through their coordinates. A fundamental property of \mathbb{RP}^2 is that any two distinct points determine a unique line, constructed as the projective span of their representing vectors in \mathbb{R}^3.^[11] Conversely, any two distinct lines intersect in a unique point, ensuring no parallel lines exist; their intersection is the projective point corresponding to the 1-dimensional intersection of their planes in \mathbb{R}^3.^[2] As an example, the real projective line \mathbb{RP}^1, which consists of points and lines in one dimension lower, is topologically equivalent to a circle S^1. This arises from identifying antipodal points on the circle itself, forming a closed loop that models the projective structure.^[2]

Duality and incidence

In the real projective plane \mathbb{RP}^2, the duality principle establishes a one-to-one correspondence between points and lines, reflecting the inherent symmetry of the space. A point with homogeneous coordinates [x : y : z] is dual to the line defined by the equation x X + y Y + z Z = 0, where (X, Y, Z) are the variables for points on the line.^[2] This mapping interchanges the roles of points and lines while preserving the geometric structure of \mathbb{RP}^2.^[12] Incidence relations are maintained under this duality: the line joining two points P and Q corresponds to the intersection point of the dual lines P^* and Q^*. Specifically, if two points lie on a common line, their dual lines intersect at the dual point of that line.^[2] This leads to the projective duality theorem, which states that the dual of a line passing through two points is the intersection of the dual lines corresponding to those points.^[12] \mathbb{RP}^2 is self-dual, meaning the duality is an isomorphism that maps the space to itself and preserves all incidence relations, such as collinearity and concurrence.^[2] This self-duality underscores the uniformity between points and lines in projective geometry.^[12] The origins of projective duality trace back to early 19th-century developments in projective geometry by Jean-Victor Poncelet and Joseph Gergonne, who introduced concepts like poles and polars for conics, laying the foundation for the general principle.^[13]

Ideal points

In the real projective plane \mathbb{RP}^2, ideal points, also referred to as points at infinity, are defined using homogeneous coordinates as [x : y : 0], where (x, y) \neq (0, 0). These points represent directions in the underlying affine plane \mathbb{R}^2 and form the projective line at infinity, denoted \mathbb{RP}^1_\infty, which is topologically a circle. The line at infinity itself corresponds to the dual point [0 : 0 : 1] in the projective sense, encapsulating all such infinite directions.^[14]^[15] These ideal points resolve the issue of parallel lines in the affine plane by providing intersection points at infinity. In \mathbb{RP}^2, every pair of distinct lines intersects at exactly one point, so parallel lines in \mathbb{R}^2—which share the same direction—converge at a unique ideal point on \mathbb{RP}^1_\infty. For instance, all vertical lines in \mathbb{R}^2, characterized by the direction (0, 1), intersect at the ideal point [0 : 1 : 0]. This construction ensures that projective geometry treats finite and infinite elements uniformly, without special cases for parallelism.^[2]^[16] The affine plane \mathbb{R}^2 is recovered from \mathbb{RP}^2 by removing the line at infinity, with the identification (x, y) \mapsto [x : y : 1] embedding \mathbb{R}^2 as an open dense subset. Thus, \mathbb{RP}^2 serves as a compactification of \mathbb{R}^2 by adjoining the circle of ideal points, transforming unbounded Euclidean directions into concrete projective points. This transition from Euclidean to projective geometry reifies directions as points, enabling a cohesive framework for incidence and intersection properties.^[14]^[15]

Visualizations

Immersions in three dimensions

An immersion of a manifold into Euclidean space is a smooth map that is locally an embedding at every point, meaning the differential is injective everywhere, but global self-intersections may occur along curves or at points where multiple sheets meet. For the real projective plane \mathbb{RP}^2, such immersions into \mathbb{R}^3 are possible, providing visualizations of this non-orientable surface despite the inherent topological constraints. The real projective plane cannot be embedded in \mathbb{R}^3 without self-intersections, as no closed non-orientable surface admits such an embedding in three-dimensional space; this follows from the orientability of \mathbb{R}^3 and the fact that an embedded closed surface would separate \mathbb{R}^3 into two components with consistent normal orientations, which contradicts the non-orientability of \mathbb{RP}^2. In contrast, all closed orientable surfaces embed in \mathbb{R}^3, highlighting \mathbb{RP}^2 as the simplest non-orientable surface that mandates self-intersections for any realization in three dimensions. A defining characteristic of generic immersions of \mathbb{RP}^2 into \mathbb{R}^3 is the occurrence of triple points, where three surface sheets intersect transversally at a single point. Banchoff's theorem establishes that, for a generic immersion of a closed surface into \mathbb{R}^3, the number of triple points is congruent modulo 2 to the Euler characteristic of the surface. With \chi(\mathbb{RP}^2) = 1 (odd), every such immersion must feature an odd number of triple points, implying at least one. This minimal triple point configuration underscores the topological obstruction, as even the simplest immersions cannot avoid these higher-order intersections.^[17] As the cross-cap surface of minimal genus among non-orientable closed surfaces, \mathbb{RP}^2 exemplifies the necessity of self-intersections in three-dimensional immersions, whereas higher-genus non-orientable surfaces (like the Klein bottle) immerse with more extensive double curves and additional triple points. Boy's surface provides a canonical example of such an immersion with precisely one triple point. Polyhedral immersions of \mathbb{RP}^2 into \mathbb{R}^3 are also feasible, realizing triangulations as piecewise-linear maps with self-intersections; unlike Steinitz's theorem, which characterizes convex polyhedral embeddings of the sphere (Euler characteristic 2), these immersions for \mathbb{RP}^2 require at least nine vertices to accommodate the topology without singularities beyond intersections.^[18]

