tiling the hyperbolic plane - long version
TRANSCRIPT
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Tiling the hyperbolic plane - longversion
Daniel Czegel
Eotvos Lorand University, Budapest
ICPS, HeidelbergAugust 14, 2014
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Definition of Gaussian curvature
1 find a normal vector N
2 rotate the normal planecontaining N
3 intersection of the surfaceand the normal plane: aplane curve, curvature:κ = 1
R
4 Gaussian curvature of thesurface:
K (r) = κmin(r) κmax(r)
5 sign!
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Definition of Gaussian curvature
1 find a normal vector N
2 rotate the normal planecontaining N
3 intersection of the surfaceand the normal plane: aplane curve, curvature:κ = 1
R
4 Gaussian curvature of thesurface:
K (r) = κmin(r) κmax(r)
5 sign!
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Definition of Gaussian curvature
1 find a normal vector N
2 rotate the normal planecontaining N
3 intersection of the surfaceand the normal plane: aplane curve, curvature:κ = 1
R
4 Gaussian curvature of thesurface:
K (r) = κmin(r) κmax(r)
5 sign!
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Definition of Gaussian curvature
1 find a normal vector N
2 rotate the normal planecontaining N
3 intersection of the surfaceand the normal plane: aplane curve, curvature:κ = 1
R
4 Gaussian curvature of thesurface:
K (r) = κmin(r) κmax(r)
5 sign!
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Definition of Gaussian curvature
1 find a normal vector N
2 rotate the normal planecontaining N
3 intersection of the surfaceand the normal plane: aplane curve, curvature:κ = 1
R
4 Gaussian curvature of thesurface:
K (r) = κmin(r) κmax(r)
5 sign!
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Theorema Egregium
Does K (r) change if we wrap, bend, twist (i.e. change theembedding in the 3 dim space)?
Theorema Egregium (Great Theorem): No!
Gaussian curvature is an intrinsic measure of the surface!
In a 2 dimensional Universe, K fully determines howspacetime curves: GR, Einstein eqs.:
Rµν −R
2gµν + Λgµν =
8πG
c4Tµν
In 2 dim:Rµν = f (K , gµν), R = 2K
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Theorema Egregium
Does K (r) change if we wrap, bend, twist (i.e. change theembedding in the 3 dim space)?
Theorema Egregium (Great Theorem): No!
Gaussian curvature is an intrinsic measure of the surface!
In a 2 dimensional Universe, K fully determines howspacetime curves: GR, Einstein eqs.:
Rµν −R
2gµν + Λgµν =
8πG
c4Tµν
In 2 dim:Rµν = f (K , gµν), R = 2K
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Theorema Egregium
Does K (r) change if we wrap, bend, twist (i.e. change theembedding in the 3 dim space)?
Theorema Egregium (Great Theorem): No!
Gaussian curvature is an intrinsic measure of the surface!
In a 2 dimensional Universe, K fully determines howspacetime curves: GR, Einstein eqs.:
Rµν −R
2gµν + Λgµν =
8πG
c4Tµν
In 2 dim:Rµν = f (K , gµν), R = 2K
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Theorema Egregium
Does K (r) change if we wrap, bend, twist (i.e. change theembedding in the 3 dim space)?
Theorema Egregium (Great Theorem): No!
Gaussian curvature is an intrinsic measure of the surface!
In a 2 dimensional Universe, K fully determines howspacetime curves: GR, Einstein eqs.:
Rµν −R
2gµν + Λgµν =
8πG
c4Tµν
In 2 dim:Rµν = f (K , gµν), R = 2K
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Theorema Egregium
Does K (r) change if we wrap, bend, twist (i.e. change theembedding in the 3 dim space)?
Theorema Egregium (Great Theorem): No!
Gaussian curvature is an intrinsic measure of the surface!
In a 2 dimensional Universe, K fully determines howspacetime curves: GR, Einstein eqs.:
Rµν −R
2gµν + Λgµν =
8πG
c4Tµν
In 2 dim:Rµν = f (K , gµν), R = 2K
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Cylinder, Oloid
⇒ a surface can be unfolded without distortion ⇔ K ≡ 0
cylinder:
κmax =1
R, κmin =
1
∞= 0 ⇒ K ≡ 0
nontrivial example: oloid: convex hull of two perpendicularcircles
every point of its surface touches the floor during rolling!
