The Fabulous Fibonacci Numbers TQW - Darksiderg

The (Fabulous)

FIBONACCINumbers

Alfred S. Posamentier
Ingmar Lehmann
Afterword by
Herbert A.
Hauptman
Lobel Laureate

~ Prometheus Bcxlks
59 John Glenn Drive
Amherst, New York 14228-2197
Published 2007 by Prometheus Books
The Fabulous Fibonacci Numbers. Copyright © 2007 by Alfred S. Posamentier

and Ingmar Lehmann. All rights reserved. No part of this publication may
be reproduced, stored in a retrieval system, or transmitted in any form or by
any means, digital, electronic, mechanical, photocopying, recording, or
otherwise, or conveyed via the Internet or a Web site without prior written
permission of the publisher, except in the case of brief quotations embodied
in critical articles and reviews.
Inquiries should be addressed to

Prometheus Books
59 John Glenn Drive
Amherst, New York 14228-2197
VOICE: 716-691-0133, ext. 207
FAX: 716-564-2711
WWW.PROMETHEUSBOOKS.COM
11 10 09 08 07 5 4 3 2 1
Library of Congress Cataloging-in-Publication Data
Posamentier, Alfred S.
The fabulous Fibonacci numbers / by Alfred S. Posamentier and
Ingmar Lehmann.
p. cm.
Includes bibliographical references and index.
ISBN 978-1-59102-475-0 (alk. paper)
1. Fibonacci numbers. I. Lehmann, Ingmar. II. Title.
QA241.P665 2007
512.7'2-dc22
2006035406
Printed in the United States of America on acid-free paper

To Barbara for her support, patience, and inspiration.
To my children and grandchildren: David, Lisa, Danny, Max,
and Sam, whose future is unbounded.
And in memory of my beloved parents, Alice and Ernest, who
never lost faith in me.
Alfred S. Posamentier
To my wife and life partner, Sabine, without whose support

and patience my work on this book would not have been possible.
And to my children and grandchildren: Maren, Claudia, Simon,
and Miriam.
Ingmar Lehmann
Contents
Acknowledgments 9
Introduction 11
Chapter 1: A History and Introduction to the

Fibonacci Numbers 17
Chapter 2: The Fibonacci Numbers in Nature 59
Chapter 3: The Fibonacci Numbers and the

Pascal Triangle 77
Chapter 4: The Fibonacci Numbers and the
Golden Ratio 107
Chapter 5: The Fibonacci Numbers and
Continued Fractions 161
Chapter 6: A Potpourri of Fibonacci Number
Applications 177
Chapter 7: The Fibonacci Numbers Found in Art
and Architecture 231
Musical Form 271
7
8 CONTENTS
Chapter 9: The Famous Binet Formula for Finding

a Particular Fibonacci Number 293

Fractals 307
Epilogue 327
Afterword by Herbert A. Hauptman 329
Appendix A: List of the First 500 Fibonacci Numbers,

with the First 200 Fibonacci Numbers
Factored 343
Appendix B: Proofs of Fibonacci Relationships 349
References 371
Index 375
Acknowledgments
T he authors wish to thank Professor Stephen Jablonsky of the

City College of New York-CUNY, who used his multifaceted
talents to present the unusual appearances of the Fibonacci num-
bers in the realm of music from the design of compositions to the
design of musical instruments.
Professor Ana Lucia B. Dias deserves our sincere thanks for
her broad discussion of the occurrence of the Fibonacci numbers in
the field of fractals.
Our dear friend Dr. Herbert A. Hauptman, the acknowledged
first mathematician to win the Nobel Prize (Chemistry, 1985), pro-
vided a stimulating afterword for the book. This capstone piece
will provide a challenge to the reader and may even have some
elements never before discovered.
It was during a conversation with our friend James S. Tisch,
president and CEO of the Loews Corporation, that he mentioned
the applications of the Fibonacci numbers in the world of eco-
nomics. He provided much information, and we took it the rest of
the way to the pages of this book. For this, we thank him.
Dr. Lehmann acknowledges the occasional support of Tristan
Vincent in helping him identify the right English words to best ex-
press his ideas.
We wish to also thank Professor Andreas Filler, of the Pada-
gogische Hochschule at Heidelberg (Germany); Heino Hellwig,
Humboldt University, Berlin; and Hans-Peter Ludtke, at the Hein-
rich-Hertz Gymnasium, Berlin, for providing some valuable com-
ments during the writing of this book.
9
10 ACKNOWLEDGMENTS
The success of a book is usually dependent on two types of

support: our faithful guide through the development of the project
for which we thank Linda Greenspan Regan, who gave us constant
advice to make the book as appealing to the general readership as
possible, and the content editor, Peggy Deemer, who, as usual, did
a magnificent job editing the manuscript-no mean feat, given the
complexity of the subject.
Naturally, we wish to thank, respectively, Barbara and Sabine,
for their encouragement, patience, and support during our work on
this book.
Introduction
The Fabulous Fibonacci Numbers
I n a remote section of the Austrian Alps, there is a long-

abandoned salt mine entrance with a cornerstone bearing the in-
scription "anno 1180." It refers to the year that the mine was estab-
lished. Clearly there is something wrong with this designation.
Scholars have determined that the first published use of the Hindu
numerals (our common numerals) in the Western world was in
1202. It was in this year that Leonardo of Pisa (Leonardo Pisano),
more commonly known as Fibonacci (pronounced: fee-boh-NACH-
ee), published his seminal work Liber Abaci, or "book of calcula-
tion." He began chapter 1 of his book with:
The nine Indian 1 figures are: 9 8 7 6 5 4 3 2 1.

With these nine figures, and with the sign 0, which
the Arabs call zephyr, any number whatsoever is written.
This constitutes the first formal mention of our base-ten numeral

system in the Western world. There is speculation, however, that
Arabs already introduced these numerals informally and locally in
Spain in the second half of the tenth century.
Unlike luminaries from the past, whose fame today is largely
based on a single work, such as Georges Bizet (1838-1875) for
Carmen, Engelbert Humperdinck (1854-1921) for Hansel und
1. Fibonacci used the term Indian figures for the Hindu numerals.
11
12 INTRODUCTION
Gretel, or J. D. Salinger (b. 1919) for The Catcher in the Rye, we