Embeddings in four dimensions

The real projective plane \mathbb{RP}^2 admits a smooth embedding into \mathbb{R}^4, which is the minimal Euclidean dimension for a self-intersection-free embedding of this compact 2-manifold. This follows from the Whitney embedding theorem, which guarantees that any smooth n-dimensional manifold embeds in \mathbb{R}^{2n}, so \mathbb{RP}^2 embeds in \mathbb{R}^4, while it cannot embed in \mathbb{R}^3 due to its non-orientability and topological obstructions such as the vanishing of the normal Euler class in lower dimensions.^[19] An explicit construction of this embedding uses homogeneous coordinates on \mathbb{RP}^2. Consider the map \Phi: \mathbb{RP}^2 \to \mathbb{R}^4 defined by

\Phi([x : y : z]) = \frac{1}{x^2 + y^2 + z^2} (x^2 - y^2, \, xy, \, xz, \, yz),

where [x : y : z] denotes projective equivalence classes under nonzero scalar multiplication. This map is well-defined because it is homogeneous of degree zero, smooth away from the origin (which is excluded in projective space), and its image avoids self-intersections. To verify it descends from the sphere model, restrict to the unit sphere S^2 \subset \mathbb{R}^3, where the antipodal quotient S^2 / \{\pm 1\} \cong \mathbb{RP}^2; the corresponding map \rho: S^2 \to \mathbb{R}^4, \rho(x,y,z) = (xy, xz, yz, x^2 - y^2), is even under (x,y,z) \mapsto (-x,-y,-z) and induces an injective immersion on the quotient, hence an embedding.^[20]^[19] Algebraically, this embedding relates to quadratic forms, as the components are quadratic monomials in the coordinates, and the image lies on an algebraic variety in \mathbb{R}^4 defined by quadratic relations derived from eliminating the parameters (e.g., relations among the pairwise products and differences). The Veronese map provides another algebraic perspective: the embedding \nu_2: \mathbb{RP}^2 \to \mathbb{RP}^5 given by [x:y:z] \mapsto [x^2 : y^2 : z^2 : \sqrt{2}xy : \sqrt{2}xz : \sqrt{2}yz] (or unnormalized equivalents) embeds \mathbb{RP}^2 as the Veronese surface, a degree-4 hypersurface in \mathbb{RP}^5 realized via rank-1 conditions on the associated symmetric matrix; projecting to a suitable \mathbb{RP}^3 and dehomogenizing yields an embedding in \mathbb{R}^4.^[21] Geometrically, \mathbb{RP}^2 is diffeomorphic to the Grassmannian \mathrm{Gr}(1,3), the manifold of 1-dimensional subspaces (lines through the origin) in \mathbb{R}^3. The above embeddings realize this Grassmannian in \mathbb{R}^4 via coordinates that parametrize line directions using quadratic invariants, ensuring no self-intersections in the higher-dimensional ambient space.^[19]

Planar projections

The real projective plane \mathbb{RP}^2 can be visualized in the Euclidean plane \mathbb{R}^2 through central projection, which maps points in \mathbb{RP}^2 to \mathbb{R}^2 by projecting lines through the origin in \mathbb{R}^3 onto a reference plane, such as z=1. Specifically, for a point represented by homogeneous coordinates [x_1 : x_2 : x_3] with x_3 \neq 0, the projection yields the affine coordinates (x_1/x_3, x_2/x_3) in \mathbb{R}^2, while points with x_3 = 0 correspond to the line at infinity in \mathbb{RP}^2, which maps to the boundary or horizon of the projected plane.^[2] This projection embeds the affine plane \mathbb{R}^2 as an open dense subset of \mathbb{RP}^2, with the line at infinity compactifying it by adding directions of parallel lines.^[2] Another common planar projection is the stereographic projection, obtained by first identifying \mathbb{RP}^2 with the quotient of the unit sphere S^2 by antipodal points and then projecting S^2 from the north pole onto the equatorial plane. For a point P = (a, b, c) \in S^2 excluding the north pole (0,0,1), the projection intersects the line from the north pole through P with the plane z=0, yielding coordinates (x,y) = \left( \frac{a}{1-c}, \frac{b}{1-c} \right) in \mathbb{R}^2.^[22] The north pole itself maps to the point at infinity, compactifying the plane to the Riemann sphere, but under the antipodal identification, this extends to model \mathbb{RP}^2.^[22] The antipodal identification on S^2 causes antipodal points to map to the same location in the plane or to reciprocally related points, introducing a branching or ramification in the projection to represent the non-orientable structure of \mathbb{RP}^2. This is evident when considering the southern hemisphere, whose projection overlaps the northern one with a twist, requiring a branch cut—such as along a chosen line in the plane—to resolve the double covering and visualize the topology.^[22] Visually, these projections depict \mathbb{RP}^2 as a disk in \mathbb{R}^2 where opposite points on the boundary are identified with a Möbius-like twist, illustrating the single non-orientable "handle" without self-intersections in the plane but capturing the projective duality.^[22]