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Cylinder, Oloid
⇒ a surface can be unfolded without distortion ⇔ K ≡ 0
cylinder:
κmax =1
R, κmin =
1
∞= 0 ⇒ K ≡ 0
nontrivial example: oloid: convex hull of two perpendicularcircles
every point of its surface touches the floor during rolling!
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Cylinder, Oloid
⇒ a surface can be unfolded without distortion ⇔ K ≡ 0
cylinder:
κmax =1
R, κmin =
1
∞= 0 ⇒ K ≡ 0
nontrivial example: oloid: convex hull of two perpendicularcircles
every point of its surface touches the floor during rolling!
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Cylinder, Oloid
⇒ a surface can be unfolded without distortion ⇔ K ≡ 0
cylinder:
κmax =1
R, κmin =
1
∞= 0 ⇒ K ≡ 0
nontrivial example: oloid: convex hull of two perpendicularcircles
every point of its surface touches the floor during rolling!
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Spherical triangles
sphere: K =?
K ≡ 1R2 > 0, a model of elliptic geometry;
”straight lines”=geodesics: great circles
What is the sum of angles Σ in a triangle?
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Spherical triangles
sphere: K =?
K ≡ 1R2 > 0, a model of elliptic geometry;
”straight lines”=geodesics: great circles
What is the sum of angles Σ in a triangle?
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Spherical triangles
sphere: K =?
K ≡ 1R2 > 0, a model of elliptic geometry;
”straight lines”=geodesics: great circles
What is the sum of angles Σ in a triangle?
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Spherical triangles
sphere: K =?
K ≡ 1R2 > 0, a model of elliptic geometry;
”straight lines”=geodesics: great circles
What is the sum of angles Σ in a triangle?
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Spherical triangles
i) Σ =3
2π, A =
π
2R2 ii) Σ = 3π, A = 2πR2
Generally?
A = (Σ− π)R2 ⇒ KA = Σ− π
The sum of angles Σ is size-independent: only if K = 0(euclidean geometry)
elliptic geometry:
KA > 0 ⇒ Σ > π
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Spherical triangles
i) Σ =3
2π, A =
π
2R2 ii) Σ = 3π, A = 2πR2
Generally?
A = (Σ− π)R2 ⇒ KA = Σ− π
The sum of angles Σ is size-independent: only if K = 0(euclidean geometry)
elliptic geometry:
KA > 0 ⇒ Σ > π
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Spherical triangles
i) Σ =3
2π, A =
π
2R2 ii) Σ = 3π, A = 2πR2
Generally?
A = (Σ− π)R2 ⇒ KA = Σ− π
The sum of angles Σ is size-independent: only if K = 0(euclidean geometry)
elliptic geometry:
KA > 0 ⇒ Σ > π
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Spherical triangles
i) Σ =3
2π, A =
π
2R2 ii) Σ = 3π, A = 2πR2
Generally?
A = (Σ− π)R2 ⇒ KA = Σ− π
The sum of angles Σ is size-independent: only if K = 0(euclidean geometry)
elliptic geometry:
KA > 0 ⇒ Σ > π
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Stereographic projection
sphere: K 6= 0 ⇒ cannot be unfolded:there is no sphere → plane map that preserves both distanceand angle!
