Geometry Lecture Note

AAU
2005
Geometry
Instructor Biadgelign
M.T
W W W. A A U . E D U . E T

Introduction to Geometry
Fundamental Concepts of Geometry
HU Department of Mathematics 1

CHAPTER I
ABSOLUTE GEOMETRY

Out of nothing I have created a strange new universe. -Jnos Bolyai

1.1 Introduction: The elements of geometry and the five groups of axioms
Geometry, like arithmetic, requires for its logical development only a small number of simple,
fundamental principles. These fundamental principles are called the axioms of geometry. The
choice of the axioms and the investigation of their relations to one another is a problem
which, since the time of Euclid, has been discussed in numerous excellent memoirs to be
found in the mathematical literature.
Geometry is the science of shape, size and symmetry. While arithmetic dealt with numerical
structures, geometry deals with metric structures. Geometry is one of the oldest mathematical
disciplines and early geometry has relations with arithmetic. Geometry was also a place,
where the axiomatic method was brought to mathematics: theorems are proved from a few
statements which are called axioms.

Absolute Geometry is Geometry which depends only on the first four of Euclid's postulates
and not on the parallel postulate. It is sometimes referred to as neutral geometry, as it is
neutral with respect to the parallel postulate.

Let us consider three distinct systems of things. The things composing the first system, well
will call points and designate them by the letter A, B, C, those of the second, we will call
straight lines and designate them by the letters a, b, c . . . and those of the third system and we
will call planes and designate them by the Greek letters . , , r | o The points are called the
elements of linear geometry; the points and straight lines, the elements of plane geometry; and
the points, lines, and planes, the elements of the geometry of space or the elements of space.


We think of these points, straight lines, and planes as having certain mutual relations, which
we indicate by means of such words as are situated, between, parallel congruent,
continuous, etc. The complete and exact description of these relations follows as a
consequence of the axioms of geometry. These axioms may be arranged in five groups. Each
of these groups expresses, by itself, certain related fundamental facts of our intuition.

I. Axioms of incidence (connection)
II. Axioms of betweenness (order)
III. Axiom of parallels (Euclids axiom).
IV. Axioms of congruence.
V. Axiom of continuity

Although point, line and plane etc. do not have formal definition, we can describe them
intuitively as follows.
Point: We represent points by dots (see Fig.1.1.1) and designate them by capital letters.
Line: We represent line by indefinitely thin and long mark (as in fig. 1.1.2). Lines are
designated by small letters. We regard line as a set of points that can be extended as
far as desired in either direction.
Plane: We think of a plane as a flat surface that has no depth (or thickness). We designate
planes by Greek letters ... , , , t | o and represent it by some appropriate figure in
space (see fig.1.1.3)

Activity: Based on the intuitive descriptions given above for point, line and plane, give
your own examples that describe each of these undefined terms
Fig. 1.1.1
Fig. 1.1.2
Fig. 1.1.3

In this subsection the axioms completely and exactly describe the properties or characteristics
that the undefined elements should possess. They also state the relationships that hold among
the undefined terms and the existence of some of these elements. We shall see them one by
one.

1.2 Axioms of Incidence
The axioms of incidence determine the properties of mutual disposition of points, lines and
planes by the term incident. Statements such as a point is incident with a line, a point
lies on a line, a line passes through a point and a line contains a point are assumed to be
equivalent. Thus we can use them interchangeably.
So if a point is incident with two lines then we say that they intersect at the point or the point
is their common point. Analogues statements will be used for a point and a plane, and for a
line and a plane.

Convention: When numbers like two, three, and four and so on are used in any
statement of this material, they will describe distinct objects. For instance by two planes,
three lines, four points we mean two distinct planes, three distinct lines and four
distinct points, respectively. But by lines m and n, we mean m and n may represent different
or the same line. The same is true for points and planes.

The group of the axioms of incidence includes the following:
AI
1
: If A and B are two points, then there is one and only one line that passes through
them.

Fig. 1.2.1
1.2.1 At least two points on any line (Exactly one line through two points)

This axiom asserts the existence and uniqueness of a line passing through any two given
distinct points A and B. Here can be described as a line determined by the two points A and
B. We denote the line passing through A and B by AB.
AI
2
: For any three points A, B, C that do not lie on the same line there exists exactly one
plane o that contains each of the points A,B,C. For every plane there exists a point
which it contains.
It follows from AI
2
that any three given distinct points not all on the same line determine a
plane passing through the three points and there is no other plane different from this
containing all the three given points.

AI
3
: If two points A and B of a line mlie in a plane o then every point of mlies in the
plane o.
AI
4
: If two planes have a point in common then they have at least one more point in
common.
AI
5
: a) There exists at least two points which lie on a given line.
b) There exist at least three points which do not lie on a line.

Fig. 1.2.2 At least three non-collinear points

c) There exist at least four points which do not lie in a plane.

Now let us see some of the immediate consequences of this group of axioms.
Theorem 1.2.1 If m
1
and m
2
are two lines then they have at most one point in
common.


Fig 1.2.3
Proof: Suppose
1
m and
2
m

have two points in common. Let these points be A and B. Thus
both
1
m and
2
m passes through A and B. But this is impossible by AI
1
. Hence they cannot
have two or more points in common. Therefore they have at most one point in common.

Remark: From the above theorem it follows that two distinct lines either intersect only at one
point or do not intersect.

Fig 1.2.4

Theorem 1.2.2 Two planes meet in a line or they do not meet at all.
Proof: Suppose two planes t
1
and t
2
have a point P in common. Then t
1
and t
2
have one
more point Q in common by AI
4
. Thus P and Q lie on t
1
and also P and Q lie on t
2
. But P and
Q determine a unique line, sayh , by AI
1
. So line h lies on both t
1
and t
2
(Why?). That is
every point on line h is common to both t
1
and t
2
. Further they cannot have any other point
not on h in common. (What will happen if they have such a point in common?). Therefore
they meet in a line if they have a point in common otherwise they do not meet.

Activity: Prove that if o is a plane then it contains at least three points that are not on the
same line.

or

Definition 1.2.1 Three or more points are called collinear iff they lie on the
same line.

Definition 1.2.2
a) Points that lie in the same plane are called Coplanar points.
b) Lines that lie on the same plane are called Coplanar lines.

Notation: If P, Q and R are three non collinear points on a plan t then we denote t as PQR.

Theorem 1.2.3: Two intersecting lines determine one and only one plane.
Proof: Left for students as an exercise.

Activity : Answer the following questions and then give a formal proof of Theorem 1.2.3 on
your own.
i) How many lines are given?
ii) Are they assumed to be intersecting?
iii) What do we need to show?
iv) How many points do two intersecting lines have in common? Why?
v) Is there a point on each of these lines different from their common point? Why?
vi) How many points do we need to determine a unique plane?

Example 1.2.1 a) Prove that a line and a point not lying on it determine one
and only one plane.
b) Show that there are at least four planes.
Solution:
a) Given a line and a point P not lying on . We need to show that there is one and only
one plane containing P and . By AI
5
(a), contains at least two points say X and Y.

Points X, Y and P are not collinear (Why?). Thus they determine a unique plane o
containing them . Moreover

lies completely ono (Why?).
So, we can conclude that P and determine plane o uniquely.
b) By AI
5
(c), there are at least four points that are not coplanar. Let us designate these points
by X, Y, Z and W. Any three of them are not collinear (why?). Thus we have four planes
namely X Y Z, X Y W, X Z W and YZW. This completes the proof.

Activity: Prove that a plane and a line not lying on it cannot have more than one point in
common.
1. Show that there are at least 6 lines.
2. Develop model for the system described by the axioms of incidence.
(An interpretation satisfying all the five axioms)

1.3 Distance Functions and the Ruler Postulate
For most common day-to-day measurements of length, we use rulers, meter sticks, or tape
measures. The distance and ruler postulates formalize our basic assumptions of these items
into a general geometric axiomatic system. The Ruler Postulate defines
a correspondence between the points on a line markings
on a meter stick) and the real numbers (units of
measurement) in such a manner that the absolute value
of the difference between the real numbers is equal to
the distance between the points (measurement of the
length of an object by a meter stick matches our usual
Euclidean distance).
The Ruler Placement Postulate basically says that it does not matter how we place a meter
stick to measure the distance between two points; that is, the origin (end of the meter stick)
does not need to be at one of the two given points.

The Ruler Postulate:
The points of a line can be placed in a correspondence with the real numbers such that:


i. To every point of the line there corresponds exactly one real number.
ii. To every real number there corresponds exactly one point of the line.
iii. The distance between two distinct points is the absolute value of the difference of the
corresponding real numbers.
Note that the first and second conditions of the Ruler Postulate imply that there exist a one-
to-one and onto function. As a reminder, we write the definitions for one-to-one and onto
functions.

Definition 1.3.1 A function f from A to B is onto B if for any b in B there is at least one a in A
such that f (a) = b.

Definition 1.3.2 A function f from A to B is one-to-one (1-1) if f(x) = f(y) then x = y for any x
and y in A. (Note that the contrapositive of this definition can be used in writing proofs.)

Definition 1.3.3 The line segment determined by A and B, denoted by , AB is the set of all
points P such that P is between A and B and the end points A and B.. In other words the (line)
segment (joining A and B) is { } { } B and A between is P P B A AB | , =

Definition 1.3.4 The length of segments ABdenoted by AB is the distance from A to B.
We
call the points A and B endpoints of
AB
.

Definition 1.3.5 For two segments ABand BC
,
BC AB BC AB ~ =
.
( ) . BC to congruent is AB e i

Axiom 1.3.1 (Ruler Postulate): for every pair of points P, Q there is a number PQ called the
distance from P to Q. For each line there is a one-to one mapping 9 : f such that if
) (P f x= and ) (Q f y = then y x PQ = is the value of the distance.

Definition 1.3.6 Let be the collection of all points. A distance (coordinate) function d is
mapping from into 9satisfying the following conditions:
1. The mapping d is a function, i.e each pair of points in assigned one and only one non
negative real number.
2. ( ) , ) , ( ) , ( e = Q P all for P Q d Q P d
3. Q P if only and if Q P d = =0 ) , (

Theorem 1.3.1 (Triangle Inequality) ) , ( ) , ( ) , ( Q R d R P d Q P d + s

Existence postulate 1.3.1 The collection of all points form a non-empty set with more than
(i.e at least two ) points.

Lemma 1.3.1 Given any two points Q P, e there exists a line containing both P and Q.
Proof. We have two cases (either ). Q P or Q P = =
Q P If = then there is exactly on line PQ = such that Q and P both line on (incidence
postulate).
If Q P= then by the existence postulate there must be a second point P R = and by incidence
postulate there is a unique line Q = P Since R. and P both contains that

then . e Q
Hence there is a line that contains both P and Q .

Definition 1.3.7 A metric is a function 9 : d (where is the set of all points)

that
satisfies:
1. ( ) , ) , ( ) , ( e = Q P all for P Q d Q P d
2. ( ) , 0 ) , ( e > Q P all for Q P d
3. Q P Q P d = =0 ) , (


Theorem 1. 3.2: Distance is a metric.
Proof: Let P and Q be points. Then we need to show that each of the following hold:
(a) QP PQ=
(b) 0 > PQ
(c) Q P PQ = =0

By lemma 1.3.1 there is line that contains both P and Q. By the ruler postulate, there is a
one to one function . : 9 f Let ) (P f x= and ) (Q f y = such that the distance is given
by y x Q f P f PQ = = ) ( ) (
To see (a),
QP y x y x PQ = = =
To see (b).
0 > = y x PQ
To see (c), first suppose that . 0 = PQ Then
y x PQ = = 0
y x =
Q P=
Where the last step follows because f is one to- one. To verify the converse of (c) suppose that
. Q P = Then y Q f P f x = = = ) ( ) ( so that 0 = = y x PQ which verifies the converse of(c).

Example 1.3.1 Let ). , ( ), , (
2 2 1 1
y x Q y x P = = Then
2
1 2
2
1 2
) ( ) ( ) , ( y y x x Q P d + =
To show that this is a metric, calculate
) ) ( ) ( ) , (
2
1 2
2
1 2
y y x x Q P d + = ) ) ( ) (
2
2 1
2
2 1
y y x x + = ) , ( P Q d =

This verifies property (1).

To get property (2), observe that the value of the square root is a non-negative number, hence
the square root is defined and positive or zero.
For property (3), first assume . Q P=

Then

) , ( ) , ( P P d Q P d =
2
1 1
2
1 1
) ( ) ( y y x x + =
.
This shows that , 0 ) , ( ) ( = = Q P d implies Q P
To get the converse, assume that ) , ( Q P d =0. Then

2
2 1
2
2 1
) ( ) ( 0 y y x x + =

2
2 1
2
2 1
) ( ) ( 0 y y x x + =
If either 0 0
2 1 2 1
= = y y or x x then the right hand side of equation 1.1 is non-zero. Hence
,
2 1 2 1
y y and x x = = which means . Q P= Thus . 0 ) , ( Q P implies Q P d = =

Example 1.3.2: In the Cartesian Plane any (non-vertical) line can be described by some
equation b mx y + = and any vertical line by a x = . Show that an arbitrary line in the
Euclidean plane satisfies the Ruler Postulate.
Solution:
2
1 ) , ( Let m x y x f + = if is non-vertical, and set y y a f y x f = ) , ( ) , ( if is vertical.
To see that f is a distance ( coordinate) function and that this works, we need to consider each
case (vertical and non-vertical) separately and to show that f is 1 -1, onto, and satisfies
) ( ) (( Q f P f PQ
In each case, suppose first that is non-vertical, and define f.

(a) To show that f is one-to-one, let ) , ( , ) , (
2 2 1 1
y x Q y x P = = and suppose that . ) ( ) ( Q f P f =
Then Since m x m x . 1 1
2
2
2
1
+ = +

0 1
2
= +m it can be cancelled out, giving .
2 1
x x = Thus

2
2
1 1
y
b mx
b mx y
=
+ =
+ =


Hence f is one-to-one

)). ( ) ( ( Q f P f Q P = =

(b) To show that f is onto, pick and 9 e z and define
2
1 m
z
x
+
= and b mx y + = .
Then e ) , ( y x P and z m x y x f P f = + = =
2
1 ) , ( ) ( . Thus f is onto.
(c) To verify the distance formula, let . ) , ( ) , (
2 2 1 1
l y x Q and y x P e = =

Then b mx y + =
1 1

b mx y + =
2 2
.
Hence

) , ( Q P d PQ=

2
1 2
2
1 2
) ( ) ( y y x x + =

2
1 2
2
1 2
) ( ) ( b mx b mx x x + + =

2
2
2
1 2
)) ( ( ) ( x x m x x + =

1 2
2
1 x x m + =

1
2
2
1 1 m x m x + + =

1 2
y y =

) ( ) ( Q f P f =
Thus if is not a vertical line, f is a coordinate function.

Now suppose that is a vertical line with equation a x = and
define 9 : f by . ) , ( y y a f =
(a) To show that f is one-to-one, e = ) , ( , ) , (
2 1
y a Q y a P Let where , Q P= hence
2 1
y y = and
) ( ) (
2 1
Q f y y P f = = = . Which shows that f is one-to-one ( ) ) ( ) ( Q f P f Q P = =
(b) To show that f is onto let 9 e y be any number. Then . ) ( ) , ( y P f and y a P = e = Hence
is onto.
(c) To verify the distance formula, let ). , ( ) , (
2 1
y a Q and y a P = = Then

2
1 2
2
) ( ) ( y y a a PQ + =

=
1 2
y y ) ( ) ( P f Q f =
The following theorem tells us that we can place the origin of the ruler at any place we want,
and orient the ruler in any direction we want.

Theorem 1.3.3 (Ruler Placement Postulate) For every pair of distinct points P, Q there is a
coordinate function. 9 PQ f : such that 0 ) ( = P f and . 0 ) ( > Q f

Lemma 1.3.2 Let 9 ; f be a coordinate function for and let . 9 e c Then 9 : g
given by c P f P g + = ) ( ) ( is also a coordinate function for .
Proof We need to show three things: g is one to one (1 1), onto , and . ) ( ) ( P g Q g PQ =
(a) Suppose g(P) = g(Q).Then c Q f c P f + = + ) ( ) ( . So ). ( ) ( Q f P f =

Since f is one to one,
P=Q. Thus g so Q P Q g P g = = )) ( ) ( is one to one.
(b) Let . 9 e x Since f is onto, e P exist there such that c x P f = ) (
.
So that
x c P f P g = + = ) ( ) (
.
Hence e 9 e P exist there x all for , such that x P g = ) ( . Thus g is
onto.

Activity: Verify the distance formula for the above lemma, i.e., show that PQ= ) ( ) ( Q g P g

Lemma 1.3.3 Let 9 : f be a coordinate function. Then ) ( ) ( x f x g = is a coordinate
function.
Proof : We need to show three things that g is one to one, onto, and ) ( ) ( P g Q f PQ = .
(a)
Let ). ( ) ( P f P g = Suppose that ). ( ) ( Q g P g = Then ) ( ) ( Q f P f = hence . Q P=

Hence g is one to one.

(b) Let 9 e x . Since f is onto, there is some point l Pe such that x P f = ) ( . Hence
there exist l Pe such that x P g = ) ( there is some l Pe such that x P g = ) ( . Hence
g is onto.
The last property is left as an activity. Thus g is a coordinate function.

Activity:
1. In lemma 1.3.3 show the last property.
2. Show that the Euclidean distance function d

satisfies the triangle inequality.

Now let us prove theorem 1.3.3
Proof.(theorem 1.3.3) Pick any two distinct points Q P= . By the incidence postulate there
is a line PQ = . By the Ruler postulate there exists a coordinate function 9 : g . Define
). (P g c =

And define 9 : h by c x g x h + = ) ( ) ( . Then h is a coordinate function by
Lemma 1.3.1. Since , 0 ) ( = P h it is must be the case that 0 ) ( = Q h because h is one to one.
We have two cases to consider. 0 ) ( 0 ) ( < > Q h or Q h . If 0 ) ( > Q h then set ) ( ) ( P h P g = and
the theorem is proven. If 0 ) ( < Q h define R g : by ), ( ) ( R h R g = which is a coordinate
function by lemma 1.3.2. Since 0 ) ( ) ( = = P h P g and ( ) ( h Q g = Q)>0, we see that g has
the desired properties. Fig. 1.3.1 Circles that intersect in the real plane do not necessarily
intersect in the rational plane.

The following examples illustrate why rulers (hence distances) require real numbers and not
rational numbers.

Example 1.3.3 The distance between the points (1, 0) and (0, 1) in . 2 is


Example 1.3.4 Find the intersection of the line x y = and the unit circle using whatever
knowledge you may already have of circles and triangles.

Example 1.3.5 Let ). 0 , 2 ( , ) 0 , 0 ( = = Q P The circles of radius 2 centered at Q and P do not
intersect in Q
2
(Q is set of rational numbers). Their intersection in ) 3 , 1 (
2
9 is (figure
1.3.1)

Remark: The Euclidean distance on
2
9 where

9is set of real numbers is given by
d(P,Q) =
2
1 2
2
1 2
) ( ) ( y y x x + for P = (x
1
, y
1
) and Q = (x
2
, y
2
). for x
1
, y
1
, x
2
, y
2
real
numbers.

Activity: Let L be a vertical line L
a
. Then P L
a
implies that P = (a, y) for some y. Define
the standard ruler f: L
a
R by f(a, y) = y. Let L is the line L
m, b
. Then P L
m, b
with P = (x, y)
implies that y = mx + b. Define the standard ruler f: L
m,b
R by f(P) = f( (x,y) ) =
2
1 m x +
In the Cartesian plane:
(a) Find the Euclidean distance between P = (2, 5) and Q = (0, 1).
(b) Find the coordinate of (2,3) with respect to the line x = 2.
(c) Find the coordinate of (2,3) with respect to the line y = 4x + 11.

Remark The distance function d satisfies the triangle inequality if
d(A, C) d(A, B) + d(B, C) for all A, B and C.


Example 1.3.6: Consider the simplest non-vertical line, y = x. The points (0, 0), (1, 1), (2, 2),
and (3, 3) are on the line. What is the distance from (0, 0) to (1, 1)? From(1, 1) to (2, 2)?
From (1, 1) to (3, 3)? Note the standard ruler for this line is
2 ) , ( x y x f = .The coordinates for the four points
determined by the standard ruler are 0, 2 , 2 2 and 3 2 ,
respectively.

Activity: By subtracting the appropriate coordinates of the ruler, can you obtain the distance
between the points?

Example 1. 3. 7: Let L be the line L
2,3
(i.e a line with slope 2 containing the point (0, 3)) in
the cartesian plane with distance function d. Show that if for an arbitrary point Q = (x, y), f(Q)
= 5x, then show that f is a ruler for L. Also, find the coordinate of R = (1,5).We first show f
satisfies the ruler equation.
Let P = (x
1
, y
1
) and Q = (x, y) be points on L
2,3
.
d(P,Q) =
2
1
2
1
) ( ) ( y y x x + why?
=
2 2
1
)) 3 2 ( 3 2 ( ) ( + + + x x x x why?
=
2
1
2
1
) 2 2 ( ) ( x x x x + why?
=
2
1
2
1
) ( 4 ) ( x x x x + why?
=
2
1
) ( 5 x x why?
= 5 x x
1
why?
= 5 x
1
5 x why?
= f(P) f(Q) why?
This proves the ruler equation.


Activity: Consider the above example
A. Why is f bijective?
B. What is the inverse function?
C. Given a real number r, set x = 5r. Then find y using the equation y = 2x + 3
D. Find the coordinate of R
1.4. The Axiom of betweenness
Here we use the undefined term between to establish some properties of an order relation
among points on a line and plane.

Notation: Point B is between points A and C will be denoted as A B C.

AB
1
: If point B is between points A and C then A, B, C are three distinct points
on a line and B is also between C and A.

Fig. 1.4.1
This axiom implies that the term between is used only for points on a line and states that the
relative position of points A and C does not affect Bs property of lying between A and C.

AB
2
: If A and B are two points on a line h then there exists at least one point C on h such
that A-B-C.
This axiom guarantees the existence of at least three points on a line.

AB
3
: If A, B and C are three collinear points then one and only one of them is between the
others.
AB
3
states that for any three collinear points A, B and C, exactly one of the following is true:

AB
4
: Any four points on a line can be labeled P, Q, R and S in such a way that P Q R
and also P Q S; and that P R S and also Q R S.

As a result of AB
4
, we have:

Fig 1.4.2

Definition 1.4.1. Let A and B be two points. The set of points on the line AB that consists of
points A and B, and all points between A and B is called a line segment determined by A and
B. We denote it by AB. Points A and B are called end points of the line segment AB.
Using set notation we write { } { } B A B X A X AB , | = .
This is means that for distinct points A, B, C; B is between A and C, and write , C B A if
AB C e and AC + CB=AB.

Definition 1.4.2 Let O be a point on a line . A set of points consisting of point O and all
points which are on one and the same side of O is called a ray. Point O is called end point of
the ray. We use point O and any other point say A, on the ray to name it.

Fig.1.4.4

In short the ray (from A in the direction of B) is { } P B A P AB AB = |

Fig. 1.4.5: Betweenness on a line segment (top), ray (middle), and line (bottom).


Activity
1. From the discussion we made so far, what do you conclude about the number of points on
a line? Give justification for your answer?
2. In how many ways we can label four points P, Q, R and S on a line if
P Q R is given.

Exercises
1. Explain why collinear is necessary in the definition of betweenness.
2 Prove that a segment has a unique midpoint.

Theorem 1.4 .1: If A-B-C, then C-B-A.
Proof We must show that CB + BA = CA.
CB + BA = BA + CA why?
= AB + AC why?
= AC why?
= CA why?

Activity
1. Suppose the intersection of AB and CD is CB . Is A-C-B-D? Explain your answer.
2. In the Euclidean plane A-B-C if and only if there is a number t with 0 < t < 1 and B =
A + t(C A).

A line lying in plane t, divides the remaining points of the plane in two parts (called half
planes), so that the line segment determined by two points in the same half planes does not
intersect , whereas the line segment determined by two points in different half planes
intersects .


In fig.14,
1
P and
2
P are on the same half plane. But
1
Q and
2
Q are on different half planes.

Theorem 1.4.2 Every segment contains at least one point different form its end points.
Proof: Let A and B be end points of a segment AB. From
AI
5
b, we have a point C not on line AB. Now taking A and
C there exists a point D on the line through A and C such
that A C D (AII
2
). Again by using AII
2
, we have a point
E on line BD such that D B E. Now consider line EC.
It divides the plane in to two half planes thus points A and
D are on different half planes and points B and D are on the
same half plane. Hence A and B are on different half planes. So AB intersects CE at some
point, say X. Point X is different from A and B (why?) and X is on AB.Consequently
ABcontains at least one point.

Activity
1. Prove that if m is a line then there exist at least three points in a plane containing m ,
which do not lie onm.
2. Let A and B be two points. Does there exist a point X on the line through A and B such
that A X B?