Specific Models

Boy's surface

Boy's surface is an immersion of the real projective plane \mathbb{RP}^2 into three-dimensional Euclidean space \mathbb{R}^3, notable for realizing this non-orientable surface without cuspidal edges or other singular points beyond self-intersections. Discovered by the German mathematician Werner Boy in his 1901 doctoral thesis under David Hilbert at the University of Göttingen, it was the first such immersion found, countering Hilbert's initial conjecture that no smooth immersion of \mathbb{RP}^2 into \mathbb{R}^3 existed. Boy constructed the surface using hand-drawn level sets, demonstrating its topological equivalence to \mathbb{RP}^2 through a series of deformations from simpler models.^[23]^[24] The surface can be described parametrically using coordinates derived from stereographic projection and angular parameters. One common form employs parameters a and b, where a = \eta \cos \phi, b = \eta \sin \phi, and \eta = r / \sqrt{1 + r^2} with r as a radial parameter and \phi \in [0, 2\pi). The coordinates are then given by:

\begin{align*} x &= \frac{(a^2 - 9 b^2)(1 + 5 b^2)}{(1 + b^2)^3}, \\ y &= \frac{2 a b (1 - b^2)(3 - 5 b^2)}{(1 + b^2)^3}, \\ z &= \frac{a (1 - b^2)(1 + b^2)(9 b^2 - 1)}{(1 + b^2)^3}. \end{align*}

This parametrization maps the disk with identified antipodal boundary points to the immersed surface. Alternatively, Boy's surface admits an algebraic description as the real zero set of a sextic polynomial, such as the one provided by François Apéry in 1986:

(x^2 + y^2 + z^2 + 1)^3 - 27(x^2 + y^2) z^2 (x^2 + y^2 + z^2 - 1) = 0,

which defines a smooth immersion except at self-intersection loci.^[25] Boy's surface features exactly one triple point, where three sheets intersect transversely with pairwise orthogonal tangent planes, and three double curves emanating from this point, each forming a figure-eight shape in projection. These double lines represent the self-intersection locus, a twisted trifolium curve of genus 3, with no additional singularities like cusps, distinguishing it from earlier models such as the Roman surface. This configuration achieves the minimal number of self-intersections for immersing \mathbb{RP}^2 in \mathbb{R}^3, as proven by subsequent classification results. Topologically, the surface is non-orientable with genus 1 (equivalent to one cross-cap), confirming its identification with \mathbb{RP}^2.^[26]^[25] Visually, Boy's surface resembles a deformed or pinched torus with threefold rotational symmetry around a vertical axis, featuring three symmetric "tunnels" or orifices that converge toward a central triple point, creating a compact, self-intersecting form that evokes a twisted disk. This structure highlights the non-orientability, as traversing certain paths reverses orientation, and models often emphasize the smooth, flowing contours away from the intersections.^[27]

Roman surface

The Roman surface, also known as the Steiner surface, was discovered by the Swiss mathematician Jakob Steiner in 1844 while he was visiting Rome, from which it derives its name.^[28] This quartic algebraic surface provides a self-intersecting singular mapping of the real projective plane \mathbb{RP}^2 into three-dimensional Euclidean space \mathbb{R}^3, capturing the non-orientable topology of \mathbb{RP}^2 through singularities and self-intersections.^[29] Steiner's construction highlights the surface's role in projective geometry, where it serves as a concrete realization of abstract projective properties in a familiar spatial setting.^[30] In homogeneous coordinates [x : y : z], the Roman surface is defined by the algebraic equation

x^2 y^2 + y^2 z^2 + z^2 x^2 - x y z (x + y + z) = 0.