should not think of Fibonacci as a mathematician who is known
only for this now-famous sequence of numbers that bears his name.
He was one of the greatest mathematical influences in Western cul-
ture and unquestionably the leading mathematical mind of his times.
Yet it was this sequence of numbers, emanating from a problem on
the regeneration of rabbits, that made him famous in today's world.
Fibonacci was a serious mathematician, who first learned
mathematics in his youth in Bugia, a town on the Barbary Coast of
Africa, which had been established by merchants from Pisa. He
traveled on business throughout the Middle East and along the way
met mathematicians with whom he entered into serious discussions.
He was familiar with the methods of Euclid (ca. 300 BeE) and used
those skills to bring to the European people mathematics in a very
usable form. His contributions include introducing a practical nu-
meration system, calculating algorithms and algebraic methods, and
a new facility with fractions, among others. The result was that the
schools in Tuscany soon began to teach Fibonacci's form of calcu-
lation. They abandoned the use of the abacus, which involved
counting beads on a string and then recording their results with Ro-
man numerals. This catapulted the discipline of mathematics for-
ward, since algorithms were not possible with these cumbersome
symbols. Through his revolutionary book and other subsequent
publications, he made dramatic changes in western Europe's use and
understanding of mathematics.
Unfortunately, Fibonacci's popularity today does not encompass
these most important innovations. Among the mathematical prob-
lems Fibonacci poses in chapter 12 of Liber Abaci, there is one
about the regeneration of rabbits. Although its statement is a bit
cumbersome, its results have paved the way for a plethora of
monumental ideas, which has resulted in his fame today. The prob-
lem, which is stated on page 25 (figure 1-2), shows the monthly
count of rabbits as the following sequence of numbers: 1,1,2,3,5,
8, 13,21,34,55, 89, 144,233,377, ... , which is known today as
the Fibonacci numbers. At first sight one may wonder what makes
this sequence of numbers so revered. A quick inspection shows that
INTRODUCTION 13
this sequence of numbers can go on infinitely, as it begins with two

1s and continues to get succeeding terms by adding, each time, the
last two numbers to get the next number (i.e., 1 + 1 = 2, 1 + 2 = 3,
2 + 3 = 5, and so on). By itself, this is not very impressive. Yet, as
you will see, there are no numbers in all of mathematics as ubiquitous
as the Fibonacci numbers. They appear in geometry, algebra, number
theory, and many other branches of mathematics. However, even
more spectacularly, they appear in nature; for example, the number of
spirals of bracts on a pinecone is always a Fibonacci number, and,
similarly, the number of spirals of bracts on a pineapple is also a
Fibonacci number. The appearances in nature seem boundless. The
Fibonacci numbers can be found in connection with the arrangement
of branches on various species of trees, as well as in the number of
ancestors at every generation of the male bee on its family tree.
There is practically no end to where these numbers appear.
Throughout these pages we will explore many manifestations, so
that you may be motivated to find other instances in nature where
these Fibonacci numbers surface. The book will also provide you with
a gentle, yet illuminating discussion of the unusual nature of these
numbers. Their interrelationships with other seemingly completely
unrelated aspects of mathematics will open the door to applications in
a variety of other fields and areas as remote as the stock market.
It is our desire that this book serve as your introduction to these
fabulous numbers. We will provide you with the development of
the Fibonacci numbers, a sort of history, one might say. Then we
will investigate the many "sightings" of these numbers by subject
area. For example, in geometry we will explore their relationship
with the most beautiful ratio, known as the "golden ratio.,,2 Here
the Fibonacci numbers, when taken as quotients in consecutive
pairs, approach the golden ratio:
l/J = 1.6180339887498948482045868343656 ...
2. The golden ratio will be discussed in reasonable depth (with ample applica-
tions in geometry and art) so that the reader will have a true appreciation for it
and, consequently, its relationship to the Fibonacci numbers. It is often symbol-
ized by the Greek letter l/J.
14 INTRODUCTION
The larger the Fibonacci numbers, the closer their quotient ap-
proaches the golden ratio.
Consider, for example, the quotient of the relatively small pair
of consecutive Fibonacci numbers:
13
- == 1.625
8
Then consider the quotient of the somewhat larger pair of con-

secutive Fibonacci numbers:
55
_ = 1.6176470588235294117 3
34
and the quotient of an even larger pair of consecutive Fibonacci

numbers:
144
- = 1.61797752808988764044943820224719101123595505
89
Notice how these increasingly larger quotients seem to sur-

round the actual value of the golden ratio, l/J. Then, when we take
much larger pairs of consecutive Fibonacci numbers, their quo-
tients get us ever closer to the actual value of the golden ratio, l/J.
For example,
4 181
, =1.6180340557275541795665634674923 ...
2,584
165 580 141

, , =1.618033988749894890909100680999418033988 ... 4
102,334,155
3. The bar above the last repeat of 16 indicates that the 16 digits repeat indefi-
nitely.
4. The fortieth Fibonacci number is 102,334,155 and the forty-first is
165,580,141.
INTRODUCTION 15
Compare this last quotient to the value of the golden ratio:

l/> == 1.6180339887498948482045868343656 ...
We will also discuss the nature and glory of the golden ratio
itself. Its appearance in architecture and art is more than coinci-
dental. If you were to sketch a rectangle encasing the front view
of the Parthenon in Athens, Greece, you will have drawn a golden
rectangle-that is, one whose length and width are in the golden
ratio. Many artists have employed the golden rectangle in their
art. The painting of Adam and Eve by the famous medieval Ger-
man artist Albrecht Durer (1471-1528), for instance, has the
subjects encased in a golden rectangle.
Interestingly enough, the Fibonacci numbers did not get their
prominence, and their name, until Edouard Lucas (1842-1891), a
French mathematician, studied them in the second half of the
nineteenth century. He mused about the start of the sequence,
wondering what would happen if the sequence had begun with 1
and 3, rather than 1 and 1. So he studied this new sequence (fol-
lowing the same additive rule) and then compared it to the
Fibonacci sequence. The Lucas numbers, which are 1, 3, 4, 7, 11,
18,29,47,76, 123, ... , interrelate with the Fibonacci numbers.
We will explore their relationship later.
There is almost no end to the places where the Fibonacci num-
bers appear and their many applications. We will present mathe-
matical recreations, as well as more serious properties of these
numbers-all of which will fascinate the uninitiated as well as the
math-savvy reader. You will be truly awed by these magnificent
numbers and most likely will be continuously consciously seeking
your own sightings of the Fibonacci numbers. Throughout this
book we hope to appeal to all types of readers, but always have the
general reader in mind. For the more mathematically advanced
reader, however, we provide an appendix with proofs of statements
made throughout this book. Our goal is to evoke the power and
beauty of mathematics for all readers.
Chapter 1
A History and Introduction to

the Fibonacci Numbers
W ith the dawn of the thirteenth century, Europe began to

wake from the long sleep of the Middle Ages and perceive
faint glimmers of the coming Renaissance. The mists rose slowly
as the forces of change impelled scholars, crusaders, artists, and
merchants to take their tentative steps into the future. Nowhere
were these stirrings more evident than in the great trading and
mercantile cities of Italy. By the end of the century, Marco Polo
(1254-1324) had journeyed the Great Silk Road to reach China,
Giotto di Bondone (1266-1337) had changed the course of paint-
ing and freed it from Byzantine conventions, and the mathemati-
cian Leonardo Pisano, best known as Fibonacci, changed forever
Western methods of calculation, which facilitated the exchange of
currency and trade. He further presented mathematicians to this
day with unsolved challenges. The Fibonacci Association, started
in 1963, is a tribute to the enduring contributions of the master.
Leonardo Pisano-or Leonardo of Pisa, Fibonacci I-his name
as recorded in history, is derived from the Latin "filius Bonacci,"
or a son of Bonacci, but it may more likely derive from "de filiis
1. It is unclear who first used the name Fibonacci~ however, it seems to be at-
tributed to Giovanni Gabriello Grimaldi (1757-1837) at around 1790, or to Pie-
tro Cossali (1748-1815).
17
18 THE FABULOUS FIBONACCI NUMBERS
Bonacci," or family of Bonacci. He was born to Guilielmo (Wil-