We have to choose; e.g.: preserves angle (conformal), butdoes not preserve distance: stereographic projection
unit sphere, equator ⊂ plane, project from the north pole
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Stereographic projection
sphere: K 6= 0 ⇒ cannot be unfolded:there is no sphere → plane map that preserves both distanceand angle!We have to choose; e.g.: preserves angle (conformal), butdoes not preserve distance: stereographic projection
unit sphere, equator ⊂ plane, project from the north pole
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Stereographic projection
sphere: K 6= 0 ⇒ cannot be unfolded:there is no sphere → plane map that preserves both distanceand angle!We have to choose; e.g.: preserves angle (conformal), butdoes not preserve distance: stereographic projection
unit sphere, equator ⊂ plane, project from the north poleDaniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Properties of stereographic projection
northern (southern) hemisphere 7→ outside (inside) of the unitcircle (north pole 7→ ∞)
circle 7→ circle
special case: geodetic (”straight line”)=great circle 7→ circle
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Properties of stereographic projection
northern (southern) hemisphere 7→ outside (inside) of the unitcircle (north pole 7→ ∞)
circle 7→ circle
special case: geodetic (”straight line”)=great circle 7→ circle
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Properties of stereographic projection
northern (southern) hemisphere 7→ outside (inside) of the unitcircle (north pole 7→ ∞)
circle 7→ circle
special case: geodetic (”straight line”)=great circle 7→ circle
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Properties of stereographic projection
northern (southern) hemisphere 7→ outside (inside) of the unitcircle (north pole 7→ ∞)
circle 7→ circle
special case: geodetic (”straight line”)=great circle 7→ circle
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Metric induced by stereographic projection
The spherical geometry can be modeled on a plane
Do not forget: this model does not preserve distance, instead,the metric:
ds2 = (dx2 + dy2)
(4
(1 + x2 + y2)2
)metric tensor:
g = λ(x , y)
(1 00 1
)⇒ isotropic scaling ⇒ conformal (angle-preserving)
distance on the sphere, if (0, 0)→∞ on the plane?∫ ∞0
ds =
∫ ∞0
2
1 + x2dx = π
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Metric induced by stereographic projection
The spherical geometry can be modeled on a plane
Do not forget: this model does not preserve distance, instead,the metric:
ds2 = (dx2 + dy2)
(4
(1 + x2 + y2)2
)
metric tensor:
g = λ(x , y)
(1 00 1
)⇒ isotropic scaling ⇒ conformal (angle-preserving)
distance on the sphere, if (0, 0)→∞ on the plane?∫ ∞0
ds =
∫ ∞0
2
1 + x2dx = π
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Metric induced by stereographic projection
The spherical geometry can be modeled on a plane
Do not forget: this model does not preserve distance, instead,the metric:
ds2 = (dx2 + dy2)
(4
(1 + x2 + y2)2
)metric tensor:
g = λ(x , y)
(1 00 1
)⇒ isotropic scaling ⇒ conformal (angle-preserving)
distance on the sphere, if (0, 0)→∞ on the plane?∫ ∞0
ds =
∫ ∞0
2
1 + x2dx = π
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Metric induced by stereographic projection
The spherical geometry can be modeled on a plane
Do not forget: this model does not preserve distance, instead,the metric:
ds2 = (dx2 + dy2)
(4
(1 + x2 + y2)2
)metric tensor:
g = λ(x , y)
(1 00 1
)⇒ isotropic scaling ⇒ conformal (angle-preserving)
distance on the sphere, if (0, 0)→∞ on the plane?
∫ ∞0
ds =
∫ ∞0
2
1 + x2dx = π
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Metric induced by stereographic projection
The spherical geometry can be modeled on a plane
Do not forget: this model does not preserve distance, instead,the metric:
ds2 = (dx2 + dy2)
(4
(1 + x2 + y2)2
)metric tensor:
g = λ(x , y)
(1 00 1
)⇒ isotropic scaling ⇒ conformal (angle-preserving)
distance on the sphere, if (0, 0)→∞ on the plane?∫ ∞0
ds =
∫ ∞0
2
1 + x2dx = π
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
The Earth
The image of the Earth under stereographic projection from thesouth pole:
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Hyperbolic geometry
Is K < 0 (negative curvature) possible at a point?
Saddle point: κmin < 0, κmax > 0
Hyperbolic plane: saddle points everywhere! (e.g. K ≡ −1)Can you imagine it? Is it possible to embed it into a 3 dimeuclidean space?
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Hyperbolic geometry
Is K < 0 (negative curvature) possible at a point?Saddle point: κmin < 0, κmax > 0
Hyperbolic plane: saddle points everywhere! (e.g. K ≡ −1)Can you imagine it? Is it possible to embed it into a 3 dimeuclidean space?
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Hyperbolic geometry
Is K < 0 (negative curvature) possible at a point?Saddle point: κmin < 0, κmax > 0
Hyperbolic plane: saddle points everywhere! (e.g. K ≡ −1)
Can you imagine it? Is it possible to embed it into a 3 dimeuclidean space?