Fig. 1.4.6

Fig. 1.4.7


Theorem 1.4.3 (Betweenness Theorem for Points) Let A, B, C be distinct points on the line.
Let 9 : f be a coordinate function for . Then B C A if and only if either
) ( ) ( ) ( B f C f A f < < or ) ( ) ( ) ( B f C f A f > >

Proof. Suppose that ) ( ) ( ) ( B f C f A f < < . Then
) ( ) ( ) ( ) ( C f B f A f C f CB AC + = +
) ( ) ( ) ( ) ( C f B f A f C f + =
) ( ) ( A f B f =
) ( ) ( A f B f =
AB =
So that A-C-B. A similar argument holds in ) ( ) ( ) ( B f C f A f > > .
Now consider the converse. Assume that A-C-B so that AC +CB= AB, i.e.,
) ( ) ( ) ( ) (( ) ( ) ( A f B f C f B f A f C f = + .
But by algebra, we also have ) ( ) ( ) ( ) ( ) ( ) ( A f B f C f B f A f C f = +
Hence, ) ( ) ( ) ( ) (( ) ( ) ( A f B f C f B f A f C f = + .
Now, let ). ( ) ( ) ( ) ( C f B f v and A f C f u = = Then v u v u + = +
From algebra we know that this means either u or v are both positive and both negative.
Assume the converse. If u> 0 and v< 0 then this gives u - v = u + v which implies v =0 or
: ) ( ) ( C f B f = but C and B are distinct points so ); ( ) ( C f B f =
If 0 < u and 0 > u then v u v u + = + which implies 0 = u or ) ( ) ( A f C f = which is
impossible because A and C are distinct points. Since u and v have the same sign, then both
) ( ) ( ) ( ) ( C f B f and A f C f have the same sign. If both
0 ) ( ) ( 0 ) ( ) ( > > C f B f and A f C f then ) ( ) ( ) ( ) ( C f B f and A f C f > > so that
). ( ) ( ) ( A f C f B f > > If both 0 ) ( ) ( 0 ) ( ) ( < < C f B f and A f C f then
) ( ) ( ) ( ) ( C f B f and A f C f < < so that ). ( ) ( ) ( A f C f B f < <


Corollary 1.4.1 If A, B, C are distinct collinear points then exactly one of them lies between
the other two.
Proof. Since A, B, C are distinct then they correspond to real numbers x, y, z. Then this is a
property of the real numbers, exactly one of x, y, and z lies between the other two.

Corollary 1.4.2 Let A, B, C be points such that . AC Be Then AC AB C B A <
Proof. By theorem 1.4.5 one the following must hold
) ( ) ( ) ( C f B f A f < < (1.1)
) ( ) ( ) ( C f B f A f > > (1.2)
If 1.1 holds then AC A f C f A f B f AB = < = ) ( ) ( ) ( ) (
If 1.2 holds then AC C f A f B f A f AB = < = ) ( ) ( ) ( ) (
To prove the converse, suppose that . AC AB< By the corollary one of A, B, C lies between
the other two. We have three possibilities B C A or C A B C B A ,
But C A B is not possible AC Be and B is distinct from A.
So suppose . B C A Then either ) ( ) ( ) ( B f C f A f < < or ). ( ) ( ) ( B f C f A f > >
If ) ( ) ( ) ( B f C f A f > > then AB A f B f A f C f AC = > = ) ( ) ( ) ( ) (
So AB>AC which is a contradiction.
If ) ( ) ( ) ( B f C f A f < < then AB A f B f A f C f AC = < = ) ( ) ( ) ( ) (
Which is also a contradiction. Hence A-C-B is also not possible.
All that is left is A-B-C.

Definition 1.4.3 The point M is the midpoint of the segment AB if A-M-B and AM= MB.

Theorem 1.4.4 If A and B are distinct points then there exists a unique point M that is the
midpoint of AB.
Proof. To prove existence, let f be a coordinate function for the line , AB and define

2
) ( ) ( B f A f
x
+
=

Since f is onto, there exists some point AB Me such that . ) ( x M f = Hence
) ( ) ( ) ( 2 B f A f M f + = or ) ( ) ( ) ( ) ( M f A f B f M f = . Thus AM = MB . To see that
, B M A let and B f A f a )) ( , ) ( ( min =

)) ( , ) ( ( max B f A f b= . Since A and B
are distinct then b a= and we have
2
b a
x
+
= with a < b. Hence b
b
x = <
2
2
and
a
b
x = >
2
2
giving b x a < < . Hence either ) ( ) ( ) ( B f M f A f < < or ) ( ) ( ) ( B f M f A f > > .
By Theorem 1.4.3 B M A . To verify uniqueness, let , AB M e ' where
. B M M A and M M ' = ' = '

Suppose that ). ( ) ( B f A f < Then both the following
hold: ) ( ) ( ) ( B f M f A f < <

and ) ( ) ( ) ( B f M f A f < ' < . Furthermore, since M and M'
are
midpoints AM M f A f = ) ( ) ( AB
2
1
= M A ' = ). ( ( M f A f ' = Since, ) ( ) ( M f A f < and
) ( ) ( M f A f ' < , this gives ) ( ) ( ) ( ) ( A f M f A f M f ' =

or ) ( ) ( M f M f ' = . Since f is one to
one then M M ' = , which proves uniqueness when ). ( ) ( B f A f > If ) ( ) ( B f A f > then the
inequalities are reversed and we get ) ( ) ( M f A f > and ) ( ) ( M f A f ' > which leads to
) ( ) ( ) ( ) ( M f A f M f A f ' = . Hence, M M ' = by the same argument. Thus, the midpoint is
unique under all cases.

Definition 1.4.4 The union of three line segments AC and BC AB, formed by three non
collinear points A, B and C is called triangle. The points A, B and C are called vertices and
segments AB, BCand AC are called sides.
We denote a triangle with vertices A, B, C as ABC A .

Fig. 1.4.8

Theorem 1.4.5 If a line in the same plane of a triangle does not pass through any vertex of
the triangle and intersects one of its sides then it intersects one and only one of the other two
sides.
Proof:
Let ABC be a triangle and be a line lying in the plane determined by A, B and C. Suppose
does not pass through any of A, B, C and intersects side AB. Then A and B are in different
half planes with respect to . Since does not pass through C, point C is in one of the two
half planes. If C is in the same half plane with A then does not intersect AC but it
intersects BC by AII
5
(as B and C are in different half planes in this case). If C is in the same
half plane with B, then does not intersect BC but it intersects AC by AII
5
(as A and C are
in different half planes in this case). Consequently, in both cases intersects one and only
one of the sides AC or BC of the triangle.

Activity
1. Restate Theorem 1.4.4 using the undefined term between
2. Prove: If A and B are two points on a line mthen there exist at least three points
which lie on m and are between A and B.
3. Prove: If A, B, C are three non collinear points and D, E are points such that A B-
D and B E C then the line through D and E has a point in common with AC .

Remarks:
1) We know that a line contains at least two points by AI
5
. Now by using AB
2
and Theorem
1.4.4 repeatedly we get the following result: A line contains infinitely many points.
2) As a line lying in a plane divides the plane into two parts called half planes, any point on
a line divides the line in to two parts. We call them half-lines.
Let O be any point on line h. Then we say that points A and B of h are on different sides of
O if A O B, otherwise we say that they are on the same side of O.


Activity
1) Examine possible cases in which two different rays can intersect.
2) Give your own definition for an angle. After having done this, compare your definition
with that given below.

1.5 The Plane Separation Postulate
Intuitively, we know that a line divides a plane into halves. These two halves are called half-
planes. We will take this observation as an axiom.

Definition 1.5.1 A set of points S is convex if for every , , S Q P e the entire segment . S PQe

Fig. 1.5.1

Axiom 1.5.1 (Plane Separation Postulate) For every line the points that do not lie on form
two disjoint convex non empty sets
2 1
H and H called half-planes bounded by such that if
1
H Pe and
2
H Qe then PQ intersects .

PQ intersects the line because P and Q
are in different half planes.

The shaded region represents half plane
denoted by H
R, l
(i.e half plane formed by
line l and point R )

Fig. 1.5.2

We can see that this postulate gives rise to the following notion. If both E and F lie in the
same set (i.e in the same half plane determined by l), then the line segment EF does not
intersect l . In this case we say that E and F lie on the same side of l .
More specifically, the plane separation postulate tells us the following (See figure 1.5.2).
us plane whole the H H min
2 1
=
| =
2 1
H H
e
1
, ( H F E ) ( ) (
1
| = _ EF and H EF
e ) , (
2
H H C ) ( ) (
2
| = _ CH and H CH
Thus | = e e PQ H Q and H P
2 1

Definition 1.5.2 Let be a line and A a point not on . Then we use
A
H to denote the half-
plane of that contains A. When the line is clear from the context we will just use the
notation
A
H

Definition 1.5.3 Two points A, B are said to be on the same side of the line if they are both
in the same have-plane. They are said to be on opposite sides of the line if they are I different
half planes.
In figure 1.5.2 points P and Q are on opposite sides of , while points P and F are on the
same side of . In terms of this notation, we can restate the plane separation postulate as
follows.

Axiom 1.5.2 (Plane Separation Postulate, Second Form) Let be a line and let A, B be
points not on . Then A and B are on the same side of if and only if | = AB and are on
opposite sides of if and only if | = AB


Definition 1.5.4 Two rays AC and AB having the same endpoint A are opposite rays if
AC AB = and = BC AC AB

-
B A C
Fig 1.5.3.

Definition 1.5.5 An angle is the union of two non-opposite rays AC and AB having the same
endpoint, and is denoted by CAB or BAC Z Z . The point A is called the vertex of the angle
and the two rays are called the sides of the angle.
Definition 1.5.6. Let A, B, C be points such that the rays AC AB= are not opposite.
The interior of BAC Z is
AB C AC B
H H
, ,
(i.e. the intersection of the two half planes)

Fig. 1.5.4: The interior of the angle BAC Z is the intersection of
the two half-planes, and is shaded darker.

Definition 1.5.7 Three points A, B,C are collinear if there exists a single line such that A,
B, and C all lie on . If no such line exists, then the points are non-collinear.

Corollary 1.5.1 If A, B, and C are non collinear, then the rays AB and AC are neither
opposite nor equal.


Definition 1.5.11 Let A, B, and C be non collinear points. Then the triangle ABC A is the
union of the three segments.
CA BC AB ABC = A
The points A, B, and C are called the vertices of the triangle, and the segments BC AB, and
CA are called the sides of the triangle.

Theorem 1.5.8 (Paschs Theorem) Let ABC A

be a triangle and suppose that is a line that
does not include A,B ,or C. Then if intersects AB then it also intersects either CA or BC

Fig. 1.5.5 Any line that intersects AB must intersect either . BC or CA

Proof. Suppose intersects AB and does not include any of the vertices A, B, or C.
Let
2 1
H and H be the two half planes determined by . Then the points A and B are in
opposite half-planes by the plane separation postulate and the hypothesis.
Suppose
2 1
H B and H A e e (this is just notation; we could have made the alternative
assignment without any loss of generality). Then either .
2 1
H C or H C e e
If ,
1
H Ce

then B and C are in opposite half planes so BC intersects , by the plane
separation postulate.
Alternatively, if
2
H Ce

then A and C are in opposite half planes, so AC intersects , also
by the plane separation postulate.

Activity: What needs to be added so that we can define the interior of triangle ABC?


1.6 Angular Measures
Recall that an angle is the union of two rays with common end point. The common end point
is called the vertex; the two rays are called sides of the angle. (See Fig.1.6.1)

Notation: The angle which is the union of the two rays OA and OB is denoted as
AOB or B O A <
.

Definition 1.6.1
i) The interior of an angle B O A
is the intersection of
(a) the half plane determined by the line OA which contains B and
(b) the half plane determined by the line OB which contains A.
The interior of an angle B O
A will be denoted by ( ) B O
A int . (See fig. 1.6.2)

ii) The exterior of an angle B O A
is the set of all points which are neither on

B O A
nor
( ) B O A
int . The exterior of an angle

B O
A
will be denoted by ( ) B O
A ext . (See Fig. 1.6.3)

Fig. 1.6.1.

Fig. 1.6.2 Fig.1.6.3

Remarks:
1) A line segment formed by any two points in the ( ) B O A
int does not intersect B O A
(That
is it lies completely in this region).A line segment whose one end point lies in the
( ) B O A
int and the other in the ( ) B O A ext
intersects B O
A .
2) In a triangle, an angle will be referred to as being included between two sides when its
sides contain those sides of the triangle. A side will be referred to as being included
between two angles when its end points are the vertices of those angles of the triangle.

In view of this, in ABC A , A AB
, and AC are two sides and the included angle while

B and AB A
are two angles and the included side. Can you mention two more triplets of
(i) two sides and the included angle (ii) Two angles and the included side.

Theorem 1.6.1 (angle addition theorem)
If C B A
and F E D
are angles such that G and H are in the int )
( C B A and
int )
( F E D respectively, H E D G B A

and F E H C B G

then F E D C B A

.
Proof: Suppose H E D G B A

and F E H C B G

with G in the int )
( C B A

And H in the int )
( F E D . To show that F E D C B A

.

Fig.1.6.4
Consider C B A
and ray ED. Then by axiom of angle construction, there exists a point I on
the half plane determined by line EDcontaining H and F such that I E D C B A

. Moreover,
we have one and only one ray EI satisfying this condition.

Fig.1.6.5

Now since G is in the int )
( C B A and I E D C B A

, there exists a unique ray EJ with J in
the int ( ) I E D
such that J E D G B A

and I E J C B G

.Why?
So H E D G B A

(hypothesis of the theorem) and J E D G B A

. Thus EH and EJ cannot
be two distinct rays by axiom of angle construction, as H and J are on the same half plane
determined by ED and J E D G B A H E D

. That is EH and EJ represent the same ray.
Points F and I are on one and the same half plane determined by EH or EJ (why?). Again
from I E J C B G F E H C B G

,

and EH is the same as that of EJ , it follows that EI
and EF represent the same ray by (why?). Hence F E D I E D

.
F E D C B A

by AC
4
( as I E D C B A

and F E D I E D

)

The proof of the following theorem is left for students.
Theorem 1.6.2 (angle subtraction theorem)
If C B A
and F E D
are angles such that point G is the int )
( C B A , point H is in the
int( ) F E D C B A , F E D

and H E D G B A

then F E H C B G

.
Proof: Left as an exercise.


Activity
1. In fig. 1.6.6, DE AE , CE BE and C E D B E A

. Prove that A C D D B A

. Use SAS
theorem, angle addition and subtraction theorems.

Fig.1.6.6

2. You are familiar with certain pairs of angles like adjacent angles, supplementary angles,
vertically opposite angles and so on using the undefined terms, axioms and theorem so far
discussed give your own definition for each of them.

Definition 1.6.2 Two angles are said to be adjacent if and only if they have the same vertex,
one side in common and neither contains an interior part of the other.

Definition 1.6.3 Two angles which are congruent, respectively, to two adjacent angles whose
non-common sides form a straight line are called supplementary angles. Each of a pair of
supplementary angles is called the supplement of the other.

Fig.1.6.7

F E D and C B A
collinear O N P
Q N P F E D
Q N O C B A

, ,

are supplementary angles.

When do we say that two adjacent angles are supplementary?

Definition 1.6.4. Non adjacent angles formed by two intersecting lines are called
vertical angles.

Definition 1.6.5. An angle is said to be a right angle if and only if it is congruent to its
supplementary angle. An angle whose two sides form a straight line is called straight angle.

Illustration: In fig.1.6.8, if lines PQ and RS
intersect at O, then Q O U
and T O Q
are adjacent
angles, T O Q
and P O T
are supplementary
angles, R O P
and S O Q
are vertical angles and

Q O P
is a straight angle.
Can you list some more pairs of adjacent,
supplementary and vertical angles?

So far we have examined different relationships that exist between line segments and between
angles. Some of these relationships are expressed in terms of the undefined terms between
and congruence. In the following group of axioms we will investigate further relationships
between angles and between line segments in terms of the notion of equality. For this we first
raise the following question: What is wrong if we say
i) CD AB = ii) T S R Z Y X

= ?

1.7 Axioms of Congruence
So far we have seen two groups of axioms. These are axioms of incidence, order axiom. Each
consists of a number of axioms that characterize the undefined terms (e.g. point, line and
plane) using the undefined relations incident, between and so on. By using these axioms,
we have stated and proved some properties concerning the undefined terms and the relations
that exist among them. In the same manner we continue our discussion on Euclidean

Fig1.6.8

geometry with a study of the ideas of congruence. The undefined term congruence will be
examined relative to segments, angles and triangles.

Notation: We use the symbol to mean is congruent to and / not congruent.

AC
1
: a) If AB is a line segment then AB AB . (Reflexivity)
b) If AB and CD, are line segments such that CD AB then AB CD .
(Symmetry).
c) If AB, CD, EF are line segments such that CD AB and EF CD then
EF AB . (Transitivity)

AC
2
: If A, B, C, D, E, F are points such that A B C, D E F, DE AB and
EF BC then DF AC . (Axiom of addition of segments)

Fig.1.7.1.

AC
3
: If AB is line segment and C is a point on a line then there exists on on one side
of C exactly one point D such that CD AB . (Axiom of segment construction)
Whenever we have given a line segment XY and a point W on a line , AC
3
enable us to
conclude the existence of a unique line segment WZ on either of the two sides of W on
such that WZ XY . That is if U and V are on the same side of W on such that
WU XY and WV XY then U and V must represent the same point i.e. U =V.

AC
4
: a) If C B A
is an angle then C B A C B A

. (Reflexivity)
b) If C B A
and F E D
are angle such that

F E D C B A

then C B A F E D

.(Symmetry)
c) If I H G F E D C B A
are angles such that F E D C B A

and I H G F E D

then
I H G C B A

(Transitive)

AC
5
: If C B A
is an angle and is a line on any plane and ED is a ray on then there is

one and only one ray EF whose all points except E lie on one of the two half planes
determined by such that F E D C B A

. (Axiom of angle construction)

Given C B A
and ray ED on line (see fig. 1.7.2). There exists exactly one angle on each
side of congruent to C B A
. That is it is not possible for C B A
to be congruent to F E D
and
G E D
, where F and G are on the same half plane determined by , unless F = G. (see fig.
1.7.3)

AC
6
. If in ABC A and DEF A , EF BC and F E D C B A DE AB

, then
F D
E C A
B and E F
D B C
A . That is,

Fig 1.7.4

Fig.1.7.2 Fig.1.7.3

By using the axioms of congruence we will prove several theorems concerning congruence of
segments, angles and triangles. We first prove two theorems about congruence of segments.

Theorem 1.7.1 If A, B, C, D, E and F are points such that A B C, D- E- F, DE AB
and DF AC then . EF BC
Proof Suppose BC is not congruent to EF . Then by AC
3
there exists a point G on ray EF
different from E and F such that BC EG. Hence either
E G F or E F G. In both cases we have D E G. Now from DE AB , BC
EG, A B C and D E G if follows that AC DG by AC
2
. But DF AC , by
assumption. So G = F by AC
3
. Thus we have
F = G and F =G (contrary). Therefore the supposition BC is not congruent to EF is
false. Consequently, BC EF .

Fig.1.7.5

Theorem 1.7.2 If A, B, C, D, E are points such that A B C and DE AC then there
exists exactly one point X such that DX AB = and D X E.
Proof: Suppose A, B, C, D, E are points such that A B C and DE AC . Then AB is
not congruent to DE by AC
3
as AC DE and B, C are on the same side of A on line AC .

Fig.1.7.6
Thus there exists a unique point X on ray DE such that DX AB by AC
3
. Again by using
AC
3
there exists a point G on the line through D and E such that DXG and XG BC .
But A B C, D X G, DX AB , XG BC implies DG AC . From DE AC ,

DG AC and G, E are on the line through D and E on the same side of D it follows that
E = G by AC
3
. Therefore there exists exactly one point X such that DX AB and D X
E. One can prove the following statement by using theorem 1.7.1 and 1.7.2. We state it as a
corollary, as it is an immediate consequence of the two theorems.

Corollary1.7.1: Given two congruent segments XZ and PR. If Y is any point on XZ
different from X and Z then there exists a unique point Q on PR different form P and R such
that PQ XY and QR YZ .
Proof: Left for students

Activity:
1. Explain why the following statements are not necessarily true.
a) If ST PQ and TU QR then SU PR
b) AIV
3
asserts that there is exactly one line segment on a given line that is congruent
to a given line segment.
c) Given angle F E D
and a line h containing point O. Then we have at most two

angles whose vertex is O and congruent to F E D
.
2. Complete the proof of the above corollary
1. PR XZ .. Hypothesis
2. X-Y-Z Hypothesis
3. __________________ by theorem 1.7.2and steps 1 & 2 .
4. ______________ ___________
________________

Now let us deal with some basic points about congruent triangles. We shall discuss more
about triangles in chapter two.

Recall that a triangle is defined as a set of points that lie on three segments which are formed
by three non collinear points. That is if A, B and C are three non collinear points then ABC A
is the union of the segments BC AB, andCA. So every triangle has three vertices, three sides
and three angles. Thus we can establish a one to one correspondence among the vertices,
sides and angles of any two given triangles ABC and DEF. We denote this by
DEF ABC A A and we have the following correspondence:

Vertices
Sides Angles
A corresponds to D
AB corresponds to DE C B
A corresponds to F E
D
B corresponds to E
BC corresponds to EF A C
B corresponds to D F
E
C corresponds to F
CA corresponds to FD B A
C corresponds to E D
F

1.8 Congruence between triangles

So far, we have proved a theorem called side angle side (SAS) congruence theorem. In this
section we will prove theorems, like SAS, that are concerned with conditions which cause one
triangle to be congruent to another triangle. These theorems, which deal with conditions for
triangle congruence, will be used to establish several theorems in this material. First let us
restate SAS theorem.

Restatement: (Side angle side theorem)
If two sides and the included angle of one triangle are congruent, respectively, to two sides
and the included angle of another triangle, then the triangles are congruent.

Theorem 1.8.1 (Angle side Angle, ASA)

If two angles and the included side of one triangle are congruent, respectively, to two angles
and the included side of another triangle, then the triangles are congruent.
Proof: Let ABC A and DEF A be triangles such that F D E C A B

, DE AB and
F E D C B A

. We need to show DEF ABC A A .

Fig.1.8.1
For this it suffices to show that EF BC (why?). Now consider EF . Then there exists a
unique point G on ray BC such that EF BG by AC
2
. Thus in ABG A and DEF A , we
have F E D G B A DE AB

, and EF BG . So DEF ABG A A by SAS, which in turn
implies, F D E G A B

, by definition of congruence of triangles. Since
F D E C A B F D E G A B

,

and G A B C A B
&
are on the same half plane determined by

AB , ray AC must be the same as AG by AC
5
and hence G =C as they are on the same line
on the same side of B. Therefore we have EF BG as ( EF BC and G = C),
F E
D C B
A and DE AB (hypothesis). Consequently, DEF ABC A A , by SAS.

Theorem 1.8.2 (SSS)
If the three sides of one triangle are congruent, respectively, to the three sides of another
triangle then the triangles are congruent,
Proof: Let ABC and DEF be triangles such that EF BC DE AB , and DF AC . To
show that DEF ABC A A .


Fig.1.8.2
On the half plane determine by EF not containing D, there exists a point H such that
H E F A B C

by AC
5
. Mark point ' D on EH so that ' ED BA , this is possible by axiom
of segment construction (AC
2
). Further '

D E F A B C why?
Thus in EF D and ABC ' A A we have EF BC and F E D C B A E D AB
'
, ' which
implies EF D ABC ' A A by SAS. It then follows from definition of congruence of triangles
that F D E C A B F D AC '

, ' and E F D B C A
'
. Now since E D DE ' and

F D DF ' , ' DED A and ' DFD A are isosceles. Hence D D E D D E '
'
and
D D F D D F '
'
(base angles of isosceles triangle are congruent). So, F D E F D E '

, by
angle addition theorem. But, C A B F D E
'
, and hence F D E C A B

, by AC
4
. Thus
from DE AB , F D
E C A
B and DF AC it follows that DEF ABC A A by SAS.

This completes the proof.

Theorem 1.8.3 If two angles of a triangle are congruent then the sides opposite these angles
are congruent.

Activity:
1. Let ABC A be isosceles such that AC AB and D is the midpoint of BC . Use SSS
congruence theorem to show that BC AD and AD is the bisector of C A B
.
2. Prove theorem 1.8.3 by using ASA theorem.

The proof of the following theorem follows identically the same pattern as that used for the
proof of the S.S.S theorem.
Theorem 1.8.4 (RHS): If the hypotenuse and a leg of a right triangle are respectively
congruent to the hypotenuse and a leg of another right triangle then the two triangles are
congruent.
Proof: (you are supposed to fill the missing statements and missing reasons in
the proof)
Statements Reasons
1) On the half plane determined by
AC not containing B there exists
a point R such that E D F R A C

________?_________

2) On AR mark point P such that
DE AP

________?_________

3) Join P with C and B Possible construction
4) DF AC
________?_________
5) ________?_________ By SAS & steps 1,2,4
6) F E D C P A

________?_________
7) C B A
?
Both are given right angles
8) C B A C P A

Transitivity and steps 6,7
9) DE AB
________?_________
10) AP AB
________?_________
11) B P A P B A

________?_________
12) B P C P B C

________?_________
13) ______?_________ _________?-
14) APC ABC A A
15) ______?_______ Transitivity and steps 5, 14

Fig.1.8.3

Fig1.8.4


Illustration: In fig 1.8.5 D is the mid point of BC , AC DE DF DE , and AB DF .
Prove that AC AB .
Proof: B F D
is a right angle as AB DF and C E D
is
a right angle as AC DE .Thus DFB A and DEC A are
right triangles with right angles at F and E. Moreover,
DE DF by assumption and DC DB as D is the
midpoint of BC. Therefore, DEC DFB A A , by RHS
and hence E C
D F B
D by definition of congruence of
triangles. That is A C B A B C

.
Consequently, AC AB . This completes the proof.

Activity:
Prove that the perpendicular line segment from the vertex to the base of an isosceles triangle
(i) bisect the vertex angle (ii) divides the base in to two congruent segments.