liam) Bonacci and his wife in the port city of Pisa, Italy, around
1175, shortly after the start of construction of the famous bell
tower, the Leaning Tower of Pisa. These were turbulent times in
Europe. The Crusades were in full swing and the Holy Roman
Empire was in conflict with the papacy. The cities of Pisa, Genoa,
Venice, and Amalfi, although frequently at war with each other,
were maritime republics with specified trade routes to the Mediter-
ranean countries and beyond. Pisa had played a powerful role in
commerce since Roman times and even before as a port of call for
Greek traders. Early on it had established outposts for its com-
merce along its colonies and trading routes.
Figure 1-1
Leonardo of Pisa (Fibonacci).
In 1192 Guilielmo Bonacci became a public clerk in the cus-

toms house for the Republic of Pisa, which was stationed in the
Pisan colony of Bugia (later Bougie, and today Bejaia, Algeria) on
the Barbary Coast of Africa. Shortly after his arrival he brought his
son, Leonardo, to join him so that the boy could learn the skill of
calculating and become a merchant. This ability was significant
since each republic had its own units of money and traders had to
A HISTORY AND INTRODUCTION TO THE FIBONACCI NUMBERS 19
calculate monies due them. This meant determining currency

equivalents on a daily basis. It was in Bugia that Fibonacci first
became acquainted with the "nine Indian figures," as he called the
Hindu numerals and "the sign 0 which the Arabs call zephyr." He
declares his fascination for the methods of calculation using these
numerals in the only source we have about his life story, the pro-
logue to his most famous book, Liber Abaci. During his time away
from Pisa, he received instruction from a Muslim teacher, who in-
troduced him to a book on algebra titled Hisab al-jabr w'al-
muqabalah2 by the Persian mathematician al-Khowarizmi (ca.
780-ca. 850), which also influenced him.
During his lifetime, Fibonacci traveled extensively to Egypt,
Syria, Greece, Sicily, and Provence, where he not only conducted
business but also met with mathematicians to learn their ways of
doing mathematics. Indeed Fibonacci sometimes referred to him-
self as "Bigollo," which could mean good-for-nothing or, more
positively, traveler. He may have liked the double meaning. When
he returned to Pisa around the tum of the century, Fibonacci began
to write about calculation methods with the Indian numerals for
commercial applications in his book, Liber Abaci. 3 The volume is
mostly loaded with algebraic problems and those "real-world"
problems that require more abstract mathematics. Fibonacci
wanted to spread these newfound techniques to his countrymen.
Bear in mind that during these times, there was no printing
press, so books had to be written by the hands of scribes, and if a
copy was to be made, that, too, had to be hand written. Therefore we
are fortunate to still have copies of Liber Abaci, which first appeared
in 1202 and was later revised in 1228.4 Among Fibonacci's other
books is Practica geometriae (1220), a book on the practice of ge-
ometry. It covers geometry and trigonometry with a rigor compara-
ble to that of Euclid's work, and with ideas presented in proof form
as well as in numerical form, using these very convenient numerals.
2. The name "algebra" comes from the title of this book.

3. The title means "book on calculation"; it is not about the abacus.
4. An English-language version of Liber Abaci-the first in a modern lan-
guage-of the 1228, Latin version of 1857 by Baldassarre Boncompagni was
recently published by Laurence E. Sigler (New York: Springer-Verlag, 2002).
Here Fibonacci uses algebraic methods to solve geometric problems

as well as the reverse. In 1225 he wrote Flos (on flowers or blos-
soms) and Liber quadratorum (or "Book of Squares"), a book that
truly distinguished Fibonacci as a talented mathematician, ranking
very high among number theorists. Fibonacci likely wrote additional
works that are lost. His book on commercial arithmetic, Di minor
guisa, is lost, as is his commentary on Book X of Euclid's Elements,
which contained a numerical treatment of irrational numbers 5 as
compared to Euclid's geometrical treatment.
The confluence of politics and scholarship brought Fibonacci
into contact with the Holy Roman Emperor Frederick II (1194-
1250) in the third decade of the century. Frederick II, who had been
crowned king of Sicily in 1198, king of Germany in 1212, and then
crowned Holy Roman Emperor by the pope in St. Peter's Cathedral
in Rome (1220), had spent the years up to 1227 consolidating his
power in Italy. He supported Pisa, then with a population of about
ten thousand, in its conflicts with Genoa at sea and with Lucca and
Florence on land. As a strong patron of science and the arts, Freder-
ick II became aware of Fibonacci's work through the scholars at his
court, who had corresponded with Fibonacci since his return to Pisa
around 1200. These scholars included Michael Scotus (ca. 1175-{;a.
1236), who was the court astrologer and to whom Fibonacci dedi-
cated his book Liber Abaci; Theodorus Physicus, the court philoso-
pher; and Dominicus Hispanus, who suggested to Frederick II that
he meet Fibonacci when Frederick's court met in Pisa around 1225.
The meeting took place as expected within the year.
Johannes of Palermo, another member of Frederick II's court,
presented a number of problems as challenges to the great mathe-
matician Fibonacci. Fibonacci solved three of these problems. He
provided solutions in Flos, which he sent to Frederick II. One of
these problems, taken from Omar Khayyam's (1048-1122) book on
algebra, was to solve the equation: x 3+ 2X2 + lOx = 20. Fibonacci
knew that this was not solvable with the numerical system then in
place-the Roman numerals. He provided an approximate answer,
5. Irrational numbers are those that cannot be expressed as a common fraction

involving integers.
since he pointed out that the answer was not an integer, nor a frac-
tion, nor the square root of a fraction. Without any explanation, he
gives the approximate solution in the form of a sexagesimal 6 number
as 1.22.7.42.33.4.40, which equals:
22 7 42 33 4 40
1 + - +2 - +3 - +4- + -
5
+-
6
60 60 60 60 60 60
However, with today's computer algebra system we can get the

proper (real) solution-by no means trivial! It is
2~3, 930 352 2~3, 930 352 2

x ==- 3 _--+ 3 +----
9 27 9 27 3
~ 1.3688081075
Another of the problems with which he was challenged is one

we can explore here, since it doesn't require anything more than
some elementary algebra. Remember that although these methods
may seem elementary to us, they were hardly known at the time of
Fibonacci, and so this was considered a real challenge. The prob-
lem was to find the perfect square 7 that remains a perfect square
when increased or decreased by 5.
Fibonacci found the number 4112
as his solution to the problem.
To check this, we must add and subtract 5 and see if the result is
still a perfect square.
~)2 +5=
( 12 1,681 + 720 =2,401 =(49)2
144 144 144 12
( _
~)2
12
5 = 1,681 _ 720 = 961 =
144 144 144
(~)2
12
We have then shown that 41

12
meets the criteria set out in the
problem. Luckily the problem asked for 5 to be added and sub-
tracted from the perfect square, for if he were asked to add or sub-
6. A number in base 60.

