Geo Me Trik On Project

Geometrikon
Paris Pamfilos
Copyright: Paris Pamfilos 2011-2017

ii
iii
To Kostas, to Michalis, to Iasonas, to Odysseas

iv
Preface
If I had to live my life again, I would have made a rule to read some
poetry and listen to some music at least once every week; for perhaps
the parts of my brain now atrophied would thus have been kept active
through use. The loss of these tastes is a loss of happiness, and may
possibly be injurious to the intellect, and more probably to the moral
character, by enfeebling the emotional part of our nature.
C. Darwin
All things are mutually intertwined, and the tie is sacred, and scarcely
anything is alien the one to the other. For all things have been ranged
side by side, and together help to order one ordered Universe.
Marcus Aurelius, To himself VII, 9
This book is the result of processed notes from courses in Geometry, Euclidean Geometry, Geometry at
School and Geometry and Computers, which I repeatedly taught during the last twenty five years at the
University of Crete in Herakleion Greece. Although the book is intended for school teachers and courses,
the material it contains is much more extensive than what can be naturally taught in school classes. The
material however is developed gradually from simple and easy to more complex and difficult subjects. This
way, in the beginning chapters, I even avoid using negative numbers and the notion of transformation, so
that the book can be used in all school courses.
In its content and organization the whole work complies with the philosophy of having one book of
reference for every school course on a specific subject: the book of Geometry, the book of Physics, the
book of Chemistry, etc. If not for the student, at least for the teacher. The book’s intention is to offer a solid
and complete foundation to both student and teacher, so that both can consult it for studying, understanding
and examining various problems and extensions of elementary geometry.
With the book material I suggest, I do not intend a structural exposition, but move instead towards a
more or less traditional, synthetic method, which aims to familiarize the reader with the basic notions and
which provides for a first contact with the shapes and the problems related to them. A more structure ori-
ented exposition would emphasize the axioms and the algebraic and analytic structures, associated with this
material. Mathematics has definitely passed from the art of calculations to the discovery and investigation
of structures. Although this is a long term accepted development, I do not necessarily consider this path ap-
propriate for a beginning and an introduction to Geometry, its related notions, the elementary simple shapes
and the problems they lead to. I reckon that the pupil must first have some elementary experience and in-
coming impressions of the simplest possible kind, without the interjection of notions of abstract structures,
which in my opinion at a beginning stage would make the student’s approach more difficult.
vi
Thus, so that the reader has a reference point, I spend a minimal time with axioms and I proceed
quickly towards their logical consequences, so as to speed up the reader’s contact with more complex and
more interesting shapes. These axioms represent very basic properties, some of which could conceivably be
proven from others, even simpler. Such a practice however would cause as a result a slightly more involved
discussion on trivial consequences and conclusions, which I view as boring and repelling the pupil from
the course. A discussion of the foundations of geometry and their detailed listing, from which all others
follow, is the work of a late wisdom and must be done after one starts to love the material. At first one
must examine what exactly is this matter, start from practice and, slowly, depending on one’s interest and
capabilities, proceed to theory. This book therefore has an introductory character, that of the first contact
with an area of knowledge, which has many levels.
I think that a teacher can use this book as a reference and roadmap of Geometry’s material in all ge-
ometry courses. Individual courses for specific classes, must and should be supported with companion aids
of instructional character (practical exercises, additional exercises, drawing exercises for consolidation of
ideas, spreadsheets, drawing software, etc.). The book contains many (over 1400) exercises, most of them
with solutions or hints, and with subjects related to the section in which they appear. Often one may also
find exercises which are related to previous sections, giving the opportunity of a quick review of the learned
material.
In writing this book, my deeper desire is to see Geometry return to its ancient respectable place in
the school. This, because, Geometry, with all these shapes, which are the main motivation and help for
inductive thinking, offers many tangible pedagogical benefits. I will mention four main ones.
The first is the realization that there are things in front of you, which you do not see. Simple things,
simple relations exposed to public view, initially invisible, which start to reveal themselves after lots of work
and effort. Attention therefore increases, as does capability of correct observation. To anyone who asks
“did I miss something” the answer is everywhere and always, “many (things)”. Not posing this question,
avoiding to answer it or answering carelessly, is completely incompatible with Geometry’s pedagogy.
The second is the all encompassing power of detail, in other words, the accuracy of thought. Real, cre-
ative work, means involving yourself with details. With Geometry, this is accomplished with the exercises.
Good intentions, visions and abstractions are void of content, when they are not the careful distillation of
the spirit of details. Speaking on generallities is characteristic of rhetorical speech, the art of words. It is not
a coincidence that Geometry has been pushed aside while flourishing the rhetorical and political speech.
The third and most important is the very essence of thought, as well as the human’s character, con-
sistency. Mathematics and Geometry especially, with the assistance of its figures, is the great teacher of
consistency. You begin from certain notions and basic properties (axioms) and start building effectively
ad infinitum, without ever diverging from these and the simple rules of logic. This way the work done is
always additive towards a building of absolute authority, which has nothing to do with lame improvisations
and products, which are the result of continuously changing rules. Man, often, to please his desire, arbitrar-
ily changes the rules of the game. He creates thus a certain culture in which, he is either an abuser, when he
himself changes the rules, or a victim, when he suffers from external arbitrary rule changes. This way the
work done is sometimes additive and other times negative, canceling the previous work done. This mode is
the widespread culture of non-thought, since clear thought is virtually synonymous to mathematical thought
and mathematical prototyping, the discovery and the respect of accepted rules.
The fourth and very crucial is the acquaintance with the general problem of knowledge, and the balance
between the quantity and quality. In the process of learning Geometry we realize with a particular intension
the infinity of the directions towards it extends, but also the unity and the intimate relations of its parts. There
results a question of approach, a question of psychology, a question of philosophy. How you can approach
this whole, an immense body of knowledge? The question is crucial and posed at an early age. A correct
or wrong attitude sets the foundation for a corresponding evolution of the whole life of the student. A big
part of the failure of the learning is due to the misconceptions about its nature. If we look at the established
practice, we realize that the mainstreem behavior is that of the hunter. You target the language, history,
chemistry, etc. you find the crucial paths and passages and you shoot. Unfortunately though, knowledge
consists not of woodcocks. Any subject we may consider consists not of isolated entities. It is not a large
herd of birds containing some rare and some exotic individuals. It is rather a connected continuous and
consistent body, accepting only one approach, through the feeling. You cannot learn something if you do
not approach it with positive feeling. Besides, every subject of knowledge is similar to a musical instrument.
vii
And as you cannot learn 10 instruments at the same time, so you cannot learn 10 subjects at the same time.
The 10 instruments you can touch, put out of their cases, taste their sounds. You have to choose though and
indulge into one. This is the characteristic property of the thinking man. He can indulge into his subject
withdrawing from everything else. Like the virtuoso of the musical instrument absorbed to its music brings
to unity the instrument, the music and his existence.
If we were asked to set the targets of education, one of the principals would be the ability to indulge into
a subject. This, though it is developped by a mental process, it has an emotional basis. The teacher at the
school, the college, the university, meets invariably the same stereotyped problem. The failure of the sudent
is not due to missing mental forces. It is due to underdeveloped or totally missing emotional basis for its
subject. The failure of the teacher is not due to what he teached or left aside. It results from not recognizing
the role of the emotional basis, not highlighting it, not cultivating it.
Motivated by the above ideas I proceed with my proposition, offering this foundation for the organiza-
tion of lessons as simple as possible or, at any rate, as simple as the subject allows. I wish that the teacher
and the studious pupil will read the book with the same or even greater satisfaction and pleasure than I had,
processing for a long time the exquisite material. A material perfect and timeless, the only responsible for
imperfections and mistakes being myself.
I wish to thank the colleagues Georgia Athanasaki, Ioanna Gazani as well as Manolis Katsoprinakis
for the correction of a plethora of mistakes and suggestions for improvements. I thank also Dimitris Kon-
tokostas for his numerous interventions on the first chapters of the book and the supervisors of the two greek
editions, John Kotsopoulos and John Papadogonas. I thank also my colleagues Stylianos Negrepontis and
George Stamou for their encouragement to continue my work and the colleague Antonis Tsolomitis for his
help on “latex”. Finally, I wish to express my gratitude to my colleague Giannis Galidakis for his assistance
in the translation from the Greek. Regarding my sources, I have included all bibliographical references,
which existed in my notes and were used to backtrack and complete the material. I believe they will be
useful to those who wish to dig deeper and compare with other sources. It is especially interesting for one
to search and enter in discussion with great minds, who came before and created material in the subject.
Looking back at the times of school, when I was initiated to Geometry by my excellent teachers, like
the late Papadimitriou, Kanellos and Mageiras, I wish to note that the book binds, hopefully worthily, to
a tradition we had on the area, which was cultivated, at that time, by the strong presence of Geometry
in high schools and gymnasiums. Many of the problems herein are the ones I encountered in books and
notes of the teachers I mentioned, as well as these found in Papanikolaou, Ioannidis, Panakis, Tsaousis.
Several other problems were collected from classic Geometry works, such as Lalesco [Lal52], Legendre
[Leg37], Lachlan [Lac93], Coxeter [Cox68], [CG67], Hadamard [Had05], the well-known book of the
Jesuits (F.G.M.) [F.G52] and many others, which are too numerous to list.
Concluding the preface, I mention, for those interested, certain books which contain historical themes
on the subject [Cou80], [Eve63], [Hea08], [Hea31], [Dan55], [Kno93], [Boy91], [Kat98], [Coo40]. Some
additional references are contained in my older notes.
In the second edition of the book there are substantial changes having to do with the material’s artic-
ulation and expansion, the addition of exercises, figures, and the return of epigrams, which were omitted
in the first edition, because of some mistaken reservations concerning the total volume of the book. Going
through the first edition, I felt somewhat guilty and had a strong feeling that these small connections with
the other non-mathematical directions of thinking and their creators are necessary and very valuable to be
omitted for the benefit of a minimal space reduction. In my older notes I used to put them also at the end or
even the middle of lengthy proofs. I had thus the feeling to be a member of the international and timeless
university and the ability to ask and discuss with these great teachers of all possible subjects. Perhaps some
of the readers will eventually keep in mind some of the many figures of the book and some of the epigrams,
which are a sort of figures in this “invisible geomery”, as the great poet says.
Herakleion, 16 February 2017

viii
Symbol index
(ABΓ ) Circle going through the points A, B, Γ

κ = (A, B, Γ , ∆) Circle going through the points A, B, Γ , ∆
X = (AB, Γ∆) Intersection point X of the lines AB, Γ∆
ΓA ∆A
(ABΓ∆) Cross ratio Γ B : ∆B
(A, B, Γ , ∆) Harmonic tetrad of collinear points
|AB| Length of AB
κ(O, ρ) Circle κ with center O and radius ρ
κ(O) Circle κ with center O
κ(ρ) Circle κ of radius ρ
O(ρ) Circle with center O, and radius ρ
p(X) or p(X, κ) or pκ (X) Power of point X relative to circle κ
’|
|ABΓ Angle measure ABΓ’
O(A, B, Γ , ∆) Pencil of four lines OA, OB, OΓ , O∆
Y = X(A, B) Y is the harmonic conjugate of X relative to A, B
(ABΓ ) Area of triangle ABΓ
(ABΓ ...) Area of polygon ABΓ ...
τ Half perimeter of triangle
α, β, γ Measures of angles of triangle ABΓ
a, b, c Lengths of sides of triangle ABΓ
r, rA , rB , rΓ Radius of inscribed and escribed circles of triangle
AB
Γ∆ √
Signed ratio of sections of the same line
5+1
φ= 2 =∼ 0.61803398874989484820... Golden section ratio
N The set {1, 2, 3, . . .} of natural numbers
Z The set of integer (positive, negative, zero) numbers
Q The set (field) of rational numbers
R The set (field) of real numbers
Contents
1 The basic notions 1

1.1 Undefined terms, axioms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Line and line segment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.3 Length, distance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.4 Angles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
1.5 Angle kinds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
1.6 Triangles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
1.7 The equality of shapes (congruence) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
1.8 Isosceles and right triangle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
1.9 Triangle congruence criteria . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
1.10 Relative measures of triangle angles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
1.11 The triangle inequality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
1.12 The orthogonal to a line . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
1.13 The parallel from a point . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
1.14 The sum of triangle angles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
1.15 The axiom of parallels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
1.16 Symmetries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
1.17 Ratios, harmonic tetrads . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
1.18 Comments and exercises for the chapter . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
2 Circle and polygons 57

2.1 The circle, the diameter, the chord . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
2.2 Circle and line . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
2.3 Two circles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
2.4 Constructions by ruler and compass . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
2.5 Parallelograms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
2.6 Quadrilaterals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
2.7 The middles of sides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
2.8 The triangle medians . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
2.9 The rectangle and the square . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
2.10 Other kinds of quadrilaterals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
2.11 Polygons, regular polygons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
2.12 Arcs, central angles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99
2.13 Inscribed angles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
2.14 Inscriptible quadrilaterals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112
2.15 Circumscribed quadrilaterals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
2.16 Geometric loci . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119
3 Areas, Thales, Pappus, Pythagoras 135

3.1 Area of polygons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135
3.2 The area of the rectangle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137
3.3 Area of parallelogram and triangle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140
3.4 Pythagoras and Pappus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146
CONTENTS xi
3.5 Similar right triangles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154

3.6 The trigonometric functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159
3.7 The theorem of Thales . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165
3.8 Pencils of lines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 170
3.9 Similar triangles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 178
3.10 Similar polygons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 185
3.11 Triangle’s sine and cosine rules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193
3.12 Stewart, medians, bisectors, altitudes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 200
3.13 Antiparallels, symmedians . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 207
4 The power of the circle 233

4.1 Power with respect to circle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 233
4.2 Golden section and regular pentagon . . . . . . . . . . . . . . . . . . . . . . . . . . . . 237
4.3 Radical axis, radical center . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 241
4.4 Apollonian circles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 245
4.5 Circle pencils . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 250
4.6 Orthogonal circles and pencils . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 259
4.7 Similarity centers of two circles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 265
4.8 Inversion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 271
4.9 Polar and pole . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 281
5 From the classical theorems 311

5.1 Escribed and excenters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 311
5.2 Heron’s formula . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 316
5.3 Euler’s circle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 318
5.4 Feuerbach’s Theorem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 322
5.5 Euler’s theorem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 324
5.6 Tangent circles of Apollonius . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 328
5.7 Theorems of Ptolemy and Brahmagupta . . . . . . . . . . . . . . . . . . . . . . . . . . . 333
5.8 Simson’s and Steiner’s lines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 340
5.9 Miquel point, pedal triangle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 344
5.10 Arbelos . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 351
5.11 Sangaku . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 356
5.12 Fermat’s and Fagnano’s theorems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 360
5.13 Morley’s theorem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 364
5.14 Signed ratio and distance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 366
5.15 Cross ratio, harmonic pencils . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 372
5.16 Theorems of Menelaus and Ceva . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 379
5.17 The complete quadrilateral . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 387
5.18 Theorem of Desargues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 393
5.19 Theorem of Pappus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 397
5.20 Theorems of Pascal and Brianchon . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 400
5.21 Castillon’s problem, homographic relations . . . . . . . . . . . . . . . . . . . . . . . . . 407
5.22 Malfatti’s problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 414
5.23 Calabi’s triangle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 417
6 Circle measurement 441

6.1 The difficulties, the limit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 441
6.2 Definition of the perimeter of the circle . . . . . . . . . . . . . . . . . . . . . . . . . . . 445
6.3 The number π . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 447
6.4 Arc length of a circle, radians . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 449
6.5 Definition of the area of the circle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 451
xii CONTENTS
6.6 The area of a circular sector . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 455

6.7 The isoperimetric inequality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 459
6.8 Anthyphairesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 463
7 Transformations of the plane 473

7.1 Transformations, isometries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 473
7.2 Reflections or mirrorings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 475
7.3 Translations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 479
7.4 Rotations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 484
7.5 Congruence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 489
7.6 Homotheties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 492
7.7 Similarities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 495
7.8 Inversions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 504
7.9 The hyperbolic plane . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 506
7.10 Archimedean tilings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 516
8 Lines and planes in space 531

8.1 Axioms for space . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 531
8.2 Parallel planes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 536
8.3 Angles in space . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 538
8.4 Skew lines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 540
8.5 Line orthogonal to plane . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 542
8.6 Angle between line and plane . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 548
8.7 Theorem of Thales in space . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 550
9 Solids 557
9.1 Dihedral angles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 557
9.2 Trihedral angles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 561
9.3 Pyramids, polyhedral angles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 569
9.4 Tetrahedra . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 571
9.5 Regular pyramids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 574
9.6 Polyhedra, Platonic solids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 577
9.7 Prisms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 581
9.8 Cylinder . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 585
9.9 Cone, conical surface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 586
9.10 Truncated cone, cone unfolding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 591
9.11 Sphere . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 593
9.12 Spherical and circumscribed polyhedra . . . . . . . . . . . . . . . . . . . . . . . . . . . . 600
9.13 Spherical lune, angle of great circles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 602
9.14 Spherical triangles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 603
9.15 The supplementary triheder . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 607
9.16 Axonometric projection, affinities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 611
9.17 Perspective projection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 616
10 Areas in space, volumes 631

10.1 Areas in space . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 631
10.2 Area of the sphere . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 637
10.3 Area of spherical polygons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 640
10.4 Euler Characteristic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 643
10.5 Volumes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 645
10.6 Volume of prisms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 647
CONTENTS xiii
10.7 Volume of pyramids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 651

10.8 Volume of cylinders . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 656
10.9 Volume of cones . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 658
10.10Volume of spheres . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 659
10.11Comments and exercises for the chapter . . . . . . . . . . . . . . . . . . . . . . . . . . . 666
11 Conic sections 673

11.1 Conic sections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 673
11.2 Dandelin’s spheres . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 676
11.3 Directrices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 679
11.4 General characteristics of conics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 686
11.5 The parabola . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 693
11.6 The ellipse . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 700
11.7 The hyperbola . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 709
12 Transformations in space 733

12.1 Isometries in space . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 733
12.2 Reflections in space . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 735
12.3 Translations in space . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 738
12.4 Rotations in space . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 739
12.5 Congruence in space . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 742
12.6 Homotheties in space . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 748
12.7 Similarities in space . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 749
12.8 Archimedean solids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 751
12.9 Epilogue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 757
Bibliography 759
Index 769
xiv CONTENTS
CHAPTER 1
The basic notions
1.1 Undefined terms, axioms

This is valid even for our own ego: we
understand it only through its properties, not
as something which can exist in and of itself.
Thomas Mann, Schopenhauer
The notions, at least the mathematical, are like the forms of matter, which are split in molecules, these
in atoms, these in elementary particles etc. In geometry the reduction to simpler and more elementary
notions comes eventually to the so called undefined terms. These are notions, simple and familiar from
our experience, for which it is difficult to find even simpler ones, with which to describe them ([Hel76]).
Such notions in geometry are the point, the plane, the space, the line, the notion of the point between two
other points and the notion of equality of two shapes.
We learn to handle these notions using their properties or axioms, which describe some of their char-
acteristics, which we accept without proof. We begin therefore with the undefined terms. We describe their
basic properties with axioms, and from there on, by combining these with logic, we deduce other properties,
theorems or propositions and corollaries (direct logical consequences of theorems). Axioms and theorems,
up to a point, are used to deduce new properties, in other words, new theorems.
Proceeding thus, we build gradually a well-organized and structured spiritual edifice, which represents
our knowledge in geometry and is, potentially, infinitely extendible. If at some point we accept a hypothesis,
for example A = B, and, relying on logic, we conclude that this leads to a contradictory result against some
axiom or some already proven theorem, then we say that our hypothesis has lead us to a contradiction, and
we are then obliged to accept that the negation of the hypothesis must necessarily hold (in this case A = B).
This method of reasoning is commonly called reduction to contradiction (Latin: Reductio ad absurdum)
or reduction to contradiction and is used very often in geometry.
Euclidean Geometry examines the properties of shapes in space and on the plane and, mainly, properties
associated with measurement. As shape we consider any collection of points on the plane (plane shape) or
space (space shape). We measure lengths, angles and areas. In space we also measure volumes. Usually
the lesson is divided into two parts. In the first part, called plane geometry, we examine properties of
plane figures, such as the triangle, the square, the circle etc. In the second part, called space geometry, we
examine properties of space figures, such as the cube, the pyramid, the cylinder, the sphere etc.
Comment-1 The axioms, we will choose as basic properties and starting point of our study, are not really
independent from each other. Some of them are consequences of others, therefore we could conceivably
2 CHAPTER 1. THE BASIC NOTIONS
start with fewer independent axioms, which would be sufficient in proving all the rest as theorems. Such
a practice however would have the disadvantage of spending considerable time on very simple properties,
which could otherwise be proven using our very few axioms.
I have therefore decided to incorporate some of these properties into the axioms, conforming to the
philosophy, that the revelation of the more hidden properties generates more interest than validating the
obvious ones. For an alternative course, where the subject of axioms is examined in detail, one may consult
the well known book of Hilbert (1862-1943) on the Foundations of Geometry [Hil03], which is devoted
solely to the discussion of axioms, their independence and their non-contradictory nature, their so-called
consistency. Most line axioms I formulate further ahead are taken from this book. I have replaced however
some of these axioms with some from Birkhoff’s (1884-1944) system ([Bir32]), which guarantee that lines
are, essentially, copies of the set of real numbers.
In any case, let it be noted that Euclidean Geometry’s foundations may be laid using very few axioms.
Hilbert in the aforementioned book, as well as Cairns (1904-1982) ([Cai33]) present systems with only four
axioms. Bachmann (1909-1982) ([Bac73]) presents a system with five axioms. Note however that these sys-
tems are involved with more complex mathematical structures (topological spaces, transformations, groups
etc.).
Comment-2 Euclid’s Elements (approximately 325-265 B.C.) ([Hei85], [Hea08]) begin with the presenta-
tion of 23 definitions, 4 of which and the last one are given as:
(1) A point is that of which there is no part.
(2) And a line is a length without breadth.
(3) And the extremities of a line are points.
(4) A straight-line is (any) one which lies evenly with points on itself.
...
(23) Parallel lines are straight-lines which, being in the same plane, and being produced to infinity in each
direction, meet with one another in neither (of these directions).
Immediately after the 23 definitions follow 5 Postulates which we name axioms:
1. Let it have been postulated to draw a straight-line from any point to any point.
2. And to produce a finite straight-line continuously in a straight-line.
3. And to draw a circle with any center and radius.
4. And that all right-angles are equal to one another.
5. And that if a straight-line falling across two (other) straight-lines makes internal angles on the same
side (of itself whose sum is) less than two right-angles, then the two (other) straight-lines, being
produced to infinity, meet on that side (of the original straight-line) that the (sum of the internal
angles) is less than two right-angles (and do not meet on the other side).
These definitions are examples of notions we have called undefined (1, 2, 4), as well as, normal definitions,
in the modern sense (3, 23). Euclid’s five axioms unfortunately are not sufficient in proving all the propo-
sitions which follow in his book. Euclid often tacitly assumes some additional properties, which are not
deduced from these axioms, but are correct nevertheless. What is needed, is the addition of more axioms,
so that the final set becomes, what we today call a complete axiomatic system, a system which is capable of
supporting the proofs of all the properties of geometric shapes we discover using logical deduction ([You17,
p. 36]).
In this book I conform to Euclid’s choice to devote relatively little time on definitions and axioms,
because I have observed repeatedly that when a student is exposet to lots of explanations and detailed
analyses of naturally intuitive notions, he usually starts doubting even the known notions and ends up
being confused instead of illuminated. Therefore caution is needed, so that the student’s natural intuition
about previous empirical knowledge is fortified rather than challenged. Following then Euclid, I will not
spend too much time on abstract notions and axioms. I will give a complete system capable of supporting
1.2. LINE AND LINE SEGMENT 3
all subsequent propositions and theorems. Trusting the reader’s intuition however, I will not especially
discuss the interdependence of these axioms or the undefined terms these axioms refer to ([You17, p. 165],
[Log80]).
1.2 Line and line segment

I think simplicity, speaking accurately, resembles very much the
non-being. The great inability of my mind consists in that it
does not have an instrument fine enough with which to grasp
this. Someone may claim that the mathematical point is simple.
In reality however, the mathematical point does not exist.
Voltaire, The Ignorant Philosopher
The plane consists of points, which we denote using capital letters A, B, Γ , ..., primed capital letters A ,
B , Γ , ... or indexed capital letters A 1 , A2 , ... etc. One of the simplest shapes of the plane is the line which
Figure 1.2.1: Line ε
we denote with small letters ε, ζ, ..., primed small letters ε , ζ , ... or indexed small letters ε 1 , ε2 , ... etc.
For lines we accept the following initial properties (axioms).
Axiom 1.2.1 Two different points A, B define exactly one line denoted by AB.
A B
Figure 1.2.2: Line AB
Axiom 1.2.2 Every line contains infinitely many points. For every line there are infinitely many points on
the plane which do not belong to the line. For every point there are infinitely many lines which do not pass
through this point.
Axiom 1.2.3 Every line divides the plane into two parts called half planes, which have no common points
with this line. A line, which has two points A and B lying in different half planes of the line ε, intersects the
line ε (the first theorem below establishes that there exists exactly one intersection point). We often use the
word side of line, meaning one of the two half planes defined by it.
A B
Figure 1.2.3: Half planes defined by a line

Axiom 1.2.4 Two points A, B of a line ε define a line segment denoted also by AB. AB consists of A,
B as well as all the points between A and B. A and B are called end points of the line segment. The
points of the line segment, excluding the end points, are called interior points of the segment. The set of
all interior points is called the interior of the line segment. The points of the line AB not belonging to the
segment AB are called exterior points of the segment.
A B
Figure 1.2.4: Line segment AB
Axiom 1.2.5 If the two points A and B belong to the same half plane defined by the line ε, then all points
of the line segment AB are contained in the same half plane. If the points A and B belong to different half
planes defined by the line ε, then the intersection point E, of the line ε and the line AB, lies between A and
B.
B
ε E
Figure 1.2.5: A and B on different half plane of the line ε
Comment-1 In Axiom 1.2.4 the word between is temporarily ambiguous. It will become clear in the next
section with the help of the notion of length of a line segment.
Comment-2 The simultaneous usage of the symbol AB for both, the line segment it represents, and the
line defined by points A and B, should not confuse us. Each time the meaning of the symbol will be clear
from context. Often we will denote the line ε = AB, considering that this symbol represents the phrase the
line ε defined by the points A and B. We may also consider the line segment AB as defining a direction
along the line AB with A being the start and B being the end of the segment AB.
A' B'
A B
Figure 1.2.6: Parallels AB and A B
Parallel we call two lines which do not intersect. A line containing a segment AB is called carrier of
the line segment. We call two line segments parallel when their corresponding carriers are parallel lines.
Secant of a line ε, is called a line ε , different from ε, which intersects ε.
ε'
Α ε
Figure 1.2.7: Intersecting lines ε and ε

1.3. LENGTH, DISTANCE 5
Proposition 1.2.1 Two different lines are either parallel or they intersect at exactly one point.
Proof: In Proposition 1.13.1 we will see that there do exist parallel lines. If the two lines ε and ε do not
intersect, then they are by definition parallel. If they do intersect they will have only one common point
A. If they had a second intersection point B, different from A, we would have two different lines ε and ε
passing through the two points A and B, which is impossible, as this contradicts Axiom 1.2.1, q.e.d.
Exercise 1.2.1 Given a line ε, show that if the line segment AB does not intersect the line ε, then the points
A and B are contained in the same half plane defined by ε.
Β
Α
ε
Figure 1.2.8: A, B on the same side of ε
Hint: Use reduction to contradiction. Suppose that AB does not intersect the line ε and points A, B are
contained in different half planes of ε. Then according to Axiom 1.2.5 the line segment AB will intersect
line ε at a point E, contradicting the supposition.
Exercise 1.2.2 Show that for each point O of the plane, there are infinitely many lines passing though it.
ε
X Y Z
Figure 1.2.9: Infinitely many lines through O
Hint: Consider a line ε, which does not pass through O. According to Axiom 1.2.2 there is at least one such
line. Next, define the lines OX, OY , ... etc., which pass through O and through the points respectively X,
Y ,..., Z of ε. Each of these points defines a different line through O.
1.3 Length, distance

No one can read two thousand books. In the four hundred years I have lived, I’ve
not read more than half a dozen. And in any case, it is not the reading that matters,
but the rereading. Printing, which is now forbidden, was one of the worst evils of
mankind, for it tended to multiply unnecessary texts to a dizzying degree.
J. L. Borges, A Weary Man’s Utopia
The axioms of this section relate lines with real numbers through the notion of distance of two points,
clear the notion of point between two other points, as well as the notion of the line segment AB, which
consists of all the points between A, B.
Axiom 1.3.1 Every pair of points A and B defines a real number |AB| ≥ 0, which we call distance of the
points and which satisfies the properties |AB| = |BA| and |AB| = 0, if and only if the two points coincide.
Α Ε Β
Figure 1.3.1: |AB| = |AE| + |EB|
Given two points A and B, we say that the point E is between them or between points A and B (Figure
1.3.1), when it is contained in the line of A, B and holds
|AE| + |EB| = |AB|.
Axiom 1.3.2 For every three different points A, B and E on the same line, one of the three necessarily lies
between the other two. If E lies between A and B then |AB| = |AE| + |EB|. Conversely, if the previous
relation holds then E lies between A and B.
δ δ
ε' ε''
B' Α B''
Figure 1.3.2: Points at distance δ from the startpoint of opposite half lines
Axiom 1.3.3 A point A on line ε separates the line in two parts ε and ε , which have only the point A in
common and are called half lines with A as startpoint . For every positive number δ, there is exactly one
point B on ε with |B A| = δ and exactly one point B on ε with |B A| = δ. A is the middle of the line
segment B B .
If the points A, B and B are contained in the same line ε and ε , ε denote the half lines of ε with start-
point A, we say that B, B are in different sides of A, when one is contained in ε and the other in ε
(Figure 1.3.2). We say that B and B are in the same side of A when they are both contained in one of ε
or ε .
We call length of the segment AB the distance |AB| of its endpoints. We call two line segments AB
and Γ∆ of the same or different lines equal, when they have the same length.
The two half lines which are defined by a point A on the line ε are called opposite. We call two half
lines parallel when they are contained in parallel lines.
Comment Line Axiom 1.3.3 means that we can construct a line segment of any length we desire. We
will consider often the problem of practically constructing a line segment of given length, limiting our-
selves to the usage of only two tools: the straight edge and the compass. For example, the construction of
the middle point M of a given line segment AB, using a straight edge and compass, requires knowledge of
the properties of the circle, which we do not have as of yet. Note however that the proof of the existence of
M relying on the aforementioned properties is simple.
Exercise 1.3.1 Let B and E be two points on the same half line AX with startpoint A. Show that if
|AE| > |AB| then B lies between A and E. Conversely, if B lies between A and E then the previous
relation holds.
Α Β Ε X
Figure 1.3.3: B between A and E

1.3. LENGTH, DISTANCE 7
Hint: Suppose B is not between A and E. Then either B is identified with E and therefore |AB| = |AE|
which is a contradiction, or E lies between A and B, which immediately implies |AB| > |AE|, contrary
to the hypothesis.
Exercise 1.3.2 (Doubling the length of line segment) Given a line segment AB, show that on the line AB
there exist points E and Z such that B is the middle of AE and A is the middle of ZB.
Ζ Α Β Ε
Figure 1.3.4: Doubling the length of AB
Hint: Take E on the half line with startpoint B which does not contain A at a distance |AB| from B.
Similarly work for Z.
Exercise 1.3.3 Show that for every line segment AB there exists exactly one point M (the middle of AB),
such that |AM | = |M B|.
Hint: If |AB| = λ then the point M at distance λ/2 from A on the side of B exists using Axiom 1.3.3 and
this is what we want.
Exercise 1.3.4 Show that if two points A and B are on the same side of line ε, then the line segment AB
does not intersect the line ε.
Hint: If AB intersected ε, then the intersection point Γ would be different from A and B, therefore it would
be between them and we have a contradiction in Axiom 1.2.5.
Exercise 1.3.5 Show that a line ε is parallel to ε, if and only if one of the two half planes of ε contains
every pair of different points of ε .
Hint: If we assume there are two points A and B of ε contained on different half planes of ε, then according
to Axiom 1.2.5, ε would intersect ε. Conversely, if one half plane of ε contains all possible pairs of points
of ε then the latter cannot intersect ε, for if it did at A, then A would define two opposite half lines on ε .
Choosing one point on each half line we would find two points of ε on different half planes of ε.
Exercise 1.3.6 Show that points B and Γ of line ε are on the same side of point A of ε, if and only if
|BΓ | = ||AB| − |AΓ ||.
Hint: If points B, Γ are on the same half line of A, then either B will be between A and Γ so that
|AΓ | = |AB| + |BΓ |, or Γ will be between A and B so that |AB| = |AΓ | + |Γ B|. Therefore in both
cases |BΓ | = ||AB| − |AΓ ||, which is what we want. A similar argument proves the converse.
Exercise 1.3.7 Let M be the middle of the line segment AB. Show that if point Γ is on the interior of
AB, then |Γ M | = 12 ||Γ A| − |Γ B||. If Γ is on line AB but outside of segment AB, then |Γ M | =
1
2 (|Γ A| + |Γ B|).
A B Γ Δ Ε ... Ω
1 2 3 4 ...
Figure 1.3.5: Zeno’s paradox
Exercise 1.3.8 In a race from A to Ω , the quickest runner (α :Achilles) can never overtake the slowest (
τ : tortoise), which starts at the middpoint B. When α reaches B, τ has advanced to Γ . When α reaches
Γ , τ has advanced to ∆, and so on. Thus, Achilles will never catch the tortoise (Zeno’s paradox 490-430
BC). How you resolve this paradox?
1.4 Angles
The Master said, “I will not open the door for a mind that is not already striving
to understand, nor will I provide words to a tongue that is not already struggling
to speak. If I hold up one corner of a problem, and the student cannot come back
to me with the other three, I will not attempt to instruct him again.”
Confucius, Analects 7.8
Two half lines OX, OY , with common startpoint O, and not contained in the same line, separate the plane
in two parts and define an acute angle or simply angle and a non-convex angle. Convex angle or simply
Y
(Ι) (ΙΙ)
Y
non-convex angle
angle
P X
O
O X
’
Figure 1.4.1: Angle XOY ’
Non-convex angle XOY
’ and consists of the two half lines OX and OY along with one
angle is called the figure we label as XOY
part of the plane we call interior of the angle. The angle’s interior (Figure 1.4.1-I) is the part of the plane
whose points P satisfy the two properties:
1. P and the half line OY are on the same side of the line OX,
2. P and the half line OX are on the same side of line OY .
Point O is called the angle’s vertex. The half lines OX, OY are called sides of the angle. Non-convex
angle is called the figure again defined by the half lines OX and OY but this time consisting of the rest of
’ and the half lines which define it (Figure 1.4.1-II). This
the plane minus the interior of the acute angle XOY
’ , or exterior of the convex angle XOY
part of the plane we call interior of the non-convex angle XOY ’.
Often we will talk about angles without regard as to whether the angle is convex or not. The exact meaning
of the case will then result from context. In the case where the two half lines, which define an angle, are
contained in the same line we define the following special angles.
180° O X Y
X
O
Figure 1.4.2: Flat angle Null angle
Flat angle or straight angle we call the shape consisting of two opposite half lines. Any of the the two half
planes, defined by the line OX, can be considered as the interior or exterior of the flat angle.
We call Angle null the shape consisting of two identical half lines OX and OY . In this case the angle
has no interior, and its exterior coinsides with the whole plane without the half line OX.
We call Full turn or complete turn or full revolution the shape consisting of two identical half lines
OX and OY . In this case the angle has no exterior, and its interior coinsides with the whole plane without
1.4. ANGLES 9
Y
O X
Figure 1.4.3: Full turn
the half line OX.
The basic properties (axioms) of angles are the following:
’ (convex or not) we associate a number | XOY

Axiom 1.4.1 To every angle XOY ’ | = |Y’ OX| ≥ 0 called
’ | = 0 if and only if the angle is the null angle.
angle measure in degrees . It holds | XOY
Β Υ
(Ι) (ΙΙ)
β Ρ
ω
Ο Χ
ω Ο α
Χ
Α
Figure 1.4.4: Equal angles on two sides of OX ’ | = |XOP

| XOY ’ | + |P
’ OY |
Axiom 1.4.2 For every number ω with 0 < ω < 180 ◦ , there exist exactly two half lines OA, OB on the
’ and XOB
two sides of the line OX such that the angles XOA ’ satisfy |XOA| ’ = |XOB| ’ = ω (Figure
1.4.4-I). A flat angle has measure 180 degrees.
Traditionally the measure ω in degrees is denoted as ω ◦ . Thus, 30◦ means angle of 30 degrees. A 1/60-th
of a degree is called minute of a degree and is denoted by a prime on the number of minutes. A 1/60-th of
a minute is called second of a degree and is denoted by two primes on the number of seconds. This way,
30◦ 23 11 denotes an angle measure of 30 + 23 11
60 + 3600 degrees.
’ and A
Two angles ABΓ ◊ ’ | = |A
B Γ are called equal if and only if, their measures are equal: | ABΓ ◊
B Γ |.
Often, in what follows, we will omit the absolute values and for two equal angles we will simply write
’ =A
ABΓ ◊BΓ .
’ (convex or not) the measures of angles XOY

Axiom 1.4.3 For every point P in the interior of angle XOY ’,
’ ’ ’ ’ ’
XOP and P OY satisfy the equation | XOY | = |XOP | + |P OY |. In all such cases we say that angle
’ is the sum of angles XOP
XOY ’ and P ’OY . Often to declare such a relation we will omit the absolute
’ ’
values and we will write XOY = XOP + P ’OY .
Two angles that have a common vertex and a common side and non-intersecting corresponding interiors
’ and P
(like XOP ’OY of figure 1.4.4) are called adjacent. By drawing a half line, from the vertex to an
interior point of the angle, and using the preceding axiom, we see that non-convex angles have a measure
ω > 180◦ . The axiom guaranties also the existence of the bisector, which is a half line through the vertex
of the angle, dividing it in two equal (adjacent) angles.
’ , there exists exactly one half line

Exercise 1.4.1 (Existence of bisector) Show that, for every angle XOY
’ ’
OZ in its interior, which divides XOY into two equal angles XOZ, ZOY ’ with |XOZ|’ = |ZOY ’| =
’ |/2.
|XOY
Δ
(Ι) (ΙΙ)
Γ
Γ
ω Β
ω
ω
ω
Β Ο Α
Α
Figure 1.4.5: Angle of segments AB and AΓ Sum of equal angles
Angle of two line segments AB and AΓ having a common endpoint A, we call the angle formed by the
corresponding half lines AB and AΓ (Figure 1.4.5-I).
’ . Show that the measure of a full turn is 360 degrees.

Exercise 1.4.2 Find the bisector of a flat angle XOY
’ of measure ω, we construct equal and successive adjacent angles

Exercise 1.4.3 Starting with angle AOB
’ ’
towards the same side BOΓ , Γ O∆, etc. For which measures ω does this procedure applied ν times defines
’ , whose side OΩ coincides with the initial segment OA (Figure 1.4.5-II)?
an angle AOΩ
’ with measure |XOY

Exercise 1.4.4 Consider angle XOY ’ | = α and point P in its interior. Show that
’ | < α. Conversely show that for every positive β < α there exists point P in the angle’s interior
|XOP
’ | = β. Additionally show that all points P are contained in some half line with endpoint
such that |XOP
O.
’ . Show that
Exercise 1.4.5 Consider points A and B contained in the interior of the convex angle XOY
’
every point of line segment AB is contained in the interior of angle XOY . Show that the corresponding
property does not hold for non-convex angles.
Υ
Ζ
Η
ω
Χ' Χ
Ρ Ο
’
Figure 1.4.6: Points in the interior of angle XOY
’ and a point P on the opposite half line OX of OX, as well

Exercise 1.4.6 Given is a convex angle XOY
as also point Z of the half line OY . Show that every point H of the half line P Z, lying outside the line
’.
segment P Z, is contained in the interior of the angle XOY
Hint: By construction, point H is contained in the side of the line OY , in which OX is also contained.
Also points Z, H are contained in the same side of the line OX, because the intersection point P of ZH
with OX is off the line segment ZH.
Comment-1 Angle Axiom 1.4.2 means that we can construct any angle we desire on either sides of a
half line. Like with line segments however, so with angles, the practical construction of an angle with
specific measure, using only the ruler and the compass, is a different problem. Thus, for example, for the
construction (by ruler and compass) of the angle of 60 degrees, we will need again properties of the circle,
1.5. ANGLE KINDS 11
α β γ
ε
Α Β Γ Δ
Figure 1.4.7: Correspondence of line segments - angles
to be discussed later.
Comment-2 It is worthwhile for the reader to observe some common properties between angles and line
segments, especially these that concern the notions of between, adjacent and measure. Figure 1.4.7 shows
how natural such a correspondence is. Using a fixed point O not belonging to the fixed line ε and for every
line segment AB belonging to this line one may construct the angle AOB. ’ Through this correspondence
between line segments and angles the notions I mentioned are transferred from the line to the angles with
vertex O. This way AΓ is the sum of AB and BΓ and the corresponding angle AOΓ ’ is the sum of AOB
’
’ . B is between A and Γ and similarly OB is between OA and OΓ , adjacent line segments
and BOΓ
correspond to adjacent angles and so on.
On the occasion of figure 1.4.7, we can state immediately two problems, but in order to solve them, we
must first learn to use some tools (for their solution see Problem 3.9.7 and 3.8.11).
Problem 1.4.1 Assume that in the figure 1.4.7 the angle AOB’ has constant measure | AOB| ’ = α and
rotates around O. Which rotational position minimizes the length of the corresponding line segment AB?
Problem 1.4.2 Assume that in the figure 1.4.7 the line segment AB slides on line ε by keeping its length
’
constant. Which position of AB maximizes the corresponding angle AOB?
1.5 Angle kinds

Many ignore the figure’s accuracy, they presuppose it, and pay attention
to the proof. Contrary to this, we will never talk about proofs: the most
important job would be to draw completely straight lines, correct, even.
To construct a perfect square, to carve a fully round circle.
Jean-Jacques Rousseau, Émile, or On Education
Two lines OX and OY intersecting at O define four angles. Taken by two they form pairs of vertical
angles, i.e. angles, such that the sides of one are extensions of the sides of the other (Figure 1.5.1). For the
X' Y
ω2
ω3 ω1
O
ω4 X
Y'
Figure 1.5.1: Angles of two lines

÷ and Y
flat angles XOX ÷ OY we have 180◦ = |XOX÷ | = |XOY
’ | + |Y÷ OX |. Also 180◦ = |Y ÷ OY | =
’ ÷ ’ ’ ÷
|Y OX| + |XOY |. Because |XOY | = |Y OX|, we conclude that the opposite vertical angles Y OX and

÷ are equal. Similarly we can also show that XOY

XOY ’ and X◊ OY equal. We have therefore proved:
Proposition 1.5.1 Vertical angles are equal.

Two successive angles adding up to 180 ◦ are called supplementary. In the previous figure every pair
of successive angles consists of supplementary angles. An angle measuring 90 ◦ is called a right angle.
Obviously a right angle is equal to its own supplementary. By extending the sides of a right angle at O, that
is, by considering the opposite half lines of the sides of the right angle, we define four right angles around
this point, every two of which are either supplementary or vertical.
Figure 1.5.2: Orthogonal lines
This way, two lines which intersect at O and form at least one (from the resulting four) right angle, will
necessarily form the other three angles also right. Two such lines are called orthogonal or vertical or
perpependicular. We call an angle XOY’ acute when its measure is | XOY ’ | < 90◦ . We call an angle
Υ Υ
acute obtuse
Χ
Χ Ο
O
Figure 1.5.3: Acute and obtuse angle
’ | > 90◦ . Obviously if an angle is acute its supplementary angle will

obtuse when its measure is |XOY
be obtuse and vice versa. We say that angle α is less/greater than angle β when their measures satisfy:
|α| < |β| (resp. |α| > |β|). Obviously every obtuse is greater than a right angle which in turn is greater
than every acute. Two angles whose measures α, β satisfy α + β = 90 ◦ are called complementary.
Υ Χ' Χ
(Ι) Χ' (ΙΙ) (ΙΙΙ)
Υ
Χ'
α
β β
Υ Χ α Χ Υ' Υ''
Ο Ο Ο
Figure 1.5.4: Supplementary, Complementary, Angles with orthogonal sides
’
Obviously two complementary angles are both acute. A point X in the interior of a right angle XOY
÷ ÷
defines two complementary angles α = XOX , β = X OY (Figure 1.5.4-II).
÷ and Y
Proposition 1.5.2 Two angles XOX ÷ OY which have sides respectively orthogonal are either equal
or supplementary (Figure 1.5.4-III).
Proof: If OY and OY are on the same side of OX , then angles α = XOX ÷ and α = Y ÷ OY are

equal, having a common complementary angle β. If OY and OY are on opposite sides of OX , then the
opposite half line OY of OY forms a supplementary of α = Y÷OY and it is on the same of OX with
◦
OY , therefore according to the previous case 180 − α = α, q.e.d.
1.6. TRIANGLES 13
Corollary 1.5.1 From a point A on the line ε passes exactly one line ζ orthogonal to ε.
ε
Α
Figure 1.5.5: Line ζ orthogonal to ε
Proof: Immediate consequence of Axiom 1.4.2, according to which there exists exactly one angle of 90
degrees with vertex at A, with one side on ε and contained in one of the half planes of ε, q.e.d.
1.6 Triangles
The great trick of regarding small departures from the truth as the truth itself - on which is
founded the entire integral calculus - is also the basis of our witty speculations, where the
whole thing would often collapse if we considered the departures with philosophical rigour.
G.J. Lichtenberg, Aphorisms
The triangle, after the line and the angle, is the simplest figure of the plane. In spite of its simplicity it has
infinitely many properties and is an object of study of well known Mathematicians throughout the ages. The
up to date properties and conclusions are so many, that they compose a special branch of geometry, the so
called Geometry of the triangle ([Lal52], [Gal13], [Yiu13]).
The triangle is the figure defined by three points A, B and Γ , not contained in a line, as well as the
line segments which join the points together. The three points are called vertices of the triangle. The line
A A
(Ι) (ΙΙ)
α Ζ
Ω
c b
Χ
γ Υ
β
B Γ
B a Γ
Figure 1.6.1: Triangle Interior and exterior of triangle
segments defined by pairs of vertices are called triangle sides. The Angles of the triangle are the (convex)
angles formed by the triangle’s sides . The lengths of the sides of a triangle are usually denoted by the Latin
letters
a = |BΓ |, b = |Γ A|, c = |AB|
and the measures of the angles of a triangle by small Greek letters
’ |, β = |Γ
α = |BAΓ ’ ’
BA|, γ = |AΓ β = B,
B| or simply α = A, “ γ=Γ
.
I will use this notation often in subsequent chapters.