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Hyperbolic geometry
Is K < 0 (negative curvature) possible at a point?Saddle point: κmin < 0, κmax > 0
Hyperbolic plane: saddle points everywhere! (e.g. K ≡ −1)Can you imagine it? Is it possible to embed it into a 3 dimeuclidean space?
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Hyperbolic paper
Cut equilateral triangles
At every vertex, glue 7 (instead of 6) triangles to each other!
What happens, if 5 triangles are glued at a vertex?
Icosahedron!
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Hyperbolic paper
Cut equilateral triangles
At every vertex, glue 7 (instead of 6) triangles to each other!
What happens, if 5 triangles are glued at a vertex?
Icosahedron!
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Hyperbolic paper
Cut equilateral triangles
At every vertex, glue 7 (instead of 6) triangles to each other!
What happens, if 5 triangles are glued at a vertex?
Icosahedron!
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Hyperbolic paper
Cut equilateral triangles
At every vertex, glue 7 (instead of 6) triangles to each other!
What happens, if 5 triangles are glued at a vertex?
Icosahedron!
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Hyperbolic paper
Cut equilateral triangles
At every vertex, glue 7 (instead of 6) triangles to each other!
What happens, if 5 triangles are glued at a vertex?
Icosahedron!
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Hyperbolic triangles
Sum of angles Σ in a triangle?
elliptic & euclidean case:
KA = Σ− π
good news: it is also valid for K < 0!
A > 0 ⇒ KA < 0 ⇒ Σ < π
special case: K = −1
A = π − Σ ≤ π
Despite the hyperbolic plane is infinite, no triangle can havelarger area than π!!
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Hyperbolic triangles
Sum of angles Σ in a triangle?
elliptic & euclidean case:
KA = Σ− π
good news: it is also valid for K < 0!
A > 0 ⇒ KA < 0 ⇒ Σ < π
special case: K = −1
A = π − Σ ≤ π
Despite the hyperbolic plane is infinite, no triangle can havelarger area than π!!
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Hyperbolic triangles
Sum of angles Σ in a triangle?
elliptic & euclidean case:
KA = Σ− π
good news: it is also valid for K < 0!
A > 0 ⇒ KA < 0 ⇒ Σ < π
special case: K = −1
A = π − Σ ≤ π
Despite the hyperbolic plane is infinite, no triangle can havelarger area than π!!
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Hyperbolic triangles
Sum of angles Σ in a triangle?
elliptic & euclidean case:
KA = Σ− π
good news: it is also valid for K < 0!
A > 0 ⇒ KA < 0 ⇒ Σ < π
special case: K = −1
A = π − Σ ≤ π
Despite the hyperbolic plane is infinite, no triangle can havelarger area than π!!
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Hyperbolic triangles
Sum of angles Σ in a triangle?
elliptic & euclidean case:
KA = Σ− π
good news: it is also valid for K < 0!
A > 0 ⇒ KA < 0 ⇒ Σ < π
special case: K = −1
A = π − Σ ≤ π
Despite the hyperbolic plane is infinite, no triangle can havelarger area than π!!
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Hyperbolic triangles
Sum of angles Σ in a triangle?
elliptic & euclidean case:
KA = Σ− π
good news: it is also valid for K < 0!
A > 0 ⇒ KA < 0 ⇒ Σ < π
special case: K = −1
A = π − Σ ≤ π
Despite the hyperbolic plane is infinite, no triangle can havelarger area than π!!