We are left with one theorem on congruence of triangles. Before dealing with this theorem,
we state and prove a theorem about a line perpendicular to a given line through a given point
not on the given line. Suppose is a line and P is a point not on . By parallel Axiom there
exists a unique line through P parallel to . What about a line though P perpendicular to ?

Theorem 1.8.5 If m is a line and A is a point not on m then there exists exactly one line
which contains A and is perpendicular to m.
Proof: Suppose m is a line and A is a point not on m. First let us show that there exists at
least one line through A perpendicular to m. Since a line contains at least two points, there
exist points B and C on m. Since A is not on m, BA and BC are two different rays. Thus by

Fig.1.8.5

angle construction axiom there exist a point E on the half plane determined by m not
containing A such that C B E C B A

.

Fig 1.8.6 Fig.1.8.7

By segment construction axiom there exists a point D on BE such that BD BA .
C B D C B A

as rays BD and BE are identical. Since A and D are on different half planes
determined by m, AD intersects m at `some point F. Now there are two possibilities: F = B
or F = B

Fig.1.8.8

(You will be asked to prove case 1 and 2, as an activity.)
From the above steps it follows that, there exists at least one line through A perpendicular to
m. Thus it remains to show that there does not exist more than one line which contains A and

perpendicular to m. To do this, suppose h (whose existence is shown above) is a line through
A perpendicular to m at O.
Now let h be any other line through A
perpendicular to m at Q different from O. Then
there exists a point ' A on h such that A O
' A and O A AO ' (Why?). Points Q, A, ' A
are not collinear, (why?). OQ A AOQ ' A A by
SAS and hence O Q A O Q A
'
. But O Q A
is a
right angle as m h ' at Q. So O Q A
' is also a
right angle. This implies A, Q, ' A lie on the
same line contradiction to that A, Q and ' A are
not collinear.
Therefore there does not exist a line ' h through A different from h perpendicular to m.
Consequently only one line exists through A perpendicular to m.
In this theorem, the given point is not on the given line. What about if the given point is on
the given line? In this case also we have the same result. We put it below as a theorem.

Theorem 1.8.6 Through a given point on a given line there exists one and only one line that is
perpendicular to the given line.
Proof: Left as an exercise

Activity: Complete the proof of Theorem 1.8.5 (prove case 1 and 2).

Theorem 1.8.7 (RHA): If the hypotenuse and a non right angle of one right triangle are
respectively congruent to the hypotenuse and a non right angle of another right triangle then
the two triangles are congruent.
Proof: Let ABC and XYZ be two right triangles with right angle at C and Z respectively
such that XY AB and Z Y X C B A

.
Fig. 1.8.9

Fig.1.8.10

Exercise: Complete the above proof (theorem 1.8.7).

Activity
1. In fig.1.8.11, AC AB and B C
D C B
D . Prove that AD bisects C A
B .
2. In fig.1.8.12, AC BE , AB CD and BE CD . Prove that AE AD .

Fig.1.8.11 Fig. 1.8.12


Exercises :
1. In Fig. 1.8.13 , AQ AP , AC AB and Q A
P C A
B . Which two triangles in the

figure are congruent? Give reasons.
2. In Fig. 1.8.14, AP bisects PN , C A
B is a perpendicular bisector of AB PX , BC and

AC PY .
i. Find three pairs of congruent triangles and prove the congruence.
ii. Prove: 2AX = AB + AC.
3. In fig. 1.8.15, AB=AD , AC =AE and BC=DE . Prove that BE =CD.
4. In fig 1.8.16, B E
A E B
A and ED BC . Prove that EC BD .

Fig.1.8.13 Fig.1.8.14

Fig.1.8.15 Fig.1.8.16

5. In fig. 1.8.17, B C
A B E
D , C B
D E B
A and B is the midpoint of EC . Prove that

DB AB .
6. In fig. 1.8.18, BD bisects C B
A and C D
A . Prove that BD is the perpendicular bisector

of AC.

Fig.1.8.17 Fig.1.8.18

7. Prove: The perpendicular bisector of the base of an isosceles triangle passes through the
point of intersection of the bisectors of the base angles. (Use ASA and RHA congruence
theorems.)
8. In Fig. 1.8.19, AC and BD bisects each other at E. Prove that E is the midpoint of RS.
9. In Fig. 1.8.20, RP and RQ are the bisectors of the congruent angles, B P
A and B Q
A . If
RP = RQ, prove that A, R and B are collinear.

Fig.1.8.19 Fig.1.8.20

10. Y is any point on the side BC of the equilateral triangle ABC, BYK is an equilateral
triangle where A and K are on the different half planes determined by BC . Prove that (i)
AY =CK and (ii) K C
Y Y A
B .


Axiom of continuity (Archimedean Axiom)
If AB and CD are any line segments then there exist points A
1
,

A
2
, ,

A
n
on the line AB .
Such that
a) A
1
lies between A and A
2
,

A
2
, lies between A
1
and A
3
, A
3
lies between A
2
and A
4
etc. and
B lies between A and A
n
.
b) CD A A A A A A AA
n n

1 3 2 2 1
1 .
Let AB and CD be any two line segments then from the axiom of continuity it follows that
there exists a natural number n such that n segments joined end to end and having equal
measure (length) with CD constructed from A, along the ray AB, will pass beyond the
point B.
For instance consider A and B on line m as shown in fig 1.8.22 take any point A
1
between A
and B. Take another point A
2
between A
1
and B such that AA
1
=A
1
A
2
and another point
A
3
if possible between A
2
and B such that A
1
A
2
= A
2
A
3
and so on until you pass over B.
The axiom of continuity assures that this process terminates after a certain number of
steps.
In fig 1.8.21 B is between A and A
7
.

Fig. 1.8.21

1.9 Geometric Inequalities
This section deals with comparison of segments and angles. The concepts of betweenness for
points and congruence for segments can be combined to develop a definition which can be
used for comparing segments. This definition can be used, along with some preceding
theorems, to obtain several theorems pertaining to the comparison of segments. Angles can be
compared in much the same manner as that of line segments. Let us see how this is possible.


Recall that, we have seen that:
Two line segments are equal in length if and only if they are congruent.
Two angles are equal in measure if and only if they are congruent.
But in this section we focus on line segments of unequal length and angles of unequal
measures. For this we put the following definitions about inequalities of line segments and
angles.

Definition 1.9.1
a) Segment AB is said to be less than segment CD if and only if there exists a point E such
that C-E-D and CE AB .
b) Angle C B A
is said to be less than angle F E D
if and only if there exists a ray EG such

that G is in the int )
( F E D and F E G C B A

.

Fig.1.9.1

Notation: AB is less than CD is symbolized as AB < CD. C B A
is less than F E D
is
symbolized as F E D C B A

< . If AB < CD we also write CD > AB and read as CD is
greater than AB. The same is true for angles.

Remarks:
1) In comparing two segments AB and CD, we have only the following three possibilities
and exactly one of them is true:
(i) AB < CD (ii) AB = CD (iii) AB > CD

2) In comparing two angles C B A
and F E D
, we have only the following possibilities and

exactly one of them is true:
(i) F E D C B A

< (ii) F E D C B A

(iii) F E D C B A

>

Now by using definition 1.9.1 and previously proved theorems let us investigate some facts
about comparison of line segments and angles.

Theorem 1.9.1 An interior angle of a triangle is less than each of its remote exterior angles.
Proof: Suppose ABC is a triangle. Consider an interior
angle C A
B .
Then by AB
2
, there exist points D and G on AC and AB,
respectively such that
A-C-D and A-B-G. Also there exist points F and E on BC
such that F B C and
B C E (see fig. 1.9.2). By definition 1.9.1b each of
D C B G B C F B A
and E C A
is a remote exterior angle

of C A B
we show F B A C A B

< , the others can be shown
analogously. Now we have only three possibilities while
comparing C A B
and F B A
:
i) F B A C A B

ii) C A B F B A

< iii) F B A C A B

<
Suppose F B A C A B

, then AC is parallel to BC (Why?). But AC and BC are not
parallel as they intersect at C. Thus the supposition is false. Therefore C A B
is not congruent
to F B A
. Suppose C A B F B A

< , there exists a point H in the int ( ) C A B
such that
H A B F B A

. Again this implies BC || AH (Why?). But this is impossible as ray AH
intersects side BC of ABC A at some point J different from B and C. Thus the supposition is
false. Therefore F B A
is not less than C A B
. Since C A B
is not congruent to F B A
and

Fig.1.9.2

F B A
is not less than C A B
, we have F B
A C A
B < . Analogously, it can be shown that

G B C C A B E C A C A B

,

< < and D C B C A B

< . Therefore an interior angle of a triangle is
less than each of its remote exterior angles.

Activity: Let O be any point inside ABC A . Prove that C O B C A B

< .
Theorem 1.9.2 If two sides of a triangle are not congruent then the angles opposite these
sides are not congruent and the lesser angle is opposite the lesser side.

Proof: Suppose ABC is a triangle with AB / AC . Then C B A
/ B C A
,
otherwise AC AB . Now since AB / AC ,
either AB< AC or AC < AB. Suppose AC <AB.
Then there exists a point D on AB such that
AD AC by definition 1.9.1a. ACD A

is isosceles as AD AC and hence
C D A D C A

. But B C A D C A

< definition
1.9.1b. Moreover C D A
is an exterior angle of CDB A and hence C D A C B A

< by
theorem 1.9.1. Thus from B C A D C A D C A C D A

,

< and C D A C B A

< it follows
that D C A C B A

< B C
A < . Therefore B C A C B A

< . Analogously it can be shown that if
AB < AC then C B A B C A

< . Thus we have proved that an angle opposite to the smallest
side is smallest.

Theorem 1.9.3 If two angles of a triangle are not congruent then their opposite sides are not
congruent and the lesser side is opposite the lesser angle.
Proof: Suppose ABC is a triangle with C B A
/ B C A
. Then either B C A C B A

< or
C B A B C A

< .
Case 1: Suppose B C A C B A

< . To prove that AC < AB.

Fig.1.9.3

If AC AB then B C A C B A

, contrary to the supposition B C A C B A

< . Thus AB is
not congruent to AC .
If AB < AC then C B A B C A

< (by theorem 1.9.2), which is contrary to the
supposition B C A C B A

< . Thus AB < AC is not true.
Therefore neither AC AB nor AB < AC. Consequently AC < AB.
Case 2: Suppose C B A B C A

< . To prove that AB < AC.
If AC AB then C B A B C A

, contrary to the supposition C B A B C A

< . Thus
AB can not be congruent to AC .
If AC < AB then B C A C B A

< , which is contrary to the supposition C B A B C A

< .
Thus AC < AB is not true.
Therefore, neither AC AB nor AC <AB. Consequently AB <AC.
This completes the proof.

The method of proof used in theorem 1.9.3 called proof by exhaustion. It consists in taking in
turn each possibilities and proving that all except one of them are untrue. It then follows that
the remaining possibility must be true.

Remarks: From theorem 1.9.2 & 1.9.3 it follows that: In any triangle:
- the angle opposite the greatest side is the greatest angle
- the side opposite the greatest angle is the greatest side.
Thus in an obtuse triangle, the greatest side is opposite to the obtuse angle; in a right triangle,
the hypotenuse is the greatest side.


Fig.1.9.4

Theorem 1.9.4 The shortest segment joining a point to a line which does not contain the
point is the segment which is perpendicular to the line.

Definition 1.9.2 The distance of a line from a point which
is not on the line is the length of the perpendicular line
segment from the point to the line.In fig. 1.9.5 PA is the
distance between P and .

Activity.
1. Prove theorem 1.9.4
2. Prove that a point is equidistant from the sides of an angle if and only if it lies on the
bisector of the angle.

Theorem 1.9.5 (Triangle inequality)
The sum of the lengths of any two sides of a triangle is greater than the length of the third
side.
Proof: Let ABC be a triangle. We need to show that BA +AC >BC,
BA +BC >AC and BC +AC >BA. We show only BA +AC >BC. The others can be shown
in similar manner. See Fig. 1.9.6.

Fig.1.9.4

Extend BA to some pint X on AB such that B-A-
X and AC AX . This is possible by Axiom of
segment construction. Join C and X. Since
A C X C X A AC AX

, . B C X A C X

< by
definition of angle comparison. Thus,
B C X C X A

< (i.e. B C X C X B

< ). Now in
BCX A we have BC < BX. But BX =BA +AX
(as B, A, X are collinear and B A - X)
BX =BA +AC as AX AC .
Therefore, BC <BA +AC.

Exercise
1. Prove that the difference of the lengths of any two sides of a triangle is less than the third
side.
2. How many triangles, not any two of them are congruent, can we have such that the lengths
of two sides for each of them are 4 and 7, and the third side is a whole number?
3. If P is an interior point of ABC A . Prove that AC +BC >AP +BP.
4. Prove: Of two segments of unequal lengths from a point to a line, the greater cuts off the
segment of greater length from the foot of the perpendicular.
5. Prove: If two sides of one triangle are congruent to two sides of another triangle but the
measures of the included angles are unequal then the lengths of the third sides are unequal
in the same order.
6. Prove: If two sides of one triangle are congruent to two sides of another triangle but the
lengths of the third sides are unequal then the measures of the angles included between the
pairs of congruent sides are unequal in the same order.
7. Prove: If an altitude of a triangle does not bisect the angle from which it is drawn then the
sides forming this angle are not congruent.
8. Let D be the midpoint of the side BC of ABC A . If AB >AC, prove that C A
D D A
B < .
(Produce AD to P so that AD =AP. Join BP.)

Fig.1.9.5

9. D is the midpoint of the side BC of ABC A . Prove that AD < (AB +AC).
10. In Fig. 1.9.7, A is the image of A in the straight
line CD, i.e., CD is the perpendicular bisector
of ' AA ; A'
EB is a straight line. Prove that
(i) D E
B C E
A , (ii) EB AE PB AP + > + .
This is called the light-path theorem. If a ray of
light from source A is reflected in a mirror CD
so as to travel to B, it follows the shortest path
AEB, and for this path the angle of incidence,
C E
A , is equal to the angle of reflection, D E
B

Example 1.9.1 Fig. 1.9.8 shows a triangle with angles of different measures. List the sides
of this triangle in order from least to greatest.

Fig. 1.9.8 List the sides of this triangle in increasing order.

Because 30 < 50 < 100, then RS < QR < QS.

Example 1.9.2. Fig. 1.9.9 Shows a triangle with
sides of different measures. List the angles of this
triangle in order from least to greatest. Because 6 <
8 < 11, then m ( N) < m ( M) < m ( P).

Fig.1.9.7

Fig. 1.9.9

Example 1.9.3 : In Fig. 1.9.10, the measures of two sides of a triangle are 7 and 12. Find the
range of possibilities for the third side.

Fig. 1.9.10

What values of x will make a triangle possible?

Using the Triangle Inequality Theorem, you can write the following:
7 + x > 12, so x > 5
7 + 12 > x, so 19 > x (or x < 19)
Therefore, the third side must be more than 5 and less than 19.

Example 1.9.4 Fig. 1.9.11 shows right ABC. Which side must be the longest?

Fig. 1.9.11

Identify the longest side of this right triangle.

Because m A + m B + m C = 180 (by Theorem 25) and m = 90, we have m A + m
C = 90. Thus, each of m A and m C is less than 90. Thus B is the angle of greatest
measure in the triangle, so its opposite side is the longest. Therefore, the hypotenuse, AC , is
the longest side in a right triangle.

1.10 Sufficient Conditions for Parallelism

Two lines are parallel if they lie in the same plane but do not intersect. We shall use the
abbreviation
2 1
L L to mean that the lines
2 1
L and L are parallel. Later, as a matter of
convenience, we shall say that two segments are parallel if the lines that contain them are

parallel. We shall apply the same term to a line and a segment, a segment and a ray, and so
on.
The Euclidean parallel postulate will be introduced in the next chapter, and used thereafter,
except in the chapter on non-Euclidean geometry. The postulate, in the form in which it is
usually stated, says that given a line and a point not on the line, there is exactly one line which
passes through the given point and is parallel to the given line.

Fig 1.10.1

We shall see, however, from Theorems 1.10.1 and 1.10.2, that half of this statement can be
proved on the basis of the postulates that we already have.

Theorem 1.10.1. If two lines lie in the same plane, and are perpendicular to the same line,
then they are parallel.
Restatement. Let T and L L
2 1
, be three lines, lying in a plane E, such
that .
2 1
T L and T L . Then
2 1
L L

Fig 1.10.2


Proof. Suppose that
2 1
L and L intersect T at points Q and P, respectively. Suppose that
2 1
L and L are not parallel, and let R be the point at which they intersect. Then there are two
perpendiculars to T through R; and this is a contradiction. Why?

Theorem 1.10.2 Given a line and a point not on the line, there is always at least one line
which passes through the given point and is parallel to the given line.

Proof. Let L be the line, let P be the point, and let E be the plane which contains them. Then
there is a line T in E which passes through P and is perpendicular to L. Then there is a line L'

in E which passes through P and is perpendicular to T. By the preceding theorem it follows
that L L '
1
, which was to be proved.

There is an easy generalization of Theorem 1.10.1, which we shall get to presently. In the
figure below, T is a transversal to the lines .
2 1
L and L

Fig 1.10.3

More precisely, if
2 1
, L L and T are three lines in the same plane, and T intersects
2 1
L and L in two (different) points P and Q, respectively, then T is a transversal to
2 1
L and L .
In the figure below 2 1 Z Z and are alternate interior angles: and 4 3 Z Z and are alternate
interior angles.


Fig 1.10.4
More precisely,
(1) if T is a transversal to ,
2 1
L and L intersecting
2 1
L and L in P and Q, respectively, and
(2) A and D are points of
2 1
L and L , respectively, lying on opposite sides of T, then
PQD and APQ Z Z are alternate interior angles. (Under a change of notation, this
definition says also that QPB and CQP Z Z are alternate interior angles.)

Theorem 1.10.3 Given two lines and a transversal. If a pair of alternate interior angles is
congruent, then the lines are parallel.
The proof uses the exterior angle theorem.

In the figure below, 1 1 ' Z Z and are corresponding angles, 2 2 ' Z Z and are corresponding
angles, and so on.

Fig 1.10.5

Definition 1.10.1 If y and x Z Z are alternate interior angles, and z and y Z Z vertical angles,
then z and x Z Z are corresponding angles.

Given two lines and a transversal. If a pair of corresponding angles are congruent, then the
lines are parallel.

Example: Prove that if n and m , then either . n m or n m=
Solution: Let n and m , be three lines such that n and m . We must prove that ether
. n m or n m = Let A be the point at which and m intersect and let B be the point at which
and n intersect (definition of perpendicular lies).There are two cases: either A = B or
. B A= If A =B, them m = n by the uniqueness of perpendiculars.

Activity: In the above example prove the case B A=

1.11 Saccheri Quadrilaterals

Quadrilaterals in Neutral Geometry
Much of the work, which goes back to Umar al-Khayyami (1048-1131) and Nasir Eddin
(1201-1274) is often attributed to the European geometers Giovanni Saccheri (1667-1733)
and Johann Lambert (1728-1777) who rediscovered the earlier results. None of these attempts
were able to successfully demonstrate that a rectangle could be constructed using only the
axioms of neutral geometry. Here we look at some of the results that can be obtained for
quadrilaterals in Neutral Geometry.

Definition 1.11.1 Let A, B, C, D be points, no three of which are collinear, such that any two
of the segments DA CD BC AB , , , either have no point in common or only have an endpoint in
common. Then the points A, B, C, D determine a quadrilateral, denoted by ABCD. The
points A, B, C, D are called the vertices of the quadrilateral. The segments DA CD BC AB , , ,
are called the sides of the quadrilateral. The diagonals of ABCD are the segments
. BD and AC

Two quadrilaterals are congruent if all four corresponding sides and all four corresponding
angles are congruent.

Fig 1.11.1: ABCD is a convex quadrilateral with diagonals BD and AC ; EFGH is a non-
convex quadrilateral with diagonals EG and ; EH IJKL is not a quadrilateral, although
IKJL (not shown) is a quadrilateral.

Definition 1.11.2 ABCD is convex if each vertex is contained in the interior of the angle
formed by the three other vertices (in their cyclic order around the quadrilateral).

Definition 1.11.3 Let ABCD be convex. Then its angle sum is given by the sum of the
measures of its interior angles:
( o BACD)= ) ( ) ( BCD m ABC m Z + Z ) ( ) ( BCD m ABC m Z + Z +

Theorem 1.11.1 (Additively of Angle Sum) Let ABCD be a convex quadrilateral with
diagonal BD. Then ( o ABCD)= ) ( ) ( BDC ABD A + A o o
.

Fig 1.11.2: The angle sum of a quadrilateral is equal to the
sum of the angle sums of the triangles defined by either diagonal


Proof. Apply the angle addition postulate to each of the angles that are split by a diagonal to
get ( o ABCD) = u | o + e + +
= q o o + + e + + +r
= ) ( ) ( o q o + e + + + +r
+ ) ( ) ( BDC ABD A + A o

Definition 1.11.4 The defect of a quadrilateral is o (ABCD)=360 - o (ABCD)

Theorem 1.11.2 (The additivity of Defect for Convex Quadrilaterals) If ABCD is
convex quadrilateral then o (ABCD)=

) ( ) ( ACD ABC A + A o
Proof. Apply theorem 1.11.1.

Corollary 1.11.1 If ABCD is convex, then ( o ABCD) s 360.
Proof. Apply theorem 1.11.1
Fig 1.11.3: The angle sum of a quadrilateral is equal to the sum
of the angle sums of the triangles defined by either diagonal (theorem 1.11.1)

Fig 1.11.4: ABCD is a parallelogram

Definition 1.11.5 ABCD is called a parallelogram if . AD BC and CD AB

Theorem 1.11.3 Every parallelogram is convex.
Proof.(Exercise)

Theorem 1.11.4 Let ABC A Be a triangle and D and E points such that
C E A and C D A
.
Then BCED is a convex quadrilateral.
Proof, (Exercise)

Fig 1.11.5: C E A and B D A BCED is a quadrilateral.

Theorem 1.11.5 BCED if and only if the diagonals have an interior point in common (i.e.
they intersect, but not at an endpoint).
Proof.
( )

Assume BCED is convex. Then by the definition of convexity C is the interior of
ZDAB. Then | = AC BD ; call the point of intersection E, where D E B
.
By similar
argument there is a point F= BD AC where C F A
.
Since BD and AC are distinct (they
correspond to opposite sides of a quadrilateral), they can meet in at most one point, we must
have E=F. Hence the diagonals intersect at E. Since D E B and C F A , the
intersection is not at an endpoint.
Conversely
( ) :

Let BCED be a quadrilateral with E= BD AC

with
. D E B and C E A

Since . C E A A and E are on the same side of the line CD

Similarly, since E and B D E B , are on the same side of . CD

Hence A and B are on the
same side of CD (Plane separation postulate), i.e. .
,CD B
H Ae

Recall that
CD B
H
,
is a half

plane determined by line CD and point B not on CD
.
(i.e
CD B
H
,
is a half plane containing
point B.) By a similar argument A and D are on the same side of
C B D
H A e i BC
'
e
,
, . ,

Hence
C B D CD B
H H A
'
e
, ,
, and thus A is in the interior of . BCD Z

By a similar argument, each of
the other vertices is in the interior of its opposite angle. Hence by definition of convexity, the
quadrilateral is convex. Assume BCED is a convex quadrilateral (i.e., Assume that R is
false). Then , | = BD AC i.e. the diagonals of BCED share an internal point. Hence
BCED is not a quadrilateral.

Example 1.11.1 Show that every parallelogram is convex.
Solution: Let ABCD be a parallelogram (hypothesis). We must prove that ABCD is a
convex quadrilateral. Since BC AD (definition of parallelogram), it follows that
. | = BC AD (definition of parallel). Hence A and D lie on the same side of BC (Plane
Separation Postulate). In the same way, the fact that CD AB can be used to prove that A and
B lie on the same side of . CD

Thus A is in the interior of BCD Z (definition of angle
interior).The remaining conditions left as an activity. Therefore, ABCD is convex
(definition of convex quadrilateral)

Activity: In the above example:
a) Show that B is in the interior of , CDA Z
b) Show that C is in the interior of and D is in the
interior of

, DAB Z
. ABC Z

Theorem 1.11.6 If ABCD is a non-convex quadrilateral then ACBD is a quadrilateral.
Proof. Since ABCD is a quadrilateral no three of the points A, B, C, D are collinear. Since
ABCD is a quadrilateral . | = AD BC SinceABCD is not convex then BD AC, are
disjoint (the diagonals do not intersect). Thus segments DA and BD CB AC , , , share at most
their endpoints. Hence BCED is a quadrilateral.

Definition 1.11.6 ACBD is a Saccheri Quadrilateral if
0
90 = Z = Z BAD ABC and AD=BC,
Segment ABis called the base and segment DC is called the summit.

Fig 1.11.6

Theorem 1.11.7 The diagonals of the Saccheri Quadrilateral are congruent.
Proof. Consider triangles . ABC and ABD A A Since , 90 , , B A and AB AB AD BC Z = = Z = = the
triangles are congruent. Hence AC BD~ .

Fig 1.11.7: the shaded triangles are congruent, hence
the corresponding diagonals of the Saccheri Quadrilateral are congruent (theorem 1.11.7).


Theorem 1.11.8 The summit angles of a Saccheri Quadrilateral are congruent.
Proof. Repeat the argument in the previous proof, but with the upper-half triangles. The
triangles are congruent by SSS-they share the same top, the diagonals are congruent; and the
sides are congruent. Hence the corner angles are congruent.