’ , BΓ
We say that the angles ABΓ ’ ’
A, Γ AB are respectively opposite to the sides AΓ , BA and Γ B. The
sum of the lengths
σ = a + b + c,
is called the triangle’s perimeter. We often also use the half-perimeter, denoted by τ = σ/2, so that
σ = 2τ .
A triangle is called (1) acute, (2) obtuse, (3) scalene, when respectively, (1) has all its angles acute, (2)
has one obtuse angle, (3) has sides with different lengths.
Two triangles ABΓ and A B Γ are called congruent or isometric or equal, when their corresponding
sides are equal (a = a , b = b , c = c ) and they have respectively equal angles (α = α , β = β , γ = γ ).
The basic properties (axioms) of the triangle are:
Axiom 1.6.1 Every triangle divides the plane into two parts, the interior and exterior. Two points X and
Y contained in the interior of the triangle define a line segment XY , which is contained entirely in the
triangle’s interior (Figure 1.6.1-II). Two points X and Z, one of which is contained in the interior and
the other in the exterior of a triangle define a line segment XZ which either contains a triangle vertex or
intersects exactly one side of the triangle in an interior point Ω of this side.
B B'
Γ'
A Γ A'
Figure 1.6.2: Triangles with corresponding sides equal
Axiom 1.6.2 (of triangle congruence) Two triangles ABΓ and A B Γ having their corresponding sides
equal (|AB| = |A B |, |BΓ | = |B Γ |, |Γ A| = |Γ A |) are congruent. In other words their corresponding
angles are also equal in measure. Moreover opposite of respectively equal sides, the corresponding angles
are equal in measure (Figure 1.6.2).
A A
(Ι) (ΙΙ)
Χ Z
X E
Y
ε Δ
B Γ B Γ
Figure 1.6.3: Pasch’s axiom Intersection on opposite side
Axiom 1.6.3 (Pasch (1843-1940)) If a line ε intersects the side AB of a triangle and does not contain any
of the triangle’s vertices, then it also intersects one of the other sides (Figure 1.6.3-I).
Comment-1 The axiom for the interior and exterior of a triangle is one of the cases I mentioned in the
beginning of the chapter. It can be inferred from the rest of the axioms, consequently it could be proved as a
theorem. The proof however contains details I don’t consider interesting for the student to involve himself
with at this stage. Hence its presentation here as an axiom.
The last axiom, Axiom 1.6.3 looks self-evident, however the property it expresses cannot be inferred from
the previous axioms. Its usefulness can be seen from the next proposition as well as from the exercise
which follows. These two propositions are presented here in order to give a taste of the details one should
pay attention to, when one desires to prove all the ,so to say, obvious properties, relying on the axioms. A
plethora of similar “self-evident” propositions can be seen in [Efi80, p. 42-84], [Bel07].
1.6. TRIANGLES 15
Proposition 1.6.1 If ∆ is an interior point of a triangle, then A∆ intersects the opposite side BΓ of the
triangle (Figure 1.6.3-II).
Proof: Consider a point E on the segment A∆. Next consider the triangle AB∆ and the secant line Γ E.
According to Axiom 1.6.3, Γ E will also meet a second side of the triangle AB∆. Side B∆ of this triangle
’
is outside of angle XΓ A, so Γ E will meet AB at a point Z. Consider then the triangle BΓ Z and the line
AE, which intersects its side of Γ Z. According to 1.6.3 AE will also meet another side of the triangle,
which cannot be BZ, because then AE would coincide with BZ. Therefore AE, which is the same as line
A∆, will meet side BΓ of triangle BΓ Z, which is also a side of the triangle ABΓ , q.e.d.
Exercise 1.6.1 Given a line ε, show that the relation between two points A and B (A and B are contained
in the same half plane of ε) is transitive. In other words if A and B are contained in the same half plane
and B and Γ are also contained in the same half plane, then A and Γ are also contained in the same half
plane.
Β
Α
ε
Ε Ζ
Figure 1.6.4: Meaning of axiom 1.6.3
Hint: Let points A, B be in the same half plane and that B, Γ are also in the same half plane, but A, Γ
are not. Then (Axiom 1.2.5) there exists a point E of line ε belonging to AΓ and lying between A and
Γ . In other words ε intersects AΓ . Because ε does not contain A, B, Γ and intersects one side of the
triangle (AΓ ) it will also intersect by Axiom 1.6.3 another side. If it intersects BΓ at point Z, we have a
contradiction, because then B, Γ will be contained in different half planes of ε. If it intersects AB we get
a similar contradiction. Therefore AΓ cannot intersect ε.
Χ
A
or
sect
r bi
oute
bise
me
atlitude
dia
ctor
B Υ Δ M Γ
Figure 1.6.5: Median, bisector, altitude
Median of the triangle is called the line segment which joins any vertex with the midpoint of the opposite
side. Bisector of the triangle is called the line segment which joins any vertex with the opposite side and
divides the angle at this vertex into two equal parts. We often call also “bisector” the half line or the
entire line, which divides the corresponding angle. Exterior angle of a triangle is called the angle which
is supplementary to a given triangle angle, for example to angle A, and which results by extending one of
the sides of the triangle, like AΓ (by considering the line AΓ ), resulting in angle BAX ’ (Figure 1.6.5).
External bisector is called the bisector of an external triangle angle. Altitude of a triangle is called the line
segment, which joins a vertex with a point belonging to the opposite side, to which it is also perpendicular
(we will guarantee the existence of altitude later in § 1.12). As we will see later, the three triangle medians
meet in point (Theorem 2.8.1), the three bisectors meet in another common point (Theorem 2.2.2) and
finally the three altitudes meet in a third point (Theorem 2.8.2).
A triangle’s medians, altitudes and bisectors (internal and external) are commonly referred to as sec-
ondary elements of the triangle.
Comment-2 Often knowledge of the length of three of these elements is sufficient for the accurate con-
struction of a triangle. For example, in Exercise 2.13.11, we will see that a triangle can be constructed
easily when we know the three lengths |AY |, |A∆| and |AM | corresponding to altitude, bisector and me-
dian from a given vertex. Usually we require the exclusive use of ruler and compass for the construction
of triangles. A relatively complex problem is to prove, that a certain geometric construction is impossible
(with only ruler and compass). For example, given the altitude |AY | and the median |AM | from the same
vertex but the bisector from a vertex different than A, the construction of the corresponding triangle from
these elements has been proven impossible ([Fur37, p. 38]). Note however that non-constructibility, using
exclusively the ruler and compass and having these three as givens does not mean that the triangle is not
constructible by using other means. Thus, for example, given three positive numbers there exists exactly
one triangle having these numbers as its bisector lengths. However this triangle cannot be constructed using
only ruler and compass ([MP94], [Oxm08]).
Exercise 1.6.2 Show that the internal and external bisectors of a given triangle vertex are orthogonal.
1.7 The equality of shapes (congruence)

It is a mistake that equality is the law of nature. Nature does not
produce equality. The supreme law is ordering and dependence.
Vauvenargues, Reflections and Maxims
A point, a line, a half line, a line segment, a triangle, are shapes. More generally, we call (plane) shape any
specific set of points on the plane. The ones we examined, up to now, are the simplest shapes. In the next
lessons we will encounter other more complex shapes and we will study properties which hold for each of
them and are the same for the so called congruent shapes or isometric shapes.
Every shape has a rule, which determines when it is equal to another. Line segments have their length.
They are equal (or congruent or isometric) exactly when they have the same length. The same for angles.
They have their measure. They are equal when they have the same measure. With triangles the notion of
equality contains more ingredients. The definition of congruence in this case requires two triangles to have
equal respective sides and respective angles. Axiom 1.6.2 gives the basic criterion for triangle congruence.
It says that, when two triangles have respective sides equal, then they are congruent. In other words, their
respective angles (the ones opposite of equal sides) will also be equal. Later (§ 1.9) we will see more criteria
for triangle congruence. The more complex the shape, the more elements of it we must compare to conclude
that it is congruent to another shape.
Euclid in his Elements does not waste any time analyzing the notion of congruence. He adopts a simple
notion of equality, under which two shapes are equal, if and only if we can displace one of the shapes
and place it upon the other, so that the two shapes become exactly coincident. But what does it mean to
displace? This notion of displacement is complex. It can be founded uppon the more general notion of
transformation and more specifically that of isometry or congruence, about which we will talk later (§ 7.1).
We initially establish congruence by giving each shape its rule, which dictates when it is congruent to
another. It doesn’t hurt however to think in Euclid’s terms. On the plane two shapes, which are congruent
according to their rule of congruence, are indeed equal according to Euclid’s notion, using displacement and
coincidence. The converse is also true: If two shapes can be moved around or displaced until they actually
become coincident, then they are congruent according to the rules we give for each case. The problem
is that, in order to actually prove this sort of equivalence, we must study various additional topics, whose
presentation here, at this point, would create some confusion. We limit ourselves therefore to Euclid’s rule.
For practical purposes this means that we conceive the plane as a transparent plastic medium and that the
shapes can be cut from their initial drawing place, displaced to the position of the second shape and placed
1.7. THE EQUALITY OF SHAPES (CONGRUENCE) 17
on top of it until coincidence is achieved. The actual precise foundation of the notion of congruence is done
much later in § 7.5, which I recommend for a second reading.
Γ Γ*
B B*
A A*
Figure 1.7.1: Congruent but with opposite orientation
I note a peculiarity of the notion of congruence which reveals itself in figure 1.7.1 and has to do with the
so called orientation of shapes. The two triangles are congruent with our notion of equality. They are
peculiar however, in that the direction of succession A → B → Γ is clockwise, while that of succession
A∗ → B ∗ → Γ ∗ is counterclockwise. Triangle ABΓ is said to be negatively oriented, while A ∗ B ∗ Γ ∗ is
said to be positively oriented. In order to achieve total coincidence between triangles ABΓ and A ∗ B ∗ Γ ∗
under the notion of displacement, we must cut one triangle and turn it upside down, in the same way we
turn a page and read the opposite side. Things become slightly more complex in space, where a similar
phenomenon occurs yet in this case there’s no outside space to perform this turning of orientation. There
the notion of congruence the way we defined it, for each figure separately, is not equivalent to the notion of
coincidence (see for example the comment in § 9.2 and on a second reading the full description of the notion
of congruence in § 12.5). Therefore, the way we handle congruence is safer than that of displacement, as
long as we refrain from going into the details of the exact definition of this notion.
Comment-1 For some shapes, congruence, under the notion of displacement, is obvious. That way, for
example, any two lines α and β are equal, in the sense that α can clearly be displaced and be put onto β, so
that the two lines become coincident. Likewise, two intersecting lines α and β forming an angle of measure
ω make a shape which is congruent with the shape of two other lines α and β making an angle of equal
measure ω.
Comment-2 A comment on notation: Often, with the congruence of two shapes, which have angles, vertices
or other similar characteristics, we create correspondences between their vertices, by denoting correspond-
ing vertices with the same letter and adding an index, prime, star or some other such symbol to the second
vertex to show the implied correspondence. This way, when we say that triangles ABΓ and A B Γ are
congruent because they have their corresponding sides equal, we mean that side AB is respectively equal
to A B , BΓ is equal to B Γ etc. We follow this rule even with space figures. Creating this notational
correspondence in shapes, which are candidates for congruence, becomes the first important step in trying
to prove their congruence, which usually reduces to equality of their respective simpler elements.
Figure 1.7.2: A simple shape
Exercise 1.7.1 The previous shape consists of a line segment AB of length δ and the lines which are
orthogonal at each of its endpoints. Show that every line segment A B of same length δ, defines by analogy
a shape congruent to the first shape under the notion of displacement.
Hint: Because of length equality, line segment A B can be displaced until it becomes coincident with AB.
Then, according to Axiom 1.4.2, the orthogonal lines at the endpoints of A B will necessarily coincide
with those of AB at its endpoints.
1.8 Isosceles and right triangle

’Tis the perception of the beautiful,
A fine extension of the faculties,
Platonic , universal, wonderful,
Drawn from the stars , and filter’d through the skies,
Without which life would be extremely dull;
In short, it is the use of our own eyes,
With one or two small senses added, just
To hint that flesh is form’d of fiery dust.
Byron, Don Juan Canto II, 212
In Geometry, as in all of Mathematics, after a general categorical definition, it is useful to examine some
special cases. It is not rare, for some general property we want to prove, to establish the result in a special
case and find through it the way to the general case. Other times, again, the properties of the special case
help in expressing the proof and properties of the general case or the rejection of some general conjecture.
The isosceles and right triangles are special cases of triangles, which we meet in the formulations and proofs
of a plethora of more general properties in all geometry chapters. Isosceles triangle is called a triangle which
Α Α
Β Γ Β Γ
Figure 1.8.1: The isosceles and right triangle
has two equal sides. The two sides, which are equal in length, are called legs of the isosceles. The third
side is called base of the isosceles. The vertex, formed by the two equal sides, is usually called apex of the
isosceles.
Right triangle (or right-angled triangle) is called a triangle which has one angle equal in measure to 90
degrees. The sides which define this angle are called orthogonal sides. The side opposite of the right angle
is called hypotenuse of the right triangle.
Theorem 1.8.1 In every isosceles triangle (|AB| = |AΓ |) the two base angles (B and Γ ) are equal.
Α
(Ι) (ΙΙ)
Ρ
Β Γ Α Β
Μ Μ
Figure 1.8.2: The theorem of the isosceles The medial line of AB
Proof: Consider the two triangles ABM and AM Γ , which are formed by drawing AM , where M is the
middle of the base BΓ (Figure 1.8.2-I). The two triangles have their sides respectively equal: |AB| = |AΓ |
by hypothesis, |BM | = |M Γ | because M is the middle of BΓ and finally AM is common. According
to triangle Axiom 1.6.2 these two triangles will be congruent, therefore their angles at B and Γ will be
respectively equal, q.e.d.
1.9. TRIANGLE CONGRUENCE CRITERIA 19
Corollary 1.8.1 In every isosceles triangle (|AB| = |AΓ |) the line which joins its apex A with the middle
M of the opposite side bisects the apex angle at A (See Exercise 1.9.8 for the converse).
Corollary 1.8.2 In every isosceles triangle (|AB| = |AΓ |) the line which joins its vertex A with the middle
M of the opposite side is orthogonal to the base and divides the triangle into two congruent right triangles
(AM B and AM Γ ) (See Exercise 1.9.10 for the converse).
Medial line of the segment AB we call the line, which is orthogonal to the segment at its middle (Figure
1.8.2-II). The previous corollary can also be expressed in the following form.
Corollary 1.8.3 In every isosceles triangle ABΓ with base AB, its apex Γ lies on the medial line of the
segment AB.
The following formulation is also equivalent:
Corollary 1.8.4 Every point P , which is equidistant from points A and B, lies on the medial line of the
segment AB.
Exercise 1.8.1 For every isosceles triangle ABΓ with base AB, its apex Γ lies on the medial line of AB.
Corollary 1.8.5 Every point P which is equidistant from the points {A, B} lies on the medial line of AB.
Exercise 1.8.2 Assuming that two lines which are orthogonal to the same line do not intersect (we will
show this later corollary 1.10.8), show that: Given three different points {A, B, Γ } on the same line ε,
there is no point X, which is equidistant from these points.
1.9 Triangle congruence criteria

Laws, taken in the broadest meaning, are the necessary relations deriving
from the nature of things; and in this sense, all beings have their laws.
Montesquieu, The Spirit of the Laws
Besides the basic Axiom 1.6.2 of triangle congruence, which is also mentioned as SSS-criterion (side-side-
side criterion) of congruence, there are two more congruence criteria, which result as theorems relying
on the SSS-criterion. These are mentioned as SAS-criterion (side-angle-side criterion) and ASA-criterion
(angle-side-angle criterion).
B Β'
A Γ A' Γ'
Figure 1.9.1: SAS-criterion
Proposition 1.9.1 (SAS-criterion) Two triangles ABΓ , A B Γ , which have two equal respective sides
’ | = |B
(|AB| = |A B |, |AΓ | = |A Γ |) and the contained in them angles also equal (| BAΓ ◊ A Γ |), are
congruent.
Proof: Place angle A onto A so that the half lines AB and A B and AΓ and A Γ become coincident.
This is possible because of the assumed equality of the angles at respectively A and A (Figure 1.9.1-I).
Because of the also assumed equality of lengths |AB| = |A B |, we will also have coincidence of B and B
(according to Axiom 1.3.3), and, for the same reason, we will have coincidence of Γ and Γ . As a result,
we have coincidence of the sides BΓ and B Γ and therefore their lengths will be equal |BΓ | = |B Γ |.
The truth of the proposition results by applying the SSS-criterion, q.e.d.
Β Β'
Α Γ Α' Γ'
Figure 1.9.2: ASA-criterion
Proposition 1.9.2 (ASA-criterion) Two triangles ABΓ , A B Γ , which have two respective angles equal
÷| = |A
(|ABΓ ◊ ’
B Γ | and |BΓ ◊
A| = |B Γ A |) and the contained in them sides also equal (|BΓ | = |B Γ |),
are congruent.
Proof: The proof is similar to the previous one. Place the triangles so that BΓ and B Γ , as well as the
angles at B, B and Γ , Γ become coincident (Figure 1.9.2-II). This is possible because of Axiom 1.4.2.
Then we get coincidence of lines BA, B A as well as Γ A, Γ A , consequently we get coincidence of their
respective intersections which define the points A and A . From this last coincidence we get |BA| = |B A |
and |Γ A| = |Γ A |. The truth of the proposition results by applying the SSS-criterion, q.e.d.
A A'
B Γ Γ' B'
Figure 1.9.3: Two equal angles produce an isosceles
Proposition 1.9.3 If a triangle has two of its angles equal then it is an isosceles.
Proof: Consider triangle A B Γ congruent to ABΓ and apply the ASA-criterion. The two triangles have
the sides BΓ and B Γ equal respectively and angles ABΓ’ and A ◊ ’
Γ B equal, as well as AΓ ◊
B and ABΓ
’
equal, therefore they are congruent. Side AΓ opposite to angle ABΓ will be equal to the side A B op-
posite to angle A◊

Γ B , which is equal to the previous angle. But, by construction A B is equal to AB,
therefore finally AB and AΓ are equal, q.e.d.
Comment-1 This proof (due to Pappus) contains a subtle and paradoxical point, where two congruent
triangles are proved again equal. The resulting wordplay has to do with the triangle’s orientation. It is true
that A B Γ is equal to ABΓ , yet it has been placed against ABΓ after reversing its orientation. Figure
1.9.4 shows the difference with a non-isosceles ABΓ . Note that, although the two triangles are congruent,
1.9. TRIANGLE CONGRUENCE CRITERIA 21
A Α'
Β=Γ' Γ=Β'
Figure 1.9.4: Placement of congruent triangles with opposite orientation
this particular placement does not make them coincident. The deeper meaning of the proposition is, that the
two congruent triangles, placed this way, become coincident, when and only when they are equal isosceli.
Corollary 1.9.1 Point Γ belongs to the medial line ε of line segment AB, if and only if it is equidistant
from A and B.
Proof: In Corollary 1.8.4 we saw that every point Γ equidistant from A and B lies on the medial line of
AB. For the converse, we take point Γ on the medial line and we show that the triangles Γ M A and Γ M B
are congruent (M being the middle of AB) by applying the SAS-criterion, q.e.d.
Comment-2 The last corollary characterizes the medial line as a geometric locus of points, which have
a specific property. We often express this as: the geometric locus of points having property x is set Y .
Thus, we will say from now on: the geometric locus of points, which are equidistant from two points A and
B is the medial line of AB. As we did in the case of the medial line, also in the general case of a geometric
locus, we must show two things: a) every point of the geometric locus has property x, b) if a point has
property x then it necessarily belongs to the geometric locus (more in § 2.16).
Exercise 1.9.1 Two right triangles which have respective perpendicular sides of the same length are con-
gruent.
Hint: Apply the SAS-criterion with respective angles the right angles of the two triangles.
Exercise 1.9.2 Two right triangles which have an orthogonal side and the adjacent acute angle respectively
equal are congruent.
Hint: Apply the ASA-criterion.
Α
Α
Μ Ν Ζ Η
Β Γ Β Γ
Figure 1.9.5: Equal medians Equal bisectors
Exercise 1.9.3 Let ABΓ be an isosceles triangle with equal angles at vertices B and Γ . Show that the
medians through these vertices are equal. Also show that the bisectors through these vertices are equal.
Hint: Let M and N be the middles of BA and Γ A respectively. Triangles BM Γ and BN Γ are congruent
as having a) BΓ common, b)BM and Γ N equal as halves of equal sides, c) angles at B and Γ equal. The
SAS-criterion therefore is applicable. The proof for bisectors is similar, only this time the ASA-criterion
is applicable. Indeed, suppose BH and Γ Z are the bisectors of the angles at B and Γ , respectively. Then
triangles BΓ H and BΓ Z are congruent as having a) BΓ common, b) angles at B and Γ equal, c) angles
’ | = |ZΓ
|HBΓ ’ B| equal as halves of equal angles.
Comment-3 The converse of the proposition in the last exercise, also holds, but in the case of medians
we need a property learned later (see Exercise 2.8.1) and in the case of bisectors the proof of the converse,
which we give in the next section (Theorem 2.5.2), is mentioned as theorem of Steiner-Lehmus and is
unexpectedly difficult. For a calculational proof see Exercise 3.12.12.
A
Α Δ
(Ι) (ΙΙ)
Z H
Δ E
Ε
Β B Γ
Γ
Figure 1.9.6: Extension of the median Equal altitudes of isosceles
Exercise 1.9.4 Let E be the middle of side AΓ of the triangle ABΓ . Extend BE (median) so as to double
its length until ∆. Show that triangles AΓ∆ and AΓ B are congruent.
Hint: Using the SAS-criterion first show that triangles AEB and Γ E∆ are congruent (Figure 1.9.6-I).
Similarly, also show that BEΓ and AE∆ are congruent. Conclude next, using the SSS-criterion, that
ABΓ and AΓ∆ are congruent.
Exercise 1.9.5 Let ABΓ be an isosceles triangle with equal angles on vertices B and Γ . Show that the
altitudes through these vertices are equal.
Hint: Let BE and Γ∆ be the altitudes respectively from angles B and Γ (Figure 1.9.6-II). Extend BE
by doubling its length until point H and Γ∆ by doubling its length until point Z. Triangles BEΓ and
HEΓ are congruent as having a) EΓ common, b) angles at E right and c) sides BE and EH equal by
construction (SAS-criterion). Consequently triangle BΓ H is isosceles. Similarly it can be proved that tri-
angle BΓ Z is also isosceles. These two triangles are congruent as having a) BΓ common, b) BZ equal to
Γ H and c) their angles at B and Γ equal as double of β and γ respectively. Thus Γ Z and BH, which are
doubling the altitudes, will be equal. For the converse of this property see Exercise 1.10.2.
Comment-4 We will later see that there is one more triangle congruence criterion, which could be called
AAS-criterion. According to this if two triangles ABΓ and A B Γ have their angles α = α , β = β and
sides a = |BΓ | = |B Γ | = a , then they are congruent. In this case it is assumed that the triangles have
two angles equal and a side respectively equal to another, but this side is opposite of α and not adjacent to
α (like in the ASA-criterion). The latter however reduces to the ASA-criterion, because from the equality
of the two angles and the relation α + β + γ = 180 ◦ , which we will show later, what results is equality of
all angles of the two triangles.
Figure 1.9.7-II shows that a supposed SSA-criterion could not be valid. In general (for a not right triangle)
there are two triangles ABΓ and A B Γ for which holds a = a , b = b and A =A . We will analyze this
figure as well later, when we will have sufficient knowledge about the circle and its properties.
Exercise 1.9.6 Show that, if the triangles ABΓ and A B Γ are congruent, then the medians/bisectors of
ABΓ are equal, respectively with those of A B Γ .
Exercise 1.9.7 Assume that the line segments AB and Γ∆ have a common medial line ε and that AΓ meets
ε in E. Show that B∆ also meets ε in E (Figure 1.9.7-II).
1.10. RELATIVE MEASURES OF TRIANGLE ANGLES 23
A'
(Ι) (ΙΙ) Α
α' Γ
c' ε Ε
Α
b'
α Δ
c b
Β
B Γ
a=a'
Figure 1.9.7: Ambiguous case Common medial line
’ , then the triangle is

Exercise 1.9.8 Show that, if the median AM of triangle ABΓ bisects the angle BAΓ
isosceles (converse of Corollary 1.8.1).
Exercise 1.9.9 Show that two congruent triangles ABΓ and A B Γ have equal respective altitudes.
Hint: Place the triangles so that their bases BΓ and B Γ coincide and vertices A, A to be found on either
side of BΓ . Then ABA is an isosceles triangle, etc.
Exercise 1.9.10 Show that, if the median AM of triangle ABΓ is perpendicular to the base BΓ , then the
triangle is isosceles (converse of Corollary 1.8.2).
1.10 Relative measures of triangle angles

Vain would be the attempt of telling all the figures of
them circling as in a dance, and their juxtapositions,
and the return of them in their revolutions upon
themselves, and their approximations. . .
Plato, Timaeus 40
Next properties of the triangle rely on the axioms we have accepted up to now and prepare the ground for
the all important triangle inequality of the next section.
Theorem 1.10.1 In every triangle ABΓ the supplementary of every angle is greater than each of the other
two angles.
A N
M
B X
Γ
Figure 1.10.1: Triangle angles comparison
’
Proof: Let us show that XΓ A (Figure 1.10.1), which is the supplementary of Γ , is greater than A.
Let M
be the middle of AΓ and assume that N lies on the half line BM , so that |BM | = |M N |. Triangles ABM
and Γ N M have: a) angles AM ÷ ÷

B and Γ M N equal, since they are orthogonal, b) sides AM and M Γ equal,
because M is the middle of AΓ , c) sides M B and M N equal by construction. Consequently, by applying
the SAS-criterion of congruence (Proposition 1.9.1), the triangles are congruent. From the congruence of
’ is equal to AΓ
the triangles results that angle BAΓ ’ ’
N . The last one however is smaller than AΓ X because
N belongs in the interior of AΓ ’ X (Exercise 1.4.6). Similarly we show that angle ABΓ ’ is smaller than
’
AΓ X, q.e.d.
Comment-1 Often angle AΓ ’ X is mentioned as an external angle (see § 1.6) and the theorem takes the
form:
Every external angle in a triangle is greater than each of the opposite internal angles.
A Α
(Ι) α E (ΙΙ) α
Δ ε Ε
δ ε
B Γ Β Γ
Figure 1.10.2: α<ε<δ Greater angle opposite greater side
’ is greater than
Proposition 1.10.1 Let the point ∆ be in the interior of triangle ABΓ . Then angle B∆Γ
’.
BAΓ
Proof: Extend one of the sides of the internal angle, for example B∆ and define the intersection point E
’ | > α = |BAΓ
with AΓ (Figure 1.10.2-I). Angle ε = | BEΓ ’ | as an external and opposite of α in triangle
’
ABE. Similarly also δ = |B∆Γ | > ε. Combining therefore, δ > α, q.e.d.
Theorem 1.10.2 In every triangle ABΓ opposite to a greater side lies a greater angle.
Proof: Let’s suppose that side AΓ is greater than AB (Figure 1.10.2-II). Then, there exists point E between
A and Γ , such that |AE| = |AB| (Axiom 1.3.3, Exercise 1.3.1). The triangle which results is isosceles
’ = |AEB|
and according to the previous theorem, | ABE| ’ > |AΓ ’ ’ | > |ABE|
B|. Also |ABΓ ’ because E
’ ’ ’
belongs in the interior of angle ABΓ . Combining therefore, | ABΓ | > |AΓ B|, q.e.d.
Theorem 1.10.3 In every triangle ABΓ opposite to a greater angle lies a greater side.
Proof: Using reduction to contradiction. Assume that in triangle ABΓ angle α > β and also side a ≤ b.
a = b is not possible, because then we would also have α = β (Theorem 1.8.1), contrary to our assumption.
a < b is also not possible, because, if this was the case, then according to the previous proposition, it would
have to be α < β, which is again contrary to our assumption. Therefore we must have a > b, q.e.d.
Corollary 1.10.1 In every triangle the sum of any two of its angles is less than 180 degrees.
Proof: Denote the angles by α, β and γ. According to Theorem 1.10.1, each of α, β is less than 180 ◦ − γ.
Therefore α + γ < 180 ◦ , β + γ < 180◦. Similarly one proves α + β < 180 ◦, q.e.d.
Corollary 1.10.2 A triangle has at most one obtuse angle.

Proof: If it had two obtuse angles, for example α and β, then it would be α + β > 180 ◦ , which is a
contradiction, q.e.d.
Corollary 1.10.3 In every isosceles triangle its two equal angles are acute.
Proof: If α and β were equal and obtuse or right, then α + β ≥ 180 ◦ , a contradiction, q.e.d.
Corollary 1.10.4 Every right triangle has its other two angles acute.
Proof: If α = 90◦ and β, γ are the other angles of the right triangle, then the exterior angle of α is also 90 ◦
and is therefore greater from both α and β, q.e.d.
Corollary 1.10.5 In a right triangle each one of the orthogonal sides is smaller than the hypotenuse .
Proof: The hypotenuse is opposite to the right angle which is greater than each of the other angles which
are acute (Corollary 1.10.4), therefore the two orthogonal sides, which are opposite to the acute angles, will
be less than the hypotenuse (Theorem 1.10.3), q.e.d.
Corollary 1.10.6 Let A∆ be orthogonal to XY , where ∆ is a point of XY . Then, points B and Γ satisfy
|B∆| < |Γ∆|, if and only if |BA| < |Γ A|.
Α Γ
(Ι) (ΙΙ)
b
c
Χ Υ
Δ Β a Γ Δ Α Β
Figure 1.10.3: Trace ∆ of the orthogonal |Γ A| < |Γ∆|
’ is acute (Figure 1.10.3-I) because it is an angle different from the right one of triangle
Proof: Angle ∆BA
’
AB∆ (Corollary 1.10.4). Similarly angle ∆Γ A is acute. Suppose now that |B∆| < |Γ∆|. Then in triangle
ABΓ , Γ lies in the extension of ∆B towards B and the angle at B is obtuse and at Γ acute. Therefore the
opposite to the obtuse side AΓ will be greater than the opposite to the acute side AB. For the converse,
suppose that |BA| < |Γ A| but |B∆| > |Γ∆|. There is a contradiction here, because, according to the
previous part of the proof, |Γ∆| < |B∆| implies |Γ A| < |BA|, contrary to the hypothesis. Likewise there
is an contradiction if we suppose |BA| < |Γ A| and |B∆| = |Γ∆|, because the latter implies that triangles
A∆B and A∆Γ are congruent. Therefore, when |BA| < |Γ A|, it must also be |B∆| < |Γ∆|, q.e.d.
Comment-2 The last corollary is equivalent to the fact that the length of the hypotenuse |AB| of the
right triangle A∆B is an increasing function of the length of the orthogonal side |∆B|, when the other
orthogonal side A∆ remains fixed.
Corollary 1.10.7 Let the isosceles triangle ABΓ have equal angles at A and B. Consider a point ∆ in
the extension of line segment AB. Then Γ∆ is greater than Γ A (Figure 1.10.3-II).
Γ Α
(Ι) (ΙΙ)
α
β
ω ε ω ε
Α Β Β Γ
Figure 1.10.4: Lines incident at angle ω Only one orthogonal from A

Corollary 1.10.8 If lines α, β intersect line ε and they form equal angles on the same side of the line and to
the same direction of it, then they are parallel. Specifically two lines perpendicular on this line are parallel.
Proof: If they were not parallel, they would intersect at say, Γ (Figure 1.10.4-I), and then they would form a
triangle with sum of angles at the vertices different from Γ : ω+(180 ◦ −ω) = 180◦, which is a contradiction
(Corollary 1.10.1), q.e.d.
Corollary 1.10.9 Trhough a point A, lying outside the line ε, there can be at most one orthogonal line to
ε.
Proof: If there were two different orthogonals from A (Figure 1.10.4-II), then they would form a triangle
and its two angles different from A would sum to 90 ◦ + 90◦ = 180◦ , which is a contradiction (Corollary
1.10.1), q.e.d.
Α
Β
(Ι) (ΙΙ) Ν Μ
Γ' Γ Β Γ
Α
Figure 1.10.5: Congruent right triangles Isosceles is characterized by equal altitudes
Exercise 1.10.1 Two right triangles, which have one orthogonal side and the hypotenuse equal, respec-
tively, are congruent.
Hint: Place the two triangles so, that the two orthogonal sides coincide with the line segment AB, with
triangles ABΓ and ABΓ lying on different sides of AB with the right angle at A (Figure 1.10.5-I). Then,
Γ , Γ and A are all lying on a line and B by hypothesis is equidistant from Γ and Γ , therefore B belongs
to the medial line of ΓΓ and A is the middle of ΓΓ .
Exercise 1.10.2 If two altitudes in a triangle are equal, then the triangle is isosceles.
Hint: Let BM and Γ N be the two altitudes (Figure 1.10.5-II). Then triangles BM Γ and BN Γ are right
triangles with common hypotenuse BΓ and the orthogonal sides BM and Γ N by hypothesis equal. Ac-
cording to the previous exercise, the right triangles are congruent and therefore their angles at B and Γ are
equal. Hence the triangle ABΓ will be isosceles (Corollary 1.9.3). Note that the property in this exercise
is the reverse of the one in Exercise 1.9.5.
Theorem 1.10.4 If two triangles ABΓ and A B Γ have equal sides respectively adjacent to A and A
>A
(|AB| = |A B |, |AΓ | = |A Γ |) and the angles at A and A are unequal ( A ), then their sides are

respectively unequal (|BΓ | > |B Γ |).
Proof: Place the triangles so that sides AB and A B become coincident and AΓ lies in the interior of
The possible cases are three, depending on the position of Γ relative to the line BΓ (I-III in figure
A.
1.10.6). Let us see the first case (Figure 1.10.6-I) and leave the other two as exercises. In this case, we
suppose that Γ and A are contained in different sides (half planes) of BΓ . Then comparing the angles
of the triangle BΓ Γ we have |BΓ÷ ’
Γ | > |AΓ ’ | > |BΓΓ
Γ | = |AΓΓ ’ |, where the last holds because
÷
|AΓ | = |A Γ | = |AΓ |. But according to the previous theorem, the inequality | BΓ ’ | implies
Γ | > |BΓΓ
|BΓ | > |BΓ |, q.e.d.
Corollary 1.10.10 , If two triangles ABΓ and A B Γ have equal sides respectively adjacent to A and A
(|AB| = |A B |, |AΓ | = |AΓ |) and the third sides are unequal (|BΓ | > |B Γ |), then their angles at A
and A are respectively unequal ( A >A ).
Α
(II)
Α
(I) Α
Β Γ (III)
Γ'
Β Γ
Β Γ' Γ
Γ'
Figure 1.10.6: Greater side opposite greater angle
Proof: If, in addition to the corollary’s hypotheses we had | BAΓ ’ | ≤ |B ◊

A Γ | then we would get an
’ | = |B
contradiction. Indeed, if | BAΓ ◊ A Γ |, then according to the SAS-criterion the triangles would
be congruent and we would have |BΓ | = |B Γ |, a contradiction. If instead | BAΓ’ | < |B ◊ A Γ |, then
according to theorem 1.10.4 we would have |BΓ | < |B Γ |, also a contradiction, q.e.d.
Exercise 1.10.3 Show that between two isosceli triangles ABΓ and ∆BΓ with the same base, the one
which has the greater angle at its apex is the one with the smaller legs and conversely.
Δ
Γ=Γ'
(Ι) (ΙΙ)
Α
ω ω
Β Μ Γ Β Α Α'
Figure 1.10.7: Greater isosceles Congruent isosceli
Hint: Show first that their corresponding third vertices, which are contained in the medial line of the base
BΓ (Figure 1.10.7-I), are closer to its middle M , when the corresponding triangle leg becomes smaller
(Corollary 1.10.6).
Exercise 1.10.4 Show that two isosceli triangles ABΓ and A B Γ , which have equal legs |Γ A| = |Γ B| =
|Γ A | = |Γ B | and equal angles adjacent to their bases, are congruent.
Hint: Place the triangles so that their angles at B and B (Figure 1.10.7-II), as well as BΓ and B Γ ,
because of equality of lengths, become coincident. If A, A were different and A was farther than B than
A, we’d have an external angle ω of triangle AA Γ at A greater than the internal ω at A (ω > ω), which is
a contradiction. Therefore the vertices A, A coincide.
Exercise 1.10.5 Given a point A not in line ε and an angle ω (0 < ω < 90), show that there is, at most,
one isosceles triangle with vertex at A, base at ε and base angles equal to ω.
Exercise 1.10.6 Show that in every triangle ABΓ the sum of its altitudes is less than its perimeter.
Exercise 1.10.7 Show that two right triangles ABΓ and A B Γ , which have equal hypotenuses and one
angle ω = 90◦ equal, are congruent.
Α Α Γ
α2
(Ι) α1
(ΙΙ)
Μ
Δ
α1
ω ω Β Α'
Β Δ
Γ
Figure 1.10.8: Congruent right triangles Median slope
Hint: Extend the adjacent to ω orghogonal BΓ by doubling its length until ∆ (Figure 1.10.8-I). The result-
ing triangle AB∆ is isosceles. Similarly results the isosceles A B ∆ from right triangle A B Γ . The two
isosceli triangles AB∆ and A B ∆ have equal legs respectively and equal angles adjacent to their base,
therefore they are congruent (Exercise 1.10.4). Then their halves, i.e. the right triangles ABΓ and A B Γ
are congruent.
Exercise 1.10.8 Show that the median AM of triangle ABΓ with unequal sides AB, AΓ , is slanted to-
wards the smaller side. Also, between angles BAM÷ and M ÷ AΓ the one adjacent to the smaller side is
greater. Conclude that between vertices B and Γ , the trace of the bisector A∆ is closer to the vertex which
belongs to the smaller side (Figure 1.10.8-II).
Exercise 1.10.9 Show the inverse of the preceding exercise, i.e. if the median AM is slanted towards B,
÷>M
then |AB| < |AΓ | and BAM ÷ AΓ .
Α Α
(Ι) Α (ΙΙ) (ΙΙΙ)
Δ Α' Β Γ
Δ
Γ=Γ' Β=Β'
Β Γ Β' Γ'
Figure 1.10.9: Comparing triangles Attaching isosceli Comparing sides
Exercise 1.10.10 The right angled triangles {ABΓ , A B Γ } have equal hypotenuses |BΓ | = |B Γ | and
’ <ω=A
angles φ = ABΓ ◊ B Γ . Show that |AΓ | < |A Γ | (Figure 1.10.9-I).
Exercise 1.10.11 Let BAΓ and Γ A∆ be isosceli triangles with apex at A and a common side AΓ . Show
that sides AB and A∆ are either contained in the same line or they form an isosceles triangle BA∆ (Figure
1.10.9-II).
Exercise 1.10.12 Extend the sides {AB, AΓ } of the triangle ABΓ to their double {AB , AΓ }. Show that
|BΓ | < |Γ B | < |B Γ | (Figure 1.10.9-III).
Exercise 1.10.13 Continuing the last exercise, show that the angles of the triangle AB Γ at {B , Γ } are
not greater from the respective angles at {B, Γ } of the triangle ABΓ . Later, after introducing the axiom
“=B
of parallels (§ 1.15), we will see that the angles are equal { B , Γ
=Γ
}.
1.11. THE TRIANGLE INEQUALITY 29
1.11 The triangle inequality

The third, to conduct my thoughts in such order that, by
commencing with objects the simplest and easiest to know, I
might ascend by little and little, and, as it were, step by step, to
the knowledge of the more complex; assigning in thought a
certain order even to those objects which in their own nature do
not stand in a relation of antecedence and sequence.
Descartes, Discourse on Method II
The triangle inequality is of fundamental importance in geometry and leads to the confirmation of intuition,
that the line segment represents the shortest path between two points.
Theorem 1.11.1 In every triangle ABΓ the sum of the lengths of two of its sides is greater than the length
of the third side.
Δ Α
(Ι) b (ΙΙ)
α
Α Δ
ω
c b δ
ω
γ
Β Γ
Β a Γ
Figure 1.11.1: Triangle inequality Inequality for the difference
Proof: Let a = |BΓ | be the greatest of all sides. It suffices to show that a < b + c. For this, extend AB by
the segment A∆ equal to AΓ . In the resulting triangle ∆BΓ , angle BΓ∆ ’ = BΓ ’ ’ is greater than
A + AΓ∆
’ therefore (Theorem 1.10.3) also side B∆, for which |B∆| = b + c, will be greater than BΓ with
Γ∆B
|BΓ | = a, q.e.d.
Theorem 1.11.2 In every triangle ABΓ the difference in length between any two sides is less than the
length of the third side.
Proof: Let AΓ be greater than AB (b > c). It suffices to show that a > b − c. For this, consider on AΓ
the segment A∆ equal to AB. In the resulting triangle ∆BΓ , angle B∆Γ ’ is greater than ∆BΓ’ . This,
because as external of the basis angle ω of the isosceles triangle BA∆, it is obtuse (Corollary 1.10.3). And
if a triangle has one obtuse angle, this angle is greater than the other two, which must be acute (Corollary
1.10.2). It follows that side BΓ , which is opposite to the obtuse δ = B∆Γ’ is greater than ∆Γ , which has
length |∆Γ | = |AB| − |AΓ |, q.e.d.
Comment The two theorems together mean that in every triangle the following inequalities hold:
|a − b| < c < a + b, |b − c| < a < b + c, |c − a| < b < c + a.
It is easy to see that all such inequalities are equivalent to the restriction a < b + c, where a is the triangle’s
greater side.
A broken line is a shape which consists of a sequence of line segments AB, BΓ ,..., XY , Y Ω in which ev-
ery pair of successive segments has exactly one common endpoint (Figure 1.11.2-I). This broken line is de-
noted by ABΓ ...Y Ω and we say that it joins points A and Ω . The broken line is called closed, when points
A and Ω coincide. Sides of the broken line we call the line segments AB, BΓ , .... Length of the broken
line is called the sum of the lengths of all line segments in the previous sequence, |AB|+ |BΓ |+ ...+ |Y Ω |.
Β Β
(Ι) Υ (ΙΙ) Δ
Α Ω Ω
Α
Γ Χ Γ Ε
Figure 1.11.2: Broken line Shortening of broken line
Corollary 1.11.1 The line segment AΩ has length |AΩ |, less than the length of any broken line which joins
A with Ω .
Proof: Given the broken line AB...Y Ω , we draw AΓ , resulting in triangle ABΓ (Figure 1.11.2-II). Because
of the triangle inequality, the new broken line which results AΓ∆E...XY Ω has smaller length than the
length of the original broken line and has one side less. Continuing this way we reduce successively the
length µ of the broken line and its number of sides. The process continues until we reach the line segment
AΩ and the corresponding reduction of length gives the inequality |AΩ | < µ, q.e.d.
Α A
(Ι)
(ΙΙ)
Ε
B Γ
Δ M
Β Γ Δ
|AB|+|AΓ |
Figure 1.11.3: |∆B| + |∆Γ | < |AB| + |AΓ | |AM | < 2
Exercise 1.11.1 Let the point ∆ be in the interior of triangle ABΓ . Show that |∆B|+|∆Γ | < |AB|+|AΓ |
(Figure 1.11.3-I).
Exercise 1.11.2 Let point ∆ be in the interior of triangle ABΓ . Show that the sum of the distances of ∆
from the triangle’s vertices is less than the triangle’s perimeter and greater than the half perimeter.
Exercise 1.11.3 Let the point ∆ be contained in the exterior of triangle ABΓ . Show that the sum of the
distances of ∆ from the triangle’s vertices is greater than the triangle’s half perimeter.
Exercise 1.11.4 Show that the median AM of triangle ABΓ has length less that the half-sum of the its two
adjacent sides.
Hint: Extend AM by doubling it to ∆ (Figure 1.11.3-II). Triangles AM B and ∆M Γ are congruent, having
|AM | = |M ∆|, |BM | = |M Γ | and contained angles AM ÷ B and Γ÷M ∆ equal. Consequently, from the
triangle inequality, we have 2|AM | = |A∆| < |AΓ | + |Γ∆| = |AΓ | + |AB|. Note that it also holds
|AM | > 12 (|AB| + |AΓ | − |BΓ |) (Corollary 3.12.3).
Exercise 1.11.5 Show that in every triangle the sum of the lengths of the medians is less than the perimeter
of the triangle.
Exercise 1.11.6 Show that in an isosceles triangle ABΓ with altitude from apex, υ = |A∆|, greater from
the basis length, b = |BΓ |, the sum of the lateral sides, 2λ = |BA| + |AΓ |, is greater than b + υ. Later,
using Pythagora’s theorem 3.4.1, we will see that 2λ ≥ b + υ, precisely when υ ≥ 23 b.
1.12. THE ORTHOGONAL TO A LINE 31
1.12 The orthogonal to a line

Let us honour if we can
The vertical man
Though we value none
But the horizontal one.
W. H. Auden, To C. Isherwoods
From Corollary 1.10.9 we know that, if there exists one orthogonal to the line ε, from point A lying outside
the line, then it will be unique. The next theorem proves the existence of such an orthogonal. For this we
use the fundamental capability of constructing a specific angle (Axiom 1.4.2) having its vertex and one of
its sides on a given line. From the same fundamental capability results also the existence of the orthogonal
ε to line ε from a point A on the line ε (Corollary 1.5.1). The practical construction of the orthogonal,
using a ruler and compass is founded on properties of the circle and will be done in the following (§ 2.4).
Α Α
(Ι) (ΙΙ)
ε'
ε
Β Δ
ε
Β Μ Γ
Γ
Figure 1.12.1: Orthogonal from point Distance |AM | of A from the line ε
Theorem 1.12.1 From a point A lying outside the line ε precisely one orthogonal to the line can be drawn.
Proof: We construct the orthogonal line using an isosceles triangle as follows (Figure 1.12.1-I). We consider
an arbitrary point B in ε and draw the line BA. In the half plane of ε, which does not contain point A,
we define the line segment BΓ which forms with ε the same angle as BA (Axiom 1.4.2). On this line
we consider the point Γ such that |BΓ | = |BA|. Line AΓ is the wanted one. Indeed, triangle ABΓ is
isosceles with base AΓ by construction and ε coincides with the bisector of its apex. Consequently ε will
be orthogonal to the base AΓ of the triangle (Corollary 1.8.2), q.e.d.
Corollary 1.12.1 (Distance of point from line) Let A be a point not lying on the line ε and AM be or-
thogonal to it at the point M belonging to ε. For every other point B of the line, AB is greater than
AM .
Proof: Any other point B (Figure 1.12.1-II) makes a right triangle AM B with hypotenuse AB, which is
always greater than the orthogonal AM (Corollary 1.10.5), q.e.d.
Point M is called (orthogonal or vertical) projection of A onto the line ε. We call the length |AM | of
AM distance of the point A from the line ε.
Exercise 1.12.1 Extend the altitude υ α of the triangle ABΓ from A, towards BΓ , to its double until point
E. The created triangle BEΓ is congruent to ABΓ .
Hint: By applying the SAS-criterion, show first that triangles B∆A and B∆E are congruent and triangles
A∆Γ and E∆Γ are also congruent. Then, by applying the SSS-criterion, show that ABΓ and EBΓ are
congruent.
1.18. COMMENTS AND EXERCISES FOR THE CHAPTER 55
Α
2
(Ι) (ΙΙ)
ν
Α
3
4
1
Β Δ Γ Β Γ
μ
Figure 1.18.3: Same angles: B∆1, B∆2, ∆Γ 3, ∆Γ 4 How many triangles?
Exercise 1.18.24 Show that for every point ∆ of the side BΓ of the triangle ABΓ , different from B and
Γ , there exists line ∆X which upon intersecting the other sides of the triangle forms triangle with angles
equal to those of ABΓ . Show that for every such ∆ there exist four different lines ∆X which have this
property (Figure 1.18.3-I).
Exercise 1.18.25 We join vertex A of the triangle ABΓ with µ different points of the opposite side. We
also join vertex B with ν different points of the opposite side (Figure 1.18.3-II). How many triangles are
contained in the resulting shape?
(µ+1)(ν+1)(µ+ν+2)
Hint: 2 .
Exercise 1.18.26 Triangle ABΓ has the sides |AΓ | > |AB| and on the side AΓ we define point B , such
that |AB | = |AB|. Show that angle B÷ BΓ = β−γ . Show also that this angle is equal to the angle between
2
the bisector and the altitude from A.
Exercise 1.18.27 Show that for every point X, not contained in the intersecting at O lines {α, β} and every
line γ passing through X, line γ intersects at least one of the two lines {α, β}.
Exercise 1.18.28 If Y , ∆, M are respectively labels for the traces on BΓ of the altitude, bisector and
median in triangle ABΓ , show that ∆ lies always between Y and M . Also show that if two of these traces
coincide, then all three traces coincide and the triangle is isosceles.
Γ
Γ
(Ι) (ΙΙ)
γ1 γ2
Δ β
α Α
Ε Ζ Β
χ1 χ2
M
ω φ Ε
Α' Ζ Β'
Α Β
’ = α+β
Figure 1.18.4: A∆B Construction of point M
Exercise 1.18.29 On the sides Γ A, Γ B of triangle ABΓ we select respectively points E and Z and we
’
draw the bisectors of the angles ω = Γ ’ , which intersect at ∆. Show that for angles
AZ and φ = EBΓ
’ ’
α = AEB and β = AZB, holds α + β = 2A∆B. ’
Hint: Suppose that Γ∆ divides angles γ = AΓ’ ’ respectively, into two parts γ 1 , γ2 and χ1 , χ2
B and A∆B,
(Figure 1.18.4-I). Observe that α + β = 2γ + ω + φ, χ 1 = γ1 + ω2 , χ2 = γ2 + φ2 .
Exercise 1.18.30 Find a point M on the side AB of the triangle ABΓ , such that |M A| + |AΓ | = |M B| +
|BΓ |.
Hint: Consider AA equal with AM on the extension of Γ A and BB equal to BM on the extension of
Γ B (Figure 1.18.4-II). The length |Γ A | = |Γ B | is known and equal to half the perimeter of the triangle.
Triangles AA M and BB M are isosceli and are constructed from the givens.
Exercise 1.18.31 Show that if the triangle ABΓ is contained entirely inside the triangle A B Γ , then each
side of ABΓ is less than the greater side of A B Γ and each side of A B Γ is greater than the smaller
side of ABΓ .
Exercise 1.18.32 Is there any line intersecting all side-lines of the triangle ABΓ under the same angle ω?
000 001 002 003 012 021
111 112 113 110 123 132
222 223 220 221 230 203
333 330 331 332 301 310
Figure 1.18.5: Trioker tiles
Equal equilateral triangles are used as building elements (tiles) in some board games, in which the end of
the game is to build shapes using these elements. Figure 1.18.5 shows the 24 building blocks of one such
game, published by Robert Laffont and called “Trioker” ([OR79]). The unique rule of the game is the
match of number of points at concurring vertices.
1 2
0
2 1 2
2
021 002 0 3 1 1
3 2 1 0 0
3 3 1
0
230 112 301
012 113 3 2 0 1
0 1 3
Figure 1.18.6: Shapes build with trioker tiles
The game is reminiscent of the classical “Stomachion of Archimedes” and “Tangram” discussed in § 3.14.
The game’s end results are shapes composed by (some or all of) these 24 tiles, characterized by the integers
at the common vertices of the tiles building the shape (Figure 1.18.6).
Exercise 1.18.33 Try to build the shaded hexagon in figure 1.18.6 using some of the trioker tiles.
CHAPTER 2
Circle and polygons
2.1 The circle, the diameter, the chord

Most people live, whether physically, intellectually, or morally,
in a very restricted circle of their potential being. They make
use of a very small portion of their possible consciousness, and
of their soul’s resources in general, much like a man who, out
of his whole bodily organism, should get into a habit of using
and moving only his little finger.
William James, Letter to Lutoslawski
Following the line and its derivatives, that is, line segments, broken lines and polygons, the circle is the
simplest curve which, like the line, has a certain homogeneity and shows the same behavior in all of its
points and parts but, contrary to the line, it does not extend to infinity. Also, the line and the circle are the
shapes for which we have the simplest drawing tools, the ruler and the compass. If we had a third, equally
simple, tool for drawing another category of curves, we would for sure include this category also in the
domain of elementary euclidean geometry.
Circle of radius ρ is called the shape of the plane consisting of all points X, whose distance from a
fixed point O is ρ. Point O is called center of the circle. Radius OX of the circle is also called the line
segment with end points the center O and the circle point X. Often we also denote this circle with O(ρ)
κ
Χ
ρ
ε
Ο Α Ο Β
Figure 2.1.1: Circle κ(O, ρ), Diameter AB
or with κ(O, ρ). Interior of the circle we call the set of points Y , whose distance from the center O is
less than the radius: |OY | < ρ. Exterior of the circle we call the set of points Z, whose distance from
58 CHAPTER 2. CIRCLE AND POLYGONS
the center O is greater than the radius: |OZ| > ρ. From Axiom 1.3.3 of lines, it follows that every line ε,
which passes through a circle’s center O, will meet the circle exactly at two points A and B. Such points of
the circle are called diametrically opposite or diametrical or diametral and the line segment AB, which
they define is called diameter of the circle. Obviously the middle of every diameter is the center O and the
diameter’s length is twice that of the radius |AB| = 2ρ. Two circles are called congruent, when their radii
are equal. Chord of a circle is called a line segment AB whose end points lie on the circle. The diameter
(I)
(II)
Ο Ο
Γ
Α Β Α Β
Figure 2.1.2: Circle chord, Common points chord and circle
is just a special case of chord. Every chord AB, which is not a diameter, defines, along with the center of
the circle O, an isosceles triangle AOB (Figure 2.1.2-I), whose legs have length ρ.
Proposition 2.1.1 A chord AB does not contain any points of the circle, other than A and B.
Proof: If the chord contained one more circle point Γ (Figure 2.1.2-II), then triangle BOΓ would be
isosceles and A would belong to the extension of the triangle’s base, therefore, according to Corollary
1.10.7 it would hold ρ = |OA| < |OΓ | = ρ which is contradictory, q.e.d.
Corollary 2.1.1 The center of a circle is always contained in the medial line of its chords.
(I) (II)
B B
O O
A A
Figure 2.1.3: Chord middle Middles of parallel chords
Proof: Obviously, since the center is equidistant from the chord end points (Figure 2.1.3-I), it will be
contained in the medial line of the chord (Corollary 1.9.1), q.e.d.
Corollary 2.1.2 The orthogonal line, from the center to a chord of a circle, goes through the chord’s middle.
Corollary 2.1.3 The middle points of parallel chords of a circle are contained in the diameter, which is
orthogonal to these chords.
Proof: From the center of the circle we draw an orthogonal line to one such chord (Figure 2.1.3-II). This
line will pass through the chord’s middle and will be orthogonal also to any other parallel chord, therefore
it will pass through the middle of any such chord (Corollary 2.1.2), q.e.d.
2.1. THE CIRCLE, THE DIAMETER, THE CHORD 59
Corollary 2.1.4 Every diameter of the circle is also an axis of symmetry of the circle.
Corollary 2.1.5 The center of a circle is also a center of symmetry of the circle.
Exercise 2.1.1 Show that an axis of symmetry of the circle, must necesserily pass through its center. Show
also, that a center of symmetry of the circle must coincide with its center.
Proposition 2.1.2 The chord length, in a circle of radius ρ, is less than or equal to the length 2ρ of the
circle’s diameter. If the chord has length 2ρ then it coincides with the circle’s diameter.
Proof: If O is the circle’s center, δ = |AB| is the length of the chord and the points A, B, O are not collinear,
then they defined a triangle ABO. This triangle is isosceles and its legs have length ρ. Consequently, using
the triangle inequality, δ = |AB| < |OA| + |OB| = 2ρ. The last inequality also implies that the chord
cannot have length 2ρ when points A, B and O are not collinear. Latter proves the second part of the claim,
q.e.d.
Theorem 2.1.1 For every triple of non-collinear points A, B and Γ , there exists a unique circle passing
through them.
Γ
(Ι) (ΙΙ)
Ρ N
O O
A
A B
M A'
B
B'
Figure 2.1.4: Circumscribed circle (ABΓ ) Equal chords, congruent isosceli triangles
Proof: The medial lines of line segments AB and BΓ respectively, intersect at point O (Figure 2.1.4-I).
Had they not intersected, then they would be parallel and from B we would have verticals BM , BN to two
parallels, therefore M , N and B would be collinear, which is contradictory. From the first medial line we
have (Corollary 1.9.1) |OA| = |OB| and from the second medial line |OB| = |OΓ |. It follows then that
|OA| = |OΓ |, therefore O also belongs to the medial line of segment AΓ . In other words, we proved that
the three medial lines of the sides of triangle ABΓ pass through the same point O, which is equidistant from
the triangle’s vertices, or |OA| = |OB| = |OΓ | = ρ. Consequently, the circle with center O and radius ρ
passes through the triangle’s vertices. The fact, that this circle is unique, follows using the same argument.
Every other circle, that would pass through the vertices of the triangle ABΓ , would have its center on the
medial lines of the triangle sides, since these sides would be circle chords (Corollary 2.1.1). Therefore the
circle’s center and radius would coincide with those of the circle constructed previously, q.e.d.
Corollary 2.1.6 For every triangle ABΓ there exists a unique circle passing through the triangle’s ver-
tices.
We call the circle of the previous corollary circumscribed or circumcircle of the triangle. The center of
this circle is called circumcenter of the triangle and, as we saw in the proof above, it coincides with the
intersection point of the medial lines of the sides of the triangle. Often the circle which passes through three
non collinear points A, B and Γ is denoted by (ABΓ ).
Exercise 2.1.2 Show that equal chords AB and A B , on the same circle with center O, define congruent
(isosceli) triangles AOB and A OB (Figure 2.1.4-II).
’ | = 90◦ , then
Corollary 2.1.7 If a point A sees the line segment BΓ under a right angle, that is, if | BAΓ
it is contained in the circle with diameter BΓ .
Proof: If BΓ is seen under a right angle from A, then, according to Corollary 1.15.18, the median from A
to BΓ will be half of it, consequently the points A, B and Γ will belong to the circle with diameter BΓ ,
q.e.d.
Corollary 2.1.8 For every point A of the circle, with diameter BΓ , the angle BAΓ ’ is a right one. Equiv-
alently: a diameter is seen from each point of the circle (except its end points) under a right angle. Con-
versely, if a circle chord is seen from one of the circle’s points under a right angle, then it is a diameter of
the circle.
Α
(Ι) (ΙΙ) Μ
γ β ε
γ 2β 2γ β
Γ Β Α Β
O
Figure 2.1.5: Diameter viewed under right angle, Projection to line
Proof: Indeed, if BΓ is a diameter and A is a point of the circle (different from B and Γ ), then the triangles
OAB and OAΓ are isosceli with base angles β and γ respectively, of triangle ABΓ (Figure 2.1.5-I). The
sum of the (external) angles of the two isosceli at O is 2β + 2γ = 180 ◦ (Corollary 1.15.2), therefore
β + γ = 90◦ , which shows that ABΓ is right angled at A. Conversely, if the chord BΓ is seen under a
right angle from A, then, according to the previous corollary, A will belong to the circle with diameter BΓ ,
which, having three common points (A, B and Γ ) with the given circle, must coincide with it, q.e.d.
Exercise 2.1.3 Show that between two circle chords α and β, the longer one lies at a lesser distance from
the center.
Exercise 2.1.4 Show that if three points X, Y and Z lie on the circle κ, then their symmetric X , Y , Z
(relative to an axis or a point) will lie on circle κ which is congruent to κ.
Exercise 2.1.5 What is the geometric locus of centers of circles κ which pass through two fixed points A
and B?
Exercise 2.1.6 Let A and B be two fixed points and ε be a line passing through B. Let also M be the
projection of A onto ε. Find the geometric locus of M as line ε rotates about B.
Hint: When line ε does not coincide with line AB or its vertical at B, then triangle ABM (Figure 2.1.5-II)
is a right triangle with fixed hypotenuse AB, therefore according to the previous corollary point M will
belong to the circle with diameter AB.
Exercise 2.1.7 Let α and β be two concentric circles. Then a line ε, which intersects both circles, defines
line segments AB, Γ∆ contained between the circles, which are equal.
Exercise 2.1.8 Let ε be a line A, B be two points not contained in ε. Find a point Γ of the line, which is
equidistant from A and B. Has this problem always a solution?
2.2. CIRCLE AND LINE 61
2.2 Circle and line

Indispensable condition for happiness is activity. One must be
busy, and if possible, create something, or at least learn something.
A. Schopenhauer, Aphorismen
In this section we examine the relative positions of a line and a circle. The key is the distance between the
center of the circle and the line. This distance, related to the radius of the circle, determines if we will have
one, two or no points in common of the two shapes.
(Ι) (ΙΙ)
Ο
Ο Α
Δ Α Γ Β
Figure 2.2.1: Intersection of line and circle Intersection of circle and half line
Proposition 2.2.1 A line AB contains at most two points of a given circle .