7. A perfect square is the square of an integer, e.g., 16, 36, or 81.
tract 1, 2, 3, or 4 instead of 5, the problem could not have been

solved. For a more general solution of this problem, see appendix B.
The third problem, also presented in Flos, is to solve the following:
Three people are to share an amount of money in the

following parts: ~,~, and ~. Each person takes some
money from this amount of money until there is nothing
left. The first person then returns ~ of what he took. The
2
second person then returns ! of what he took, and the

3
third person returns ! of what he took. When the total of

6
what was returned is divided equally among the three,
each has his correct share, namely, ~,~, and ~. How
much money was in the original amount and how much
did each person get from the original amount of money?
Although none of Fibonacci's competitors could solve any of

these three problems, he gave as an answer of 47 as the smallest
amount, yet he claimed the problem was indeterminate.
The last mention of Fibonacci was in 1240, when he was hon-
ored with a lifetime salary by the Republic of Pisa for his service to
the people, whom he advised on matters of accounting, often pro
bono.
The Book Liber Abaci
Although Fibonacci wrote several books, the one we will focus on

is Liber Abaci. This extensive volume is full of very interesting
problems. Based on the arithmetic and algebra that Fibonacci had
accumulated during his travels, it was widely copied and imitated,
and, as noted, introduced into Europe the Hindu-Arabic place-
valued decimal system along with the use of Hindu-Arabic numer-
als. The book was increasingly widely used for the better part of
the next two centuries-a best seller!
He begins Liber Abaci with the following:
The nine Indian 8 figures are: 9 8 7 6 5 4 3 2 1.9

With these nine figures, and with the sign 0, which
the Arabs call zephyr, any number whatsoever is written,
as demonstrated below. A number is a sum of units, and
through the addition of them the number increases by
steps without end. First one composes those numbers,
which are from one to ten. Second, from the tens are
made those numbers, which are from ten up to one hun-
dred. Third, from the hundreds are made those numbers,
which are from one hundred up to one thousand ... and
thus by an unending sequence of steps, any number
whatsoever is constructed by joining the preceding num-
bers. The first place in the writing of the numbers is at
the right. The second follows the first to the left.
Despite their relative facility, these numerals were not widely

accepted by merchants, who were suspicious of those who knew
how to use them. They were simply afraid of being cheated. We
can safely say that it took the same three hundred years for these
numerals to catch on as it did for the completion of the Leaning
Tower of Pisa.
Interestingly, Liber Abaci also contains simultaneous linear
equations. Many of the problems that Fibonacci considers, how-
ever, were similar to those appearing in Arab sources. This does
not detract from the value of the book, since it is the collection of
solutions to these problems that makes the major contribution to
our development of mathematics. As a matter of fact, a number of
mathematical terms-common in today's usage-were first intro-
duced in Liber Abaci. Fibonacci referred to "factus ex multiplica-
tione,,,IO and from this first sighting of the word, we speak of the
"factors of a number" or the "factors of a multiplication." Another
example of words whose introduction into the current mathematics
vocabulary seems to stem from this famous book are the words
"numerator" and "denominator."
8. Fibonacci used the term Indian figures for the Hindu numerals.
9. It is assumed that Fibonacci wrote the numerals in order from right to left,
since he took them from the Arabs, who write in this direction.
1O. David Eugene Smith, History of Mathematics, vol. 2 (New York: Dover,
1958), p. 105.
The second section of Liber Abaci includes a large collection

of problems aimed at merchants. They relate to the price of goods,
how to convert between the various currencies in use in Mediterra-
nean countries, how to calculate profit on transactions, and prob-
lems that had probably originated in China.
Fibonacci was aware of a merchant's desire to circumvent the
church's ban on charging interest on loans. So he devised a way to
hide the interest in a higher initial sum than the actual loan, and
based his calculations on compound interest.
The third section of the book contains many problems, such as:
A hound whose speed increases arithmetically chases a
hare whose speed also increases arithmetically. How far
do they travel before the hound catches the hare?
A spider climbs so many feet up a wall each day and
slips back a fixed number each night. How many days
does it take him to climb the wall?
Calculate the amount of money two people have after a
certain amount changes hands and the proportional in-
crease and decrease are given.
There are also problems involving perfect numbers, II problems

involving the Chinese remainder theorem, and problems involving
the sums of arithmetic and geometric series. Fibonacci treats num-
bers such as J10 in the fourth section, both with rational 12 ap-
proximations and with geometric constructions.
Some of the classical problems, which are considered recrea-
tional mathematics today, first appeared in the Western world in
Liber Abaci. Yet the technique for solution was always the chief
concern for introducing the problem. This book is of interest to us,
not only because it was the first publication in Western culture to
use the Hindu numerals to replace the clumsy Roman numerals, or
because Fibonacci was the first to use a horizontal fraction bar, but
1 1. Perfect numbers are those where the sum of the number's proper factors is
equal to the number itself. For example, the proper factors of 6 are 1, 2, and 3.
The sum of these factors is 1 + 2 + 3 = 6, and therefore 6 is a perfect number.
The next larger perfect number is 28.
12. Rational numbers are those that can be expressed as common fractions, in-
volving integers.
because it casually includes a recreational mathematics problem in

chapter 12 that has made Fibonacci famous for posterity. This is the
problem on the regeneration of rabbits.
The Rabbit Problem
Figure 1-2 shows how the problem was stated (with the marginal
note included):
Beginning 1 A certain man had one pair of rabbits together in a certain enclosed
place, and one wishes to know how many are created from the pair
in one year when it is the nature of them in a single month to bear
another pair, and in the second month those born to bear also. Be-
cause the above written pair in the first month bore, you will dou-
First 2 ble it; there will be two pairs in one month. One of these, namely
the first, bears in the second month, and thus there are in the sec-
Second 3 ond month 3 pairs; of these in one month two are pregnant and in
Third 5 the third month 2 pairs of rabbits are born and thus there are 5
pairs in the month; in this month 3 pairs are pregnant and in the
Fourth 8 fourth month there are 8 pairs, of wh ich 5 pairs bear another 5
Fifth 13 pairs; these are added to the 8 pairs making 13 pairs in the fifth
month; these 5 pairs that are born in this month do not mate in this
month, but another 8 pairs are pregnant, and thus there are in the
Sixth 21 sixth month 21 pairs; to these are added the 13 pairs that are born
Seventh 34 in the seventh month; there will be 34 pairs in this month; to this
are added the 21 pairs that are born in the eighth month; there will
Eighth 55 be 55 pairs in this month; to these are added the 34 pairs that are
Ninth 89 born in the ninth month; there will be 89 pairs in this month; to
these are added again the 55 pairs that are born in the tenth month;
Tenth 144 there will be 144 pairs in this month; to these are added again the
Eleventh 233 89 pairs that are born in the eleventh month; there will be 233
pairs in this month. To these are still added the 144 pairs that are
Twelfth 377 born in the last month; there will be 377 pairs and this many pairs
are produced from the above-written pair in the mentioned place at
the end of one year.
You can indeed see in the margin how we operated, namely, that
we added the first number to the second, namely the 1 to the 2, and
the second to the third and the third to the fourth and the fourth to
the fifth, and thus one after another until we added the tenth to the
eleventh, namely the 144 to the 233, and we had the above-written
sum of rabbits, namely 377 and thus you can in order find it for an
unending number of months."
Figure 1-2
To see how this problem would look on a monthly basis, con-

sider the chart in figure 1-3. If we assume that a pair of baby (B)
rabbits matures in one month to become offspring-producing adults