Theorem 1.11.9 Let ABCD be a Saccheri quadrilateral. Then the segment joining the
midpoints of the base and summit is perpendicular to the base and summit.
Proof. Let M be the bisector of AB and N the bisector of CD (see fig 1.11.8). Since M is a
bisector, AM = MB; by definition of the Saccheri quadrilateral AD = BC and . B A Z = Z
Hence CBM DAM A ~ A by SAS.

Fig. 1.11.8

Hence DM=CM; by definition of bisector DN=CN; hence CMN DMN A ~ A by SSS (they
share a common side). By congruence, ; CNM DNM Z = Z they must each be right angles. A
similar argument is used to show that by BMN AMN : 90 = Z = Z SAS ; NDA NCB A ~ A hence
; NA BN~ hence by SSS, ; NMA NMB A ~ A hence by congruency, . BMN AMN Z ~ Z

Example 1.11.2
Prove the following: Let ABC A be a triangle and D and E points such that
. C E A and B D A Then BCED is convex.
Solution:
Let ABC A be a triangle, let D be a point on AC and let E be a point on BC (hypothesis).
We must prove that BCED is a convex quadrilateral.

First note that E is in the interior of BAD Z and D is in the interior of ABCE Z
.
Now A and
D are on the same side of BE BC= . In addition, A and C are on opposite sides of DE
(because AC intersects DE at D). Furthermore, B and C are on opposite sides of
DE(because BC intersects DEat E). Hence A and B are on the same side of DE (Plane
Separation Postulate). Therefore, A is in the interior of BED Z (definition of angle interior). A
similar proof can be used to show that B is in the interior of . EDA Z

Thus ABED is
convex (definition of convex quadrilateral.).

Theorem 1.11.10 If ABCD is a Saccheri quadrilateral, then it is a parallelogram.
Proof. See fig 1.11.9 . DC MN and AB MN By the alternate interior angles theorem,
. AB CD Similarly, . BC AD

Fig 1.11.9

Theorem 1.11.11 Saccheri quadrilaterals are convex.
Proof: This follows because every Saccheri quadrilateral is a parallelogram (theorem 1.11.10)
and every parallelogram is convex.

Theorem 1.11.12 The summit angles of a Saccheri quadrilaterals are either right angles or
acute.
Proof. Since ABCD is convex (theorem 1.11.11). o (ABCD) s360. By definition of a
Saccheri Quadrilateral, . 90 = Z = Z B A Hence 180 s Z + Z D C
By theorem 1.11.8 the summit angles are congruent, hence . D C Z = Z
Hence . 90 90 s Z s Z D and C

Definition 1.11.7 A Lambert quadrilateral is a quadrilateral in which three of the interior
angels are right angles.

Fig 1.11.10: A Lambert Quadrilateral. Like the Saccheri Quadrilateral, it is not possible to
prove that this figure is a rectangle using only the axioms of neutral geometry-one must
accept Euclids fifth postulate to do so.

Corollary 1.11.2 Let ABCD be a Lambert quadrilateral. Then it is a parallelogram.
Proof. This follows because a Lambert quadrilateral is also a Saccheri quadrilateral, and
hence it is a parallelogram (theorem 1.11.11).

Corollary 1.11.3 Let ABCD be a Lambert quadrilateral. Then it is convex.
Proof. It is a parallelogram and all parallelograms are convex.

Corollary 1.11.4 Let ABCD be a Lambert quadrilateral with right angles at vertices A, B,
and C. Then D Z is either a right angle or it is acute.
Proof. This follows because it is a summit angle of a Saccheri quadrilateral.


Corollary 1.11.5 Let ABCD be a Lambert quadrilateral with angles at vertices
A, B, and C then AD BCs

Fig 1.11.11: In a Lambert quadrilateral, AD BCs

Proof Suppose BC > AD. Then there exists a point E with C E B such that AD BD=
(ruler postulate).ABED is a Saccheri quadrilateral (def. of Saccheri quadrilateral). Hence
90 s ZBED Angle BED Z is an exterior angle for ECD A
.
Angle , 90 = ZC it is a remote
angle of , BED Z

By the exterior angle theorem, BED C Z < Z (strict inequality). This leads to
the result 90 < 90; therefore we must conclude that . AD BCs

1.12 The Angle-Sum Inequality for Triangles
If we only assume Euclids first four postulates, along with the axioms of incidence,
congruence, continuity and betweenness, the angle sum of a triangle is always less than or
equal to 180. This geometry is called neutral (or absolute) geometry.
We will also consider some important consequences of this theorem.

Theorem 1.12.1 (Exterior Angle Inequality) The measure of an exterior angle of a triangle
is greater than the measure of either remote interior angle.
Proof. Given , ABC A extend side BC to ray BC and choose a point D on this ray so that
. D C B we claim that A m ACD m Z > Z and > ZACD m B mZ . Let M be the midpoint of
AC and extend the median BM so that M is the midpoint of BE.

Fig 1.12.1
Then AMB Z and CME Z are congruent vertical angles and CME AMB A ~ A by SAS.
Consequently, CAB m ACE m Z = Z . Now, E lies in the half-plane of A and CD
,
since A and E
are on the same side of CD. Also, E lies in the half-plane of D and AC since D and E are on
the same side of AC. Therefore E lies in the interior of ACD Z , which is the intersection of
these two half-planes. Finally, . A m CAB m ACE m ECD m ACE m ACD m Z = Z = Z > Z + Z = Z

Activity: In the above theorem (theorem 1.12.1) prove the case B m ACD m Z > Z

Corollary 1.12.1 The sum of the measures of any two interior angles of a triangle is less than
180.
Proof. Given , ABC A extend side BC to BC and choose points E and D on BC so that
D C B E (see fig 1.12.2).

Fig 1.12.2

By Theorem1.12.

ABE m A m and ACD m B m ACD m A m Z < Z Z < Z Z Z Z , , , 1 . By adding
ACB m C m Z = Z

to both sides of the first two inequalities, and by adding ABC m B m Z = Z to
both side of the third we obtain
180 = Z + Z < Z + Z ACB m ACD m C A m

180 = Z + Z < Z + Z ACB m ACD m C B m
180 = Z + Z < Z + Z C AB m ABE m B A m

Theorem 1.12.2 If two lines are cut by a transversal and a pair of alternate interior angles are
congruent, then the lines are parallel.
Proof. We prove the contrapositive. Assume that lines l and m intersect at the point , R and
suppose that a transversal t cuts line l at the point A and cuts line m at the point B. Let
2 1 Z Z and be a pair of alternate interior angles. Then either 1 Z is an exterior angle of
2 Z A and ABR is a remote interior angle or vise versa.

Fig 1.12.3
In either case 2 1 Z = Z m m

by the Exterior angle Inequality (Theorem 1.12.1).

Theorem1.12.3 (Saccheri-Legendre Theorem) The angle sum of a triangle is less than or
equal to 180.
Proof. Assume, on the contrary, that the angle sum of p ABC + = A 180
,
for some
. 0 > p Construct the midpoint M of side ACthen extend BM its own length to point E such
that E M B . Note that CEM ABM A ~ A by SAS. Therefore
the angle sum of ABC A = angle sum of + AABM angle sum of BMC A

=
angle sum of + ACEM angle sum of BMC A

= angle sum of . BEC A

Furthermore . ABE m BEC m Z = Z Therefore, either ,
2
1
ABC m BEC m Z s Z
or .
2
1
ABC m EBC m Z s Z

Thus we may replace ABC A with BEC A having the same angle sum
as ABC A and one angle whose measure is less than or equal to ABC mZ
2
1
.

Fig 1.12.4

Now repeat this construction in EBC A
. If
,
2
1
ABC m EBC m Z s < construct the midpoint N
of CE and extend BNits own length to point F such that B-N-F. Then BEC A

and BFC A
have the same angle sum and either EBC m FBC m or EBC m BFC m Z s Z Z s Z
2
1
2
1
. Replace
EBC A with FBC A having the same angle sum as ABC A and one angle whose measure is
.
4
1
ABC mZ s On the other hand, if ABC m BEC m Z s Z
2
1
, do same construction with N as
the midpoint of BC and replace EBC A with . FEC A Continue this process indefinitely; the
Archimedian property of real numbers guarantees that for sufficiently large n, the triangle
obtained after the n
th
iteration has the same angle sum as ABC A and one angle whose measure
is p ABC m
n
< Z s
2
1
, in which case the sum of its other two angles is greater than 180
0
,
contradicting Corollary 1.12.1.

Example 1.12.1: Prove that the sum of the measures of two interior angles of a triangle is
less than or equal to the measure of their remote exterior angle.
Solution: Let ABC A be a triangle and let D be a point on AB such that D B A
(hypothesis). We must prove that ) ( ) ( ) ( CBD m CAB m BCA m Z s Z + Z


In the figure above, we know that 180 ) ( ) ( ) ( s Z + Z + Z BCA m ABC m CBA m (Saccheri-
Legendre Theorem). We also know that 180 ) ( ) ( = Z + Z CBD ABC (Linear Pair Theorem).
Hence from algebra ) ( ) ( 180 ) ( ) ( CBD m ABC m BCA m CAB m Z = Z s Z + Z

Definition 1.12.1 The defect of . 180
0
C m B m B m A m ABC is ABC Z Z Z Z = A o

Corollary1.12.2 Every triangle has non-negative defect.
Proof.If
0 0
0 180 < Z Z Z = C m B m A m ABC o , then the angle sum of
0
180 > AABC contradicting theorem 1.12.3.

Theorem 1.12.4 (Additivity of defect) Given any triangle ABC A and any point D between A
and B, . BCD ACD ABC o o o + =

Fig 1.12.5
Proof. Since BDC and ADC Z Z are supplementary,
0
180 = = Z + Z CDB m CDA m . Since DC
is in the interior of . , BCD m ACD m ACB m ACB Z + Z = Z Z
Therefore,


DAC m CDA m ACD m BCD ACD Z Z Z = +
0
180 o o

DBC m CDB m BCD m Z Z Z +
0
180
DBC m BCD ACD m DAC m CDB m CDA m Z Z + Z Z Z + Z = ) ( ) ( 360
0

= CDA m BAC m ABC m Z Z Z +
0
180
. ABC o =

Corollary1.12.3 Given any triangle ABC A and any point D between A and B, the angle sum
of
0
180 = AABC if and only if the angle sums of BCD and ACD A A both equal 180
0.
Proof. If the angle sums of both BCD and ACD A A equal 180
0
, then . 0
0
= = BCD ACD o o
By Theorem 1.12.4,
0
0 = ABC o so that the angle sum of , 180
0
= AABC
. 0
0
= + = BCD ACD o o But by corollary 1.12.2,
0 0
0 0 > > = BCD and ACD o o . Therefore
0
0 = = BCD ACD o o and both angle sums equal 180
0
.

Theorem 1.12.5 If there is a triangle with angle sum 180
0
, then a rectangle exists.
Proof. Consider a triangle ABC A with angle sum 180
0
, by corollary 1.12.1, the sum of the
measures of any two interior angles is less than 180
0
, so at most one angle is obtuse. Suppose
B and A Z Z are acute and construct the altitude , CD we claim that A D -B. But if not, then
either D A B or A B D. Suppose D A B and consider . DAC A

Fig 1.12.6

Then the remote interior angle CDA Z has measure 90
0
, which is greater than the measure of
exterior angle , CAB Z contradicting Theorem 1.12.1, Assuming the A-B-D leads to a similar
contradiction, proving the claim. Then by corollary 1.12.3
0
0 = = BDC ADC o o . Let us

construct a rectangle from right triangle .. BCD A By the congruence axioms, there is a unique
ray CX with X on the opposite side of BC from D such that BCX CBD Z ~ Z , and there is a
unique point E on CX

such that . BD CE~

Fig 1.12.7
Then BCE CBD A ~ A by SAS; therefore BCE A is a right triangle with
0
0 = BCE o and right
angle at E. Also, since
0
90 = Z + Z BCD m DBC m , substituting corresponding parts gives
0
90 = Z Z BCD m and ECB m and . 90
0
= Z + Z EBC m DBC m Furthermore, since alternate
interior angles DBC and ECB Z Z are congruent DB CE by Theorem 1.12.2. Therefore B is
an interior point of . ECD Z

By the same argument, ED CD and C is an interior point of
. EBD Z Therefore
0
90 = Z = Z EBD m ECD m and CDEB is a rectangle.

Theorem 1.12.6 If a rectangle exists, and then the angle sum of every triangle is 180
0

Proof. We first prove that every right triangle has angle sum 180
0
. Given a rectangle, we can
use the Archimedian property to lengthen or shorten the sides and obtain a rectangle AFBC
with sides AC and BC of any prescribed length. Now given a right triangle D C E ' ' ' A , construct
a rectangle AFBC such that C E BC and C D AC ' ' > ' ' > . There is a unique point D on AC
and a unique point E on BC such that D C E ECD ' ' ' A ~ A as shown in fig 1.12.8

Fig 1.12.8

We claim
0
0 = ABC o .If not, then
0
0 > ABC o by Corollary 1.12.2 and consequently
. 90
0
< Z + Z BAC m ABC m

But
0 0
90 90 = Z + Z = Z = Z + Z = Z BAF m BAC m CAF m and ABF m ABC m CBF m
Therefore BAC m BAF m and ABC m ABF m Z = Z Z = Z
0 0
90 90 so that
BAF m ABF m ABF Z Z =
0 0
90 180 o
) 90 ( ) 90 ( 90
0 0 0
BAC m ABC m Z Z =

0 0
0 90 < Z + Z = BAC m ABC m ,
Contradicting corollary 1.12.2 and proving the claim. Now by repeated application of
corollary1.12. 3 we have
0 0
0 0 = = ECD and BCD o o . But D C E ECD ' ' ' A ~ A implies
0
0 = ' ' ' D C E o . Thus every right triangle has zero defect. Now by the construction in Theorem
1.12.5, an arbitrary triangle ABC A can be appropriately labeled so that its altitude CD lies in
the interior of
ABC A and subdivides the triangle into two right triangles (see fig 1.12.9), each having zero
defect. Thus
0
0 = ABC o by Corollary 1.12.3.

Fig 1.12.9

Corollary 1.12.4 A rectangle exists if and only if every triangle has angle sum 180
0
.

1.13 The Critical Function
In this subunit, we shall make heavy use of the incidence and separation theorems. For
convenience, we briefly restate two of them:


The Postulate of Pasch: Given ABC A and a line L (in the same plane). If L intersects ABat
a point between A and B, then L also intersects either AC or . BC

Fig 1.13.1
The Crossbar Theorem. If D is in the interior of , BAC Z then AD intersects . BC

Fig 1.13.2

Given a line L and an external point P. Let A be the foot of the perpendicular from P to L, and
let B be any other point of L(fig 1.13.3). For each number r between 0 and 180 there is
exactly one ray D with PD, on the same side of AP as B , such that
r APD m = Z

Fig 1.13.3

Obviously, for some numbers PD r , will intersect . AB (For example, take .) APB m r Z = For
r > 90, PD will not intersect AB. Let { } AB PD r K intersects = . Then K is nonempty,
and has an upper bound. Therefore K has a supremum. Let . sup
0
K r =

The number
0
r is
called the critical number for P and AB
.
The angle APD Z with measure equals
0
r is called
the angle of parallelism of AB and P.

Theorem 1.13.1. If PD then r APD m ,
0
= Z does not intersect AB.
Proof. Suppose that PD intersects AB at. Q:

Fig 1.13.4
If R is any point such that A-Q-R, then
0
r APR m > Z , so that
0
r is not an upper bound K.

Theorem 1.13.2. If . sec int
0
AB ts er PD then r APD m < Z

Fig 1.13.5
Proof. Since
0
r = sup K, and ,
0
r APD m < Z it follows that APD mZ is not an upper bound of
K. Therefore some r in K is greater than . APD mZ Let D' be such that

. r D AP m = ' Z then D P ' intersects AB in a point F. But PD is in the interior of
I
APD Z
.
Therefore by the crossbar theorem PD intersects AF . Therefore PD intersect AB. Thus there
is a certain critical ray PD, with PD r APD m ;
0
= Z does not intersect AB. But if F is in
the interior of APD Z , then PF does intersect AB.
[If F is in the interior of , APD Z we shall say that AF is an interior ray of .] APD Z

Figure 1.13.6
Note that
0
r was defined in terms of P, A, and B. It turns out, however, that
0
r depends only
on the distance AP.

Theorem 1.13.3. Let P, A, B and also B A P ' ' ' , , be as in the definition of the critical
number. If P A AP ' ' = , then the critical numbers
0 0
, r r ' are the same.

Fig 1.13.7

Proof. Let { } AB ts er PD r K sec int = and let { } B A ts er D P r K ' ' ' ' = ' sec int

If , K r e
let Q be the point where PD intersects , AB and let Q' be the point of B A ' '

for which
AQ Q A = ' ' .Then . r D P A m = ' ' ' Z (Why?) Therefore . K r ' e Thus ; K K ' _ and similarly
. K K _ ' therefore . s u p s u p . K K and K K = ' = '

We now have a function

.
0
r AP

We shall denote this function by c, and call it the critical function. Thus, for every
a > 0, c( a ) denotes the critical number corresponding to AP = a . Thus PD intersect
AB when ), (a c APD m < Z but PD does not intersect AB when APD mZ

| c ( a ).

Fig 1.13.8

We shall now investigate the function c.
Theorem 1.13.4. c never increases as a increases. That is, if , a a > ' then ) (a c '

c(a).

Fig 1.13.9
Proof. Given P P ' , , with , , P A a AP a ' = ' = as fig 1.13.9. Take PD so that ) (a c APD m = Z

and take D P ' ' so that ) (a c D P A m = ' ' ' Z . Then PD and D P ' ' are parallel. Therefore all
points of D P ' ' are on the side of PD that contains P'
. And all points of AB are on the sidof
PD that contains A. Therefore D P ' ' does not interest AB. Now let
{ } AB ts er D P r K sec int ' ' = '

as in the definition of the critical angle, so that
. sup ) ( K a c ' = '

Then ) (a c is an upper bound of , K' because D P ' ' does not intersect AB.
And ) (a c ' is the least upper bound of K '
. Therefore s ') (a c ) (a c , which was to be proved.

In the Euclidean case, this theorem cannot be strengthened to give the strict inequality
) ( ) ( a c a c < ' for . a a > ' (The reason, obviously, is that in Euclidean geometry we have

c(a) =90 for every a .) In hyperbolic geometry, however, not only do we have the strict
inequality but we actually have . 0 ) ( a as a c
Theorem 1.12.4 allows the possibility that 90 ) ( < a c when a is large, but 90 ) ( = a c when a is
sufficiently small. But in fact this cannot happen, as the following two theorems show.

Theorem 1.13.5 If 90 ) 2 / ( 90 ) ( < < a c then a c

Fig 1.13.10
Proof. Given P P ' , as in the figures, with . 2 / , a P A a AP = ' = Take PD so that
90 ) ( < = Z a c APD m and take . P at AP E P ' ' If PD fails to intersect , E P' then
obviously . 90 ) 2 / ( < a c

Suppose, then, that PD
does intersect
E P' , at a point F. Let G be
any point such that P F G. Then G P A ' Z is acute. Now
I) AB cannot intersect G P'
except perhaps in a point of ; G P' the reason is that all other
points of G P'
lie on the wrong side of PG . And
II) AB does not contain P'
or G. Finally
III) AB does not contain a point between P'
and G; if so, it would follow from the
postulate of Pasch that ABintersects , FG or F P' which is false.
Therefore G P'
does not intersect AB, and , 90 ) 2 / ( < a c which was to be proved.


Theorem 1.13.6. If , 90 ) (
0
< a c for some
0
a then 90 ) ( < a c for every a
Proof. For each , n let
n
n
a
a
2
0
=

By induction based on Theorem 1.13.5, we have
90 ) ( <
n
a c for every . n

Suppose now that 90 ) ( = b c for some b. Since
, 0
lim
=
n n
a

We have b a
k
< ; for some k. Thus , ) ( ) ( b c a c but b a
k k
< < and this contradicts Theorem
1.13.4.

This theorem clarifies the meaning of the parallel postulate; it tells us that the situation
described in the postulate holds either always or never.

Theorem 1.13.7 (The All or - None Theorem) If parallels are unique for one line and one
external point, then parallels are unique for all lines and all external points.
Proof. Given points P, P'
and lines L, , L' with a AP= and , a P A ' = ' ' as in the figure 1.13.11.

Fig 1.13.11

It is now easy to see that each of the statements below is equivalent to the next (i.e. the
following statements are equivalent).
1. There is only one parallel to L through P.
2. . 90 ) ( = a c
3. . 90 ) ( = ' a c
4. There is only one parallel to L'
through P'

Therefore (1) and (4) are equivalent, which was to be proved.

Thus we can state our two possible parallel postulates in seemingly weak but actually quite
adequate forms.
(I) Euclidean. For some line and some external point, parallels are unique.
(II) Lobachevskian. For some line and some external point, parallels are not unique.

If the formula 180 = Z + Z + Z C m B m A m holds for even one triangle, then it holds for all
triangles; if even one pair of triangles are similar without being congruent, then the geometry
is Euclidean; and so on.

1.14 Open Triangles and Critically Parallel Rays

Given rays , AB , PD and the segment , AP no two of these figure being collinear. Suppose
that B and D are on the same side of , AP and that . PD AB
Then AB PA PD is called an open triangle, and is denoted by . DPAB A

Fig 1.14.1
Here, when we write , PD AB we mean that the lines are parallel in the usual sense of not
intersecting one another.
Suppose now that DPAB A is an open triangle and every interior ray of APD Z
intersects AB:

Fig 1.14.2

We then say that PD is critically parallel to , AB and we write . AB PD Here the single
vertical stroke is supposed to suggest that PD is parallel to , AB with no room to spare.
Note that PD and AB do not appear symmetrically in this definition. Thus if , AB PD it
does not immediately follow that . PD AB Note also that the relation AB PD (as we have
defined it) depends not only on the directions of the two rays, but also on the initial points:

Fig 1.14.3
Thus, if AB PD (as we have defined it) depends not only on the initial points:
Thus, if , AB PD we cannot conclude immediate that B A D P ' ' . We shall see, however, in
the next few theorems, that the conclusion is true

Theorem 1.14.1 If AB PD , and C-P-D , then . AB CD

Fig 1.14.4
Proof. Let CE be an interior ray of , ACD Z and suppose that CE does not intersect AB. By
the exterior angle theorem, we know that ACD APD Z > Z .
Therefore there is an interior ray PF of APD Z such that . DCE DPF Z ~ Z

Therefore . CE PF Therefore PF does not intersect AB, because these rays lie on
opposite sides of CE. This contradicts the hypothesis . AB PD

Theorem 1.14.2 If , AB PD and P-C-D, . AB CD

Fig 1.14.5
We give the proof briefly. Suppose that there is an interior ray ACD of CE Z such that CE
does not intersect AB. Let F be any point of , C CE and take G so that
P-F-G. Then
(1) F is in the interior of APC Z
(2) PF does not intersect AB;
(3) FG does not intersect AB;
(4) PF does not intersect AB.
Statement (1) and (4) contradict the hypothesis . AB PD
Two rays R and R'
are called equivalent if one of them contains the other. We then write R
~ R'
. Obviously the symbol ~

represents an equivalence relation. Fitting together the
preceding two theorems, we get:


Theorem 1.14.3 If R and R and AB R ' , are equivalent, then . AB R'
Somewhat easier proofs show that the relation AB PD depends only on the equivalence
class of . AB We leave these proofs to you.

Theorem 1.14.4.
1 2 1
, R R R ' ~
2 1
R and R ' ~
2
R then
2 1
R R ' ' .
Given , AB PD let C be the foot of the perpendicular from P to , AB and let a PC =

Fig 1.14.6
Then CB PD (Providing, of course, that B is chosen so that A-C-B, as in the figure).
Therefore . ) (a c CPD m = Z Now on the side of PC that contains B there is only one ray PD

for which ). (a c CPD m = Z Thus we have:

Theorem 1.14.5. The critical parallel to a given ray, through a give external point, is unique.
Two open triangles are called equivalent if the rays that form their sides are equivalent. An
open triangle DPAB A is called isosceles if . A P Z ~ Z

Fig 1.14.7

Theorem 1.14.6. If AB PD , then DPAB A is equivalent to an isosceles open triangle which
has P as a vertex.

Fig 1.14.8
Proof. Since AB PD , the bisecting ray of APD Z intersects AB in a point Q. By the
crossbar theorem, the bisecting ray of PAB Z

intersects PQ at a point R. Let S, T, and U be
the feet of the perpendicular from R to . , , AP and AB PD

Then
. RS RU and RT RU = = Therefore . , RTS RST and RT RS Z ~ Z = Hence (by addition or
subtraction) DSTB and BTS DST A Z ~ Z ; is isosceles. To make P a vertex, we take V on the
ray opposite to , TB such that TV = SP.

Fig 1.14.9

Theorem 1.14.7. Critical parallelism is a symmetri relation. That is, if , AB PD then . PD AB
Proof. By Theorems 1.14.4 and 1.14.6, we may suppose that DPAB A is an isosceles open
triangle:


Fig 1.14.10
Let AE be any interior ray of . PAB Z Let PF be an interior ray of , APD Z such that
. BAE DPF Z ~ Z Then PF intersects AB at a point Q. It follows that AE intersects PD

at
the point S where PS = AQ.

Theorem 1.14.8. If two nonequivalent rays are critically parallel to a third ray, then they are
critically parallel to each other.
Restatement. If EF and AB and EF CD CD AB , , are not equivalent, then . EF AB
1) Suppose that EF and AB lie on opposite sides of . CD Then AE intersects , CD and by
Theorem 1.14.4 we can assume that the point of intersection is C.