This degree-4 equation describes a bounded surface symmetric under permutations of the coordinates and contained within the unit sphere.^[29] Key features include three double lines along which the surface self-intersects, creating six pinch points at their endpoints, and a single triple point at the origin where three sheets meet.^[28] These singularities reflect the mapping's inability to smoothly realize \mathbb{RP}^2 without intersections in \mathbb{R}^3, with the double lines forming the skeleton of a tetrahedron that outlines the surface's global structure.^[31] Geometrically, the Roman surface arises as the locus of intersection points of corresponding planes associated with a complete quadrangle in projective space—a configuration of four points, no three collinear, and their six connecting lines.^[28] Steiner's method pairs opposite sides of the quadrangle to define planes whose intersections trace the surface, illustrating how projective duality generates the mapping. This construction underscores the surface's degree-4 nature and its faithful representation of \mathbb{RP}^2, where points correspond to lines through the origin in \mathbb{R}^3 modulo scaling.^[30]

Cross-capped disk

The cross-capped disk provides a simple model of the real projective plane \mathbb{RP}^2 by starting with a closed disk and performing boundary identification where antipodal points on the circumference are glued together in an orientation-reversing manner, equivalent to a Möbius identification or half-twist.^[4] This construction yields a quotient space that captures the topology of \mathbb{RP}^2, where the interior of the disk corresponds to an affine plane and the boundary represents lines at infinity.^[5] An alternative perspective views the cross-capped disk as a Möbius strip sewn to the edge of another disk along their boundaries, effectively closing the Möbius strip to form a compact surface.^[32] This attachment highlights the non-orientable nature of the model, as the [Möbius strip](/page/Möbius strip) introduces the twist that prevents consistent orientation.^[33] In visualization, the cross-capped disk features a central disk region that remains flat, with the boundary twisted inward to simulate the projective identification, often resulting in a self-intersecting immersion when realized in three dimensions.^[32] This model emphasizes the projective plane as arising from a single cross-cap attached to a sphere (or equivalently, the disk construction), serving as the fundamental generator of non-orientable closed surfaces with Euler characteristic \chi = 1.^[4]^[5] Topologically, the cross-capped disk is equivalent to the quotient of the 2-sphere S^2 by the antipodal map, where each pair of opposite points is identified.^[4]

Hemi-polyhedra

Hemi-polyhedra are discrete polyhedral models of the real projective plane \mathbb{RP}^2, constructed by applying the antipodal identification to the vertices, edges, and faces of Platonic solids, effectively quotienting the sphere by the antipodal map to yield a tessellation of \mathbb{RP}^2.^[34] This process identifies opposite points on the bounding sphere, resulting in structures where faces pass through the center and are paired via the identification.^[35] These models provide combinatorial realizations of \mathbb{RP}^2 as cell complexes, emphasizing its non-orientable topology through faceted approximations rather than smooth embeddings. Prominent examples include the hemi-cube, derived from the cube by antipodal quotienting, which features 4 vertices, 6 edges, and 3 square faces, with each face corresponding to a pair of opposite cube faces identified.^[34] The hemi-dodecahedron, obtained similarly from the dodecahedron, has 10 vertices, 15 edges, and 6 pentagonal faces, where the edge graph forms the Petersen graph and the faces include both planar and non-planar pentagons under tetrahedral symmetry.^[35] The hemi-icosahedron arises from the icosahedron via the same quotient, yielding 6 vertices, 15 edges, and 10 triangular faces, with its skeleton being the complete graph K_6 and existing in two enantiomorphic forms.^[34] A defining property of hemi-polyhedra is their Euler characteristic V - E + F = 1, consistent with the topology of \mathbb{RP}^2, as computed for the hemi-cube ($4 - 6 + 3 = 1), hemi-dodecahedron ($10 - 15 + 6 = 1), and hemi-icosahedron ($6 - 15 + 10 = 1).^[35] In these constructions, faces are generally identified in antipodal pairs, leading to half the original number of faces from the Platonic solid, though the resulting polyhedron may require higher-dimensional space for symmetric realization without self-intersection.^[34] Combinatorially, hemi-polyhedra realize \mathbb{RP}^2 as a CW-complex, with the minimal such structure consisting of one 0-cell, one 1-cell, and one 2-cell, where the 2-cell attaches via a degree-2 map to the 1-cell to reflect the non-trivial fundamental group \mathbb{Z}/2\mathbb{Z}.^[36] This quotient construction from Platonic solids under the antipodal map preserves the regular face types while embedding the projective geometry in a polyhedral framework.^[34]