(A), then we can set up the following chart:
No. of Pairs of No. of Pairs Total
Month Pairs
Adults ~A} of Babies ~B! Pairs
Jan. 1 0
/""
Feb. 1
/'" ~ 2
Mar. 1
Apr. 1 A
/\
/:) t:::-,
B
\ ""
2
3 2
3
5
May 1
June 1
A J /t t\
A B A A B A
\
B
B
\
A
t\ \ \"'"
A B A A B
5
8
3
5
8
13
July 1 13 8 21
Aug. 1 21 13 34
Sept. 1 34 21 55
Oct. 1 55 34 89
Nov. I 89 55 144
Dec. 1 144 89 233
Jan. I 233 144 377
Figure 1-3
This problem generated the sequence of numbers:

1, 1, 2, 3, 5, 8, 13, 21, 34, 55, 89, 144, 233, 377, ... ,
which is known today as the Fibonacci numbers. At first glance
there is nothing spectacular about these numbers beyond the rela-
tionship that would allow us to generate additional numbers of the
sequence quite easily. We notice that every number in the sequence
(after the first two) is the sum of the two preceding numbers.
The Fibonacci sequence can be written in a way that its recursive
defmition becomes clear: each number is the sum of the two preceding
ones.
1+1=2
1+2=3
2+3=5
3+5=8
5+8= 13
8+ 13=21
13 + 21 = 34
21 + 34 = 55
34 + 55 = 89
55 + 89 = 144
89 + 144 = 233
144 + 233 = 377
233 + 377 = 610
377 + 610 = 987
610+987= 1,597 ...
The Fibonacci sequence is the oldest known (recursive) recurrent

sequence. There is no evidence that Fibonacci knew of this rela-
tionship, but it is securely assumed that a man of his talents and
insight knew the recursive l3 relationship. It took another four hun-
dred years before this relationship appeared in print.
Introducing the Fibonacci Numbers
These numbers were not identified as anything special during the

time Fibonacci wrote Liber Abaci. As a matter of fact, the famous
German mathematician and astronomer Johannes Kepler (1571-
1630) mentioned these numbers in a 1611 publication 14 when he
wrote of the ratios: "as 5 is to 8, so is 8 to 13, so is 13 to 21 al-
most." Centuries passed and the numbers still went unnoticed. In
the 1830s C. F. Schimper and A. Braun noticed that the numbers
appeared as the number of spirals of bracts on a pinecone. In the
mid 1800s the Fibonacci numbers began to capture the fascination
of mathematicians. They took on their current name ("Fibonacci
, 15
numbers") from Franyois-Edouard-Anatole Lucas (1842-1891),
the French mathematician usually referred to as "Edouard Lucas,"
who later devised his own sequence by following the pattern set by
Fibonacci. Lucas numbers form a sequence of numbers much like
the Fibonacci numbers and also closely related to the Fibonacci
numbers. Instead of starting with 1, 1, 2, 3, ... , Lucas thought of
beginning with 1, 3, 4, 7, 11 .... Had he started with 1, 2, 3, 5, 8,
... , he would have ended up with a somewhat truncated version of
13. The recursive relationship comes from the notion that each number is gener-
ated from its two predecessors. This will be explored in more detail later in the
book-in chapter 9.
14. Maxey Brooke, "Fibonacci Numbers and Their History through 1900," Fi-
bonacci Quarterly 2 (April 1964): 149.
15. Lucas is also well known for his invention of the Tower of Hanoi puzzle and
other mathematical recreations. The Tower of Hanoi puzzle appeared in 1883
under the name of M. Claus. Notice that Claus is an anagram of Lucas.
His four-volume work on recreational mathematics (1882-1894) has be-
come a classic. Lucas died as the result of a freak accident at a banquet when a
plate was dropped and a piece flew up and cut his cheek. He died of erysipelas a
few days later.
28 THE FABULOUS FiBONACCI NUMBERS
the Fibonacci sequence. (We will inspect the Lucas numbers later
in this chapter.)
At about this time the French mathematician Jacques-Philippe-
Marie Binet (1786- 1856) developed a fonnula for finding any Fi-
bonacci number given its position in the sequence. That is, with
Binet's fonnula we can find the I I 8th Fibonacci number without
having to list the previous 117 numbers. (We will have the oppor-
tunity to use this fonnula in chapter 9.)
Figure 1-4
Fran~ois- Edouard-Anatole Lucas.'6
Today, these numbers still hold the fascination of mathemati-

cians around the world. The Fibonacci Association was created in
the early 1960s to provide enthusiasts an opportunity to share ideas
about these intriguing numbers and their applications. Through
The Fibonacci Quarterly,17 their official publication, many new
facts, applications, and relationships about them can be shared
worldwide. According to its official Web site, http://www.mses
.dal.ca/Fibonacci/, "The Fibonacci Quarterly is meant to serve as a
focal point for interest in Fibonacci numbers and related questions,
especially with respect to new results, research proposals, chal-
lenging problems, and innovative proofs of old ideas."
Still one may ask, what is so special about these numbers?
That is what we hope to be able to demonstrate throughout this
16. Pennission of Francis Lucas http://edouardlucas.free.fr/.

17. Gerald E. Bergum, editor, Department of Computer Science, South Dakota
State University, Brookings, SD 57007.
book. Before we explore the vast variety of examples in which we

find the Fibonacci numbers, let us begin by simply inspecting this
famous Fibonacci number sequence and some of the remarkable
properties it has.
We will use the symbol F, to represent the seventh Fibonacci
number, and Fn to represent the nth Fibonacci number, or, as we
say, the general Fibonacci number-that is, any Fibonacci number.
Let us look at the first thirty Fibonacci numbers (figure 1-5). No-
tice that after we begin with the first two 1s, in every case after
them, each number is the sum of the two preceding numbers.
F I =1 Fi6 = 987
F; == 1 Fi7 = 1,597
F; = 2 Fi8 = 2,584
Fi9 = 4,181
F4 =3
F20 = 6,765
~ ==5
F21 = 10,946
~=8 F22 = 17, 711
F, = 13 F23 = 28,657
F8 = 21 F24 = 46,368
F9 = 34 F2S = 75,025
Fio = 55 F26 = 121,393
Fil = 89 F27 = 196,418
Fi2 = 144 F28 = 31 7 , 811
Fi3 == 233 F29 = 514,229
Fi4 = 377 F10 = 832,040
Fis =610
Figure 1-5
In appendix A we provide you with a list of the first five hun-

