<!-- Generated from https://www.teclamairme.com/#group-T12 by readviews.py; never edited by hand. -->

# Glossary cards — Chapter ᚊ(Q): The gallery of shape

*Teclam Áirme, the Mathematical Compendium — a copy for AI readers of https://www.teclamairme.com/#group-T12, as of 2026-09-30. Links are relative to https://www.teclamairme.com/fortheaireaders/.*

[Guide for AI readers](index.md) · [The chapter](ch12.md)

*Topology, knot theory, differential geometry, and Euclidean geometry: Euler characteristic, Betti numbers, the torus’s genus, the unknot, curvature over the sphere, and the 3–4–5 triangle. Each card: its symbol, type, one-line headline, the longer note, the reference entries it cites and the cards it links to.*

### the letter Ceirt — `ᚊ` {#g-og-cert}

*Ogham letter*

Ogham letter name meaning ‘a bush’, confused early with ceirt, ‘rag’; the medieval glossators read it as the apple.

Cert is cognate with Welsh perth, ‘bush’, and with Latin quercus, ‘oak’, from Indo-European *perkwos. It was confused with a different Old Irish word, ceirt, ‘rag’, and the kennings carry the confusion. The three Old Irish word-oghams, the Bríatharogaim ascribed to Morann mac Moín, to Mac ind Óc, and to Cú Chulainn, give clithar baiscill, “shelter of a lunatic” (the reading is uncertain); bríg anduini, “substance of an insignificant person”; and dígu fethail, “dregs of clothing.” The Auraicept glosses aball, apple, and reads the first kenning as shelter for a wild hind. Modern tree lore and ogham divination descend from the medieval plant glosses by way of Robert Graves’s The White Goddess of 1948, which read the letter order as an ancient seasonal calendar of tree magic. Celtic scholarship rejects that reconstruction; the associations in circulation today are modern, not ancient.

Cites: [2.15](ch02-refs.md#r-2.15), [2.16](ch02-refs.md#r-2.16), [2.17](ch02-refs.md#r-2.17), [2.19](ch02-refs.md#r-2.19)

### Euler characteristic of the torus — `χ(T²)` {#g-chiT}

*Integer or rational*

V − E + F = 0 for any cutting of the torus into polygons.

The Euler characteristic χ = V − E + F was found by Euler in 1750 for convex polyhedra and generalized by Poincaré in 1895 to spaces of any dimension, the Euler–Poincaré characteristic; for a closed orientable surface of genus g it equals 2 − 2g, so the torus, with one handle, has χ = 0. A square with opposite edges glued, one vertex, two edges, and one face, gives 1 − 2 + 1 = 0 directly. The Klein bottle and Möbius band share the value. An integer.

Cites: [12.1](ch12-refs.md#r-12.1), [12.21](ch12-refs.md#r-12.21)

See also: [the classification of surfaces](ch12-cards.md#g-classsurf), [the Euler characteristic](ch12-cards.md#g-eulerchar), [the Klein bottle](ch12-cards.md#g-kleinbottle)

### Euler characteristic of the sphere — `χ(S²)` {#g-chiS}

*Integer or rational*

2, Euler’s polyhedron formula.

Euler’s polyhedron formula, V − E + F = 2, announced in 1750 and proved by Legendre in 1794 and Cauchy in 1811, holds for any polyhedron whose surface is a sphere; the cube gives 8 − 12 + 6 and the icosahedron 12 − 30 + 20. Descartes had a closely related result around 1630; his manuscript is lost, and it survives through a copy Leibniz made in 1676, itself lost until about 1860. Whether it amounts to the same formula is disputed. The 2 is the sphere’s Euler characteristic, and it is why exactly five Platonic solids exist. An integer.

Cites: [12.1](ch12-refs.md#r-12.1), [12.21](ch12-refs.md#r-12.21)

See also: [the Euler characteristic](ch12-cards.md#g-eulerchar), [the number of Platonic solids](ch12-cards.md#g-platonic)

### Euler characteristic of the projective plane — `χ(ℝP²)` {#g-chiRP}

*Integer or rational*

1: half the sphere’s 2, because identifying antipodal points halves every count.