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Poincare model of the hyperbolic plane
K 6= 0 ⇒ There is no distance-preserving ANDangle-preserving map to the euclidean plane
Poincare model: angle preserving, but not distance preserving(like the stereographic projection)
infinite hyperbolic plane → unit disc
infinity 7→ edge of the unit disk
geodesics (”straight lines”) 7→ circles that meet the edge ofthe unit disc at 90◦
parallel lines 7→ not intersecting such circles
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Poincare model of the hyperbolic plane
K 6= 0 ⇒ There is no distance-preserving ANDangle-preserving map to the euclidean plane
Poincare model: angle preserving, but not distance preserving(like the stereographic projection)
infinite hyperbolic plane → unit disc
infinity 7→ edge of the unit disk
geodesics (”straight lines”) 7→ circles that meet the edge ofthe unit disc at 90◦
parallel lines 7→ not intersecting such circles
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Poincare model of the hyperbolic plane
K 6= 0 ⇒ There is no distance-preserving ANDangle-preserving map to the euclidean plane
Poincare model: angle preserving, but not distance preserving(like the stereographic projection)
infinite hyperbolic plane → unit disc
infinity 7→ edge of the unit disk
geodesics (”straight lines”) 7→ circles that meet the edge ofthe unit disc at 90◦
parallel lines 7→ not intersecting such circles
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Poincare model of the hyperbolic plane
K 6= 0 ⇒ There is no distance-preserving ANDangle-preserving map to the euclidean plane
Poincare model: angle preserving, but not distance preserving(like the stereographic projection)
infinite hyperbolic plane → unit disc
infinity 7→ edge of the unit disk
geodesics (”straight lines”) 7→ circles that meet the edge ofthe unit disc at 90◦
parallel lines 7→ not intersecting such circles
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Poincare model of the hyperbolic plane
K 6= 0 ⇒ There is no distance-preserving ANDangle-preserving map to the euclidean plane
Poincare model: angle preserving, but not distance preserving(like the stereographic projection)
infinite hyperbolic plane → unit disc
infinity 7→ edge of the unit disk
geodesics (”straight lines”) 7→ circles that meet the edge ofthe unit disc at 90◦
parallel lines 7→ not intersecting such circles
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Poincare model of the hyperbolic plane
K 6= 0 ⇒ There is no distance-preserving ANDangle-preserving map to the euclidean plane
Poincare model: angle preserving, but not distance preserving(like the stereographic projection)
infinite hyperbolic plane → unit disc
infinity 7→ edge of the unit disk
geodesics (”straight lines”) 7→ circles that meet the edge ofthe unit disc at 90◦
parallel lines 7→ not intersecting such circles
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Poincare model of the hyperbolic plane
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Ideal triangles
Triangles with largest area?
A = π − Σ
largest: if Σ = 0 ( ⇔ α = β = γ = 0). How does it look like?
Figure 1 : Ideal triangles having all verices at infinity. Note thatthese triangles are congruent (having the same area A = π)!
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Ideal triangles
Triangles with largest area?
A = π − Σ
largest: if Σ = 0 ( ⇔ α = β = γ = 0). How does it look like?
Figure 1 : Ideal triangles having all verices at infinity. Note thatthese triangles are congruent (having the same area A = π)!
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Ideal triangles
Triangles with largest area?
A = π − Σ
largest: if Σ = 0 ( ⇔ α = β = γ = 0). How does it look like?
Figure 1 : Ideal triangles having all verices at infinity. Note thatthese triangles are congruent (having the same area A = π)!
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Ideal triangles
Triangles with largest area?
A = π − Σ
largest: if Σ = 0 ( ⇔ α = β = γ = 0). How does it look like?
Figure 1 : Ideal triangles having all verices at infinity. Note thatthese triangles are congruent (having the same area A = π)!
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Hyperbolic metric
Metric?
ds2 = (dx2 + dy2)
(4
(1−(x2 + y2))2
)(compare with stereographic projection!)
again: isotropic scaling ⇒ angle-preserving
Near the edge (x2 + y2 ≈ 1): ds2 >> dx2 + dy2
center → edge in the Poincare model: ∞ distance in thehyperbolic plane!∫ 1
0ds =
∫ 1
0
2
1− x2dx =∞
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Hyperbolic metric
Metric?
ds2 = (dx2 + dy2)
(4
(1−(x2 + y2))2
)(compare with stereographic projection!)
again: isotropic scaling ⇒ angle-preserving
Near the edge (x2 + y2 ≈ 1): ds2 >> dx2 + dy2
center → edge in the Poincare model: ∞ distance in thehyperbolic plane!∫ 1
0ds =
∫ 1
0
2
1− x2dx =∞
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Hyperbolic metric
Metric?
ds2 = (dx2 + dy2)
(4
(1−(x2 + y2))2
)(compare with stereographic projection!)