Topological Properties

Non-orientability

A surface is non-orientable if it contains an embedded Möbius strip as a subspace.^[37] Equivalently, for a smooth manifold, orientability holds if and only if the first Stiefel-Whitney class w_1(TM) vanishes in cohomology with \mathbb{Z}/2\mathbb{Z} coefficients. For the real projective plane \mathbb{RP}^2, w_1(T\mathbb{RP}^2) is the nonzero generator of H^1(\mathbb{RP}^2; \mathbb{Z}/2\mathbb{Z}), confirming its non-orientability.^[38] The real projective plane arises as the quotient space \mathbb{RP}^2 = S^2 / \sim, where \sim identifies antipodal points via the map a: S^2 \to S^2 given by a(x) = -x. This antipodal map is orientation-reversing, as it corresponds to multiplication by -1 on \mathbb{R}^3, which has determinant -1. In the quotient, a loop in \mathbb{RP}^2 lifting to a path on S^2 connecting a point to its antipode will reverse orientation upon closing, yielding an orientation-reversing loop in \mathbb{RP}^2. In the cross-capped disk model of \mathbb{RP}^2, a closed curve traversing the self-intersection line of the cross-cap reverses the handedness of a local frame, providing a concrete illustration of this orientation reversal.^[4] The real projective plane is the simplest closed non-orientable surface, being compact, connected, without boundary, and having Euler characteristic \chi(\mathbb{RP}^2) = 1.^[37] While \mathbb{RP}^2 is locally orientable—admitting consistent orientations in neighborhoods of points—global orientability fails due to the existence of these orientation-reversing loops. In the classification of compact surfaces, closed non-orientable surfaces are homeomorphic to the connected sum of the sphere S^2 with k cross-caps for k \geq 1, and \mathbb{RP}^2 corresponds to the case k=1.^[37]

Fundamental group and homology

The fundamental group of the real projective plane, denoted \pi_1(\mathbb{RP}^2), is isomorphic to \mathbb{Z}/2\mathbb{Z}.^[7] This group is generated by a loop that traverses the cross-cap once, and every non-trivial loop in \mathbb{RP}^2 has order two, reflecting the space's non-trivial topology.^[7] To compute \pi_1(\mathbb{RP}^2), consider the universal covering space S^2 \to \mathbb{RP}^2, which is a two-sheeted cover induced by the antipodal map on the sphere. Since S^2 is simply connected, the fundamental group of \mathbb{RP}^2 is the deck transformation group of this cover, which is \mathbb{Z}/2\mathbb{Z}.^[7] Alternatively, using van Kampen's theorem on a cell decomposition of \mathbb{RP}^2 as a disk with antipodal boundary identification yields the same result, with the generator satisfying a relation of order two.^[7] The integer homology groups of \mathbb{RP}^2 are H_0(\mathbb{RP}^2; \mathbb{Z}) \cong \mathbb{Z}, H_1(\mathbb{RP}^2; \mathbb{Z}) \cong \mathbb{Z}/2\mathbb{Z}, and H_2(\mathbb{RP}^2; \mathbb{Z}) \cong 0, with all higher groups vanishing.^[39] These can be computed via cellular homology on the CW-complex structure of \mathbb{RP}^2, which has one cell in each dimension 0, 1, and 2. The chain complex is $0 \to \mathbb{Z} \xrightarrow{\times 2} \mathbb{Z} \to \mathbb{Z} \to 0, where the boundary map from the 2-cell to the 1-cell has degree 2, producing the torsion in H_1 and vanishing top homology.^[7] With \mathbb{Z}/2\mathbb{Z} coefficients, the homology groups are H_n(\mathbb{RP}^2; \mathbb{Z}/2\mathbb{Z}) \cong \mathbb{Z}/2\mathbb{Z} for n = 0, 1, 2, as all boundary maps become zero modulo 2.^[7] The Euler characteristic of \mathbb{RP}^2 is \chi(\mathbb{RP}^2) = 1, obtained as the alternating sum of the ranks of the homology groups: $1 - 0 + 0 = 1.^[7] This matches the direct computation from the CW structure: one 0-cell, one 1-cell, and one 2-cell, yielding $1 - 1 + 1 = 1.^[7] With \mathbb{Z}/2\mathbb{Z} coefficients, the Betti numbers are 1 in each dimension, yielding Euler characteristic $1 - 1 + 1 = 1.^[7] The \mathbb{Z}/2\mathbb{Z} torsion in H_1(\mathbb{RP}^2; \mathbb{Z}) arises from the non-orientability of the space, distinguishing it from orientable surfaces.^[7] Furthermore, \mathbb{RP}^2 is not aspherical, as its universal cover S^2 has non-trivial second homotopy group \pi_2(S^2) \cong \mathbb{Z}, which projects non-trivially.^[7]

Relations to Other Structures

Comparison to other non-orientable surfaces

The real projective plane, denoted RP², serves as the fundamental building block for all closed non-orientable surfaces in topology, as every such surface is homeomorphic to a connected sum of g copies of RP² for some positive integer g, where g is the non-orientable genus.^[40] This classification theorem underscores RP²'s role, distinguishing it from orientable surfaces, which are connected sums of tori.^[41] A key example is the Klein bottle, which is homeomorphic to the connected sum RP² # RP², corresponding to g=2.^[41] Unlike RP², which has Euler characteristic χ=1, the Klein bottle has χ=0, computed via the connected sum formula χ(M # N) = χ(M) + χ(N) - 2.^[42] The fundamental group of RP² is ℤ/2ℤ, abelian and of order 2, reflecting its simple loop structure, whereas the Klein bottle's fundamental group is non-abelian, presented as ⟨a, b | aba = b⁻¹⟩, indicating more complex homotopy.^[43]^[44] In terms of cross-cap decompositions, RP² is equivalent to a single cross-cap, while the Klein bottle corresponds to two cross-caps, and the connected sum of three RP² yields Dyck's surface (g=3) with χ=-1.^[45] These higher-genus surfaces, like the Klein bottle, can be embedded in ℝ⁴ without self-intersection, matching RP²'s embedding dimension, but none embed in ℝ³; instead, they admit immersions in ℝ³ with self-intersections, such as the standard bottle model for the Klein bottle or Boy's surface for RP².^[46] RP²'s models often exhibit self-intersections due to its projective nature, differing from immersions of higher cross-cap sums that may avoid certain singularities.