The real projective plane is the sphere with antipodal points identified, or a disk with its boundary glued to a Möbius band; it cannot be embedded in three-dimensional space without crossing itself. Its Euler characteristic is half the sphere’s, 1, because the identification halves every count. Nonorientable surfaces have χ = 2 − k for k cross-caps, so ℝP² is the k = 1 case. An integer.

Cites: [12.1](ch12-refs.md#r-12.1)

See also: [the classification of surfaces](ch12-cards.md#g-classsurf), [the Euler characteristic](ch12-cards.md#g-eulerchar)

### Euler characteristic of the double torus — `χ(Σ₂)` {#g-chiS2g}

*Integer or rational*

−2, since χ = 2 − 2g.

The double torus, or genus-two surface, is a sphere with two handles, and χ = 2 − 2g gives −2; it is the first orientable surface whose Euler characteristic is negative, and by Gauss–Bonnet the first that must carry negative curvature on average, which is why its natural geometry is hyperbolic. The denominator uses its absolute value so that every term stays positive. An integer, −2, entered as 2.

Cites: [12.1](ch12-refs.md#r-12.1)

See also: [the Euler characteristic](ch12-cards.md#g-eulerchar)

### Euler characteristic of a point — `χ(•)` {#g-chipt}

*Integer or rational*

1: one vertex, no edges, no faces.

A single point has one vertex, no edges, and no faces, so V − E + F = 1; it is the simplest space there is. Any contractible space, a disk or a solid ball, has the same value. An integer.

Cites: [12.1](ch12-refs.md#r-12.1)

See also: [the Euler characteristic](ch12-cards.md#g-eulerchar)

### first Betti number of the sphere — `b_{1}(S²)` {#g-b1S2}

*Integer or rational*

0: every loop on a sphere contracts.

The first Betti number counts independent loops that cannot be shrunk to a point; on a sphere every loop slides off to a point, so b₁(S²) = 0, a fact Poincaré made rigorous in 1895. The sphere is simply connected as well, which is the stronger statement: a trivial fundamental group forces b₁ = 0, but b₁ = 0 does not force a trivial fundamental group, and in 1904 Poincaré himself built the counterexample, a closed 3-manifold with the homology of the 3-sphere and a fundamental group of 120 elements. That is why the Poincaré conjecture in the epilogue, that a simply connected closed 3-manifold is a sphere, is stated about loops and not about Betti numbers. An integer.

Cites: [12.17](ch12-refs.md#r-12.17), [12.21](ch12-refs.md#r-12.21), [12.3](ch12-refs.md#r-12.3)

See also: [Betti numbers](ch12-cards.md#g-bettinum), [homology groups](ch12-cards.md#g-homology), [the fundamental group](ch12-cards.md#g-pi1), [the Poincaré conjecture, as a bit](ch28-cards.md#g-poincare), [a closed three-manifold](ch28-cards.md#g-threemfld)

### one component and one cavity — `b_{0}(S²), b_{2}(S²)` {#g-b02S2}

*Integer or rational*

1 and 1: one connected piece and one enclosed cavity.

The sphere has one connected piece, b₀ = 1, and encloses one cavity, b₂ = 1, and these are the only nonzero Betti numbers it has; their alternating sum with b₁ = 0 gives χ = 2, recovering Euler’s polyhedron formula. Enrico Betti introduced these counts in his paper of 1871, and Poincaré, who named them after him, gave them a more precise definition in Analysis Situs in 1895. Integers, 1 and 1.

Cites: [12.2](ch12-refs.md#r-12.2), [12.3](ch12-refs.md#r-12.3)

See also: [first Betti number of the sphere](ch12-cards.md#g-b1S2), [Betti numbers](ch12-cards.md#g-bettinum), [Euler characteristic of the sphere](ch12-cards.md#g-chiS), [the Euler characteristic](ch12-cards.md#g-eulerchar)

### first Betti number of the torus — `b_{1}(T²)` {#g-b1T2}

*Integer or rational*

2 independent loops.