again: isotropic scaling ⇒ angle-preserving
Near the edge (x2 + y2 ≈ 1): ds2 >> dx2 + dy2
center → edge in the Poincare model: ∞ distance in thehyperbolic plane!∫ 1
0ds =
∫ 1
0
2
1− x2dx =∞
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Hyperbolic metric
Metric?
ds2 = (dx2 + dy2)
(4
(1−(x2 + y2))2
)(compare with stereographic projection!)
again: isotropic scaling ⇒ angle-preserving
Near the edge (x2 + y2 ≈ 1): ds2 >> dx2 + dy2
center → edge in the Poincare model: ∞ distance in thehyperbolic plane!∫ 1
0ds =
∫ 1
0
2
1− x2dx =∞
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Hyperbolic metric
Metric?
ds2 = (dx2 + dy2)
(4
(1−(x2 + y2))2
)(compare with stereographic projection!)
again: isotropic scaling ⇒ angle-preserving
Near the edge (x2 + y2 ≈ 1): ds2 >> dx2 + dy2
center → edge in the Poincare model: ∞ distance in thehyperbolic plane!∫ 1
0ds =
∫ 1
0
2
1− x2dx =∞
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Figure 2 : Hyperbolic man takes a walk to infinity
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Schlafli-symbol
Regular tiling: tiling by i) congruent regular poligons (n-gons),ii) at every vertex, m poligons meet
Euclidean plane?
Schlafli-symbol: {n,m}Any relationship between these three Schlafli-symbols?
1
n+
1
m=
1
2There is no other such n,m ∈ N ⇔ no other regulareuclidean tiling
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Schlafli-symbol
Regular tiling: tiling by i) congruent regular poligons (n-gons),ii) at every vertex, m poligons meet
Euclidean plane?
Schlafli-symbol: {n,m}Any relationship between these three Schlafli-symbols?
1
n+
1
m=
1
2There is no other such n,m ∈ N ⇔ no other regulareuclidean tiling
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Schlafli-symbol
Regular tiling: tiling by i) congruent regular poligons (n-gons),ii) at every vertex, m poligons meet
Euclidean plane?
Schlafli-symbol: {n,m}Any relationship between these three Schlafli-symbols?
1
n+
1
m=
1
2There is no other such n,m ∈ N ⇔ no other regulareuclidean tiling
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Schlafli-symbol
Regular tiling: tiling by i) congruent regular poligons (n-gons),ii) at every vertex, m poligons meet
Euclidean plane?
Schlafli-symbol: {n,m}
Any relationship between these three Schlafli-symbols?
1
n+
1
m=
1
2There is no other such n,m ∈ N ⇔ no other regulareuclidean tiling
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Schlafli-symbol
Regular tiling: tiling by i) congruent regular poligons (n-gons),ii) at every vertex, m poligons meet
Euclidean plane?
Schlafli-symbol: {n,m}Any relationship between these three Schlafli-symbols?
1
n+
1
m=
1
2There is no other such n,m ∈ N ⇔ no other regulareuclidean tiling
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Schlafli-symbol
Regular tiling: tiling by i) congruent regular poligons (n-gons),ii) at every vertex, m poligons meet
Euclidean plane?
Schlafli-symbol: {n,m}Any relationship between these three Schlafli-symbols?
1
n+
1
m=
1
2There is no other such n,m ∈ N ⇔ no other regulareuclidean tiling
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Tiling on the sphere
Regular tilings on the sphere? (assume n,m ≥ 3, i.e.nondegenrate cases)
n = 3, m = 3 ?
A blown tetrahedron!
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Tiling on the sphere
Regular tilings on the sphere? (assume n,m ≥ 3, i.e.nondegenrate cases)
n = 3, m = 3 ?
A blown tetrahedron!
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Tiling on the sphere
Regular tilings on the sphere? (assume n,m ≥ 3, i.e.nondegenrate cases)
n = 3, m = 3 ?
A blown tetrahedron!
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Tiling on the sphere
Regular tilings on the sphere? (assume n,m ≥ 3, i.e.nondegenrate cases)
n = 3, m = 3 ?
A blown tetrahedron!