Proof: We show that a line cannot contain three or more points of a given circle κ. Indeed, if the line had
three points A, B and Γ in common with κ, then one of the three would lie between the other two which
define a circle chord anyway (Figure 2.2.1-I). This way we would have a chord which would contain a third
circle point. Something which we have excluded with Proposition 2.1.1, q.e.d.
Corollary 2.2.1 Each half line with end point coincident with the center of a circle, intersects this circle at
exactly one point (Figure 2.2.1-II).
Tangent of a circle is called a line, which has exactly one common point with the circle. This point is called
contact point of the circle and the line.
(Ι) (ΙΙ)
Ο Ο
Γ
Δ Β
Α Β Γ Α
Figure 2.2.2: Circle tangent Lines through point A of the circle
Theorem 2.2.1 A line which has only one common point A with a circle, is orthogonal at A to the radius
OA of the circle.
Proof: Using reduction to contradiction. If the line was not orthogonal to OA, then suppose OB is the
orthogonal and define Γ on the line, so that B is the middle of AΓ (Figure 2.2.2-I). Then triangles OBA
and OBΓ would be right and congruent, as having OB common, BA and BΓ equal and the contained
’ and OBΓ
angles OBA ’ equal as right. Then triangle OAΓ would be isosceles and we would have |OΓ | =
|OA| = ρ, in other words Γ would belong to the circle and the line would have another common point with
the circle besides A, which is contradictory, q.e.d.
Corollary 2.2.2 Every line which passes through a point A of a circle, except the orthogonal to the radius
OA, intersects the circle at a second point B. The excepted orthogonal to OA at A is the unique tangent to
the circle at A (Figure 2.2.2-II).
Α
(Ι) (ΙΙ)
Ζ
Η
Ο Ι
ρ
ε Γ
Β
Β Ε
Figure 2.2.3: Intersection of line and circle Inscribed circle in triangle
Corollary 2.2.3 Given a line ε and a point O, not lying on it, the circle O(ρ) intersects the line, if and only
if ρ > |OB| where B is the projection of O onto ε (|OB| is the distance of point O from the line ε). When
ρ = |OB|, then the circle is tangent to the line at B. Finally when ρ < |OB| the circle does not intersect
the line (Figure 2.2.3-I).
Theorem 2.2.2 The three angle bisectors of triangle ABΓ pass through a common point I, which is the
center of a circle I(r) simultaneously tangent to all three of the triangle’s sides.
Proof: Let I be the intersection point of two angle bisectors and specifically those of angles at B “ and Γ

(Figure 2.2.3-II). We will show that the third bisector of angle A also passes through I. Indeed, according
to proposition 1.15.16, I’s distances from the sides of angle B: “ IE and IH will be equal. Similarly I’s
distances from the sides of angle Γ : IE and IZ will be equal. Consequently the three distances IE, IH
and IZ will be equal, therefore they are radii of the circle with center I and radius r = |IE|. The equality
of distances IH and IZ from the sides of angle A shows that I also belongs to the bisector of angle A

(Proposition 1.15.16). The orthogonality of the sides to these radii at their end points shows (2.2.2) that the
circle is tangent to all three sides of the triangle, q.e.d.
Point I, ensured by the previous proposition, is called triangle incenter. The circle with center I, which
is tangent to all three sides of the triangle, is called inscribed circle or incircle of the triangle. Its radius
r = |IE| = |IZ| = |IH| is equal to the distance of I from any of the triangle’s sides. Note that from
the congruence of right triangles EIΓ , ZIΓ follows that |Γ E| = |Γ Z| and similarly |BE| = |BH|,
|AH| = |AZ|.
3ω 4ω Y
ω 2ω Θ K
Ζ
Δ
Β
Ο
Α Γ Ε Η I X
Figure 2.2.4: Multiples of the angle ω
’ we define alternatively points A, B, Γ , . . . at equal

Exercise 2.2.1 On the sides of the acute angle XOY
distances |OA| = |AB| = |BΓ | = . . . . Show that, in these isosceli triangles, the base angles are ω, 2ω,
’ |, (Figure 2.2.4). Can we construct infinitely many such triangles?
3ω, . . . , where ω = |XOY
2.2. CIRCLE AND LINE 63
(Ι) (ΙΙ) A
O Μ
Υ κ κ'
ε
N
A B B
M
Γ
Figure 2.2.5: Minimal chord through point Application of symmetry
Exercise 2.2.2 Show that the chord of least length XY , of circle κ(O, ρ), which passes through a fixed
point M (other than O) in the interior of the circle, is the orthogonal AB to OM (Figure 2.2.5-I).
Hint: Triangle OM N has hypotenuse |OM | > |ON |, therefore the isosceles OAB has a smaller base than
that of the isosceles OXY (Exercise 2.1.3).
Often the distance |OM | of the middle of the chord from the center is called apothem of the chord.
Exercise 2.2.3 Show that the symmetric X of the points X of the circle κ, relative to a fixed point M , are
contained in a circle κ congruent to κ.
Exercise 2.2.4 Given is a circle κ, a point M and a line ε. Construct a line segment AB, with end points
respectively on κ and ε, and having its middle at M (Figure 2.2.5-II). When does the problem have (or
doesn’t have) a solution?
(Ι) κ (ΙΙ)
Α β
x Γ'
Γ'
λ Β' Κ
y z μ
I
Γ κ
Β Γ
Α' α
Ο Α Β
Figure 2.2.6: Three tangent circles Minimization of diameter
Exercise 2.2.5 Construct circles κ, λ, µ, which are pairwise tangent and have their centers respectively at
the triangle vertices A, B and Γ (Figure 2.2.6-I). Show that the half-perimeter τ of the triangle, and the
respective radii x, y, z of these circles, satisfy the equalities (with a = |BΓ |, b = |Γ A|, c = |AB|):
τ = 12 (a + b + c), y = 12 (c + a − b) = τ − b,
x = 12 (b + c − a) = τ − a, z = 12 (a + b − c) = τ − c.
Exercise 2.2.6 In the previous exercise, show that the circle contact points A , B , Γ lie on the sides of
triangle ABΓ , whose incenter I coincides with the circumcenter of triangle A B Γ .
Exercise 2.2.7 The lines {α, β} are orthogonal at O, points {A, B} are fixed on α and point Γ varies on β
(Figure 2.2.6-II). In the circle κ = (ABΓ ) consider the diametrically opposite Γ of Γ . Find the geometric
locus of Γ and the position of the circle which minimizes |ΓΓ |.
Exercise 2.2.8 Show that point I, on the bisector A∆ of triangle ABΓ , is the intersection point of the
‘
‘ | = 90◦ + |B AΓ |
angle bisectors, if and only if | BIΓ 2 .
2.3 Two circles

The child searching a melody on the piano is
the prototype of every true composition.
T.W. Adorno, The sociology of music
We examine here the relative positions of two circles. The key to the subject is the distance between the
centers of the circles in relation to the magnitude of their radii. The next theorem shows that there exist
three possibilities: (a) two different intersection points, (b) one intersection point, (c) none. Of fundamental
importance also is theorem 2.4.1, which completes the triangular inequality (§ 1.11), showing the existence
of a triangle with given lengths of sides {a, b, c}, which satisfy this inequality.
Theorem 2.3.1 Two different circles have at most two common points.
Proof: Because if they had three or more, they would be coincident, according to Theorem 2.1.1, q.e.d.
(I) (II)
B
P A B
O
E
Figure 2.3.1: Common chord and centerline Contact point on centerline
The line segment AB of the common points A and B of two circles, which intersect at two different points
(Figure 2.3.1-I), is called common chord of the two circles. Two circles which have the same center
are called concentric. The line which joins the centers O and P of two non-concentric circles is called
centerline of the circles. We often also call centerline the line segment OP , which joins the centers. What
exactly we mean in each case, will be made clear from context.
Corollary 2.3.1 The centerline OP of two circles, which intersect at two different points A and B, coin-
cides with the medial line of their common chord AB.
Proof: According to Corollary 2.1.1 the centers of the two circles will belong to the medial line of their
chord AB, q.e.d.
Two circles are called tangent when they have exactly one common point A (Figure 2.3.1-II). This point is
called contact point of the two circles.
2.3. TWO CIRCLES 65
Proposition 2.3.1 The contact point of two tangent circles belongs to their centerline. Conversely, if two
different circles have one intersection point, belonging to their centerline, then this point is unique and the
circles are tangent.
O M P
Figure 2.3.2: Symmetry relative to centerline
Proof: Using reduction to contradiction. Suppose that the contact point A of the two circles does not belong
to their centerline OP (Figure 2.3.2). Draw then the orthogonal AM from A to the centerline OP and
extend it by doubling it until B. The triangles OM A and OM B are congruent (SAS-criterion). Similarly
triangles P M A and P M B are congruent. It follows, that |OB| = |OA| = ρ and |P B| = |P A| = ρ ,
where ρ and ρ are the radii of the two circles. Consequently, besides A, we find another point B common
between the two circles, a contradiction. Conversely, if the circles have two different intersection points A
and B, then their centerline coincides with the medial line of AB (Corollary 2.3.1) and none of the two
points may belong to the centerline. Hence, if there is an intersection point on the centerline, then it is
unique, q.e.d.
ρ'
Ο Ν Ο'
ρ ρ' Ν Ο' Ο
Figure 2.3.3: ρ + ρ = |OO | |ρ − ρ | = |OO |
Corollary 2.3.2 Two circles O(ρ) and O (ρ ) have exactly one common point (they are tangent), if and
only if one of the following equations hold:
ρ + ρ = |OO |, |ρ − ρ | = |OO |.
In the first case we say that the circles are tangent externally and in the second we say they are tangent
internally (Figure 2.3.3).
Theorem 2.3.2 Two circles O(ρ) and O (ρ ) intersect at two different points, if and only if their centerline
|OO | and their radii satisfy the triangle inequalities
|ρ − ρ | < |OO | < ρ + ρ .
Proof: If the circles intersect and one (of the two) intersection point(s) is A (Figure 2.3.4-I), then this point
forms triangle OAO , with sides of length ρ, ρ and |OO | (Proposition 2.3.1). According to Theorem 1.11.1
and Theorem 1.11.2, the triangle inequalities will be satisfied. Conversely, if these inequalities are satisfied,
then the two circles cannot be tangent, because then we would have ρ + ρ = |OO | or |ρ − ρ | = |OO |.
Δ Γ Δ
(I) (II)
Ι
Ζ Ε
Θ Ε
Η
Α Β A B Γ
Figure 2.17.15: Parallelogram and rhombus Locus of ∆ for equal circles
Exercise 2.17.41 AB, BΓ are successive line segments of the same line. We construct congruent circles
which have, respectively, AB, BΓ as chords. Find the geometric locus of their other intersection point ∆
(Figure 2.17.15-II).
Exercise 2.17.42 Construct a triangle ABΓ , which has medials of its sides three given lines α, β, γ,
passing through a common point O.
Exercise 2.17.43 The circles {κ 1 , κ2 } intersect at points {A, B}, through which pass correspondingly lines
{α, β}. These intersect the circles a second time at points {A 1 , A2 } the line α and at points {B 1 , B2 } the
line β. Show that the lines {A 1 B1 , A2 B2 } are parallel. What happens when {A 1 , B1 } coincide (Exercise
2.13.18);
(Ι) (ΙΙ)
κ
Γ Β Ε Β
60° Δ Δ Ε
κ
Α Γ
120°
Figure 2.17.16: Seven congruent regular hexagons inside a circle
Exercise 2.17.44 In a given circle κ inscribe seven congruent regular hexagons (from Pappus’ Synagogue
[Pap76, p. 1097]) (Figure 2.17.16).
Hint: In figure 2.17.16-I, point A is constructible because it sees B∆ under the angle of 60 ◦ and |∆B|/|BΓ | =
2 (proof of Pappus). Similar proof in figure 2.17.16(II), in which |AΓ |/|Γ B| = 2.
Exercise 2.17.45 Given is a line ε, a point M of it and a point H outside this line. To construct a triangle
ABΓ whose base BΓ is contained in ε, has its middle at M , its orthocenter at H and the angles at the
base satisfy the relation β − α = 90 ◦ .
Exercise 2.17.46 The line ε passes through the apex A of the isosceles triangle ABΓ . To find a point M
on ε for which the distance |M B| + |M Γ | is the least possible. Then find the geometric locus of these
points M , when ε rotates about A.
Exercise 2.17.47 Given two triangles ABΓ and A B Γ , we construct a third triangle A B Γ as follows
(Figure 2.17.17): From the vertex Γ we draw the segment Γ B parallel and equal to A B , from A we draw
the segment AΓ parallel and equal to B Γ and from B we draw the segment BA parallel and equal to
Γ A . Show that the triangle A B Γ has the same centroid M as ABΓ .
Β'' Β'
Γ'' Ο Ν
Ρ
Μ Γ'
Β Γ
Α'
Α''
Figure 2.17.17: Triangles with common center of mass
Hint: From A draw the segment A N parallel and equalto Γ B . The triangle BA N is congruent to
Γ A B , therefore BN is parallel and equal to Γ A. The center of mass of A B Γ and ABΓ coincides
with that of Γ Γ N .
(Ι) (ΙΙ)
Α Α
Η Θ Η
Δ
Ε Ζ Ε
Β Γ
Β Ζ Γ
Figure 2.17.18: Quadrilateral of tangent circles Equilaterals from three congruent circles
Exercise 2.17.48 Two circles are tangent at a point E. Show that the quadrilateral ABΓ∆, with vertices
their points of contact with their common tangents, is circumscriptible to a circle with center E (Figure
2.17.18-I).
Exercise 2.17.49 With radius the leg r = |AB| of an isosceles triangle ABΓ we draw circles with centers
points A, B, Γ . Show that the intersection points of these circles form two equilateral triangles ∆EZ and
∆HΘ (Figure 2.17.18-II).
Hint: Because of symmetry ZE∆ is isosceles, |∆Z| = |∆E|. Also show that the triangles EΓ Z and EΓ∆
’
are congruent and the angle Γ EZ is 30◦ .
Exercise 2.17.50 Construct a quadrilateral for which are given the measures of two opposite of its angles,
the lengths of its diagonals and the measure of their angle.
Exercise 2.17.51 Given the triangle ABΓ , points A , B and Γ are defined as the intersections of its
bisectors with the circumcircle. Show that the triangle A B Γ has the bisectors of ABΓ as altitudes
(Figure 2.17.19-I).
Ζ
(Ι) A (ΙΙ) Η Ε
Β'
Γ'
Θ Ο Δ
ε Ρ Ρ'
B Γ Ι P0
Γ
Α' Α Β
Figure 2.17.19: Bisector traces Sum of distances
Exercise 2.17.52 Show that for an arbitrary point P in the interior of a regular polygon the sum of its
distances from the sides of the polygon is fixed.
Hint: Show first, using exercise 2.5.1, that this sum of distances does not change if the point P moves on a
line ε, which is parallel to a side of the polygon (Figure 2.17.19-II). Then let the point P take the position
P0 of the projection on ε of the center O of the polygon, parallel to another side of the polygon. Then,
apply the same remark for the line P 0 O, thus, showing this sum to be equal to the sum of distances of the
center O from the sides of the polygon (alternatively, see exercise 3.14.27).
dP tP
(Ι) (ΙΙ) A
Ο
Ν ε' P
O ε
ε
κ
Μ
Α
Figure 2.17.20: Secant of parallels Geometric locus
Exercise 2.17.53 From the point O outside the two parallels {ε, ε } we draw lines intersecting them at
points correspondingly {M, N }. To find the secant for which the distances {|AM |, |AN |} from a fixed
point are equal (Figure 2.17.20-I).
Exercise 2.17.54 Line ε passes through the center of the circle κ(O). For every point P of ε outside the
circle, we draw the tangent t P of κ from P always on the same side of ε. To find the locus of the projections
A of the center O of κ on the bisector d P of the angle of lines {ε, t P } (Figure 2.17.20-II).
Exercise 2.17.55 Show that a quadrilateral is a parallelogram, if and only if it admits a center of symmetry.
Hint: Combine the Exercise 2.6.7 and Exercise 2.9.3.
Exercise 2.17.56 In the triangle ABΓ , two points {X, Y } on its base BΓ , have equal sums of distances
from the other sides. Show that the triangle is isosceles with base BΓ .
Exercise 2.17.57 In the triangle ABΓ , from a point X on its base BΓ draw parallels to its sides {AB, AΓ }
intersecting the other sides respectively at {B X , ΓX } and define d X = |XBX | + |XΓX |. Show that, if for
two points {X, Y } of the base BΓ : d X = dY , then the triangle is isosceles.
Α
(Ι) Α (II)
κ
Ε Ζ
Β' Δ
γ/2
κ3 Ι
Γ'' κ2
Γ' Δ
Α'' β/2
Β''
γ/2
Β Γ Β Θ Η Γ
Α' κ1
Figure 2.17.21: Fixed angles Intersection on circle
Exercise 2.17.58 Triangles {ABΓ , A B Γ } have a common circumcircle κ. The first remains fixed and
the second is rotating inside the circle without to change its shape. Show that the lines {AA , BB , ΓΓ }
are side-lines of a triangle A B Γ with constant angles and its vertices are moving on three circles
{κ1 , κ2 , κ3 } which pass through a common point ∆ (Figure 2.17.21-I).
Hint: Apply exercise 2.13.17.
Exercise 2.17.59 In the triangle ABΓ the bisector from B and line ∆E, joining the contact point of
{AB, AΓ } with the inscribed circle, intersect at point Z. Show that Z is on the circle with diameter BΓ
Ι A
Δ Ε Γ
(Ι) (ΙΙ)
Μ Z λ E
Ν Ο
Υ K
κ Η
M
I μ
K
Ε' ε
A Χ Β B X Δ Γ
Figure 2.17.22: Variable circles passing through two fixed points
Exercise 2.17.60 Given is a rectangle ABΓ∆, and a fixed point E of the side Γ∆ (Figure 2.17.22-I).
For every point X of the line AB we define the orthogonal projection Y of B on line EX and the circle
κ = (AXY ). Show that
1. All the circles κ pass through a fixed point I of the diagonal AΓ .
2. If H is the intersection point of the lines {EE , BY } and Z is the second intersection point of the
circles {κ, λ = (EY H)}, then line ZY passes through a fixed point K of the circle µ = (EΓ B).
Hint: For (1) define the intersection point M of the lines {A∆, BY } and see that M is contained in
κ = (AXY ). Point Y is also contained in the circle µ = (BΓ EE ). If I denotes the second intersection
point of the circles {κ, µ}, show that I is contained in AΓ .
For (2) observe that IY’ K =Γ ’AE, which is a fixed angle.
Exercise 2.17.61 Points {N, E} on the sides {AB, AΓ } of the triangle ABΓ are fixed and X moves on the
line BΓ (Figure 2.17.22-II). Show that the common chord XK of the circles {β = (BXN ), γ = (XΓ E)}
passes through a fixed point I, lying on the parallel of BΓ from A.
CHAPTER 3
Areas, Thales, Pappus, Pythagoras
3.1 Area of polygons

That I cannot know any substance, as I can have no ideas but of their
qualities, and that a thousand qualities of a thing cannot reveal the
intimate nature of this thing, which may possess a hundred thousand
other qualities that I am unaware of.
Voltaire, The ignorant Philosopher p. 44
After the lengths of line segments and the measures of angles, the third magnitude we measure in plane
Euclidean Geometry is the area of polygons. The area of a convex polygon Π is a positive number (Π ),
of which we require the following properties.
Property 3.1.1 Two congruent polygons have equal areas.
I
(Ι) (ΙΙ) K
H Z
Δ
A A
E Z
B Γ
B
Γ Δ H Θ
Figure 3.1.1: (ABΓZH )+ (Γ∆EZ ) = (ABΓ∆EZH ) (ABΓ∆) < (ZHΘIK )
Property 3.1.2 A polygon Π , which is composed from other, not overlapping, finite in number polygons
Π , Π , ... , has area the sum of the areas of the polygons
(Π ) = (Π ) + (Π ) + ...
.
136 CHAPTER 3. AREAS, THALES, PAPPUS, PYTHAGORAS
In figure-3.1.1-I the two polygons ABΓ ZH and Γ∆EZ have the side Γ Z in common and form the new
polygon ABΓ∆EZH.
Property 3.1.3 A polygon Π contained in another one Π (Figure 3.1.1-II)has less area: (Π ) < (Π ).
Property 3.1.4 The area of a square with unit side is one.

Comment-1 As we will see below, these four properties we require for the area, completely determine its
form, and for rectangles lead to the well known expression for area, as the product of the lengths of its sides.
Its existence, therefore, is a consequence of the axioms. For this reason I speak about the “properties” and
not about the “axioms” of area. I could speak about “axioms” satisfied by the area-function, which, however
do not belong to the “axioms of geometry”.
Comment-2 Property 3.1.2 could be reduced inductively to the corresponding additivity of areas for poly-
gons which are composed of only two other polygons. Also the property 3.1.3 is a consequence of property
3.1.2 and could be proved using the other properties. However the proof contains several subtle points,
which in the first stages of acquaintance with geometry it is not necessary for them to be analyzed further.
Comment-3 It is noteworthy to observe the similarities apparent in measurements of lengths and areas.
They are special cases of measures, in other words mechanisms of measuring length and more generally,
content (area, volume), which are particular cases of the, so called, Jordan measures and more gener-
ally Lebesque measures (([KF70, p. 254])). These measures, beginning with measurements of very simple
shapes, extend and allow the measurement of more complex shapes than these which concern us in this
lesson. Special cases of the previous are also the volume of polyhedrons which we will meet in solid ge-
ometry (§ 10.5) as well as the area of spherical polygons (§ 10.3). In all cases the measurement mechanism
is constructed the same way. We require of it certain properties and we prove that these requirements imply
the existence of a unique mechanism which satisfies them.
Exercise 3.1.1 Show that every parallelogram is divided by one of its diagonals into two triangles of equal
area. Show more generally, that every line which passes through the center of a parallelogram divides it
into two triangles of equal area.
Hint: The two triangles defined by the diagonal are congruent, therefore they have equal areas.
A B
O M
Δ Γ
Figure 3.1.2: Triangles of equal area
Exercise 3.1.2 Show that the two diagonals of a parallelogram divide it into four triangles of equal area.
Hint: Triangles AOB and ∆OΓ , where O is the center of the parallelogram ABΓ∆ (point of intersection
of the diagonals), are congruent, therefore have equal areas. Consider the symmetric points M , N of O
relative to the middles of BΓ and ∆Γ respectively. The parallelograms BOΓ M and O∆N Γ are congruent,
therefore their halves according to area (according to the previous exercise), which are BOΓ and ∆OΓ will
have equal areas.
Exercise 3.1.3 Show that the median A∆ of a triangle ABΓ divides it into two triangles of equal area.
Hint: Combination of the two previous exercises (see also Corollary 3.3.6).
3.2. THE AREA OF THE RECTANGLE 137
3.2 The area of the rectangle

Hold fast the time! Guard it, watch over it, every hour, every
minute! Unregarded it slips away, like a lizard, smooth,
slippery, faithless, a pixy wife. Hold every moment sacred.
Give each clarity and meaning, each the weight of thine
awareness, each its true and due fulfillment.
Thomas Mann, The Beloved Returns
The key for the calculation of areas of polygons is the area of the rectangle. In this section we begin from the
area of rectangles whose side lengths are integers and stepping by gradually, with the help of the properties
of areas, we arrive at the fundamental expression of the area = ab, as the product of the lengths of its
sides, for any values of a and b, integer or not.
Lemma 3.2.1 We divide the two opposite sides of the unit square (having side of length 1) into µ equal
parts and the other two opposite sides into ν equal parts and we draw parallels which join opposite lying
1
points. This forms µν parallelograms, each of them having area µν .
Δ Γ
H Z
A E B
1
Figure 3.2.1: (AEZH)= µν
Proof: Obviously there are formed µν congruent rectangles which according to the property 3.1.1 will have
the same area E. According to property 3.1.2 the area of the square which is 1 (Property 3.1.4) will be the
sum of the areas 1 = µνE, q.e.d.
Lemma 3.2.2 The area of a rectangle ABΓ∆, with sides AB and A∆, whose lengths are rational num-
bers, is equal to the product of these lengths = |AB||A∆|.
Δ Γ
A B
Figure 3.2.2: |AB| = α

β, |A∆| = γ
δ ⇒ (ABΓ∆) = |AB||A∆|
Proof: We assume that in the rationals |AB| = α β , |A∆| = δ the α, β, γ and δ are positive integers and
γ
we divide the unit square into βδ congruent rectangles as in the previous lemma, according to which each
1
has area = βδ . By the hypothesis the side AB of the rectangle can be divided into α number segments
of length β and the side A∆ can be divided into γ number segments of length 1δ . Drawing parallels from
1
the points of division of these sides we therefore form αγ number congruent rectangles, each of which has
1 1
area βδ . The sum of the areas of these rectangles is αγ βδ and is equal, according to the property 3.1.2, to
the area of the rectangle, q.e.d.
Lemma 3.2.3 For every positive number θ and every natural number ν there exists another natural number
µ (or zero), such that the following inequality holds
µ 1

θ − ≤ .
ν ν
0 1/ν 2/ν ............. μ/ν θ (μ+1)/ν
Figure 3.2.3: Archimedean axiom
Proof: In essence the lemma coincides with the so called Archimedean axiom (Archimedes 287-212 B.C.)
for line segments of one line and is pictured in figure-3.2.3. For the proof, starting at the beginning of a
segment of length θ, we place successive intervals of length ν1 . According to the Archimedean axiom, there
is a first critical µ, such that (µ + 1) ν1 exceeds the length θ, in other words θ finds itself in the interval (of
length ν1 ) between µν and µ+1 ν , q.e.d.
Theorem 3.2.1 The area of the rectangle ABΓ∆, with sides AB and A∆, is equal to the product of the
lengths of the sides = |AB||A∆|.
Δ
Γ
θ'
........
0 1/ν 2/ν 3/ν μ/ν (μ+1)/ν

θ
Figure 3.2.4: Rectangle approximation
Proof: For the proof we consider the lengths of the sides θ = |AB| and θ = |A∆| and a (fairly large) ν.
According to the previous lemma, there exist two corresponding integers µ and µ satisfying

µ 1 µ 1
θ − ≤
and θ − ≤ .
ν ν ν ν
This way, the area of the rectangle is greater than
µ µ
≤ ,
ν ν

since it contains the rectangle with sides of rational length µν and µν . Thinking similarly, we see that the
µ +1
same area is simultaneously less than the area of the rectangle with sides of rational length µ+1
ν and ν ,
in other words
µ + 1 µ + 1
≤ .
ν ν
3.2. THE AREA OF THE RECTANGLE 139
In total therefore we see that the numbers θθ and satisfy the same inequalities
µ µ µ + 1 µ + 1 µ µ µ + 1 µ + 1
≤ θθ ≤ and ≤≤ .
ν ν ν ν ν ν ν ν
Therefore their difference will satisfy the inequality
µ + µ + 1 1 µ 1 µ 1 1 1 1 θ + θ + 1
| − θθ | ≤ 2
= + + 2 ≤ θ + θ + 2 ≤ .
ν νν ν ν ν ν ν ν ν
Because the right quantity can become arbitrarily small, provided we choose a big ν, it follows that the
(fixed) quantity on the left cannot be strictly positive, q.e.d.
Corollary 3.2.1 The area of a square with side length α is equal to α 2 .
ab b2
a2
a2 ab
a
a b
Figure 3.2.5: Area of square Square of sum geometrically
Exercise 3.2.1 Given the positive numbers a and b, consider the square with side length a + b (and a − b)
(Figure 3.2.5) and show, the identities (a + b) 2 = a2 + b2 + 2ab .
Δ H Γ
(Ι) 1 1 (ΙΙ) (III)
2 M Μ
4 Λ
Λ Z
5
Θ I
2 4 Ι
K Κ
3 3
A E B
Figure 3.2.6: Square division
Exercise 3.2.2 If E, Z, H, Θ are the side middles of the square ABΓ∆, show that Γ E, ∆Z, AH and
BΘ define the square IKΛM , which has area 1/5 that of ABΓ∆ (Figure 3.2.6-I). Compute analogously
the area of the central square IKΛM , formed by dividing each side of a square in four equal parts (Figure
3.2.6-II). Generalize by dividing in 2ν parts (Figure 3.2.6-III).
Exercise 3.2.3 Similarly to exercise 3.2.1, show geometrically the identity (a − b) 2 = a2 + b2 − 2ab.
The Tangram consists of a partition of the square into 7 tiles and was a game in ancient China, with main
problem, the rearrangement of the tiles so that a polygon of given outline results ([Tia12]). Relatively re-
Figure 3.14.26: The 13 convex polygons which are constructed using Tangram
cently (1942), for example, it was proved that there are only 13 convex polygons which can be constructed
with these tiles. Figure 3.14.26 shows the initial partition of the square into 7 tiles, as well as, rearrange-
ments of the tiles which form each of these 13 polygons ([FTW42]).
Θ Δ Η
Ε Β Ζ
Figure 3.14.27: Theorem of Pick for parallelograms
Exercise 3.14.62 (Theorem of Pick (1859-1942), for parallelograms) Show, that every parallelogram p
with its vertices at nodal points of squared paper and containing in its interior κ nodal points and at its
boundary λ nodal points, has area given by the formula
λ
(p) = κ + −1
2
.
Hint: Three steps. (1) We easily see that the formula holds for parallelograms like EZHΘ , with horizontal
and vertical sides. (2) relying on (1), we see again easily that the formula holds for right triangles, like AEB
with one side horizontal and the other vertical. (3) Combination of the previous and the fact that each par-
allelogram ABΓ∆ defines another one which encloses it and has horizontal and vertical sides, like EZHΘ .
The theorem holds more generally for a polygon with vertices at nodal points ([AO12, p. 277], [RWG76]).
In the interesting collection of problems of Arnold “for children from 5 to 15 years old”, the simpler prob-
lem of the next exercise is proposed [Arn04, p. 10], without mentioning the formula of Pick. A consequence
of this formula is that of the one to next exercise.
(I) (IΙ)
Δ
Β
α
Α
Figure 3.14.28: Parallelogram of area 1 Tangent circles
Exercise 3.14.63 Show that on squared paper, a parallelogram with its vertices at nodal points (Figure
3.14.28-I), which does not contain in its interior or on its side another nodal point, has area equal to that
of the nodal square.
Exercise 3.14.64 Show that the formula of Pick holds also for triangles and that there is no equilateral
triangle with vertices at nodal points.
Hint: The proof for triangles reduces to that for parallelograms. For the equilateral triangle with vertices at
nodal points, the square of the altitude υ 2 is calculated easily when it is rational. The
√
area , according to
Pick, is also rational, therefore also υ
2 will be rational, which is however equal to 23 .
Exercise 3.14.65 Show that the radii of two externally tangent circles and tangent to two lines which inter-
sect at angle α (Figure 3.14.28-II), is given by the formula

1 + sin α2
r = r α.
1 − sin 2
Conclude the formula which connects the radii of the first and the last circle of a chain of ν circles, as in
figure 3.14.28-II.
Exercise 3.14.66 Given are three fixed points {A 1 , A2 , A3 } and positive constants {λ 1 , λ2 , λ3 , µ}. Show
that the geometric locus of points X, for which the sum is
λ1 |XA1 |2 + λ2 |XA2 |2 + λ3 |XA3 |2 = µ,
is a circle. Show the corresponding property for four points {A 1 , A2 , A3 , A4 }.
Hint: Use exercise 3.12.9. Replace λ 1 |XA1 |2 +λ2 |XA2 |2 with a term of the form ν|XB| 2 , using a suitable
fixed point B and a fixed number ν.
Exercise 3.14.67 For a point X, not lying on the sides of the rectangle ABΓ∆, show that |XA| 2 +|XΓ |2 =
|XB|2 + |X∆|2 .
The two next exercises come from, respectively, Euler and Fermat. The first one, which is solved easily
using the figure, is used by Euler to solve the second ([San15, p. 10], [Whi07, p. 314]). Here, the hint of the
second one, leads to a simpler solution which doesn’t use the first one.
Exercise 3.14.68 Show that for four collinear points {A, P, Σ , B} in this order (Figure 3.14.29-I), the
following relation holds |AB| · |P Σ | + |AP | · |Σ B| = |AΣ | · |P B|.
(I) Ι (ΙΙ)
Δ Γ
Α φ ω Β
ω Ρ Σ φ
Η Ζ
Κ Ε
A Ρ Σ B Α' Ρ' Σ' Β'
Figure 3.14.29: Relation of lengths problem of Fermat
√
Exercise 3.14.69 The rectangle ABΣ P has sides a = |AP |, b = |AB| = a 2 and point M is found
on the half circle with diameter AB (Figure 3.14.29-II). Show that the lines {M P , M Σ } intersect AB at
points {P, Σ }, for which holds |AΣ | 2 + |P B|2 = |AB|2 .
Hint: {AA , BB } form a right triangle with altitude equal to a = |AP |. Consequently holds (Proposition
3.5.2) |A P ||Σ B | = a2 = b2 /2 (∗ ). The next relations prove the proposition for the, proportional to the
given, segments defined by the points {A , P , Σ , B }.
|A Σ |2 + |P B |2 = (|A B | − |Σ B |)2 + (|A B | − |A P |)2

= 2|A B |2 + |Σ B |2 + |A P |2 − 2|A B |(|Σ B | + |A P |)
= 2|A B |2 + |Σ B |2 + |A P |2 − 2|A B |(|A B | − b)
= 2|A B |2 + (|Σ B | + |A P |)2 − [2|Σ B ||A P |] − 2|A B |(|A B | − b)
(∗)
= 2|A B |2 + (|A B | − b)2 − [b2 ] − 2|A B |(|A B | − b) = |A B |2 .
(Ι) (ΙΙ) (ΙΙΙ)

Θ Η
B Γ B Γ
Τ1 Τ1 Τ2
Τ1 = Τ2
b Δ Δ
y Τ2 y
a x
A x A
Ε Ζ
Figure 3.14.30: Table which unfolds and doubles its surface
Exercise 3.14.70 The table EZHΘ has two rectangular plates {T 1 , T2 } lying one on top of the other and
connected with a join along EZ (Figure 3.14.30-I). At a certain point ∆ on the table’s frame, beneath of
T1 , plate T1 is mounted so that it can turn around ∆ (Figure 3.14.30-II) and unfolding plate T 2 , after a turn
by 90◦ makes a rectangle of the double area, sides parallel to the initial rectangle and the same symmetry
center (Figure 3.14.30-III). At which point of T 1 must be located point ∆?
Hint: {x = (2a − b)/4, y = b/4}.
Exercise 3.14.71 From the point A, outside cirlce κ(O), we draw a line intersecting it at points {B, Γ },
and draw also the tangents {τ B , τΓ } of κ at these points. Assume that these tangents intersect the orthog-
onal δ of AO at A at points {B , Γ }. Show that A is the middle of B Γ .
Exercise 3.14.72 With diameters the sides {A B , B Γ , Γ A } of the orthic A B Γ of the triangle ABΓ
we draw circles correspondingly {γ, α, β} intersecting triangle’s sides {AB, AΓ } the circle α at {A 1 , A2 },
sides {BA, BΓ } the circle β at {B1 , B2 }, sides {Γ A, Γ B} the circle γ at {Γ1 , Γ2 }. Show that points
{A1 , A2 , B1 , B2 , Γ1 , Γ2 } are concyclic.
Α ζ
Ξ
δ
Μ Ν
η
Κ Θ
Η
Ε
Ζ Ι γ
β ε
Ρ
Β Δ Γ
Figure 3.14.31: Circles related to bisectors
Exercise 3.14.73 Consider the circles {β = (AB∆), γ = (AΓ∆)}, where ∆ the trace of the bisector A∆
on side BΓ of triangle ABΓ (Figure 3.14.31). Prove the following relations:
1. If {Z, Θ} are the centers of {β, γ} and {E, H} the intersections of the bisectors {BI, Γ I} with the
circles {β, γ}, then the points {E, H} are on the line ZΘ .
2. The points {B, Γ , E, H} are contained in the circle ε.
3. If {M, N } are the intersection points of {∆H, ∆E} with the corresponding sides {AB, AΓ }, then
the 6 points {A, M, H, I, E, N } are contained in a circle δ, where I is the incenter of the triangle
ABΓ .
4. The pairs {(KI, BΓ ), (IΓ , ∆E), (IB, ∆H)} consist of orthogonal lines, where K is the center of
the circle δ.
˘ AH∆}
Hint: (1) is obvious, since {E, H} will be the middles of the arcs { AE∆, ˘ of the circles {β, γ}.
’
(2) results immediately by showing that BEH = Γ /2.
= ∆AΓ
For (3) notice that A/2 ’ = BE∆ ’ = ∆HΓ ’ , which shows that {AIEN, AM HI} are cyclic
’ ’ ’ ’
quadrilaterals. Also HA∆ = HΓ∆ and EA∆ = EB∆ imply that AHIE is a cyclic quadrilateral.
For (4), the orthogonality of {IK, BΓ } follows from the fact that η = ZΘ is parallel to the external
’ = AΞ
bisector ζ at A and angle B∆I ‘I, where Ξ is the intersection of IK with δ. The two other pairs of
orthogonals result by easy angle measurements.
Exercise 3.14.74 Continuing the previous exercise prove also the following relations:
1. The quadrilaterals {ZKHA, Θ EKA, ZHBΛ, ΓΘ EΛ} are cyclic, where Λ is the center of ε.
2. The quadrilateral Θ ZΛK is also cyclic and its circumcenter coincides with the circumcenter of the
triangle ABΓ .
3. The line pairs {(KH, EΛ), (KE, ΛH), (ΓΘ , BZ)} intersect on the external bisector ζ.
Exercise 3.14.75 An isosceles trapezium has fixed basis a diameter of its circumscribed circle κ and the
side parallel to its base is a chord of κ. Find the place of the chord which maximizes the area of the
trapezium.
Exercise 3.14.76 Given is a triangle ABΓ . To find a fourth point ∆ so that the quadrilateral ABΓ∆ is
simultaneously inscriptible and circumscribable.
’ at corresponding
Exercise 3.14.77 Point P is projected on the sides and the bisector of the angle XOY
points {A, B, ∆}. Show that |M A| = |M B|.
Exercise 3.14.78 To find the geometric locus of points P whose projections on the sides of a fixed angle
’ at corresponding points {A, B} define lines AB of a fixed direction.
XOY
E Υ
H
A
I
Χ φ O
Δ B K
Z
Θ Ρ
Figure 3.14.32: Reflected lines
Exercise 3.14.79 Given is an angle XOY ’ and an inner point A of it (Figure 3.14.32). We think of the sides
as mirors and point A as a light source and AB a light beam with B moving on OX. The beam AB is
reflected to the BΓ , this to Γ∆, this to ∆E etc. Show that:
1. All lines BΓ pass through a fixed point Z.
2. All lines Γ∆ pass through a fixed point H.
3. All lines ∆E pass through a fixed point Θ .
4. The intersection point P of lines BΓ and ∆E belongs to circle (OZΘ ),
5. This circle passes through the intersection poin I of OY with HΘ .
Hint: Point Z is the symmetric of A with respect to OX. Point H is the symmetric of Z with respect to
OY . Point Θ is the symmetric of H with respect to OX. Angle Θ ’ ’ − Γ∆B
P Z = ABK ’ = (∆BΓ ’ −
’ ’ ’ ’ ’ ’
BΓ O) + (∆Γ E − E∆X) = 2φ. Similarly ZIΘ = 2ZHΘ = 2φ and ZKΘ = 2ZHΘ = 2φ. ’
Exercise 3.14.80 Given is an angle XOY ’ and points A and B in its interior. To find the trajectory of a
ball (broken line) which starting from A and reflecting alternatively on sides OX, OY and finally again on
OX, passes through B.
(I) Γ (II) A
D F
Δ G
x Ε
y H
z
Α Ζ Β B E C
Figure 3.14.33: z independent of |AB| Similar triangles
Exercise 3.14.81 Point E is the intersection of the hypotenuses of the two right triangles {ABΓ , AB∆}
and Z its projection on AB (Figure 3.14.33-I). Show that z = |EZ| does not depend on |AB| but only on
{x = |AΓ |, y = |B∆|}.
Exercise 3.14.82 Divide a triangle ABΓ in parts, which when reasambled, build two triangles similar to
ABΓ (Figure 3.14.33-II).
Exercise 3.14.83 Angle ω = XOY ’ of constant measure is turning about the center O of the circle κ(O) in-
tersecting it at points {A, B}. To find the geometric locus of intersection points P of the lines {AA , BB },
where A B is a fixed diameter of the circle.
Exercise 3.14.84 Points {A, B} move on parallel lines {α, β}, so that the segment AB has a constant
’
direction. From point P outside the band of the parallels we form the angle ω = AP B. To locate the place
of AB, for which this angle becomes maximal.
Exercise 3.14.85 Construct circles {α, β, γ} with diameters the medians {AA , BB , ΓΓ } of triangle
ABΓ . Show that these circles intersect by pairs at points {A 1 , A2 , B1 , B2 , Γ1 , Γ2 } contained in the alti-
tudes of the triangle, so that triangle A 1 A2 A3 is homothetic to ABΓ and triangle B 1 B2 B3 is homothetic
of the orthic of ABΓ .
Ε''
Α'
Ε'
Δ''
Α''
Α Ε
Β Δ
Β'
Γ Δ'
Γ' Γ''
Β''
Figure 3.14.34: Miquel’s theorem for pentagons
Exercise 3.14.86 (Miquel’s theorem for pentagons) Given is the pentagon ABΓ∆E, of which we extend
the sides, so that five triangles are formed: {ABA , BΓ B , Γ∆Γ , ∆E∆ , EAE }, the circumcircles of
which intersect at second points {A , B , Γ , ∆ , E }. Show that these points are concyclic (Figure 3.14.34).
Hint: A solution results, by a, so called, angle chasing. With this we show that the resulting five quadrilat-
erals which result by avoiding a vertex, e.g. ∆ are cyclic. For A B Γ E for example, we show that the
opposite angles { B , E
} are supplementary. We notice first, that pentagon B Γ ∆E E is inscriptible. In
fact, consider the circumcirlce κ of the triangle B ∆E and its angles at Γ :
◊
B ◊
Γ ∆ = B ’
Γ Γ + ΓΓ ’
∆ = Γ ÷
BA + ΓΓ ÷
∆ = 180◦ − AE E.
This shows that B Γ ∆E is cyclic. Similarly we show that B ∆E E is cyclic. Then, in B Γ E A we

notice that the opposite angles B = A ÷ ÷
B B + BB Γ and E = 180◦ − A ◊ ◊
E E − E E X, for which it
÷ ◊ ◊ ÷
is valid A B B = A AE = A E E and BB Γ = BB Γ = E E X. ◊ ◊
CHAPTER 4
The power of the circle
4.1 Power with respect to circle

Socrates: And that the good life, the beautiful life, and
the just life are the same; does that still hold, or not?
Crito: It does hold.
Plato, Crito
An important and with many applications quantity, depending on a circle and a point is that of the power
of point with respect to a circle. This quantity resides on the two next propositions, which correspond
to theorems 35, 36 and 37 of the third book of the Euclid’s elements ([Hea08, p. 73, II]). Extended use of
it appears for the first time in the work of Louis Gaultier (1803, [Gau13]), who introduced the notion of
radical axis (and radical plane of two spheres). The full development and its systematic use though, is due
to Jacob Steiner [Ste71, p. 17-76,I], who introduced the word power.
Theorem 4.1.1 Given a circle κ and a point X, assume that a line through X intersects the circle at points
A and B. Then the product |XA||XB| is independent of the direction of the line and depends only on the
position of the point relative to the circle.
A A
κ
κ
Γ Γ
Χ
B
Δ
B
Δ
Figure 4.1.1: Power of a point relative to a circle |XA||XB|
Proof: Consider two different lines which pass through point X and intersect the circle κ at points A, B
and Γ , ∆ respectively. The triangles XAΓ and X∆B are similar (Corollary 2.14.1). Consequently they
234 CHAPTER 4. THE POWER OF THE CIRCLE
have proportional sides:
|XA| |X∆|
= ⇒ |XA||XB| = |XΓ ||X∆|, q.e.d.
|XΓ | |XB|
The number p(X) = |XA||XB| if the point is external, p(X) = −|XA||XB| if the point is internal and
p(X) = 0 if X lies on the circle, is called power of the point X relative to the circle κ. Often we will
use also the symbols p(X, κ) or/and p κ (X). In the case where X is an external point we have two special
positions of the line: the tangents from X. We may consider this as a limiting case, in which A and B
coincide. Yet again similar triangles XAΓ and X∆A are formed (Figure 4.1.2-I), because of the fact that
A A
(I) (II)
ρ
Δ Γ Χ
O Χ
κ
κ
Figure 4.1.2: |XA| 2 = |XΓ ||X∆| p(X) = |XA|2 = |XO|2 − ρ2
’ is equal to Γ’
the inscribed angle X∆A AX which is formed from the chord Γ A and the tangent at its end
A (Theorem 2.13.3). Therefore it follows that |XA| |X∆|
|XΓ | = |XA| , which implies p(X) = |XΓ ||X∆| = |XA| .
2
We proved then the following corollary.
Corollary 4.1.1 For every point X, external to the circle κ, the power p(X) is equal to the square of the
tangent to κ from X.
A third expression for the power p(X) is found by drawing XO, which joins the point X with the center
O of the circle κ. If A is the point of contact of the tangent XA from X, then from the right triangle
OAX we have p(X) = |XA|2 = |XO|2 − ρ2 , where ρ is the radius of the circle (Figure 4.1.2-II). For
Χ
(I) κ E (II) B
ρ A
A B
O Χ O
Γ
Δ Υ
κ
Z
Figure 4.1.3: p(X) = |XO| 2 − ρ2 Concyclicity criterion
points X internal to the circle there are no tangents to it. However the previous expression, this time with
opposite sign, again gives the power of the point. This is seen by drawing the chord of the circle which
passes through X and is orthogonal to XO (Figure 4.1.3-I). The definition of the power of X and the right
triangle OXE lead to:−p(X) = |XA||XB| = |XE|2 = ρ2 − |OX|2 . Taking into account the fact that for
points on the circle holds p(X) = ρ 2 − |XO|2 = 0, we have the next proposition.
Proposition 4.1.1 Given a circle κ(O, ρ) of center O and radius ρ, the power of any point X on the plane
relative to the circle κ is given by the formula
p(X) = |OX|2 − ρ2 .
4.1. POWER WITH RESPECT TO CIRCLE 235
Proposition 4.1.2 Given two intersecting lines OX and OY the points A, B on OX and Γ , ∆ on OY are
concyclic, if and only if |OA||OB| = |OΓ ||O∆|.
Proof: As we saw, when points A, B, Γ and ∆ are on the same circle (concyclic)(Figure 4.1.3-II), then the
equation holds. Conversely, suppose that the equation holds and κ is the circle which passes through the
three points: A, B and Γ (Theorem 2.1.1). Suppose ∆ is the second intersection point of this circle with
OY . Then according to Theorem 4.1.1 we have |OA||OB| = |OΓ ||O∆ |. But from hypothesis we also
have |OA||OB| = |OΓ ||O∆| and therefore |O∆ | = |O∆| and points ∆ and ∆ coincide, q.e.d.
Proposition 4.1.3 Given are three non collinear points, A on the line OX and B, Γ on line OY . The circle
(ABΓ ) is tangent to OX at A, if and only if
|OA|2 = |OB||OΓ |.
Υ
(I) (II)
Β Γ
κ
Γ Y A X B
Ο A
Χ
Figure 4.1.4: κ = (ABΓ ) tangent to OX Characterization of harmonic tetrad
Proof: If OA is tangent to the circle κ = (ABΓ ) at A (Figure 4.1.4-I), then the relation holds according to
Corollary 4.1.1. Conversely, if the relation holds and we suppose that OX intersects κ at points A and A ,
then, according to Proposition 4.1.2 will hold |OA||OA | = |OB||OΓ |. However by hypothesis also holds
|OA|2 = |OB||OΓ |, which combined with the previous gives |OA| = |OA |, which means that points A
and A coincide, q.e.d.
Proposition 4.1.4 The pairs of collinear points (A, B) and (X, Y ) are harmonic conjugate, if and only if
p(Y ) + p(X) = |XY |2 ,
where p(X), p(Y ) are the powers of the points relative to the circle with diameter AB (Figure 4.1.4-II).
Proof: It follows from Corollary 1.17.5 that
p(Y ) = |Y A||Y B| = |Y Γ |2 , −p(X) = |XA||XB| = |XΓ |2 ,
where Y Γ is the tangent to the circle from Y , by hypothesis in the exterior of the circle, q.e.d.
(I) (II)
B A B
κ A λ
ε ε
E Γ E' E
Figure 4.1.5: Circle through A, B tangent to line ε

Exercise 4.1.1 Construct a circle κ, such that it passes through two given points A and B and is tangent
to a given line ε.
Hint: Assume that the wanted circle was constructed. Assume further that Γ is the point of intersection
of AB with the given line ε and E is the point of contact of the circle with ε (Figure 4.1.5-I). Point Γ is
determined from the givens of the problem and its power relative to the circle will be p(Γ ) = |Γ A||Γ B| =
|Γ E|2 . Therefore the length |Γ E| and consequently the position of E is also determined from the givens.
Consequently one more point (E) of the circle is determined. In general there are two solutions which
correspond to two positions of E symmetric relative to Γ . Note that the problem doesn’t have a solution
when the points A and B belong to different sides of ε.
Also note, that in the case where Γ doesn’t exist, in other words, when the line AB is parallel to ε
(Figure 4.1.5-II), the problem has only one solution and point E is determined by intersecting line ε with
the medial line of AB.
Exercise 4.1.2 Given is a circle κ and a constant δ. Find the geometric locus of points X for which the
power p(X) relative to the circle κ is equal to δ.
(I) (II) Β
Γ Γ
B
A Α
ω ε'
Χ Α' Β' Γ'
Α' ε
Δ
Figure 4.1.6: Point X : |XA| = |XB| Collinear points A , B , Γ
Exercise 4.1.3 Given is a convex angle ω and a point A in its interior. Find a point X on one side of the
angle which is equidistant from the point A and its other side (Figure 4.1.6-I).
Exercise 4.1.4 From a point ∆ on the circle κ are drawn three chords ∆A, ∆B, ∆Γ , as well as a parallel
ε of the tangent to ∆, which intersects the chords at A , B and Γ respectively. Show that |∆A||∆A | =
|∆B||∆B | = |∆Γ ||∆Γ | (Figure 4.1.6-II).
Exercise 4.1.5 The median AM of a given triangle ABΓ intersects the circumcircle of the triangle at ∆.
Show that
1
|AM ||A∆| = (|AB|2 + |AΓ |2 ).
2
(Ι) (ΙΙ)
tB μΒ α
μ
κ Α Ο Β Α
Β
λ
Δ κ
Κ ε Z
Δ Α' Ε
Γ Β'
Figure 4.1.7: Locus of point ∆ Locus of centers of circles

4.2. GOLDEN SECTION AND REGULAR PENTAGON 237
Exercise 4.1.6 The variable circle α is tangent to circle κ(K) at B and passes through the fixed point A.
To find the locus of intersections Γ of the medial line µ B of AB with the tangent t B to κ at B (Figure
4.1.7-I).
Exercise 4.1.7 The variable points {∆, E} of line ε have constant product of distances from the point Z of
ε and we consider circles λ, passing through {∆, E} and the fixed point A, not lying on ε. Show that: (1)
The circles κ with diameter ∆E pass through two fixed points {B, B }. (2) The circles λ pass also through
two fixed points {A, A }. (3) The lines {AA , BB } intersect at Z.
4.2 Golden section and regular pentagon

Such is always the pursuit of knowledge. The celestial
fruits, the golden apples of the Hesperides, are ever
guarded by a hundred headed dragon which never
sleeps, so that it is an Herculean labor to pluck them.
Henry David Thoreau, Writings v.9