Higher-dimensional projective planes

The real projective space \mathbb{RP}^n is defined as the space of all one-dimensional subspaces (lines through the origin) of the vector space \mathbb{R}^{n+1}. Equivalently, it is the quotient space S^n / \sim, where S^n is the n-sphere and \sim identifies each point with its antipode -x.^[47]^[48] Unlike \mathbb{RP}^2, which is non-orientable, the orientability of \mathbb{RP}^n depends on the parity of n: it is orientable if and only if n is odd, and non-orientable if n is even.^[49]^[50] For example, \mathbb{RP}^3 is diffeomorphic to the special orthogonal group SO(3), which is orientable.^[36] The fundamental group of \mathbb{RP}^n is \mathbb{Z}/2\mathbb{Z} for all n \geq 2, reflecting the double cover by the sphere S^n.^[51]^[47] The integer homology groups of \mathbb{RP}^n exhibit torsion in intermediate dimensions: H_k(\mathbb{RP}^n; \mathbb{Z}) = \mathbb{Z} for k=0 and for k=n when n is odd; H_k(\mathbb{RP}^n; \mathbb{Z}) = \mathbb{Z}/2\mathbb{Z} for odd k with $1 \leq k < n; and H_k(\mathbb{RP}^n; \mathbb{Z}) = 0 otherwise.^[52]^[36] This structure arises from the cellular chain complex of \mathbb{RP}^n, which has one cell per dimension up to n and differentials of degree 2 modulo 2.^[36] While \mathbb{RP}^2 does not embed smoothly in \mathbb{R}^3, higher-dimensional \mathbb{RP}^n for n > 2 embed in \mathbb{R}^{2n} by the Whitney embedding theorem, though the minimal embedding dimension can be lower in some cases (e.g., \mathbb{RP}^3 embeds in \mathbb{R}^4) but is constrained by orientability for even n.^[53]^[54] In algebraic geometry, \mathbb{RP}^n serves as the parameter space for lines in \mathbb{R}^{n+1}, facilitating the study of projective varieties and their intersections.^[2] It admits a standard Riemannian metric induced as a quotient of the round metric on S^n, analogous to the Fubini-Study metric on complex projective space but adapted to the real setting.^[48]^[55]