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Tiling on the sphere
{4, 3} ?
A cube:
{3, 4}, {3, 5}, {5, 3} ?
Octahedron, icosahedron, dodecahedron.
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Tiling on the sphere
{4, 3} ?
A cube:
{3, 4}, {3, 5}, {5, 3} ?
Octahedron, icosahedron, dodecahedron.
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Tiling on the sphere
{4, 3} ?
A cube:
{3, 4}, {3, 5}, {5, 3} ?
Octahedron, icosahedron, dodecahedron.
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Tiling on the sphere
{4, 3} ?
A cube:
{3, 4}, {3, 5}, {5, 3} ?
Octahedron, icosahedron, dodecahedron.
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Tiling on the sphere
Any more?
No. Why?
1
n+
1
m>
1
2
only for these five!
Figure 3 : The five Platonic solids
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Tiling on the sphere
Any more?
No. Why?
1
n+
1
m>
1
2
only for these five!
Figure 3 : The five Platonic solids
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Tiling on the sphere
Any more?
No. Why?
1
n+
1
m>
1
2
only for these five!
Figure 3 : The five Platonic solids
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Hyperbolic tiling
Hyperbolic tiling?
If 1n + 1
m < 12
How many such tilings?Infinite!
Figure 4 : {3, 7} Figure 5 : {7, 3}
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Hyperbolic tiling
Hyperbolic tiling?If 1
n + 1m < 1
2
How many such tilings?Infinite!
Figure 4 : {3, 7} Figure 5 : {7, 3}
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Hyperbolic tiling
Hyperbolic tiling?If 1
n + 1m < 1
2How many such tilings?
Infinite!
Figure 4 : {3, 7} Figure 5 : {7, 3}
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Hyperbolic tiling
Hyperbolic tiling?If 1
n + 1m < 1
2How many such tilings?Infinite!
Figure 4 : {3, 7} Figure 5 : {7, 3}
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Hyperbolic tiling
Hyperbolic tiling?If 1
n + 1m < 1
2How many such tilings?Infinite!
Figure 4 : {3, 7} Figure 5 : {7, 3}
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Hyperbolic tiling
n or m can even be infinite!
Figure 6 : {3,∞},for every triangle:α = β = γ = 0 ⇒ A = π
Figure 7 : {∞, 3},”aperiogon”
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Hyperbolic tiling
n or m can even be infinite!
Figure 6 : {3,∞},for every triangle:α = β = γ = 0 ⇒ A = π
Figure 7 : {∞, 3},”aperiogon”
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Hyperbolic tiling
Or both!
Figure 8 : {∞,∞}Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Classification of regular tilings
Figure 9 : Classification of regular tilings
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Figure 10 : M.C. Escher
Figure 11 : H.S.M.Coxeter
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Figure 12 : Escher’s Circle Limit I. (1958)
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Figure 13 : Circle Limit I.: nonregular tiling of the hyperbolic plane(m = 4 and 6)
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Any regular tiling?
Figure 14 : Escher’s Circle Limit III. (1959).
Schlafli symbol?
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Any regular tiling?
Figure 14 : Escher’s Circle Limit III. (1959).
Schlafli symbol?
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Any regular tiling?
Figure 14 : Escher’s Circle Limit III. (1959).
Schlafli symbol?
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
Figure 15 : Schlafli symbol of Circle Limit III.: {8, 3}!
Daniel Czegel Tiling the hyperbolic plane - long version
Gaussian curvatureElliptic geometry
Hyperbolic geometryTiling
M.C. Escher’s work
References
Weeks, J. R. (2001). The shape of space. CRC press.
Dirnbock, H., & Stachel, H. (1997). The development of theoloid. Journal for Geometry and Graphics, 1(2), 105-118.
http://aleph0.clarku.edu/
~djoyce/poincare/poincare.html
http://en.wikipedia.org/wiki/
Uniform_tilings_in_hyperbolic_plane
http://euler.slu.edu/escher/index.php/
Math_and_the_Art_of_M._C._Escher
http://www.reed.edu/reed_magazine/march2010/
features/capturing_infinity/3.html
Thank You!
Daniel Czegel Tiling the hyperbolic plane - long version