√
Two famous numbers, which √ satisfy a quadratic equation are: 2, which satisfies the equation x 2 − 2 = 0,
and the golden section φ = 5+1 2
2 , which satisfies equation x − x − 1 = 0. Geometrically, the number φ
is defined indirectly from the division of a line segment into two parts, such that:
The ratio of the whole to the larger one is equal to the ratio of the larger to the smaller one.
This is traditionally expressed using the terminology of the next construction.
Construction 4.2.1 (Golden section) On a given line segment AB find a point Γ , which divides it into
mean and extreme ratio, in other words, such that
|AB| |AΓ |
= ⇔ |AΓ |2 = |AB||Γ B|.
|AΓ | |BΓ |
Z
x
E
Δ
δ/2
A M Γ B
Figure 4.2.1: Golden section
Construction: Construct the right triangle ABE with |BE| = δ2 , where δ = |AB| (Figure 4.2.1). On the
hypotenuse EA consider the point ∆: |E∆| = |EB|. The circle κ with center A and radius A∆ intersects
AB at the wanted point Γ . This results by computing the power of A relative to κ.
|AB|2 = |A∆||AZ| ⇔ δ 2 = x(x + δ),
Exercise 4.10.91 Given is a triangle ABΓ . To find a fourth point ∆ so that the quadrilateral ABΓ∆ is
simultaneously inscriptible and circumscribable.
’ at corresponding
Exercise 4.10.92 Point P is projected on the sides and the bisector of the angle XOY
points {A, B, ∆}. Show that |M A| = |M B|.
Exercise 4.10.93 To find the geometric locus of points P whose projections on the sides of a fixed angle
’ at corresponding points {A, B} define lines AB of a fixed direction.
XOY
Exercise 4.10.94 Angle ω = XOY ’ of constant measure is turning about the center O of the circle κ(O) in-
tersecting it at points {A, B}. To find the geometric locus of intersection points P of the lines {AA , BB },
where A B is a fixed diameter of the circle.
Exercise 4.10.95 Points {A, B} move on parallel lines {α, β}, so that the segment AB has a constant
’
direction. From point P outside the band of the parallels we form the angle ω = AP B. To locate the place
of AB, for which this angle becomes maximal.
Exercise 4.10.96 Construct circles {α, β, γ} with diameters the medians {AA , BB , ΓΓ } of triangle
ABΓ . Show that these circles intersect by pairs at points {A 1 , A2 , B1 , B2 , Γ1 , Γ2 } contained in the alti-
tudes of the triangle, so that triangle A 1 A2 A3 is homothetic to ABΓ and triangle B 1 B2 B3 is homothetic
of the orthic of ABΓ .
μ
λ
δ
Ιμ κ
Δμ
Ιλ Δλ
Δκ
γ
Ικ
Ρμ
Ρκ Ρλ Γκ Γλ Γμ
A B
Figure 4.10.38: Generalization of Thales for pencils of circles
Next exercises discuss a generalization of the theorem of Thales for circles. Instead of a pencil of lines, we
have here a pencil of circles {κ, λ, µ, . . .} passing through two points {A, B}. On every line γ through A
there are defined the second intersection points {Γ κ , Γλ , Γµ , . . .}. The basic property is formulated in the
next exercise (Figure 4.10.38).
Exercise 4.10.97 With the previous assumptions and notations, the ratios of the lengths {|Γ κ Γλ |/|Γλ Γµ |}
does not depend on the direction of the line γ passing through A. In other words, for another line δ through
A will be defined corresponding points {∆ κ , ∆λ , ∆µ , . . .} and the corresponding ratios will be equal
|Γκ Γλ | |∆κ ∆λ |
= .
|Γλ Γµ | |∆λ ∆µ |
Hint: Project the centers {I κ , Iλ , Iµ , . . .} of the circles to the corresponding points {P κ , Pλ , Pµ , . . .} of line
δ and see that |∆κ ∆λ | = 2|Pκ Pλ | = 2|Iκ Iλ | cos(φ), where φ is the angle of the line of centers {I κ } with
the line δ.
Exercise 4.10.98 With the previous assumptions and notations, show that, for constant lines {γ, δ} passing
through A, the ratio |∆ κ ∆λ |
|Γκ Γλ | does not depend on the special member-circles {κ, λ} of the pencil and is the
same for all pairs of circles passing through {A, B}.
Α
Γκ Δκ
Δλ Δ
Γλ μ
Γμ
δ
Figure 4.10.39: Inverse of the generalization of Thales
Exercise 4.10.99 Show the inverse of the property of the exercise 4.10.97. If on two intersecting at A lines
{γ, δ} the corresponding points {Γ κ , Γλ , Γµ } and {∆κ , ∆λ , ∆µ } satisfy the relation |Γ κ Γλ | |∆κ ∆λ |
|Γλ Γµ | = |∆λ ∆µ | ,
then the circles {κ = (AΓκ ∆κ ), λ = (AΓλ ∆λ ), ν = (AΓν ∆ν )} pass through a fixed point B.
In chapter 11 we will see that the lines {Γ κ ∆κ , Γλ ∆λ , . . .} are tangents to a parabola, which has its focus
at B (Figure 4.10.39).
Α
Ε
κ κ'
B X
Γ
Y
Δ
I
Figure 4.10.40: Circle power curve
Figure 4.10.40 shows the two additional geometric loci of the vertices of a rectangle, which is constructed
using the power p κ (A) of point A relative to circle κ. For a variable secant XY through A, we take AE
equal to AX and orthogonal to it. There results the rectangle AY ∆E of constant area, equal to p κ (A).
Although the locus of E is a simple circle, the diagonal to the fixed A lying vertex ∆ of the rectangle
describes a quite complicated curve of 4th degree, which could be called circle power curve.
Exercise 4.10.100 Using the previous construction, show that the locus of E is a circle κ equal to κ.
CHAPTER 5
From the classical theorems
5.1 Escribed and excenters

As the whole of nature is akin, and the soul has learned everything, nothing
prevents a man, after recalling one thing only - a process men call learning -
discovering everything else for himself, if he is brave and does not tire of the
search, for searching and learning are, as a whole, recollection.
Plato, Meno 81 d
Escribed or tritangent circles of the triangle ABΓ are called the circles which are outside the triangle and
are tangent to its sides. Their centers are called excenters of the triangle of the triangle. The existence of
Α
ΙΒ
ΙΓ
Ι
Ε
Β Γ Δ
ΙΑ
Figure 5.1.1: The escribed (tritangent) circles of a triangle
these circles relies on next proposition, which is similar to the one for inscribed circles (Theorem 2.2.2).
312 CHAPTER 5. FROM THE CLASSICAL THEOREMS
Theorem 5.1.1 The internal bisector of one triangle angle and the external bisectors of the other two
angles pass through a common point.
Proof: We will show that the internal bisector at B and the external bisectors at A and Γ intersect at the
same point, which we denote by I B . Similar things will hold also for the other angles, which will define
the points IA and IΓ (Figure 5.1.1). The proof is transferred almost verbatim from Theorem 2.2.2. Let I B
be the intersection point of two out from the three bisectors of the triangle and specifically of the external
bisectors of the angles A and Γ . We will show that the internal bisector of angle B “ also passes through
IB . Indeed, according to Proposition 1.15.16, the distances of I B from the sides of angle A are equal
|IB E| = |IB ∆|. Similarly, also the distances of I B from the sides of angle Γ are equal |IB E| = |IB ∆|.
Consequently the three distances will all be equal |I B Z| = |IB E| = |IB ∆|, therefore they are radii of a
circle with center IB and radius rB = |IB ∆|. The equality of the distances |I B Z| = |IB ∆| from the sides
of angle B“ show that IB is to be found also on the bisector of angle B “ (Proposition 1.15.16). The fact that
the sides are orthogonal to these radii of this circle, shows (Corollary 2.2.2) that the circle is tangent to all
three sides of the triangle, q.e.d.
Exercise 5.1.1 Let {I, IA , IB , IΓ } be respectively the incenter and the excenters of the triangle ABΓ .
Show that:
1. Each of the triples (A, I, IA ), (B, I, IB ), (Γ , I, IΓ ) consists of collinear points.
2. The line defined by each of these triples is an altitude of the triangle I A IB IΓ .
3. Point I is the orthocenter of the previous triangle.
4. Triangle ABΓ is the orthic of triangle I A IB IΓ .
Proposition 5.1.1 The length of the tangent B∆, from the vertex B to the corresponding escribed circle
with center IB , is equal to half the perimeter τ of the triangle ABΓ
1
|B∆| = (a + b + c) = τ.
2
Proof: Here, as usual, with {a, b, c} we denote the lengths of the sides of the triangle. The proof follows
directly from the equality of the tangents from B: |BZ| = |B∆|, as well as from A and Γ : |AZ| =
|AE|, |Γ E| = |Γ∆| (Figure 5.1.1). The perimeter therefore is written as
a+b+c = (|BΓ | + |Γ E|) + (|AE| + |BA|)
= (|BΓ | + |Γ∆|) + (|BA| + |AZ|)
= 2(|BΓ | + |Γ∆|), q.e.d.
A Α
(I) (ΙΙ)
τ-a
Η Θ
B' Ε
Γ' τ-c
A'
B Γ Β Δ Ζ Γ
τ-b
Figure 5.1.2: Tangents from the vertices Tangent parallel to base
Proposition 5.1.2 The tangents A∆, AZ from the vertex A of triangle ABΓ to its inscribed circle have
length
|A∆| = |AZ| = τ − a,
where τ = 12 (a + b + c) is the half perimeter of the triangle.
5.1. ESCRIBED AND EXCENTERS 313
Proof: As the proof of the previous proposition, so this one is also relying on the equality of the tangents
from one point to a circle: |A∆| = |AZ|, |B∆| = |BH|, |Γ H| = |Γ Z| (Figure 5.1.2-I). It suffices
therefore to write the perimeter as
a + b + c = 2τ = 2(|A∆| + |BH| + |HΓ |) = 2(|A∆| + a),
from which the wanted equality follows directly, q.e.d.
Exercise 5.1.2 In the triangle ABΓ , with |AΓ | ≥ |AB|, the circle κ is tangent to the sides {AB, AΓ }
and passes through the point A of the base BΓ (Figure 5.1.2-I). Show that κ coincides with the inscribed
circle of the triangle ABΓ , if and only if, it holds
|A Γ | − |A B| = |AΓ | − |AB|.
Then point A coincides with the contact point of the circle with the base BΓ .
Hint: For {x = |A B|, y = |A Γ |}, the above relation is equivalent with {y − x = b − c, y + x = a}.
Exercise 5.1.3 Let E be the contact point of the tangent HΘ of the inscribed circle, which is parallel
to the base BΓ of the triangle ABΓ . Show that the line AE intersects BΓ at a point ∆, such that
|AB| + |B∆| = |∆Γ | + |Γ A| = τ . Conclude that |B∆| = |ZΓ |, where Z the contact point of the incircle
with BΓ (Figure 5.1.2-II).
Ρ ΙΒ
Α
ΙΓ
Κ Ζ
Δ Ι Ο
Λ Β Η Θ Γ Ξ
Ν
ΙΑ
Figure 5.1.3: Segments on the sides
Proposition 5.1.3 Next table gives the centers of the inscribed and escribed circles of triangle ABΓ as
well as their respective projections on the sides AB, BΓ and Γ A (Figure 5.1.3).
AB BΓ ΓA
I ∆ H Z
IA M Θ N
IB Π Ξ O
IΓ K Λ P
The following relations are valid:
τ −a = |A∆| = |AZ| = |ΓΞ | = |Γ O| = |BΛ| = |BK|

τ −b = |B∆| = |BH| = |ΓΘ | = |Γ N | = |AK| = |AP |
τ −c = |Γ Z| = |Γ H| = |AΠ | = |AZ| = |BM | = |BΘ |
|HΘ | = |c − b|, |OZ| = |a − c|, |K∆| = |b − a|.
Proof: That it holds τ − a = |A∆| = |AZ|, we saw in the previous proposition. For the other equalities on
the same line write
|ΓΞ | = |BΞ | − |BΓ | = τ − a.
Similarly follow also the equalities in the second and third line. Equality |HΘ | = |c − b| follows from the
previous
|HΘ | = |BΓ | − |BΘ | − |Γ H| = a − (τ − c) − (τ − c) = c − b.
Similarly follow also the two last equalities, q.e.d.
Exercise 5.1.4 Using the Figure: 5.1.3 show that

1. |ΛΘ | = |KM | = b, |ΘΞ | = |ON | = c, |OP | = |Π K| = a.
2. |∆M | = |ZN | = a, |HΞ | = |∆Π | = b, |ZP | = |HΛ| = c.
3. |P N | = a + c, |ΛΞ | = b + c, |M Π | = a + b.
Exercise 5.1.5 Given are two not congruent and external to each other circles κ(O) and λ(P ). Show that
the relations suggeste by the figure 5.1.4.
Ζ
Ε
Λ Δ λ
Κ ν μ ξ
Ν
κ Μ Ρ
Ο
Ξ
Η Γ
Ι Β Α
Θ
Figure 5.1.4: Common tangents of two circles
1. |KΛ| = |ΛN | = |∆E| = |EZ| etc.

2. |ΛE| = |N Γ | etc.
3. The circles ν = (Γ , ∆, N, H), µ = (B, E, Λ, Θ ), ξ = (A, Z, K, I) are concentric.
4. The circle µ passes through the centers O and P .
Exercise 5.1.6 Find a point M on the side AB of triangle ABΓ (Figure 5.1.5-I), such that
|M A| + |AΓ | = |M B| + |BΓ |.
5.1. ESCRIBED AND EXCENTERS 315
(Ι) ρΓ (ΙΙ)
Α Α
ΟΓ
c b
M M
a
Β Γ Β Γ
Figure 5.1.5: M bisects the perimeter (ABΓ ) = 12 (a + b − c) · ρΓ
Exercise 5.1.7 Show that for the radii {r, r A , rB , rΓ } of the inscribed and escribed circles, the altitudes
{hA , hB , hC } and the area = (ABΓ ) of the triangle ABΓ holds
1 1 1 1 1 1 1
= (τ − a)rA = (τ − b)rB = (τ − c)rΓ , + + = = + + .
rA rB rΓ r hA hB hC
1 1
Hint: For the first three equalities see 5.1.5-II. For the rest see that hA = a
2
, which implies that hA + h1B +
1 1
hC =
= r .
τ
Exercise 5.1.8 Show that, if the incenter of the triangle ABΓ coincides with its centroid or its orthocenter,
then the triangle is equilateral.
Α
(I) (II)
B
ΙΑ Ο
I
rA Β Γ
r Μ
A α Κ
Η
Δ Γ Ε
Θ Ι
a
Figure 5.1.6: A, r(r A ), a - triangle construction Triangle from traces H, Θ , I
’ , its side a = |BΓ | and the radius

Exercise 5.1.9 Construct a triangle ABΓ from its angle α = BAΓ
r(rA ) of the inscribed (Escribed) circle.
Hint: Let ∆ be the contact point of the inscribed circle I(r) with side AB (Figure 5.1.6-I). The right triangle
A∆I can be constructed, because we know the orthogonal I∆ of length r and the angle |∆AI| = α2 .
Extend A∆ by a until E. E is the contact point of AB with the escribed circle I A (rA ) (Exercise (2) 5.1.4).
Therefore this too can be constructed by drawing the orthogonal to AB at E and finding the intersection
point of IA with AI. Draw next the common tangent BΓ of the two circles (Construction 2.4.10). Triangle
ABΓ is the wanted one.
’ , its perimeter 2τ and

Exercise 5.1.10 Construct a triangle from its side a = |BΓ | or its angle α = BAΓ
its area E = ρ · τ .
Exercise 5.1.11 Construct a triangle ABΓ for which are given the radii of the inscribed r, escribed r A
and the difference of sides |b − c|.
Exercise 5.1.12 Construct a triangle ABΓ for which are given the positions of the traces on its circumcir-
cle of the altitude, bisector and median from vertex A.
Hint: If {H, Θ , I} are the respective traces (Figure 5.1.6-II), then these determine the circumcircle as well
as the diametrically opposite K of A, on this circle. Then the right triangle AHK can be constructed, and
from this the wanted ABΓ .
5.2 Heron’s formula

One ought, every day at least, to hear a little song, read a good poem, see a
fine picture, and, if it were possible, to speak a few reasonable words.
J. Goethe, Wilhelm Meister’s Apprenticeship, V, 1
The theorem (or formula) of Heron (approximately 10-75 A.D.) expresses the area of the triangle as a
function of the lengths of its sides. With the help of one factorization exercise it can result as an application
of the formula of Stewart (Theorem 3.12.1). In what follows I give this proof, as well as the more elegant
one which is attributed to Heron himself ([Dan55, p. 160]).
υα
Β Γ
Δ
2 1
Figure 5.2.1: υA = 4a2 (a + b + c)(b + c − a)(c + a − b)(a + b − c)
Theorem 5.2.1 (Heron’s formula) Let τ = 12 (a + b + c) be the half perimeter of triangle ABΓ . Then the
triangle’s area E is given by the formula
»
E = τ (τ − a)(τ − b)(τ − c).
Proof: We apply the formula for the altitude υ A2

= 4a12 (a + b + c)(b + c − a)(c + a − b)(a + b − c)
(Proposition 3.12.1), which holds even for obtuse triangles, provided we consider the altitude from its
obtuse angle (supposing it to be at A, such that ∆ falls into BΓ . Then the factors which show up in the
previous proposition are written as
(a + b + c) = 2τ
(b + c − a) = (b + c + a − 2a) = 2(τ − a)
(c + a − b) = (c + a + b − 2b) = 2(τ − b)
(a + b − c) = (a + b + c − 2c) = 2(τ − c)
1
Heron’s formula then follows from the known formula for area E = 2 υA · a, q.e.d.
The second proof of Heron’s formula uses the calculation of the radii of the inscribed and escribed cir-
cles of the triangle (§ 5.1).
Proposition 5.2.1 The radius r of the inscribed and r A of the escribed circle of the triangle ABΓ are given
respectively by the formulas
(τ − a)(τ − b)(τ − c) τ (τ − b)(τ − c)

r2 = 2
rA = .
τ τ −a
5.2. HERON’S FORMULA 317
Γ
τ-c
IA
τ-a Ζ Η
Ι
Θ rA
r
Α Δ Β Ε
τ-b
τ
Figure 5.2.2: r, r A as functions of the sides
Proof: The figure 5.2.2is familiar. We met it in § 5.1. Two circles are shown, the inscribed I(r) and the
escribed IA (ra ). The proof uses the relations of the aforementioned section and the similarity between two
pairs of triangles. The first pair of triangles is (A∆I, AEI A ). The second pair is (∆IB, EBI A ). Both
pairs consist of right triangles and we have:
|∆I| |A∆| r τ −a
= ⇔ = ,
|EIA | |AE| rA τ
|∆I| |EB| r τ −c
= ⇔ = .
|∆B| |EIA | τ −b rA
Solving the second relative to r A and substituting into the first expression, we find the formula for r 2 .
Squaring the first formula and substituting with the found expression for r 2 , we prove the second formula
as well, q.e.d. The second proof of Heron’s formula follows by substituting in the formula for the area
Α
κ
(Ι) (ΙΙ) Γ' Β'
A
Γ''
Β''
r r
b
c I
Β Α' Α'' Γ
r
B a Γ
Figure 5.2.3: Heron’s proof Symmetric inscribed hexagon
(ABΓ ) = r · τ (Corollary 3.3.3) the radius r through the formula of Proposition 5.2.1 (Figure 5.2.3-I).
Exercise 5.2.1 Let {B Γ , Γ A , A B } be tangents of the inscribed circle κ(r) of the triangle ABΓ ,
respectively parallel to the sides {BΓ , Γ A, AB} (Figure 5.2.3-II). Show that the hexagon A A B B Γ Γ
is symmetric and has its opposite sides equal and parallel. Show also that the sum of the inradii of the small
circles rA + rB + rC = r. Finally show that the circumcircles of the small triangles are tangent to the
circumcircle of ABΓ .
Exercise 5.2.2 Let R be the radius of the circumcircle of the triangle with side lengths {a, b, c}. Prove the
following formulas:
E = r · τ = rA · (τ − a) = rB · (τ − b) = rC · (τ − c) (1)
4R + r = rA + rB + rC , (2)
2 2 2
a b c
R2 = . (3)
2(b2 c2 + c2 a2 + a2 b2 ) − (a4 + b4 + c4 )
Hint: Use Heron’s formula and Corollary 3.11.2.
Exercise 5.2.3 Show that the area E of the triangle ABC is expressed with the help of its altitudes h A ,
hB , hC through the formula:
Å ã Å ã Å ã Å ã
1 1 1 1 1 1 1 1 1 1 1 1 1
= + + · − + + · − + · + − .
E2 hA hB hC hA hB hC hA hB hC hA hB hC
Comment In the articles of Baker [Bak85a], [Bak85b] are contained 110 formulas for the area E of the
triangle.
5.3 Euler’s circle

The world is a king, and, like a king, desires flattery in return for favor; but
true art is selfish and perverse - it will not submit to the mold of flattery.
Ludwig van Beethoven, Conversations, 1820
The Euler (1707-1783) circle of the triangle, is the one which passes through the three middles of the sides.
Its particularity lies in the fact that it also passes through six more noteworthy points of the triangle, that’s
why it is often called circle of nine points of the triangle.
ρ
E
Ρ
Ξ N
Z O
Π
H
Σ
T
B Δ M Γ
κ
Figure 5.3.1: Euler’s circle κ of the triangle ABΓ
Theorem 5.3.1 The circle κ, which passes through the middles M , N , Ξ of the sides of triangle ABΓ ,
has the following properties:
1. It passes also through the traces {∆, E, Z} of the altitudes of the triangle.
2. It passes also through the middles {P, Σ , T } of the line segments which join the vertices with the
orthocenter H of the triangle.
5.3. EULER’S CIRCLE 319
3. Its center Π is the middle of the segment which joins the orthocenter H with the circumcenter O of
the triangle.
4. Its radius |Π P | is half that of the radius ρ = |OA| of the circumscribed circle λ of the triangle.
5. Point H is a center of similarity of κ and the circumcircle of the triangle, with similarity ratio 1 : 2.
Proof: The proof relies on the existence of three rectangles which have, by two, a common diagonal. The
rectangles are ΣTNΞ , ΣMNP and PΞMT . First let us see that these rectangles exist. I show that Σ T N Ξ
is such a rectangle. The proof for rectangles ΣMNP and PΞMT is similar.
In ΣTNΞ then, ΣΞ joins the middles of sides of the triangle BHA. Therefore it is parallel and the half
of HA. Similarly T N joins the middles of sides of triangle AHΓ . Therefore it is parallel and the half of
HA. Consequently ΣΞ and T N , being parallel and equal, they define a parallelogram ΣTNΞ . That this
is actually a rectangle, follows from the fact that ΣT joins middles of sides of triangle HBΓ , therefore it
is parallel and the half of BΓ . Since A∆ and BΓ are mutually orthogonal, the same will happen also with
their parallels ΣΞ and ΣT .
The three rectangles have by two a common diagonal, which is the diameter of their circumscribed
circle. This implies that the three circumscribed circles of these rectangles coincide. This completes the
proof of the first two claims of the proposition.
For the proof of the next two claims, it suffices to observe that in the triangle HOA the segment ΠP
joins the middles of the sides of triangle HOA, therefore it is parallel and the half of OA. However OA
is a radius of the circumscribed circle λ and ΠP is a radius of circle κ. The last follows from the fact
|AH| = 2|OM | (Proposition 3.9.4), therefore P HM O is a parallelogram and its diagonals are bisected at
Π . However P M , as seen previously, is a diameter of the circle κ. The last claim is a consequence of the
two previous ones, q.e.d.
ΙΒ
Α''
Α
ΙΓ
Ο'
Ο
Ι
Β Γ
Α'
ΙΑ
Figure 5.3.2: Circumcircle of ABΓ as Euler circle of I A IB IΓ
Exercise 5.3.1 Show that the triangle ABΓ coincides with the orthic triangle of triangle I A IB IΓ , with
vertices the excenters of ABΓ . Conclude that the circumcircle of the triangle ABΓ coincides with the
Euler circle of IA IB IΓ (Figure 5.3.2).
Exercise 5.3.2 Construct a triangle ABΓ for which are given the position of the vertex A, the position of
the projection ∆ of A on the opposite side BΓ and the position of the center Π of its Euler circle.
internal
Exercise 5.3.3 The orthocenter H of the triangle ABΓ is projected on the bisectors of angle A,
and external, at the points Σ and P . Show that the line Σ P passes through the middle of BΓ and the
center of its Euler circle.
Hint: The first part of the exercise is Exercise 3.14.50.
κ Γ
(I) Ρ (II)
y
β z
Ο Γ'
Α' Β'
α x
Α
Α Β
Β υ
α
Γ
Figure 5.24.29: Concurring lines Triangle construction
Hint: On the sides {AB, AΓ } of the angle α we define using coordinates the homographic relation y =
(ux+v)/(rx+s) and locate the fixed point P through which pass all the lines {B(x)Γ (y} (Figure 5.24.29-
II). With center at A and radius υ we draw a circle κ and draw the tangent to it from P .
Exercise 5.24.51 To construct a triangle ABΓ , whose sides {x = |AB|, y = |AΓ |} satisfy the homo-
’ and the internal,
graphic relation y = (ux + v)/(rx + s) and of whose is given the angle α = BAΓ
or external bisector, or the inradius.
graphic relation y = (ux + v)/(rx + s) and of which are given the angles.
graphic relation of the form ux + vy = w and of which is given also the angle α = BAΓ’ and the
circumradius R or the median from A.
The homographic relation in the last exercise has not the general form of the previous exercises. If the
exercise was formulated using the general homographic relation y = (ux + v)/(rx + s), then the corre-
sponding problems would be not solvable with ruler and compass only. For example, if the last relation is
Γ'
ζ Ρ
x B
A ω ε
y z
Figure 5.24.30: Conchoid of Nicomedes
’ and the circumradius R, then the side z = |BΓ | is determined

valid and given are also the angle ω = BAΓ
z = 2R sin(ω) and the problem reduces to that of intersecting a known angle BAΓ ’ from P , by a secant
P BΓ through P , so that the segment BΓ has a given length. Point P is determined from the angle ω and
the homographic relation, as in the previous exercises. In the figure 5.24.30 we see the geometric locus
of points, which result from the fixed point P and the fixed line ε, taking on all lines through P segments
|BΓ | = |BΓ | = z. This is the known from ancient times conchoid of Nicomedes (about 200 b.C), which
is intimately related to the problem of neusis (or verging) (Exercise 5.24.28) and was used to solve the
problem√of trisection of an angle and the doubling of the volume of the cube, equivalently the construction
of the 3 2. The requested Γ is precisely an intersection point of the other side ζ of the angle ω and the
curve.
Α
α
Δ
Θ Ν Μ
Ο P γ
κ
Ε
ε'
β
ε Β
Figure 5.24.31: Geometric locus of the middles
Exercise 5.24.54 Let {α, β, γ} be three lines through the point O and P a point of γ. We consider all the
circles κ passing through {O, P } and intersecting a second time the lines {α, β} at corresponding points
{A, B}. Show that the middle M of the segment AB is contained in a fixed line ε and teh centroid N of the
triangle OAB is contained in a line ε , which is parallel to ε (Figure 5.24.31).
Hint: The triangle P AB has constant angles, its vertex P is fixed and points {A, B} move on fixed lines.
Apply theorem 2.14.3.
Exercise 5.24.55 In the figure 5.24.31 of the previous exercise, consider the medial line of OP and its
intersections {∆, E} respectively with the lines {α, β} and the center Θ of the circle κ. Show that P is a
Miquel point of {A, B, Θ } relative to the triangle O∆E and is also a pivot of ABΘ relative to O∆E.
Γ'
A
Β'
B Γ
p(K)
Α'
Figure 5.24.32: Symmedian point K and the polar p(K)
Exercise 5.24.56 In the figure 5.24.32 K is the symmedian point of triangle ABΓ , and A B Γ is the
tangential triangle of ABΓ . The line p(K) is the polar of K with respect to the circumcircle of ABΓ ,
coinciding with the trilinear polar of K with respect to ABΓ and with respect to A B Γ . Study the
various relations suggested by this figure. Find all harmonic quadruples of points contained in the lines of
this figure.
Exercise 5.24.57 Construct the pedals of the orthocenter, the centroid and the circumcenter of the triangle
ABΓ and determine other inscribed triangles A B Γ of ABΓ , which have these points as Miquel points.
Exercise 5.24.58 Construct a right triangle for which are given the length of the hypotenuse and the length
of the bisector of one of its acute angles.
Exercise 5.24.59 Construct square ABΓ∆ for which is given the center O and two points Z, H on the
sides, respectively, Γ∆ and ∆A, with |ZO| = |HO|.
Exercise 5.24.60 The vertices of polygon A 1 A2 ...Aν are projected on a line ε at points B 1 , B2 , ..., Bν .
Show that for the signed lengths holds
B1 B2 + B2 B3 + ... + Bν−1 Bν + Bν B1 = 0.
Exercise 5.24.61 Let P be a point on the circumcircle κ(O, R) of the triangle ABΓ and {A , B , Γ }
be the projections respectively on lines {BΓ , Γ A, AB, s P }, where sP the Simson line of P . Prove the
formulas ([Lal52, p. 14]):
1. |P A||P A | = |P B||P B | = |P Γ ||P Γ | = 2R|P P |.
2. |P A||P B||P Γ | = 4R2 |P P |.
3. |P A ||P B ||P Γ | = 2R|P P |2 .
A
B''
A1
Γ''
H
B' 90°+γ-β
Γ'
Π
Ρ K Γ1
Λ M N
B1
Γ
B A'
A''
Figure 5.24.33: Six collinear points
Exercise 5.24.62 Of the triangle ABΓ the middles of its sides are respetively A , B , Γ , the traces of its
altitudes are A , B , Γ , the orthocenter is H and A 1 , B1 , Γ1 are the middles of HA, HB, HΓ . Show
that the next 6 points, which are defined as intersections of lines, are contained in a line parallel to BΓ :
P = (AB, A B1 ), K = (BB , A Γ ), L = (A Γ , B B1 ), N = (ΓΓ , B A ), M = (A B , Γ Γ1 )
and Π = (A Γ1 , AΓ ). Show also that point M is the center of the circle with diameter Π P and that the
quadrilaterals A KHN , A P AΠ are inscriptible and their circumcircles are tangent to the Euler circle
at point A .
Hint: Begin with the inscriptible quadrilaterals, for which the equality of the angles, suggested by the figure
5.24.33, is crucial.
5.24.62 is a typical exercise sample, which results with the help of the computer. The specific one comes
from a program, which starts from a given shape in which a few points have bin singled out, called first gen-
eration points (in this case they are 13: {A, B, Γ , A , B , Γ , A , B , Γ , A1 , B1 , Γ1 , H}). The program
calculates all the different lines defined by these points, the so called first generation lines (specifically 699).
Subsequently, it determines points of second generation, which are the intersections of the first generation
lines (specifically 16470). Then, the program investigates how many lines contain more than two points. In
the specific case there are 5286 such lines which carry 3, 4, 5, 6 and 7 points of second generation. Each
such line defines an exercise, like the previous one, where it is requested to prove the collinearity of certain
points. In the specific case most of the 5286 lines carry 3 second generation points. The more collinears
on a line we require, the less lines we find. This way, in the specific example, from the 5286 lines only 38
carry six points of a second generation (and consequently define 38 exercises similar to the previous one),
while there exist also 9, which carry 7 points each. In the specific example this is the maximum number of
collinears. In other words, each of the 5286 lines contains no more that 7 second generation points.
Obviously the process could be continued ad infinitum, defining similarly lines of second generation,
respectively points of third generation and so on and so forth, examining again the collinearities, finding the
maximum number of collinear points in each generation etc. The actual program, for reasons of exhaustion
of memory, stops at the collinearities of second generation points, but examines the corresponding process
by substituting circles in the place of lines. These are defined from all possible triples of non-collinear
points, the so called first generation circles. There the corresponding investigation concerns 4 or more con-
cyclic points and in the specific example, among other things, determines the Euler circle, which contains 9
points. The following exercise resulted from a corresponding investigation for the same basic figure of 13
points of the previous exercise. The exercise shows that the aforementioned process produces problems of
non trivial geometric content. However, for reasons of consistency to historical evolution, I do not include
in the book more problems produced this way. It is obvious that, by beginning with a small number of
points from one shape, we can produce an infinity of problems. However, these problems do not seem to
possess, generally speaking, an importance similar to that of the classical problems, which were developed
in a timely fashion and are useful as tools for the solution of other problems.
Α
Ρ
Β''
Γ'' Z
Λ Γ' Β'
Κ Ε
Β Γ
Α'' Α'
Figure 5.24.34: Polar AZ of point Λ
Exercise 5.24.63 The middles of sides of the triangle ABΓ are respectively A , B , Γ and the traces of its
altitudes are A , B , Γ . Let Z be the intersection point of the lines B Γ , Γ B , and K be the second
intersection point of the circles (B B Z), (Γ Γ Z). Show that the lines B Γ , B Γ , EK, where E is
the center of the Euler circle, pass through the same point Λ.
Hint: Let Λ = (Γ B , Γ B ) be the pole of the line AZ relative to the Euler circle of the triangle ABΓ
(Proposition 5.17.1). We easily see, that the triangles AΓ B , AB Γ are similar isosceli (Figure 5.24.34).
The intersection point P = (B Γ , AZ) is the harmonic conjugate P = Λ(Γ , B ). Also the angles
Γ÷ ◊
KP , P KB are equal, because of the inscriptible quadrilaterals ZKΓ Γ , ZKB B . It follows, that
ΛK is the external bisector of triangle KB Γ , therefore it is orthogonal to AZ, which is the polar of Λ.
The orthogonal from the pole Λ to the polar AZ passes through the center E. Note that line EZ is the Euler
line of the triangle ABΓ (Exercise 5.19.5).
CHAPTER 6
Circle measurement
6.1 The difficulties, the limit

Upon the whole, I am inclined to think that the far greater part, if not
all, of those difficulties which have hitherto amused philosophers, and
blocked up the way to knowledge, are entirely owing to our selves.
That we have first raised a dust and then complain we cannot see.
G. Berkeley, Concerning the Principles of Human Knowledge
The problem with the circle is that, before measuring it, one has to prove that it has length (we often call it
perimeter). This appears difficult to understand by the novice, it is however a problem which results from
the fact that we don’t have, up to now, a definition for the length of a curve. Currently, we know how to
measure length only for line segments and its derivatives which are broken lines and polygons. The circle
however is something different. We have to clarify things, to define what exactly we mean by length of
a circle and to prove that our definition is meaningful and doesn’t contradict our axioms. After all these
are done, we can subsequently proceed in the determination of the length of the circle. In the subject of
the existence of length we need to resort to an axiom of the real numbers (called axiom of completeness)
which has to do with sequences and limits. This axiom says:
Axiom 6.1.1 Every increasing sequence of real numbers α 1 , α2 , α3 , ... which is also bounded has limit a
real number A.
The word sequence (of numbers) means a set of numbers, each one of which is characterized by an integer
number (index):
α1 , α2 , α2 , ... αν , ....
Each one of these numbers is called member of the sequence. The word increasing means that these
numbers are increasing with each step:
α1 ≤ α2 ≤ α3 ≤ α4 ....
The fact that the sequence is bounded means that there exists a specific number M , such that all the sequence
numbers are less than M . We could write briefly:
α1 ≤ α2 ≤ α3 ≤ α4 ... ≤ M.
442 CHAPTER 6. CIRCLE MEASUREMENT
M is called an (upper) bound of the sequence. A number M > M is also such a bound, and usually the
exact value of M is immaterial. It suffices to find some M , which satisfies the previous inequalities. For
example, the numbers
1 2 3 4 5
0 < < < < < < ...
2 3 4 5 6
define an increasing sequence, which is also bounded, since all numbers are less than M = 1. Notice that
the terms of this sequence are given by the formula
ν−1
αν = ,
ν
by substituting in it the values 1, 2, 3, 4, 5, ... etc.
The last word which requires explanation, in this short formulation of the axiom, is the limit. I will
not give here its general definition rather only its special, in other words, that of the limit of an increasing
sequence. The limit of an increasing sequence, then, is a number A for which are satisfied two inequalities
A − ε < αν ≤ A,
the right one for all α ν , the left, for any ε > 0 we choose, is satisfied for all α ν , with the exception of some,
which are nevertheless finite in number (and depend on the size of ε). In the previous example say, we can
easily see that the limit is A = 1. Indeed the previous inequality becomes
ν −1
1−ε< ≤ 1.
ν
The right inequality holds true for all ν =1, 2, 3, 4, ... The left inequality can be written
1
1−ε<1− ,
ν
and is equivalent to
1 1
< ε which in turn is equivalent to < ν.
ν ε
The last however is satisfied exactly as required by the definition of the limit, that is for all ν with the
exception of some, whose number is finite (and depends on the magnitude of ε). This way for example, for
ε = 0.001 therefore 1ε = 1000 the inequality is satisfied for all ν > 1000 and is not satisfied for ν < 1000,
which are however finite in number. In practice, to show that A is the limit of the increasing sequence we
proceed as follows:
1. We find the (candidate) limit A,
2. We show that αν ≤ A holds for all ν,
3. We consider ε > 0 and we “solve” the inequality A − ε < α ν for ν,
4. We verify that the α ν which are not solutions to the previous inequality are finite in number.
The first difficult step is to find or guess A. The second step usually follows from the first. On the third step
we must, like we do with equations, manage to bring to the one side of the inequality the ε and to the other
the ν and to show that the inequality is equivalent to one of the form
N (ε) < ν.
N (ε) is usually a number which depends on ε and having (3) in this form, means automatically: (a) that (3)
is satisfied for all ν which are greater than N (ε) and (b) the inequality is not satisfied for all ν which are
less than N (ε), which are nevertheless finite in number.
Comment We need the notion of limit, because through it are defined the length of the circle and its
area. As we will see below, beginning with the square we construct regular polygons Π 1 , Π2 , Π3 , ... each
of which has double the number of sides than the previous and are all inscribed in the same circle κ. The
perimeter of the circle is then defined as the limit of the perimeters of these polygons. Similarly the area
also is defined as the limit of the areas of these polygons.
6.1. THE DIFFICULTIES, THE LIMIT 443
Often, in the lessons of Geometry, the discussion on the difficulties which are involved with the circle,
when we try to measure its perimeter and its area, are omitted. I think however that this is not correct. At
this point we are faced with a difficulty, whose resolution is one of the major cultural achievements. In the
investigation of circles appears the need for approximations, for limits, and the foundations of calculus are
invented (from Archimedes), which will be developed much later. It is a pity for the student to not have a
small idea about these notions. Even if one has problems in understanding fully the notions, it is useful for
one’s education to feel, some more some less, that a difficulty leads to its transcendence, which opens new
fields and new horizons. The connection between the three notions “difficulty”, “limit” and “transcendence”
is noteworthy.
Exercise 6.1.1 Show that if the increasing sequence α 1 ≤ α2 ≤ α3 ... has as limit the number A, then if
A < A, then there exists an αi with
A < αi < A.
Hint: Write A = A − ε with ε = A − A > 0. Next apply the definition of the limit of an increasing
sequence, according to which, there are infinitely many α i which satisfy A − ε < αi < A. Choose one
of them.
Exercise 6.1.2 Show that if the increasing sequence α 1 ≤ α2 ≤ α3 ... has as limit the number A, then the
sequence ρα1 ≤ ρα2 ≤ ρα3 ..., where ρ > 0, has as limit the number ρA.
Hint: The right inequality of the definition: ρα i ≤ ρA is a consequence of the corresponding α i ≤ A
which holds by assumption. The left inequality of the definition of limit: ρA − ε < ρα i is equivalent to
A − ρε < αi . However, by assumption the last one holds for all α i except for finitely many, consequently
the same will happen with the equivalent to the previous one. Finally then, for every ε > 0 the inequalities
ρA − ε < ραi ≤ ρA will hold for all the α i with the exception of some, something which shows the truth
of the claim.
Exercise 6.1.3 Show that the sequence

3ν + 2
αν =
5ν + 10
3
is increasing. Show further that the limit of the sequence is A = 5
Hint: The fact that it is increasing is equivalent to

3ν + 2 3(ν + 1) + 2
< ,
5ν + 10 5(ν + 1) + 10
which is proved easily. For the fact that the limit is the specific number we show first that for each ν =1, 2,
3, ... holds
3ν + 2 3
< .
5ν + 10 5
Next we “solve”
A − ε < αν ⇔ A − αν < ε ⇔
3 3ν + 2 4 1
− < ε ⇔ · < ε ⇔
5 5ν + 10 5 ν +2
4 1 4 1
· < ν +2 ⇔ · −2 < ν
5 ε 5 ε
The last holds for all ν except finitely many, the number of which depends on the magnitude of ε.
Exercise 6.1.4 Consider the sequence whose general term is given by the formula
αν + β
αν = ,
γν + δ
where α, β, γ and δ are constants. Show that this sequence is increasing, if and only if αδ − βγ > 0. Also
show that when it is increasing, then the sequence has limit αγ .
Besides sequences of numbers, we will also consider next sequences of polygons. As with numbers so with
polygons, the word sequence means a set of polygons each one of which is characterized by an integer:
Π1 , Π2 , Π2 , ... Πν , ....
Each one of the polygons is a term of the sequence. It is fairly obvious that the notion of the sequence and
its terms can be transferred to sets of similar objects. This way we could define in some way sequences of
points, sequences of squares, sequences of circles etc. Figure 6.1.1-I, for example, shows a few terms of a
Δ Γ
(Ι) (ΙΙ) (ΙΙΙ)
τ2
τ3
τ1
Α Β
Figure 6.1.1: Triangles sequence Squares sequence Hexagons sequence
sequence of triangles τ 1 , τ2 , τ3 , ... each defined by the middles of the sides of the previous one, beginning
with a specific triangle τ1 . Figure 6.1.1-III, shows a sequence of polygons. The sequence is constructed
using always the same orientation and taking on each side a point dividing the side into the ratio κ. Joining
the points on the various sides results in a new polygon, which is inscribed in the previous one. Repeating
this process we create a sequence. The previous sequence of triangles, is a special case of that more general
sequence.
Exercise 6.1.5 In the square τ 1 = ABΓ∆ we inscribe the square τ2 = A B Γ ∆ , the vertices of which
A A B B
divide the sides {AB, BΓ , Γ∆, ∆A} in the ratio k < 0 : k = A B = B Γ = ΓΓ ∆
Γ
= ∆ ∆
∆ A . The
same procedure is repeated with τ 2 and continuing this way we define the sequence {τ n } (Figure 6.1.1-II).
Compute the area of the square τ n , as well as the distance εn of its vertices from the center of τ1 . Show that
εn , for appropriate n, becomes as small as we wish.
The archetype of a sequence is certainly the one of natural numbers N = {1, 2, 3, . . . }. One of the fun-
damental properties of this sequence is formulated by the so called principle of mathematical induction
([CR96, p. 9]), according to which, if we have a sequence of propositions {A 1 , A2 , . . .}, each depending on
the associated index, then to prove the validity of all these propositions it suffices:
1. To prove the validity of A 1 .
2. With the assumption that A k is true, for an arbitrary k, to prove that A k+1 is true.
A typical example of such a sequence of propositions is the so called Bernoulli inequality
An : (1 + x)n ≥ 1 + nx, valid for x > −1 and all integers n ≥ 1. (*)
The proof results by applying the principle of induction:

(1) For n = 1, A1 : (1 + x)1 = (1 + x), is true.
(2) For n = k, assuming A k : (1 + x)k ≥ (1 + kx) is true, multiply it by (1 + x) > 0, resulting in
(1 + x)k+1 ≥ (1 + kx)(1 + x) = 1 + (k + 1)x + kx2 > 1 + (k + 1)x, hence Ak+1 is also true.
A typical application of this inequality is in the study of the so called geometric progression, which is
the sequence of numbers {s n = 1 + x + · · · + xn }, depending on x. Using induction we can prove that

1 − xn+1
sn − 1 = |x|
n+1
sn = 1 + x + · · · + xn = ⇒ , for x > −1. (**)
1−x 1−x 1−x
1 1 1 ∗
Numbers 0 < x < 1 can be written x = 1+θ with θ > 0. Then, xn = (1+θ) n ≤ 1+nθ , because of ( ).
∗∗
This, in view of ( ) implies easily, that, for 0 < x < 1, the geometric progression converges to the limit
1/(1 − x).
6.2. DEFINITION OF THE PERIMETER OF THE CIRCLE 445
Exercise 6.1.6 Supply the details of the proof, that the geometric progression for 0 < x < 1, converges to
1/(1 − x).
6.2 Definition of the perimeter of the circle

The second, to divide each of the difficulties I will examine
into as many parts as possible and needed to solve it better.
Descartes, Discourse on the Method, II
In this section we consider a circle of radius ρ and some regular polygons Π 1 , Π2 , Π3 , ... inscribed in it, each
of which has as vertices those of the predecessor plus the middles of the arcs defined from the successive
vertices of the predecessor. We begin then with the square Π 1 . If ABΓ∆ labels the square, then the middles
B B
(Ι) (ΙΙ) φ/2
Z E Z E
M
ρ φ
Γ O A Γ A
O ρ
H Θ Θ
H
Δ Δ
1
Figure 6.2.1: Square and octagon |AE|: µ of the perimeter
of the arcs AB, BΓ , Γ∆, ∆A define four additional points E, Z, H, Θ respectively and AEBZΓH∆Θ is
a regular octagon: Π 2 . Π3 will result similarly and will have 16 sides, Π 4 will have 32 and, in general, Π ν
will have 2ν+1 sides.
Lemma 6.2.1 The regular polygon with µ sides, inscribed in the circle of radius ρ has perimeter
Å ã
180◦
2µρ sin .
µ
Proof: Because all the sides of a regular polygon are equal, its perimeter will be µ · |AE| where AE is one
of its sides. If O is the center of the circle, then triangle AOE is isosceles with angle at O equal to
360◦ φ 180◦
φ= ⇒ =
µ 2 µ
since the sum of µ such equal angles will give a full turn about point O. Also the median AM will
be orthogonal at the middle M of AE (Corollary 1.8.2) and the triangle OM E will be right, therefore
according to Theorem 3.6.1 Å ã
|AE| φ
= |M E| = ρ · sin .
2 2
Combining the previous relations gives the wanted, q.e.d.
Lemma 6.2.2 The sequence of perimeters p 1 , p2 , p3 , ... of the polygons Π 1 , Π2 , Π3 , ... is increasing.
Proof: Because each such polygon has sides double those of its previous, it suffices to show that the ratio
of the perimeters pp of two regular polygons Π and Π from which the first has µ sides and the second has
2µ sides is less than 1. However according to the previous, taking into account the formula for the sine of
the double angle (Exercise 3.6.5), we have:
2µρ sin( 180

◦
sin( 180
◦
2 sin( 180 180 ◦ Å ã ◦
p µ ) µ ) 2µ ) · cos( 2µ ) 180◦
= ◦ = ◦ = ◦ = cos < 1, q.e.d.
p 2(2µ)ρ sin( 180
2µ ) 2 sin( 180
2µ ) 2 sin( 180
2µ )
2µ
Lemma 6.2.3 Every regular polygon with µ sides inscribed in a circle of radius ρ has perimeter less than
that of the circumscribed square on the same circle.
Ρ Δ' Γ' Ξ
Γ B'
(Ι) Δ (ΙΙ)
Y'
E' B Z
E O Y
A'
A
Z' Z I
O
H Θ I' X X'
M H' Θ' N
Figure 6.2.2: Projection on the square Comparison of triangle and its projection
Proof: The proof results by projecting the regular polygon from its center O onto a circumscribed square.
To this, for each of the vertices A, B, Γ , ... of the inscribed polyhon we draw the corresponding half line
OA, OB, OΓ , ... and we define its intersection point with the circumscribed square as A , B , Γ ,... This
defines a polygon A B Γ ... which has its vertices on the square and each of its sides is greater than the
corresponding side of the initial polygon. This results from the comparison of the corresponding triangles
(AOB, A OB ), (BOΓ , B OΓ ), ... For each such pair of triangles, let me call one (XOY, X OY ), point
X is on the extension of OX and Y is on the extension of OY , consequently X Y is greater than XY .
For the proof of the last claim, suppose that from the distances XX and Y Y , XX is the less one and
draw the parallel X Z to XY from X . It obviously holds
|XY | ≤ |X Z| < |X Y |.
|XY | |OX|
The first inequality holds because of similarity, |X Z| = |OX | ≤ 1. The second inequality holds because

the angle X ZY is obtuse (Corollary 1.10.3), therefore opposite to it lies the greater side of the triangle
X ZY . It follows then that the perimeter of A B Γ ... is greater than that of ABΓ ... but also less than the
perimeter of the square. The latter because the sides of A B Γ ... are either parts of the sides of the square
(like A B in the figure) or they have their parts in different sides of the square (like B Γ in the figure) and
they are hypotenuses of right triangles, which are less than the sum of the two orthogonal sides. Concluding
then, the perimeters p, p and p of the regular polygon, of its projection to the square and of the square,
satisfy:
p < p < p , q.e.d.
Lemmas 6.2.2 and 6.2.3 guarantee, relying on Axiom 6.1.1, that the sequence α 1 , α2 , ... of the perimeters
of the polygons Π 1 , Π2 , ... has a limit A. This A we define as length or perimeter of the circle.
Comment-1 The numbers α i , which are mentioned above as perimeters of the Π i , result from Lemma
6.2.1 by setting to it
µ = 2i+1 .
6.3. THE NUMBER π 447
This way the perimeter of Π i is Å ã

180◦
αi = 2(2i+1 ) · ρ · sin .
2i+1
According to the definition of the limit, for all ε > 0 it follows, that for the perimeter of the circle A and
the above numbers α i hold the inequalities
A − ε < αi < A,
the right one for all α i and the left for all the α i except finitely many. How many α i don’t satisfy the left
inequality depends on the magnitude of ε. Their number however is always finite.
Comment-2 The definition of the perimeter of the circle we gave is connected to the definition of length of
a broken line and it can be proved that these two notions of length (broken and circle) can result as special
cases of a more unified definition of curve length. Unfortunately or fortunately this definition belongs to the
area of calculus and goes beyond the purposes of this lesson. An idea about the details which are involved
is given in the next exercises.
Exercise 6.2.1 Given is a circle κ. Show that every inscribed in κ regular polygon has perimeter less than
that of the circle κ.
Hint: Here the problem is that we consider a general regular polygon and not the special regular polygons
Πi with 2i+1 sides, for which we showed that their perimeters α i make an increasing sequence with limit
the perimeter of the circle A and α i < A. The proof in the general case follows from this special case and
Exercise 2.13.9, according to which the perimeter of the regular polygon with ν + 1 sides is greater than
this of the regular polygon with ν sides. This way, for given ν we find a power of 2 for which ν ≤ 2 i+1 ,
something which is always possible. Then however the perimeter p of the regular polygon with ν sides will
be less than the corresponding perimeter α i of Πi , which in turn is less than A.
Exercise 6.2.2 Show that the sequence of perimeters β 1 , β2 , β3 , ... of the regular polygons with ν = 3, 4,
5, ... sides and inscribed in the circle κ is increasing and bounded consequently has limit B. Show also
that B = A, where A is the limit of the perimeters α1 , α2 , α3 , ... of the regular polygons Π 1 , Π2 , Π3 , ...
inscribed in κ.
Hint: The fact, that the sequence β 1 , β2 , β3 , ... of the perimeters of all the polygons is increasing, is
proved in Exercise 2.13.9. The fact, that this sequence is bounded, is proved exactly as in Lemma 6.2.3.
The existence of the limit B is then a consequence of Axiom 6.1.1. The fact, that B = A, is proved by
excluding the cases B < A and A < B. Indeed, if we assume that B < A, then, according to the definition
of the limit, there will exist α i , such that B < αi < A. Then, however, and for each polygon with number
of sides ν > 2i+1 we will have corresponding perimeter α i < βν , which implies that B < αi < βν . The
last thought is contradictory, because for all the β ν holds βν < B. Similarly we prove that A < B leads to
a contradiction.
6.3 The number π

My method to overcome a difficulty is to go round it.
George Polya, How to solve it
Lemma 6.3.1 For every regular polygon with µ sides the ratio of its perimeter to the diameter of its cir-
cumcircle is independent of its size, it depends only on µ and is given by the formula:
Å ã
180◦
µ · sin .
µ
1/(2q2)
0/1 1/3 1/2 3/5 2/3 3/4 4/5 5/6 1/1

p/q
Figure 6.9.4: Ford circles
Exercise 6.9.10 The Ford circles (1886-1967) are tangent to a line ε on the same side. They are con-
structed by selecting a system of coordinates on ε and considering the points with coordinates the irre-
ducible fractions pq of the interval [0, 1]. To every such point a circle κ p,q is tangent (we often say “the
circle pq ”) of radius 2q12 (Figure 6.9.4). Show that
1. Two Ford circles are tangent, if and only if d 2 = 4r ·r , where d is the distance of the points of contact
and {r, r } are the radii of the circles.
2. The previous equality for the circles {κ p,q , κp ,q } is equivalent to p · q − q · p = −1.
3. Two Ford circles are either disjoint or they are tangent.