Fig 1.14.11
Let AG be any interior ray of . EAB Z Then AG intersects CD at a point H. Take I so that C-
H-I and take J so that A-H-J. Then EF HI , by Theorem 1.14.4; and HJ is an interior ray of
EHI Z Therefore HJ intersects EF at a point K. Therefore AG

intersects

, EF which was
to be proved.

2) If EF and CD are on opposite sides of , AB then the same conclusion follows.
Here we may suppose that , A EC AB = for the same reasons as in the first case. Through E
there is exactly one ray F E ' critically parallel to , AB by the result in case (1), . CD F E ' Since
critical parallels are unique , AB EF and EF F E = ' which was to be proved.

Fig 1.14.12

Given three nonintersecting lines, it can easily happen that every two of them are on the same
side the third. Therefore the conditions EF CD CD AB , are not enough for our
purpose; to get a valid proof, we need to use the full force of the hypothesis
. , EF CD CD AB We shall show, under these conditions, that
(3) some line intersects all three of the ray . , , EF CD AB (Surely this will be enough.)

Fig 1.14.13
If A and E are on opposite sides of , CD then AE intersects , CD and (3) follows. Suppose,
then, that

a) A and E are on the same sides of CD . If A and D are on the same side of EC, then CA
is an interior ray of , C Z so that CA intersects , EF and(3) follows. If A lies on CE then
(3) holds. We may therefore suppose that
b) A and D are on opposite sides of CE. Therefore AD intersects CE at a point G. Take H
so that C-D-H. Then . AB DH By the exterior angle theorem, . C HDA Z > Z Therefore
there is an interior ray DI of HDA Z such that . C HDI Z ~ Z Then CE DI but DI
intersects AB at a Point . J

Now CE intersects AD at G. Therefore CE interests another
side of . ADJ A Since CE doe not intersect CE DJ , intersects AJ at a point K. Now (3)
follows; the line that we wanted is . CE

Exercises
By the interior of an open triangle , DPAB A we mean the intersection of the interiors of
A and P Z Z . If a line intersects the interior of an open triangle, does it follow that the line
intersects one of the sides? Why or why not?
1. Same question, for the case where . AB PD
2. In a Euclidean plane, if a line intersects the interior of an angle, does it follow that the line
intersects the angle?


CHAPTER TWO
Euclidean Geometry

Though in schools most students learn plane Geometry/Euclidean geometry, there are actually
many different types. These different types were developed by other mathematicians who
developed theories and research that may have contradicted the work of other.
But here our concern is Euclidean geometry which is based on rules called postulates as stated
below. It is different from other geometries, such as Absolute/ Neutral geometry, hyperbolic
geometry, Elliptic geometry and the like where no parallel lines exist, because of the parallel
postulate.
Euclids Axioms for geometry: The Euclidean geometry is based on the following postulates
Postulate 1: We can draw a unique line segment between any two points.
Postulate 2: Any line segment can be continued indefinitely.
Postulate 3: A circle of any radius and any center can be drawn.
Postulate 4: Any two right angles are congruent.
Postulate 5: Let and mbe two lines cut by a transversal in such a way that the sum of
the measures of the two interior angles on one side of t is less than 180. Then
and mintersect on that side of t .

2.1 Euclidean parallel postulate and some consequences
The parallel postulate was the most controversial of Euclids postulate for geometry.
Many mathematicians felt that it should be possible to be deduced the parallel postulate from
Euclids other postulates. It was later proved to be impossible to deduce the parallel postulate
from the other postulates, efforts to do so led the invention of various non- Euclidean
geometries in which the parallel postulate is violated. Here below it will be given the
statement of Euclidean parallel postulate and some of its consequences.


Definition 2.1: Two distinct lines are parallel if they have no points in common. We also say
that any line is parallel to itself.
The word parallel simply means that two lines have no points in common or are equal. It
doesnt say any thing about being in the same direction, or being equidistance from each
other, or any thing else.

Euclidean parallel postulate: For every line and for every point p that does not lie on
there is exactly one line m such that p is on m and m// . The parallel postulate in its
equivalent form:
[P](Play fairs Axiom) For each point P and each line , there exists at most
one line through P parallel to . Indeed, in Euclids development of geometry;
this is not an Axiom, but a theorem that can be proved from the axioms.
However, some mathematicians like to take the statement [p] as an axiom
instead of using Euclids parallel postulate. As a result, it is very important to
explain in what sense we can say that Euclids parallel postulate is equivalent
to Playfairs axiom. Since the parallel postulate plays such a special role in
Euclids geometry, let us make a special point of being aware when we use
this postulate, and which theorems are dependent on its use. Let us recall
neutral geometry the collection of all postulates and common notions except
parallel postulate together with all theorems that can be proved without using
parallel postulate. If we take neutral geometry, and add back the parallel
postulate, then we recover the ordinary Euclidean geometry and we can prove
[p] as a theorem.

Euclid has proved, using the parallel postulate, that the angle sum in triangle is always two
right angles. This property of triangles is equivalent to the parallel postulate, i.e., one can also
prove that the converse implication, that if the angle sum is assumed to be two right angles,
then the parallel postulate follows. Thus, proving the parallel postulate is equivalent to
proving the angle sum theorem.


Theorem 2.1.1: Given two lines and a transversal. If the lines are parallel, then each pair of
alternate interior angles is congruent.

Figure 2.1
Proof. There is exactly one line 1
'
L through P, for which the alternate interior angles are
congruent, and by Theorem in chapter one, we have
2 1
L L
'
. Since there is only one such
parallel line, we have .
2 1
L L =
'
Therefore . 2 1 Z ~ Z which was to be proved.

The proof of the following theorem is entirely analogous.
Theorem 2.1.2: Given two lines and a transversal. If the lines are parallel, then each pair of
corresponding angles is congruent. The inequality 180 s Z + Z + Z C m B m A m now becomes
an equation.
Theorem.2.1.3: In any triangle ABC A we have 180 = Z + Z + Z C m B m A m
.

Figure 2.2
Proof. Let L be the parallel to AC though B. Let D and E be points of L such that D-B-E
and such that D and A are on the same side of BC. Then DBC m B m m Z = Z + Z2

and
0
180 1= Z + Z m C DB m . Therefore, 180 2 1 = Z + Z + Z m B m m . By Theorem 2.1
; 2 1 A m m and C m m Z = Z Z = Z

therefore 180 = Z + Z + Z C m B m A m . This was to be
proved.

The following theorems are an immediate consequence.
Theorem 2.1.4: The acute angles of a right triangle are complementary
Theorem 2.1.5: Every Saccheri quadrilateral is a rectangle

Figure 2.3
Proof: By Theorem 1, . 4 2 Z ~ Z since AB= DC and AC=AC, it follows that . DCA BAC A ~ A
Therefore , D B Z ~ Z is a right angle. The proof that C Z is a right angle is obtained merely
by permuting the notation. Thus we have finally shown that rectangles exist.
Note that in this proof we are using a figure to explain the notation. If the reader (or the
writer) sees no other way to explain, say, the idea of alternate interior angles, then it is
worthwhile to fight our way through the problem as we did in the previous chapter. But once
we have done this, we have earned the right to speak in the abbreviated language of pictures.
A quadrilateral is a trapezoid if at least one par of opposite sides are parallel (It is sometimes
required that the other pair of opposite sides be nonparallel, but this is artificial, just as it
would be artificial to require that an isosceles triangle be nonequidrilateral.) If both pairs of
opposite sides of a quadrilateral are parallel, then the quadrilateral is a parallelogram. If two
adjacent sides of a parallelogram are congruent, then the quadrilateral is a rhombus. The
proofs of the following theorems are omitted. (They are not much harder to write than to
read.)

Theorem 2.1.6: For any triangle, the measure of an exterior angle is the sum of the measures
of its two remote interior angles.

Theorem 2.1.7: In a plane, any two lines parallel to a third line are parallel to each other.

Theorem 2.1.8: If a transversal is perpendicular to one of two parallel lines, it is
perpendicular to the other.

Theorem 2.1.9: Either diagonal divides a parallelogram into two congruent triangles. More
precisely: If ABCD is a parallelogram, then . CDA ABC A ~ A

Theorem 2.1.10: In a parallelogram, each pair of opposite sides is congruent.

Theorem 2.1.11: The diagonals of a parallelogram bisect each other. That is, they intersect at
a point which is the bisector of each of them. Thus the proof must begin with a proof that the
diagonals intersect each other.

2.2 Equivalent form of the Euclidean parallel postulate
In this section we consider some statements that are equivalent to Euclids parallel postulate.
When we say that two statements are equivalent in this sense we mean that if we add either
statement to the axioms of neutral Geometry, we can prove the other statement. It does not
mean that the two statements are precisely logically equivalent.

Euclidean Parallel Postulate: For every line and for every point P that does not lie on
there is exactly one line m such that m pe and m .

Equivalent Axiom (Euclids Fifth Postulate) 2.2.1: Let and mbe two lines cut by a
transversal in such a way that the sum of the measures of the two interior angles on one side
of t is less than 180. Then and mintersect on that side of t .

Euclids Fifth postulates states that if 180 < + | o , then intersects mat a point C that is
on the same of t as oand | .

Figure 2.4
Proof: (Euclids 5
th
postulate is equivalent to the Euclidean Parallel Postulate.)
( ) [The Euclidean Parallel Postulate Euclids 5
th
Postulate.] Let | o, , , , t m be as
indicated in figure 2.2, i.e., construct the lines , , m and n as shown;
then 180 < + | o (why?). There is a line n through B such that 180 = +o (by the
protractor postulate). By the linear pair theorem, then 180 = +o c and 180 = +o .
Hence,
o c o + = + 180 180 ) ( 360 o + = 180 360 = 180 = . (*)
Thus both pairs of non-alternating interior angles formed by t sum to 180. By
assumption 180 < + | o Substituting equation (*) gives 180 180 < + | c ,
c | < . In particular, since , c | = then n m = . Since n , 180 o o = = (alternate
interior angles theorem). Since n m = this means mis not parallel to (this is
because we are assuming the Euclidean parallel postulate, that there is only one line
through B that is parallel to ). Since mis not parallel to , they intersect at a point
C, and there must be such a point C on the same side of t as the angles o and | .
This is Euclids 5
th
postulate.
( ) : [Euclids 5
th
postulate the Euclidean Parallel Postulate.]
See proof that Euclids Fifth postulate implies the Euclidean parallel postulate.


Figure 2.5
Assume Euclids 5
th
postulate. Let be a line and Pbe a pint such that e p .Drop a
perpendicular line from P to , and call the foot of the line Q. Construct mthrough
Psuch that PQ m . By the alternate interior angles theorem, m . Assumer m n =
is a second line through P such that n . Then PQ is a transversal to nand .
Since m n = , the interior angles 90 = and 90 = o . Since they form a linear
pair, 180 = +o .Hence one of o , is less than 90 and the other is greater than
90.By Euclids fifth postulate, lines n and meet on whichever side of PQ the
smaller of angles and o lies (e.g., on the same side as in figure 2.5). Thus,
to parallel not is n .Hence there is only one line through P that is parallel to .
Hence, the Euclidean parallel Postulate follows from Euclids Fifth Postulate.

Equivalent Axiom (converse of Alternate Interior Angles Theorem) 2.2.2:
If two parallel lines are cut by a transversal, then both pairs of alternate interior angles are
congruent.
Proof. (The converse of the Alternate Interior Angles Theorem is Equivalent to the Euclidean
Parallels Postulate).
( ) [Assume that the converse of the Alternate Interior Angles Theorem is true and show
that the Euclidean parallel Postulate follows.]

Figure 2.6
Let be a line and let p be a point such that e p . From result we have in previous
section, we can drop a perpendicular line t from p to and then construct a second
line mthrough p such that m t . By the alternate interior angles theorem m
(because a pair of interior angles is congruent; it just happens that by construction, all
of the interior angles are all right angles, figure 2.4). Suppose there is another line
m n = such that n p e and n . Thent is a transversal to and n.Then by the
alternate interior angles theorem we have o = .Since 90 = , then 90 = o , and
consequently t n .By the uniqueness part of the protractor postulate, there is only
one line through P that is perpendicular tot . Hence m n= .Thus there is only one line
parallel to through P. Thus the proof follows.
( ) : [Assume the Euclidean Parallel Postulate and show that that the converse to the
Alternate Interior Angles Theorem follows.] Let and m be parallel lines that are cut
by a transversal t (hypothesis). Figure 2.4 shows that the converse of the alternate
interior angles theorem is equivalent to the Euclidean Parallel Postulate.

Figure 2.7

We need to show that | o = and o = .Figure2.5 shows that the Euclidean parallel
postulate implies the converse of the Alternate Interior angles.

The next axioms expresses the relationships between parallel and perpendicular lines in
Euclidean geometry and their proofs are left as exercises.
Equivalent Axiom 2.2.3: If //m and t is a transversal such that t , then t m .
Equivalent Axiom (transitivity of parallels) 2.2.4: If //m and m//n, then either =n or //
n.

Transitivity of parallelism has the following two corollaries and their proofs are left as
exercises.
Corollary2.2.1: The diagonals of a parallelogram are congruent.
Corollary 2.2.2: The diagonals of a parallelogram bisect one another.

The next theorem is another most important fact in Euclidean geometry where the Euclidean
parallel postulate is equivalent to proving the angle sum theorem.
Equivalent form :( Angle-Sum theorem) 2.2.5: If the angle sum of every triangle is 180
0
,
then Euclids Parallel Postulate holds.

2.3 Euclidean Parallel Projection
We know that the perpendicular from a point to a line always exists and is unique. Further
more, the parallel projection theorem is one consequences of Euclidean parallel postulate. We
will discuss the general notion of parallel projection in plane as follows. And we also show
that parallel projection preserves betweeness, congruence and ratios, let us first consider the
special case indicated in the following figure, and treated in next the theorem.

Figure 2.8
Thus, given two lines L, L'
in the same plane, we can define the vertical projection of L
into L'
. This is the function L L f ' :

under which to each point P of L there corresponds the
foot ) (P f P = ' of the perpendicular from P to L'
. In fact, the vertical projection can be
defined equally well for the case where the lines are not necessarily coplanar; and the
definition is exactly the same. This degree of generality, however, will not concern us. Note
also that the existence and uniqueness of the vertical projection do not depend on the parallel
postulate. We do, however, need this postulate to define and investigate the more general
notion of parallel projection. Under this more general scheme, instead of following the
perpendicular, to get from P to , ) (P f P = ' we proceed in any direction we want, providing,
however, that we always go in the same direction for every point P of L.
More precisely, the definition of parallel projections is as follows. Given two lines L
,
L'
and a
transversal T. (By definition of a transversal, this means that all three of our lines are
coplanar.) Let T intersect L and L'
in points Q and Q ' , respectively.

Figure 2.9

Let Q Q f ' = ) ( . For every other point P of L, let
p
T be the line through P, parallel to T.
1) If
p
T were parallel to L'
, then it would follow that L T ' , which is false, because T is a
transversal to L and L'
in at least one point P'
.
2)
If , L T
p
' = it follows that , T L' which is false. Therefore
p
T intersects L'
in at most
one point P'
. This defines a function . : L L f '

The function f is called the projection of L onto L in the direction T.

Theorem 2.3.1: Every parallel projection is a one-to-one correspondence.
Proof: Given
L L f ' :
,
the projection of L onto L'
in the direction T. (See Fig. 2.9).
Let g be the projection L'
onto L in the direction T. Obviously g reverses the action of f.
That is, if ). ( . ) ( P f P then P g P = ' ' =

. :
1
L L g f ' =

Therefore f is a one-to-one
correspondence , L L ' which was to be proved.(Another way of putting it is to say that
every point ) (P f is L of P = ' ' for one and only one point P of L.

Theorem (The parallel projection theorem) 2.3.2: Let , m, and n be distinct parallel lines;
let t and t be transversals that cut these three lines at points A, B, C, and A, B, C
respectively.

Figure 2.10


Theorem 2.3.3: Parallel projections preserve betweeness.
Restatement. Let L L f ' : be a parallel projection. If P-Q-R on L, then L on R Q P ' ' ' '

Figure 2.11

Here, of course, ). ( , ) ( , ) ( R f R and Q f Q P f P = ' = ' = '
PROOF:. Let
n Q p
T T T , , be as in the definition of a parallel projection, so that
n Q p
T T T // // .
Then R and R'
are on the same side of
Q
T , because R R ' does not intersect
Q
T . Similarly, P
and P'
are on the same side of
Q
T . But P and R are on opposite sides of
Q
T , because P-Q-R,
and P'
and R'
are on opposite sides of
Q
T (why?). Therefore R P ' ' intersects
Q
T in a point X.
since , L T
Q
' = there is only one such point of intersection. Therefore X=Q' . Therefore Q'
lies on , R P ' ' and R Q P ' ' ' , which was to be proved.

Theorem 2.3.4: Parallel projections preserve congruence.
Restatement. Let L L f ' : be a parallel projection. If CD AB ~ on L., then D C B A ' ' ~ ' '
on . L'

Proof:
(1) If L L ' , then B A and AB ' ' are opposite sides of a parallelogram. From this, it follows
that B A AB ' ' ~ (why?). Similarly . D C CD ' ' ~

Therefore , D C B A ' ' ~ ' ' as desired.
(2) Suppose that L and L ' are not parallel, as in fig.11.7. Let V be the line through A,
parallel to L'
, intersecting T
D
at E
. Let W be a line through C, parallel to L,

intersecting
F at T
D
.
Now L and W V , // is a transversal. For appropriate choice of
the notation for C and D, BAE Z = Z1

and DCF Z = ' Z1 are corresponding angles.
(This is true for the case shown in the figure; if it isnt true; we interchange the letters
C and D). We have 1 1 Z = ' Z .

Figure 2.12

For the same reasons . 2 2 ' Z ~ Z since CD AB ~ by hypothesis, it follows by ASA that
. CDF ABE A ~ A
Therefore . CF AE~ But ., D C CF and B A AE ' ' ~ ' ' ~ because these segments are opposite
sides of parallelograms. Therefore , D C B A ' ' ~ ' ' this was to be proved.
Consider now three lines with two common transversals, like this:

Figure 2.13

It ought to be true that .
BC
AB
C B
B A
=
' '
' '

In the style of our previous theorems, we could convey this by saying that parallel projections
preserve ratios. In fact, this is true, but the proof is omitted.
The theorem is worth working for; it is the foundation of the whole theory of similarity for
triangles. The proof will depend on the Euclidean parallel postulate, as one might expect: if
parallels are not unique, then parallel projections are not even well defined.

2.4 Basic similarity Theorem
Here we will revise the preliminary notions that could be used in this section like ratio and
proportion while we study similarity of triangles.

Definition 2.4.1: A comparison of the magnitudes of two quantities of the same kind in the
same unit is called a ratio. It is usually expressed as quotient of two numbers. For instance, if
we are given lengths of two line segments as AB =16cm and DE=7cm, then the ratio of their
lengths is 16:7.

Definition2.4.2An equality of two ratios is called a proportion.
Remark: A proportion is usually expressed as
d
c
b
a
= or a: b = c: d
The constant ration = k
d
c
b
a
= is called the proportionality constant
(common value of each ratio.)
If three quantities a, b, c are such that
c
b
b
a
= , then b is called a mean
proportional between a and c. Thus if b is the mean proportional between a
and c, then b
2
= ac.


Ratio of segments of a line: Let P be a point on line segment AB.
If APB, then AB is said to be divided internally at P in the ratio AP: PB. (See figure 2.14i)
If PAB or ABP, then AB is said to be divided externally at P in the ratio AP: PB.
(See Fig 2.14 ii &iii)

Figure 2.14

Theorem: 2.4.1 If a line parallel to one side of a triangle intersects the other two sides (at
points that divides the sides internally), then it divides each of these sides in segments
which are proportional.
Let us investigate some important proportions that can be deduced form this theorem.
In Fig. 2.15, let BC || DE . Then it follows that ) 1 .........( ..........
EC
AE
DB
AD
=

Figure 2.15
But
EC
EC AE
DB
DB AD
EC
AE
DB
AD
EC
AE
DB
AD +
=
+
+ = + = 1 1

Also from (1), it follows that

AE
EC
AD
DB
=
AE
EC
AD
DB
+ = + 1 1

AE
EC AE
AD
DB AD +
=
+

Thus if D and E are two points on sides AB and AC of ABC A such that BC DE||
then
EC
AE
DB
AD
= ,
EC
AC
DB
AB
= and
AE
AC
AD
AB
= .

Theorem (Basic Similarity Theorem) 2.4.2: If
1
,
2
and
3
are three parallel lines, with
common transversals mand n, then
DE
EF
AB
BC
=

Figure 2.16
Here
1
,
2
and
3
are parallel lines, with common transversals m and n. We want to prove
that
DE
EF
AB
BC
=


Proof: Let m and n be transversals to
3 2 1
, , where
1
3

To show that
EF
DE
BC
AB
= . (See Fig 2.16)
Join A with F and apply theorem 2.4.1 in ACF A and FDA A , to get
XF
AX
BC
AB
=
and
ED
FE
XA
FX
= (X is the point of intersection of AF and
2
).
But

EF
DE
BC
AB
=

Note: The theorem stated above is one of the basic theorems in proving similarity of two
triangles.

Theorem 2.4.3: If M and N are two points on sides XY and XZ of XYZ A , respectively
such that YZ MN || , then
XZ
XN
YZ
MN
XY
XM
= =
Proof: By theorem 2.4.1, we have
XZ
XN
XY
XM
= .

Figure 2.17
Fig. 2.6.4

Now since N is not on , XY there exists a unique line through N parallel to XY . Let this
line intersect YZ at O. (see fig 2.17). Then by theorem 2.4.1
YO
YZ
XN
XZ
= .
That is ,
YZ
OY
XZ
XN
= . But NM OY (why?).
Hence,
YZ
MN
XZ
XN
= . Therefore,
YZ
MN
XZ
XN
XY
XM
= = .

Theorem2.4.4: If points D and E are respectively on sides AB and AC of ABC A such
that
AC
AE
AB
AD
= , then BC DE|| .

Figure 2.18
Proof: Suppose DE is not parallel to BC . Then by parallel axiom there exists a point F on
AC different from E such that BC DF || . Hence
AC
AF
AB
AD
= by Theorem 2.4.1
But from the hypothesis of the theorem we have
AC
AE
AB
AD
= . Thus,
AC
AE
AC
AF
= and
hence AF = AE. This in turn implies AE AF , contrary to axiom of segment construction
as E F = . Therefore the supposition DE is not parallel to BC is wrong.
Consequently BC DE|| .


2.5 Similarities between Triangles
We will discuss the mathematical notion of similarity which describes the idea of change of
scale that is found in such forms as map making, perspective drawings, photographical
enlargement and indirect measurements of distance. In this section we will mainly discuss
some important theorems that could be used to prove similarity of triangles. The proofs of
similarity theorems are based on the use of basic similarity theorem. Recall from high school
that geometric figures are similar when they have the same shape, but not necessarily same
size.

Definition 2.5.1 2.4.1: Two triangles A ABC and A DEF are said to be similar, written as
A ABC
~
A DEF, if and only if i) all three parts of corresponding angles are congruent
and ii) lengths of all three pairs of corresponding sides are
proportional.

Note: To establish similarity of triangles, however, it is not necessary to establish congruence
of all pairs of angles and proportionality of all pairs of sides. It is also important to note that
triangle similarity do require Euclids parallel postulate.

Note: 1) If ABC A is similar to DEF A , we denote this by ABC A ~ DEF A .
2) Similar triangles should always be named in such way that so that the order of the
letters indicates the correspondence between the two triangles.
3) ABC A ~ DEF A if and only if (i) F C E B D A Z Z Z Z Z Z , , and
(ii)
FD
CA
EF
BC
DE
AB
= =
4) The common value of each ratio in (ii) is called proportionality constant.
5) Intuitively speaking, two triangles are similar if they have the same shape, although
not necessary the same size. It looks as if the shape ought to be determined by the
angles alone, and this is true.


Theorem2.5.1 The AAA Similarity Theorem.
Given a correspondence between two triangles. If corresponding angles are congruent, then
the correspondence is a similarity.
Restatement: Given ABC A , DEF A and a correspondence ABC DEF.
If A Z ~ D Z , B Z ~ E Z , and C Z ~ F Z , then ABC A ~ DEF A .
Proof: Let E' and F' be points of AB and AC shown in Fig 2.5.1.
By SAS, we have ABC A ~ DEF A .
Therefore F AE
'
Z ~ E Z . Since B E Z ~ Z , we have
' '
F AE Z B Z ~ ;
Thus
' '
F E //BC, and A, F
'
,C correspond to A, E
'
, and B under a parallel projection.

Since parallel projection preserve ratios, we have
AC
AF
AB
AE
' '
=

Figure 2.19
In exactly the same, merely changing the notation, we can show that
BC
EF
AC
AF
=
'

Therefore,
AB
AE
AC
AF
BC
EF
' '
= = .


Hence, corresponding angles are congruent and corresponding sides are proportional.
By definition, ABC A ~ DEF A .

Theorem 2.5.2: AA similarity theorem. If two angles of one triangle are congruent to the
corresponding two angles of another triangle, then the triangles are similar.
Proof: Let A ABC and A XYZ be two triangles such that X A Z Z and Y B Z Z . We
need to show that ABC A ~ XYZ A . Since X A Z = Z and Y B Z = Z , then
Z C Z Z .So it remains to show that the corresponding sides are proportional.

