Publ. Math.

Debrecen
51 / 1-2 (1997), 13–20

A polar-coordinate model of the hyperbolic space

By L. NE METH (Sopron)

Abstract. This article presents a d-dimensional polar-coordinate model, a new

model of the d-dimensional hyperbolic space. Since we can establish a bijection between
this model and the Weierstrass model with a proper projection, we can give a simple
discussion of the geometric configurations, which is the even simpler in case d = 2 than
in [Wiegand, 1992].

1. Introduction

The different models of the d-dimensional (d > 2) hyperbolic space

(the Cayley-Klein model, the Klein-Poincaré model, Poincaré’s hemisphere
model, the Weierstrass model, etc.) can be found in several books ([Bolyai,
1987], [Faber, 1983], [Liebmann, 1992], [Rozenfeld, 1969], [Szász,
1973]), but a complete survey is difficult today. We can find a polar-
coordinate model of the hyperbolic plane in the article of [Wiegand,
1992]. As a generalisation we give a polar-coordinate (P d ) model of the
d-dimensional hyperbolic space. The examination of the geometric config-
urations becomes more simple even in the case d = 2 when we establish
a bijection between the Weierstrass model (W d ) and our polar-coordinate
model by a projection [Németh]. We deal with the reflection in a hyper-
plane, as the generating element of the congruence transformations.

2. The Weierstrass and the polar-coordinate

models of the hyperbolic space

Let us consider the sheet of a hyperboloid H d given by the equation

x21 +x22 +. . .+x2d −x2d+1 = −1, xd+1 > 0 in the (d+1)-dimensional Euclidean
Mathematics Subject Classification: 51M09.
14 L. Németh

space. The Weierstrass model (W d ) of the d-dimensional hyperbolic space

can be built on this surface. Let the points of the hyperbolic space be the
points of the surface H d , and each k-dimensional (k < d) plane corresponds
to a (k + 1)-dimensional Euclidean space section of the surface, which
contains the point O(0, 0, . . . , 0). Thus the lines are intersections of H d
with 2-dimensional planes containing the point O, namely branches of
hyperbolas. Incidence and order are Euclidean. The definitions of distance
and angle are the same as in the case d = 2 ([Faber, 1983], [Rozenfeld,
1969]). Let Ad be the half asymptotic cone of the surface H d . Two planes
with dimensions s and r are called parallel if their intersecting (s + 1) and
(r + 1)-dimensional Euclidean planes have one and only one common half
generator of Ad .
We can define a polar-coordinate model (P d ) of the d-dimensional
hyperbolic space on the base of a polar-coordinate model of the hyper-
bolic plane ([Wiegand, 1992]) the following way. We denote by Hd the
hyperbolic space and by Ed the Euclidean space. Let us consider a rect-
angular coordinate system (O h , xh1 , xh2 , . . . , xhd ) in Hd and assign the rect-
angular coordinate system (O e , xe1 , xe2 , . . . , xed ) in Ed to the previous coor-
dinate system. Let us assign the point P e ∈ Ed with polar coordinates
(sinh p, ϕ1 , . . . , ϕd−1 ) to the point P h ∈ Hd (P h 6= O h ) with polar co-
ordinates (p, ϕ1 , . . . , ϕd−1 ). The point O e corresponds to the point O h .
This assignment is a bijection between the spaces Hd and Ed because of
the invertibility of the function hyperbolic sine. Hence we can model the
d-dimensional hyperbolic space in the space Ed .

3. The discussion of the d-dimensional polar-coordinate model

By the help of the Weierstrass model and of an orthogonal projection

we are able to describe our d-dimensional polar-coordinate model in a
simple way (in the case d = 2 see [Németh]).
3.1. Points, lines, planes and hyperplanes
Lemma 1. Let the hyperbolic distance between the points
K(0, . . . , 0, 1) ∈ Ed+1 and P ∈ W d be ρ and the orthogonal projections of
the points K and P onto the x1 x2 . . . xd -hyperplane be O(0, 0, . . . , 0) and
P 0 , respectively. Then the Euclidean distance between O and P 0 is equal
to sinh ρ.
Proof. As the surface Hd has rotational symmetry (in the Euclidean
sense) about the xd+1 -axis, we can rotate it so, that the points P and P 0
 A polar-coordinate model of the hyperbolic space 15

be in the coordinate³ plane xd xd+1 (Fig. 1).´ Then the orthogonal projec-
p
tion of the point P 0, . . . , 0, pd , pd + 1 is the point P 0 (0, . . . , 0, pd , 0).
2