dred Fibonacci numbers. You may discover certain wonders there.
For example, if you look at the list, you will notice that (as you
would expect) the number of digits of the Fibonacci numbers in-
creases steadily as the numbers get larger. The French mathemati-
cian Gabriel Lame (1795-1870) proved that in the Fibonacci se-
quence there must be either at least four numbers and at most five
numbers with the same number of digits. In other words, there

would never be only three Fibonacci numbers with a certain num-
ber of digits, nor would there be as many as six Fibonacci numbers
with a specified number of digits. Interestingly, Lame did not use
the name "Fibonacci numbers." So, as a consequence, they were
often called "Lame numbers" because of his proof.
Another oddity that is easily noticeable by a quick inspection of
the Fibonacci numbers is seen if you look at the sixtieth through the
seventieth Fibonacci numbers, which we call F60 and F,o . You will
notice a curious pattern emerging with their last digits. Yes, they
form the Fibonacci sequence (last digits) starting from the beginning:
0, 1, 1, 2, 3, 5, 8, 13, 21, 34, 55. 18 This will continue on and on.
This form of a pattern of repetition, which mathematicians call
"periodicity," can be seen most readily with rational numbers
-those that can be written as common fractions. A fraction such as
1
- == 0.142857142857142857142857142857 ...
7
which shows a continuous repetition of the digits 142857, can also

be written as ~ = 0.142857 , with the bar above the digits representing
the endless repetition of the digits it covers. Here the period is one
involving six repeating digits. The Fibonacci numbers also exhibit
a less obvious periodicity. Let's look at the first thirty-one Fibo-
nacci numbers (plus the one preceding them, which we will call
Fo). See figure 1-6. When Fo through F31 are each divided by 7,
we obtain the following sequence displaying the quotients and re-
mainders. 19
18. We can also begin the Fibonacci sequence with Fa = 0 and F; = 1, instead of
F; = 1 and F; = 1.
19. We are writing quotients in a multiplication form. That is, instead of writing,
say, 55 divided by 7 is 7, with a remainder of 6, we will write this fact as
55 = 7 ·7 + 6 .
Fo =0=0·7+0 F16 =987=141·7+0
F=I=0·7+1
I
F 17 = 1,597 = 228 . 7 + 1
F=I=0·7+1
2
FIX = 2,584 = 369 . 7 + 1
F=2=0·7+2
I
F19 =4,181=597·7+2
F=3=0·7+3
4
F20 = 6, 765 = 977 . 7 + 3
F =5=0·7+5
<;
F21 = 10,946= 1,573-7+5
F=8=1·7+1
6
F22 = 17,711 = 2,530 . 7 + 1
F=13=1·7+6
7
F21 =28,657=4,093·7+6
F=21=3·7+0
X
F24 = 46,368 = 6,624 . 7 + 0
F = 34 = 4·7 + 6
<)
F25 =75,025=10,717·7+6
F10 = 55 = 7·7 + 6 F26 =121,393=17,341·7+6
FII = 89 = 12 - 7 + 5 F27 =196,418=28,059·7+5

F12 = 144 = 20 . 7 + 4 F2X = 3 17 ,8 1 1 = 47, 401 . 7 + 4
FIl = 233 = 33 . 7 + 2 F29 =514,229=73,461·7+2
F14 = 377 = 53 . 7 + 6 F.,0 = 832,040 = 118,872·7 + 6
F15 =610=87·7+1 F = 1,346,269 = 192,324·7 + 1

"
Figure 1-6
You can see where we have 0 remainders; that is, where 7 di-
vides the numbers, as with: Fa, Fg, F:.6, and F24 in this list.
Notice the pattern of the remainders (since each Fn is the sum
of the two preceding Fibonacci numbers):
0, 1, 1, 2, 3, 5, 1, 6, 0, 6, 6, 5, 4, 2, 6, 1
There is a sort-of-Fibonacci recursive relationship here, al-

though a bit hidden. If you do the same recursive procedure as we
used on the original Fibonacci numbers, but taking only the re-
mainders when the number generated is divided by 7, we get the
following. This is the pattern seen in figure 1-6.
0, 1, 1, 2, 3, 5, (8 = 7 + 1), 6, (1 + 6 = 7 + 0), 6, 6, (6 + 6 = 7 + 5),

(5 + 6 = 7 + 4), (5 + 4 = 7 + 2), 6, (2 + 6 = 7 + 1), ...
The bold numbers in the line above are precisely the numbers that
we see as the sequence of remainders. That is, they form a recursive
sequence much the same as the Fibonacci number sequence.
n Fn
60 1548008755920
61 2504730781961
62 4052739537881
63 6557470319842
64 10610209857723
65 17167680177565
66 27777890035288
67 44945570212853
68 72723460248141
69 117669030460994
70 190392490709135
Figure 1-7
When you get to the 120th (F;20)' the same will be true (see
figure 1-8). You might want to search in appendix A for other evi-
dences of these Fibonacci numbers. And there will be other sight-
ings! (Hint: look at F; 80 .)
n Fn
120 5358359254990966640871840
121 8670007398507948658051921
122 14028366653498915298923761
123 22698374052006863956975682
124 36726740705505779255899443
125 59425114757512643212875125
126 96151855463018422468774568
127 155576970220531065681649693
128 251728825683549488150424261
129 407305795904080553832073954
130 659034621587630041982498215
Figure 1-8
Some Properties of the Fibonacci Numbers
To warm you up, we will present some of the countless curious

characteristics of the Fibonacci numbers-working inductively.
For those who care to see proofs of these relationships, we refer
you to appendix B. So let's begin our inspection of these numbers.
1. The sum of any ten consecutive Fibonacci numbers is divisi-

ble by 11. We can convince ourselves that this may be true by
considering some randomly chosen examples. Take, for ex-
ample, the sum of the following ten consecutive Fibonacci
numbers:
13+ 21 + 34 + 55+ 89 + 144 + 233+ 377 + 610 + 987 = 2,563
which just happens to be divisible by 11, since

11 . 233 == 2,563.
Let's take another example: the sum of another ten con-
secutive Fibonacci numbers, say, from F21 to F30:
10,946+ 17,711 +28,657+46,368+75,025+ 121,393
+ 196,418 + 317,811 + 514,229 + 832,040 = 2,160,598
which also is a multiple of 11, since 11 . 196,418 = 2,160,598.