Figure 2.20
If XY AB , then ABC A ~ XYZ A (why?).
If Z C Z Z XY AB
/
, then either AB < XY or XY < AB. With out loss of
generality assume that XY < AB. Then there exists a point D on AB such that A-D-B
and XY AD . By axiom of angle construction there exists a point F on the half plane
determined by AB containing C such that XYZ ADF Z Z .

Figure 2.21

Since ABC ADF XYZ ABC Z Z Z Z , by transitivity. Hence BC DF ||
(why?).
Since ray DF does not pass through the vertices of ABC A and does not
intersect BC, it must intersect side AC at some point E. Thus, XYZ ADE A A by
ASA. Hence YZ DE XY AD , and ZX EA . But BC DF || as
DF E e and BC DF || . It then follows from theorem 2.4.3 that this in turn
implies
CA
ZX
BC
YZ
AB
XY
= = . Therefore ABC A ~ XYZ A .

Theorem 2.5.3: SAS similarity theorem. Given a correspondence between two triangles. If
two pairs of corresponding sides are proportional, and the included angles are
congruent, then the correspondence is a similarity.
Proof: Given two triangles ABC A and PQR A such that P A Z Z and
PR
AC
PQ
AB
= .

Figure 2.22
To show that ABC A ~ PQR A . Since it is given that P A Z Z , it is sufficient to show that
Q B Z Z . Let D and E be two points on AB and ACrespectively such
that PQ AD and PR AE (This is possible by axiom of segment construction).
Then PQR ADE A A by SAS. Hence PQR ADE Z Z

Figure 2.23
From PQ AD
PR
AC
PQ
AB
= , and PR AE it follows that
AE
AC
AD
AB
= .
Thus, BC DE|| by theorem 2.4.4 and hence ABC ADE Z Z .
So PQR ABC Z Z , by transitivity. Therefore ABC A ~ PQR A by AA similarity as
P A Z Z and Q B Z Z .

Theorem 2.5.4: The SSS similarity theorem. If two triangles are such that the
corresponding sides are proportional, then the two triangles are similar.
Restatement: Given ABC A , DEF A , and a correspondence ABC DEF.
If
DF
AC
EF
BC
DE
AB
= = , then ~ ABC A DEF A .
Proof: Let E
'
be the point of AB. Let be the line through E
'
parallel to BC. If // AC, then
BC//AC, which is false. Therefore intersects AC at a point F
'
. (In the figure, b=AC, not
AF) Now ,
' '
B F AE Z ~ Z because these are corresponding angles; and A A Z ~ Z .
Therefore, ABC F AE A A ~
' '
. Hence
AC
AF
BC
F E
AB
AE
' ' ' '
= = .
Therefore
' ' '
'
'
'
F E
BC
AF
AF
AE
BE
= = .
Thus in the figure AE
'
DE ~ and ABC A ~ DEF A by SSS (Why?)
Therefore ABC DEF A A ~ .


Figure 2.24

Theorem2.5.5: The bisector of an angle of a triangle divides the opposite side into segments
which are proportional to the adjacent sides.
Restatement: If in ABC A , AD is the bisector of BAC Z where D is point on BC ,
then
CD
BD
AC
AB
=
Proof: Left as an exercise. (Hint: Draw a line through C parallel to AD)

Theorem2.5.6: If an external bisector of an angle of a triangle intersects the line containing
the opposite side, then the point of intersection divides the opposite side externally
into segments which are proportional to the adjacent sides.

Figure 2.25

Proof: Let the external bisector of angle X Z of XYZ A intersect the opposite side YZ at W
(Externally).Draw a line through W parallel to XZ . Then and XY are not
parallel (why?). So they intersect at some point, say T. TWX ZXW Z Z since they
are alternate interior angles.
Also TXW ZXW Z Z so we have TWX TXW Z Z by transitivity. Thus, XT = TW.
Since TW XZ || , we have
TW
YT
TX
YT
WZ
YW
= = (1)
From YXZ YTW A A ~ it follows that
XZ
TW
YZ
YW
YX
YT
= = , which implies that

XZ
YX
TW
YT
= ... (2)
From (1) and (2) we conclude
WZ
YW
XZ
YX
= . This completes the proof.

Illustration: In fig.2.4.7 BE and DC are angle bisectors of CBF Z and ACB Z
respectively. If AD = 21cm, AC = 30 cm and BC = 20 cm, find DB and EC.

Figure 2.26
Solution: Since CD is the bisector of B C A
,
DB
AD
BC
AC
= (why?).
14
30
20 21
=
=
AC
BC AD
DB

Since BE is the external bisector of angle B Z of
EC
AE
BC
AB
ABC = A , (why?).
Thus
EC
EC +
=
30
20
35
. 40 = EC .

Theorem 2.5.7: In a right triangle, if an altitude is drawn to the hypotenuse, then
(i) The triangle is divided into two similar right triangles and, which are also similar
to each other.
(ii) The altitude is the mean proportional between the segments of the hypotenuse.
(iii)Either leg is the mean proportional between the hypotenuse and the segment of the
hypotenuse adjacent to the leg.
Proof: Let ABC A be a right triangle with right angle at C and CD be altitude to the
hypotenuse AB.

Figure 2.27
i) Then ACD Z and CBD Z are congruent as they are complements of the same
angle CAB Z . Similarly BCD CAD Z Z . Thus ACD ABC A A ~ and
CBD ABC A A ~ by AA. Then ACD A ~ CBD A .
ii) From ACD A ~ CBD A , it follows that,
CD
AD
BD
CD
= . That is CD
2
= AD. BD

iii) From ABC A ~ ACD A if follows that,
AD
AC
AC
AB
= .
That is AC
2
= AB. AD. From ABC A ~ CBD A , we have

BD
BC
CB
AB
= . That is BC
2
= AB . BD

2.6 Pythagorean Theorem

In mathematics, the Pythagorean Theorem is a relation in Euclidean geometry among the
three sides of a right triangle (right-angled triangle).
In terms of areas, it states that in any right triangle, the area of the square whose side is the
hypotenuse (the side opposite the right angle) is equal to the sum of the areas of the squares
whose sides are the two legs (the two sides that meet at a right angle).
Euclids version of Pythagorean Theorem: The sum of the areas of the two squares on the
legs (a and b) equals the area of the square on the hypotenuse(c).
The theorem can be written as an equation relating the lengths of the sides a, b and c, often
called the Pythagorean equation
2 2 2
c b a = + where c represents the length of the
hypotenuse, and a and b represent the lengths of the other two sides.
These two formulations show two fundamental aspects of this theorem: it is both a statement
about areas and about lengths. The Pythagorean Theorem has been modified to apply outside
its original domain. A number of these generalizations are found in more advanced
mathematics courses including extension to many-dimensional Euclidean spaces, to spaces
that are not Euclidean, to objects that are not right triangles, and indeed, to objects that are not
triangles at all, but n-dimensional solids.

Theorem 2.6.1 (Pythagoras theorem): Let A ABC be a right triangle with right angle at
vertex C. The square of the hypotenuse of a right triangle is equal to the sum of the
squares of the other two sides.

Figure 2.28
Proof: Let ABC A be a right triangle with right angle at C.
To show that
2 2 2
BC AC AB + =
. Draw altitude
CD
to
AB
.
Then from theorem 2.4.1 (iii) we have, AC
2
+ BC
2
= (AB) . (AD) + (AB) . (BD)
= AB (AD + BD)
= AB.AB as A D B

2 2 2
BC AC AB + =

Theorem 2.6.2: (Converse of Pythagorean Theorem): If
2 2 2
c b a = + , then c Z is a right
angle.

Figure 2.29
Proof. We are given ABC A with
2 2 2
c b a = + .
Construct a right angle at point F on rays FH and FG . Define point E
e
FH such that
, a FE= and define point FG De such that . b FD= Then DEF A is a right triangle.

By the Pythagorean Theore
2 2 2 2 2 2
c b a e d f = + = + =
. This means that c f = and hence by
SSS. . DEF ABC A ~ A Hence . 90 = Z = Z F C

Definition 2.6.1: (Trigonometry) Let ABC A be a right triangle with right angle at vertex C,
and let
u
=
. CAB Z
Then if u is a acute, we define

AB
AC
and
AB
BC
= = u u cos sin
If u is obtuse, then let u u = ' 180 and define
u u ' =sin sin and u u ' = cos cos
Also, define 0 0 sin = and 1 0 cos = ; 0 90 sin = and 0 90 cos =

Theorem 2.6.4 (Pythagorean Identity)
1 cos sin
2 2
= + u u

Proof. (Exercise.)

Theorem 2.6.3 (Law of Sines): Let ABC A be any triangle with sides a, b, c, opposite vertices
A, B,C. Then,

C
c
B
b
A
a
Z
=
Z
=
Z sin sin sin

Proof. (Exercise)

Theorem 2.6.6 (Law of Cosines): Let
ABC A
be any triangle with sides a, b, c opposite
vertices A,B,C. Then
C ab b a c Z + = cos 2
2 2 2

Proof (Exercise.)

Euclid, of course, did not state the Pythagoreans theorem in terms of the sum of the squares of
the edges; to do so would have required algebra, which was not invented for another thousand
years after Euclid. Instead, the theorem was expressed in terms of area.

Theorem 2.6.4 (Euclids Version of the Pythagorean Theorem): The area of the square on
the hypotenuse of a right triangle is equal to the sum of areas of the squares on the legs.

Theorem 2.6.5: In any triangle, the product of a base and the corresponding altitude is
independent of the choice of the base.
Restatement. Given ABC A . Let be the altitude from A to , and let be the altitude
from B to . Then

Figure 2.30
PROOF. Suppose that and , as shown in the figure. Then and
, because both are right angles. Therefore .
Hence . Thus , and which was to be
proved. If E = C, then is a right triangle with its right angle C and we also have D = C.

Figure 2.31
In this case, the theorem says trivially that ab = ba.

Theorem 2.6.5: For similar triangles, the ratio of any two corresponding altitudes is equal to
the ratio of any two corresponding sides.
Restatement. Suppose that . Let h be the altitude from A to , and let h be
the altitude from A to . Then, .


Figure 2.32
PROOF: Let and be the altitudes whose lengths are h and h.
If D = B, then D = B, and there is nothing to prove. If not, and the
theorem follows.

Theorem 2.6.6: The area of a right triangle is half the product of the length of its legs.

Figure 2.33
PROOF. Given , with a right angle at C. Let D be the point such that ADBC is a
rectangle. By the additivity postulate, Area (ADBC) = Area (
A
ABC) + Area (
A
ABD).
By the congruence postulate, Area (
A
ABC) = Area (
A
ABD).
By the rectangle formula, Area (ADBC) = ab.
Therefore ab ABC Area = A ) ( 2 and Area (
A
ab ABC
2
1
) = which was to be proved.

Theorem 2.6.7: The area of a triangle is half the product of any base and the corresponding
altitude

Figure 2.34


PROOF. Given . Let D be the foot of the perpendicular from B to ; let AC = b and
let BD =h (as in each of the figures). There are essentially, three cases to consider.
(1) If A = D, then is a right triangle and , by Theorem 1.
(2) A-D-C. Let and . By Theorem 1, , .
By the additivity postulate
BDC BDA ABC A + A = A
.
Therefore, , which was to be proved.
(3) D-A-C. let . By Theorem 2.6.5, Area ( ) BDC A = ( )h b b +
'
2
1
.
Also by Theorem 2.6.5 Area ( ) BDA A = ( )h b b +
'
2
1
.
By the additivity postulate, .
Therefore, . which was to be
proved.

Theorem 2.6.8: If two triangles have the same altitude, then the ratio of their areas is equal to
the ratio of their bases.

This theorem follows immediately from the area formula. If the triangles and
have bases , and the corresponding altitude for each of them is h, then
which was to be proved. In the same way, we get the following theorem.

Theorem 2.6.9: If two triangles have the same base, then the ratio of their areas is the ratio of
their corresponding altitudes. The next theorem is a corollary of each preceding theorems.

Theorem 2.6.10: If two triangles have the same base and the same altitude, then they have the
same area.


Figure 2.35

Theorem 2.6.11. If two triangles are similar, then the ratio of their areas is the square of the
ratio of any two corresponding sides. That is, if , then
.

Figure 2.36

PROOF. If the altitudes to are h and h, as in the figure above, then we know from
Theorem 2.6.7 that .
Now, which was to be proved.


CHAPTER THREE
Hyperbolic Geometry
Introduction
Until the 19th century Euclidean geometry was the only known system of geometry
concerned with measurement and the concepts of congruence, parallelism and
perpendicularity. Then, early in that century, a new system dealing with the same concepts
was discovered. The new system, called non-Euclidean geometry, contained theorems that
disagreed with the Euclidean theorems.
Non-Euclidean Geometry is not not Euclidean Geometry. The term is usually applied only to
the special geometries that are obtained by negating the parallel postulate but keeping the
other axioms of Euclidean Geometry.
Since the first 28 postulates of Euclids Elements do not use the Parallel Postulate, then these
results will also be valid in our first example of non-Euclidean geometry called hyperbolic
geometry.
Recall that one of Euclids unstated assumptions was that lines are infinite. This will not
be the case in our other version of non-Euclidean geometry called elliptic geometry and so not
all 28 propositions will hold there (for example, in elliptic geometry the sum of the angles of a
triangle is always more than two right angles and two of the angles together can be greater
than two right angles, contradicting Proposition 17).
Hyperbolic geometry is the geometry you get by assuming all the postulates of Euclid,
except the fifth one, which is replaced by its negation.

3.1. The Poincare Model
In this section we shall assume that there is a mathematical system satisfying the postulates of
Euclidean plane geometry, and we shall use Euclidean geometry to describe a mathematical
system in which the Euclidean parallel postulate fail, but in which the other postulates of
Euclidean geometry hold.

Consider a fixed circle C in a Euclidean plane. We assume, merely for the sake of
convenience, that the radius of C is 1. Let E be the interior of C.

Figure 3.1
By an L-circle (L for Hyperbolic) we mean a Circle C' which is orthogonal to C. When we
say that two circles are orthogonal, we mean that their tangents at each intersection point are
perpendicular. If this happens at one intersection point R, then it happens at the other
intersection point S. But we shall not stop to prove this; this chapter is purely descriptive and
proofs will come later.

The points of our L-plane will be the points of the interior E of C. By an L-line we mean
(1) the intersection of E and an L-circle, or
(2) The intersection of E and a diameter of C.
It is a fact that:

Every two points of E lie on exactly one L-line. (*)
We are going to define a kind of plane geometry in which the plane is the set E and the
lines are the L-lines. In our new geometry we already know what is meant by point and line.
We need next to define distance and angular measure.
For each pair of points X, Y, either on C or in the interior of C, let XY be the usual Euclidean
distance. Notice that if R, S, T, and U are as in the figure, then R and S are not points of our

L-plane, but they are points of the Euclidean plane that we started with. Therefore, all of the
distances TS, TR, US, UR are defined, and (*) tells us that R and S are determined when T
and U are named. There is one and only one L-line through T and U, and this L-line cuts the
circle C in the points R and S. We shall use these four distances TS, TR, US, UR to define a
new distance d (T, U) in our plane E, by the following formula:

US UR
TS TR
U T d
e
/
/
log ) , ( = ,
Evidently we have the following postulate:
: d is function
9 E E d : .
Let us now look at the ruler postulate in chapter 1. On any L-line L, take a point U and regard
this point as fixed. For every point T of L, let

US UR
TS TR
T f
e
/
/
log ) ( = .
That is, ) (T f is what we get by omitting the absolute value signs in the formula for d (T, U).
We now have a function,
9 L f : .
We shall show that f is a coordinate system for L.
If V is any other point of L, then

US UR
VS VR
V f
e
/
/
log ) ( = .
Let ) (T f x = and ) (V f y = . Then

VS VR
TS TR
US UR
VS VR
US UR
TS TR
y x
e e e
/
/
log
/
/
log
/
/
log = = , because the
difference of the logarithms is the logarithm of the quotient.
Therefore;
) , ( V T d y x = , which means that our new distance function atisfies the ruler postulate.
Since the ruler postulate in chapter 1holds, the other distance postulates automatically hold..
We define betweenness, segment, rays, and so on, exactly as in chapter 1. All of the theorems
of chapter 1 hold in our new geometry, because the new geometry satisfies the postulates on

which the proofs of the theorems were based. It is rather easy to convince yourself that the
plane-separation postulate holds E.
To discuss congruence of angles, we need to define an angular-measure function. Given an
L-angle in our new geometry, we form an angle in the old geometry by using the two
tangent rays:

Figure 3.2
We then define the measure BAC m < of BAC < to be the measure (in the old sense) of the
Euclidean angle C A B ' ' < .
It is a fact that the resulting structure
| | m d L E , , ,
satisfies all the postulates of chapters 1, including the SAS postulate. The proof of this takes
time, however, and it requires the use of more Euclidean geometry than we know so far.
Granted that the postulates hold, it follows that the theorems also hold. Therefore, the whole
theory of congruence, and of geometric inequalities, applies to the Poincare model of
hyperbolic geometry.

Figure 3.3

On the other hand, the Euclidean parallel postulate obviously does not hold for the Poincare
model. Consider, for example, an L-line L which does not pass through the center P of G
(Figure 3.3). Through P there are infinitely many L-lines which are parallel to L.

3.2. The Hyperbolic Parallel Postulate
Hyperbolic geometry (also called Lobachevskian geometry) is the kind represented by the
Poincare model. In such geometry, when the familiar parallel postulate fails, it pulls down a
great many familiar theorems with it. A few samples of theorems in hyperbolic geometry
which are quite different from the analogous theorems of Euclidean geometry follow.
1) No quadrilateral is a rectangle. In fact, if a quadrilateral has three right angles, the
fourth angle is always acute.
2) For any triangle, the sum of the measures of the angles is always strictly less than 180.
3) No two triangles are ever similar, except in the case where they are also congruent.
The third of these theorems means that two figures cannot have exactly the same
shape, unless they also have exactly the same size. Thus, in hyperbolic geometry,
exact scale models are impossible.
In fact, each of the above three theorems characterizes hyperbolic geometry. If the
angle-sum inequality,
, 180 < < + < + < C m B m A m
holds, even for one triangle, then the geometry is hyperbolic; if the angle-sum equality
holds, even for one triangle, then the geometry is Euclidean; and similarly for (1) and (3).
This has a curious consequence in connection with our knowledge of physical space. If
physical space is hyperbolic, which it may be, it is theoretically possible for the fact to be
demonstrated by measurement. For example, suppose that you measure the angles of a
triangle, with an error less than 0.0001 for each angle. Suppose that the sum of the measures
turns out to be 179
0
59
59.999
. The difference between this and 180

0
is 0.001
. This
discrepancy could not be due to errors in measurement, because the greatest possible
cumulative error is only 0.0003. Our experiment would therefore prove that the space that
we live in is hyperbolic.

On the other hand, no measurement however exact can prove that space is Euclidean. The
point is that every physical measurement involves some possible error. Therefore we can
never show by measurement that an equation,
, 180 = + + t s r
holds exactly; and this is what we would have to do to prove that the space we live in is
Euclidean.
Thus there are two possibilities:
1) The Euclidean parallel postulate does not hold in physical space, or
2) The truth about physical space will never be known.
The Hyperbolic Parallel Postulate: Given a line L and a point P not on L, there are at least
two lines L L ' ' ', which contain P and are parallel to L.

Figure 3.4

3.3. Closed triangles and angle Sum
In this section we deal specifically with the hyperbolic case. To avoid confusion, throughout
this chapter, we shall mention the hyperbolic parallel postulate in every theorem whose proof
requires it.

Figure 3.5
If

AB PD
, then PAD A is called a closed triangle.

Note that every closed triangle is an open triangle, but under hyperbolic parallel postulate the
converse is false, because through P there is more than one line parallel to

AB.
Closed triangles have important properties in common with genuine triangles.
Theorem 3.3.1: The Exterior Angle Theorem: Under hyperbolic parallel postulate, in every
closed triangle, each exterior angle is greater than its remote interior angle.
That is,
If

AB PD
and Q A B, then < QAP > <P.

Figure 3.6

Figure 3.7
Proof: If DPAB A is isosceles, this is obvious. Here, if hyperbolic parallel postulate holds,
then < P and < PAB are acute (because c(a) < 90 for every a), and therefore < QAP is obtuse.
Suppose then that DPAB A is not isosceles. By the theorem: If

AB PD , then DPAB A
DPAB A is equivalent to an isosceles open triangle DPCB A , and this open triangle is also
closed:

Figure 3.8
If C = A, there is nothing to prove. For the case A C B, let the degree measures of the
various angles be as in the figure. Then
p > r, because c(a) < 90.
And
180 s + + s q p , by theorems in chapter one.
Therefore,
s r s p q t + > + > =180 , and
s r t + > , which proves half of our theorem.
To prove the other half, we need to show that q u > . This follows from
s r u q t + = > = 180 180 .
We found, in Chapter one, that the critical function c was non-increasing. That is, if a a > ' ,
then ) ( ) ( a c a c s ' . Using the exterior angle theorem, we can sharpen this result.

Figure 3.9

Theorem 3.3.2: Under hyperbolic parallel postulate, the critical function is strictly
decreasing. That is, if a a > ' , then ) ( ) ( a c a c s ' .
Proof: In the figure, AP = a and a P A ' = ' ,

AB PD
and

' ' AB D P
, so that

' 'D P PD
.
Therefore PD P D ' ' A is a closed triangle. Therefore ) ( ) ( a c a c ' > , this was to be proved.
Theorem 3.3.3: Under hyperbolic parallel postulate, the upper base angles of a Saccheri
quadrilateral are always acute.

Figure 3.10
(We already know, from chapter 1, that they are congruent, and cannot be obtuse.)
In the figure,
BQand
CP are the critical parallels to
AD , through B and C.
Therefore
DCP m a c ABQ m < = = < ) ( , as indicated. Applying the exterior angle theorem to the
closed triangle PCBQ A , we see that
. s t >
Therefore
). ( ) ( a c s a c t + > +
Therefore
90 ) ( < + a c s , which proves our theorem.

Figure 3.11
Theorem 3.3.4: Under hyperbolic parallel postulate, in every right triangle ABC A , we have
180 < < + < + < C m B m A m
Proof: Suppose not. Then, if < A is the right angle, < B and < C must be complementary.
Take D on the opposite side of
BC from A, so that ABC BCD ~< < and CD = AB.

Then DCB ABC A ~ A , by SAS; and ABDC is a Saccheri quadrilateral. This is impossible,
because < D is a right angle.
Theorem 3.3.5: Under hyperbolic parallel postulate, for every triangle ABC A , we have
108 < < + < + < C m B m A m

Figure 3.12
Proof: Let
AC be a longest side of ABC A , and let
BDbe the altitude from B to
AC . Then
, 180 90< + + s r

And
. 180 90 < + + u t

Therefore
, 180 ) ( < + + + u t s r which proves the theorem.

Soon we shall see that under hyperbolic parallel postulate this theorem has a true converse:
for every number 180 < x there is a triangle for which the angle sum is x . Thus 180 is not an
upper bound for the angle sums of triangles, but is precisely their supremum.

3.4. The defect of a triangle and the collapse of similarity Theory
The defect of ABC A is defined to be C m B m A m < < < 180

The defect of ABC A is denoted by ABC A o

under hyperbolic parallel postulate we know that
the defect of any triangle is positive, and obviously it is less than 180. (Later we shall see that
the converse holds: every number between 0 and 180 is the defect of some triangle.)
The following theorem is easy to check, regardless of under hyperbolic parallel postulate.

Theorem 3.4.1: Given ABC A , with B D C. Then
ADC ABD ABC A + A = A o o o

Figure 3.13
It has, however, an important consequence.
Theorem 3.4.2:

Under hyperbolic parallel postulate, every similarity is congruence. That is, if
DEF ABC A A ~ , then DEF ABC A ~ A .

Figure 3.14

First we take G on
AB so that DE AG = ; and we take H and
AC so that AH = DF. We then

have EDF AGH A ~ A , by SAS; therefore

ABC AGH A A ~

If G = B, then H = C, and the theorem follows. We shall show that the contrary assumption
B G = , C H = ( as shown in the figure ) leads to a contraction.
Let the defects of GHC AGH A A ~ , and GBC A be d
1
, d
2
, and d
3
, as indicated in the figure;
let d be the defect of ABC A . By two applications of the preceding theorem, we have :
d = d
1
+ d
2
+ d
3
. This is impossible, because the angle congruences given by the similarity
AGH ABC A A ~ tell us that d = d
1
.
The additivity of the defect, described in theorem 3.4.1, gives us more information about the
critical function. What we know so far is that
1. 90 ) ( 0 < < a c for every a>0, and
2. c decreases as a increases.
There remains the question of how small the numbers ) (a c eventually become when a is very
large. We might have either of the following situations:

Figure 3.15
In each case, { } ) ( inf a c e = , that is, the greatest lower bound of the numbers ) (a c . In each
case, it follows from (2) that e a c
a
=

) ( lim . To prove the following theorem, therefore, we
need merely show that 0 > e is impossible.


Theorem 3.4.3: 0 ) ( lim =

a c
a
.
Proof: Suppose that 0 ) ( > > e a c for every a.