4. If the circles which correspond to the fractions { pq , pq } are tangent, then the circle which corresponds
p p p+p
to q , is tangent to two others, if and only if q = q+q .
0/1 1/3 1/2 3/5 2/3 3/4 4/5 5/6 1/1
Figure 6.9.5: Collinear, concyclic points on Ford circles
Exercise 6.9.11 How many and which relations exactly may one notice in figure 6.9.5 ? For example, that
the circle κ is orthogonal to the circles which correspond to the fractions { 10 , 13 , 12 }. That the common
tangent ε of the circles { 32 , 11 } passes through the contact point 5

6 and its diametrically opposite 1
2 relative
to the circle 12 .
The sequence of coordinates of the points of contact of the Ford circles with the line is called sequence
of Haros - Farey. Its terms are ordered in a natural way in the so called Haros-Farey series of order N ,
which are subsets {F1 , F2 , ...} of the sequence and are constructed, each one from the previous through the
operation, which in two fractions assigns the so called middle term (which is different from the geometric

middle of the interval [ pq , pq ]).
p p p + p
⊕ = .
q q q + q
The sets FN consist of the irreducible fractions 0 ≤ p
q ≤ 1 with q ≤ N .
ß ™
0 1
F1 = , ,
1 1
ß ™
0 1 1
F2 = , , ,
1 2 1
ß ™
0 1 1 2 1
F3 = , , , , ,
1 3 2 3 1
ß ™
0 1 1 1 2 3 1
F4 = , , , , , , , ...
1 4 3 2 3 4 1
The next exercise ([RG95, p. 114]), which presents the basic properties of the Haros - Farey sequence, is
for those who are attracted to the theory of numbers.

p p+p
Exercise 6.9.12 1. For two elements pq , pq of FN , p
q ⊕ q = q+q is irreducible.
¶ ©
p p p p p
2. If q , q , q , are consecutive terms of FN , then q = p
q ⊕ q .
3. Each FN contains the previous F N −1 , and results from FN −1 by taking some of its middles.
4. Each term of the Haros - Farey sequence is contained in some F N .
5. For every N the number of elements |F N | of FN satisfies the equation |F N | = |FN −1 |+φ(N ), where
φ(N ) represents the number of integers 1 ≤ x ≤ N , which share no common factor with N ([RG95,
p. 133]).
(I) Γ (II)
h/2
Α Β A B
Figure 6.9.6: Broken line of constant length Curves of constant length
Exercise 6.9.13 Show that for every line segment AB and every ε > 0, there is a broken line from A to B,
the length of whose is as big as we please and whose all of its points X satisfy |X − X | < ε, where X the
projection of X on AB.
Hint: Start with a triangle ABΓ and construct two similars to it on the two halfs of its base AB (Figure
6.9.6). Repeat this procedure with the new triangles etc. At each stage, joining the sides of the resulting
triangles, except the bases, results to a broken line of length |AΓ | + |Γ B|. After ν steps the distance of the
points X of this broken line from AB is less than h/2 ν−1 , where h is the altitude of the initial triangle.
Exercise 6.9.14 Show that for every line segment AB and every ε > 0, there is a curve with end points
{A, B}, consisting of successive semicircles, having constant length π|AB| and such that its points X
satisfy |X − X | < ε, where X the projection of X on AB.
Hint: The construction is similar to the one of the previous exercise (Figure 6.9.6-II). See also the exercise
6.4.3.
Γ Β''
(Ι) (ΙΙ) Γ
Ε1
Β'
Ε2
Δ1
Ε Α' Δ'
Δ2 Δ
O
Α''
Δ κ
Α Ε3 Δ3 Β Α Γ' Β Γ''
Figure 6.9.7: Pedal triangle and isogonal points, Pedal triangle of inverse point
Exercise 6.9.15 Let point ∆ be outside the side-lines of triangle ABΓ and ∆ 1 ∆2 ∆2 be the pedal triangle
of ∆, whose vertices are the projections of ∆ on the sides of the triangle (Figure 6.9.7-I). Compute the area
of the pedal triangle ∆ 1 ∆3 ∆3 , the area of its circumcircle κ and show that the orthogonals to the sides, at
their second intersection point with κ, pass through a point E, which is the isogonal conjugate of ∆.
Exercise 6.9.16 Let ∆ be a point not lying on the side-lines of the triangle ABΓ . Show that the pedal
triangle t = A B Γ of ∆ and the pedal triangle t = A B Γ of the inverse ∆ of ∆ relative to the
circumcircle of ABΓ are similar triangles (Figure 6.9.7-II). Compute the ratio of the areas of the triangles
{t, t }, as well as the ratio of areas of their circumcircles.
One of the best known sequences of integer numbers is the Fibonacci sequence (1175-1250)
{1, 1, 2, 3, 5, 8, 13, 21, 34, 55, 89, 144, 233, 377, 610, 987, 1597, 2584...}.
Its terms {ai } are determined by the relations:
a1 = a2 = 1,
an = an−1 + an−2 , for n > 2.
This sequence appears in various problems, practical
√
and theoretical ([PL07], [Wal01]). We encounter it
5+1
also in the study of the golden section φ = 2 , when we try to express the powers of φ in terms of φ
using the basic equation of its definition: φ 2 = φ + 1 (§ 4.2):
φ2 = φ + 1
3
φ = φ2 + φ = 2 · φ + 1
φ4 = 2φ2 + φ = 3 · φ + 2
φ5 = 3φ2 + 2φ = 5 · φ + 3
φ6 = 5φ2 + 3φ = 8 · φ + 5
φ7 = 8φ2 + 5φ = 13 · φ + 8
... = ...
φn = an · φ + an−1 .
√
1 5−1 2
A similar relation holds true for the inverse of φ, x = φ = 2 , which satisfies the equation x +x−1 = 0,
from which it results analogously that
(−x)n = an · (−x) + an−1 .
Combining the two relations, we obtain the Binet formula (1786-1856):

√
φn − (−x)n = an · (φ + x) = an · 5 ⇒
ÇÇ √ ån Ç √ ån å
1 1 1+ 5 1− 5
an = √ (φ − (−x) ) = √
n n
− .
5 5 2 2
The figure 6.9.8 shows a sequence of rectangles, which start with a square of side-length 1, and whose
.....
Figure 6.9.8: Approximation of the golden rectangle
side-lengths are expressed with two consecutive Fibonacci numbers. Selecting the rectangles of odd order,
i.e. those having their short side horizontal, we obtain the sequence of the ratios of their side-lengths:
a2n
bn = , n = 1, 2, . . . .
a2n+1
It is proved (Exercise 6.9.19), that this sequence is bounded and increasing, and its limit is the number
x = φ1 . Consequently, the ratios of these side-lengths approximate the ratio of the sides of the golden
rectangle. The corresponding rectangles, while becomming bigger and bigger, simultaneously tend to be
similar to the golden rectangle.
Exercise 6.9.17 Show that for every n > 1 holds true the formula a n+1 an−1 = a2n + (−1)n .
√
Hint: Write an = k(φn − (−x)n ), where k = 1/ 5 and observe that φ m · (−x)m = (−1)m , for every
m ≥ 1. Then,
an+1 an−1 − a2n = k 2 ((φn+1 − (−x)n+1 )(φn−1 − (−x)n−1 ) − (φn − (−x)n )2 )

= −k 2 φn−1 (−x)n−1 (φ − x)2 = (−1)n .
Exercise 6.9.18 Show that the sequence {b n } is bounded and increasing, hence converges to a limit.
Hint: The sequence is bounded by 1. To see that it is increasing, observe that

a2n+2 a2n a2n+2 a2n+1 − a2n+3 a2n
bn+1 − bn = − =
a2n+3 a2n+1 a2n+1 a2n+3
a2n+2 a2n+1 − (a2n+2 + a2n+1 )a2n a2n+1 (a2n+2 − a2n ) − (−1)2n+1
= = > 0.
a2n+1 a2n+3 a2n+1 a2n+3
Exercise 6.9.19 Show that the sequence {b n } converges to x = 1/φ.
Hint: Since x < 1 ⇒ xn < 1, we have:

(−x)n an−1 an−1 1
(−x)n = an (−x) + an−1 ⇔ = −x ⇒ − x < .
an an an an
Exercise 6.9.20 Show that the numbers x which satisfy an equation of the form x 2 − nx + 1 = 0, where n
a positive integer, can be expanded into an infinite continued fraction of the form x = [n; n, n, n, . . . ].
Hint: x = n + x1 , . . .
√
Exercise 6.9.21 Show that the numbers of the form x = n2 + 1, where n is a positive integer, can be
expanded in continued fraction of the form x = [n; 2n, 2n, 2n, . . . ].
√ √ √ √ √
Hint: n2 + 1 − n = ( n2 + 1 − n)( n2 + 1 + n)/( n2 + 1 + n) = 1/( n2 + 1 + n). This implies
1 1
x=n+ ⇒ x=n+ 1 etc.
x+n 2n + x+n
Exercise 6.9.22 Determine the irrational number x, whose expansion into continued fraction is the peri-
odic one x = [2; 5, 1, 3, 5, 1, 3, 5, 1, 3 . . .] = [2; 5, 1, 3].
Hint: ([Moo64, p. 46]) Consider the periodic part of the fraction
1 1 3y + 1
y =5+ = 5+ = 5+ ⇔ 2y 2 − 11y − 3 = 0,
1 1+ y
3y+1 4y + 1
1+
1
3+
y
and use the relation x = 2 + y1 .
3x+2
Exercise 6.9.23 Show that the positive number x, which satisfies x = 13x+9 , has the expansion in contin-
ued fraction x = [0; 4, 2, 1]. Construct and study examples of expansions of numbers y, which satisfy an
equation of the form y = ay+b
cy+d , where {a, b, c, d} positive integers satisfying ad − bc = 1.
Exercise 6.9.24 Given the positive integers {a, b}. Determine the form of numbers x whose expansion in
continued fraction has the form x = [0; a, b].
Knowing the decimal representation of an irrational number up to to a certain precision, we can find its
expansion in a continued fraction up to a certain order. The number π for example, has an infinite and, not
a kind of regularity showing, expansion in a continued fraction ([Per57, p. 18, I]) whose first 40 terms are
[3; 7, 15, 1, 292, 1, 1, 1, 2, 1, 3, 1, 14, 2, 1, 1, 2, 2, 2, 2, 1, 84, 2, 1, 1, 15, 3, 13, 1, 4, 2, 6, 6, 99, 1, 2, 2, 6, 3, 5].
Β
Δ
1/2
Ε 7/8
Γ Ζ Η Α 1
Figure 6.9.9: Approximation of π
355 42
Figure 6.9.9 shows a geometric construction of the approximation π ≈ [3; 7, 15, 1] = 113 = 3+ 72 +82
= 3.141592... due to Jacob de Gelder (1764-1848).
Exercise 6.9.25 Prove the construction of the segment |Γ Z| = 4 2 /(72 + 82 ) suggested by the figure, in
which |AB| = 1, |A∆| = 7/8, |Γ E| = 1/2, E is projected on AΓ to H and EZ is parallel to BH.
CHAPTER 7
Transformations of the plane
7.1 Transformations, isometries

To properly know the truth is to be in the truth; it is to have the truth for
one’s life. This always costs a struggle. Any other kind of knowledge is a
falsification. In short, the truth, if it is really there, is a being, a life.
Kierkegaard, Truth is the way
Transformation of the plane is called a process f , which assigns to every point X of the plane, with a
possible exception of some special points, another point Y of the plane which we denote by f (X). We
write Y = f (X) and we call X a preimage of the transformation and Y the image of X through the
transformation. We often say that the transformation f maps X to Y . For the process f we accept that it
satisfies the requirement
X = X ⇒ f (X) = f (X ).
In other words, different points also have different images. Equivalently, this means that, if for two points X,
X holds f (X) = f (X ), then it will also hold X = X . The set of points X, on which the transformation f
is defined, is called domain of the transformation f , while the set consisting of all Y , such that Y = f (X),
when X varies in the domain of f , is called range of f . For transformation examples the reader may consult
the beginnings of the following sections. Here we limit ourselves in a description of common characteristics
of these notions.
For every shape of the plane Σ the set of images f (X), where X runs through Σ , is called image of Σ
and is denoted as f (Σ ).
Applying the processes one after the other, we create the notion of composition of transformations.
For two given transformations f and g, we call composition of f and g, the transformation whose process
results by the successive application of the processes of f and g. The composition of transformations is
denoted by
g ◦ f.
By definition, the process of composition g ◦ f first corresponds Y = f (X) to X and then Z = g(Y ) =
g(f (X)) to Y . Totally then, it corresponds Z = g(f (X)) to X. There are some details, which we must be
careful with in compositions. These have to do with the domains and ranges of the transformations, which
participate in the composition. For all of it to be meaningful, the range of the first transformation (f ) must
be contained in the domain of the second (g). Things are considerably simplified for transformations which
have domain and range the entire plane.
474 CHAPTER 7. TRANSFORMATIONS OF THE PLANE
Since the composition g ◦ f is a new transformation, we may consider its composition with a third transfor-
mation h :
h ◦ (g ◦ f ),
and more generally we can define the composition of as many transformations f 1 , f2 , f3 ,..., fk we want,
which, for simplicity, let us consider that they are defined on the entire plane:
fk ◦ fk−1 ◦ ... ◦ f1 .
The meaning of such a composition of transformations is that we apply successively the processes of the
transformations which participate in sequence from right to left. f 1 maps point X1 to X2 = f1 (X1 ), f2
next maps X2 to X3 = f2 (X2 ), and so on and so forth. This process can be denoted pictorially by the
diagram
f1 f2 f3 fk
X1 −→ X2 −→ X3 −→ X4 ... −→ Xk+1 .
A very simple and insignificant, regarding its action, transformation is the so called identity transforma-
tion, which we denote with e and which, to every point X corresponds X itself. This one resembles the
unit in the familiar multiplication, which leaves numbers unchanged. The same way, this transformation
does nothing. It leaves every point fixed. Its structural meaning however is as important as that of the mul-
tiplication unit. With its help we can define immediately the inverse transformation of a transformation f
which we denote with f −1 . This one performs exactly the opposite process to that of f and by definition
holds
f −1 ◦ f = e.
If we confine ourselves to transformations f , g, h, ..., defined for all points on the plane, then their totality
together with composition, presents a noteworthy similarity with the set of positive numbers and multipli-
cation. I list the similarities (and one difference) in two parallel columns:
Numbers Transformations
z = x · y (product) h = g ◦ f (composition)
x · y = y · x (commutativity) g ◦ f = f ◦ g (in general)
x · (y · z) = (x · y) · z (associativity) h ◦ (g ◦ f ) = (h ◦ g) ◦ f
1 (unit) e (identity transformation)
y = x−1 ⇔ y · x = 1 (inverse) g = f −1 ⇔ g ◦ f = e
We will apply these rules in the next sections. I underline here the associative property h◦(g◦f ) = (h◦g)◦f ,
which holds for transformations. This is due to their very nature as a correspondence process, which re-
mains the same, any way we choose to group them (insert parentheses) in a particular composition of more
than one transformations.
A special category of transformations, we will deal with in the next sections, is that of isometries or
congruences of the plane . With this naming we mean transformations, which are defined on the entire
plane and additionally have the property of preserving distances ([Cox61, p. 39], [Sin95]). In other words,
transformations X = f (X), such that, for every pair of points X, Y and their images X , Y will hold
|X Y | = |XY |.
Theorem 7.1.1 An isometry preserves angles.

Proof: The short wording means that for three points X, Y , Z and their images X , Y , Z through the
transformation the angles Y’ XZ and Y◊ X Z are equal. This however is a consequence of the property of
the isometry of the transformation, on the basis of which |X Y | = |XY |, |Y Z | = |Y Z|, |Z X | = |ZX|.
In other words the triangles X Y Z and XY Z are congruent. From the congruence of triangles follows
the equality of the angles as well, q.e.d.
Exercise 7.1.1 Show that the composition g ◦ f of two isometries f and g is again an isometry. Show also
that the inverse transformation f −1 of an isometry is an isometry.
7.2. REFLECTIONS OR MIRRORINGS 475
Theorem 7.1.2 An isometry maps a line ε onto a line ε . If the isometry fixes two points A and B of ε, then
it also fixes all the points of ε.
|XA|
Proof: The position of a point X of the line AB is completely determined by the ratio |XB| and the fact
that ||XA| ± |XB|| = |AB|. The latter gives the necessary and sufficient condition so that X is on the line.
If therefore the isometry f fixes points A, B then for the images X , A , B will hold:
|X A | |XA|
= , and |X A | ± |X B | = |XA| ± |XB|.
|X B | |XB|
Consequently if X is contained in line ε = AB, then also X will be contained in line ε = A B . The
second part follows immediately from the previous equalities, if we take into account that A = A, B = B.
Then from these follows that for every point X of ε point X coincides with X, q.e.d.
Theorem 7.1.3 An isometry, which fixes three non collinear points, coincides with the identity transforma-
tion.
Proof: Assume that the isometry f fixes the points A, B and Γ . Then, according to the previous theorem, it
also fixes the lines AB and AΓ . If X is a point not lying on these lines, we draw a line ε through X, which
intersects AB and AΓ respectively at ∆ and E, which are fixed by f . By the previous theorem f fixes all
the points of ∆E therefore also X, q.e.d.
Corollary 7.1.1 Two isometries coincident at three points are coincident everywhere.
Proof: Indeed, if f , g are the two isometries, then g −1 ◦ f will fix the three points, therefore it will coincide
with the identity transformation g −1 ◦ f = e ⇔ f = g, q.e.d.
Exercise 7.1.2 Show that an isometry f maps a circle κ onto a circle κ = f (κ) of equal radius.
Proposition 7.1.1 If an isometry f satisfies the relation f ◦ f = e, then it fixes at least one point.
Proof: Obviously the identity transformation has the property of the psoposition. Let us suppose then that
f is not coincident with the identity transformation and let us denote with X the point f (X), so that,
according to the hypothesis X = X. We consider now an arbitrary point X such that X = f (X) = X.
Such a point exists, for otherwise f would be the identity. Suppose M is the middle of XX . We show that
f fixes M . By hypothesis f exchanges X and X , therefore maps the line XX to itself (Theorem 7.1.2).
Also
|X M | = |XM | = |X M | = |X M | = |XM |.
The first equality holds because f is an isometry. The second because M is the middle of XX , the third
because we have |XM | = |X M |. This means that point M lies on the medial line of XX , but also, as
we noted, it is a point of the line XX , therefore it coincides with M , q.e.d.
7.2 Reflections or mirrorings

The thing that doesn’t fit is the thing that’s the most interesting,
the part that doesn’t go according to what you expected.
R. Feynman, The Pleasure of Finding Things Out
A line of the plane ε defines a simple transformation called reflection or mirroring relative to the line ε,
which is called axis or mirror of the reflection . The process for this transformation is described as follows:
a) To every point X not contained in the line ε corresponds the point Y , such that ε is the medial line of
ε Μ
Figure 7.2.1: Mirroring or reflection relative to ε
XY . In other words, point X is projected orthogonally to ε at its point M and XM is extended to its
double towards M , until Y .
b) To every point X contained in the line ε the process corresponds the point X itself. In this case then,
point X is, as we say, a fixed point of the transformation.
Thus, the reflection is well defined for every point of the plane or, in other words, its domain is the entire
plane. The same happens also with its range. It also coincides with the entire plane, since for every Y there
exists one X such that f (X) = Y . The line ε, through which a reflection is defined, consists of all the fixed
points of the reflection. Every point X not contained in ε is in correspondence with Y which is on the other
side of ε than that where X is to be found. A reflection then interchanges the two sides of ε and leaves the
points of ε fixed.
The reflection underlies the notion of axial symmetry which we examined in § 1.16: The shape Σ is
axially symmetric, if there exists a reflection f , such that f (Σ ) = Σ . The next theorem is proved in exactly
the same way as the corresponding Theorem 1.16.1.
Theorem 7.2.1 Every reflection is a plane isometry.
Γ
X'
Χ A
B
ε
Β'
Υ Α'
Y'
Γ'
’ =B
Figure 7.2.2: |XX | = |Y Y | and BAΓ ◊
A Γ
Proposition 7.2.1 For every reflection f holds f ◦ f = e, in other words the inverse of a reflection is the
same transformation of the reflection.
Proof: Indeed, if Y = f (X), then point Y is the symmetric of X relative to the line ε, which defines the
reflection. Then, however, point X is also the symmetric of Y relative to ε, consequently X = f (Y ) hence,
for every X will hold f (Y ) = f (f (X) = X, which means that the composition f ◦ f coincides with the
identity transformation e, q.e.d.
The important characteristic of reflections is that, as we say, they generate all the isometries of the plane.
In other words, every isometry of the plane may be written as a composition of reflections. With a little
more work (Exercise 7.4.9) we will prove later the next theorem ([Cox61, p. 46]):
7.2. REFLECTIONS OR MIRRORINGS 477
Theorem 7.2.2 Every isometry of the plane is either a reflection or a composition of two or three reflec-
tions.
Closely connected with reflections is also the other simple transformation we met, the point symmetry. A
point O of the plane defines the transformation f of the point symmetry relative to O (section 1.16). This
one for every X = 0 corresponds point X = f (X) to X, which is the symmetric of X relative to O. In
other words the point X for which O is the middle of segment XX .
X O
X'
Figure 7.2.3: Transformation of point symmetry relative to O
Theorem 7.2.3 The composition g ◦ f of two reflections whose axes intersect orthogonally at O coincides
with the point symmetry relative to O.
β
Χ
Ο
α
Υ Ζ
Figure 7.2.4: Point symmetry as composition of two reflections
Proof: The proof, suggested by the figure, is the same as that of Theorem 1.16.2. If Y is symmetric relative
to line α and Z is symmetric of Y relative to line β, which intersects line α orthogonally at O, then Z is
also symmetric of X relative to point O, q.e.d.
Theorem 7.2.4 If an isometry f fixes two points A and B, then it also fixes all the points of the line AB
and coincides with either the identity transformation or the reflection relative to the line AB.
Proof: The first part of the theorem follows from Theorem 7.1.2. For the second, assume that f = e. It
suffices to consider one point X off the line AB and see what is the point X = f (X). Triangles ABX
A Y B
X'
Figure 7.2.5: Isometry with two fixed points
and ABX will be congruent and we see easily that they either coincide or they will be symmetric relative
to AB. The first is excluded by assumption, therefore point X will be the symmetric of X relative to AB,
q.e.d.
Proposition 7.2.2 If an isometry of the plane f , different from the identity, satisfies the relation f ◦ f = e
and has exactly one fixed point O, then it is coincident with the point symmetry relative to O.
Proof: The proof is contained in Proposition 7.1.1. We showed there that for every X with X = f (X) = X
the middle M of the segment XX is fixed. Then if there exists exactly one fixed point O, then all XX
will have the same middle O, which is the essense of point symmetry, q.e.d.
Corollary 7.2.1 An isometry of the plane f , different from the identity, which satisfies the relation f ◦f = e
is coincident with a symmetry relative to point O or with a reflection relative to line ε.
Comment-1 This section repeats, in essence, the material related to symmetry of shapes, which we studied
in § 1.16. It brings, however, in the foreground the abstract notion of transformation, with which we can
describe more general shape symmetries. Indeed, for a given transformation f , the shape Σ of the plane is
called symmetric relative to f , when f (Σ ) = Σ . The shapes Σ which are symmetric relative to a point O
are precisely those which satisfy f (Σ ) = Σ , where f is the symmetry transformation relative to O. The
shapes Σ which are symmetric relative to an axis ε are precisely those which satisfy f (Σ ) = Σ , where f
is the reflection relative to the line ε.
Comment-2 Transformations which coincide with their inverse (f −1 = f ⇔ f ◦ f = e) are called

involutions and play an important role in Euclidean, as well as in other Geometries ([VY10, p. 102 (I)]).
Exercise 7.2.1 Show that, for two different lines of the plane ε and ε , there always exists a reflection f
which transforms one to the other (f (ε) = ε ). In fact, depending on the position of the lines there exist
exactly two or exactly one reflection which has this property. When exactly do these cases happen?
Exercise 7.2.2 Show that for two different line segments AB, Γ∆ of the same length, sometimes there exists
a reflection which transforms one to the other f (AB) = Γ∆ and sometimes there doesn’t. When exactly
does either case happen?
Exercise 7.2.3 Given two different lines (equal segments, congruent circles) examine when there exists a
point symmetry which transforms one to the other.
Α ε Α'
Β Β'
Γ Γ'
Figure 7.2.6: Reflections reverse orientation
Reflections are closely connected to the reversal of orientation of triangles and more generally of polygons,
which we noted in § 1.7. Every reflection f maps a triangle ABΓ onto a triangle A B Γ , which has
the opposite orientation (Figure 7.2.6). The transformations we consider in this book (are, as it is said,
Α
Γ' Β'
O
Β Γ
Α'
Figure 7.2.7: Point symmetries of the plane preserve orientation
“continuous” and) have the property that, if they preserve the orientation of a triangle, then they preserve
(Ι) A' (ΙΙ)

ωΔ Γ
c'
a Δ'
E
d Γ' Δ
Β A'
Β Δ B' Δ2
ωΒ ωΔ
Δ1
b c
Β1
Β2
Γ Α
Figure 7.11.10: 4 reflections composition 4 rotations composition
“ Because of
reflection on ∆B and f 4 is the reflection on the line which forms with B∆ an angle equal to ∆.
“ “
the hypothesis B + ∆ = π, the reflection-axes of {f 1 , f4 } are parallel and the composition
f = (f4 ◦ f3 ) ◦ (f2 ◦ f1 ) = (f4 ◦ f3 ) ◦ (f2 ◦ f1 ) = f4 ◦ f1 ,
is the translation by the double of the distance of these parallels. The rest follows from simple calculations.
Noticable is the symmetry of the expression in terms of the lengths of sides, which, by Ptolemy’s theorem,
can be represented also through the product of the diagonals.
Exercise 7.11.14 Let f = f 4 ◦ f3 ◦ f2 ◦ f1 be the composition of the rotations about the vertices, corre-
spondingly, {A, B, Γ , ∆} of the quadrilateral ABΓ∆ by the respective positively oriented angles of the
quadrilateral. Show that f is a translation by the oriented segment ∆ 2 B2 of the side ∆A. This segment is
the double of the projection ∆ 1 B1 on ∆A of the diagonal of a quadrilateral A B Γ ∆ (Figure 7.11.10-II).
This quadrilateral is cyclic and is formed by the intersection points of the inner bisectors of the angles of
the given quadrilateral ABΓ∆.
Hint: The rotation f 1 about A can be represented as a composition of two reflections f 1 = h1 ◦ g1 . g1 is the
h1 is the reflection in the side AB. Similarly, the rotation
refelection with respect to the bisector AA of A.
f2 about B can be represented as a composition f 2 = h2 ◦ g2 , where g2 = h1 is, again, the reflection on the
“ Then the composition of the rotations
side AB and h2 is the reflection on the bisector BΓ of B.
f2 ◦ f1 = (h2 ◦ g2 ) ◦ (h1 ◦ g1 ) = h2 ◦ g1 ,
is the composition of the two reflections in the two consecutive sides {A ∆ , ∆ Γ } of the quadrilateral
A B Γ ∆ . A similar argument shows that the composition of the rotations about Γ and ∆ coincides
with the composition of the reflections in B Γ and Γ ∆ . Apply the previous exercise to the inscriptible
quadrilateral A B Γ ∆ .
Exercise 7.11.15 Show that a convex quadrilateral ABΓ∆ is circumscriptible if and only if, the composi-
tion f = f4 ◦ f3 ◦ f2 ◦ f1 of the rotations by the positive oriented angles of the quadrilateral, represents the
identity transformation.
Exercise 7.11.16 (Hjelmslev’s theorem) Prove that for a line α and its isometric image β = f (α) by an
isometry f of the plane, the middpoints of all the segments XX , where X = f (X), are either collinear
on the same line γ or coincident to a point O ([Cox61, p. 47]).
Hint: From theorem 7.4.2 follows that there are two types of isometries between two lines: rotations and
translations preserving the orientation, and reflections and glidereflections reversing it. Examine in each
case what happens with the middles of XX . In the case of rotation apply exercise 3.9.9. Show that for
glidereflections all these middles are on its axis.
Α
(I) (II)
A
p15
p8
B
(III)
Figure 7.11.11: Tilings with tiles of type p n
Exercise 7.11.17 The tile of type p n is defined by a regular n−gon q. For even n we consider the symmetric
q of q relative to the second greater diagonal AB and delete from q the common with q part. What remains
is the tile pn , like the p8 in figure 7.11.11-I. For an odd n we do the same procedure relative to a maximal
diagonal AB, from which results p n , like the p15 in figure 7.11.11-II. Show that these non-convex tiles can
tile the plane in concentric circular rings, like the p 8 in 7.11.11-I and the p 15 in figure 7.11.11-II. Compute
the number of tiles contained in each ring.
CHAPTER 8
Lines and planes in space
8.1 Axioms for space

The eye of understanding is like the eye of the sense; for as you may
see great objects through small crannies or levels, so you may see great
axioms of nature through small and contemptible instances.
F. Bacon, Sylva Sylvarum
For the study of shapes of the space we need the abstract notion of the plane and some additional axioms
to these we met in the previous chapters. Axioms which describe the basic properties of planes, as well as,
properties of lines in space and their relation to planes.
Axiom 8.1.1 Three non-collinear points define precisely one plane. Every plane ε contains infinitely many
points and, additionally, there also exist infinitely many points not contained in the plane ε.
Axiom 8.1.2 If a line has two common points with a plane ε, then it is contained in the plane ε.
Axiom 8.1.3 If two planes have a point in common, then they also have an entire line in common.
Axiom 8.1.4 Every plane ε separates the space into two parts, which have no points in common and are
called half spaces. A line segment, which has its end points in different half spaces intersects the plane at
exactly one point. A line segment, which has both its end points in the same half space, does not intersect
the plane.
Planes play the same role in space as lines on the plane. In plane geometry we isolate one of the planes of
space and we examine the shapes which are contained in that plane. In stereometry or solid geometry we
examine shapes, like for example, the cube and the sphere, which extend in space and can not be confined
only on some plane.
Theorem 8.1.1 If two different planes intersect, then their intersection coincides with a line.
Proof: According to the last axiom their intersection includes at least a line AB. It cannot include however
points not lying on the line AB. This because two points on the line, for example, points A, B and a point
X not lying on it, define exactly one plane. Therefore if X belonged to both planes then they, according to
the first axiom, would have to be coincident, contrary to the hypothesis, q.e.d.
532 CHAPTER 8. LINES AND PLANES IN SPACE
B
X
A
Figure 8.1.1: Intersecting planes
Theorem 8.1.2 Two different intersecting lines define one and only plane which contains them both.
Proof: Consider the point of intersection A of two lines α and β. Also consider two points X and Y , the
first belonging to α and the second to β (Figure 8.1.2). According to the first axiom this defines exactly one
α O
β Y
ε X
Figure 8.1.2: Plane of two intersecting lines
plane ε which contains the three points A, X and Y . According to the second axiom the plane will contain
both lines, q.e.d.
Proposition 8.1.1 A line ε and a point X not lying on it define exactly one plane which contains them.
ε Β
α Α Χ
Figure 8.1.3: Plane of line and point
Proof: Consider two points A and B of ε. The plane α, which contains the three points A, B and X also
contains the entire line ε (Axiom 8.1.2). Each plane which contains ε and X will contain also points A, B
and X, therefore it will coincide with α (Axiom 8.1.1), q.e.d.
Proposition 8.1.2 If two lines ε and ζ are contained in a plane, then this plane is unique.
Proof: The same as the previous, considering two points A, B of ε and a point X of ζ etc, q.e.d.
Parallel planes are called two planes which do not intersect. A line is called parallel to a plane when
it does not intersect it. A plane is called parallel to a line when it does not intersect it. Two lines are called
β
α
Figure 8.1.4: Parallel planes
parallel (in space) when (1) they are contained in a plane α and (2) they are parallel lines in that plane.
Two lines α and β not contained in a plane are called non coplanar or skew lines.
I clarify first a subtle point, which has to do with two parallel lines in space. The definition of parallels
in space and on the plane has a subtle difference. While on the plane two parallels are defined as two non-
intersecting lines, in space this property is insufficient, as can be seen with two skew lines, which, while
8.1. AXIOMS FOR SPACE 533
Figure 8.1.5: Non coplanar or “skew” lines
it is true that they do not intersect, they are not however parallel. Parallels in space are, by assumption,
contained in the same plane and their definition, as parallels in space, reduces to the definition of them
being parallel on their containing plane.
Theorem 8.1.3 From a point A not lying on the line ε, one and only line ζ parallel to ε can be drawn.
Α ζ
α ε
Figure 8.1.6: Unique parallel postulate in space
Proof: On the plane α, which is defined by the line ε and the point A, consider the unique parallel ζ to ε
(Axiom of parallel 1.15.1). Any other plane, which contains ε and point A will coincide with α, therefore
the parallel from A to ε will be unique, q.e.d.
Corollary 8.1.1 For every point X not lying on the line ε, the plane which contains point X and line ε also
contains the parallel ζ to ε from X.
Α Β
ε
α Χ ζ
Figure 8.1.7: Plane of two parallel lines
Proof: Consider the unique plane α which contains ε and X. In this plane the parallel ζ to ε from X can
be defined. Every plane which contains ε and X will coincide with α and the parallel of ε from X will
coincide with ζ, q.e.d.
Proposition 8.1.3 If a line ε is not parallel to a plane α and is not contained in it, then it intersects it at
exactly one point.
X
α
Figure 8.1.8: Line intersecting plane
Proof: Obviously ε must intersect α. Otherwise it would be parallel to it. Further, if it intersected it at more
than one points then, according to the second axiom, it would have to be entirely contained in α, contrary
to the assumption. Therefore ε intersects α at exactly one point, q.e.d.
β
ε ζ
η
α
Figure 8.1.9: Parallel lines intersecting plane
Proposition 8.1.4 If the line ε intersects plane α at a point, then also every parallel ζ to it will intersect
the plane at one point.
Proof: Consider the plane β which contains the parallels ε and ζ. It will intersect plane α along a line η. On
plane β lines ε and ζ are parallel and ε intersects η. Therefore line ζ also intersects η at a point (Corollary
1.15.1), which also happens to be a point of α, q.e.d.
Construction 8.1.1 From a point A not lying on the plane α, draw a parallel line ζ to the plane α.
Α ζ
β
ε
α
Figure 8.1.10: Construction of a line parallel to a plane
Construction: Consider an arbitrary line ε of the plane α and define the plane β which contains the line ε
and the point A. On this plane draw the parallel ζ of ε from A. ζ cannot intersect α, because if it did, then
the common point Σ of ζ with α would also be a point of ε, which is contradictory, q.e.c.
Proposition 8.1.5 If the tplane β passes through line XY and intersects plane α, relative to which line
XY is parallel, then the intersection ε of α and β is a line parallel to XY .
β
Υ
Χ ε
α
Figure 8.1.11: Plane through a line intersecting a parallel plane
Proof: If line ε is not parallel but intersected XY at a point Σ , then Σ would also be on XY and on α,
therefore XY would not be parallel to α, contrary to the hypothesis. This contradiction shows that ε is
parallel to XY , q.e.d.
Theorem 8.1.4 A line XY is parallel to a plane α, if and only if α contains a line ε parallel to XY .
Proof: If line XY is parallel to the plane α, consider a point Z of α and the plane β, which contains line
XY and point Z. According to the previous proposition, plane β would intersect α along a line parallel
to XY . Conversely, suppose that line XY is parallel to line ε of the plane α and suppose that line XY
intersects the plane α at point Σ . According to the definition, the lines XY and ε are contained in plane
β and line ε coincides with the intersection of α and β. Also point Σ will be contained in the intersection
8.1. AXIOMS FOR SPACE 535
Χ Υ
β
ε
α Z
Figure 8.1.12: Line parallel to plane
of α and β, therefore it will be a point of ε. This however is contradictory, because line ε was supposed
parallel to XY . Therefore line XY cannot intersect α, q.e.d.
Proposition 8.1.6 Two different intersecting planes, which are both parallel to line ε, intersect along a line
ζ parallel to ε.
β
Χ
ε α
ζ
Figure 8.1.13: Intersection of planes parallel to line
Proof: Assume that the planes α and β are parallel to the line ε and intersect along line ζ. Let then X be
a point of ζ. According to Proposition 8.1.5 plane γ, which is defined from ε and point X, will intersect α
and β along lines parallel to ε and pass through point X. This implies that these two lines will coincide and
will also coincide with the intersection of α and β, q.e.d.
Exercise 8.1.1 Assume that the lines α and β are different and parallel as well as parallels to plane ε.
Show that the plane of α and β either is parallel to plane ε or intersects it along a line γ parallel to α and
β.
Exercise 8.1.2 Assume that planes α and β pass through the line ε, which is parallel to plane ζ. Also
assume that α and β intersect the plane ζ. Show that their intersections are two parallel lines of ζ.
ε
α
Figure 8.1.14: Planes α through line ε and their intersection with ζ
Exercise 8.1.3 Consider all the planes α, which pass through the line ε. Let also ζ be a plane not contain-
ing ε and not parallel to it (Figure 8.1.14). Show that the lines which are defined as intersections of the
planes α and ζ all pass through a fixed point of ζ. How is this point defined from the givens?
8.2 Parallel planes

In other endeavors it comes after they are completed. In philosophy, on
the contrary, delight and knowledge are combined, because pleasure
doesn’t follow learning, but they happen simultaneously.
Epicurus, Address
Proposition 8.2.1 If two planes α and β are parallel, then every line of one is parallel to the other plane.
ε
β
Figure 8.2.1: Line parallel to plane
Proof: If the line ε of α intersected β at X, then α and β would have X as a common point, contrary to the
assumption. Therefore line ε cannot intersect β, q.e.d.
Proposition 8.2.2 The plane β of two intersecting lines, which are parallel to plane α, is parallel to α.
ε A
Ω
ζ
X Υ
β
Figure 8.2.2: Plane of two parallels to plane
Proof: Assume that the plane β of the intersecting lines XY and XΩ intersects the plane α along line η.
We will show that this is contradictory. To this, consider a point A outside both planes and draw from A
the parallel ε to XY and the parallel ζ to XΩ . Because ε is parallel to XY it intersects neither β nor α
(Proposition 8.1.4), (Theorem 8.1.4). Therefore, according to Proposition 8.1.6 the intersection η of α and
β will be parallel to ε. Similarly we also show for ζ, that it will be parallel to η. From point A then we will
β
α
Figure 8.2.3: Plane intersecting parallel planes
have two parallels to ζ, which is contradictory and shows that α and β do not intersect, q.e.d.
8.2. PARALLEL PLANES 537
Proposition 8.2.3 A plane γ intersecting two parallel planes α and β, intersects them along parallel lines
(Figure 8.2.3).
Proof: The two lines-intersections of γ with α and β cannot intersect, because then the planes would, too.
Also they are contained in plane γ, therefore, by definition, they are parallel, q.e.d.
Theorem 8.2.1 From a point X, not lying on the plane α, one and only one plane β parallel to α can be
laid.
γ Ω
X Υ
β
Z H
α
Figure 8.2.4: Plane from point, parallel to an other plane
Proof: From Proposition 8.2.2 follows that there exists a plane β through X parallel to α. This plane is
unique. Indeed, if there was also a second plane β through X then β and β would intersect along a line
η passing through X, which without loss of generality we may assume different from XY (Figure 8.2.4).
Consider then a point Z on α and consider the plane γ which contains points X, Y and Z. This intersects
plane β along XY , β along a line XY (not drawn) and α along line ZH. Because the planes are parallel,
the lines XY and XY will be two different parallels from point X to line ZH of plane γ (Proposition
8.2.3). This is contradictory and shows the uniqueness of the plane β parallel to α from X, q.e.d.
Corollary 8.2.1 A plane γ, which intersects plane β, will also intersect every other plane α parallel to β.
Proof: Assume that the plane γ intersects β but not its parallel α. Then from one point X of the intersection
of β and γ can be drawn two different planes parallel to α: β and γ, something contradictory according to
Theorem 8.2.1. Therefore if γ intersects β it must also intersect α, q.e.d.
Corollary 8.2.2 For every line ε parallel to plane β, there exists one exactly plane α which contains ε and
is parallel to β.
Proof: From a point X of ε draw a parallel ζ to β different from ε. According to Proposition 8.2.2 the plane
α of ε and ζ is parallel to β. According to Theorem 8.2.1 α is unique with this property, q.e.d.
Proposition 8.2.4 If the line ε intersects plane α, then it will intersect every other plane β parallel to α.
ε
γ
X
A
α Υ
β B
Figure 8.2.5: Line intersecting parallel planes
Proof: Assume that the line ε intersects plane α at point X but does not intersect its parallel plane β. Draw
an arbitrary plane γ containing ε and intersecting α along the line XY . Because planes α, β are parallel and
γ intersects α it will also intersect β (Corollary 8.2.1) and in fact by parallel lines XY and AB (Proposition
8.2.3). It follows, that on plane γ we have from point X two parallels to AB: line XY and line ε, for
which we assumed that it does not intersect β, therefore also AB. This contradiction shows that if the line
ε intersects α, then it must also intersect β, q.e.d.
Corollary 8.2.3 If the line ε is parallel to the plane α, then it is also parallel to every plane β parallel to
α.
Corollary 8.2.4 From a point X not lying on the plane β are drawn infinitely many lines parallel to plane
β. All these lines are contained in the unique plane α which passes through X and is parallel to β.
Proposition 8.2.5 Parallel planes excise on parallel lines equal line segments.
X' Y'
β
X Y
α
Figure 8.2.6: Parallels between parallels
Proof: Assume that the parallel planes α and β excise on parallel lines the segments XX and Y Y . Then
XX Y Y is a parallelogram, therefore XX and Y Y are equal, q.e.d.
Proposition 8.2.6 If plane β is parallel to α and plane γ is parallel to β, then plane γ is also parallel to α.
Proof: If planes α and γ were not parallel then from their intersection, which is a line ε not intersecting β,
therefore parallel to it, we would have two different planes parallel to β, which is contradictory (Corollary
8.2.2). Consequently α and γ must be parallel, q.e.d.
Exercise 8.2.1 Show that two circles contained in two different planes α and β have at most two common
points.
8.3 Angles in space

What is easy and obvious is never valued; and even what is in itself
difficult, if we come to the knowledge of it without difficulty, and without
any stretch of thought or judgment, is but little regarded.
D. Hume, A Treatise of Human Nature
We measure angles in space the same way we measure them on the plane. Two different and intersecting
lines α and β define exactly one plane ε (Theorem 8.1.2) and on this plane we measure the angles formed
O β
α
ε
Figure 8.3.1: Angle in space
by the lines the same way we do for the angles of that plane (§ 1.4). In particular, two lines in space are
8.3. ANGLES IN SPACE 539
called orthogonal, when they are contained in the same plane ε and they are orthogonal lines of this plane ε.
Comment-1 In the next section we will see that the notion of angle in space is extended to include non
coplanar (skew) lines, which, by definition, are not parallel yet they do not intersect.
X'
O'
β Y'
X
α O Y
Figure 8.3.2: Angles with parallel sides
Theorem 8.3.1 Angles with parallel and equally oriented sides are equal.
Proof: Assume that the two angles XOY ’ and X ◊ O Y have respective sides parallel. Then they are either
contained in different or coincident planes. If they are contained in the same plane, then according to
Corollary 1.15.11 they form equal angles. If they are contained in different planes then these planes are
parallel (Proposition 8.2.2). In this case define on their sides the points such that |OX| = |O X | and
|OY | = |O Y |. Then OXX O and OY Y O are parallelograms and consequently XX and Y Y are
parallel, therefore XY Y X also is a parallelogram. It follows that the triangles OXY and O X Y have
equal respective sides, therefore they are congruent and consequently their angles at O and O are equal,
q.e.d.
Corollary 8.3.1 The two planes XOO and Y OO intersect along line OO . Then two parallel planes
α and β, which intersect OO , intersect the planes XOO and Y OO along lines forming equal angles
’ =X
(XOY ◊ O Y in figure-8.3.2).
Corollary 8.3.2 Let ABΓ∆... be a polygon on the plane α. Consider also a pencil of parallel lines
through its vertices. On every plane β, parallel to α the pencil defines a polygon A B Γ ∆ ... congru-
ent to ABΓ∆....
Δ
Γ
α A B
Δ' Γ'
β Α' Β'
Figure 8.3.3: Congruent polygons defined through a pencil of parallel lines
Proof: According to Proposition 8.2.5, ABB A , BΓΓ B , ..., are all parallelograms, consequently the
polygons have respective sides equal |AB| = |A B |, |BΓ | = |B Γ |, ... According to Theorem 8.3.1 the
polygons also have respective angles equal, q.e.d.
Β Α ζ
ε
Figure 8.3.4: Line orthogonal to line from point
Theorem 8.3.2 From a point A in space, not lying on the line ε, exactly one line AB orthogonal to it can
be drawn.
ζ
Ρ
O B
Υ
ε
A Σ
Χ
Figure 8.8.1: Plane of equal distances from sides of angle
Exercise 8.8.16 Show that the geometric locus of the points P , for which the lines OP form equal angles
’ is the plane ζ which is defined from the bisector of the angle and the line
with the sides of angle XOY
which is orthogonal to its plane ε at point O (Figure 8.8.1).
Exercise 8.8.17 Show that if a plane ε passes through one diagonal of a parallelogram, then the distances
from ε of the end points of the other diagonal are equal.
Exercise 8.8.18 Given is a plane ε and one of its lines η. Find the geometric locus of the points X in space,
which are at a fixed distance d from η and at a fixed distance d from ε. For which d, d there is no solution?
(Ι) Δ (ΙΙ)
β γ
Γ Β
Α ε Γ α
Α
Β Δ
η
ε
Figure 8.8.2: Angle variation Constant cross ratio
Exercise 8.8.19 Given is a plane ε and a line η, orthogonal to the plane at vertex A of the isosceles triangle
’ is greater
ABΓ of the plane (Figure 8.8.2-I). Show that, for every point ∆ = A of the line η, the angle BAΓ
’
than B∆Γ .
Exercise 8.8.20 The three planes {α, β, γ} are fixed and pass through the same line ε (resp. are parallel).
The variable line η intersects these planes respectively at points {A, B, Γ }. Find the locus of a fourth point
∆ of the line η, which defines a constant cross ratio k = (ABΓ∆) (Figure 8.8.2-II).
Exercise 8.8.21 Given are three pairwise skew lines α, β and γ. Construct a line δ, parallel to γ and
intersecting α and β.
Hint: From a point A of α draw the parallel γ to γ and define the plane ε 1 = αγ which contains them.
Similarly, from point B of β draw the parallel γ to γ and define the plane ε 2 = βγ which contains them.
The intersection of the planes ε 1 , ε2 is the wanted line.
Exercise 8.8.22 Given are two skew lines α and β and a plane ε which intersects both. Construct a line
segment AB with its end points respectively on α and β, which is parallel to ε and has a given length δ.
Hint: Planes εα , εβ , which contain respectively α, β and are orthogonal to ε intersect, in general, along a
line γ orthogonal to ε (Figure 8.8.3-I). Triangle ABΓ , formed by the intersections of α, β, γ with a plane
parallel to ε, can be calculated, as a function of the distance x = |A Γ |, such that |AB| = δ. Examine also
γ
B'' Α
β
Γ
(Ι) (II)
α Β
α x
Β Α β
ε Α' Α1
Α'
ε η
Β' δ Β'
Β1 δ Α1 Β1
Figure 8.8.3: Construction of segment of given length
the case where εα , εβ are parallel. In this case ε may be considered as containing the line segment η of the
minimal distance η of the two skew lines (Figure 8.8.3-II). The rectangle with diagonal A 1 B1 , constructed
through the projections of {A, B} on ε, is constructible from the givens. Then, the triangle A AB of the
projections onto ε α is constructible, since its angles and the side AB are determined from the givens.
Exercise 8.8.23 For two different points {A, B} of space, consider alle planes/lines α through A and on
each such plane/line the projection B α of B. Show that all these points {B α } have the same distance
r = |AB|/2 from the middle M of AB.
Exercise 8.8.24 The planes {α, β}, intersect along the line ε and contain corresponding points {A, B}.
Show that line AB forms with the planes the same angle, if and only if the points are at equal distance from
line ε.
Hint: Project A on line ε at the point A and to plane β at the point A and analogously project point B to
{B , B } onto {ε, α}. The equality of angles of AB with the planes is equivalent with the equality of the
orthogonal triangles {A BA, B AB}. This is equivalent with the equality of {|AA |, |BB |}, which in
turn, is equivalent with the equality of the triangles {AA A , BB B }.
Exercise 8.8.25 Show that the medial planes of the sides of a skew quadrilateral pass, all four, through a
common point.
Exercise 8.8.26 Show that if two opposite sides of a skew quadrilateral are equal, then their projections on
the line which joins the middles of the two other sides are also equal.
β
Γ
ζ
γ
Ν
Δ
ε Μ Χ
α
A B
Figure 8.8.4: Points X with equal distances from skew lines {α, β}
Exercise 8.8.27 Show that there are infinite many lines γ, whose points X are equidistant from two given
skew lines {α, β}.
Hint: Take two arbitrary but equal segments {AB, Γ∆} respectively on {α, β}. Consider the medial planes
{ε, ζ} of the segments respectively {A∆, BΓ }. The intersection line γ of two such planes does the work
(Figure 8.8.4). The triangles {XAB, X∆Γ } and their altitudes from X are equal.
CHAPTER 9
Solids
9.1 Dihedral angles

Narrative is linear, but action has breadth
and depth as well as height and is solid.
Thomas Carlyle
Two planes α and β, which intersect, separate space into four parts, called quadrants. They also define
shapes analogous to the angles which are defined by two intersecting lines of the plane. Each pair of half
(Ι) (2) (1) (ΙΙ)

Y β
β
Υ
α
Χ
X
(3) (4) α
Figure 9.1.1: Quadrants Diheder
planes which are parts of different planes defines a convex dihedral angle or diheder. The line XY of the
intersection of α and β is called edge of the diheder. The half planes (as well as the planes also) which
define the diheder are called faces of the diheder.
To the notion of the diheder corresponds, in plane geometry, that of the usual angle. This is emphasized
also with the next definition which gives the way we measure the magnitude of a dihedral angle. The
measure of diheder is defined by drawing a plane orthogonal to the edge. This plane intersects the faces
along lines OA and OB and as measure of the diheder we define the measure of the angle AOB. ’ This way
’
the convex diheder corresponds to the convex angle AOB with measure less than π (in radians) and the
non-convex diheder corresponds to the non-convex angle AOB ’ with measure greater than π.
We must note that the measure of the diheder is independent of the position of O on its edge XY . If we
draw the orhthogonal to XY plane at another point O of the edge, then the corresponding angle A ◊ O B
’ consequently the two angles will be equal

will have its sides parallel and equally oriented to those of AOB,
(Theorem 8.3.1).
558 CHAPTER 9. SOLIDS
(I) (II) Y
Y B'
O β
β B
O B
X A
X A
A'
α
α
’ < A
Figure 9.1.2: Diheder angle measure AOB ÷ OB
Theorem 9.1.1 The measure AOB ’ of a diheder is less from that of every other angle A ÷ OB , formed by
intersecting the diheder with a plane which is not orthogonal to its edge (Figure 9.1.2-II).
Proof: The proof of this is identical with that of the exercise 8.6.2, q.e.d.
We call two dihedral angles congruent when they have the same measure. More generally, by transfer-
ring the notions of from plane angles to the diheders, we talk about right diheder, obtuse diheder, acute
diheder, complementary diheder, supplementary diheder, orthogonal diheder, etc. Properties similar
to these of angles also hold. It suffices to re-examine the properties of plane angles and substitute in them
the phrase side of angle with the phrase face of diheder. The next proposition gives an example of these
analogies.
Y
Y'
β
O B β'
O' B'
X A
X' A'
α
α'
Figure 9.1.3: Congruent dihedral angles
Proposition 9.1.1 Dihedral angles, whose respective faces are parallel planes are either congruent or
supplementary.
Proof: Suppose that the faces α , β of one diheder are parallel planes respectively to the faces α, β of
the other. Then their edges are parallel. This can be seen by extending one face, for example β until it
intersects α along the line ε. According to Proposition 8.2.3, ε and XY will be parallel as excised by the
parallel planes β and β through α. Similarly also ε and X Y will be parallel as excised by the parallel
planes α and α through β. Consequently the edges XY and X Y will be parallel lines. Therefore a plane
’ of the one diheder will also be orthogonal to the
γ orthogonal to the edge XY which defines the angle AOB

edge X Y of the other diheder (Proposition 8.5.2) and will define the angle A ◊ O B of the other diheder.
’ ◊

Because of the parallel planes angles AOB and A O B will have parallel sides, therefore they will either
be congruent or supplementary (Corollary 1.15.11), q.e.d.
We call two planes which form a right dihedral angle, orthogonal planes.
Theorem 9.1.2 If a line AB is orthogonal to the plane α, then every other plane β which contains AB will
be orthogonal to plane α.
Proof: Line AB (with B on α) is orthogonal to every line of α which passes through B (Theorem 8.5.1),
therefore it will be orthogonal also to line BΓ , which is orthogonal to the intersection XY of the planes
and is contained in α. This shows that the angle of the diheder between α and β is right, q.e.d.
9.1. DIHEDRAL ANGLES 559
X Y
Γ B
α
Figure 9.1.4: orthogonal planes
Proposition 9.1.2 For two planes α and β, intersecting along line XY , the points in space Γ , which
are equidistant from the planes are to be found on two other planes γ, δ, which pass through XY , form
congruent diheders with α and β and are orthogonal between them.
Y
B β
O
γ
X
Γ
A
α
Figure 9.1.5: Bisector plane of diheder
Proof: Let the point Γ be equidistant from planes α and β. Draw the orthogonals Γ A and Γ B to them.
Plane ABΓ is orthogonal to XY , because XY , as a line of plane β, is orthogonal to BΓ and, as line of
’ where O is the point of intersection of XY
plane α, is orthogonal to AΓ . It follows, that angle AOB,
with plane ABΓ , is the dihedral angle between α and β which contains Γ . Also the right triangles OΓ A
and OΓ B are congruent, having the hypotenuse OΓ common and the orthogonals Γ A and Γ B equal by
assumption. It follows that plane γ, which contains Γ and the line XY forms equal angles with the planes
α and β. Similarly we show that if Γ belongs to one of the other quadrants defined by the two planes, it is
contained either in γ or in plane δ which forms a right angle with γ (bisects the external diheder), q.e.d.
In the previous proposition we have again a property of the angles of the plane (Corollary 1.15.17) which is
transferred to dihedral angles. The planes, defined by the previous proposition, are called bisecting planes
or bisectors of the diheder.
Exercise 9.1.1 Find the geometric locus of the points X in space, which are at a given distance δ from the
faces of a diheder angle.
Theorem 9.1.3 A line α, which is not contained in and is not orthogonal to the plane ε, is contained in
exactly one plane ζ orthogonal to ε.
Proof: From an arbitrary point X of α draw the orthogonal XY to ε. The plane ζ which contains α and
XY is orthogonal to ε (Proposition 9.1.2). Every other plane ζ which contains α and is orthogonal to ε,
will intersect ε along a line β. From X draw the orthogonal XY to β on plane ζ . Line XY will be
also orthogonal to ε because besides line β it will also be orthogonal to a orthogonal to it contained in ε,
because of the assumption that ζ and ε intersect orthogonally. Consequently line XY will coincide with
the previous XY and plane ζ will contain α and XY , therefore it will be coincident with plane ζ, q.e.d.
Χ α
ε O Υ β
Figure 9.1.6: Plane through line orthogonal to other
Line β, defined in the previous proposition, as the intersection of the unique plane ζ which contains the
line α and intersects orthogonally the plane ε, is called projection of the line α to the plane ε.
Proposition 9.1.3 The projections α , β of two parallel lines α and β on the plane ε are either coincident
or they are two points or they are two parallel lines of ε.
X Y
α β
Y' β'
Χ' α'
ε
Figure 9.1.7: Projection of parallels to parallels
Proof: If lines α, β are orthogonal to ε, then their projections are two points: the points at which they
intersect plane ε. If plane ε is orthogonal to their plane, but not orthogonal to α, β then the lines project on
the same line of ε. In the case where none of the above happens, considering a point X on α and a point
Y on β and projecting them onto ε, we define the planes orthogonal to ε which contain the projections α
and β of the two lines. These two planes, as containing two pairs of parallel lines are parallel (Proposition
8.2.2) therefore lines α , β by which they intersect plane ε will be parallel (Proposition 8.2.3), q.e.d.
Corollary 9.1.1 The projection of a parallelogram of the plane ε onto another plane ζ, which is not or-
thogonal to ε, is a parallelogram.
Corollary 9.1.2 Two triangles ABΓ and ∆EZ, contained in plane α, which are congruent and with par-
allel sides are projected onto plane β, not orthogonal to α, to two triangles A B Γ and ∆ E Z , which
are also congruent and have parallel sides.
Y
α
Γ Z
A Δ
B β
E
Α' Γ' Z'
X Δ'
Β'
Ε'
Figure 9.1.8: Projection of congruent triangles onto congruent triangles
Proof: Follows from the previous corollary, since the parallelogram, for example BΓ ZE, which is defined
by two equal and parallel sides of the triangles on α, will be projected onto a parallelogram B Γ Z E ,
forming equal and parallel sides B Γ and E Z , q.e.d.
9.2. TRIHEDRAL ANGLES 561
Corollary 9.1.3 Two polygons ABΓ∆E..., Π P Σ T..., contained in plane α, which are congruent and
with parallel sides, are projected onto plane β, not orthogonal to α, as two polygons A B Γ ∆ E ...,
Π P Σ T ..., which are also congruent and have parallel sides.
Proof: Follows from the previous corollary by dividing the two polygons into triangles and projecting,
q.e.d.
Exercise 9.1.2 Show that the projection M of the middle M of line segment AB on a plane ε is the middle
of the line segment A B , where A and B are the projections of A and B on that plane.
Exercise 9.1.3 If A B Γ is the projection of triangle ABΓ on the plane ε, show that the projection of the
centroid of ABΓ is the centroid of A B Γ .
Exercise 9.1.4 If line ε is not orthogonal to plane α and ω is an angle 0 < ω < π/2, show that there are
two planes containing line ε and making with α dihedral angles equal to ω. Construct these planes and the
bisector of the diheder they form.
9.2 Trihedral angles

If by chance ten or twenty centuries of history were
eliminated, our knowledge relative to human nature
wouldn’t change a bit. The only irreversible loss would
be if we lost the works of art these centuries bore.
Claude Levi Strauss, With an eye on things
Three half lines OA, OB, OΓ , not contained in a plane and passing through a common point O, define
pairwise, planes and create a solid shape, which divides space in two parts {Σ , Σ }. From these two parts
one, Σ say, is convex (i.e. for every pair of points X, Y in Σ the entire segment XY is contained in Σ ),
and the other non-convex. Selecting one of these parts, usually the convex part, defines a trihedral angle
Γ
α
β
B
α
β γ γ
O
A
Figure 9.2.1: Trihedral angle or triheder
or triheder. The part of space selected is called interior of the triheder. The half lines which define the
triheder are called edges of the triheder and the angles α = BOΓ ’, β = Γ ’ ’ as well as the
OA, γ = AOB,
planes containing them, are called faces of the triheder.
The dihedral angles with edges the lines OA, OB, and OΓ , are called diheders of the triheder and
are often denoted below respectively by A, B,
“ Γ . Often with the same letter we denote also the measure
of the angle. This way, in the various propositions below α, β and γ may denote the faces or/and even the
measures of the faces of the triheder. Similarly A, B,
“ Γ may denote the diheders or/and their measures,
which are defined by measuring the angles which these diheders excise on planes orthogonal to their edges.
◊and O
Two trihedrals OABΓ Ÿ
A B Γ are called congruent when they have respective faces congruent and
respective diheders also congruent.