The torus has two independent loops that cannot be shrunk or slid into each other: one around the tube and one through the hole. So b₁(T²) = 2, and with b₀ = b₂ = 1 the alternating sum 1 − 2 + 1 = 0 is the torus’s Euler characteristic, the numerator’s zero seen from the other side. An integer.

Cites: [12.3](ch12-refs.md#r-12.3)

See also: [the Euler characteristic](ch12-cards.md#g-eulerchar)

### first Betti number of the double torus — `b_{1}(Σ₂)` {#g-b1S2g}

*Integer or rational*

4: two independent loops per handle, and the double torus has two handles.

A surface of genus g has b₁ = 2g, one around and one through each handle, so the double torus has four; the alternating sum 1 − 4 + 1 = −2 matches its Euler characteristic. These loops generate the fundamental group’s abelianization, which is how homology and homotopy meet. An integer.

Cites: [12.3](ch12-refs.md#r-12.3)

See also: [the Euler characteristic](ch12-cards.md#g-eulerchar), [homology groups](ch12-cards.md#g-homology), [the fundamental group](ch12-cards.md#g-pi1)

### Euler characteristic of the Klein bottle — `χ(Klein bottle)` {#g-chiK}

*Integer or rational*

0, the same as the torus; the Klein bottle is glued from a square with one pair of edges joined straight and the other pair joined with a twist.

Felix Klein described the bottle in 1882; it is a closed nonorientable surface with two cross-caps, so χ = 2 − 2 = 0, and like the torus it is built from one square, V − E + F = 1 − 2 + 1. It cannot sit in three-dimensional space without passing through itself. An integer.

Cites: [12.1](ch12-refs.md#r-12.1), [12.21](ch12-refs.md#r-12.21)

See also: [the Euler characteristic](ch12-cards.md#g-eulerchar)

### the boundary of a boundary — `∂(∂M)` {#g-bdbd}

*Integer or rational*

The boundary of a boundary is the empty chain: ∂∂ = 0. It is an operator identity, and this empty chain comes home as the integer 0.

The boundary of a region is a closed surface, and a closed surface has no boundary. Poincaré introduced homology in Analysis Situs (1895) and its complements (1899–1904), and ∂∘∂ = 0 is its foundation: a chain complex is exactly a sequence of maps whose successive compositions vanish, and homology measures the gap between cycles (∂c = 0) and boundaries (c = ∂d). It is the dual of d² = 0 in Chapter ᚉ(C). The value is the empty chain; as a mixed-type term, it comes home as the integer 0.

Cites: [12.15](ch12-refs.md#r-12.15), [12.21](ch12-refs.md#r-12.21), [12.4](ch12-refs.md#r-12.4)

See also: [the exterior derivative applied twice](ch11-cards.md#g-ddzero), [homology groups](ch12-cards.md#g-homology)

### the genus of the torus — `g(T²)` {#g-genusT}

*Integer or rational*

1: one handle, one hole, and χ = 2 − 2g gives the numerator’s zero.

The genus of a closed orientable surface is its number of handles; the classification of surfaces — begun by Möbius (1863, or 1861 by one specialist account) and Jordan (1866), carried on by von Dyck (1888) and by Dehn and Heegaard (1907), and proved rigorously by Brahana (1921) — says the genus determines the surface. The torus has genus 1, and the formula χ = 2 − 2g ties the whole chapter together. An integer.

Cites: [12.5](ch12-refs.md#r-12.5)

See also: [the classification of surfaces](ch12-cards.md#g-classsurf), [the Euler characteristic](ch12-cards.md#g-eulerchar)

### cohomology — `Hⁿ(X)` {#g-cohomology}

*Function or operator*

Homology read backwards: instead of the cycles in a space, the functions on them—which is what lets cohomology classes be multiplied.

Where homology builds groups out of the cycles in a space, cohomology builds them out of cochains—functions on chains—and the reversal buys something homology does not have: a product. The cup product makes the cohomology of a space into a graded ring, so two classes can be multiplied and not merely counted. It arrived about forty years after homology: Alexander had a notion of cochain by 1930, he and Kolmogorov both introduced cohomology at a Moscow conference in 1935, and Whitney and Čech built the cup and cap products between 1936 and 1938. This is the language the epilogue’s Hodge question is asked in — a Hodge class is a cohomology class, and the conjecture asks whether every one of them comes from subvarieties. A tool of the framework, uncolored.