Figure 1

The Euclidean distance between O and P 0 is |pd |, the hyperbolic dis-

tance ρ between
p the points K and P is given by the expression
p cosh ρ =
−hK, P i = p2d + 1 [Faber, 1983], from which ρ = arcosh p2d + 1.
Using cosh2 x − sinh2 x = 1, we obtain that
µ q ¶ s µ q ¶
2 2 2
sinh ρ = sinh arcosh pd + 1 = cosh arcosh pd + 1 − 1 = |pd |.
¤
Theorem 1. The orthogonal projection of the model W d onto the co-
ordinate plane x1 x2 . . . xd is the model P d of the d-dimensional hyperbolic
space.
Proof. Take the rectangular coordinate system in the model W d ,
whose orthogonal projection is the rectangular coordinate system
(O, x1 , x2 , . . . , xd ) in the x1 x2 . . . xd -hyperplane. The centres (K and O)
and the axes of the coordinate systems correspond to each other. The
angle of the lines passing through the point K in the model W d is equal
to the Euclidean angle of the tangent vectors at the point, and is equal
to the Euclidean angle of the projections of the tangent vectors, since the
tangent vectors are parallel to the x1 x2 . . . xd -hyperplane. Then the hy-
perbolic angle of the lines passing through the point K is equal to the
16 L. Németh

Euclidean angle of their projections (these are lines passing through the
point O in the x1 x2 . . . xd -hyperplane). Using Lemma 1 a point P 0 with
Euclidean polar coordinates (sinh p, ϕ1 , . . . , ϕd−1 ) is assigned to a point
P ∈ W d with polar coordinates (p, ϕ1 , . . . , ϕd−1 ). This assignment is a
bijection. ¤
On the base of Theorem 1, we define our polar-coordinate model as
the orthogonal projection of the Weierstrass model. This definition and
the previous definition (see 1.) are equivalent. So the points of P d are the
points of the x1 x2 . . . xd -hyperplane.
Definition. We call the imaginary throat-sphere of a two-sheeted hy-
perboloid of revolution Dd the sphere, which is the intersection of Dd
and the hyperplane that is orthogonal to the axis of revolution of Dd and
contains the center of Dd .
Theorem 2. The k-dimensional planes in our polar-coordinate model
are the Euclidean k-dimensional planes passing through the point O, and
the sheets of those two-sheeted hyperboloids of revolution whose throat-
sphere’s radius is the imaginary unit and the vertices of their asymptotic
cones fall into the point O.
Proof. The k-dimensional planes of the model W d are the plane
sections of the surface H d with (k + 1)-dimensional planes containing the
point O. Then the vertices of the asymptotic cones of the plane sec-
tions are the point O. If the (k + 1)-dimensional (Euclidean) subspace
belonging to the k-dimensional (hyperbolic) plane L ∈ W d is perpendic-
ular to the x1 x2 . . . xd -hyperplane, then the orthogonal projection L0 is a
k-dimensional plane containing the point O in the x1 x2 . . . xd -hyperplane.
If the subspace belonging to L is not perpendicular to the x1 x2 . . . xd -
hyperplane, then the orthogonal projection L0 is a k-dimensional sheet of
a hyperboloid. Since the orthogonal projection preserves incidence, the or-
thogonal projection of the half asymptotic cone of L is the half asymptotic
cone of L0 , and its vertex is the point O. The common points of H d and
the x1 x2 . . . xd -plane satisfy the equation x21 + x22 + . . . + x2d = −1, it is the
throat-sphere of H d . The subspace belonging to L intersects this throat-
sphere in an imaginary-sphere (throat-sphere of the surface L). This sphere
lies in the x1 x2 . . . xd -hyperplane, therefore its orthogonal projection is it-
self and because of the preservation of incidence it is the throat-sphere of
 A polar-coordinate model of the hyperbolic space 17