◊the sum of two of its faces is greater than the third.

Theorem 9.2.1 In every triheder OABΓ
O β
γ α
ω
A Γ
Δ
Χ Z
B
Υ
Figure 9.2.2: Inequalities between faces
’ be the greater of the three faces α = Y’

Proof: Let β = XOZ ’ γ = ZOX.
OZ, β = XOZ, ’ We will show
that β < α + γ. Consider two arbitrary points B and Γ on the edges OY and OZ respectively. Then draw
on the plane XOZ the triangle ∆OΓ , so that it is congruent to BOΓ and extend ∆Γ up to its intersection
point A with the third edge. From the way the triangles ABO and A∆O were defined, they have AO
common and BO and ∆O equal. Also from the triangle inequality we have
|AΓ | = |A∆| + |∆Γ | < |AB| + |BΓ |.
However |∆Γ | = |BΓ | ⇒ |A∆| < |AB|. From Corollary 1.10.10 we have for the angle
’ < AOB
ω = AO∆ ’ ⇔ AOΓ
’ − BOΓ
’ < AOB,
’ q.e.d.
Theorem 9.2.2 In every triheder OABΓ the sum of its faces is less than 2π.
Proof: Consider arbitrary points A, B and Γ respectively on the edges OX, OY and OZ of the trihedral
(figure 9.2.2). This forms three more triheders with vertices the points A, B and Γ . Applying the previous
proposition to these three triheders we have
α+β+γ ’ − OΓ
= (π − OBΓ ’ ’ − OΓ
B) + (π − OAΓ ’ ’ − OBA)
A) + (π − OAB ’
’ + OBA)
= 3π − (OBΓ ’ − (OΓ ’ ’
B + OΓ ’ + OAB)
A) − (OAΓ ’
’ + BΓ
< 3π − (ABΓ ’ ’
A+Γ AB)
= 3π − π = 2π, q.e.d.
Considering equal line segments |OA| = |OB| = |OΓ | = δ respectively on the edges OX, OY , OZ,
we define a triangle ABΓ , which, together with the length δ, describes the triheder completely with the
assistance of a plane figure which is called unfolding of the triheder. This figure results by cutting the
triheder along its edges and rotating the isosceli triangles about their bases, which are the sides of the
triangle ABΓ , until the planes of the isosceli coincide with that of ABΓ . The shape (Figure 9.2.3), which
results can be used also for the construction of the trihedral from paper. It suffices to rotate, inversely to
the previous direction, the isosceli triangles about their bases, until their three vertices coincide at the same
point of space O, which defines then the vertex of the triheder. The next two propositions show that the
inequalities of the two previous propositions are also sufficient conditions, which guarantee that such a
process always leads to a triheder.
Lemma 9.2.1 Given three isosceli triangles with equal legs, whose apical angles, pairwise, have sum of
angles greater than the third, we can construct a triangle, whose sides coincide with the bases of the isosceli.
Π
O Ρ
A
B Γ
A Γ
Z
Χ B
Υ
O
Figure 9.2.3: Unfolding of triheder
Proof: Assume that the three triangles with equal legs are OBΓ , ΠΓ A and P A B . We place two of them,
e.g. the first two, so that the angles at their apexes become adjacent. From the triangle inequality we have
|BΓ | + |Γ A| > |BA|.
Triangles P A B and OAB are isosceli (Figure 9.2.4) with the same length of legs and, by assumption, we
B' B Γ
A'
A
Ρ O=Π
Figure 9.2.4: Trihedral inequality
have that their apical angles satisfy

÷
A ’
P B < BOA.
From Theorem 1.10.4 follows

|A B | < |AB| < |BΓ | + |Γ A|,
consequently the triangle inequality is satisfied and a triangle with sides the bases of the isosceli can be
constructed (Theorem 2.4.1), q.e.d.
Theorem 9.2.3 For every triple of angles α, β and γ with α + β + γ < 2π and such that the greatest of
these is less than the sum of the two others, there exists a triheder having these angles as faces.
Proof: Construct isosceli triangles with apical angles α, β and γ and with their legs all equal to an arbitrary
positive number δ. The assumptions for the angles imply, according to the previous lemma, that the bases
of these isosceles make the sides of a triangle ABΓ . Consider the circumscribed circle of this triangle and
draw the orthogonal to its center K. On this orthogonal consider then a line segment KO of length

κ = |KO| = δ 2 − ρ2 ,
Γ
K
ε A B
Figure 9.2.5: Construction of triheder
where ρ is the radius of the circumscribed circle. From the pythagorean theorem it follows that |OA| =
|OB| = |OΓ | = δ and the isosceli triangles BOΓ , Γ OA, AOB have equal legs and their bases are respec-
tively equal to those of the initial isosceli triangles, therefore they are congruent to them and a trihedral with
the given angles (faces) is formed at O. The fact that point O is outside the plane ε of the triangle ABΓ is
guaranteed by the assumption α + β + γ < 2π, from which follows δ > ρ. Indeed if point O coincided
with K, that is if δ = ρ, then we would have α + β + γ = AKB ’ + BKΓ ’ + Γ’ KA = 2π, contrary to
the assumption. Also if it was δ < ρ, then angle AOB ’ would be on plane ε of triangle ABΓ and less than
’ because the isosceles AOB would be inside the isosceles AKB (Proposition 1.10.1). Similar things
AKB,
would hold also for the other angles BOΓ ’ and Γ ’ OA. Consequently α + β + γ = BOΓ ’ +Γ ’ ’<
OB + AOΓ
’ + Γ’
BKΓ KA + AKB ’ = 2π, again contrary to the assumption, q.e.d.
Γ'
A'
Β'
B
A
Figure 9.2.6: Vertical triheders
Comment-1 The last proposition implies that a trihedral angle is determined completely from its faces.
Before to proceed to the proof of the basic congruence theorem for triheders (alternatively: congruence
β β
γ γ
α α
Figure 9.2.7: Opposite orientation of triheders
of spherical triangles, comment-2 in § 9.15), I will note that congruence among triheders, as we defined
it, is not coincident with coincidence in space, in the sense of moving the one until it coincides with the
other. If the previous action can be done, then, of course, the triheders are congruent. Our definition of
congruence, however, includes also another case in which coincidence is not possible. This is the case of
◊and OA
the vertical triheders. Here (figure-9.2.7) the two triheders OABΓ Ÿ B Γ have respective faces
congruent and respective diheders also congruent. However it is not possible to place one onto the other so
that they are coincident. This because their orientations are different. The two triheders result from three
isosceli triangles through the previous proposition but are placed in such a way so that the base triangles
have different orientation.
Ο Ο'
Ζ Ζ'
Ε Ε'
Γ A Γ' Α'
Δ Β' Δ'
Β
Figure 9.2.8: Congruence of triheders with equal faces
Theorem 9.2.4 Two triheders with respective equal faces are congruent, i.e. they have equal also the
corresponding diheders lying opposite to equal faces.
Proof: Replacing, if necessary, a triheder with its vertical, we may assume that the triheders have the same
orientation. Adopting this assumption, we show that the diheder A is equal to A
. Analogously is proved
the congruence of the other diheders. For the proof we define on the edges equal segments
OA = OB = OΓ = O A = O B = O Γ ,
by which are defined corresponding equal isosceli triangles (Figure 9.2.8)
OAB = O A B , OBΓ = O B Γ , OΓ A = O Γ A .
We take also equal segments OZ = O Z respectively on {OA, O A } and draw at points {Z, Z } the
orthogonal planes to these edges. There are then defined the triangles {∆EZ, ∆ E Z } proven easily to be
’ ∆
congruent. From the congruence of the triangles follows that their angles { ∆ZE, ◊ Z E }, which coincide

with the corresponding diheders { A, A } are equal, as we wanted.
The congruence of triangles {∆EZ, ∆ E Z } results by comparing the right triangles {AZ∆, A Z ∆ }
and showing them congruent. Similarly the triangles {AZE, A Z E } are congruent, and {ABΓ , A B Γ }
are congruent and {A∆E, A ∆ E } are congruent. The proof of these congruences is trivial, q.e.d.
Comment-2 In § 9.15 we will see that the knowledge of the diheders of a triheder also determines com-
pletely the triheder. There exists therefore an equivalence between faces and diheders of a trihedral angle:
The former determine completely the latter and vice versa. Next theorem corresponds to the SAS congru-
ence criterion for triangles (Proposition 1.9.1).
Theorem 9.2.5 Two triheders having one diheder equal and the adjacent to it corresponding faces equal
are congruent.
Proof: We define again on the two triheders equal segments on the edges
OA = OB = OΓ = O A = O B = O Γ ,
Β'
Α Α
(Ι) Β Β' A (ΙΙ) Β (ΙΙΙ) Β
Γ'
Γ'' Γ''
Γ1 Δ' Δ'
Γ''
Γ1 Β''
Δ' Δ
Γ Γ1
Δ
Figure 9.18.13: Acute diheders ⇒ acute faces
Hint: Use contradiction, assuming that one face e.g. BA∆ ’ is obtuse or right. Then, by hypothesis, the
projection of AΓ onto the plane BA∆ falls inside the angle ABΓ (Figure 9.18.13-I). From a point Γ 1 of
this projection we draw the plane which is orthogonal to AΓ and whose intersection with the diheder AΓ
defines the measure B ÿ Γ ∆ of the diheder. They result the following possibilities.
(1) This plane, orthogonal to AΓ , intersects both other edges and defines a triangle Γ ∆ B . Rotating
this triangle about B∆ we define the equal to it triangle B Γ ∆ , with Γ on line AΓ1 . Since the angle
◊
AΓ ∆ is a right one, from the right triangle AΓ ∆ results that |Γ ∆ | = |Γ ∆ | < |A∆ |. Analogoysly
and |B Γ | < |AB |. Hence, for the angles B ◊Γ Γ > B÷ A∆ , which means that the diheder AΓ is obtuse,
a contradiction.
(2) If the orthogonal to AΓ plane at Γ is parallel to one side e.g. to AB of the angle BA∆ ’ (Figure
9.18.13-II), then its intersection with the plane BA∆ is a line Γ 1 ∆ parallel to AB and its diheder AΓ is
the supplementary of Γ ∆ Γ1 , which is smaller than the diheder A∆ (Theorem 9.1.1), which is acute, a
contradiction.
(3) If the orthogonal to AΓ plane at Γ intersects AB externally at the point B , then the angle
ÿ
B Γ ∆ , being external of the triangle B Γ ∆ , is greater than the angle Γ ∆ B , which is supplemen-
tary of ÿ Γ ∆ Γ1 , which, in turn is smaller from the diheder of A∆ (Theorem 9.1.1), which is obtuse, a
contradiction (see also exercise 9.4.2).
Figure 9.18.14: The shadow of the cube
Exercise 9.18.36 Show that the shadow of a cube (Figure 9.18.14), resulting by illuminating it with light
beams parallel to a given direction, together with its base, makes a hexagon with the properties: (1) It is
symmetric. (2) Has two opposite angles right and the adjacent to them sides equal to the side of the cube.
(3) Its area is the sum of the area of the square face of the cube and the area of a parallelogram whose one
side is the diagonal of the face of the cube. Determine the directions of illumination such that the hexagon
has {2, 3, . . .} times the area of the cube-face. Show that if the perimeter of the above hexagon remains
constant, then its vertices, which are different from the cube-vertices are contained in circles with centers at
the vertices of the cube and equal radii, their measure depending from the perimeter. How a point of one of
the previous circles determines the direction of the corresponding light beam? Find the area of the shadow,
when the direction of the beam is parallel to a diagonal of the cube. If the perimeter of the hexagon remains
constant, find the direction of the beam for which the shadow has maximal area.
(Ι) (ΙΙ)
Α
y Ε H
Β a
Γ I
Δ x Ζ
b ε
z
Γ Δ Β
Ε
Figure 9.18.15: Projection to equilateral Projection to two equilaterals
Exercise 9.18.37 The right angled triangle ABΓ of the space is projected orthogonally to the plane ε,
onto the equilateral triangle Γ∆E with side-length d. To find the relation satisfied by its orthogonal sides
(Figure 9.18.15-I).
Hint: If {x = |AΓ |, y = |AB|} the orthogonal sides and z = |BΓ | the hypotenuse, we can assume, with-
out affecting the generality that the vertex Γ lies on the plane ε and the vertices {A, B} are projected
correspondingly to {∆, E} of the equilateral Γ∆E. We get the relations:
z 2 = |BΓ |2 = x2 + y 2 , |BE|2 = z 2 − d2 , |A∆|2 = x2 − d2 ⇒ 4(x2 − d2 )(y 2 − d2 ) = d4 ,

b = |BE| = (2a2 + d2 )/(2a), where a = |A∆|.
Exercise 9.18.38 The right angled triangle ABΓ , with the hypotenuse-vertices {A, B} lying correspond-
ingly on two intersecting planes {α, β}, is projected orthogonally to these planes and its projections are
equilateral triangles of side-length d. To determine the dimensions of the right triangle.
Hint: If {I, H} are the projections of {A, B} on the planes (Figure 9.18.15-II), then the right angled
triangles {ABI, ABH}, by assumption will have in common the hypotenuse and their orthogonal sides
|IB| = |HA| = d. Hence they will be congruent. Applying then the previous exercise, we see that for
b = |AI| = |BH| and a = |Γ∆| = |Γ E|. Hence also the right angled triangles {Γ∆B, Γ EA} are equal,
hence the right angled triangle ABΓ is isosceles. To compute |AB| it suffices to apply the relation of the
orthogonal sides of the previous exercise for x = y.
Α2
Δ'
Β' Γ2 Γ1
Δ1
Γ' X
Γ
Δ Β2
Α' Β1
Y
Figure 9.18.16: Constant sum of distances of X from the faces
Exercise 9.18.39 Show that the sum of the distances of a point X of the interior of a regular tetraheder
ABΓ∆ from its faces is constant and equal to the altitude of the tetraheder.
Hint: If {XA , XB , XΓ , X∆ } are the orthogonals from X to the faces (Figure 9.18.16), then the length
of |XA | does not change when X moves on the plane B 1 Γ1 ∆1 parallel to BΓ∆, and passing through X.
Similarly the length |X∆ | does not change when X moves on the plane A 2 B2 Γ2 parallel to ABΓ and
passing through X.
σ
(Ι) (ΙΙ) Δ
Α
α
α
Α
ω
Δ β γ
Β Γ
Γ Β
β
Figure 9.18.17: Right from diheder Intersection of three planes
Exercise 9.18.40 Given is a diheder of measure ω > 90 ◦ . Find a plane intersecting the diheder along a
right angle (Figure 9.18.17-I).
Exercise 9.18.41 The triangle AB∆ is isosceles and orthogonal to the intersection line Γ∆ of the planes
{α, β}, which define a diheder of measure ω (Figure 9.18.17-I). We consider also the sphere σ with diameter
the base AB of the isosceles. Show that the sphere intersects the two planes along equal circles {κ 1 , κ2 }
with a common chord contained in line Γ∆. Find the angle under which this chord is seen from the points
of the circles {κ1 , κ2 } in terms of ω and d = |AB|.
Exercise 9.18.42 Points {A, B, Γ , ∆} are on the sphere σ (Figure 9.18.17-II). The planes {α, β, γ} are
respectively orthogonal to the lines {∆A, ∆B, ∆Γ } respectively at {A, B, Γ }. Locate the intersection
point of the three planes. Locate also the intersection lines of the pairs of these planes.
Exercise 9.18.43 Continuing the previous exercise, consider the three circles {κ 1 , κ2 , κ3 } intersected from
the sphere σ respectively by the three planes {α, β, γ}. Do these circles have a point in common? Do they
pairwise intersect? Which are the intersection points?
Exercise 9.18.44 Two circles {κ 1 (O1 , r1 ), κ2 (O2 , r2 )} in space intersect at a unique point A. Find a
condition which guaranties the existence of a sphere σ containing both circles.
Exercise 9.18.45 Show that three circles of the same sphere cannot be pairwise tangent at the same point
of the sphere.
Figure 9.18.18: Sphere tangent to three other spheres
Exercise 9.18.46 Three spheres are pairwise externally tangent. Find the radius of the biggest possible
sphere which passes through the hole formed by the three spheres (Figure 9.18.18).
CHAPTER 10
Areas in space, volumes
10.1 Areas in space

Accordingly, things arranged in a fixed order, like the successive
demonstrations in geometry, are easy to remember (or recollect) while
badly arranged subjects are remembered with difficulty.
Aristotle, On Memory and Reminiscence
The area of polyhedral surfaces relies on the definitions and the properties of area of plane figures. This
way the area of the surface of a rectangular parallepiped with sides equal to α, β and γ is 2(αβ + βγ + γα),
the area of the surface of a cube of size δ is 6δ 2 and similar calculations give the areas of any polyhedron,
by adding the areas of the corresponding faces on them. The area of polyhedra, therefore, is not difficult to
calculate and consequently possesses little theoretical interest (see however the exercises below).
Difficulties show up when trying to calculate areas of curved surfaces, like the cylinder, the cone, the
sphere, as well as shapes contained in these surfaces. The definition of the area of these surfaces or/and
their subsets in a general way, analogously to the definition of length for general curves, is the object of
Differential Geometry (area of surface region [dC76, p. 114], volume element of manifold [Spi65, p. 130])
and is beyond the scope of these lessons. It is proved, however, that the area of the cone, the cylinder and
the sphere can be defined as a limit of areas of polyhedra inscribed in these surfaces. The process is, in
essence, the same with that we used to define the length and the area of the circle § 6.5. The cone area is
defined as the limit of the areas of regular pyramids inscribed in the cone. Similarly the cylinder area is
defined as the limit of the areas of regular prisms inscribed in the cylinder. In these two objects the area
of the bases is that of the circle, consequently the interest is focused on calculating the area of the lateral
surface.
As the lateral area of the cylinder we consider the limit of the lateral areas of regular prisms inscribed in
the cylinder, while we increase the number of their sides. Let us denote again with Π 1 , Π2 , Π3 , ... the
regular polygons which we used also in § 6.2. They begin with the square (Π 1 ) inscribed in the circle κ
and each one is also inscribed in κ and has twice the number of sides of the previous. Denoting with p ν the
perimeter of Π ν , the area of the lateral surface of the regular prism with base Π ν is
υ · pν ,
where υ is the altitude of the prism. From Lemma 6.2.2 it follows that the sequence
υ · p1 , υ · p2 , υ · p3 , ...
632 CHAPTER 10. AREAS IN SPACE, VOLUMES
Z E Δ
A
ε B Γ
Figure 10.1.1: Area of lateral surface of cylinder
is increasing. And from Lemma 6.2.3 follows that the sequence is bounded. We call the limit of this
sequence the lateral area of the cylinder. From the previous (Corollary 6.3.1), follows then the next
proposition.
Theorem 10.1.1 The area of the lateral surface of the cylinder is equal to
= 2π · ρ · υ,
where ρ is the radius of the base of the cylinder and υ its altitude.
The calculation for the lateral area of the cone is done in a similar way. As lateral area of the cone we
A Δ Χ
E
B Γ
Figure 10.1.2: Area of lateral surface of cone
consider the limit of the lateral areas of the regular pyramids inscribed in the cone as we increase the number
of their sides. Using the previous notation, the area of the lateral surface of the regular cone with base Π ν is
1
· υν · p ν ,
2
where υν is the altitude of the triangular face of the prism (not the altitude of the cone, neither its generator,
but the length of the segment like OE in figure-10.1.2). We again verify easily that the sequence
1 1 1
· υ1 · p 1 , · υ2 · p 2 , · υ3 · p3 , ...
2 2 2
is increasing and bounded by the number
1
· υ · (2πρ),
2
where υ = |OX| is the length of the generator of the cone and ρ is the radius of the circle of its base. Again
according to Axiom 6.1.1 the sequence will converge. Exactly this limit of this sequence we call lateral
area of the cone.
10.1. AREAS IN SPACE 633
Theorem 10.1.2 The area of the lateral surface of the cone is equal to
= π · υ · ρ,
where ρ is the radius of the base of the cone and υ the length of its generator.
Proof: I sketch the proof because I don’t want to enter to details about limits. We then show first that the
sequence
υ1 , υ2 , υ3 , ...
is increasing and bounded, therefore converges, and in fact to υ = |OX|. For the perimeters p ν , we know
already that they converge to 2πρ. To complete the proof we need the rule for limits, according to which
the limit of a product of convergent sequences is the product of the limits. According to this rule then the
limit of the sequence (υ ν )(pν ) will be υ · (2πρ), from which the formula for the area results, q.e.d.
Λ λ
O
μ
η
K κ
Χ
Figure 10.1.3: Area of lateral surface of truncated cone
The calculation of the area of the lateral surface of a truncated cone is reduced to the corresponding calcu-
lation for cones. The lateral surface of the truncated cone can be seen as the difference of the lateral area
of the cone relative to the larger base (circle κ in figure-10.1.3) minus the lateral area of the cone relative
to the small base (circle λ in figure-10.1.3). Applying the formula, we consequently have that the area is
given by the formula
= π(υ2 ρ2 − υ1 ρ1 ),
where ρ1 , ρ2 , υ1 , υ2 are the lengths of the radii and the generators of the cones whose difference is the
truncated cone.
υ1 υ2
ρ1 υ
δ
ρ2
Figure 10.1.4: Section of truncated cone
Setting υ = υ2 − υ1 , δ = ρ2 − ρ1 and substituting in the formula, we have that

υ 2 ρ2 − υ 1 ρ1 = (υ1 + υ)(ρ1 + δ) − υ1 ρ1
= υρ1 + υ1 δ + υδ

ρ
1
= υρ1 + υ δ + υδ
δ
= υ(ρ1 + ρ1 + δ)
= υ(ρ1 + ρ2 ).
The substitution in the middle equality follows from the similarity of the corresponding right triangles. The
half sum ρ = 12 (ρ1 + ρ2 ) is the radius of the middle section of the truncated cone and as such we have the
following proposition.
Proposition 10.1.1 The area of the lateral surface of the truncated cone is given by the formula
= 2π · ρ · υ,
where ρ is the radius of its middle section and υ is the length of its generator.
Corollary 10.1.1 The area of the lateral surface of a cylinder, a cone and a truncated cone is given by the
formula
= 2πρυ,
where υ is the length of the generator and ρ is the radius of the middle section of the solid.
ρ
υ
υ υ
ρ
ρ
Figure 10.1.5: Area of lateral surface, universal formula
Proof: The corollary simply unifies the specific results of the three previous propositions, relying on the
observation that, for the cone the middle section is the circle with radius half that of the base, while for the
cylinder the middle cut is a circle congruent to that of the base, q.e.d.
κ O
ρ
M
Figure 10.1.6: Area of lateral independent of the inclination of AB
Comment-1 The last formula shows that the line segment AB rotated around its middle M generates the
lateral surface of a truncated cone with the same area, independent of its inclination to the plane of the
rotation circle κ (Figure 10.1.6). When AB is orthogonal to the plane of κ, then the lateral surface is that
of a cylinder. When AB is on the plane of κ and smaller than the radius of κ, then this generates a circular
B
M ρ
A
κ O
Figure 10.1.7: Annulus when AB is contained in ε
plane annulus (figure-10.1.7).

10.1. AREAS IN SPACE 635
Corollary 10.1.2 The area of the lateral surface of a cylinder, cone as well as of a truncated cone is given
by the formula
= 2πση,
where σ is the length of the vertical segment at the middle of the generator up to the axis and η is the
altitude of the solid.
η η η
σ σ
Figure 10.1.8: Area of lateral surface, universal formula II
Proof: Proof by the figure. In the case of the cylinder the corollary is identical to the previous. In the cases
of the cone and the truncated cone, from the similarity of the right triangles with sides (ρ, σ) and (η, υ)
respectively follows
ρ η
= ⇒ ρυ = ση.
σ υ
Substituting ρυ in the formula of the previous corollary we prove the claim, q.e.d.
σ
σ
υ υ
η η
ρ
ρ
Figure 10.1.9: Area of lateral surface, alternative universal formula
Comment-2 If we use the “unfolding” terminology, mentioned in sections 9.9, 9.8, and specifically the
fact that the process of unfolding preserves lengths and areas, the formulas, which were proved here with
the usage of limits, are reduced to simple plane area calculations. Next figure shows the unfoldings of a
cone, a cylinder and a truncated cone and their areas. In all three cases υ is the length of the generator and ρ
is the radius of the middle section and is found from the corresponding equality already mentioned above.
In it the right side is, in the cases of the cone and the truncated cone, the length of the arc AXB and in the
case of the cylinder the length of the line segment AB.
Exercise 10.1.1 Show that from all parallepipeds inscribed in a given cylinder the one with square base
has the maximum possible lateral area.
Hint: Use Exercise 2.13.8.

υ υ
υ
A B A B
X
A X
B
2πρ=|AΧB| 2πρ=|AB| 2πρ=|AΧB|
Figure 10.1.10: Area through unfolding
Exercise 10.1.2 Find the cylinder with the lateral surface of maximal area which can be inscribed in a
sphere.
Α
Β' Β
ρ
Γ Β η
σ Γ Α
ρ
Figure 10.1.11: Inscribed cylinder of maximal lateral area
Hint: The quantity σ · η

2 is maximized when σ = η
2 (Figure 10.1.11).
(I) (II) A Γ
A α
Δ
α
Α' B
Γ'
A' β
B ω Δ'
β Β'
Δ
Γ
Figure 10.1.12: Projection of triangle to other plane
Exercise 10.1.3 Show that given two planes α and β and a triangle ABΓ of plane α, such that its side BΓ
belongs to the intersection of α and β, the area of the projection A BΓ of ABΓ on β is
(A BΓ ) = cos(ω)(ABΓ ),
where ω is the angle of the planes α and β (Figure 10.1.12-I).
Exercise 10.1.4 Show that the previous formula holds also when the triangle ABΓ has its side BΓ parallel
to the intersection of α and β.
10.2. AREA OF THE SPHERE 637
Hint: Apply Corollary 9.1.2 and transfer ABΓ to the plane α in a parallel manner to itself, such that its
edge BΓ coincides with the intersection of α and β.
Exercise 10.1.5 Show that the previous formula holds for every triangle ABΓ of the plane α and its pro-
jection A B Γ on the plane β. That is
(A B Γ ) = cos(ω)(ABΓ ),
where ω is the angle of the planes α and β (Figure 10.1.12-II).

Hint: From one vertex of ABΓ for example A draw the parallel A∆ to the intersection of α and β. Apply
the previous exercise to the two triangles AB∆ and AΓ∆ and add or subtract, depending on whether ∆ is
inside or outside of BΓ .
Exercise 10.1.6 Show that for every convex polygon ABΓ∆... of the plane α and its projection A B Γ ∆ ...
on the plane β the following relation holds for their areas:
(A B Γ ∆ ...) = cos(ω)(ABΓ∆...),
where ω is the angle of the planes α and β.

Hint: From one of the vertices of ABΓ∆... draw all its diagonals and decompose it into triangles. Then,
apply the previous exercise and add.
10.2 Area of the sphere

The first was to never accept something as real, unless I know it obviously as real.
Descartes, Discourse on the Method, b’ part
The area of the sphere is calculated through the areas of solids which result from inscribed regular polygons.
Here we think that the sphere is generated through the rotation of a circle κ about one of its diameters AE.
An approximation of the sphere is produced through the rotation of a regular polygon inscribed in the circle
κ. Let us denote again with Π 1 , Π2 , Π3 , ... the regular polygons we used also in § 6.2. They begin from the
square (Π1 ) inscribed in the circle κ and each polygon is inscribed in κ and has double the number of sides
Z Δ
M
H Λ
Γ
K
Θ B
N
A
Figure 10.2.1: Approximation of the surface of sphere
of its predecessor, while all have the diagonal AE in common (Figure 10.2.1), which is also a diameter of
the circle κ. From the rotation of one such polygon about the diameter AE, is produced a surface consisting
of cones and truncated cones.
The lateral surface of these cones and truncated cones was calculated in the previous section and, ac-
cording to the formulas which were proved there, (Corollary 10.1.2) gives a total surface, which in the case
of the octagon (Π 2 , see figure-10.2.1) is
2 = (2π · η2 ) · (|AK| + |KΛ| + |ΛM | + |M E|) = (2π · η2 ) · |AE| = (2π · η2 )(2ρ)
where η2 = |ΛN | is the distance of the side of the octagon (Π 2 ) from the center of the circle, which is
equal to the radius of the inscribed circle of the regular polygon and ρ is the radius of the circle κ in which
the polygons are inscribed. We consider then that the area of the surface of the sphere is the limit of the
sequence which results similarly
1 = (4πρ)η1 , 2 = (4πρ)η2 , 3 = (4πρ)η3 , ....
The fact that this sequence converges follows from the fact that the sequence of radii
η1 , η2 , η3 , ...
of the circles inscribed to the polygons Π ν (which in turn are inscribed in the circle with diameter |AE| =
2ρ) converges to ρ. Consequently the limit of these areas ν is
= 4πρ2 .
Thus, after several assumptions relative to what we call area of sphere, we arrive at the proposition.
Theorem 10.2.1 The area of the surface of the sphere of radius ρ is = 4πρ 2 .
τ
σ
Figure 10.2.2: Sphere and enveloping cylinder
Corollary 10.2.1 A sphere σ has the same area with the lateral surface of its enveloping cylinder τ .
We call spherical zone/cap the part of the sphere which is included between two parallel planes which
intersect it (or are tangent to the sphere) (Figure 10.2.3). The distance υ between the planes is called height
of the spherical zone. The area of the spherical zone is defined, similarly to that of the sphere, as a limit of
areas of solids inscribed in the spherical zone. We consider the spherical zone as being generated through
the rotation of an arc AB˜ of the circle κ about a diameter XY of the circle, which does not intersect the
arc (Figure 10.2.4). The area of the surface produced is defined as the limit of the sum of lateral areas of the
truncated cones which result through division of the arc into ν equal parts and considering the corresponding
polygonal lines with vertices at the dividing points. As first polygonal line P 1 we take the line segment AB,
which through the rotation generates a truncated cone. The second polygonal line results by considering the
middle Γ of the arc and the polygonal line AΓ B. Continuing this way, we produce a sequence of polygonal
Exercise 10.11.17 Show that the area ε of a spherical equilateral triangle of a sphere of radius ρ and its
angle α are connected with the formula
π
α= 2 + .
3ρ 3
Conclude that the angle of the spherical equilateral is always greater than 60 degrees. (Also see Exercise
9.14.4)
The next table records the ratio of the area of the surface (Σ )/a 2 , as well as also the ratio of volume
o(Σ )/a3 , respectively, to the square/cube of the edge of the Platonic solid.
Exercise 10.11.18 Show that the numbers of the first three lines in the following table are valid.
Platonic solid Σ (Σ )/a2 o(Σ )/a3

√ 1
√
Tetrahedron 3 12 2
Cube 6
√ 1
√
1
Octahedron 2 3 √ 3 3√
1
Dodecahedron 3 25 √
+ 10 5 4 (15 + 7√ 5)
5
Icosahedron 5 3 12 (3 + 5)
σ1 σ2
κ
Figure 10.11.10: Volume defined by two spheres
Exercise 10.11.19 Two spheres of different radii {σ 1 (r1 ), σ2 (r2 )} are tangent externally and the cone κ
is tangent to both. Compute the volume of the external domain of the spheres and internal to the cone
included between the circles of contact of the spheres and the cone, in dependence of the radii {r 1 , r2 }
(Figure 10.11.10).
Exercise 10.11.20 Discuss the similar to the previous problem for two spheres {σ 1 (r1 ), σ2 (r2 )} with dif-
feren radii, external to each other and the cone κ tangent to both. Compute the volume of the external
domain of the spheres and internal to the cone included between the circles of contact of the spheres and
the cone, in dependence of the radii {r 1 , r2 } and the distance d of the centers.
Exercise 10.11.21 Let ABΓ∆ be the base of a cube. On the orthogonal edge from Γ we consider a line
segment ΓΘ of length x (Figure 10.11.11-I). From the point H of the orthogonal edge at B and at distance
|BH| = y we draw the parallel HI of the diagonal B∆. For some values of x, y the plane Θ HI intersects
the cube along a pentagon. Given x, find for which other values of y this happens. Also calculate the area
of the resulting pentagon and the volumes of the two parts of the cube.
Hint: One way is through Exercise 10.1.6.
Exercise 10.11.22 In a cube ABΓ∆EZHΘ draw the diagonals {AΓ , BZ} (Figure 10.11.11-II). Show
that the dihedral angle between the planes {BZΓ , ZHΓ } e’inai 60 ◦ .
E E Δ
(Ι) (ΙΙ)
Z
I Ζ A
H
y Θ Θ Γ
Δ
A x
B Γ Η B
Figure 10.11.11: Cube’s pentagonal section Dihedral angle
Hint: The tetraheders {BΓ ZH, ABΓ Z} are symmetric relative to the plane BΓ Z and the tetraheders
{ABΓ Z, AΓ∆Z} are symmetric relative to the plane AΓ Z.
Δ
Ι
Γ
Κ
Μ
Β Ο
Figure 10.11.12: Cutting with plane through the middles of opposite sides
Exercise 10.11.23 The plane ε passes through the middles {M, N } of the opposite edges of the tetraheder
ABΓ∆ (Figure 10.11.12). Show that the intersection points {I, K} defined on the two other edges, corre-
spondingly, {AΓ , B∆} divide them in proportional parts : K∆
KB
= IΓ
IA .
Hint: Consider the intersection point O of the intersection plane with ∆Γ and apply the theorem of
Menelaus (1) to the triangle A∆Γ with secant line ON and (2) on the triangle ∆BΓ with secant line
OM . It follows that
KB O∆ M Γ IΓ O∆ N A
· · = 1 = · · ,
K∆ OΓ M B IA OΓ N ∆
and the relation follows through simplification.
Exercise 10.11.24 Continuing the previous exercise, show that a plane ε, passing through the middles
{M, N } of two opposite edges of the tetraheder ABΓ∆, divides it in two solids {σ 1 , σ2 } of equal volumes.
Hint: Each of the two solids {σ 1 , σ2 } is the union of a quadrangular pyramid τ i and a tetraheder δ i . σ1
consists of {τ1 = ∆M IN K, δ1 = ∆M Γ I} and σ2 of {τ2 = AKM IN, δ2 = KBM A}. The pyramids
{τ1 , τ2 } have the same volume because they have the same base and the same height.
Tetraheders {δ1 , δ2 } have also equal volumes. This results by comparing first the volumes of the half of
tetraheders o = o(∆ABM ) = o(∆M Γ A).
o(δ1 ) (∆IΓ ) |IΓ | o(δ2 ) (BM K) |BK|
= = , = = .
o (∆Γ A) |AΓ | o (BM ∆) |B∆|
The conclusion results, using the previous exercise, from the equality of the ratios of the segments ([Cat52,
p. 241]).
Ο
β
α
ε
Figure 10.11.13: Cylinder with oblique cut
Exercise 10.11.25 A right circular cylindrical surface of radius r is cut with a plane ε orthogonal to its
axis and a plane ε oblique to its axis, such that in the solid Σ , bounded by the cylinder and the two planes,
the longest generator has length β and the shortest has length α (Figure 10.11.13). Invent a method to
calculate the lateral area of the solid as well as its volume in terms of {r, α, β}. Show that these quantities
do not change if {α, β} vary but their sum remains constant.
Γ
(Ι) (ΙΙ) Α
Ο Δ
ω
Β
Figure 10.11.14: Cylindrical frame
Exercise 10.11.26 Devise a method to construct a rectangular cylindrical frame of outer dimensions {α, β}
(Figure 10.11.14-I). The sides of the frame are parts of a cylindrical surface of radius r. Compute the total
surface and volume of the frame in terms of {r, α, β}.
Exercise 10.11.27 Suppose you would like to build a frame like the previous one, but without these sharp
corners. Devise a method to construct the same frame with rounded corners. You could use as a building
element the one shown in figure 10.11.14-II. In this we use a solid Σ like the one of exercise 10.11.25, which
we cut with a plane orthogonal to the axis at its intersection O with the oblique plane. Then we replace the
upper cylindrical part with a spherical lune AOBΓ∆ (Figure 10.11.14-II). Calculate the lateral surface
and volume of such a solid building element, especially when the angle of the lune ω = 45 ◦ . Then calculate
the surface and volume of the corresponding frame with rounded corners. Note that, for the time being, we
cannot compute the area of the oblique base of this building element, which consists of a half-circle and a
half-ellipse glued together along the diameter AB of the half-circle.
CHAPTER 11
Conic sections
11.1 Conic sections

It will be thought, no doubt, that the goal and mark of knowledge
which I myself set up (the very point which I object to in others) is
not the true or the best; for that the contemplation of truth is a thing
worthier and loftier than all the utility and magnitude of works;
Bacon, Novum Organum IV 110
Intersections of a conic surface Σ with a plane ε, orthogonal to its axis and not passing through its apex O,
are circles (Proposition 9.9.2). When the plane is not orthogonal to the axis of Σ , the resulting intersections
are ellipses, parabolas and hyperbolas. The ellipses result, as we will see, through intersections with planes
cutting all the generators of Σ . An ellipse is also characterized as the geometric locus of points X, which
(I) (II) (III)
Υ Χ
Χ
X
ξ
M
Α Ξ Α
Β B
A
Figure 11.1.1: Ellipse |XA| + |XB| = λ, Parabola |XA| = |XY |, Hyperbola ||XA| − |XB|| = λ
have fixed sum of distances from two fixed points A and B (Figure 11.1.1-I).
|XA| + |XB| = λ.
Points A and B are called foci (sing. focus) or focal points of the ellipse. Ellipses also result when a right
circular cylinder is intersected with a plane non-orthogonal (and not parallel) to its generators. Circles can
be considered as a special case of ellipses, whose focal points A and B coincide.
674 CHAPTER 11. CONIC SECTIONS
Parabolas result through intersections with planes which are parallel to some plane tangent to the conic
surface (§ 9.9). The parabola is characterized also as the geometric locus of the points X, which have
the same distance from a fixed point A and fixed line ξ (Figure 11.1.1-II). Point A is called focus of the
parabola and the line ξ is called directrix of the parabola.
Finally hyperbolas result through intersections with planes, which are parallel to exactly two generators
of the conic surface. The hyperbola is characterized also as the geometric locus of the points X, which
have fixed difference of distances from two fixed points A and B (Figure 11.1.1-III).
||XA| − |XB|| = λ.
Again the points A and B are called foci or focal points of the hyperbola.
These three kinds of curves (ellipses, parabolas, hyperbolas) are collectively called proper conic sec-
tions and result as intersections of a conic surface with a plane not passing through the vertex O. Ellipses
and hyperbolas admit, as we will see, two axes of symmetry as well as a center of symmetry and are called
conics with center. The parabola however admits only one axis of symmetry but not a center of symmetry.
When the plane ε, with which we intersect the conic surface, passes through the vertex O of the conic
surface, then, depending on the inclination of ε, result the following shapes: (a) a point (plane ε contains
(α) (β) (γ) (δ)
Figure 11.1.2: The four kinds of singular conic sections
then point O and no other point of Σ ), (b) two intersecting lines (ε intersects Σ along two generators) and
(c) a line (ε is tangent to Σ ).
Collectively, these particular cases of sections are called singular conic sections (Figure 11.1.2). To
these are added also pairs of parallel lines, considering them as intersections of a cylinder with a plane
parallel to its generators.
Comment-1 Usually the student meets conic sections, long before he meets them from the geometric view-
point, in the form of graphs of simple functions. For example, the parabola as the graph of the function
2
y = ax2 + bx + c, whose focus is at point x 0 = − 2a
b
, y1 = 4ac−b
4a
+1
and the directrix is the parallel to the
3
x0
O 2
y1
Γ 1
y0
A y2
B 3 -2 -1 1 2
(1) y = ax2 + bx + c (2) y = x2 , y1 = 14 , y2 = − 14
Figure 11.1.3: Graphic representation of trinomial

11.1. CONIC SECTIONS 675
4ac−b2 −1
x-axis from the point of the y-axis y 2 = 4a (Figure 11.1.3).
The student meets also the hyperbola as the graph of the function y = x1 , which is a hyperbola passing
√ √
through points ±(1, 1) and with foci at points ±( 2, 2). More generally, the function y = ax+b
cx+d , where
a, b, c and d are constants, also represents a hyperbola. (Figure 11.1.4).
1
-d/c
-3 -2 -1 1 2 3
a/c
-1
-2
-3
1
(1) y = x (2) y = ax+b
cx+d
Figure 11.1.4: Graphic representation of hyperbola
Exercise 11.1.1 Construct the points of the ellipse γ(A, B, λ), which satisfy |XA| + |XB| = λ and which
lie on the orthogonals of the line which joins the foci ε = AB at points A and B.
Hint: If x = |XA| is the distance of one such point of the orthogonal of AB at A, then d 2 + x2 = (λ − x)2 ,
where d = |AB|.
Exercise 11.1.2 Construct the points of the parabola γ(Π , ξ) with focus Π and directric ξ, by finding
intersection points B of γ with lines orthogonal to ξ.
Exercise 11.1.3 Construct points of the parabola γ(Π , ξ) with focus Π and directrix ξ, by finding inter-
section points B of γ with lines parallel to ξ and on the same side of ξ as point Π .
Hint: If ε is one such parallel of ξ, consider the symmetric Π of Π relative to ε and the projection O of Π
on ξ. If Γ is the projection of the wanted point B of ε, then must hold |OA| · |OA | = |OΓ |2 .
Exercise 11.1.4 Construct the points of the hyperbola γ(A, B, λ), which satisfy ||XA| − |XB|| = λ and
which are found on the orthogonals of the line through the foci ε = AB at points A and B.
Hint: If x = |XA| the distance of one such point of the orthogonal to AB at A, then λ(2x + λ) = d 2 ,
where d = |AB|.
Comment-2 The hyperbola (a special case of it, called “rectangular”) is the only kind of conic section
which represents the graph of an invertible function. This function, which has the form y = ax+b
cx+d , is used
for the definition of the homographic relation of {x, y} (§ 5.21).
11.2 Dandelin’s spheres

Words should be simple and straight. Why were the words of our Sage so
subtle (ao)? That is because truth itself is subtle and intriguing; to
communicate this kind of truth, the Sage could not but use subtle words.
Xun Yue, Shenjian (Extended Reflections)
The Dandelin’s spheres (1794-1847) readily explain the relationship between conic sections and the geo-
metric loci, which were mentioned in the previous section. They are spheres inscribed in conic or cylindrical
surfaces which are simultaneously tangent to another plane ε. The next propositions show that every plane
ε, which intersects such a surface defines two (or one in some cases, as we will see in the next section)
spheres called Dandelin’s spheres of plane ε. Of decisive importance in this section is Theorem 9.9.1,
according to which every plane ε, not passing through the vertex O of a conic surface, intersects (a) either
all generators, (b) or all except two, (c) or all except one. Here we will examine the first two cases. In the
next paragraph we will discuss the third as well.
Proposition 11.2.1 Let ε be a plane, which intersects all the generators of the conic surface K. Then, there
exist two spheres tangent to the conic surface and simultaneously tangent to the plane ε.
κ
Υ Z
ε ζ
Ρ Υ Ζ'
ζ Ρ
Π Ω
A Ω A
E Π
B T B
T
Χ Η'
μ Χ
Σ
H
Figure 11.2.1: Dandelin’s spheres of the intersection of a conic surface with a plane
Proof: Consider the plane θ, which is defined from the axis OX of the conic surface K and the orthogonal
OΩ from the apex O of the cone to the plane ε. Let also line ζ be the intersection of the planes ε and θ.
By construction, line ζ is orthogonal to OΩ and intersects the two generators of the conic surface K which
are contained in θ at points A and B (Figure 11.2.1). This defines then on plane θ the triangle OAB and
consequently its inscribed and escribed at the angle AOB’ circles with centers, respectively, P and Σ . The
spheres with centers points P and Σ and radii, respectively, those of the inscribed and escribed circles, P Π
and Σ T , are the wanted ones.
I show that one of them, P (P Π ) is tangent to the cone. Indeed, if OH is another generator and Z is the
projection of P onto it, then the right triangles P ZO and P Z O, where Z is the projection of P on OA,
are congruent. This because P lies on the axis of the conic surface, therefore has the same distance from the
generators, hence |P Z| = |P Z |. Also the two right triangles have the hypotenuse OP in common, they
are therefore congruent. This implies that OH is tangent to the sphere at Z and proves the claim. Similarly
it is proved that the sphere Σ (Σ T ) is also tangent to the conic surface. That these two spheres are also
11.2. DANDELIN’S SPHERES 677
tangent to the plane ε follows directly, e.g. for P (P Π ), follows from the fact that P Π is also orthogonal
to the plane ε, as parallel to OΩ , which has this property. Consequently the plane ε is tangent to the sphere
P (P Π ), being orthogonal at the end point of one of its radii, q.e.d.
Theorem 11.2.1 The intersection of a conic surface K with a plane ε, which intersects all generators and
does not pass through the apex, defines an ellipse with foci at the contact points of the plane with the
corresponding two Dandelin’s spheres.
Proof: The proof results from the equality of the tangents from a point to a sphere (Theorem 9.11.2). If
E is the point of intersection of ε and K (Figure 11.2.1), then the generator OE is simultaneously tangent
to both spheres. Also, if Π and T are the contact points of the two spheres P (P Π ) and Σ (Σ T ) with the
plane ε, then EΠ and ET are respectively also tangent to these spheres. It follows that |EΠ | = |EZ| and
|ET | = |EH|, where E and Z are the contact points of the generator OE with the spheres. Then however
the sum
|ET | + |EΠ | = |EZ| + |EH| = |ZH|,
is fixed and equal to the length of the generator of the truncated cone formed from the circles κ and µ, along
which the two spheres are tangent to the conic surface, q.e.d.
Proposition 11.2.2 Let ε be a plane intersecting all the generators of a cylindrical surface K. There exist
then two spheres tangent to K and simultaneously tangent to the plane ε.
Ρ Ρ
μ Z Ζ'
Π
α I
Ω
Π
T E
T β
K
ε
ζ
ζ
κ H Η'
Σ Σ
Figure 11.2.2: Dandelin’s spheres of the intersection of a plane and a cylindrical surface
Proof: The proof is similar to that of Proposition 11.2.1. The plane of the paper in figure-11.2.2 is the plane
θ containing the axis η of the cylindrical surface and the orthogonal OΩ on the plane ε from an arbitrary
point O of the axis. Plane θ intersects the cylindrical surface K along two (parallel) generators α and β.
It also intersects plane ε along a line ζ. Drawing the bisectors KΣ , IP of the angles formed by ζ with
α, β and their intersections P , Σ with the axis η, we define the circles P (P Π ) and Σ (Σ T ), which are
simultaneously tangent to α, β and ζ. The spheres P (P Π ) and Σ (Σ T ) are the wanted ones. The proof is
similar to that of Proposition 11.2.1, q.e.d.
Theorem 11.2.2 The itnersection of a cylindrical surface K, with a plane which intersects all the genera-
tors of K, is an ellipse with foci at the contact points of the Dandelin’s spheres with the plane ε.
Proof: The proof results from the equality of the tangents from a point to a sphere (Theorem 9.11.2). If E
is an intersection point of ε with K, then the generator EZ will be simultaneously tangent to both spheres.
Also, if Π and T are the points of contact of the spheres P (P Π ) and Σ (Σ T ) with the plane ε, then EΠ
and ET are also respectively tangent to these spheres. It follows that |EΠ | = |EZ| and |ET | = |EH|,
where E and Z are the points of contact of the generator OE with the spheres. Then however the sum
|ET | + |EΠ | = |EZ| + |EH| = |ZH|,
is fixed and equal to the length of the generator of the cylinder which is formed from the circles κ and µ,
along which the two spheres are tangent to the cylindrical surface, q.e.d.
Proposition 11.2.3 Let the plane ε intersect all the generators of a conic surface K except two. There exist
then two spheres tangent to K and simultaneously tangent to the plane ε.
E
Π
Ρ
A
Z κ
Ζ'
Ω
O
Η' H μ
B
T
Figure 11.2.3: Spheres of Dandelin II
Proof: The proof is similar to that of Proposition 11.2.1. The plane of the paper in figure-11.2.3 is plane
θ containing the axis of the conic surface K and the orthogonal OΩ from the apex O of the conic surface
to the plane ε. Let ζ be the line intersection of the planes ε and θ. By construction, ζ is orthogonal to OΩ
and intersects the two generators of the conic surface contained in θ at points A and B. This defines then
on plane θ the triangle OAB and consequently its escribed circles with centers P and Σ . The spheres with
centers points P and Σ and radii those of the escribed circles respectively, P Π and Σ T , are the wanted
ones.
I show that one of them, P (P Π ) is tangent to the conic surface. Indeed, if OH is another generator and
Z is the projection of P to it, then the right triangles P ZO and P Z O, where Z is the projection of P on
OA, are congruent. This, because point P is on the axis of the cone, therefore has the same distance from
the generators, consequently |P Z| = |P Z |. Also the two triangles have the hypotenuse OP in common,
they are therefore congruent. This implies that OH is tangent to the sphere at point Z and proves the claim.
Similarly it is proved also that the sphere Σ (Σ T ) is tangent to the conic surface. The fact that these spheres
are also tangent to the plane ε follows directly, e.g. for P (P Π ), from the fact that P Π is also orthogonal
to the plane ε, as parallel to OΩ , which has this property. Consequently plane ε is tangent to the sphere
P (P Π ), being orthogonal to the end point of one of its radii, q.e.d.
11.3. DIRECTRICES 679
Theorem 11.2.3 The intersection points of the plane ε, which intersects all the generators of a conic sur-
face K except two, define a hyperbola on ε with foci at the contact points of the Dandelin’s spheres with
the plane ε.
Proof: The proof results from the equality of the tangents from a point to a sphere (Theorem 9.11.2). If E is
an intersection point of the plane ε and K, then the generator OE is simultaneously tangent to both spheres.
Also, if Π and T are the contact points of the spheres P (P Π ) and Σ (Σ T ) with the plane ε, then also EΠ
and ET are respectively tangents to these spheres. It follows that |EΠ | = |EZ| and |ET | = |EH|, where
E and Z are the contact points of the generator OE with the spheres. Then however the sum
|ET | − |EΠ | = |EH| − |EZ| = |ZH|,
is fixed and equal to the length of the generator of the truncated cone (union of the two cones with common
vertex O) formed from the circles κ and µ, along which the two spheres are tangent to the conic surface,
q.e.d.
11.3 Directrices
Man is a tool-using animal. Without
tools he is nothing, with tools he is all.
Thomas Carlyle, Sartor Resartus
Besides the parabola, in which the directrix plays an important role for its usual characterization as a
geometric locus, directrices are defined also for the other kinds of conic sections with properties similar
to that of the directrix of the parabola. In this section we relate first the directrix of the parabola with the
corresponding sphere of Dandelin and next we examine its generalization for the other conic sections.
Proposition 11.3.1 Let ε be a plane parallel to a tangent plane of a conic surface Σ . There exists then a
sphere tangent to Σ and simultaneously tangent to the plane ε.
Ζ' Υ
A
Ρ
Π β
α
Figure 11.3.1: Dandelin’s sphere for a plane parallel to a tangent plane of Σ
Proof: As in the propositions of the previous section, we consider the plane θ containing the axis of the
conic surface Σ and the orthogonal OΩ from the apex O of Σ to the given plane ε (in figure-11.3.1 θ
coincides with the plane of the paper). Planes θ and ε intersect along a line ζ. This line is parallel to the
generator β, which is contained in the tangent plane parallel to ε. Plane θ also contains the symmetric
generator α of β relative to the axis. Also, because ζ and β are parallel, the triangle with sides α, ζ and the
axis is isosceles. Thus, passing to the plane figure contained in θ, we find a circle P (P Π ) with center on
the axis and simultaneously tangent to the two parallels ζ and β, as well as to the other generator α of Σ
contained in θ. For this it suffices to consider the orthogonal to the axis OP from the intersection point of
α and ζ. The latter bisects the angle at A of lines α and ζ and defines the circle P (P Π ), simultaneously
tangent to α, β and ζ. The sphere P (P Π ) is the wanted one. The proof of this claim is the same with that
of Proposition 11.2.1, q.e.d.
Theorem 11.3.1 The intersection of a conic surface Σ and a plane ε, parallel to a tangent plane of Σ ,
define on ε a parabola with focus at the contact point Π of the Dandelin’s sphere with the plane ε and
directrix the line ξ, along which ε intersects the plane η of the circle of the contact points of the sphere with
the conic surface.
ε
O
ζ
ξ
Θ
Δ Z
η Γ
Π
μ
K H
υ E
Figure 11.3.2: Parabola and Dandelin’s sphere
Proof: The proof results from the equality of the tangents from a point to a sphere (Theorem 9.11.2). To
begin with, the circle, along which are tangent the sphere and the conic surface Σ , is contained in a plane η
orthogonal to the axis of the conic surface, which intersects plane ε along a line ξ (Figure 11.3.2). Line ξ
is orthogonal to the plane θ which contains the axis and the generator β = OΘ , which defines the tangent
plane of Σ , which, by assumption, is parallel to ε. From an arbitrary point E of the intersection of the
plane ε with Σ , passes a plane υ parallel to η and forming with that a truncated cone. Suppose EZ is the
generator of this cone which passes through point E and HΘ the equal to EZ generator contained in plane
θ. From the equality of the tangents from E to the sphere, follows that EZ and EΠ have the same length,
therefore EΠ and HΘ also have the same length. It suffices then to show that HΘ and EΓ have the same
length. Here EΓ is the orthogonal from E to ξ. Let line µ be the intersection of the planes ε and υ. Let also
K be the intersection of µ with the plane θ. Because planes η and υ are parallel, the lines ξ and µ, which are
intersected from them by the plane ε, are parallel. Because planes ε and η are orthogonal to θ, their inter-
section ξ will also be orthogonal to θ. Consequently, if ∆ is the intersection point of ξ with θ, then EΓ∆K
is a rectangle and K∆ and EΓ have the same length. Because however ζ and the generator OΘ are paral-
lel, K∆ and HΘ will also have the same length, as segments of parallel lines between parallel planes, q.e.d.
In what follows we show one more characteristic property of ellipses and hyperbolas in which appear
lines with similar properties to the directrix of the parabola.
Theorem 11.3.2 Given an ellipse, to each focus Π corresponds a line ξ such that, for every point E of the
Β1 Ι Β1 Ι
Η Η Δ
(Ι) (ΙΙ)
Α1 Ν Α1
Β E2 Ε Β E2 Ε
Α Ρ Μ Α
α α
Β2 Ο Α2 Β2 Ο Α2
Δ'
Κ
Α' Α'
Β' Β'
β β
Figure 11.8.30: Conjugate diameters Determination of axes
Exercise 11.8.36 Point A of a diameter AOA of an ellipse defines point A 1 of the major circle, through
its intersection with the orthogonal AA 2 on the major axis α. Subsequently A 1 is projected to point E 2 of
the minor axis β. This forms a rectangle OA 2 A1 E2 , where O is the center of the ellipse. This rectangle
is divided into four triangles: OA 2 A, OA1 , OA1 E and OEE2 , where E is defined through the ratio
|A1 E2 |
|EE2 | = b . We rotate the rectangle OA 2 A1 E2 by 2 so that it takes the position OHB 1 B2 (Figure
a π
11.8.30-I). Show that after the rotation the segment E 2 EA1 passes to the segment B2 BB1 , where BOB
is the diameter conjugate to AOA .
Hint: Application of Theorem 11.6.3 and of Theorem 11.6.1.
Exercise 11.8.37 Determine the direction and the length of the axes of an ellipse, for which are given, by
position and length, two conjugate diameters AOA and BOB .
Hint: Using the previous exercise and drawing the parallel and equal to OE, forms the parallelogram
OA∆E (Figure 11.8.30-II), whose sides are equal to half the given diameters, and OE is orthogonal to
OB. Therefore ∆ can be constructed from the givens. Let ∆ be the symmetric of ∆ relative to A. Then
∆ AEO is a parallelogram and O∆ is parallel to EA. It follows that ∆◊ OA = AEA
2
÷1 and A ◊2 OA1 =
÷ |A E
1 2 | |A A |
EA1 O. Because, by construction, |EE2 | = |AA2 | = b , it follows that AA1 EM is a rectangle. Therefore
1 2 a
÷1 = EA
AEA ◊ 1 M . This, because of the previous equality between angles, implies that OA 2 is a bisector
◊
of the angle ∆ OA , which can be constructed from the givens. This way the direction of the axes α and
1
β is determined. Because A1 is constructible, the length a = |OA 1 | determines the major axis and b is
|AA2 |
determined from a and the ratio |A 1 A2 |
= ab .
Exercise 11.8.38 For a given ellipse, show that the geometric locus of points A, from which tangents
√ drawn which make at A a right angle, is a circle with center the center of the ellipse and radius
can be
r = a2 + b2 , where a and b are the axes of the ellipse.
Hint: According to Theorem 11.6.2(6) the projections M , N of the foci Π , T on the tangent A∆ are
contained in the major circle σ(O, a) of the ellipse (Figure 11.8.31-I). Therefore the power of the point
A relative to this circle will be |AI| 2 = |AM ||AN | = |Π N ||T M | = b2 . Consequently |AO| 2 =
|AI|2 + |OI|2 = a2 + b2 .
The circle τ , defined in the previous exercise is called director circle of the ellipse. Considering the
symmetric Γ of A relative to O and drawing therefrom the tangents, we build a circumscribed rectangle of
the ellipse. Circle τ is also the locus of the vertices of all circumscribed rectangles of the ellipse.
Exercise 11.8.39 Show that, for a given parabola, the geometric locus of points A from which tangents can
be drawn, which intersect orthogonally at A is the directrix ξ of the parabola (Figure 11.8.31-II).
A
M I
(I) τ (II)
M' σ
B
a N ξ
N' Δ
Τ
Π A
B Ο
Γ
Γ
Figure 11.8.31: Director circle of ellipse Directrix of parabola
Hint: Follows from proposition 11.4.4, in combination with properties (5,6) of the Archimedian triangle of
Theorem 11.5.4.
Exercise 11.8.40 Construct 6 points of the hyperbola, which admits the triagle AOB as asymptotic with
its side AB tangent to the hyperbola.
(I) (II) Α
Β
α M2 E'
Ε Α'
Ο
Μ Δ
β Ζ Β' Γ'
M1
Z' Β Γ
Α
Figure 11.8.32: Hyperbola with given asymptotic Division into parts of equal area
Hint: One point of the hyperbola is the middle M of AB (Figure 11.8.32-I). Two additional points of the
hyperbola are the middles E , Z of the segments BE and AZ, where E, Z are the middles of the sides
of the parallelogram OA∆B. The symmetric of M , E , Z relative to O are also points of the hyperbola.
How many additional points of the hyperbola can be constructed from these six points?
Exercise 11.8.41 Show that every line, which divides the triangle ABΓ into two polygons of equal area,
is tangent to a hyperbola, which admits an asymptotic triangle consisting of two sides and a median of
the triangle, the median being tangent to the hyperbola. Every triangle defines this way three hyperbolas
Exercise 11.8.42 Let Γ be a point of the directrix ξ of the ellipse κ, relative to its focus A. Show that every
line ε through Γ intersecting the ellipse (conic) at the points {∆, E}, forms the angle ∆AE ’ at A which is
bisected by the polar p(Γ ) of Γ relative to the ellipse (conic).
Exercise 11.8.43 Given two concentric circles {κ(O, b), λ(O, a)} with a > b, show that the ellipse, which
has these as respective minor and major circles, has its foci {A, B} at the middles of the smaller sides of a
rectangle which is inscribed in λ and has its bigger sides tangent to κ (Figure 11.8.33-I). For which ellipses
the circle κ is tangent to all the sides of the rectangle?
(Ι) (ΙΙ) A
a
2a
A B
κ b
b
λ 2a B
Figure 11.8.33: Ellipse defined by a rectangle Cylinder cut for a given ellipse
Exercise 11.8.44 Given two positive numbers {a > b}, construct a cylinder and two equal spheres in-
scribed in it, such that the intersection of the cylinder with a plane tangent to both spheres is an ellipse with
axes {a, b} (Figure 11.8.33-II).
Exercise 11.8.45 Given are the circle κ, the constant k > 0, the angle ω and a point A outside the circle.
For every point B of the circles, consider the point Γ of the segment AB, for which |AΓ |/|Γ B| = k and
the line εB passing through Γ and making the angle ω with AB (Figure 11.8.34). Show that all these lines
Ζ'
Ρ
Θ
Μ
Ε'
λ Ε A
Ζ
Ο
Κ
Δ ω
Η
Γ
κ
B εΒ
Figure 11.8.34: Hyperbola generated from a variable angle
εB are tangent to a hyperbola, of which we request the location of the foci and the asymptotes.
Hint: Consider the symmetric ∆ of A relative to the line ε B and show that the triangle AB∆ has constant
angles, hence ∆ is contained in a circle λ(P ) (Theorem 3.10.3) and the lines ε B coincide with the medial
lines KΓ of the segments A∆, hence they are tangent to a hyperbola (Exercise 11.7.1), which has its foci at
{A, P }. The asymptotes of the hyperbola result from the positions which takes line ε B , when B coincides
with the contact points of the tangents to κ from A.
Many properties of the conics can be studied in a synthetic geometric way, as we did in the previous
exercises ([Ask03], [Bes95], [CrFW91], [Tay81]). Many more, however, can be studied using the meth-
ods of analytic and projective Geometry ([Cha65], [Sal17]). Historically, the geometric method came first
([Coo68]). The methods of analytic and projective geometry followed, which eventually led to, the today
dominant, methods of Algebraic Geometry ( [Per08]).
CHAPTER 12
Transformations in space
12.1 Isometries in space