Cites: [12.20](ch12-refs.md#r-12.20)

See also: [the Hodge conjecture, as a bit](ch28-cards.md#g-hodge), [a Hodge class](ch28-cards.md#g-hodgeclass), [homology groups](ch12-cards.md#g-homology)

### Betti numbers — `bₙ(X)` {#g-bettinum}

*Function or operator*

The ranks of a space’s homology groups, counting independent features of each dimension: connected pieces, independent loops, enclosed cavities, and so on.

Enrico Betti introduced these counts in a paper of 1871, Henri Poincaré named them after him and gave them a rigorous definition in Analysis situs of 1895, and Emmy Noether recast them in the 1920s as the ranks of the abelian homology groups Hₙ(X). For the sphere b₀ = 1, b₁ = 0, b₂ = 1; for the torus b₁ = 2; for a genus-two surface b₁ = 4. Their alternating sum is the Euler characteristic, χ = b₀ − b₁ + b₂ − ⋯, which is why every Euler characteristic in this chapter can be read off as a homology count. Specific values, such as the sphere’s and the tori’s first Betti numbers, have cards of their own; this card is the general notion.

Cites: [12.23](ch12-refs.md#r-12.23)

See also: [one component and one cavity](ch12-cards.md#g-b02S2), [first Betti number of the sphere](ch12-cards.md#g-b1S2), [the Euler characteristic](ch12-cards.md#g-eulerchar), [homology groups](ch12-cards.md#g-homology)

### the Klein bottle — `N₂` {#g-kleinbottle}

*Non-real object*

A closed nonorientable surface, glued from a square with one pair of edges joined straight and the other joined with a twist, that cannot be embedded in ordinary three-dimensional space without self-intersection.

Felix Klein first described the Klein bottle in 1882. Like the torus it has Euler characteristic 0, but unlike the torus it is one-sided: cut it in half and two Möbius bands fall out, and no amount of stretching turns the Klein bottle into the torus, since the torus is orientable and the Klein bottle is not. It sits in the classification of closed surfaces as N₂, the nonorientable surface with two cross-caps, alongside the projective plane N₁ and the orientable surfaces Σ_{g}; it can be pictured as a tube passed back through itself to join its two ends from the inside, which is why a physical model must let the surface pass through itself, something a true Klein bottle in four dimensions does not need to do.

Cites: [12.24](ch12-refs.md#r-12.24)

See also: [the Euler characteristic](ch12-cards.md#g-eulerchar)

### crossing number of the unknot — `c(○)` {#g-unknot}

*Integer or rational*

The unknot can be drawn with no crossings at all.

The crossing number of a knot is the fewest crossings in any drawing of it; the unknot, a plain loop, can be drawn with none, so c(○) = 0, while the trefoil needs three and no knot needs one or two. Peter Guthrie Tait tabulated knots in the 1870s and 1880s for Lord Kelvin’s theory that atoms were knots in the ether; the theory was wrong and the tables outlived it. Whether a given tangle is the unknot is decidable, but no polynomial-time method is known. An integer.

Cites: [12.22](ch12-refs.md#r-12.22), [12.6](ch12-refs.md#r-12.6)

See also: [crossing number](ch12-cards.md#g-crossnum), [crossing number of the trefoil](ch12-cards.md#g-trefoil)

### crossing number of the trefoil — `c(3₁)` {#g-trefoil}

*Integer or rational*

3: the simplest knot that is really knotted needs three crossings, and no knot needs one or two.

The trefoil is the first entry in Tait’s tables of the 1870s, the (2, 3) torus knot; its crossing number is 3, proved by tricoloring or by the Jones polynomial, and it is chiral, distinct from its mirror image, which Max Dehn showed in 1914. An integer.

Cites: [12.22](ch12-refs.md#r-12.22), [12.6](ch12-refs.md#r-12.6)

See also: [crossing number](ch12-cards.md#g-crossnum), [the Jones polynomial](ch12-cards.md#g-jonespoly)

### the Jones polynomial of the unknot — `V_{○}(t)` {#g-jonesunknot}

*Integer or rational*

1: the invariant Vaughan Jones found in 1984 gives the unknot the constant polynomial 1, and it is unknown whether any other knot does.