L0 as well. This throat-sphere is really a sphere, then L0 is a hyperboloid of

revolution. The imaginary radius of the throat-sphere of L0 is equal to 1.
If L0 is a branch of a hyperbola in the Euclidean sense, then its conjugate
axis corresponds to its throat-sphere. ¤
If the hyperbolic distance between a hyperplane perpendicular to the
x1 -axis and the point O is equal to t (t 6= 0) and the hyperplane is in the
x1 > 0 halfspace, then the equation of the hyperplane in the model (from
Lemma 1 and Theorem 2) is:

x21
− + x22 + . . . + x2d = −1, x1 > 0.
sinh2 t
Corollary. In the Euclidean sense the lines in our polar-coordinate
model are the lines passing through the point O and the branches of hy-
perbolas with unit semiconjugate axes whose asymptotes pass through the
point O.
3.2. Incidence, order and parallelism
Since incidence and order are Euclidean in the model W d and the
orthogonal projection preserves incidence, furthermore, it preserves order
for the corresponding branches of hyperbolas, the incidence and order are
Euclidean in the model P d too.
In the model W d two elements of the space are parallel if and only
if there exists a common half-generator of their asymptotic cones. Then
parallelism is the following in the model P d :
Definition. Two planes (they may have different dimensions) are
called parallel in the model P d if their half-asymptotic cones have one
and only one common half-generator. In this case we consider the half-
asymptotes of a line (as a branch of a hyperbola) to be the half-asymptotic
cone of the line, and the plane (or line) containing the point O itself is
considered to be the asymptotic cone of the plane (or line).
3.3. Reflection in a hyperplane
Definition. A reflection in a hyperplane L in the model W d is the
affinity whose fixed plane is the d-dimensional Euclidean hyperplane that
intersects the surface H d in the hyperbolic hyperplane L. The direction
of the affinity is the conjugate direction of this fixed plane with respect to
the surface H d and its ratio is equal to −1.
We can prove that this affinity assigns the surface H d to itself and the
above definition satisfies the properties of the reflection in a hyperplane.
18 L. Németh

Definition. The reflection in a hyperplane in the model P d is the

orthogonal projection of the reflection in a hyperplane in the model W d .
3.4. Sphere, parasurface and equidistant surface
The (d − 1)-dimensional spheres of radius r in the model W d are the
hyperplane sections of the surface H d with hyperplanes not containing the
point O and intersecting each element of Ad (in the case d = 2 see [Faber,
1983]). The centre of a sphere is the common point of H d and of the
tangent plane of H d that is parallel to the sphere’s intersecting hyperplane.
These surfaces in the Euclidean sense are (d − 1)-dimensional spheres or
ellipsoids. The orthogonal projections of these surfaces in the x1 x2 . . . xd -
hyperplane are the (d − 1)-dimensional spheres of P d , as the assignment
between the two models is a bijection, preserves incidence and the centres
of the corresponding spheres correspond to each other. The spheres of P d
in the Euclidean sense are the spheres with the centre O, or ellipsoids of
revolution whose line of major axis passes through the point O. These
surfaces are surfaces of revolution because the two-dimensional planes in
the model P d containing the point O (that are two-dimensional Euclidean
planes as well) and the centre of the sphere intersects this ellipsoid in
congruent ellipses. Their line of major axis passes through the point O
because the circles are determined by the same center and radius.
If the coordinates of the centre of a d-dimensional sphere of radius r
are C(ρ, 0, . . . , 0) in the hyperbolic space, then its equation is (based on
the case d = 2, when we got the equations of the circles in P 2 from the
equations of the circles in W 2 [Németh]):

(x1 − sinh ρ · cosh r)2 x22 x2d

+ + . . . + = 1,
cosh2 ρ · sinh2 r sinh2 r sinh2 r
if ρ = 0, then:

x21 + x22 + . . . + x2d = sinh2 r.