Figure 3.16
The markings in the figure should be self-explanatory. For each n,

n n
Q P intersects

1 0
R P ,
because ) (n c e < . The right triangles
1 1 + +
A
n n n
Q P P all are congruent, and therefore have the
same defect d
0
. Consider now what happens to the defect d
n
of
n n
R P P
0
A where n is increased
by 1. In the figure below, the letters in the interiors of the triangles denote their defects. We
have:
,
1 0
y d R P P
n n n
+ = A
+
o
,
1 0
x d R P P
n n n
+ = A
+
o
), ( ) (
0 1
x d y d d
n n
+ + + =
+


Figure 3.17
by theorem 3.4.1 in each case. Therefore
.
0 1
d d d
n n
+ >
+

Thus ,
0 1 2
d d d + > ; 2
0 1 0 2 3
d d d d d + > + >
And by induction, we have . ) 1 (
0 1
d n d d
n
+ >
When n is sufficiently large, we have 180 >
n
d , by the Archimedean postulate. This is
impossible, because the defect of a triangle is 180 minus the angle sum. Therefore
0 ) ( > > e a c is impossible, which was to be proved.
Consider now what happens to the measure ) (a r of the base angles of an isosceles right
triangle, as the length a of the legs becomes large.

Figure 3.18 Figure 3.19

Here

AC BD . Therefore we always have ) ( ) ( a c a r < . Therefore 0 ) ( lim =

a r
a
. Let us now
make the figure symmetrical by copying ABC A on the other side of

AB. For DBC A , the
angle sum is ) ( 4 a r . Therefore the defect ) ( 4 180 a r can be made as close to 180 as we
please; we merely need to take a sufficiently large. Thus 180 is not merely an upper bound of
the numbers which are the defects of triangles; 180 is precisely their supremum.
Theorem 3.4.4: For every number 180 < x there is a triangle whose defect is greater than x.


CHAPTER FOUR
The consistency of the hyperbolic geometry
In this chapter, we shall show that the Poincare Model satisfies all the postulates of hyperbolic
geometry. Our analysis of the model will depend, of course, on Euclidean geometry, and so
our consistency proof will be conditional. At the end of the chapter we shall know not that the
hyperbolic postulates are consistent, but merely that they are as consistent as the Euclidean
postulates.
4.1. Inversions of a Punctured Plane
Given a point A of a Euclidean plane E and a circle C with center at A and radius a.
The set E A is called a punctured plane. The inversion of E A about C is a function,
, : A E A E f
defined in the following way. For each point P of E A, let ) (P f P = ' be the point of
APfor
which
AP
a
P A
2
= ' .

Figure 4.1

(Thus, for , 1 = a we have .) / 1 AP P A = ' since ,
2
a a a = we have the following theorems.
Theorem 4.1.1: If C Pe
,
then . ) ( P P f =
Theorem 4.1.2: If P is in the interior of , C then ) (P f is in the exterior of C, and conversely.

Theorem 4.1.3: For every . )) ( ( , P P f f P =
That is, when we apply an inversion twice, this gets us back to wherever we started.
Proof: ) (P f is the point of APfor which )) ( ( , )
2
P f f and AP a AfP = is the point of the
same ray for which
AP
AP a
a
P Af
a
fP Af = = =
/ ) (
)) (
2
2 2

Therefore . ) ( ( P p f f =
Theorem 4.1.4: If L s a line through A, then . ) ( A L A L f =
Here by ) ( A L f we mean the set of all image point ), ( p f where e P A. In general, if
, A E K c
Then
( ) { } K P P f P K f e = ' = ) ( .

It is also easy to see that if P close to A, then P'
is far from A, and conversely; the reason is
that AP a / "
2
is large when AP is small In studying less obvious properties of inversions, it
will be convenient to use both rectangular and polar coordinates, taking the origin of each
coordinate system at A.

The advantage of polar coordinates is that they allow us to describe the inversion in the
simple form

). , ( ) , ( :
. :
u u s r
A E A E f

where

r
a
s
2
=

and
s
a
r
2
=
In rectangular coordinates, we have
), sin , cos ( ) ( u u r r y x P = =

), sin , cos ( ) ( ) ( u u s s v u P f = =
Where r and s are related by the same equation as before evidently.

2 2 2
v u s + =
just as
2 2 2
y x r + = .
These equations will enable use to tell what happens to lines and circles under inversions. We
allow the cases in which the lines and circles contain the origin A, so that they appear in E-A
as punctured lines and punctured circles. Thus we shall be dealing with four types of
figures, namely, lines and circles, punctured and unpunctured. For short, we shall refer to such
figures as k-sets. The rest of this section will be devoted to the proof that if K is a k-set, then
so also is ). (K f Let us look first, however, at a special case.
Let K be the line x=a.

Figure 4.2
Then K is the graph of the polar equation
a r = u cos
Since ,
2
s a r = where ( ), , ) , ( u u s r f = it follows that ) (K f is the graph of the condition
0 . cos
2
= = s a
s
a
u
Or 0 , cos = = s a s u
or 0 , cos
2
= = s as s u

In rectangular form, this is
. 0 ,
2 2 2 2
= + = + v u au v u
Replacing u and v by x and y (to match the labels on the axes), we see that ) (K f is the
graph of
, 0 , 0
2 2 2 2
= + = + y x y ax x

And is hence the punctured circle with center at ) 0 , 2 / (a and radius . 2 / a thus f has pulled the
upper half of the line K onto the upper semicircle, and the lower half onto the lower
semicircle. It is to see that points far from the x-axis either above or below) go onto points
near the origin.
More generally, we have the following theorem.
Theorem 4.1.5: If K is a line in E-A, then ) (K is a punctured circle.
Proof: since we can choose the axes any way we want, we are free to assume that K is the
graph of a rectangular equation
. 0 > =b x
And hence of a polar equation
. 0 cos > =b r u
As before, setting s a r
2
= , we conclude that ) (K f is the graph of
. cos
2
b
x
a
= u . 0 = s
Or , cos
2
2
u s
b
a
s = . 0 = s
Or . 0 , 0
2 2 2
2
2
= + = + v u v u
b
a
u
Or . 0 , 0
2 2 2
2
2
= + = + y x v u
b
a
u
Therefore ) (k f is a punctured circle, with center at ) 0 , 2 / (
2
b a and radius b a 2 /
2


It is easy to see that (1) every punctured circle is described by the above formula for some
choice of b and some choice of the axes. Therefore (2) every punctured circle L is = ) (K f for
some line K. Theorem 3 tells us that P P f f = )) ( ( for P. Therefore
. )) ( ( ) ( K K f f L f = =
Thus we have the following theorem.
Theorem 4.1.6: It L is a punctured circle, then ) (L f is a line in . A E
We now know, from Theorem 4, that under f , punctured lines go onto punctured lines; and
we know, by theorems 5 and 6, that lines go onto punctured circles and vice-versa. Now we
must see what happens to circles.
Theorem 4.1.7: If M is a circle in E-A, then ) (M f is a circle in E-A.
Proof: M is the graph of a rectangular equation
, 0
2 2
= + + + + C By Ax y x
where 0 = C because the circle is not punctured. In polar form, this is
, 0 sin cos
2
= + + + C Br Ax x u u

Since ,
2
s a r = this tells us that ) (M f is the graph of the equation
. 0 sin . cos .
2 2
2
4
= + + + C
s
a
B
s
a
A
s
a
u u
Or , sin cos
2 2 2 4
o C s Ba Aa a
s
= + + + u u
or , ) ( /
2 2 2 2 4
o v u C v Ba u Aa a = + + + +

Replacing u and v by x and y, to match the labels on the axes, we get an equation for f(M)in
the form
. 0
2 2 2
2 2
= + + + +
C
a
y
C
Ba
x
C
Aa
y x
The graph ) (M f is a circle; this circle is punctured, because . 0 /
2
= C a
Theorem 4.1.8: If K is a k-set, then so also is f(K).


4.2. Cross Ratio and Inversions
We recall, from Chapter 1, the definition of distance in the Poincare model.

Figure 4.3
If T and U are points of the L-line with end points R, S on the boundary circle C, then the
non-Euclidean distance is defined by the formula.
.
/
/
log, ) , (
US UR
TS TR
U R d =
The fraction whose logarithm gets taken in this formula is called the cross ration the
quadruplet R,S,T,U, and is commonly denoted by (R,S,T,U).

Thus
(R,A,T,U)= ,
/
/
log
US UR
TS TR

And changing the notation slightly, we have

3 2 4 1
4 2 3 1
4 3 2 1
.
.
) , , , (
P P P P
P P P P
P P P P = .
We shall show that inversions preserve the cross ratio. In the following theorem, f is an
inversion of a punctured plane E-A about a circle with center at A and radius a, as in the
preceding section.
Theorem 4.2.1: If P:=
), 4 , 3 , 2 , 1 )( (
1
= = ' i P f P
i
then
). , , , ( ) , , , (
4 3 2 1 4 3 2 1
P P P P P P P P ' ' ' ' =

Proof: For each i from 1 to 4, let the polar coordinates of P
i
be ). , (
i i
r u By the usual polar
distance formula, we have

). cos( 2
2 2 2
j i j i j i j i
r r r r P P u u + =

Now
. , ) , (
2
|
|
.
|
\
|
= = '
i
i
i i i
r
a
s P u u

Therefore
)] cos( 2 )][ ( cos 2 [
)] cos( 2 )][ cos( 2 [
) , , , (
3 2 3 2
2
3
2
2 4 1 4 1
2
4
2
1
4 2 4 2
2
4
2
2 3 1 3 1
2
3
2
1 2
4 3 2 1
u u u u
u u u u
+ +
+ +
=
r r r r r r r r
r r r r r r r r
P P P P
,

And
2
4 3 2 1
) , , , P P P P ' ' ' '
(
+
(
(
+
(
+
(
+
=
) cos( 2 ) cos( 2
) cos( 2 ) cos( 2
3 2
3 2
4
2
3
4
2
2
4
4 1
4 1
4
2
4
4
2
1
4
4 2
4 2
4
2
4
4
2
2
4
3 1
3 1
4
2
3
4
2
1
4
u u u u
u u u u
r r
a
r
a
r
a
r r
a
r
a
r
a
r r
a
r
a
r
a
r r
a
r
a
r
a

To reduce the second of these fractions to the first, we multiply in both the numerator and
denominator by

.
8
2
4
2
3
2
2
2
1
a
r r r r

This theorem will tell us, in due course, that inversions applied to the Poincare model are
isometries, relative to the non-Euclidean distance.

4.3. Angular Measure and Inversions
A reexamination of Section 4.1 will indicate that the image of an angle, under an inversion, is
never an angle. The point is that every angle in E-A has at least one side lying on a
nonpunctured line, and the image of a nonpunctured line is always a punctured circle.
Therefore the following theorem does not mean what it might seem to mean.

Figure 4.4
Theorem 4.3.1: If A, and Q are non-collinear ), ( ) ( Q f Q and P f P = ' = '
. P Q A m APQ m ' ' Z = Z
Proof: Consider P A Q and PAQ ' ' A A . They have the angel A Z in common.
Since ,
2 2
AQ
a
Q A
P A
a
P A = '
'
= '
We have , . .
2
a Q A AQ P A AP = ' = '
So that .
P A
AQ
Q A
AP
'
=
'

By the SAS similarity theorem,
PAQ A . P A Q ' ' A
(Note the reversal of order of vertices here.) Since P Q A and APQ ' ' Z Z are corresponding
angles, they have the same measure.

Figure 4.5
In the figure above, ) ( ) ( Q f Q and P f P = ' = ' as before. Here we have

.
) 180 (
180
o
o
o
=
=
=
s
r
r u

Therefore
o u s
The order of s and u depends on the order in which P and P'
appear on appear on the ray.
If P and P'
are interchanged, we should interchange s and u, getting
o = s u
Thus in general we have

. o = u s

Consider next the situation illustrated in the figure below:

Figure 4.6
Here B is the center of a circular arc;
PQ
is a line intersect in the arc at c at P; PS is a tangent
ray at P: . a P R =
o
We assert that
.
lim
0
SPQ m PQ R m
a a
Z = Z

(Proof? The first step is to show that . 0 lim = = Z
PS R m
a o a

Consider now a circular arc QS with end point at a point Q .

Figure 4.7
For small positive numbers a, let
o
R be the point of the arc for which
. a QR =
o

Let S Q ' ' be the image of ; QS that is
; ) (QS f S Q = ' '
Let R Q ' ' be the tangent ray at Q' .
We assert that
. TQR R Q A Z ~ ' ' Z
To see this, we observe that
a a
R Q A m and TQR m ' ' Z Z are the s and u that we discussed just
after Theorem 4.3.1. Therefore
o = ' ' Z Z
a a
R Q A m TQR m

Now ,
lim
0
TQR m TQR m
a a
Z = Z

And ,
lim
0
R Q A m R Q A m
a a
' ' Z = ' ' Z

Therefore . [
lim
0
R Q A m TQR m TQR m
a a
' ' Z Z Z

But the absolute value of the quantity indicated in square brackets is ; o = and
0 0 a and a . Therefore

R Q A m TQR m ' ' Z = Z .
Given two intersecting circles or lines, the tangent rays give us tangent angles, like this:

Figure 4.8
By the preceding result, we have the following theorem.
Theorem 4.3.2: Under inversions, corresponding tangent angles are congruent.
That is, if AC and AB are arcs with a tangent angle of measure r , then their images
) ( ) ( AC f and AB f have a tangent angle of measure r. Similarly for an arc and a segment or a
segment and a segment.

4.4. Reflections across L-Lines in the Poincare Model
We recall that the points in the Poincare model are the points of the interior E of a circle C
with center at P; the L-lines are : (1) the intersection of E with lines through P and
(2) the intersection E with circles C' orthogonal to C.

Figure 4.9

If L is an L-line of the first type, then the reflection of E across L is defined in the familiar
fashion as a one correspondence.
, : E E f
Such that for each point Q of E, Q and f(Q) are symmetric across L.
If L is an L-line of the second type, then the reflection of E across L is the inversion of E
about . C'
To Justify this definition of course, we have to show that if f is an inversion about a circle
C' orthogonal to C, then f(E)=E. But this is not hard to show. In the next few theorems, it
should be understood that f is an inversion about C' ; C' has center at A, and intersects C
orthogonally at R and . : C E L and S ' =

Figure 4.10
Theorem 4.4.1: F(C)=C.
Proof: ) (c f is a circle. This circle contains R and S, because R R f = ) ( and . ) ( S S f = By
Theorem 2 of the preceding section, C and C f ' ) ( are orthogonal. But there is only one circle
C which crosses C' orthogonally at R and S. (Proof? Show that P must be the center of any
such circle.) therefore , ) ( C C f = which was to be proved.
Theorem 4.4.2: E E f = ) (
Proof: Let X be any point of E. Then AX intersects C at points T and U. Since f(C)=C, we
have U=f(T) and T=f(U). But inversions preserve betweenness on rays starting at A.
Therefore TU TU f = ) ( , and E X f e ) ( . Thus . ) ( E E f c

We need to show, conversely, that ). (E f E c This is trivial: given that E E f c ) ( , we have
). ( )) ( ( E f E f f c Since ). ( , )) ( ( E f E gives this E E f f c =
Theorem 4.4.3: If M is an L-line, then so also is ) (M f .
Proof: M is the intersection , D E where D is either a circle orthogonal to E a line
orthogonal to C. Now ) (D f is orthogonal to C, and is a line or circle (punctured or
unpunctured). Let D'
be the corresponding complete line or circle. (Thus
.) ) ( ) ( A D f D or D f D = ' = ' Then

.
) ( ) (
E D
E D f M f
' =
=

which is an L-line.
We recall that an L-angle is the angle formed by two ray in the Poincare model.

Figure 4.11
The measure of an L-angle is the measure of the angle formed by the tangent rays.
We can now sum up nearly all of the preceding discussion in the following theorem.
Theorem 4.4.4: Let f be a reflection E across and L-line. Then,
(1) f is a one-to-one correspondence ; E E
(2) f preserves the non-Euclidean distances between points;
(3) f preserves L-lines;
(4) f preserves measures of L-angles.
For L-lines of the first kind (passing through P) all this is trivial, because in this case f is an
isometry in the Euclidean sense. It therefore preserves distances of both kinds, lines, circles,
orthogonality, and angular measure. For L-lines of the second kind, Condition (1) through (4)
follows from the theorems of this section and the preceding two sections.

4.5. Uniqueness of the L-line Through Two Points
Given the center P of C, and some other point Q of E. We know that P and Q lie on only one
(straight) line in the Euclidean plane. Therefore P and Q lie on only one L-line of the first
kind. But P does not lie on any L-line of the second kind. (The reason is that on the right
triangle ARP A in the figure, the hypotenuse. , AP is the longest side.) It follows that the L-line
through two points of E is unique, in the case where one of the points is P.

Figure 4.12
To prove that uniqueness always holds, holds, we need the following theorem.
Theorem 4.5.1: For each point Q of E there is a reflection f such that . ) ( P Q f =

Figure 4.13

Proof: We start by the method of wishful thinking. If the inversion f about C' give
, ) ( p Q f = then
AQ
a
AP
2
=
We recall that the radius PR=1. Let K = QP, and let x be the unknown distance

AP (Figure
4.13).

Then the equation

2
. a AQ AP =
takes the form 1 ) (
2
= x k x x
Or
k
x or kx
1
1 = =
Since Q is in E, we know that k<1. Therefore, x >1, and A is outside D. If C' is the circle with
center at A, orthogonal to C, then the reflection across C E ' is the one that we wanted.
We can now prove the following theorem.
Theorem 4.5.2: In the Poincare model, every two points lie on exactly one L-line.
Proof: Let Q and R be points of E. Let f be a reflection across an L-line such that f (Q) = P
and f (R) = R'
. We know that P and R'
. We know that P and R'
lie on an L-line L .Therefore
Q and R lie on the L-line f (L). If there were two L-lines and f (L
2
) would be different L-lines
containing P and which is impossible.
By Theorem 4.5.1, we can speak of the L-line containing Q and r. We shall, denote this by
; QR and to avoid confusion, we shall agree not to use this notation in the rest of this chapter,
to denote Euclidean lines.

4.6. The Ruler Postulate; betweenness: Plane Separation and angular measure
Our strategy in this chapter is to verify statements about L-lines, first for the each case of L-
lines through P, and then to use inversions to show that the curved L-lines behave in the
same way as the straight ones. In this spirit, we first check the ruler postulate first check the
ruler postulate the L-lines through P.
Theorem 4.6.1: Every L-line through P has a coordinate system.

Figure 4.14

Proof: Suppose that L Passes through P, and let its end points on C be R and S. For Every
point Q of L., let

PS PR
QS QR
Q f
e
/
/
log ) ( =

QS
QR
e
log = .
(Because PR=PS.). Let QS = x. Then
QR = 2- QS = 2 - x,
And we have :
x
x
Q f
e

=
2
log ) (
Obviously f is a function R L into the real numbers. We need to verify that f is a one-to-
one correspondence R L . The we need to show that every real number K is =f (Q) for
exactly one point Q. Thus we want

x
x
k
e

=
2
log
or
x
x
e
k

=
2

or 2 ) 1 ( = + x e
k

or
1
2
+
=
k
e
x
Every k there is exactly one such x, and 0<x<2, as it should be. There every k is = f(Q) for
exactly one point Q of L.
We have already checked, in Chapter 9, that when the coordinate system f defined in this way,
the distance formula
) ( ) ( ) , ( U f T f U T d =
Ways satisfied.
Before proceeding to generalize Theorem 1, we observe that the formulas se give us some
more information.


Figure 4.15
The figure S Q x
i
= for i=1,2,3. It is easy to check that (2-x)/x is a function. (Its derivative is
). 0 / 2
2
< x and the logarithm is an in sing function. Function.
Therefore, if ,
3 2 1
x x x < < as in the figure, it follows that
), ( ) ( ) (
3 2 1
Q f Q f Q f < <
and conversely. We recall that betweenness is defined in terms of distance, and one point of a
line is between two others if and only if its coordinate is between their coordinates. Thus we
have:
Theorem 4.6.2: Let
3 2 1
, , Q Q Q be points of an L-line through P. Then Q
2
-Q
3
under the non-
Euclidean distance if and only if
3 2 1
, , Q Q Q in the Euclidean plane.
Theorem 4.6.3: Every L-line has a coordinate system.
Proof: Given an L-line L. If L contains P, we use Theorem 4.6.1. If not, let Q be point of L;
let g be a reflection such that g(Q)=P; let ), (L g L = ' and let
9 ' L f :

be a coordinate system for . L' For each point T of L, let
)). ( ( ) ( T g f T f = '

That is, the coordinate of T is the coordinate of the corresponding point g(T) of L'
. Since f
and g are one-to-one correspondences, so also is their composition f(g). Given points T, U of
U of L., we know that
)), ( ), ( ( ) , ( U g T g d U T d =
Because inversions preserve the non-Euclidean distance. This in turn is
, )) ( ( )) ( ( U g f T g f =
Because f f is a coordinate system for . L' Therefore
, ) ( ) ( ) , ( U f T f U T d ' ' =
which was to be proved.
Theorem 4.6.4: Every l-line through P separates E into two sets H
1
and H
2
such that
(1)H
1
,and H
2
are convex, and (2) if QR then H R and H Q
2 1
e e intersects L.

Here QR means of course the non-Euclidean segment

Figure 4.16
Proof: We know that the Euclidean line containing L separates the Euclidean plane into two
half-planes . ,
1 1
H H ' ' Let
2 1
H and H be the intersections E H '
1
and E H '
2
, as indicated in
the figure.
Suppose that , ,
1
H R Q e and suppose that QR intersects L in a point S. Let f be an inversion
, E E about a circle with center A on the line containing L such that . ) ( P S f = Then
) (QR f is an L-line through P, and ) (Q f and ) (R f bel ong to .
1
H Since , R S Q in the
non-Euclidean sense, because f preserves the non-Euclidean distance. Therefore
) ( ) ( R f P Q f in the Euclidean sense, which is impossible, because ) (Q f and ) (R f are in
the same Euclidean half plane.
It follows, in the same way, that
1
H is convex. Thus we have verified half of
Suppose now that
1
H Qe and .
2
H Re Let C' be the Euclidean circle that contains the
: QR line L

Figure 4.17

Then L contains a point S of the Euclidean segment from Q to R, and S is in the interior of C'
. It follows that the Euclidean line containing L intersects C' in points, one of which is a point
T of L. Now we must verify that Q-T-R in non-Euclidean sense. [Hint: Use an inversion
, , , , :
2 2 1 1
P T H H H H E E f and then apply theorem 4.6.2.]
To extend this result to L-lines in general, we observe that:
Theorem 4.6.5: Reflections preserve betweenness.
Because they preserve lines and distance.
Theorem 4.6.6: Reflections preserve segments.
Because they preserve betweenness.
Theorem 4.6.7: Reflections preserve convexity.
Because they preserve segments.
Theorem 4.6.8: The plane separation postulate holds in the Poincare model.
Proof: LET L be any L-line, and let Q be any point of L. Let f be a reflection such that
); ( ( ; ) ( L f L let P Q f = ' = and let
2 1
H and H ' ' be the half-planes in E determined by . L' Let
) (
1
1
1
H f H ' =

and ). (
2
1
2
H f H ' =

1
f is also a reflection ,and reflections preserve convexity, it follows that and
2
H are
convex. This proves half of the plane separation postulate for L. It remains to show that if
1
H Re and
2
H S ' e , then RS intersects L.
If
2 1
), ( ) ( H S and H R then S f S and R f R ' e ' ' e ' = ' = ' so that S R ' ' intersects L'
at a point
. T' Therefore RS intersects L at ). (
1
T f T ' =

Theorem 4.6.9: Reflections preserve segments Reflections preserve half planes.
That is, if
2 1
H and H are the half planes determined by L., then ) (
1
H f and ) (
2
H f are the
half planes determined by ). (L f Proof?
Theorem 4.6.10: Reflections preserve segments Reflections preserve interiors of angles.
Proof: The interior of ABC Z is the intersection of (1) the side of AB that contains C, and (2)
the side of ABthat contains A. Since reflections preserve half planes, they preserve
intersections of half planes.

We have defined the measure of an (non-Euclidean) angle as the measure of
the (Euclidean) angle formed by the two tangent rays. We need to check whether this measure
function satisfies the postulates of section 1.5. For angles with vertex at P this is obvious. To
verify it for angles with vertex at some other point Q, we throw Q onto P by a reflection f.
Now f preserves angles, angular measure, lines, and interiors of angles. It is therefore trivial
to check that if Postulates M-l through M-5 hold at P, then they hold at Q.


CHAPTER FIVE
The Consistency of Euclidean Geometry
Our proof of the consistency of hyperbolic geometry, in the preceding chapter, was
conditional. We should that if there is a mathematical system satisfying the postulates for
Euclidean geometry, then there is a system satisfying the postulates for hyperbolic geometry.
We shall now investigate the if, by describing model for the Euclidean postulates. Here again
our consistency proof will be conditional. To set up our model, we shall need to assume that
the real number system is given.

5.1. The Coordinate Plane and Isometries
Definition 5.1.1: E= R x R, where R is the real number system is called a Cartesian model
(or a coordinate system).
Then a point in a Cartesian model E is defined to be an ordered pair of real numbers.
Definition 5.1.2: A line in the Cartesian model E is a subset of E which has the form
{ }. 0 , 0 ) , (
2 2
> + = + + = B A C By Ax y x L
That is, a line is defined to be the graph of a linear equation in x and y.
Definition 5.1.3: If ), , ( ) (
2 2 1 1
y x Q and y x P = = are two points in a Cartesian model E then the
distance between these two points from analytic geometry is given by
. ) ( ) ( ) , (
2
1 2
2
1 2
y y x x Q P d =

We define betweenness in terms of distance. (As usual, we abbreviate .) ) , ( PQ as Q P d
Segments and rays are defined in terms of betweenness; and angles are defined when rays are
known.
It turns out that setting up an angular measure function is a formidable technical chore. We
hope, therefore, that the reader will settle for a congruence relation ~ for angles, satisfying
the congruence postulates for angles. This relation is defined in the following way.