Vaughan Jones discovered his polynomial in 1984 while working on von Neumann algebras, and won the Fields Medal for it in 1990; it distinguishes the trefoil from its mirror image, which the older Alexander polynomial of 1928 could not. The unknot has V = 1, and whether the Jones polynomial detects the unknot, i.e. whether V = 1 forces a knot to be trivial, is an open problem. An integer, 1.

Cites: [12.22](ch12-refs.md#r-12.22), [12.25](ch12-refs.md#r-12.25), [12.7](ch12-refs.md#r-12.7), [12.8](ch12-refs.md#r-12.8)

See also: [the Jones polynomial](ch12-cards.md#g-jonespoly)

### the Jones polynomial — `V(t)` {#g-jonespoly}

*Function or operator*

A knot invariant, a Laurent polynomial in t assigned to every knot or link, that distinguishes knots the older Alexander polynomial could not.

Vaughan Jones discovered the polynomial in 1984 while studying von Neumann algebras, and the discovery won him the 1990 Fields Medal. It is normalized so the unknot has V(t) = 1, and it can tell a knot from its mirror image—unlike the Alexander polynomial, it distinguishes the trefoil from its reflection. Whether any nontrivial knot can also have V(t) = 1, meaning the polynomial would fail to detect the unknot, remains an open problem. The unknot’s value has a card of its own; this card is the polynomial itself.

Cites: [12.25](ch12-refs.md#r-12.25)

### crossing number — `c(K)` {#g-crossnum}

*Function or operator*

The fewest crossings in any diagram of a knot: 0 for the unknot, and at least 3 for every knot that is genuinely knotted.

The crossing number of a knot K is the minimum number of crossings over every possible planar diagram of it; the unknot achieves c(○) = 0, and no nontrivial knot can be drawn with fewer than three crossings, a bound the trefoil meets exactly. Peter Guthrie Tait tabulated knots by crossing number in the 1870s and 1880s while pursuing Lord Kelvin’s theory that atoms were knots in the ether, and the tables outlived the physics they were built for. Whether a given diagram already achieves the true minimum is decidable in principle but has no known fast general method. The unknot’s and the trefoil’s values have cards of their own; this card is the function itself.

Cites: [12.27](ch12-refs.md#r-12.27)

See also: [crossing number of the unknot](ch12-cards.md#g-unknot)

### Gaussian curvature of a cylinder — `K(cylinder)` {#g-cylinder}

*Integer or rational*

A cylinder is flat: it unrolls onto a plane, so its Gaussian curvature is 0 everywhere.

Gaussian curvature, defined by Gauss in his Disquisitiones generales circa superficies curvas of 1827, is the product of the two principal curvatures at a point; a cylinder curves one way and is straight the other, so the product is 0 everywhere, and it can be unrolled flat without stretching. Gauss’s Theorema Egregium says this curvature is intrinsic, measurable from within the surface, which is why a flat map of the round Earth must distort. An integer.

Cites: [12.9](ch12-refs.md#r-12.9)

### Gauss–Bonnet for the sphere — `∫_{S²} K dA` {#g-gaussbonnet}

*Proven transcendental*

4π: total curvature equals 2π times the Euler characteristic, so geometry and topology agree on every closed surface.

The Gauss–Bonnet theorem says ∫K dA = 2πχ for any closed surface; the unit sphere has K = 1 and area 4π, and 2π·2 = 4π. Gauss had a version he never published and Pierre Ossian Bonnet published a special case in 1848, but the closed-surface statement is credited to Walther von Dyck in 1888. It is why the torus, with χ = 0, must be flat on average, and the double torus negatively curved on average. Transcendental, because π is.

Cites: [12.10](ch12-refs.md#r-12.10), [12.16](ch12-refs.md#r-12.16)

See also: [Euler characteristic of the torus](ch12-cards.md#g-chiT), [the Euler characteristic](ch12-cards.md#g-eulerchar)

### the number of Platonic solids — `5` {#g-platonic}

*Integer or rational*

Exactly five regular convex polyhedra exist, a fact Euclid’s Elements ends with; Euler’s formula caps the count at five in a line.