The (d − 1)-dimensional parasurfaces in the model W d are the hyper-

plane sections of H d which are (d−1)-dimensional paraboloids (in the case
d = 2 see [Faber, 1983]). Their orthogonal projections in the x1 x2 . . . xd -
hyperplane are also paraboloids, that are the (d − 1)-dimensional parasur-
faces in the model P d because of the properties of the assignment between
the two models. Since the axes of the surfaces contain the point O and
 A polar-coordinate model of the hyperbolic space 19

each hyperplane section of these surfaces containing their axes is congru-

ent, therefore these surfaces are paraboloids of revolution (that are the
parasurfaces), whose axes contain the point O. If the axis is the x1 -axis,
its equation has the form (based on the case d = 2):
1
· (x22 + . . . + x2d ) + sinh t = x1 .
2(cosh t − sinh t)

The equidistant surfaces of a hyperplane L in the model W d are the

hyperplane sections of H d which are sheets of hyperboloids, and its hy-
perplanes are parallel to the hyperplane that intersects H d in the given
plane L (in case d = 2 see [Faber, 1983]). Their orthogonal projections in
the x1 x2 . . . xd -hyperplane are (d − 1)-dimensional Euclidean planes (if L0
contains the point O), or sheets of hyperboloids of revolution, which will
be the equidistant surfaces in the model P d too, because of the properties
of the assignment between the two models. If the plane L0 contains the
point O, its equidistant surfaces are planes parallel to L0 in the Euclidean
sense. If L0 does not contain the point O, then the axes of revolution of the
equidistant surfaces coincide with the axis of revolution of the hyperboloid
L0 and their asymptotic cones will be parallel to each other.
If the axis of the equidistant surface is the x1 -axis, the parameter of
the plane is t, and the distance is ρ, then its equation has the form (based
on the case d = 2):

(x1 ± sinh t · sinh ρ)2 x22 x2d

− + + . . . + = −1.
sinh2 t · cosh2 ρ cosh2 ρ cosh2 ρ

The intersections of (d−1)-dimensional cycles and (k +1)-dimensional

(hyperbolic) planes in the model are k-dimensional cycles.

References
[1] E. Benedetti and C. Petronio, Lectures on Hyperbolic Geometry, Springer –
Verlag, Berlin, 1992.
[2] J. Bolyai, Appendix the Theory of Space (with introduction, comment, and ad-
denda edited by F. Kárteszi), Akadémiai Kiadó, Budapest, 1987.
[3] K. Borsuk and W. Szmielew, Foundations of Geometry, North-Holland Publish-
ing Company, Amsterdam, 1960.
[4] J. Bohm and E. Hertel, Polyedergeometrie in n-dimensionalen Räumen konstan-
ter Krümmung, Birkhäuser Verlag, Basel–Boston–Stuttgart, 1981.
20 L. Németh : A polar-coordinate model of the hyperbolic space

[5] R. L. Faber, Foundations of Euclidean and Non-Euclidean Geometry, Marcel

Dekker Inc., New York, 1983.
[6] H. Liebmann, Nichteuklidische Geometrie, Berlin und Leipzig, 1992.
[7] L. Nemeth, Connections of the different models of the hyperbolic plane, BDTF.
Tud. Közl., (to appear ).
[8] I. Reiman, Geometry and its Boundaries, Gondolat, Budapest, 1986. (in Hungar-
ian)
[9] B. A. Rozenfeld, Non-Euclidean Spaces, Izdatelstvo Nauka, Moszkva, 1969. (in
Russian)
[10] P. Szasz, An Introduction to Bolyai-Lobatchevski Geometry, Akadémiai Kiadó,
Budapest, 1973.
[11] T. Wiegand, A polar-coordinate model of the hyperbolic plane, Publ. Math.
Debrecen (1992), 161–171.

LÁSZLÓ NÉMETH
MATH. INST.
UNIVERSITY OF SOPRON
SOPRON

E-mail : lnemeth@efe.hu

(Received September 15, 1995, revised January 31, 1997)