There is a curve peculiar to hyperbolic geometry, called the
horocycle .
Consider two limiting parallel lines,
and m, with a common direction,
say
. Let P be a point on one of these lines
. If
there exists a point
such that the
trilateral ,
, has the property that

then we say that Q corresponds to P. If the
trilateral
has the above property we shall say that it is
equiangular . Note that it is obvious from
the definition that if Q
corresponds to P, then P corresponds to Q. The points P and Q are
called a pair of corresponding points .
Theorem 18.1: If points P and Q lie on two limiting parallel lines in
the direction of the ideal point,
, they are corresponding points on
these lines if and only if the perpendicular bisector of PQ is limiting
parallel to the lines in the direction of
.
Theorem 18.2: Given any two limiting parallel lines, there exists a line each of whose points is equidistant from them. The line is limiting parallel to them in their common direction.
Proof: Let
and m be limiting parallel lines with common
direction
. Let
and
. The bisector of
in the trilateral
meets
side
in a point X and the bisector of
meets side
AX of the triangle
in a point C. Thus the bisectors of the
angles of the trilateral
meet in a point
C. Drop perpendiculars from C to each of
and m, say P and Q,
respectively. By Hypothesis-Angle
(M is the midpoint of AB) and
. Thus,
. Thus, by SAS for trilaterals, we
have that

and thus the angles at C are congruent. Now, consider the line
and let F be any point on it other than C.
By SAS we have
. If S and T are
the feet of F in
and m, then we get that
and
. Thus, every point on the
line
is equidistant from
and
m.
This line is called the equidistant line.
Theorem 18.3: Given any point on one of two limiting parallel lines, there is a unique point on the other which corresponds to it.
Theorem 18.4: If three points P, Q, and R lie on three parallels in the same direction so that P and Q are corresponding points on their parallels and Q and R are corresponding points on theirs, then P, Q, and R are noncollinear.
Theorem 18.5: If three points P, Q, and R lie on three parallels in the same direction so that P and Q are corresponding points on their parallels and Q and R are corresponding points on theirs, then P and R are corresponding points on their parallels.
Consider any line
, any point
, and an ideal point in one
direction of
, say
. On each line parallel to
in the
direction
there is a unique point Q that corresponds to P. The
set consisting of P and all such points Q is called a
horocycle , or,
more precisely, the horocycle determined by
, P, and
. The
lines parallel to
in the direction
, together with
, are
called the radii of the horocycle. Since
may be denoted by
, we may regard the horocycle as determined
simply by P and
, and hence call it the horocycle through P
with direction
, or in symbols, the horocycle
.
All the points of this horocycle are mutually corresponding points by
Theorem 19.5 , so the horocycle is equally well determined by any one
of them and
. In other words if Q is any point of horocycle
other than P, then horocycle
is the same as
horocycle
. If, however, P' is any point of
other than
P, then horocycle
is different from horocycle
,
even though they have the same direction and the same radii. Such horocycles,
having the same direction and the same radii, are called codirectional
horocycles.
There are analogies between horocycles and circles. We will mention a few.
Lemma 18.1: There is a unique horocycle with a given direction which passes through a given point. (There is a unique circle with a given center which passes through a given point.)
Lemma 18.2: Two codirectional horocycles have no common point. (Two concentric circles have no common point.)
Lemma 18.3: A unique radius is associated with each point of a horocycle. (A unique radius is associated with each point of a circle.)
A tangent to a horocycle at a point on the horocycle is defined to be the line through the point which is perpendicular to the radius associated with the point.
No line can meet a horocycle in more than two points. This is a consequence of the fact that no three points of a horocycle are collinear inasmuch as it is a set of mutually corresponding points, cf. Theorem 19.4.
Theorem 18.6: The tangent at any point A of a horocycle meets
the horocycle only in A. Every other line through A except the radius meets
the horocycle in one further point B. If
is the acute angle between
this line and the radius, then d(A,B) is twice the segment which corresponds
to
as angle of parallelism.
Proof Let t be the tangent to the horocycle at A and let
be the direction of the horocycle. If t met the horocycle in another point
B, we would have a trilateral with two right angles, since
A and B are corresponding points. In fact the entire horocycle, except for
A, lies on the same side of t, namely, the side containing the ray
.
Let k be any line through A other than the tangent or radius. We need to
show that k meets the horocycle in some other point. Let
be the
acute angle between k and the ray
. Let C be the point of
k, on the side of t containing the horocycle, such that AC is a segment
corresponding to
as angle of parallelism. (RECALL:
). The line perpendicular to k at C is then parallel to
in the direction
. Let B be the point
of k such that C is the midpoint of AB. The trilaterals
and
are congruent. Hence
, B corresponds to A, and
.
A chord of a horocycle is a segment joining two points of the horocycle.
Theorem 18.7: The line which bisects a chord of a horocycle at right angles is a radius of the horocycle.

Figure 19.1: A horocycle in the Poincaré model
We can visualize a horocycle in the Poincaré model as follows. Let
be the diameter of the euclidean circle
whose interior
represents the hyperbolic plane, and let O be the center of
. It is
a fact that the hyperbolic circle with hyperbolic center P is
represented by a euclidean circle whose euclidean center R lies between P
and A.
As P recedes from A towards the ideal point
, R is pulled up to
the euclidean midpoint of
, so that the horocycle
is a
euclidean circle tangent to
at
and tangent to
at A.
It can be shown that all horocycles are represented in the Poincaré
model by euclidean circles inside
and tangent to
. Moreover,
all the Poincaré lines passing through the ideal point
are
orthogonal to the horocycle
.
In the Poincaré model two horocycles tangent to
at
are
said to be concentric. It is by studying the ratio of corresponding arcs
on concentric horocycles that the Bolyai-Lobachevsky formula can be derived.

Figure 19.2: Concentric horocycles in the Poincaré model
Another curve found specifically in the hyperbolic plane and nowhere else is
the equidistant curve, or
hypercycle.
Given a line
and a
point P not on
, consider the set of all points Q on one side of
so that the perpendicular distance from Q to
is the same as the
perpendicular distance from P to
.
The line
is called the axis, or
base line, and the common
length of the perpendicular segments is called the distance. The
perpendicular segments defining the hypercycle are called its radii.
The following statements about hypercycles are analogous to statements about
regular euclidean circles.
In the Poincaré model let P and Q be the ideal end points of
. It
can be shown that the hypercycle to
through P is represented
by the arc of the euclidean circle passing through A, B, and P. This
curve is orthogonal to all Poincaré lines perpendicular to the line
.

Figure 19.3: A hypercycle in the Poincaré model
In the Poincaré model a euclidean circle represents:
;
except for one point
where it is tangent to
;
nonorthogonally in two
points;
orthogonally.
It follows that in the hyperbolic plane three non-collinear points lie either on a circle, a horocycle, or a hypercycle accordingly, as the perpendicular bisectors of the triangle are concurrent in an ordinary point, an ideal point, or an ultraideal point.