Book XIII of Euclid’s Elements, c. 300 BC, closes by proving that the five regular convex solids are the only ones, and appears to be based largely on an earlier treatise by Theaetetus; Euler’s polyhedron formula gives the modern proof, since V − E + F = 2 with p-gon faces and q faces at each vertex forces 1/p + 1/q > 1/2, which has five solutions. An integer.

Cites: [12.12](ch12-refs.md#r-12.12), [12.22](ch12-refs.md#r-12.22)

See also: [Euler characteristic of the sphere](ch12-cards.md#g-chiS), [the Euler characteristic](ch12-cards.md#g-eulerchar)

### the 3–4–5 right triangle — `3²+4²−5²` {#g-pythag345}

*Integer or rational*

Exactly 0: the most familiar integer instance of the Pythagorean theorem.

A right triangle with legs 3 and 4 has hypotenuse 5, so 3²+4²−5²=9+16−25=0. Euclid’s Elements I.47 gives the classical proof around 300 BC, while the theorem and integer triples were known in older mathematical cultures and attribution to Pythagoras is historically uncertain. The identity gives elementary Euclidean geometry a visible carrier.

Cites: [12.11](ch12-refs.md#r-12.11), [12.12](ch12-refs.md#r-12.12), [12.22](ch12-refs.md#r-12.22)

### the angle sum of a triangle — `∠A + ∠B + ∠C` {#g-anglesum}

*Proven transcendental*

π, or 180 degrees: Euclid I.32, true in the flat plane and false on the sphere; on the unit sphere the excess over π is the triangle’s area.

Euclid’s Elements I.32 proves that the three angles of a triangle add to two right angles, using the parallel postulate. On a sphere of radius 1 the sum exceeds π by exactly the triangle’s area (Girard, ~1626: the triangle formula is in his 1626 corrected edition of Stevin, and his 1629 main work generalized it to spherical polygons), and on a saddle it falls short, which is why the angle sum decides which geometry you are in and why this chapter’s Gauss–Bonnet integral is the same fact for a whole surface. Transcendental, being π.

Cites: [12.12](ch12-refs.md#r-12.12), [12.13](ch12-refs.md#r-12.13)

See also: [Gauss–Bonnet for the sphere](ch12-cards.md#g-gaussbonnet), [the parallel postulate](ch12-cards.md#g-parallel)

### homeomorphism — `X ≅ Y` {#g-homeo}

*Function or operator*

Two spaces are the same to a topologist when one can be stretched into the other without cutting or gluing: the coffee cup and the doughnut.

A homeomorphism is a continuous bijection with continuous inverse; Poincaré’s Analysis Situs of 1895 made topology the study of what such maps preserve, and every number in the topology brace, χ, the Betti numbers, genus, is a homeomorphism invariant. A relation, not a number.

Cites: [12.2](ch12-refs.md#r-12.2)

See also: [Betti numbers](ch12-cards.md#g-bettinum), [a one-to-one correspondence](ch27-cards.md#g-bijection), [the Euler characteristic](ch12-cards.md#g-eulerchar)

### homology groups — `H_{n}(X)` {#g-homology}

*Function or operator*

The groups that count n-dimensional holes; their ranks are the Betti numbers of the denominator.

Homology assigns to a space X a sequence of abelian groups H₀, H₁, H₂, …, built from the chain complex with ∂∂ = 0; Poincaré defined the idea in 1895, Emmy Noether recognized the groups behind Betti’s numbers in the 1920s, and b_{n} = rank H_{n}(X). For the torus H₁ ≅ ℤ², which is the 2 in the denominator. A functor, not a number.

Cites: [12.2](ch12-refs.md#r-12.2), [12.3](ch12-refs.md#r-12.3), [12.4](ch12-refs.md#r-12.4)

See also: [first Betti number of the torus](ch12-cards.md#g-b1T2), [the boundary of a boundary](ch12-cards.md#g-bdbd), [Betti numbers](ch12-cards.md#g-bettinum)

### the fundamental group — `π₁(X)` {#g-pi1}

*Function or operator*

The loops in a space up to deformation; π₁(S²) is trivial, π₁(S¹) ≅ ℤ, and the Poincaré conjecture is a statement about π₁.