Are you convinced that the sum of any ten consecutive
Fibonacci numbers is divisible by II? You really shouldn't be
convinced without further evidence. One way to go about per-
suading yourself of the truth in this "conjecture" is to keep on
taking the sum of groups of ten consecutive Fibonacci num-
bers. You could also try to prove the statement, mathemati-
cally. (See appendix B.)
2. Two consecutive Fibonacci numbers do not have any com-
mon factors, which means that they are said to be relatively
prime. 20 This can be seen by mere inspection. Or take any two
consecutive Fibonacci numbers and factor them, and you will
see that they have no common factors. Take a look at figure
1-9, where we have listed the first forty Fibonacci numbers
20. Two numbers (integers) are relatively prime if they have no common factors,
other than 1.
and factored those that are not prime. Notice that no two con-
secutive Fibonacci numbers have any common factors.
Enough examples of this should convince you. (However, for
those who require a proof to be convinced that this is true for
any pair of consecutive Fibonacci numbers, see appendix B.)
n Fn Factors
1 1 Unit
2 1 Unit
3 2 prime
4 3 Prime
5 5 prime
6 8 23
7 13 prime
8 21 3·7
9 34 2· 17
10 55 5 . 11
11 89 prime
12 144 24·32
13 233 prime
14 377 13·29
15 610 2·5·61
16 987 3·7·47
17 1,597 prime
18 2,584 23 . 17 . 19
19 4,181 37 . 113
20 6,765 3,5,11-41
21 10,946 2· 13 - 421
22 17,711 89· 199
23 28,657 prime
24 46,368 25 . 32·7·23
25 75,025 52 - 3,001
26 121,393 233 - 521
27 196,,418 2 . 17 - 53 . 109
28 317811 3 . 13 - 29 . 281
29 514,229 prime
30 832,040 23 . 5 . 11 . 3 1 . 61
31 1,346,269 557· 2,417
32 2,178,309 3 . 7 . 47 . 2,207
33 3,524,578 2·89· 19,801
34 5,702,887 1,597 . 3,571
35 9,227,465 5 . 13 . 141,961
36 14,930,352 24 . 33 . 17 . 19 . 107
37 24,157,817 73 . 149 . 2,221
38 39,088,169 37 . 1 13 - 9,349
39 63,245,986 2·233'135,721
40 102,334,155 3 - 5 . 7 . 1 1 . 41 . 2,161
Figure 1-9
3. Consider the Fibonacci numbers in the composite-number21

positions (with the exception of the 4th Fibonacci number).
By that we mean, the 6th, 8th, 9th, 10th, 12th, 14th, 15th,
16th, 18th, 20th, and so on, Fibonacci numbers-which are
all nonprimes. A quick inspection (figure 1-10) shows that
the Fibonacci numbers in the composite-number (i.e., non-
prime) positions are also composite numbers. For the fac-
tors of these Fibonacci numbers, we refer you to figure 1-9
and for further evidence see the much longer list in appen-
dix A. Again, you may be convinced of this property, but to
really make it a fact, we have to prove it mathematically.
(See appendix B.)
F6 =8 Fio = 6, 765
Fg = 21 Fil = 10,946
F9 = 34 F22 = 17, 711
Fio = 55 Fi4 = 46,368
Fi2 = 144 F25 = 75,025
Fi4 = 377 F26 = 121,393
Fi5 = 610 F27 = 196,418
Fi6 = 987 F28 = 3 17 , 81 1
Fi8 = 2,584 F10 = 832,040
Figure 1-10
One could conjecture at this point that the analog of this is

also true-namely, that the Fibonacci numbers in prime-
number22 positions are also prime. That is, if we look at the
list of the first thirty Fibonacci numbers above, those in the
prime positions, namely, the 2nd, 3rd, 5th, 7th, 11th, 13th,
17th, 19th, 23rd, and 29th Fibonacci numbers, would then
also have to be prime.
21. A composite number is one that is not prime, that is, it is divisible by num-
bers other than itself and 1.
22. A prime number is a number (other than 0 or ±1) that is divisible only by
itself and 1.
They are:
F2 = 1, not a prime FI3 = 233

F3 =2 FI7 = 1,597
~ =5 F I9 =4,181 =37·113
F 23 = 28,657
F7 = 13
F 29 = 514,229
~1=89
However, you will notice that the 2nd and the 19th
Fibonacci numbers are not primes, and therefore this ana-
log situation does not hold true. Nothing further to prove.
One counterexample suffices to draw this conclusion.
4. With all these lovely relationships embracing the Fibonacci
numbers, there must be a simple way to get the sum of a
specified number of these Fibonacci numbers. A simple for-
mula would be helpful as opposed to actually adding all the
Fibonacci numbers to a certain point. To derive such a for-
mula for the sum of the first n Fibonacci numbers, we will use
a nice little technique that will help us generate a fonnula.
Remember the basic rule (or definition of a Fibonacci
number. We can write this definition formally as
F
n+
2 == F
n+
I + F
n
,where n > 1
Or
F==F
n n+2
-Fn+l
By substituting increasing values for n, we get:
F;=~-~
~=~-~
~=~-~
F4=~-~
F11-1 =F11+1 -F11

Fn =Fn+2 -Fn+1
By adding these equations, you will notice that there

will be many terms on the right side of the equations that
will disappear (because their sum is zero-since you will
be adding and subtracting the same number). What will re-
main on the right side will be Fn+2 - F2 = Fn+2 - 1.
On the left side, we have the sum of the first n Fibo-
nacci numbers: F. + F2 + ~1 + F4 + .. . + Fn ' which is what
we are looking for.
Therefore, we get the following:
which says that the sum of the first n Fibonacci numbers is

equal to the Fibonacci number two further along the se-
quence minus 1.
There is a shortcut notation that we can use to signify
n
the sum: F; + F2 + F; + F4 + . . . + Fn , and that is L F; .
i=1
This reads: "The sum of all the F I

terms where i takes on
the values from 1 to n." So we can then write this result as:
L~ = F. + F; + F; + F4 + ... + Fn = Fn+2 -1
i=1
or simply: L ~ = Fn+2 - 1
i=1
5. Suppose we now consider the sum of the consecutive even-

positioned Fibonacci numbers, beginning with the first such
Fibonacci number, F2 • Again, let's see if we can discover a
pattern when adding these consecutive even-positioned
Fibonacci numbers.
F2 +F4 =1+3=4
F; + F4 + F6 = 1+3+8= 12
F2 + F4 + ~ + ~ = 1 + 3 + 8 + 21 = 33
F2 + F4 + F6 + ~ + ~o = 1+ 3 + 8 + 21 + 55 = 88
We may notice that each sum is 1 less than a Fibonacci

number-more particularly, the Fibonacci number that
follows the last even number in the sum. We can write this
symbolically as:
F2 + F4 + ~ + ~ + ... + F 2n - 2 + F2n = F2n + 1 - 1, where n > 1

n
or L F;i = F;n+1 - 1, where n > 1
i=l
Again, this looks to be the case for all Fibonacci numbers,

but remember we only looked at the first few. To accept this
as general truth, we would have to prove it. (See appendix B.)
6. Now that we have established a short-cut method to get the
sum of the initial consecutive even-positioned Fibonacci
numbers, it is only fitting that we inspect the analog: the
sum of the initial consecutive odd-positioned Fibonacci
numbers. Once again, by considering a few examples of
these sums, we will look for a pattern.
F+F
I 3
=1+2=3
~+F;+~=1+2+5=8
~ +F; +~ +F, = 1+2+5+13=21
~ +F; +~ +F, +~ = 1+2+5+13+34=55
These sums appear to be Fibonacci numbers. But how

do these sums relate to the series that generated them? In
each case the sums are the next Fibonacci number after the
last term in the sum of odd-positioned Fibonacci numbers.
This can be symbolically written as
or L F21 _ = F211
1=1
1
(See appendix B.)
If we add the sum of the initial consecutive even-

positioned Fibonacci numbers and the sum of the initial
consecutive odd-positioned Fibonacci numbers, we should
get the sum of the initial consecutive Fibonacci numbers:
F2 + F4 + ~ + ~ + ... + F;11 = F2n + 1 - 1, where n > 1