And the last, to perform enumerations everywhere so complete, and
reviews so general, that I am certain I don’t exclude anything.
Descartes, Discourse on the Method, II
Transformations in space, analogously to those on the plane (§ 7.1), are processes through which to every
point X of space corresponds another point Y of space, which we denote by f (X). The terminology here,
Image, Prototype, Maps, Domain, Range, Composition of Transformations, ..., etc. is the same with that of
the aforementioned section and I do not repeat it.
Analogous is also the definition of the isometry. Formally, it is the same with that of the plane. The
only thing that changes is the location of the points, which now are in space. We call then isometry of
space, a transformation f in space which preserves the distances between points, in other words, for every
pair of points in space X, Y and their images X = f (X), Y = f (Y ) holds
|X Y | = |XY |.
We can immediately see some consequences of the definition, completely analogous to similar properties
of the plane, some of which I present as exercises.
Theorem 12.1.1 An isometry maps a triangle ABΓ to a congruent triangle A B Γ , an angle to an equal
angle, a line to a line and a plane to a plane.
Proof: The first part is obvious, because by definition the isometry preserves the lengths of line segments,
therefore the two triangles will have equal respective sides. The same way it is proved also that an angle
’ maps to an equal angle X
XOY ◊ O Y . For the second part we show first that the isometry maps lines to
lines. This follows directly from the previous part. Indeed, if X is a point of a line BΓ and point X is
not on the corresponding line B Γ (Figure 12.1.1), then a genuine triangle B X Γ is created and for its
lengths will hold
|B X | + |X Γ | = |BX| + |XΓ | = |BΓ | = |B Γ |,
which contradicts the triangle inequality. Thus, the isometry maps lines to lines. From this follows that it
also maps a plane ε onto another plane ε . Indeed if AB, AΓ are two lines of plane ε and A B , A Γ are
their images, then for every other point Y of the plane ε we consider a line through Y which intersects AB,
734 CHAPTER 12. TRANSFORMATIONS IN SPACE
Υ'
Α' Ε Γ' ε'
Δ
Α Υ
Ε
Δ Β'
Χ'
Β Γ
Χ ε
Figure 12.1.1: Isometry preserves length, angles, lines, planes
AΓ respectively at points ∆, E. The images ∆ , E of these points through the isometry will be contained,
according to the previous, respectively in lines A B , A Γ , therefore the line ∆ E which contains Y will
also be contained in the plane of A B , A Γ , q.e.d.
Exercise 12.1.1 An isometry, which leaves two different points A, B fixed, leaves fixed also all the points
of the line AB.
Exercise 12.1.2 An isometry, which leaves three non-collinear points fixed, leaves fixed also all the points
of the plane, which passes through these three points.
Exercise 12.1.3 An isometry, which leaves four non-coplanar points fixed, leaves fixed also all the points
of space, in other words it is the identity transformation e of space.
Exercise 12.1.4 Two isometries coincident at four non-coplanar points are coincident everywhere.
Exercise 12.1.5 An isometry f preserves parallel lines and planes, as well as parallel planes. In other
words, if α, ε, ε are respectively a line and two planes, then if α is parallel to ε then f (α) will also be
parallel to f (ε). Also, if the planes ε and ε are parallel, then the planes f (ε) and f (ε ) will also be
parallel.
Exercise 12.1.6 An isometry f preserves the angle between a line α and a plane ε.
Exercise 12.1.7 An isometry f maps a sphere Σ onto a sphere Σ of equal radius. It also maps a circle κ
onto a circle κ of equal radius. If circle κ is contained in Σ , then κ is also contained in Σ .
Exercise 12.1.8 An isometry f maps a dihedral angle onto a congruent dihedral and a tetrahedron onto a
congruent tetrahedron. More generally, it maps a polyhedron onto a congruent polyhedron.
As on the plane (§ 1.7), so in space we also have a notion of orientation, which can be defined through
tetrahedra and depends on the order we use to place letters on the vertices. The tetrahedron ABΓ∆ is
called positively oriented, if after placing a clock on the plane ABΓ with its face towards ∆ the direction
A → B → Γ is the same as that of its hands. If the direction A → B → Γ is opposite to that of the hands
of the clock, then the tetrahedron is called negatively oriented. We say that the transformation f reverses
the orientation of a tetrahedron, when the tetrahedron A B Γ ∆ , which has vertices the images of the
vertices of the tetrahedron ABΓ∆ through f has opposite orientation from that of ABΓ∆. If the two tetra-
hedrons ABΓ∆ and A B Γ ∆ have the same orientation, then we say that f preserves their orientation.
The transformations we consider in this chapter (are, as it is called, “continuous” and) have the property
that, if they preserve the orientation of a tetrahedron, then they will preserve the orientation of every other.
Respectively, if they reverse the orientation of a tetrahedron, they will reverse the orientation of every other
tetrahedron. It suffices then to examine what happens to the orientation of one and only tetrahedron, in order
to infer from it, the answer to whether the specific transformation preserves or reverses the orientation.
For the transformations which preserve the orientation of tetrahedra we also say often that they preserve the
12.2. REFLECTIONS IN SPACE 735
orientation of space, while for those which reverse the orientation of tetrahedrons we say that they reverse
the orientation of space.
12.2 Reflections in space

And that can happen, you know, with the cleverest man, the
psychologist, the literary man. The temperament reflects
everything like a mirror! Gaze into it and admire what you see!
F. Dostoevsky, Crime and punishment
A plane ε defines a transformation in space, called reflection or mirroring in ε, or relative to ε. Often ε is

called the mirror of the reflection. This is the analogue of reflections in the plane, relative to lines (§ 7.2).
This reflection corresponds to every point X of the space: a) X itself, if the point is contained in the plane
Μ
ε
Υ
Figure 12.2.1: Reflection in space
ε, b) point Y of space, such that ε will be the medial plane of XY , if X is not contained in ε. Equivalently,
if M is the projection of X onto ε, then Y will be in the extension of XM towards M and at equal distance
with X from M .
Theorem 12.2.1 A reflection in a plane ε is an isometry of the space, which reverses the orientation of
space.
Χ Δ
(Ι) Y (ΙΙ)
A
Β
M N Γ
ε B ε Α
Y'
X' Δ'
Figure 12.2.2: Reflections in planes are isometries
Proof: Let us assume that the points X, Y in space are not contained in plane ε, which defines the reflection,
and X , Y are their images (Figure 12.2.2-I). Then, by definition the line segments XX , Y Y will be
orthogonal to ε, therefore parallel and they will define the trapezium XY Y X . We show that this trapezium
is isosceles. For this, we draw parallels from the middle N of Y Y respectively to Y X and Y X , which
intersect XX respectively at points A and B. This creates the parallelograms XAN Y and X BN Y , for
which hold |AX| = |Y N | = |N Y | = |BX |. However, if M is the middle of XX , then |M X| = |M X |
and consequently that |AM | = |M B|. Also AB is, by its definition, orthogonal to M N , therefore the
triangle AN B is isosceles and consequently we have |XY | = |AN | = |BN | = |X Y |. The proof for the
case when one or both points X, Y are contained in ε or in the case where the isosceles degenerates to the
line segment M N is even simpler.
That the reflection reverses the orientation, is seen easily by taking a tetraheder ABΓ∆, of whose the
basis ABΓ is contained in the plane ε (Figure 12.2.2-II). By the reflection in ε, this tetraheder maps to the
tetraheder ABΓ∆ , where ∆ is the symmetric of ∆ and we see immediately that these two tetraheders have
opposite orientations, q.e.d.
Proposition 12.2.1 For every reflection f holds f ◦ f = e, in other words, the inverse of a reflection is the
reflection itself.
Proof: The same as that of Proposition 7.2.1, q.e.d.
Proposition 12.2.2 If an isometry in space f , other than the identity, satisfies the relation f ◦ f = e and
admits at least three non-collinear fixed points, then it coincides with the reflection relative to the plane
defined by these three points.
Exercise 12.2.1 Show that an isometry in space, which leaves fixed the points of a plane ε, and only these,
coincides with the reflection relative to ε.
By analogy to the shapes of the plane, a shape in space Σ is called symmetric relative to the plane ε, when
the reflection f relative to ε maps Σ to itself (f (Σ ) = Σ ). A plane ε, relative to which the shape Σ is
symmetric, is called plane of symmetry of Σ . The most symmetric shape in space is the sphere.
Exercise 12.2.2 Show that every plane, that passes through the center of the sphere, is a plane of symmetry
of the sphere. Show that shapes Σ , for which there exists a point O, such that every plane through O is a
plane of symmetry of Σ , are unions of spheres centerred at O.
Exercise 12.2.3 How many and which are the planes of symmetry of a specific plane, or a dihedral angle?
Exercise 12.2.4 When does a trihedral angle has a plane of symmetry?
Exercise 12.2.5 Show that, if a triheder has two planes of symmetry, then it also has a third and, conse-
quently, has congruent faces and congruent dihedral angles.
Exercise 12.2.6 How many and which planes of symmetry has the cylinder and the cone?
Exercise 12.2.7 Show that, for two spheres of equal radii, there exists a reflection which maps the one onto
the other.
As on the plane, so also in space a point O defines a point symmetry in O, or with respect to O, or relative
to O. This transformation, to the point O corresponds O itself, and to every other point X = O corresponds
point Y , so that O will coincide with the middle of XY . We say that Y is the symmetric of X relative to
O. The symmetry f relative to O is, like the reflection, its own inverse:
f ◦ f = e.
A shape Σ is called symmetric in O, or relative to point O, when the symmetry f relative to O maps the
shape to itself (f (Σ ) = Σ ). We also say then that O is a center of symmetry of the shape. The typical
example is again the sphere.
Exercise 12.2.8 Show that in a line and in a plane any of their points is a center of symmetry for them.
Exercise 12.2.9 Which, among the Platonic solids, admits a center of symmetry?
12.2. REFLECTIONS IN SPACE 737
Figure 12.2.3: Axial symmetry in space
Exercise 12.2.10 Show that the symmetry relative to a point is an isometry.
In space we also have a kind of symmetry which is defined through a line α. We call axial symmetry
relative to the line α in space the transformation which: a) to every point X of the line α corresponds X
itself and b) to every point X outside the line α corresponds point Y , so that α will be the medial line of
XY . In other words point Y results by projecting point X onto α at M and extending XM towards M to
the double distance of |XM |.
Every plane ε, orthogonal to the line α, maps through the axial symmetry relative to α to itself and
consequently this defines a transformation of this plane to itself, which coincides with the symmetry relative
to M , where M is the intersection point of the plane ε with the line α. The sphere admits every line through
its center as an axis of symmetry.
Exercise 12.2.11 Show that a line, a plane and a sphere have infinitely many axes of symmetry.
Exercise 12.2.12 Find the axes of symmetry of a rectangular parallepiped.
Exercise 12.2.13 Which Platonic solids admit axes of symmetry and how many?
Exercise 12.2.14 Show that the axial symmetry relative to the line α is equal to the composition of two
reflections relative to two orthogonal planes passing through α.
Hint: See also the proof of Theorem 7.2.3.
γ
Ζ
O Υ α
Χ
β
Figure 12.2.4: Composition of axial symmetries of orthogonal axes
Exercise 12.2.15 Show that the composition of two axial symmetries relative to axes α and β, which inter-
sect orthogonally at O is the axial symmetry relative to the axis γ, which is orthogonal to the plane of α, β
at O.
Exercise 12.2.16 Show that the composition of three reflections, relative to three planes, which pass through
point O and intersect pairwise orthogonally, is the symmetry relative to O.
12.3 Translations in space

At all times and in all places numerous men have existed who “did not
believe in anything,” precisely because “they did not question anything”;
living, for them, meant simply abandoning oneself from one moment to the
next, without any inner response or position in the face of any dilemma.
Ortega y Gasset, The origin of philosophy
As in the plane (§ 7.3), so in space, an oriented line segment AB defines the transformation of translation
by AB. This one corresponds, to every point X in space, a point Y , such that XY and AB are equal, parallel
and equally oriented line segments. Equivalently: ABY X is a parallelogram. Here too we consider the
identity transformation as a null translation, i.e. a translation by a segment, whose end points coincide.
All the properties of the translations of the plane, which are examined in the aforementioned section, can
be carried over almost verbatim, along with their proofs, into corresponding properties of translations in
space. I include some of them as exercises.
Exercise 12.3.1 Every translation is an isometry of space (Theorem 7.3.1).
Exercise 12.3.2 The composition of two or more translations is again a translation (Theorem 7.3.2, Corol-
lary 7.3.1).
Exercise 12.3.3 The composition of translations along the oriented sides A 1 A2 , A2 A3 , ..., Ak−1 Ak , Ak A1
of a (skew) closed polygon A 1 A2 ...Ak is the identity transformation in space (Corollary 7.3.2).
Exercise 12.3.4 The composition of two reflections relative to two planes, whose distance is δ, is a trans-
lation along the line segment AB of length 2δ and direction orthogonal to the two planes (Theorem 7.3.3).
Exercise 12.3.5 The composition of two symmetries relative to axes, which are parallel and at distance δ,
is a translation by a line segment AB of the plane of the parallels, orthogonal to the parallel axes and of
length 2δ.
Ζ Y
(I) a a (II)
O' a O'
a/2 O
Υ 2a
Ο X Z
Χ
Figure 12.3.1: Symmetry and translation Symmetry and symmetry
Exercise 12.3.6 The composition of a symmetry relative to a point O and a translation α, is a symmetry to
a point O (Theorem 7.3.4).
Exercise 12.3.7 The composition of two symmetries relative to two different points O, O is a translation
by the double of OO (Theorem 7.3.5).
Exercise 12.3.8 The composition of ν symmetries relative to ν points A 1 , A2 , ..., Aν is, for even ν, a
translation and for odd ν a symmetry (Corollary 7.3.3).
Exercise 12.3.9 Given ν different points A 1 , A2 , ..., Aν , there exists exactly one (skew) polygon which has
these points as middles of its successive sides if ν is odd. If ν is even, in general, there is no such polygon.
If however there is one, there are infinitely many and in fact each point in space may be considered a vertex
of such a polygon (Theorem 7.3.6).
Exercise 12.3.10 Given the tetraheder ABΓ∆, determine the result of the composition of symmetries f ∆ ◦
fΓ ◦ fB ◦ fA on its vertices. How the order we take the vertices influences the result of the composition?
12.4. ROTATIONS IN SPACE 739
12.4 Rotations in space

All that is distinctive of man, marking him off from the clay he walks upon or the potatoes
he eats, occurs in his thought and emotions, in what we have agreed to call consciousness.
J. Dewey, The Role of Philosophy in the History of Civilization
The main difference between rotations in space and those of the plane (§ 7.4) is that the former have an axis
of rotation instead of a center of rotation. There exists, in other words, an entire line, which remains fixed
during the rotation (its points are fixed points of the transformation). We also consider this line (axis) as
being oriented, by choosing (arbitrarily) one of its directions as positive and its opposite as negative. The
two notions of rotation, of plane and space, are however intimately related. One may say that the rotation
in space is a kind of extension of the rotation on the plane. This is reflected in the property of a rotation in
space, according to which, it defines a plane rotation on every plane orthogonal to its axis.
Μ
Χ Υ
ω
Figure 12.4.1: Rotation in space
We call then rotation in space relative to the oriented line or axis α and oriented angle ω, the transforma-
tion which is defined as follows: a) The points of the axis remain fixed. b) For every point X outside the
axis, we draw the plane ε through X, which is orthogonal to the axis and intersects it at point M . We choose
its side (ε+ ), which sees towards the positive direction of the axis. On this plane we define as positive the
orientation of rotation which is opposite to the direction of the clock, the face of which sees towards the
positive direction of the axis. To X the transformation corresponds Y , so that for the oriented angle holds
(XM Y ) = ω and also |M Y | = |M X|. In other words, after we determine the orientation in ε, we rotate
X by applying to it the plane rotation of ε, with center M and angle ω.
Comment-1 In the case where the rotation angle is |ω| = π the rotation relative to axis α coincides with
the symmetry relative to axis α. In every other case, of a rotation f by an angle which is not an integral
multiple of π, changing, if necessary, the direction of the axis, we can define a rotation g, which gives the
same result as f , for every point in space, and its angle is 0 < (ω) < π.
Ζ
ε Υ
Χ
ω/2
η
ζ
Figure 12.4.2: Rotation as composition of two reflections
As in the case of the plane, so in space the rotation is a composition of two reflections, this time relative to
two planes ζ, η. The next proposition is proved exactly the same way as the corresponding one for planes
(Proposition 7.4.2).
Proposition 12.4.1 The composition of two reflections f = h ◦ g, whose planes intersect along the line α
making a dihedral angle of oriented measure ω with |ω| ≤ π2 , is a rotation with axis α and angle of rotation
2ω.
Theorem 12.4.1 Every rotation in space is an isometry in space.
Proof: A simple proof follows from the previous proposition and the fact that reflections relative to planes
are isometries in space, q.e.d.
Proposition 12.4.2 The composition of two rotations with the same axis η and angles α and β is a rotation
with the same axis and angle of rotation α + β.
Proposition 12.4.3 Every rotation f in space, relative to the axis α, defines on every plane ε, orthogonal
to the line α a transformation f ε of this plane. The transformation f ε , of plane ε on itself, coincides with
the rotation of the plane ε with center the point of intersection of ε with α and angle the rotation angle of
f.
Proof: Indeed, from the definition follows that f maps every plane ε orthogonal to the axis α to itself,
leaving fixed the intersection point O of ε and α. It also defines an orientation of the plane considering as
positive rotation the direction which is opposite to that of the clock, which has been placed on the plane
with its face towards the positive direction of the axis α. f then defines a transformation f ε of this plane.
This transformation, according to the definition of rotation f , defines on ε exactly the same correspondence
which is effected also by the rotation of ε relative to O and by the angle of f , q.e.d.
Proposition 12.4.4 Every rotation f ε , defined initially only for points of the plane ε, can be extended to a
rotation f of the space, relative to axis α coinciding with the line which is orthogonal to ε at the center O
of the rotation f ε and angle that of f ε .
Proof: This proposition, which is the converse of the previous one, presupposes that the plane ε is oriented.
In other words one of the sides of the plane, ε + has been chosen and on it has been defined the positive
rotation angle. The orientation of ε defines an orientation on the line α considering as positive the one
towards the chosen side ε + . Under these assumptions, f is defined as the rotation in space about axis α
with angle that of f ε , q.e.d.
Proposition 12.4.5 The composition of two rotations relative to the axes α, β, which pass through point O,
is a rotation relative to an axis γ which also passes through O.
β α
γ
Β
Α Γ
Figure 12.4.3: Composition of rotations with intersecting axes
Proof: Let us represent the first rotation f as a composition of two reflections f ζ , fη relative to planes ζ, η
which intersect along α. We can choose these planes so that the second η coincides with the plane of lines
1. Two kinds of polygons on each vertex ii. Triangles + 2 n-gons

(a) Triangles + squares (e) Squares + n-gons (n ≥ 5)
i. Triangles + 1 square only i. Squares + 1 n-gon only
ii. Triangle + 2 squares ii. Squares + 2 n-gons
iii. Triangles + n > 2 squares (f) Pentagons + n-gons (n ≥ 6)
(b) Triangles + pentagons i. Pentagons + 1 n-gon only
i. Triangles + 1 pentagon only ii. Pentagons + 2 n-gons
ii. Triangles + 2 pentagons
(c) Triangles + hexagons 2. Three kinds of polygons on each vertex
i. Triangles + 1 hexagon only (a) Triangles + squares + n-gons
ii. Triangles + 2 hexagons (b) m-gons + n-gons + p-gons, without tri-
(d) Triangles + n-gons (n ≥ 7)
i. Triangles + 1 n-gon only angles
Figure 12.8.6: Archimedean solids-I
• (1.α . i) The square has angles of 90 ◦ and consequently it can coexist with, at most 4 triangles.
Otherwise, the total sum of the faces around the vertex will exceed 360 ◦ . The possible cases are:
(3, 3, 3, 3, 4), (3, 3, 3, 4) and (3, 3, 4). From these the last one is impossible (restriction 12.8.3). The
other two are possible and are realized in the so-called truncated cube and the square antiprism
respectively.
• (1.α , ii) The two squares around a vertex have sum of angles 180 ◦ consequently they can coexist
with at most 2 triangles. This results in cases (3, 3, 4, 4), (3, 4, 3, 4) and (3, 4, 4). The case (3, 4, 3, 4)
is rejected because of restriction 12.8.2. The other two are possible and realize in the so-called
cubeoctahedron and the regular triangular prism respectively.
• (1.α , iii) Three squares around a vertex have sum of angles 270 ◦ consequently they can coexist with
only one triangle. The only possibility then in this category is (3, 4, 4, 4), which is realized in the
so-called rhombicuboctahedron.
• (1.β , i) The pentagon with angle 108 ◦ can coexist with 4, at most triangles. Otherwise, the total sum
of the faces around the vertex will exceed 360 ◦. Then result the possibilities (3, 3, 3, 3, 5), (3, 3, 3, 5),
(3, 3, 5), of which the last one is impossible according to restriction 12.8.3. The other two possibilities
are realized in the so-called truncated dodecahedron and the pentagonal antiprism.
• (1.β , ii) Two pentagons around a vertex have sum of angles 216 ◦ and can coexist with 2, at most,
equilateral triangles. There result the possibilities (3, 3, 5, 5), (3, 5, 3, 5), (3, 5, 5), of which the last
one is impossible according to restriction 12.8.3. The first one is also rejected because of restriction
12.8. ARCHIMEDEAN SOLIDS 755
12.8.2. The only possibility in this category is, therefore, (3, 5, 3, 5), which is realized in the so-called
icosidodecahedron.
• (1.γ , i) The hexagon has angles of 120 ◦ and can coexist with 3, at most, equilateral triangles around
a vertex. There result the possibilities (3, 3, 3, 6) and (3, 3, 6), of which the last one is impossible
according to restriction 12.8.3. The only possibility in this category then is (3, 3, 3, 6), which is
realized in the hexagonal antiprism.
Figure 12.8.7: Archimedean solids - II
• (1.γ , ii) Two hexagons, with sum of angles 240 ◦ can coexist around a vertex with only one equilat-
eral triangle. The only possibility in this case then, is, (3, 6, 6), which is realized in the truncated
tetrahedron.
• (1.δ , i) The case of one exactly n-gon and triangles, with n ≥ 7 and respective angles greater than
128◦ allows the coexistence with 3, at most equilateral triangles around a vertex. The possibilities,
consequently, in this category are (3, 3, 3, n), (3, 3, n), of which the second is impossible because of
restriction 12.8.3 and the first one is realized in the anriprism with base the regular n-gon.
• (1.δ , ii) The case of exactly two n-gons and triangles, with n ≥ 7 and respective angles greater than
128◦ allows the coexistence with one, at most, equilateral triangle around a vertex. The possibilities,
then, in this category correspond to the symbol (3, n, n), which, for odd n are excluded, because of
restriction 12.8.3. Also, even numbers greater than 10 are excluded, since then the angle sum around
a vertex exceeds 360 ◦ . Then, in this category, remain the possibilities (3, 8, 8) and (3, 10, 10), which
are realized respectively in the truncated cube and the truncated dodecahedron.
• (1. , i) The case of squares and exactly one n-gon around a vertex, with n ≥ 5 is (4, 4, n), which
corresponds to the regular prisms with base a regular n-gon.
• (1. , ii) The case of squares and two exactly n-gons around a vertex, with n ≥ 5 allows only one
square and only solids with symbols of type (4, n, n). These for n ≥ 8 are rejected because they give
a sum of angles around the vertex greater than 360 ◦. Also the type (3, 7, 7) is rejected because of
restriction 12.8.3. The only case in this category then is (4, 6, 6), which is realized in the truncated
octahedron.
• (1.στ , i) The case of pentagons and exactly one n-gon, with n ≥ 6 around a vertex, is rejected
already from the respective type (5, 5, n) because such polygons give a sum of angles around the
vertex greater than 360 ◦ .
• (1.στ , ii) For the same reason the case of pentagons and exactly two n-gons, with n ≥ 6, is possible
only for the solid with corresponding symbol (5, 6, 6), which is realized in the truncated icosahe-
dron. The previous categories exhaust all the cases for vertices of semi-regular polyhedra where
exactly two kinds of polygons show up.
Figure 12.8.8: Archimedean solids - III
• (2.α ) This category includes polyhedra around whose vertices show up three kinds of polygons:
triangles, squares and n-gons with n ≥ 5. We again consider the cases:
1. Only a square ⇒ and at most 2 triangles. The symbols (3, 3, 4, n), (3, 4, n) result, which are
rejected, because of the restrictions 12.8.2 and 12.8.3 respectively.
2. Exactly 2 squares ⇒ and at most 1 triangle. The symbols (3, 4, 4, n), (3, 4, 5, 4) result, of which
the first is rejected, because of the restriction 12.8.2. The symbol (3, 4, 5, 4) is realized in the
so-called rhombicosidodecahedron.
• (2.β ) This last category includes polyhedra around the vertices of which show up three kinds of
polygons, but not triangles. We see immediately that the symbol (4, 4, 5, 6) defines angles around a
vertex, whose sum is greater than 360 ◦ . It follows that the polygons around one vertex must all be dif-
ferent and according to the restriction 12.8.3, none of them contains an odd number. There result the
possibilities (4, 6, 8) and (4, 6, 10), which are realized, respectively, in the truncated cuboctahedron
and truncated icosidodecahedron.
12.9. EPILOGUE 757
The previous analysis exhausts all possibilities of convex polyhedral angles, which may show up in semi-
regular polyhedra, and shows that no more than the already mentioned exist, q.e.d.
Comment-1 The theorem deals with the determination of all possible semi-regular polyhedra. It does
not address, however, the problem of existence. To take care of this matter, one must, like Euclid does for
the Platonic solids, construct each solid, relying on the length of its edge or the radius of its circumscribed
sphere. For most constructions, and specifically these which are characterized by the adjective truncated,
but also for the cuboctahedron and icosidodechedron, the possibility of reduction to the construction of
a Platonic body exists. For example, the truncated terahedron results by cutting off in a simple way the
vertices of a regular tetrahedron. Also the truncated cuboctahedron results by cutting off vertices of the
cuboctahedron, which results similarly from the cube.
Comment-2 Two of the Archimedean solids, the truncated cube and the truncated dodecahedron show
up in two, as they are called, enantiomorphic forms, which are nothing but their reflected images relative
to a plane. According to our discussion for congruence in space (§ 12.5), the two enantiomorphic forms of
each of these solids are isometric, but they cannot be placed the one on top of the other.
Figure 12.8.9: Fullerene C 60 , carbon atoms distributed on the vertices of a truncated icosahedron
Comment-3 The Archimedean solids have come again to the foreground with the discovery in Chemistry
(1996) of the fullerens, which are carbon molecules, whose atoms are on the vertices of polyhedra. The
most well known is C60 , which corresponds to the truncated icosahedron ([FC11]).
12.9 Epilogue
Books have the same enemies as men: fire,
humidity, nonsense, time, and their own content.
Paul Valery, Thoughts and Aphorisms
Looking back, at the material we discussed, in its entirety, we realize that most of the times we were con-
cerned, not so much with properties which remain invariant relative to isometries, but rather with properties
which remain invariant relative to similarities. Even measures of magnitudes which are preserved by isome-
tries, like lengths, areas and volumes, change through similarities in a simple way, by multiplying the initial
magnitudes respectively with κ, κ 2 , κ3 , where κ is the ratio of similarity. This way, one could express
the general rule, that Euclidean Geometry has as its object the set of properties of shapes in space, which
remain invariant relative to similarities in space ([Yag62, p. 4,II]). There exist many interesting articles and
books on the evolution of the idea of “Geometry” and the meaning of the word Geometry. I will mention
one of the most important ones, that of S.S. Chern ([Che90]), wherein are contained references to relevant
articles, as well as the books of Berger [Ber10] and of Scriba and Schreiber [CS05].
Regarding the infrastructures, on which the discussed material relies, these are to be found in many
places. To begin with, the basic structure in the entire system of Euclidean geometry is the one of the set of
real numbers. Many of the weaknesses of Euclid’s Elements stem from the fact, that the precise description
of the structure of real numbers has been achieved after more than two milenia have been elapsed. The
euclidean lines are, in essence, copies of the set of real numbers. A euclidean line is a “model” of the set of
real numbers, and inversely, the set R of real numbers is an abstract mathematical model of the euclidean
line. On this bidirectional correspondence relies the idea of the “analytic geometry” of Descartes, which,
taking two orthogonal lines on the plane, represents it with the product R × R. The structure of “vector
space” emerges then at once. A structure we have been tacitly using all the time, working with directed
segments AB, broken lines, parallelograms, polygons, middles of sides, centroids, which are notions related
to “sums” of such directed segments = “vectors”.
Another domain sustained by a multitude of infrastructures, is the one of relations and equations, which
result in almost every theorem and exercise of the euclidean geometry. What are these equations? Which
of them can be solved? Why there are solvable and non-solvable equations? How can be proved the
unsolvability? Questions of this kind are related also to the “constructibility” problem of a segment of√a
given length with the exclusive use of the rule and compass. Whether, for example, a segment of length π
or the side of the regular n−gon can be constructed. These questions lead to algebra, the “theory of fields”
and the “theory of groups”, which in turn lead to a multitude of other algebraic structures.
We encountered the notion of “limit” when trying to measure areas/volumes of polygons/polyhedra and
lengths of simple curves, such as the circle. The “length”, the “area”, the “volume”, in combination with the
“limit”, lead to the “theory of measure”. Underneath the notion of “limit” leaves the whole “infinitesimal
calculus”.
Finally another domain, which leads to substantial structures, is the one of the “symmetries of a shape”.
The underlying structure here is that of the “group”. Many of the properties of the transformations, we have
studied, in “translations”, “rotations”, “similarities”, can be expressed in a unified way in the language of
the “theory of groups”. For example, the restriction of the different kinds of the platonic or/and archimedean
solids to a small number, results from the fact that the group of isometries of the sphere has few concrete
finite “subgroups”. Thus, in this book, we make a good start. In some cases, underneath a simple theorem
or exercise, we can find, if we dig a bit deeper, an important theory, a new world.
Bibliography
[AA06] Andreescu, Titu and Dorin Andrica: Complex Numbers from A to ... Z. Birkhaeuser, Berlin,
2006.
[Aar08] Aarts, J.M.: Plane and Solid Geometry. Springer Verlag, Berlin, 2008.
[AB83] A. Bruen, J. Fisher, J. Wilker: Apollonius by Inversion. Mathematics Magazine, 56:97–103,
1983.
[Adl06] Adler, August: Theorie der Geometrischen Konstruktionen. Goeschensche Verlagshandlung,
Leipzig, 1906.
[Alp00] Alperin, Roger: A Mathematical Theory of Origami Constructions and Numbers. New York
Journal of Mathematics, 6:119–133, 2000.
[AO12] Alexander Ostermann, Gerhard Wanner: Geometry by its history. Springer, Berlin, 1912.
[AP88] Agazzi, Evandro and Dario Palladino: Le Geometrie non Euclidee e i Fondamenti della Ge-
ometria. La scuola, Brescia, 1988.
[Arn04] Arnlold, Vladimir: Problems for children from 5 to 15. MCCME, Moscow, 2004.
[Arn05] Arnlold, Vladimir: Arnold’s Problems. Springer, Berlin, 2005.
[Ask03] Askwith, E.H.: A course of pure Geometry. Cambridge University Press, Cambridge, 1903.
[Aud02] Audin, Michele: Geometry. Springer Verlag, 2002.
[Aym04] Ayme, Jean Louis: A purely Synthetic Proof of the Droz-Farny Line Theorem. Forum Geomet-
ricorum, 4:219–224, 2004.
[Bac73] Bachmann, Friedrich: Aufbau der Geometrie aus dem Spiegelungsbegriff. Springer Verlag,
Heidelberg, 1973.
[Bai88] Bailey, David: The computation of π to 29,360,000 Decimal Digits Using Borweins’ Quarti-
cally Convergent Algorithm. Mathematics of Computation, 50:283–296, 1988.
[Bak85a] Baker, Marcus: A collection of formulae for the area of a plane triangle. Annals of Mathemat-
ics, 1:134–138, 1885.
[Bak85b] Baker, Marcus: A collection of formulae for the area of a plane triangle. Annals of Mathemat-
ics, 2:11–18, 1885.
[Ban88] Bankoff, Leon: The Metamorphosis of the Butterfly Problem. Mathematics Magazine, 60:195–
210, 1988.
760 BIBLIOGRAPHY
[Bea83] Beardon, Alan: The Geometry of Discrete Groups. Springer, Berlin, 1983.
[Bel07] Belyaev, Oleg: Fundamentals of Geometry. Preprint (Internet), 2007.
[Ben07] Benko, David: A New Approach to Hilbert’s Third Problem. The American Mathematical
Monthly, 114:665–676, 2007.
[Ber87] Berger, Marcel: Geometry vols I, II. Springer Verlag, Heidelberg, 1987.
[Ber02] Berger, Marcel: A Panoramic View of Riemannian Geometry. Springer Verlag, Heidelberg,
2002.
[Ber10] Berger, Marcel: Geometry Revealed. Springer Verlag, Heidelberg, 2010.
[Bes95] Besant, W. H.: Conic Sections Treated Geometrically. George Bell and Sons, London, 1895.
[BH07] Barker, William and Roger Howe: Continuous Symmetry, From Euclid to Klein. American
Mathematical Society, 2007.
[Bir32] Birkhoff, George: A set of postulates for plane Geometry, Based on Scale and Protractor.
Annals of Mathematics, 33:329–345, 1932.
[Bir15] Birsan, Temistocle: Bounds for Elements of a Triangle Expressed by R,r and s. Forum Geo-
metricorum, 15:99–103, 2015.
[Bla36] Blaschke, Wilhelm: Kreis und Kugel. Walter de Gruyter, Berlin, 1936.
[Bly00] Blythe, W.: Geometrical Drawing. Cambridge University Press, Cambridge, 1900.
[BM57] B., Argunov and Balk M.: Geometritseskie Postroenia na ploskosti (Russian). Gasudarstvenoe
Isdatelstvo, Moscva, 1957.
[Boa06] Boas, Harold: Reflections on the Arbelos. The American Mathematical Monthly, 113:236–
249, 2006.
[Boc15] Bocher, Maxime: Plane Analytic Geometry. Henry Holt and Company, New York, 1915.
[Bol82] Bold, Benjamin: Famous Problems of Geometry. Dover Publications, Inc., New York, 1982.
[Bon12] Bonola, Roberto: Non-Euclidean Geometry. The Open Court Publishing Company, Chicago,
1912.
[Bot84] Bottema, Oene: Archimedes Revisited. Mathematics Magazine, 57:224–225, 1984.
[Bot01] Bottema, Oene: The Malfatti Problem. Forum Geometricorum, 1:43–50, 2001.
[Bot07] Bottema, Oene: Topics in Elementary Geometry. Springer Verlag, Heidelberg, 2007.
[Boy91] Boyer, Carl: A History of Mathematics, 2nd Edition. John Wiley, New York, 1991.
[Bur86] Burnside, William: Theory of equations. Longmans, Green and Co., London, 1886.
[Cai33] Cairns, Stewart: An axiomatic basis for plane Geometry. Transactions of the American Math-
ematical Society, 35:234–244, 1933.
[Car16] Carslaw, H.S.: The elements of non-euclidean plane geometry and trigonometry. Longmans,
Green and Co., London, 1916.
[Cat52] Catalan, E.: Theoremes et problemes de Geometrie Elementaire. Carilian-Coeury, Paris, 1852.
[Cay76] Cayley, Arthur: On three-bar motion. Proceedings of the London Mathematical Society,
7:136–166, 1876.
BIBLIOGRAPHY 761
[CG67] Coxeter, H and L Greitzer: Geometry Revisited. Math. Assoc. Amer. Washington DC, 1967.
[Cha52] Chasles, M.: Traite de geometrie suprerieure. Bachelier, Paris, 1852.
[Cha65] Chasles, M.: Traite de Sections Coniques. Gauthier-Villars, Paris, 1865.
[Che90] Chern, Shiing Shen: What Is Geometry? American Mathematical Monthly, 97:679–686, 1990.
[Chu02] Church, Albert: Elements of Descriptive Geometry. American Book Company, New York,
1902.
[Cla19] Clawson, Wentworth: The Complete Quadrilateral. Annals of Mathematics, 20:232–261,

1919.
[Con98] Connes, Alain: A new proof of Morley’s theorem. Les relations entre les mathematiques et la
physique theorique, IHES, 40:43–46, 1998.
[Coo16] Coolidge, Julian Lowell: A treatise on the circle and the sphere. Oxford University Press,
Oxford, 1916.
[Coo40] Coolidge, Julian Lowell: A history of the Geometrical Methods. Oxford University Press,
Oxford, 1940.
[Coo68] Coolidge, Julian Lowell: A history of the Conic Sections and Quadric Surfaces. Dover Publi-
cations, New York, 1968.
[Cou80] Court, Nathan Altshiller: College Geometry. Dover Publications Inc., New York, 1980.
[Cox48] Coxeter, H S M: A Problem of Collinear Points. American Math. Monthly, 55:26–28, 1948.
[Cox49] Coxeter, H: Projective Geometry. Mathematics Magazine, 23:79–97, 1949.
[Cox61] Coxeter, H: Introduction to Geometry. John Wiley and Sons Inc., New York, 1961.
[Cox67] Coxeter, H: CUPM Geometry Conference. Math. Assoc. Amer. Washington DC, 1967.
[Cox68] Coxeter, H S M: The problem of Apollonius. American Math. Monthly, 75:5–15, 1968.
[Cox12] Cox, David: Galois Theory, 2nd ed. Wiley., New York, 2012.
[CR96] Courant, Richard and Herbert Robbins: What is Mathematics. Oxford University Press, Ox-
ford, 1996.
[Cra03] Crabbs, Robert: Gaspar Monge and the Monge Point of the Tetrahedron. Mathematics Maga-
zine, 76:193–203, 2003.
[CrFW91] Cockshott, Arthur and rev. F.B. Walters: Geometrical Conics. MacMillan and Co., London,
1891.
[Cro97] Cromwell, Peter: Polyhedra. Cambridge University Press, Cambridge, 1997.
[CS05] Christoph Scriba, Peter Schreiber: 5000 Jahre Geometrie. Springer Verlag, Heidelberg, 2005.
[CSC94] Chou Shang-Ching, Gao Xiao-Shan, Zhang Jing Zhong: Machine Proofs in Geometry. World
Scientific, Singapore, 1994.
[Dan55] Dantzig, Tobias: The bequest of the Greeks. George Allen and Unwin Ltd., London, 1955.
[Dar04] Darling, David: The Universal Book of Mathematics. Wilen, New York, 2004.
[Dav93] Davis, Philip: Spirals, From Theodorus to Chaos. A K Peters, Wellesley Massachusetts, 1993.
[Dav06] Davis, Thom: Geometry with Computers. Free internet edition, 2006.
762 BIBLIOGRAPHY
[dB81] Bruijn, N.G de: Algebraic theory of Penrose non-periodic tilings of the plane I, II. Proceedings
of the AMS, 84:39–66, 1981.
[dB11] Berg, Mark de: Computational Geometry. University of Crete Editions, Heraclion, 2011.
[dC76] Carmo, Manfredo do: Differential Geometry of Curves and Surfaces. Prentice-Hall Inc., New
Jersey, 1976.
[DD06] Dusan Djukic, Vladimir Jankovic, Ivan Matic Nikola Petrovic: The IMO Compedium, Inter-
national Mathematical Olympiads: 1959-2004. Springer, New York, 2006.
[Dea56] Deaux, Roland: Introduction to the Geometry of Complex Numbers. Dover, New York, 1956.
[Del89] Deltheil, R Caire D.: Geometrie et complements. Editions Jaques Gabay, Paris, 1989.
[Der07] Dergiades, Nikolaos: The Soddy Circles. Forum Geometricorum, 7:191–197, 2007.
[DF06] Dmitry Fuchs, Serg Tabchnikov: Mathematical Omnibus, Thirty lectures on Classic Mathe-
matics. Department of Mathematics, Davis, 2006.
[Dij87] Dijksterhuis, E. J.: Archimedes. Princeton University Press, Princeton, 1987.
[Doe65] Doerrie, Heinrich: 100 Great Problems of Elementary Mathematics. Dover Publications, Inc.
New York, 1965.
[DR81] DeTemple, Duane and Jack Robertson: A Billiard Path Characterization of Regular Polygons.
Mathematics Magazine, 54:73–75, 1981.
[Dri11] Driankos, Socrates: Irrational Numbers and Continued Fractions. Diploma Dissertation, Uni-
versity of Athens, 2011.
[dV98] Villiers, Michael de: Dual Generalizations of Van Aubel’s theorem. The Mathematical Gazette,
11:405–412, 1998.
[Efi80] Efimov, N. V.: Higher Geometry. Mir Publishers, Moscow, 1980.
[Ehr04] Ehrmann, Jean Pierre: Steiner’s Theorems on the Complete Quadrilateral. Forum Geometri-
corum, 4:35–52, 2004.
[Eng98] Engel, Arthur: Problem-Solving Strategies. Springer, Berlin, 1998.
[ERH94] Eddy R. H, Fritsch R.: The conics of Ludwig Kiepert: A Comprehensive Lesson in the Geom-
etry of the triangle. Mathematics Magazine, 67:188–205, 1994.
[EV51] Eves, Howard and V.Hoggatt: Hyperbolic Trigonometry Derived from the Poincare Model.
American Mathematical Monthly, 58:469–474, 1951.
[Eve63] Eves, Howard: A survey of Geometry. Allyn and Bacon, Inc., Boston, 1963.
[FC11] F. Cataldo, A. Graovac, O. Ori: The Mathematics and topology of fullerenes. Springer, Berlin,
2011.
[Fet77] Fetisov, A: Proof In Geometry. Mir Publishers, Moscow, 1977.
[F.G52] F.G.M.: (Jesuits’) Geometry Exercises. Karavias Editions, Athens, 1952.
[Fou00] Fourray, E.: Curiosites Geometriques. Vuibert editeurs, Paris, 1900.
[Fow79] Fowler, D. H.: Ratio in Early Greek Mathematics. Bulletin of the American Mathematical
Society, 1:807–846, 1979.
[Fra89] Fraleigh, John: A first course in Abstract Algebra. Addison-Wesley, New York, 1989.
BIBLIOGRAPHY 763
[FTW42] Fu Traing Wang, Chuan Chih Hsiung: A theorem on the Tangram. The Mathematical Gazette,
49:596–599, 1942.
[Fur37] Fursenko, V.B.: Lexicographical account of constructional problems of triangle geometry.