References

[1]
[PDF] MATH431: Real Projective 2-Space - UMD MATH
Oct 6, 2021 · Definition 2.1.1. Define real projective 2-space denoted RP2 is the set (re- ally a topological space) of straight lines through the origin in ...
[2]
[PDF] Basics of Projective Geometry - UPenn CIS
Three projectively independent points define a (unique) projective plane. A closer look at projective subspaces will show some of the advantages of pro- jective ...
[3]
[PDF] Real Projective Space: An Abstract Manifold
Mar 10, 2017 · This talk will focus on one kind of abstract manifold, namely real projective space RPn. To see why this space has some interesting properties ...
[4]
Surfaces: 3.3 The projective plane | OpenLearn - The Open University
The geometric approach is to define the projective plane as the set of all infinite lines through the origin in Euclidean three-dimensional space.
[5]
Glossary: Real Projective Plane - The Geometry Center
The real projective plane is a non-orientable surface with Euler characteristic 1, and is a Möbius band with a disk attached along its boundary curve.
[6]
[PDF] Projective Geometry - Purdue Computer Science
▷ The projective plane is not orientable. ▷ Lines have one side: removing a line leaves a connected set. ▷ Segments are ambiguous: two points split their line ...
[7]
[PDF] Algebraic Topology - Cornell Mathematics
This book was written to be a readable introduction to algebraic topology with rather broad coverage of the subject. The viewpoint is quite classical in ...
[8]
The real projective plane in homogeneous coordinates - Plus Maths
The POINTS and LINES of the real projective plane are just the lines and planes of Euclidean xyz-space that pass through (0, 0, 0).Missing: definition | Show results with:definition
[9]
[PDF] Homogeneous Coordinates∗
The projective plane P2 is defined to be the set of all equivalence classes, that is, {[(x, y, z)] | x 6= 0,y 6= 0, orz 6= 0}. An equivalence class is ...
[10]
[PDF] Math 149 W02 M. Homogeneous coordinates and the real projective ...
W ith this definition, geometry in the real pro j ective plane obeys very simple rules: ul e . very two lines meet in exactly one point. ul e . very two ...
[11]
[PDF] The Real Projective Plane RP - maths.nuigalway.ie
Projective lines are planes through the origin in R3. Projective points are represented by homogeneous coordinates. [x, y, z]. For example (1, 2,−3) and ...<|control11|><|separator|>
[12]
https://math.libretexts.org/Bookshelves/Geometry/Euclidean_Plane_and_its_Relatives_(Petrunin)/15:_Projective_Geometry/15.07:_Duality
[13]
Duality Principle -- from Wolfram MathWorld
A similar duality exists for reciprocation as first enunciated by Poncelet (1817-1818; Casey 1893; Lachlan 1893; Cremona 1960, p. x). Examples of dual geometric ...
[14]
[PDF] 18.900 Spring 2023 Lecture 26: Projective Geometry
(26a) The projective plane. The projective plane ... if (a, b) = (0, 0), we have the line at infinity z = 0, which consists of all points at infinity.
[15]
[PDF] 1 The Projective Plane - Brown Math
A line in the projective plane is the set of equivalence classes of ... line at infinity. Exercise 8: Prove that VP ∩ L∞ consists of the points pk ...
[16]
[PDF] Mathematical Principles in Vision and Graphics: Projective Geometry
Ideal points in the projective 3-space are located at infinity, and have homogeneous coordinates of the form (x, y, z, 0). ▫ These points are also called ...
[17]
Immersions of the projective plane with one triple point - ScienceDirect
We further show that any generic immersion of the projective plane with one triple point ... Models of the Real Projective Plane: Computer Graphics of Steiner and ...
[18]
[PDF] 20 POLYHEDRAL MAPS - CSUN
The minimum number of vertices for polyhedral maps that admit polyhedral immersions in R3 is 9 for the real projective plane RP2 [Bre90b], the Klein bottle, and ...
[19]
[PDF] Embedding the Real Projective Plane into R4 Define S2 to be {(x, y ...
We define a quotient space by identifying antidopal points on S2. That is two points (x1,y1,z1) and (x2,y2,z2) on S2 are identified if a(x1,y1,z1)=(x2,y2,z2). ...
[20]
[PDF] SOLUTIONS
Jun 9, 2016 · In this exercise we shall construct a smooth embedding of the real projective plane RP2 into R4 by means of the function F : R3 ! R4, where.
[21]
[PDF] What is the Veronese map? - OSU Math
Jul 26, 2016 · Abstract. We introduce the Veronese map νd : Pn → PN(n,d), a popular example in algebraic geometry. We consider the d=2 and n=2 case as the ...
[22]
[PDF] A Perspective on Conics in the Real Projective Plane
Before we depart, let us make the following definition. Definition 3.0.3. A conic is the set of points in P2 satisfying a homogeneous quadratic equation in x = ...
[23]
The Boy Surface at Oberwolfach — MFO
The Boy surface is named after Werner Boy, who constructed this surface, which is an immersion of the real projective plane in Euclidean 3-space, in 1901 in ...
[24]
[PDF] Coloring 3D Line Fields Using Boy's Real Projective Plane Immersion
Coloring 3D Line Fields Using Boy's Real Projective Plane Immersion. C¸ a ... the self intersection curve of Boy's surface. While we have applied our.
[25]
Boy Surface -- from Wolfram MathWorld
The Boy surface is a nonorientable surface that is one possible parametrization of the surface obtained by sewing a Möbius strip to the edge of a disk.
[26]
Boy's surface - MATHCURVE.COM
Boy's surface was discovered after the search of a model in of the projective plane that would not have other singularities than self-intersections along which ...
[27]
Steiner surface - Paul Bourke
The Steiner surface is attributed to Jacob Steiner, created while he was visiting Rome in 1844. Points on the so called Steiner surface satisfy the following ...
[28]
Roman Surface -- from Wolfram MathWorld