Poincaré defined the fundamental group in ~1894 as the group of loops based at a point, composed by concatenation, up to continuous deformation; its abelianization is H₁, so the first Betti number is its rank. The sphere’s π₁ is trivial, which is the numerator’s b₁(S²) = 0 seen through homotopy; the epilogue’s Poincaré conjecture says trivial π₁ characterizes the 3-sphere. A group, not a number.

Cites: [12.2](ch12-refs.md#r-12.2), [12.3](ch12-refs.md#r-12.3)

See also: [first Betti number of the sphere](ch12-cards.md#g-b1S2), [homology groups](ch12-cards.md#g-homology), [the Poincaré conjecture, as a bit](ch28-cards.md#g-poincare)

### the classification of surfaces — `Σ_{g}, N_{k}` {#g-classsurf}

*Function or operator*

Every closed surface is a sphere with g handles (χ = 2 − 2g) or with k cross-caps (χ = 2 − k), and nothing else.

The classification theorem, stated by Möbius in 1863, or 1861 by one specialist account, carried further by Dehn and Heegaard in 1907, and first proved rigorously by Brahana in 1921, lists every closed surface once: orientable Σ_{g} with χ = 2 − 2g, nonorientable N_{k} with χ = 2 − k. Every surface in this chapter is on the list: S² = Σ₀, T² = Σ₁, Σ₂, ℝP² = N₁, the Klein bottle N₂. A theorem, not a number.

Cites: [12.1](ch12-refs.md#r-12.1), [12.5](ch12-refs.md#r-12.5)

See also: [the Klein bottle](ch12-cards.md#g-kleinbottle)

### the Euler characteristic — `χ` {#g-eulerchar}

*Function or operator*

V − E + F for a polyhedron, and the alternating sum of the Betti numbers in general; it does not change under continuous deformation.

For a polyhedral decomposition, χ = V − E + F, and Euler announced the value 2 for convex polyhedra in a 1750 letter to Goldbach, publishing the details in two papers of ~1752—written in 1750 and 1751, and printed in the Novi Commentarii volume for 1752–53, which itself appeared only in 1758. The modern definition is the alternating sum of the Betti numbers, χ = b₀ − b₁ + b₂ − b₃ + ⋯. Because homology is a topological and even a homotopy invariant, so is χ—two spaces that can be deformed into one another share it. For a closed orientable surface of genus g it is 2 − 2g, which is why the Euler characteristics of this chapter’s orientable surfaces can be read off the number of handles. An invariant, not a number of this chapter’s own, so uncolored.

Cites: [12.1](ch12-refs.md#r-12.1), [12.18](ch12-refs.md#r-12.18)

See also: [Betti numbers](ch12-cards.md#g-bettinum), [Euler characteristic of the sphere](ch12-cards.md#g-chiS), [the classification of surfaces](ch12-cards.md#g-classsurf), [homology groups](ch12-cards.md#g-homology)

### the parallel postulate — `Postulate 5` {#g-parallel}

*Function or operator*

Euclid’s fifth postulate, the assumption the angle sum of a triangle rests on; without it the sum need not be π: it rises above π on the sphere and falls below it on the saddle.

Euclid’s proof that the three angles of a triangle add to two right angles, Elements I.32, runs through I.31 and I.29 and so depends on the parallel postulate. That dependence is what makes the angle sum a test of which geometry you are standing in: on the unit sphere the sum exceeds π by exactly the triangle’s area, and a hyperbolic triangle’s angles add to less than π. The same seam, taken over a whole surface instead of one triangle, is this chapter’s Gauss–Bonnet integral. A postulate, not a number.

Cites: [12.12](ch12-refs.md#r-12.12), [12.13](ch12-refs.md#r-12.13)

See also: [the angle sum of a triangle](ch12-cards.md#g-anglesum), [Gauss–Bonnet for the sphere](ch12-cards.md#g-gaussbonnet)