Mathematics in school (in Russian), 5:4–30, 1937.
[Gal13] Gallatly, William: The modern geometry of the triangle. Francis Hodgsonn, London, 1913.
[Gau13] Gaultier, Louis: Sur le moyens generaux de construire graphiquement un cercle determine par
trois conditions. Journal de l’Ecole polytechnique, 16:124–214, 1813.
[Gla20] Glaser, Robert: Stereometrie. Walther de Gruyter, 1920.
[GR04] Gisch, David and Jason Ribando: Apolloniu’s Problem: A Study of Solutions and Their Con-
nections. American Journal of Undergraduate Research, 3:15–26, 2004.
[Gra59] Graham, L.: Ingenious Mathematical Problems and Methods. Dover Publications, New York,
1959.
[Gre88] Greitzer, Samuel: Arbelos, Special Geometry issue. Arbelos, Mathematical Association of
America, 6:1–51, 1988.
[Gro04] Gronau, Detlef: The Spiral of Theodorus. The American Mathematical Monthly, 111:230–
237, 2004.
[GS87] Gruenbaum, Branko and G.S. Shephard: Tilings and Patterns. W.H. Freeman and company,
New York, 1987.
[GS95] Grunbaum, Branko and G. C. Shephard: Ceva, Menelaus, and the Area Principle. Mathematics
Magazine, 68:254–268, 1995.
[GT12] George Thomas, Ross Finney: Calculus. Addisson Wesley, New York, 2012.
[Had05] Hadamard, Jacques: Lecons de Geometrie elementaire I, II. Librairie Armand Colin, Paris,
1905.
[Har56] Hart, Andrew: Geometrical investigation of Steiner’s solution of Malfatti’s problem. Quarterly
journal of pure and applied Mathematics, 1:219–222, 1856.
[Haw62] Hawk, Minor Clyde: Theory and problems of descriptive geometry. McGraw-Hill Book Com-
pany, New York, 1962.
[Hea96] Heath, T: Apollonius of Perga, Treatise on conic sections. Cambridge University Press, Cam-
bridge, 1896.
[Hea97] Heath, T: The works of Archimedes. Cambridge University Press, Cambridge, 1897.
[Hea08] Heath, T: The thirteen books of Euclid’s elements vol. I, II, III. Cambridge University Press,
Cambridge, 1908.
[Hea31] Heath, T: A manual of Greek Mahtematics. Oxford University Press, Oxford, 1931.
[Hei85] Heiberg, Johan: Euclidis Elementa. Teubner Verlag, 1885.
[Hel76] Helmholtz, H.: The Origin and Meaning of Geometrical Axioms. Mind, 1:301–321, 1876.
[Hes12] Hess, Albrecht: A highway from Heron to Brahmagupta. Forum Geometricorum, 12:191–192,
2012.
[HF08] Hidetoshi Fukagawa, Tony Rothman: Sacred Mathematics. Princeton University Press, Prince-
ton, 2008.
764 BIBLIOGRAPHY
[Hil03] Hilbert, David: Grundlagen der Geometrie. Teubner Verlag, Leipzig, 1903.
[Hon95] Honsberger, Ross: Episodes in Nineteenth and Twentieth Century Euclidean Geometry. The
Mathematical Association of America, New Library, Washington, 1995.
[Hon97] Honsberger, Ross: In Polya’s Footsteps. The Mathematical Association of America, The Dol-
ciani Mathematical Expositions, Washington, 1997.
[Joh16] Johnson, Roger: Relating to the Simson line or Wallace line. American Math. Monthly, 23:61–
62, 1916.
[Joh60] Johnson, Roger: Advanced Euclidean Geometry. Dover Publications, New York, 1960.
[Kar13] Karpenkov, Oleg: Geometry of Continued Fractions. Springer, Berlin, 2013.
[Kat98] Katz, Victor: A History of Mathematics. An introduction, Second Edition. Addison-Wesley

Longman Inc., New York, 1998.
[Kat10] Katsigiannis, Kostas: From Archimedes’ Stomachion to Pick’s theorem Diploma Dissertation.
University of Athens, University of Cyprus, 2010.
[Kaz68] Kazarinoff, Nicholas: On who first proved the impossibility... ruler and compass alone. Amer-
ican Mathematical Monthly, 75:647, 1968.
[Kaz70] Kazarinoff, Nicholas: Ruler and the Round. Dover, New York, 1970.
[Kep96] Keppler, Johannes: Nova Stereometria doliorum vinariorum. Frankfurt, 1596.
[Kep19] Keppler, Johannes: Harmonices Mundi Libri V. Linz, 1619.
[KF70] Kolmogorov, A. and S. Fomin: Introductory Real Analysis. Dover, New York, 1970.
[Khi61] Khinchin, A.: Continued Fractions. The University of Chicago press, Chicago, 1961.
[Kim14] Kimberling, Clark: Encyclopedia of triangle centers.

http://faculty.evansville.edu/ck6/encyclopedia/ETC.html, 2014.
[KL07] Kaplan, Wilfred and Donald Lewis: Calculus and Linear Algebra, vol. I, II. Sholarly Publish-
ing Office, Ann Arbor, 2007.
[Kle97] Klein, Felix: Famous Problems of Elementary Geometry. Ginn and Company, Boston, 1897.
[Kno93] Knorr, Wilbur Richard: The ancient tradition of geometric problems. Dover, New York, 1993.
[KW91] Klee, Victor and Stan Wagon: Old and New Unsolved Problems in Plane Geometry and Num-
ber Theory. The Mathematical Association of America, Washington, 1991.
[Lac93] Lachlan, R.: Modern Pure Geometry. Macmillan and Co., London, 1893.
[Lal52] Lalesco, Trajan: La Geometrie du triangle. Librairie Vuibert, Paris, 1952.
[Lan96] Lang, Robert J.: Origami and Geometric Constructions. Preprint, 1996.
[LC65] Lebosse C., Hemery C.: Geometrie. Fernand Nathan, Paris, 1965.
[Leg37] Legendre, Adrien Marie: Elements de Geometrie, suivis d’ un traite de Trigonometrie. Langlet
et compagnie, Bruxelles, 1837.
[Log80] Logothetti, Dave: An Interview with H.S.M. Coxeter, the King of Geometry. The Two-Year
College Mathematics Journal, 11:2–19, 1980.
[Loh82] Lohwater, Arthur: Introduction to inequalities. Internet published in pdf, 1982.

BIBLIOGRAPHY 765
[Loo68] Loomis, Elisha Scott: The Pythagorean Proposition. The National Council of Teachers of
Mathematics, Washington, 1968.
[LWBM06] Leah Wrenn Berman, Gordon Ian Williams and Bradley James Molnar: The Cross Ratio Is the
Ratio of Cross Products! Mathematics Magazine, 79:54–59, 2006.
[Mac93] Mackay, J. S.: Early history of the symmedian point. Proceedings of the Edinburgh Mathemat-
ical Society, 11:92–103, 1893.
[Mac95] Mackay, J. S.: Symmedians of a triangle and their concomitant circles. Proceedings of the
Edinburgh Mathematical Society, 14:37–103, 1895.
[Mar87] Martin, George E.: Transformation Geometry. Springer-Verlag, Berlin, 1987.
[Mar02] Marchisotto, Elena Anne: The theorem of pappus: A bridge between algebra and geometry.
Amer. Math. Monthly, 109:497–516, 2002.
[Mcl91] Mclelland, William: A treatise on the Geometry of the Circle. Macmillan and Co, London,
1891.
[Moo64] Moore, Charles: An Introduction to Continued Fractions. The National Council of Teachers
of Mathematics, Washington, 1964.
[MP77] Millman, Righard and George Parker: Elements of differential geometry. Prentice-Hall Inc.,
New Jersey, 1977.
[MP94] Mironescu, P. and L. Panaitopol: The existence of a triangle with prescribed angle bisector
lengths. Amer. Math. Monthly, 101:58–60, 1994.
[Mui95] Muirhead, R.: On the number and nature of the solutions of the Apollonian contact problem.
Proc. Edinburgh Math. Society, 14:135–147, 1895.
[Nah06] Nahin, Paul J.: Dr. Euler’s Fabulous Formula. Princeton University Press, Princeton, 2006.
[Nah07] Nahin, Paul J.: Chases and Escapes, The Mathematics of Pursuit and Evasion. Princeton
University Press, Princeton, 2007.
[Nar09] Naraniengar, M. T.: Solution to Morley’s problem. Educational Times (New Series), 15:47,
1909.
[Neg06] Negrepontis, S.: Plato’s theory of ideas is the philosophic equivalent of the theory of continued
fraction expansions of lines commensurable in power only. Manuscript, 1:1–70, 2006.
[Nie11] Niemeyer, Jo: A simple Construction of the Golden Section. Forum Geometricorum, 11:53,
2011.
[Niv78] Niven, Ivan: Convex Polygons that Cannot tile the Plane. The American Mathematical
Monthly, 85:785–792, 1978.
[Ogi69] Ogivly, Stanley: Excursions in Geometry. Oxford University Press, New York, 1969.
[O’L10] O’Leary, Michael: Revolutions of Geometry. Wiley, 2010.
[OR79] Odier, Y. and Y. Roussel: Trioker mathematisch gespielt. Friedr. Vieweg, Braunschweig, 1979.
[O’R11] O’Rourke, Joseph: How to fold it. Cambridge University Press, Cambridge, 2011.
[Oxm08] Oxman, Victor: A Purely Geometric Proof of the Uniqueness of a Triangle With Prescribed
Angle Bisectors. Forum Geometricorum, 8:197–200, 2008.
[Pap76] Pappus: Congregation. Weidmann, Berlin, 1876.

766 BIBLIOGRAPHY
[Pap96] Papelier, Georges: Exercices de Geometrie moderne. Editions Jacques Gabay, 1996.
[Pap10] Papadopoulos, Athanase: Nikolai I. Lobachevsky, Pangeometry. European Mathematical So-

ciety, 2010.
[PE99] Pierre Eymard, Jean Pierre Lafon: Autour du nombre π. Hermann Editeurs, Paris, 1999.
[Ped90] Pedoe, D: A course of Geometry. Dover, New York, 1990.
[Per57] Perron, Oskar: Die Lehre der Kettenbruechen, Band I, II. B.G Teubner Verlagsgesellschaft,
Stuttgart, 1957.
[Per08] Perrin, Daniel: Algebraic Geometry An Introduction. Springer, New York, 2008.
[Pet01] Petersen, Julius: Constructions Geometriques. Gauthier-Villars, Paris, 1901.
[Pet41] Peters, J.W.: The theorem of Morley. National Mathematics Magazine, 16:119–126, 1941.
[Pet09] Petkovic, Miodrag: Famous Puzzles. American Mathematical Society, Providence, 2009.
[Pin05] Pinelis, Iosif: Cyclic polygons with given edge lengths. Journal of Geometry, 82:156–171,
2005.
[PL07] Posamentier, Alfred and Ingmar Lehmann: The (Fabulous) FIBONACCI Numbers.
Prometheus Books, New York, 2007.
[PR06] Poonen, Bjorn and Michael Rubinstein: The number of intersection points made by the diago-
nals of a regular polygon. SIAM J. Discrete Mathematics, 11:135–156, 2006.
[Pre11] Pressley, Andrew: Elementary Differential Geometry. University of Crete Editions, Heraclion,
2011.
[PS88] Posamentier, Alfred and Charles Salkind: Challenging Problems in Geometry. Dover, New
York, 1988.
[Qui04] Quinn, John James: A linkage for describing the conic sections by continuous motion. Ameri-
can Mathematical Monthly, 11:12–13, 1904.
[RA03] Rashid, M. A and A. O Ajibade: Two conditions for a quadrilateral to be cyclic expressed in
terms of the lengths of its sides. International Journal of Mathematical Education in Science
and Technology, 34:739•799, 2003.
[R.B09] R.B.Manfrino, J.A.G. Ortega, R.V.Delgado: Inequalities, A Mathematical Olympiad Ap-

proach. Birkhaeuser, Berlin, 2009.
[RG95] Ronald Graham, Donald Knuth, Oren Patashnik: Concrete Mathematics. Addison-Wesley,
New York, 1995.
[Ria62] Riaz, M.: Geometric Solutions of Algebraic Equations. American Mathematical Monthly,
69:654–658, 1962.
[Ric08] Richeson, David: Euler’s gem. Princeton University Press, Princeton, 2008.
[Rob71] Robinson, R.: Undecidability and nonperiodicity of tilings of the plane. Inventiones Mathe-
maticae, 12:177–209, 1971.
[Ros88] Rosenfeld, B.A.: A history of Non-Euclidean Geometry. Springer-Verlag, New York, 1988.
[Row17] Row, Sundara: Geometric Exercises in Paper Folding. The open court publishing company,
Chicago, 1917.
[RT33] Rademacher, Hans and Otto Toeplitz: Von Zahlen und Figuren. Springer Verlag, 1933.
BIBLIOGRAPHY 767
[RWG76] R. W. Gaskell, M. S. Klamkin, P. Watson: Triangulations and Pick’s Theorem. Mathematics

Magazine, 49:35–37, 1976.
[Sal17] Salmon, George: A treatise on Conic Sections. Longmans, Green and Co., London, 1917.
[San15] Sandifer, Edward: How Euler Did Even More. Mathematical Association of America, New
York, 2015.
[Sch64] Schwarz, Hermann Amandus: Elementarer Beweis des Pohlkeschen Fundamentalsatzes der
Axonometrie. Journal fuer die reine und angewandte Mathematik, 63:309–314, 1864.
[Sch78] Schattschneider, Doris: Tilling the Plane with Congruent Pentagons. Mathematics Magazine,
51:29–44, 1978.
[Sch93] Schreiber, Peter: On the Existence and Constructibility of Inscribed Polygons. Beitraege zur
Algebra und Geometrie, 34:195–199, 1993.
[Scu28] Scudder, H.T.: How to trisect an angle with a carpenter’s square. American Mathematical
Monthly, 35:250–251, 1928.
[sH94] Hahn, Liang shin: Complex Numbers and Geometry. Mathematical Association of America,
1994.
[Sha86] Sharygin, I. F.: Problems in Solid Geometry. Mir, Moscow, 1986.
[Sin95] Singer, David: Isometries of the Plane. The American Mathematical Monthly, 102:628–631,
1995.
[Sin01] Sinefakopoulos, Achilleas: Circumscribing an Inscribed Quadrilateral. The American Math-

ematical Monthly, 108:378, 2001.
[Smi64] Smirnov, V.I.: A course of higher Mathematics I,II,III. Pergamon Press, New York, 1964.
[Spi65] Spivak, Michael: Calculus on Manifolds. Addison-Wesley, New York, 1965.
[Spi94] Spivak, Michael: Calculus, Third Edition. Publish or Perish, Houston, 1994.
[ST76] Singer, I and J Thorpe: Lecture notes on Elementary Topology and Geometry. Springer-Verlag,
Berlin, 1976.
[Sta00] Stamou, Georgios: Relating to Archimedes’ Twin Circles. Educational Considerations, 8:31–
32, 2000.
[Ste71] Steiner, Jacob: Gesammelte Werke vol. I, II. Chelsea Publishing Company, New York, 1971.
[Ste83] Steinhaus, Hugo: Mathematical Snapshots. Oxford University Press, 3rd Edition, Oxford,
1983.
[Sti47] Stiefel, Eduard: Lehrbuch der Darstellenden Geometrie. Springer Verlag, Basel, 1947.
[Sup01] Suppa, Italo D’ Ignazio Ercole: Il Problema Geometrico, dal compasso al cabri. interlinea
editrice, Teramo, 2001.
[TA12] Titu Andreescu, Cosmin Pohoata: Back to Eucldiean Geometry: Droz-Farny Demystified.
Mathematical Reflections, 3:1–5, 2012.
[Tan67] Tan, Kaidy: Different Proofs of Desargues’ Theorem. Mathematics Magazine, 40:14–25, 1967.
[Tay81] Taylor, Charles: Geometry of Conics. Deighton Bell and Co, Cambridge, 1881.
[Thi03] Thiele, Ruedger: Hilbert’s Twenty-Fourth Problem. The American Mathematical Monthly,
110:1–24, 2003.
768 BIBLIOGRAPHY
[Tho57] Thomas, Ivor: Greek Mathematics (selections) 2 vols. Harvard University Press, Cambridge,
1957.
[Tia12] Tian, Xiaoxi: The Art and Mathematics of Tangrams. Bridges 2012: Mathematics, Music, Art,
Architecture, Culture, 1:553–556, 2012.
[Var00] Vardi, Ilan: Mekh-Mat Entrance Examination Problems. Institut des Hautes Etudes Scien-
tifiques, 6:1–47, 2000.
[Var05] Varverakis, Antreas: A Maximal Property of Cyclic Quadrilaterals. Forum Geometricorum,
5:63–64, 2005.
[Vas69] Vasic, O. Bottema R. Djordjevic R. Janic D. Mitrinovic P.: Geometric Inequalities. Wolters-
Noordhoff Publishers, Groningen, 1969.
[VG88] V. Gusev, V. Litvinenko, A. Mondkovich: Solving Problems in Geometry. Mir Publishers,
Moscow, 1988.
[VY10] Veblen, Oswald and John Young: Projective Geometry vol. I, II. Ginn and Company, New
York, 1910.
[vY93] Yzeren, Jan van: A Simple Proof of Pascal’s Hexagon Theorem. The American Mathematical
Monthly, 100:930–931, 1993.
[Wal01] Walser, Hans: The Golden Section. The Mathematical Association of America, 2001.
[Wei03] Weisstein, Eric: CRC concise encyclopedia of mathematics. Chapman and Hall/CRC, Boca
Raton, 2003.
[Wel91] Wells, David: Dictionary of Curious and Interesting Geometry. Penguin Books, 1991.
[Whi07] White, Homer: The Geometry of Leonhard Euler, Life Work and Legacy. Studies In the history
and philosophy of mathematics, Elsevier, 5:303–323, 2007.
[Wil69] Wilker, J.B.: Four Proofs of a Generalization of the Descartes Circle Theorem. American
Mathematical Monthly, 76:278–282, 1969.
[WY99] Woo, Dodge Schoch and Yiu: Those Ubiquitous Archimedean Circles. Mathematics Magazine,
72:202–213, 1999.
[Yag62] Yaglom, I. M.: Geometric Transformations I, II, III. The Mathematical Association of Amer-
ica, 1962.
[Yiu13] Yiu, Paul: Introduction to the Geometry of the Triangle .
http://math.fau.edu/Yiu/Geometry.html, 2013.
[You17] Young, John Wesley: Lectures on Fundamental Concepts of Algebra and Geometry. Macmil-
lan Company, New York, 1917.
Index
O(ρ), 57, 62, 65 Angle of similarity, 495

κ(O, ρ), 57 Angle of skew lines, 541
π, 447 Angle of spherical lune, 602
π approximation, 449, 472 Angle of spherical polygon, 606, 640
π, value, 449 Angle oriented, 484
(ABΓ ), 59 Angle sides, 8
Angle straight, 8
Absolute Geometry, 33, 35, 38 Angle sum, 9
Acute, 12, 24 Angle trisection, 98
Acute triangle, 14, 116, 204, 206, 222, 290, 301, Angle vertex, 8
322, 361, 573 Angles adjacent, 9
Additivity of areas, 136 Angles equal, 9
Adjacent angles, 21, 39, 74, 94, 626 Angles of great circles, 602
Affinity, 727 Angles vertical, 11
Altitude, 15, 22, 23, 26, 27, 30, 53–55, 72, 74, 76, Anthyphairesis, 148, 464, 465
86, 102, 109, 115, 117, 120, 121, 123– Anti-inversion, 295, 297
126, 129, 132, 140, 142, 151, 157, 179, Anti-inversion interchanging, 296
194, 200, 204, 206, 210, 214, 217, 222– Anti-isometry, 479
224, 232, 250, 264, 301, 308, 312, 315, Anticenter, 116
316, 318, 322, 357, 420, 421, 436, 439, Anticomplementary, 85
553, 570, 620, 622, 629, 691 Antihomologous points, 266, 269, 277, 296, 297
Altitude of cone, 586, 658, 659 Antihomothetic, 180
Altitude of prism, 581, 632, 648, 649 Antihomothety, 436, 492, 748, 749
Altitude of pyramid, 569, 651, 652 Antiparallel, 207, 210, 212
Altitude of truncated cone, 591, 659 Antiparallels, 527
Analysis, 70, 396, 415, 571 Antiprism, 625, 752, 754, 755
Analysis, Synthesis, Discussion, 70 Antisimilarity, 495, 498
Angle, 8 Apex, 18, 27, 28, 30, 31, 40, 53, 80, 86, 95, 131,
Angle acute, 44 144, 158, 323, 363, 553, 569, 574, 610,
Angle central, 100, 103 640, 673
Angle congruence, 16, 20, 26, 32 Apollonian circle, 245–247, 250, 262, 291, 294, 339,
Angle exterior, 8 427, 497, 498
Angle inscribed, 103 Apollonius, 246, 249, 255, 328, 330, 624, 708
Angle interior, 8 Apothem, 63
Angle measure, 9, 220, 452 Arbelos, 351, 355, 422, 430, 466
Angle non-convex, 8 Arc, 99, 102
Angle null, 8 Arc complementary, 99
Angle of circles, 259, 274, 280, 295 Arc length, 449, 451, 592
Angle of line and plane, 548 Archimedean axiom, 138, 641, 646
Angle of parallelism, 515 Archimedean solid, 580, 625, 751, 752, 754, 758
770 INDEX
Archimedean tiling, 516, 518, 521 Blundon, 328

Archimedean triangle, 694, 697, 700, 721, 731 Bodenmiller, 392
Archimedes, 108, 138, 351, 443, 449, 751 Bolyai, 35
Arcs congruent, 100 Brahmagupta, 334, 338
Arcs equal, 450 Bretschneider, 338, 339, 427
Area, 638, 728 Brianchon, 401, 403
Area of parallelogram, 140 Brianchon point, 404
Area of polygon, 135, 189 Broken line, 29
Area of quadrilateral, 164, 195 Butterfly theorem, 407
Area of rectangle, 138
Area of trapezium, 142 Cairns, 2
Area of triangle, 140, 194, 318 Cairo tiling, 522
Area, axioms of, 136 Calabi, 417
Arithmetic mean, 156, 656 Calabi’s triangle, 417
ASA-criterion, 19–22, 40, 74, 93, 95, 96 Carrier, 4
Asymptote, 690, 711, 714, 715, 717, 732 Castillon, 289, 407, 411
Asymptotic triangle, 513, 715, 717, 731 Cavalieri, 666
Autopolar triangle, 392, 421 Cayley, 372
Auxiliary circles, 712 Center, 57, 89
Axes of ellipse, 725 Center line, 214, 241, 251, 256, 259, 331, 505, 623,
Axial symmetry, 45, 86, 737 706
Axiom, 3, 6, 8, 14, 38, 138 Center of conic, 690
Axiom of completeness, 441 Center of mass = centroid, 84
Axiom of parallels, 38, 152, 163, 187 Center of symmetry, 47, 133, 736, 737
Axioms, 1, 2, 5, 14, 33, 35, 44, 100, 136, 187, 441, Centerline, 64
506, 540 Centers of similarity, 383
Axioms, space, 531 Central projection, 616
Axis of antisimilarity, 495 Centroid, 84, 86, 181, 213, 218, 466
Axis of cone, 586, 659, 673 Centroid of tetrahedron, 576
Axis of cylindrical surface, 585, 672 Ceva, 380, 383
Axis of parabola, 699 Chord, 57, 99–101, 103, 152, 351
Axis of rotation, 660, 665, 740 Chord of conic, 688
Axis of symmetry, 45, 48, 59, 66, 73, 89, 94, 674, Chord of sphere, 593
737 Circle, 57
Axonometric projection = Parallel projection, 612 Circle area, 452, 455
Circle chain, 524
Bachmann, 2 Circle construction, 329
Base, 18 Circle escribed, 311
Base of cone, 586 Circle exterior, 57
Base of pyramid, 569 Circle inscribed, 311, 313
Bases of a prism, 581 Circle interior, 57
Bases of truncated cone, 591 Circle of inversion, 271, 504
Between, 1, 4, 5 Circle of similitude, 269, 297
Bicentric quadrilateral, 119, 164 Circle orthogonal, 261, 275
Billiard, 158, 215, 523 Circle pencil, 250, 260, 274, 286, 350
Binet, 471 Circle perimeter, 447
Birkhoff, 2 Circle power curve, 310
Bisecting plane, 666 Circle squaring, 98
Bisecting planes, 559 Circle tangent, 236
Bisector, 9, 15, 21, 34, 44, 52, 62, 68, 86, 95, 101, Circle tritangent, 311
106, 110, 115, 132, 143, 158, 200, 202, Circular section, 456
203, 223, 250, 312, 322, 325, 526, 555, Circular sector, 455, 662
559 Circular sector area, 455
Bisector external, 430 Circular segment, 663
INDEX 771
Circumcenter, 59, 86, 325 Conic singular, 674

Circumcircle, 59, 95, 114, 236 Conic surface, 587
Circumradius, 95 Conics with center, 674
Circumscircle, 89 Conjugate chords of conic, 690
Circumscribed, 95 Conjugate diameters, 690
Circumscribed circle = Circumcircle, 59 Conjugate directions, 690, 707, 708
Circumscribed conical surface, 590 Contact point, 61, 64, 594
Circumscribed parallepiped, 573 Continued fraction, 464, 472
Circumscribed polyhedron, 601 Convex, 77
Circumscribed quadrilateral, 118 Convex angle, 8
Circumscribed sphere, 601 Convex hull, 461
Circumscriptible, 118 Convex polyheder, 751
Clifford, 372 Convex polyhedron, 577
Common harmonics, 263 Convex prism, 581
Common orthogonal, 547 Convex pyramid, 569
Common tangents, 266, 314 Convex spherical polygons, 640
Commuting transformations, 482 Coordinate, 367
Complementary, 12 Cosine rule, 193, 195
Complementary triangle, 181 Cosine spherical 1st formula, 606
Complete quadrilateral, 387, 693 Cosine spherical 2nd formula, 608
Composition of transformations, 473 Cross ratio, 373, 504, 555, 693
Concentric, 64, 101 cross ratio of four lines, 374
Conchoid of Nicomedes, 437 Cube, 578, 582, 626, 667
Concyclic, 235 Cubeoctahedron, 580, 754
Cone, 586 Cyclic = Inscriptible, 112
Cone angle, 587 Cylinder, 585, 635
Cone apex, 586, 587 Cylinder altitude, 585, 656
Cone area, 631 Cylinder area, 631, 632
Cone circular, 604 Cylinder base, 585
Cone frustum, 591 Cylinder oblique circular, 586
Cone generator, 588 Cylinder radius, 585
Cone oblique circular, 626 Cylinder unfolding, 585
Cone opening, 626 Cylinder volume, 656
Cone opening = Cone angle, 587 Cylinders congruent, 585
Cone right circular, 586 Cylindrical surface, 585, 586, 677
Cone truncated, 591
Cone vertex = Cone apex, 586 Dandelin, 676, 685, 732
Cones congruent, 586 Dandelin spheres, 676, 679, 682, 684, 685
Congruence, 474, 489, 742 Degrees, 9
Congruent circles, 58 Delian problem, 98
Congruent polygons, 94 Deltoid, 343
Congruent prisms, 581 Desargues, 393, 408, 540, 619
Congruent quadrilaterals, 77 Descartes, 434
Congruent shapes, 16 Descartes formula, 434
Congruent spherical lunes, 602 Diagonal, 74, 94, 96, 99, 136, 387
Congruent triangles, 20, 23, 54, 74, 114, 490, 500, Diagonal of parallepiped, 581
612, 627 Diameter, 57, 58, 60
Congruent truncated cones, 591 Diametral, 108, 593
Conic exterior, 688 Diametral = Diametrically opposite, 58
Conic interior, 688 Diametrical, 258
Conic line intersection, 250 Diametrically opposite, 58, 181, 195, 206, 324, 325,
Conic section, 676 643
Conic section, proper, 674 Diametrically opposite = Diametral, 593
Conic sections, 405, 674 Diheder, 557, 620
772 INDEX
Diheder acute, 558 Ellipse, secondary circle, 702

Diheder complementary, 558 Elliptic pencil = Pencil intersecting, 251
Diheder congruent, 558 Enantiomorphic, 520, 757
Diheder convex, 557 End, 4
Diheder face, 557 End point, 4
Diheder measure, 557 Envelope, 343, 694
Diheder obtuse, 558 Equal segments, 6
Diheder right, 558 Equality, 1, 16, 460
Diheder supplementary, 558 Equation cubic, 327
Diheders, 581 Equilateral, 40, 53, 56, 68, 75, 86, 89, 96, 108, 109,
Diheders of polyhedron, 577 114, 115, 132, 143, 146, 149, 221, 258,
Diheders of the triheder, 561 294, 315, 323, 335, 347, 360, 362–364,
Dihedral, 561 576, 606, 610, 668, 670
Dihedral angle = Diheder, 557 Equilateral inscribed, 347
Direct isometry, 479, 748 Escribed, 313, 326
Direct similarity, 495 Escribed circles, 315
Direction, 4 Escribed polyhedron, 601
Director circle, 730 Euler, 228, 318, 324, 644
Directrices of ellipse, 683 Euler characteristic, 644
Directrices of hyperbola, 683 Euler circle, 318, 319, 321, 343, 407, 439, 440
Directrix, 674, 679, 680, 682, 683 Euler line, 181, 303, 400
Discussion, 70 Euler’s theorem, 324
Distance, 5 Excenter, 311, 312
Distance from line, 31 Exterior, 4
Distance from plane, 544, 545 Exterior angle, 15
Distance line, 513–515 Exterior points, 688
Distance of two planes, 545 External angle, 24, 39
Distance of parallel from plane, 545 External bisector, 15, 144
Distance of parallels, 42, 140 External similarity center, 266
Distance of skew lines, 545
Dodecahedron, 579, 611, 624 Face, 581
Domain, 473 Face of polyhedron, 577
Droz-Farny, 300 Faces of triheder, 561
Dual platonic solid, 580 Fagnano, 360, 361
Fermat, 228, 360
Eccentric angle, 702, 707, 726 Fermat numbers, 97
Eccentricity, 683, 686, 690, 700, 702, 709, 724 Feuerbach, 322, 323, 337
Edge of diheder, 557 Feuerbach point, 323
Edges of polyhedral angle, 569 Fibonacci, 470
Edges of polyhedron, 577 Fixed point, 47
Edges of pyramid, 569 Fixed point of transformation, 476
Edges of triheder, 561 Fixed points, 45, 271
Electromagnetic field, 264 Flank, 145
Ellipse, 673, 675, 677, 681, 700, 704, 730 Flat angle, 8
Ellipse, area, 728 Focal distance, 700, 709
Ellipse, auxiliary circles, 704, 706, 724 Focal parameter, 692, 693, 700, 702, 709
Ellipse, axes, 701, 705 Focal point, 673
Ellipse, center, 701 Foci, 674
Ellipse, diameter, 701 Foci = Focal points, 673
Ellipse, half parameter, 702 Focus, 674
Ellipse, major axis, 701 Focus = Focal point, 673
Ellipse, major circle, 702 Ford, 468
Ellipse, minor axis, 701 Ford circles, 468
Ellipse, minor circle, 702 Full turn, 8
INDEX 773
Fundamental invariants, 328 Hexagon, 211, 399, 403

Hexagon regular, 96, 131, 146
Galileo, 454 Hexagon symmetric, 517
Gauss, 97, 392, 591 Hilbert, 2
Gelder, 472 Hippocrates, 458
Generator, 581 Hjelmslev, 529
Generators of cone, 586 Homographic relation, 408, 409, 412, 436
Generators of surface, 585 Homologous, 180, 393
Genus of polyhedron, 644 Homologous points, 266
Geometric locus, 21, 43, 60, 64, 88, 90, 102, 105, Homothetic polygons, 552
119, 124, 128, 133, 151, 158, 174, 181, Homothetic triangles, 180
182, 193, 201, 205, 218, 222, 224, 231, Homothety, 492, 524, 748, 750
236, 241, 245, 246, 253, 254, 256, 257, Homothety center, 180, 186, 395, 492, 748
263, 270, 279, 280, 285, 287, 289, 290, Homothety ratio, 180, 186, 492, 748
294, 298, 305, 309, 340, 351, 421, 543, Horizon, 506, 514
548, 552, 554, 559, 623, 673, 679, 684, Horocycle, 514
693, 699, 705, 715, 723, 724, 730 Horocycle, limit point, 514
Geometric mean, 156 Hyperbola, 674, 679, 682, 690, 709, 713, 724
Gergonne, 383, 384, 405 Hyperbola, axes, 710
Glide reflection, 490, 743, 746
Hyperbola, center, 710
Golden ratio section, 238
Hyperbola, conjugate axis, 712
Golden rectangle, 238
Hyperbola, conjugate hyperbola, 717
Golden section, 96, 237, 470, 512, 624
Hyperbola, diameters, 710
Golden section ratio, 466
Hyperbola, focal points, 710
Great circle, 594
Hyperbola, major circle, 712
H-circle, 509 Hyperbola, primary circle = major circle, 712
H-isometry, 511 Hyperbola, rectangular, 675, 717
H-line, 506 Hyperbola, transverse axis, 712, 714
H-line segment, 508 Hyperbola, vertices, 712
H-medial line, 510 Hyperbolic Geometry, 33–35
H-middle, 510 Hyperbolic pencil = Pencil non-intersecting, 251
H-plane, 506, 511 Hyperparallels, 507
H-point, 506 Hyperparallels, common orthogonal, 512
H-point at infinity, 506 Hypotenuse, 18, 25, 71
H-reflection, 507, 511, 514
H-regular polygon, 512 Icosaehedron, 611
H-triangle, 511 Icosahedron, 579, 625
H-trigonometry, 515 Icosidodecahedron, 580, 755
Half circumference, 100 Identity transformation, 474
Half line, 6, 41, 61, 100 Image, 473
Half planes, 3 Incenter, 62, 63, 312, 324
Half spaces, 531 Incidence angle, 32
Half turn, 486 Incircle, 62, 96
Half-perimeter, 14 Incircle = Inscribed circle, 62
Harmonic conjugate, 50, 263, 281, 289, 391 Indirect isometry, 748
Harmonic division = Harmonic tetrad, 49 Inscribable = Inscriptible, 112
Harmonic mean, 156, 213 Inscribed, 96
Harmonic pencil , 374 Inscribed circle, 62, 118
Harmonic quadrilateral, 378, 506 Inscribed circle of the arbelos, 352
Harmonic tetrad, 49, 50, 173 Inscribed conical surface, 590
Height of spherical zone, 638 Inscribed quadrilateral, 112
Helicoidal displacement, 744, 746 Inscriptible, 116, 119
Helicoidal displacements, 743 Inscriptible in sphere, 601
Heron, 316, 318 Inscriptible quadrilateral, 195
774 INDEX
Interior, 4 Line at infinity, 283, 382, 383, 394, 399, 683

Interior of triheder, 561 Line parallel to plane, 538
Interior points, 688 Line, in space, 531
Internal similarity center, 266 Lobatsevsky, 35, 506
Intersection of planes, 536 Logarithmic spiral, 502
Intervals separating each other, 263 Lune, 456, 458
Invariant by similarity, 348
Inverse pencil, 275 Malfatti, 414
Inverse points, 271 Maps = transforms, 473
Inverse points distance, 279 Maximal altitude, 293
Inverse transformation, 474 Maximal angle, 177, 232, 258, 309
Inversion, 271, 289, 290, 308, 354, 504, 508 Maximal area, 146, 184, 199, 222, 294, 417, 453,
Inversion center, 271, 504 461, 467, 525, 592, 628, 636, 647
Inversion interchanging, 277 Maximal distance, 192
Involution, 478 Maximal lateral area, 635
Iscriptible, 112 Maximal perimeter, 108, 221
Isogonal conjugate, 386, 387, 432 Maximal product, 184
Isogonal conjugation, 432 Mean proportional, 156, 223
Isometric shapes = congruent shapes, 16 Measure in radians, 450
Isometries, 752 Measures, 136
Isometries of the plane = Concgruences of the plane, Medial line, 19, 21, 58, 67, 101, 245, 248, 510, 543,
474 694, 705, 713
Isometry, 16, 475, 486, 733, 734, 736, 738, 740, 747, Medial plane, 543, 547, 556, 577, 595
750 Median, 15, 21, 30, 43, 83, 136, 141, 145, 200, 201,
Isoperimetric inequality, 459 232, 236, 421
Isosceles, 18, 20, 21, 25, 26, 30, 40, 44, 48, 53, 55, Medians of tetrahedron, 576
58, 59, 63, 68, 72–74, 76, 80, 84, 86, 108, Medians triangle, 85
117, 129, 133, 142, 146, 153, 157, 179, Menelaus, 379, 401
206, 238, 302, 323, 325, 340, 417, 453, Menelaus theorem, 493
553, 661, 694 Middle, 6, 67
Isosceles trapezium, 115, 126, 143, 181, 182, 212, Middle cut, 591
230, 240, 291, 308, 342, 365, 659, 668, Middle parallel, 43, 81, 93, 552
725 Middle sphere, 601, 602
Isotomic conjugate, 386 Middle term, of fractions, 469
Minimal angle, 258, 550
Kepler, 666
Minimal area, 346, 650, 708
Lateral area, 631–636, 639, 672 Minimal chord, 63
Lateral edges, 569, 581 Minimal distance, 31, 32, 79, 80, 182, 192, 200, 503,
Lateral faces, 569 549, 555, 647
Lateral surface, 586, 631, 656 Minimal perimeter, 116, 143
Latus rectum of conic, 692 Minimal sum, 219, 256, 360
Legendre, 35, 654 Minor circle = Small cirlce, 594
Legs, 18 Minute (of degree), 9
Lehmus, 76, 206 Miquel, 232, 344–346, 348, 422, 432
Lemoine, 209, 403 Miquel pentagon theorem, 232
Lemoine circles, 209, 211 Miquel point, 344, 348, 349, 438, 439, 720
Lemoine line, 403 Miquel point of four lines, 348
Length, 6 Mirror, 735
Length of broken line, 29 Mirror = Reflection axis, 475
Lhuilier, 583 Mirroring = Reflection, 475, 735
Limit, 663 Monge’s point, 620, 622
Limiting parallels, 507 Monohedral tiling, 516
Lindemann, 448 Morley, 364
Line, 1, 3 Morley’s theorem, 364
INDEX 775
Napoleon’s theorem, 362 Parabola, axis, 694, 695, 719, 723

Negative half line, 367 Parabola, diameter, 694
Negatively oriented, 485, 734 Parabola, directrix, 696, 722
Neusis, 98, 428, 437 Parabola, parameter, 693
Newton, 82 Parabola, vertex, 693
Newton line, 82, 144, 168, 202, 393 Parabolic pencil = Pencil tangent, 251
Nicomedes, 437 Parallel, 4, 6
Nine point circle = Euler circle, 318 Parallel chords, 688
Non coplanar lines, 539 Parallel circle = Small circle, 594
Non coplanar lines = Skew lines, 532, 540 Parallel distance, 171
Non-convex angle, 8 Parallel lines, 26, 33, 38, 40, 41, 68
Null translation, 479, 738 Parallel planes, 532, 538, 551
Parallel projection, 612
Oblique circular cone, 626 Parallelogram, 74, 75, 77, 78, 81, 82, 87, 133, 136,
Obtuse, 12, 24, 86 140, 202, 289, 400
Obtuse triangle, 14 Parallels in space, 533
Opposite, 6, 13, 77, 387, 581 Parallepiped, 581, 622, 635, 737
Order of tiling, 518 Pascal, 401, 408
Orientation, 17, 20, 478, 479, 484, 488, 497, 499, Pascal line, 401
564, 734, 739, 748, 749
Pasch, 14
Orientation preserving transformation, 479
Pedal, 439
Orientation reversing transformation, 479
Pedal triangle, 345
Oriented angle, 484, 739
Pencil center, 170
Oriented equally, 41
Pencil direction, 170
Oriented negatively, 17
Pencil generated, 252
Oriented opposite, 41
Pencil intersecting, 251, 284, 307
Oriented positively, 17
Pencil non-conventional, 257, 276
Origami, 46, 306
Pencil non-intersecting, 251
Origin, 367
Pencil of circles, 248, 293
Orthic, 116, 232, 309, 433
Pencil of lines, 170
Orthic triangle, 312
Pencil orthogonal, 292
Orthocenter, 15, 86, 107, 111, 121, 124, 131, 184,
Pencil parallel, 170
217, 264, 286, 290, 299, 308, 315, 319,
Pencil tangent, 251
321, 323, 326, 342, 349, 351, 363, 386,
Pencil’s basic points, 251
392, 400, 420, 430, 438, 439, 567, 573,
Pencil’s limit points, 251, 292
620, 622, 696, 718
Orthocentric tetrad, 111 Pencil’s radical axis, 251
Orthocentric tetraheder, 622 Penrose tiling, 522
Orthodiagonal, 152, 298, 466 Pentagon, 298, 402, 521
Orthogonal, 12 Pentagon regular, 96, 239
Orthogonal axonometric projection, 612 Perimeter, 14, 71, 109, 128, 441, 446, 448
Orthogonal circles, 259 Perpendicular = Orthogonal, 12
Orthogonal cut, 581, 582, 584, 656, 657 Perspective horizon, 617
Orthogonal from point, 510 Perspective projection, 617, 619
Orthogonal lines, 554 Perspective projection = Central projection, 616
Orthogonal pencil, 261, 276, 286, 509, 514 Perspective relative to line, 393
Orthogonal planes, 558, 737 Perspective triangles, 393, 396
Orthogonal sides, 18, 25, 42, 146, 155, 185, 629 Perspectivity axis, 393
Orthogonal to line, 31, 68 Perspectivity center, 393
Orthogonal to plane, 542, 545, 549 Pick, 227
Orthogonals in space, 539 Pivot, 348, 438
Ovoid, 457 Pivoting, 422
Plane, 1, 531
Pappus, 20, 131, 135, 149, 351, 353, 397, 408 Plane Geometry, 1
Parabola, 674, 675, 680, 684, 693, 700, 719, 723 Plane of symmetry, 736
776 INDEX
Platonic body = Platonic solid, 578 Projection of shape, 544

Platonic solid, 601, 611, 623, 670, 736, 737, 752, Projection onto a plane, 540
758 Projective geometry, 400
Platonic solids, 576 Proper conic sections, 674
Pohlke, 613 Proper similarity, 495, 749
Pohlke’s theorem, 613 Ptolemy, 333, 334
Poincare, 506 Pursuit problems, 503
Point, 1, 3 Pyramid, 569, 570, 601, 651
Point of contact, 596 Pyramid frustum, 570
Point of line at infinity, 283, 399 Pyramid volume, 651, 652
Point of the line at infinity, 382 Pythagoras, 135, 146
Point symmetric, 47, 48, 88 Pythagorean theorem, 163
Point symmetry, 47, 86, 477, 484, 736
Polar, 281, 283–285, 289, 391, 392 Quadrangle = Quadrilateral, 77
Polar circle, 286, 392 Quadrant, 557
Polar relative to conic, 693, 706 Quadratic equation, 292
Pole, 283 Quadrilateral, 77, 80–82, 144, 164
Pole of circle, 594 Quadrilateral bicentric, 298
Pole relative to conic, 693 Quadrilateral complete, 387
Quadrilateral construction, 80, 94, 124, 129, 132,
Polygon, 94
193, 293, 339, 340, 427
Polygon angles, 94
Quadrilateral convex, 77, 113
Polygon convex, 94
Quadrilateral harmonic, 428
Polygon partition, 225
Quadrilateral inscriptible, 335, 505
Polygon perimeter, 95
Quardatic equation, 291
Polygon pivoting, 348
Polygon regular, 95, 97, 99, 108, 109, 115, 125, 133, Radians, 449, 450
218, 241, 447, 451, 460 Radical axis, 241, 264, 434, 623
Polygon regular, perimeter, 447 Radical center, 243, 286, 291, 303, 404
Polygon self-intersecting, 94 Radical plane, 623
Polygon sides, 94 Radius, 57
Polygons congruent, 186 Range, 473
Polygons homothetic, 186 Ratio, 49, 495, 749
Polygons similar, 185, 187, 189 Real numbers R, 441
Polyhedral angle, 569 Rectangle, 87, 137, 308
Polyhedral angles of polyhedron, 577 Rectangular hyperbola, 717, 718
Polyhedron, 577, 601, 645, 656 Rectangular parallepiped, 582, 601, 645
Polyhedron area, 631 Reduction to contradiction, 1, 5, 24, 35
Polyhedron circumscribed, 665 Reflection, 475, 482, 492, 735, 744, 747
Positive half line, 367 Reflection angle, 32
Positively oriented, 485, 734 Reflection axis, 475, 500
Power, 233 Reflective similarity = Antisimilarity, 495
Power of inversion, 271 Regular dodecahedron, 578
Power of point, 623 Regular icosahedron, 578
Power ratio, 255 Regular octahedron, 578
Preimage, 473 Regular pentagon, 625
Preserves the orientation, 734 Regular polygon, 516
Primary, 702 Regular polygon area, 452
Prism, 581, 647 Regular polyhedral angles, 575
Prism volume, 647 Regular polyhedron, 578
Prismatic surface, 581–583, 650 Regular prism, 581
Prismatoid, 669 Regular pyramid, 574
Projection, 31, 60 Regular tetrahedron, 578
Projection of line, 560 Regular tiling, 516
Projection of point, 544 Relative position, 187, 188
INDEX 777
Reverses orientation, 734 Similarity rotational, 495

Rhombicosidodecahedron, 756 Similarly invariant, 188
Rhombicuboctahedron, 754 Simson, 340
Rhombus, 91–94, 719 Simson line, 340, 341, 343–345, 349, 432, 433, 439,
Right angle, 12, 60, 105 720
Right angle rotating, 102, 215, 298 Simson lines angle, 341
Right circular cylindrical, 585 Sine rule, 193, 194
Right prism, 581 Sine spherical formula, 609
Right triangle, 18, 21, 25, 39, 43, 53, 69, 90, 116, Singular conic sections, 674
124, 129, 141, 146–148, 151, 154, 161, Skew lines, 532, 540, 545, 547, 548, 554, 555, 744
167, 189, 287, 724 Skew lines distance, 545
Rotation, 348, 485, 486, 489, 496, 528 Skew orthogonal, 545, 554, 567
Rotation in space, 739 Skew orthogonals, 541, 552
Rotational similarity, 495 Skew quadrilateral, 542
Rotoreflection, 743, 747 Small circle, 594
Soddy, 331
Saccheri, 34, 35 Solid angle, 569
Salinon, 458 Space, 1, 531
Sangaku, 356, 359 Space Geometry, 1
SAS-criterion, 19, 21, 22, 27, 35, 43, 65, 71, 88 Sphere, 580, 593, 596, 598, 601, 603, 610, 622, 627,
Scale, 495 636, 639, 640, 643, 656, 659, 667, 670,
Scalene triangle, 14 676, 685, 732, 734, 736, 742, 749
Secant, 4 Sphere area, 637
Second (of degree), 9 Sphere center, 593
Secondary elements, 15 Sphere diameter, 593
Segment, 4 Sphere exterior, 593
Self-intersecting, 77 Sphere interior, 593
Semi latus rectum, 692 Sphere radius, 593
Semiregular tiling, 518 Sphere volume, 659, 663
Sequence, 441 Spheres congruent, 593
Sequence bounded, 441 Spherical area axioms, 640
Sequence Fibonacci, 470 Spherical cap, 638
Sequence increasing, 441, 443, 447, 451, 471 Spherical cosine 1st formula, 606
Sequence limit, 449, 471 Spherical cosine 2nd formula, 608
Sequence member, 441 Spherical excess, 643
Sequence of polygons, 444 Spherical lune, 602, 642
Sequence term, 444 Spherical polygon, 603, 604, 643
Sequence, Haros-Farey, 469 Spherical polygon vertex, 604
Sequence, limit of, 441–443, 446 Spherical polygon, angle, 603
Sequence, upper bound, 442 Spherical polygons, 640
Shape, 1 Spherical polyhedra, 601
Side, 3 Spherical ring, 663
Side proportionality, 154 Spherical sector, 662
Sides, 74, 387 Spherical sine formula, 609
Signed distance, 367 Spherical triangle, 603, 606, 611, 642
Signed measure of angle, 484 Spherical triangle, apex, 603
Signed ratio, 203, 366 Spherical triangle, right, 609
Similar triangles, 154, 178, 346 Spherical triangle, side, 603
Similarity, 495, 501, 749, 751 Spherical triangle, side length, 603
Similarity center, 265, 285, 527 Spherical triangle, vertex, 603
Similarity centers, 266 Spherical triangles, congruent, 603
Similarity direct, 495 Spherical zone, 638, 664
Similarity proper, 495, 749 Spherical zone/cap area, 639
Similarity reflective = Antisimilarity, 495 Spiral of Theodorus, 148
778 INDEX
Square, 87, 89, 139, 145 Theorem of sines, 194

Square inscription, 422 Theorem of Stewart, 200
Squaring the circle, impossible, 448 Theorem of Thales, 165, 167
SSS-criterion, 14, 19, 20, 22 Theorem of three bars, 372
Start, 4 Three orthogonals theorem, 546
Startpoint, 6 Tile of type pn , 530
Steiner, 22, 76, 206, 351, 353 Tiles, 226
Steiner chains, 354 Tiling, 516, 643
Steiner line, 343, 430, 433, 434, 696 Tiling pentagonal, 518
Stereometry, 531 Transcendental, 448
Stewart, 200, 316 Transformation, 16, 49, 473, 733
Sum of squares, 202 Translation, 479, 484, 494, 738, 741, 744
Sum of triangle angles, 35 Trapezium, 33, 91, 94, 101, 107, 121, 142, 143, 173,
Supplementary, 12, 23, 40, 41 213, 217, 218, 427, 570, 582
Sylvester, 127 Trapezium isosceles, 91, 92, 102
Symmedian, 207, 209, 386, 696 Trapezoid = Trapezium, 91
Symmedian center, 208–210 Triangle, 13
Symmetric, 478, 736 Triangle angle sum, 38
Symmetric relative to axis, 45 Triangle circumscribed, 221
Symmetries, 477 Triangle construction, 16, 67, 69, 72, 86, 90, 109,
Symmetry, 45, 79, 737, 738, 745 115, 120, 123, 126, 129, 130, 146, 157,
Symmetry axis = Axis of symmetry, 45 158, 202, 205, 206, 214, 220, 221, 238,
Synthesis, 70 250, 284, 287, 288, 315, 323, 325, 420,
436, 437, 439, 502
Tangent, 61, 69, 72, 234 Triangle formulas, 212, 420
Tangent circles, 64 Triangle inequalities, 462
Tangent cone, 597 Triangle inequality, 29, 59, 65, 71, 80, 200, 462,
Tangent plane, 588, 594 704, 712
Tangent to conic, 686, 693, 718 Triangle sides, 13
Tangent to sphere, 596 Triangle vertices, 13
Tangent to the conic, 686 Triangle’s exterior, 14
Tangential triangle, 208, 265, 284, 402 Triangle’s interior, 14
Tangents common, 70 Triangles congruent, 14
Taylor, 211 Triangles equal = Triangles congruent, 14
Taylor circle, 211 Triangles homothetic, 493
Tetraheder, 619, 627, 734, 738, 751 Triangles isometric = Triangles congruent, 14
Tetrahedron, 571, 601, 620, 666 Triangles similar, 127, 186, 216
Thales, 135, 165 Trigonometric formulas, 197, 198, 240
Thales in space, 550 Trigonometric functions, 159
Thales theorem, 309 Trigonometric functions: sin, cos, tan, cot, 159
Theodorus of Kyrenia, 148 Triheder, 561, 605, 736
Theorem of Apollonius, 246 Triheder supplementary, 607, 608, 610
Theorem of Ceva, 380 Triheders of the pyramid, 569
Theorem of cosines, 195 Triheders vertical, 565
Theorem of Desargues, 394 Trihedral angle, 590
Theorem of Droz-Farny, 300 Trihedral angle = Triheder, 561
Theorem of Heron, 316 Trilinear polar, 381, 394
Theorem of Lhuilier, 397 Trilinear pole, 382
Theorem of Menelaus, 379 Trioker game, 56
Theorem of Pappus, 149, 397 Triply orthogonal tetraheder, 622
Theorem of Pascal, 401 Trisector, 364
Theorem of Pick, 227 Tritangent circle, 316
Theorem of Ptolemy, 334, 336, 397 Truncated cone, 634
Theorem of Pythagoras, 146, 168 Truncated cone volume, 659
INDEX 779
Truncated cube, 755

Truncated cuboctahedron, 756
Truncated dodecahedron, 754, 755
Truncated icosahedron, 756
Truncated icosidodecahedron, 756
Truncated octahedron, 756
Truncated prism, 582, 620
Truncated pyramid, 570
Truncated pyramid volume, 655
Truncated tetrahedron, 751, 755
Tucker, 211
Tucker circles, 211
Ultraparallels = Hyperparallels, 507

Undefined terms, 1
Unfolding, 635
Unfolding of cone, 591
Unfolding of plane onto other, 583
Unfolding of pyramid, 572
Unfolding of tetrahedron, 572
Unfolding of triheder, 562
Van Aubel, 153

Vanishing line, 617
Vecten, 363
Vecten point, 363
Vecten triangle, 363
Vertex, 387
Vertical = Orthogonal, 12
Vertical triheders, 565
Vertices, 74
Vertices of polyhedron, 577
Viete, 331
Volume, 645
Volume axioms, 645
Volume of cone, 658
Volume of oblique prism, 648
Volume of spherical zone, 664
Wallace, 340
Wallace line = Simson line, 340
Zeno’s paradox, 7

Geo Me Trik On Project

Uploaded by

Copyright:

Available Formats

Geo Me Trik On Project

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Geo Me Trik On Project

Uploaded by

Copyright:

Available Formats

Geometrikon

Copyright: Paris Pamfilos 2011-2017

To Kostas, to Michalis, to Iasonas, to Odysseas

Marcus Aurelius, To himself VII, 9

Herakleion, 16 February 2017

(ABΓ ) Circle going through the points A, B, Γ

1 The basic notions 1

2 Circle and polygons 57

3 Areas, Thales, Pappus, Pythagoras 135

3.5 Similar right triangles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154

4 The power of the circle 233

5 From the classical theorems 311

6 Circle measurement 441

6.6 The area of a circular sector . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 455

7 Transformations of the plane 473

8 Lines and planes in space 531

10 Areas in space, volumes 631

10.7 Volume of pyramids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 651

11 Conic sections 673

12 Transformations in space 733

The basic notions

1.1 Undefined terms, axioms

Thomas Mann, Schopenhauer

(1) A point is that of which there is no part.

(2) And a line is a length without breadth.

(3) And the extremities of a line are points.

Immediately after the 23 definitions follow 5 Postulates which we name axioms:

1.2 Line and line segment

Voltaire, The Ignorant Philosopher

Figure 1.2.1: Line ε

Figure 1.2.2: Line AB

Figure 1.2.3: Half planes defined by a line

Figure 1.2.4: Line segment AB

Figure 1.2.5: A and B on different half plane of the line ε

Figure 1.2.6: Parallels AB and A B

Figure 1.2.7: Intersecting lines ε and ε

Figure 1.2.8: A, B on the same side of ε

Figure 1.2.9: Infinitely many lines through O

1.3 Length, distance

J. L. Borges, A Weary Man’s Utopia

Figure 1.3.1: |AB| = |AE| + |EB|

|AE| + |EB| = |AB|.

Figure 1.3.3: B between A and E

Figure 1.3.4: Doubling the length of AB

Figure 1.3.5: Zeno’s paradox

Confucius, Analects 7.8

Figure 1.4.2: Flat angle Null angle

Figure 1.4.3: Full turn

the half line OX.

The basic properties (axioms) of angles are the following:

’ (convex or not) we associate a number | XOY

Figure 1.4.4: Equal angles on two sides of OX ’ | = |XOP

’ (convex or not) the measures of angles XOY

’ , there exists exactly one half line

Figure 1.4.5: Angle of segments AB and AΓ Sum of equal angles

’ . Show that the measure of a full turn is 360 degrees.

’ of measure ω, we construct equal and successive adjacent angles

’ with measure |XOY

’ and a point P on the opposite half line OX of OX, as well