The Roman surface is essentially six cross-caps stuck together and contains a double infinity of conics. The Roman surface can given by the equation. (x^2+y^2+z ...
[29]
Steiner Surfaces
It has two real pinch points and three double lines meeting at a triple point, and, unlike the Roman or Cross Cap, is not compact in any affine neighborhood.Missing: figure- eight
[30]
The Roman Surface - The Geometry Center
Oct 10, 1995 · The Roman surface is an image of the real projective plane. It contains three segments of double points each of which terminates in two pinch points.
[31]
Cross-Cap -- from Wolfram MathWorld
" A cross-handle is homeomorphic to two cross-caps (Francis and Weeks 1999). A sphere with one cross-cap has traditionally been called a real projective plane.
[32]
[PDF] rubber sheets and crazy bottles - City Tech OpenLab
We create a projective plane by adding one cross cap to a sphere. Stated another way, a projective plane is a Möbius band with a disk glued to its boundary.
[33]
[PDF] Polyhedral Models of the Projective Plane - The Bridges Archive
The tetrahemihexahedron is the only uniform polyhedron that is topologically equivalent to the projective plane. Many other polyhedra having the same topology ...
[34]
[PDF] A 10-Dimensional Jewel - People @EECS
Both the hemi-icosahedron and the hemi-dodecahedron can be used as building blocks to construct two additional intriguing and difficult-to-visualize self- ...
[35]
[PDF] Algebraic Topology I: Lecture 17 Real Projective Space
Cn(X). We have a CW structure on RPn with Skk(RPn) = RPk; there is one k-cell – which we'll denote by ek – for each k between 0 and n. So the cellular chain ...
[36]
[PDF] A Guide to the Classification Theorem for Compact Surfaces
Sep 19, 2012 · Indeed, a rigorous proof requires, among other things, a precise definition of a surface and of orientability, a precise notion of triangulation ...
[37]
The Stiefel-Whitney classes of projective space
Dec 17, 2010 · We want to compute the Stiefel-Whitney classes of the tangent bundle {T(\mathbb{RP}^n)}. The cohomology ring of {\mathbb{RP}^n} with {\mathbb{Z}/2} -
[38]
[PDF] elementary homology theory with computations
Sep 18, 2018 · This paper aims to expose the reader unfamiliar with algebraic topology to elementary topics in homology theory. In particular, we intend to.
[39]
Classifying Surfaces (CliffsNotes Version) - Math3ma
Mar 16, 2016 · A non-orientable surface of genus g is a connected sum of g g ... projective planes can be written as a connected sum of projective planes only.
[40]
Connected sum of two surfaces - MATHCURVE.COM
the connected sum of two real projective planes is a Klein bottle: . - the connected sum of three real projective planes (thus of a Klein bottle and a ...
[41]
Klein Bottle -- from Wolfram MathWorld
The Klein bottle is a closed nonorientable surface of Euler characteristic 0 (Dodson and Parker 1997, p. 125) that has no inside or outside.
[42]
The Fundamental Group of the Real Projective Plane - Math3ma
Feb 1, 2016 · As a quotient space, this is the same as a sphere whose antipodal points are identified. Equivalently, RP2 R P 2 is the quotient of a disc ...
[43]
Fundamental group of Klein bottle - Groupprops
Feb 1, 2012 · The fundamental group of the Klein bottle is a particular group, unique up to isomorphism.
[44]
Dyck's Surface -- from Wolfram MathWorld
Dyck's Surface: The surface with three cross-caps (Francis and Collins 1993, Francis and Weeks 1999).
[45]
Embedding Klein bottles in 4-space - MathOverflow
Mar 18, 2020 · A nicely parametrized embedding into 4-space, very similar to the standard parametrization of the torus in 3-space, can be found here.Why can't the Klein bottle embed in R3? - MathOverflowLagrangian Kleinian bottles - MathOverflowMore results from mathoverflow.net
[46]
real projective space in nLab
Mar 12, 2024 · Proof. Use that ℝ P n ≃ S n / ( ℤ / 2 ) is the quotient space of the Euclidean n-sphere by the ℤ / 2 -action which identifies antipodal points.Properties · Cell structure · Homotopy groups · Relation to the ℤ / 2 \mathbb{Z...
[47]
[PDF] notes on the course “algebraic topology” - University of Oregon
Real projective spaces. A real projective space RPn is a set of all lines in Rn+1 going through 0 ∈ Rn+1 . Let ℓ ∈ RPn be a line, then we define a basis ...
[48]
[PDF] Manifold Theory Peter Petersen - UCLA Mathematics
Thus we see that RPn is orientable iff n is odd. Using the double covering lemma show that the Klein bottle and the Möbius band are non-orientable. Manifolds ...
[49]
[PDF] Homework 5
(b) The real projective space is defined by RPn = Rn+1 − {0} ∼, where we ... n is orientable if and only if n is odd. 1. Page 2. Problem 3. Suppose M ...
[50]
[PDF] Fundamental group fact sheet Let X be a topological space. The set ...
From Corollary 0.8 it follows that the fundamental group of RPn is Z/2Z (this is the only group with 2 elements). Example 4. We can extend Corollary 0.8 to a ...
[51]
Homology of real projective space - Topospaces
Sep 30, 2011 · This article describes the value (and the process used to compute it) of some homotopy invariant(s) for a topological space or family of topological spaces.
[52]
[PDF] Imbedding of Manifolds in Euclidean Space
Let M be an n-manifold and let P c int Mbe an (mi-1)- dimensional polyhedron (O < 2m < n) such that the inclusion map i: P -. M is homotopic in M to a constant.
[53]
Is it true that all real projective space $RP^n$ can not be smoothly ...
May 10, 2012 · So first for n even, RPn is not orientable, hence can not be embedded in Rn+1. For odd n, RPn is orientable, hence the normal bundle is trivial.
[54]
[PDF] Projective spaces, the Fubini-Study metric and a little bit more
If V were a real space, this classifies all the totally geodesic submanifolds of PV. It is possible to modify the above argument to classify all totally ...