Recall that: A one-to-one correspondence f from a set A to a set B is a function
f : A B such that, for each b B, there is a unique a A for which f(a) = b. This is
equivalent to the mapping f : A B being both one-to-one and onto. In other words, we have
a pairing between elements of A and elements of B

Definition 5.1.4: An isometry is a one-to-one correspondence

Preserving distance.
Definition 5.1.5: Two angles DEF and ABC Z Z are congruent if there is an isometry
E E f : such that . ) ( DEF ABC f Z = Z
We have now given definitions, in the Cartesian model, for the terms used in the Euclidean
postulates. Each of these postulates thus becomes a statement about a question of fact; and our
task is to show that all of these statements are true.

5.2 The Ruler Postulate
Recall the following:
1. The ruler postulate: Every line has a coordinate system.
2. A vertical line is a line which is the graph of an equation . a x =
3. Every non-vertical line is the graph of an equation b mx y + = .
4. If b x and a x = = are equations of the same line, then . b a =
5. If
2 2 1 1
b x m y and b x m y + = + = are equations of the same line, then
.
2 1 2 1
b b and m m = =
Definition 5.2.1: A coordinate system f on a line L is a one-to-one correspondence

f: L R.

Definition 5.2.2: (Distance function)
For each line L in the plane fix a coordinate system f
L
: L R. Then

the distance function on the plane E is the function d : E E R
which assigns to any two points P, Q a real number d(P, Q) = PQ defined
by
d(P,Q) = PQ =

Theorem 5.2.1: Every vertical line L has a coordinate system.
Proof: Let . a x = be the vertical line and for each point let L of y a P , ) , ( =
y P f = ) ( .
Then f is a one-to-one correspondence . R L If ) , ( ) , (
2 1
y a Q and y a P = = then

2
1 2
2
) ( ) ( ) , ( y y a a Q P d PQ + = =

2
1 2
) ( y y =

1 2
y y =
, ) ( ) ( P f Q f =
as desired .
Theorem 5.2.2: Every non-vertical line has a coordinate system.

Figure 5.1

Proof: Let L be the graph of y=mx+b. If(x
1,
y
1
) and (x
2
, y
2
)
, L e
then it is easy to check that
). ( ,
1 2 1 2
1 2
1 2
x x m y y m
x x
y y
= =


and
. ) 1 ( ) ( ) (
1 2
2 2
1 2
2 2
1 2
x x m x x m x x PQ + = + =
From this we see how to define a coordinate system for L.For each point
.Let
. 1 ) , (
2
m x y x f + =
Then for ) , ( ), , (
2 2 1 1
y x Q y x P = =
We have

as it should be.
These two theorems give us:
Theorem 5.2.3: In the Cartesian model, the ruler postulate holds.

5.3. Incidence and Parallelism
Theorem 5.3.1: Every two points of the Cartesian model lie on a line.
Proof: Given ). , ( ), , (
2 2 1 1
y x Q y x P = = If ,
2 1
x x = then P and Q lie on the vertical line
.
1
x a x = = If not, then P and Q lies on the graph of the equation
) (
1
1 2
1 2
1
x x
x x
y y
y y
=
which is easily seen to be a line.
Theorem 5.3.2: Two lines intersect in at most one point.
Proof: Given L
1
and L
2
with L
1 =
L
2
. If both are vertical, then they do not intersect at all. If
one is vertical and the other is not, then the graphs of
b mx y a x + = = ,
Intersect at the unique point ). , ( b ma a + Suppose finally, that L
1
and L
2
are the graph of
. ,
2 2 1 1
b x m y b x m y + = + =

If
2 1
m m = , very elementary algebra gives us exactly one common solution and hence exactly
one intersection point. If
2 1
m m =
,
then
2 1
b b = ,and the graphs do not intersect at all.
We have already observed that if L is the graph of y=mx+b, then for every two points
( ) ( )
2 2 1 1
, , , y x y x of L., we have
.
1 2
1 2
m
x x
y y
=

Thus m is determined by the non-vertical L. As usual, we call m the slop of L.
Theorem 5.3.3: Every vertical line intersects every non-vertical line.[At the
Point (a, ma +b).]
Proof: Let L
1
be a vertical line and L
2
be a non vertical line
then by theorem 2, L
1
and L
2
intersect at the point (a , ma +b).
Theorem 5.3.4: Two lines are parallel if and only if (1) both are vertical, or (2) neither is
vertical, and they have the same slope.
Proof: Give
2 1
L L = . If both are vertical, then
2 1
L L . If neither is vertical, and they have the
same slope, then the equations
) ( ,
2 1 2 1
b b b mx y b mx y = + = + =
have no common solution, and
2 1
L L .
Suppose, conversely, that
2 1
L L . If both are vertical, then (1) holds. It remains only to show
that if neither line is vertical, they have the same slope.
Suppose not. Then
) ( : , :
2 1 2 2 2 1 1 1
m m b x m y L b x m y L = + = + =
We can now solve for x and y:
), ( ) ( 0
2 1 2 1
b b x m m + =
,
2 1
2 1
m m
b b
x
=
.
1
2 1
2 1
1
b
m m
b b
m y +
|
|
.
|
\
|
=

We got this value of y by substituting in the equation of
1
L . But our x and y also satisfy the
equation of
2
L
.
This contradicts the hypothesis .
2 1
L L
Theorem 5.3.5: Given a point ( )
1 1
, y x P = and a number m,there is exactly one line which
passes through P and has slope = m.
Proof: The lines L with slope m are the graphs of equations . b mx y + =
If L contains ( ), ,
1 1
y x then ,
1 1
mx y b = and conversely. Therefore our line exists and is
unique.
Theorem 5.3.6: In the Cartesian model, the Euclidean parallel postulate holds.
Proof: Given a line L and a Point ( )
1 1
, y x P= not on L.
1. If L is the graph of x =a, then the line
1
: x x L = ' is the only vertical line through P, and,
by Theorem 3, no non-vertical line is parallel to L. Thus the Parallel L through P is
unique.
2. If L is the graph of , b mx y + = then the only parallel to L through P is the line through
P with slope= m. This is unique.

5.4 Translations and Rotations
By a translation of the Cartesian model, we mean a non-to-one correspondence

). , ( ) , ( :
. :
b y a x y x
E E f
+ +

Merely by substitution in the distance formula, and observing that a and b cancel out, we
have:
Theorem 5.4.1: Translations are isometries.
If L is the graph of the equation
, 0 = + + C By Ax
then the points ) , ( ) , ( b y a x y x + + = ' ' of ) (L f satisfy the equation
, 0 ) ( ) ( = + ' + ' C b y B a x A
or 0 ) ( = + + ' + ' C bB aA y B x A
This is linear. Thus we have proven the theorem.

Theorem 5.4.2: Translation preserves lines.
Since translations preserve lines and distance, they preserve everything defined in terms of
lines and distance.
Theorem 5.4.3: Translations preserve betweenness, segments, rays, angles, triangles, and
angle congruences.
Rotations are harder to describe, because at this stage we have no trigonometry to work with.
Let us first try using trigonometry, wishfully, to find out what we ought to be doing, and then
find a way to do something equivalent , using only the primitive apparatus that we now have
at our disposal in our study of the Cartesian model.

Figure 5.2
We want to rotate the Cartesian model through and angle of measure | (Fig. 6.2).
Trigonometrically, this can be done by a one-to-one correspondence,
, : E E f
defined as the labels in the figure suggest.
Now . sin sin cos cos ) cos( | u | u | u = +
. sin cos cos sin ) sin( | u | u | u + = +
Let | | sin , cos = = b a
Now ,
2 2
y x r + =
, cos
2 2
y x
x
+
= u
, sin
2 2
y x
y
+
= u


We can therefore rewrite our formulas in the form
), , ( ) , ( : y x y x f ' '

Where ) ( cos | u + = ' r x

|
|
.
|
\
|
+
+
+ = b
y x
y
a
y x
x
y x
2 2 2 2
2 2

, by ax =
And
|
|
.
|
\
|
+
+
+
+ = ' b
y x
x
a
y x
y
y x y
2 2 2 2
2 2

. bx ay+ =
Any correspondence of this form, with 1
2 2
= +b a , is called a rotation of the Cartesian model.
Theorem 5.4.4: Rotations preserve distance
Proof: We have
. ) , (
1 1
y x P=
, ) , (
2 2
y x Q=
), , ( ) (
1 1 1 1
bx ay by ax P f P + = = '
), , ( ) (
2 2 2 2
'
bx ay by ax Q f Q + = =

It is merely an exercise in patience to substitute in the distance formula, calculate
, Q P ' ' simplify with the aid of the equation , 1
2 2
= +b a and observe that . PQ Q P = ' ' (The
reader is warned that ( )
2
Q P ' ' appears as a sum of twenty terms. )Solving for y and x in terms
of y and x ' ' , we get
x b y a y y b x a x ' ' = ' + ' = ,
Comparing the formulas
ay bx y by ax x + = ' = ' ,
For f and the corresponding formulas for ,
1
f we see that these have the same form
, ,
' '
x b y a y y b x a x ' ' + ' ' = ' ' =

Where . b b and a a = ' = '

Therefore we have the following theorem.
Theorem 5.4.5: The inverse of a rotation is a rotation.
Theorem 5.4.6: Rotations preserve lines.
Proof: L is the graph of an equation
, ) 1 ( k x=
(2) , k y =
Or (3) ). 0 ( = + = m k mx y

In Case (1), If (L) is the graph of
, k y b x a = ' + '
Where a and b are not both = 0, because 1
2 2
= +b a . Therefore L is a line
In Case (2),f(L)is the graph of
. , line a is which k x b y a = ' '
In Case(3),f(L) is the graph of
, max
' '
k mby x b y a + + = ' '

Or
. 0 ) ( ) ( = + ' + ' + k y a mb x b ma
If we had both
, 0 = +b ma 0 = a mb ,
Then 0 , 0
2 2
= = + ab mb b a ma
So that , 0 ) (
2 2
= +b a m
and m= 0, contradicting our hypothesis.
As for translations, once we know that rotations preserve lines and distance, it follows that
they preserve everything that is defined in terms of lines and distance.
Therefore we have:
Theorem 5.4.7: Rotations preserve betweenness, segments, rays, angles, triangles, and angle
congruences.


We are going to use rotations in the Cartesian model in much the same way that we used
reflections in the Poincare model, to show that postulates for angle congruence hold. To do
this, we shall need to know that every ray starting at the origin (0, 0) can be rotated onto the
positive end of the x-axis, and vice versa. By Theorem 5, it will be sufficient to prove the
following theorem.
Theorem 5.4.8: Let , ) , ( , ) 0 ( ) 0 , (
1 1 0 0
y x letQ x x P = > =
and suppose that

2
1
2
1 0
y x x + =
.
Then there is a function f such f(P)=Q

Figure 5.3
The equation in the hypothesis says, of course, such P and Q are equidistant from the origin.
As a guide in setting up such a rotation, we note unofficially that we want to rotate E through
an angle of measure , where

Thus the rotation ought to be

Where

Obviously 1
2 2
= +b a in these equations, and so f is a rotation. And
) , ( , ) 0 , (
1 1 0
2
1
2
1
1
0
2
1
2
1
1
0
y x x
y x
y
x
y x
x
x f =
|
|
.
|
\
|
+ +
=
which is the result that we wanted.

5.5. Plane Separation
We shall show first that the plane-separation postulate holds for the case in which the given
line is the x-axis. It will then be easy to get the general case.
Definition 5.5.1: A subset E
+
of the plane E is convex if, whenever P and Q are two points of
E
+
, then the line segment PQ joining P to Q is also contained in E
+
.
Definition 5.5.2: The two non-empty convex sets E
+
and E
-
formed by removing the line
L from the plane are called half planes, and the line L is the edge of each half plane.
Let E
+
be the upper half plane. That is,
E
+
= ( ) { }. 0 , > y y x

Theorem 5.5.1: E
+
is convex.
Proof: Remember that, says that if A, B, and C are points of a line, with coordinates x, y and
z, and , z y x < < then A-B-C. (This was proved merely on the basis of the ruler postulate, and
we can therefore apply it now.) Since only one of the points A,B,C is between the other two,
the lemma has a true converse: if A-B-C, then . x y z or z y x < < < <
Consider now two points, A =
+
= E of y x C y x ) , ( , ) , (
2 2 1 1

Figure 5.4

We need to show that AC lies in E
+
. That is, if A-B-C., with ), , (
3 3
y x B= then . 0
3
> y
Obviously, for the case
2 1
x x = we may assume that
2 1
x x <
,
as in figure; and for the case
x
1
=x
2,
we may assume that .
2 1
y y >
In the first case, the line ACis the graph of an equation
. b mx y + =
And has a coordinate system of the form
) , ( y x f = . 1
2
x m +
In the second case, the line is the graph of the equation

1
x x =

And has a coordinate system of the form
. ) , ( y y x f =
It is easy to check that in the first case
). ( ) ( ) ( C f B f A f < <
So that
.
2 3 1
x x x < <
For . 0 > m
;
2 3 1
b mx b mx b mx + < + < +
For , 0 < m the inequalities run the other way; but in either case
2
y lies between two positive
numbers. In the second case , ) (
2 1
x x = the same result follows even more easily.
Let

E be the lower half plane. That is.
}. 0 ) , {( < =
y y x E
Since the function,
. ) , ( ) , ( : y x y x f
Is obviously an isometry, it preserves segments. Therefore it preserves convexity. Since
, ) (
+
= E E f we have the following theorem.

Theorem 5.5.2:
E is convex.
It is an easy exercise in algebra to show that if
+
e = e = E y x B and E y x A ) , ( . } , {
2 2 1 1
then
ABcontains a point ) 0 , (x of the . axis x Thus:
Theorem 5.5.3: E and the line 0 = y satisfy that conditions for E and L in the plane separation
postulate.
Now le L be any line in E, and let ) , (
1 1
y x A= be any point of L. By a translation f, we can
move A to the origin. By a rotation g, we can move the resulting line onto the x-axis. Let
. ) ( ), (
1 1
2
1 1
1
+
= = E f g H E f g H

Since all of the conditions of the plane separation postulate are preserved under isometries, we
have the following theorems.
Theorem 5.5.4: E satisfies the conditions of the plane separation postulate
Theorem 5.5.5: Isometries preserve half planes
Proof: Let
1
H be a half plane with edge L, and let H
2
be the other side of L. If f is an
isometry, then f(L) is a line L'
. Let
). ( ), (
2 2 1 1
H f H H f H = ' = '
The
1
H' and
2
H' are convex, and every segment between two points
1
) ( H of A f ' and f(B) of
2
H' must intersect f(L). Therefore
1
H' is a half plane with L'
as edge.
From Theorem 5 it follows that:
Theorem 5.5.6: Isometries preserve interiors of angles.
That is, if I is the interior of , ABC Z then ) (I f is the interior of ). ( ABC f Z

5.6. Angle Congruence
We want to verify that angle congruence, defined by means of isometries of E onto itself,
satisfies the postulates of angle congruence, and also satisfies SAS. Only one of these
verifications is trivial.


Statement 1: For angles, congruence is an equivalence relation.
Proof:
(1) A A Z ~ Z always, because the identity function E E is an isometry
(2) If , B A Z ~ Z then , A B Z ~ Z because the inverse of an isometry is an isometry.
(3)If , , C B and B A Z ~ Z Z ~ Z then , C A Z ~ Z because the composition of the isometries for
which C B and B A Z Z Z Z is always an isometry for which . C A Z Z
The other verifications are more difficult. We begin with a lemma.
Lemma 5.6.1: Let f be an isometry of E onto itself. If P P f and E E f = =
+ +
) ( , ) ( for every
point P of the axis x , then f is the identity.
Proof:. Let A be the origin (0,0), and let . ) 0 , 1 ( = B Let ) , ( b a Q= be any point, and let
). , ( ) ( d c Q f = Then
). ( ) ( ), ( ) ( Q f B f BQ Q f A f AQ = =
Taking the square of each of these distances, we get
,
2 2 2 2
d c b a + = +
, ) 1 ( ) 1 (
2 2 2 2
d c b a + = +
, 1 2 1 2
2 2 2 2
+ + = + + c d c a b a
So that c a = . Therefore .
2 2
d b = Since d and b E E f , ) (
+ +
= are both positive, both zero,or
both negative. Therefore b =d. Thus f(Q)=Q for every Q, which was to be proved.
Lemma 5.6.2: Let A be the origin; Let ) 0 ( , ) 0 , ( > = a a B be a point of the
) , ( ) , ( ; e d D and c b C let and axis x = = be points of
+
E and E such that
AC=AD, BC=BD.

Then there is an isometry
E E f :
Such that . ) ( , ) ( ) ( , ) ( C D f and D C f and B B f A A f = = = =

Figure 5.5
Proof: We shall show that d =b and e =-c. the desired isometry f will then be the function
). , ( ) , ( y x y x
Given
, ) ( ) (
,
2 2 2 2
2 2 2 2
e a d c a b
e d c b
+ = +
+ = +

We have . 2 2 ad ab = Since , 0 > a this gives . d b= Therefore .
2 2
e c = since
, 0 0 < > e and c we have . c e =
Lemma 5.6.3: Given , ABC Z there is an isometry f of E onto itself such that
. ) ( ) ( BA BC f and BC BA f = = That is, the sides of the angle can be in changed by an
isometry.
In the proof, we may suppose that BA=BC, since A and C can always be chosen so as to
satisfy this condition.

Figure 5.6
Let D be the midpoint of . AC Using a translation followed by a rotation, we get an isometry
E E g : such that ) (BD g is the positive end of the x-axis (Fig. 6.7). (First we translate B to
the origin, and then we rotate.) By the preceding lemma there is an isometry , : E E h
interchanging and C and A , ' '

leaving B'
and D
fixed . Let

.
1
hg g f

=
That is, f is the composition of . , ,
1
g and h g Then f is an isometry;
. ) ( , ) ( , ) ( A C f and C A f B B f = = = .

Figure 5.7
It is now easy to verify the rest of our congruence postulates. Oddly enough, the easiest is
SAS. We put this in the style of a restatement.
SAS. Given , , C B A ABC ' ' ' A A and a correspondence
. C B A ABC ' ' '
If (1) , ) 2 ( , B B B A AB ' Z ~ Z ' ' = and(3)BC= , C B ' ' then (4) , A A ' Z ~ Z
(5) . ) 6 ( ,
'
C A AC C C ' ' = Z ~ Z

Figure 5.8
Proof: By hypothesis (2), there is an isometry . , B B E E ' Z Z . By Lemma 3 it follows
that there is an isometry

'
'
' '
'
:
. :
:
:
C B BC
A B BA
B B
E E f


(If the given isometry moves B onto B ' Z Z in the wrong way, then we follow it by an
isometry which interchanges the sides of .) B' Z . From (1) it follows that
). ( ) ( C f C and A f A = ' = '

Therefore , . ), ( ; , ) (
' '
C A AC also C C C f C A A and A f A = Z ~ ' Z Z = ' Z Z ~ ' Z Z = ' Z because f
is an isometry.
This proof bears a certain resemblance to Euclids proof of SAS by superposition.
Statement 2: The angle Construction Postulate: Let ABC Z be an angle, let
'
C B' be a ray,
and let H be a half plane whose edge contains C B ' ' . Then there is exactly one ray A B ' ' ,
with A'
in H, such that . C B A ABC ' ' ' Z = Z
We give the proof merely in outline. It should be understood that all of the functions
mentioned are isometries of E onto E, and that the ray R is the positive x-axis.
(1) Take
1
f so that . ) (
1
R C B f = ' '
(2) Take . ) ( ) (
1 2 2 2
+
= = E H f f and R R f that so f (Of course, if ) (
1
H f is already
+
= E , we let
2
f be the identity.)
(3) Take
1
g so that . ) (
1
R BC g =
(4) Take
2
g so that ) ( ) (
1 2 2
A g g and R R g = is in .
+
E
(5) Let ). (
1 2
1
2
1
1
ABC g g f f x Z = Z

Then x Z is the C B A ' ' ' Z that we wanted.
(6) Suppose that there are two rays A B A B ' ' ' ' ' , satisfying these conditions.

Figure 5.9


Then , ) ( ) (
' '
1 2 1 2
KM A B f f A B f f = ' = ' '

Where K and M are in KM and KL and E
+
are different rays. Since
, MKN ABC LKN Z ~ Z ~ Z
We have
. MKN LKN Z ~ Z
Thus there is an isometry f, of E onto itself, such that
. ) ( MKN LKN f Z = Z
By Lemma 3, f can be chosen so that . ) ( ) ( KM KL f and KN KN f = = It follows that for each
point P of the x-axis, . ) ( P P f = Since isometries preserve half-planes, and f(L) is in E
+
, we
have . ) (
+ +
= E E f By Lemma 1 it follows that f is the identity. This contradicts the hypothesis
. ) ( KL KM KL f = =
Statement 3: The angle addition postulate: If (1) D is in the interior of D BAC ' Z ) 2 ( , is in
the interior of , C A B ' ' ' Z
(3) , ) 4 ( , C A D DAC and D A B BAD ' ' ' Z ~ Z ' ' ' Z ~ Z then (5) C A B BAC ' ' ' Z ~ Z .

Figure 5.10

Proof: (1) By an isometry f, we move AD onto R and B into .
+
E (For this we need a
translation, followed by a rotation and perhaps a reflection ).) , ( ) , ( y x y x
(2) By an isometry , g we move D A ' ' onto R and B' into .
+
E
(3) By the uniqueness condition in the preceding postulate, we know
that ). ( ) ( ) ( ) ( C A g AC f and B A g AB f ' ' = ' ' =

(4) Therefore ). ( ) ( C A B g BAC f ' ' ' Z = Z Therefore, ; C A B BAC ' ' ' Z ~ Z the required
isometry is .
1
f g

Statement 4: The angle Subtraction Postulate: If (1) D is in the interior of D BAC ' Z ) 2 ( , is
in the interior of , C A B ' ' ' Z
(3) D A B BAD ' ' ' Z ~ Z and (4) C A B BAC ' ' ' Z ~ Z , then (5) . C A D DAC ' ' ' Z ~ Z

Figure 5.11
Proof: Let f be the isometry given by (4), so that C A B BAC ' ' ' Z ~ Z . By Lemma 3 we may
suppose that B A AB f ' ' = ) ( and C A AC f ' ' = ) ( . Then surely . ) ( BAD BAD f Z ~ Z The
uniqueness condition in C-7 therefore tells us that D A AD f ' ' = ) ( . Therefore
, ) ( C A D DAC f ' ' ' Z = Z and , C A D DAC ' ' ' Z = Z which was to be proved.

Exercise:
1. Let L be a line and let A, B, and C be three distinct points of L with coordinates x, y,
and
z, respectively. If the point B is between the points A and C, then the number y is
between the numbers x and z.

2. Show that there are at least three points in a plane E which are not contained in any
single line.
3. Given two distinct points p = (x1, y1) and Q = (x2, y2), then show that there is exactly
one line L in a plane E containing both points.
4. Suppose L is a line in the real Cartesian plane defined by the equation

Ax + By + C = 0 and f : L R
2
is the function given by

f((x, y)) =
(a) Prove that f is a coordinate system on L.
(b) If every line L in R
2
is given the coordinate system L as defined in
(a), prove that the distance function defined on R
2
is the standard
distance function studied in analytic geometry:
d((x
1
, y
1
), (x
2
, y
2
)) =
5. Let L be a line and let A, B,C be three distinct points of L with coordinates
x, y, z, respectively. Then the point B is between the points A and C if
and only if the number y is between the numbers x and z.

Geometry Lecture Note

Caricato da

Informazioni sul documento

Descrizione originale:

Copyright

Formati disponibili

Condividi questo documento

Condividi o incorpora il documento

Opzioni di condivisione

Hai trovato utile questo documento?

Questo contenuto è inappropriato?

Copyright:

Formati disponibili

Geometry Lecture Note

Caricato da

Copyright:

Formati disponibili

AAU

A int . (See fig. 1.6.2)

is the set of all points which are neither on

int . The exterior of an angle

A ext . (See Fig. 1.6.3)

int does not intersect B O A

int and the other in the ( ) B O A ext

, and AC are two sides and the included angle while

are angles such that G and H are in the int )

are angles such that point G is the int )

are supplementary angles.

are vertical angles and

are angle such that

are angles such that F E D C B A

is an angle and is a line on any plane and ED is a ray on then there is

. That is it is not possible for C B A

and a line h containing point O. Then we have at most two

are on the same half plane determined by

A and DE AB (hypothesis). Consequently, DEF ABC A A , by SAS.

. Now since E D DE ' and

(base angles of isosceles triangle are congruent). So, F D E F D E '

B and DF AC it follows that DEF ABC A A by SAS.

is a right angle as AB DF and C E D

D . Prove that AD bisects C A

B . Which two triangles in the

B is a perpendicular bisector of AB PX , BC and

A and ED BC . Prove that EC BD .

A and B is the midpoint of EC . Prove that

A . Prove that BD is the perpendicular bisector

is said to be less than angle F E D

if and only if there exists a ray EG such

, we have only the following possibilities and

is a remote exterior angle

is not less than C A B

is not less than C A B

B < . Analogously, it can be shown that

is an exterior angle of CDB A and hence C D A C B A

A , is equal to the angle of reflection, D E

. The difference between this and 180

CP are the critical parallels to

BC from A, so that ABC BCD ~< < and CD = AB.

AC be a longest side of ABC A , and let

BDbe the altitude from B to

AB so that DE AG = ; and we take H and

AC so that AH = DF. We then

Potrebbero piacerti anche