~ + F; + ~ + F7 + ... + F2n _ 1 = F;n ,where n > 1
The sum of these sequences is:
or
which is consistent with what we concluded In item 4

(above).
7. Having established relationships for various sums of
Fibonacci numbers, we shall now consider the sum of the
squares of the initial Fibonacci numbers. Here we will see
another astonishing relationship that continues to make the
Fibonacci numbers special. Since we are talking about
"squares," it is fitting for us to look at them geometrically.
21
13
21
13
3
2
2
1 T 1
5
3
8
Figure 1-11
We find that, beginning with a 1 x 1 square (see figure

1-11), we can generate a series of squares, the sides of which
are Fibonacci numbers. We can continue on this way indefi-
nitely. Let us now express the area of each rectangle as the
sum of its component squares.
12+ 12+22+3 2 +5 2 +8 2 + 13 2= 13·21
If we express the sum of the squares making up the

smaller rectangles, we get this pattern:
12 +1 2 =1.2
12 + 12 + 22 = 2·3
12 + 12 + 22 + 3 2 = 3·5
12 + 12 + 22 + 3 2 + 52 = 5·8
12 + 12 + 22 + 3 2 + 52 + 8 2 = 8·13
12+ 12+22+3 2 +5 2 +8 2 + 13 2 = 13·21
From this pattern we can establish a rule: the sum of the

squares of the Fibonacci numbers, to a certain point in
the sequence, is equal to the product of the last number and
the next number in the sequence. That is, if, for example, we
choose to find the sum of the squares of the following por-

tion of the sequence of Fibonacci numbers: 1, 1, 2, 3, 5, 8,
13, 21, 34, we would get:
12 + 12 +22 +3 2 +5 2 +8 2 + 132 +212 +34 2 = 1870

,
However, we can apply this amazing rule that tells us

that this sum is merely the product of the last number and
the one that would come after it (sometimes known as its
immediate successor), in the Fibonacci sequence. That
would mean that this sum can be found by multiplying 34
by 55. Indeed, this product gives us 1,870 (= 34 . 55) .
We can write this with our summation notation as
n
".£..J F2
I
= F n Fn+l
i=1
Imagine you would want to find the sum of the squares

of the first thirty Fibonacci numbers. This neat relationship
makes the task very simple. Instead of finding the squares
of each and then finding their sum-a rather laborious task
-all we need to do is multiply the thirtieth by the thirty-
first Fibonacci number.
That is, the sum of the squares:
30
L F;2 = F;2 + F;2 + .. . + F;~ + F;~
;=1
= 12 + 12 + 22 + 32 + 52 + 8 2 + 132 + ... + 514,229 2

+ 832,040 2
= 1,120,149,658,760
Or simply, using our newly established formula,
30
LF;2 = F;o· F;l = 832,040 ·1,346,269 = 1,120,149,658,760
;=1
(For the more advanced reader we offer a proof of this re-

lationship in appendix B.)
8. While we are talking about the squares of Fibonacci num-
bers, consider the following relationship. Let's take the
square of a Fibonacci number, say, 34, and the square of the
Fibonacci number that is two before it, 13. Subtracting these
2 2
two squares, we get: 34 - 13 = 1,156 - 169 = 987 , which
is also a Fibonacci number! So we squared the ninth
Fibonacci number (~ = 34), and subtracted the square of
the seventh Fibonacci number (F; = 13) to get the sixteenth
Fibonacci number (~6 = 987). This can be written symboli-
cally as: ~2 F7 2 = ~6'
-
It is nice to see that by subtracting the squares of two

alternating 23 Fibonacci numbers, we seem to get another
Fibonacci number. Will this work for any pair of alternat-
ing Fibonacci numbers? To answer this we would have to
do a mathematical proof (which is provided in appendix B).
However, we can begin to convince ourselves by inspecting
some examples that this result was no fluke, but rather, that
it actually works for any appropriate pair of Fibonacci
numbers. Let's look at a few now and see if we can deter-
mine a pattern:
F6 2 - F4 2 = 8 2 - 32 = 55 = F10
F7 2 - F2
5
= 132 - 52 = 144 = F12
~52 -~32 =610 -233 =317,811= F2X
2 2
By inspecting the three subscripts for each of these ex-

amples, we notice that the sum of the first two equals the
third. This would lead us to make the following generaliza-
tion: Fn2 - F}_2 = F2n - 2 . You might want to try to apply this
23. Alternating members of a sequence will be those that are one member apart
from one another. That is, for example, the 4th and 6th members or the 15th and
17th members are called alternating members.
rule to a few more such pairs to further convince yourself if

you wish. (A proof is in appendix B.)
9. The next natural question would be to look at the sum of
the squares of two Fibonacci numbers. Suppose we con-
sider two consecutive Fibonacci numbers, say, F7 (==13)
and Fg (== 21). Their squares, 169 and 441, have a sum of
610, which is also a Fibonacci number, Fis. Perhaps we are
onto something here. Let us we try another consecutive pair
of Fibonacci numbers and see what the sum of the squares
of these numbers is. To establish a pattern, we will look at
the two Fibonacci numbers Fio (== 55) and Fi I (== 89). Their
squares, 3,025 and 7,921, have a sum of 10,946, which is
again a Fibonacci number, F21 • It appears that as we try this
for several other pairs of consecutive Fibonacci numbers,
we arrive each time at another Fibonacci number. But what
is the pattern? To be able to predict which Fibonacci num-
ber will result from the sum of the squares of a consecutive
pair of Fibonacci numbers, we will want to inspect their
position in the sequence. In our first example, above, we
used the Fibonacci numbers F7 and Fs , and the sum of
their squares was Fis. In the next example we used the
Fibonacci numbers Fio and Fi I , and we found that the sum
of their squares was F;I . It appears that the sum of the sub-
scripts of the two Fibonacci numbers whose squares we are
adding will give us the subscript of the Fibonacci number
representing their sum. That is, the sum of the squares of
the Fibonacci numbers in positions nand n + 1 (consecu-
tive positions) is the Fibonacci number in place n + (n + 1)
2
= 2n + 1, or Fn + F;,2+1 = F 2n + l • (The proof may be found in
appendix B.)
10. Here is an engaging relationship that takes another look at
the unexpected patterns we find among the numbers in the
Fibonacci sequence. Take any four consecutive numbers in
the sequence, say, 3, 5, 8, 13, and then find the difference
of the squares of the middle two numbers: 82 - 52 == 64 - 25
== 39. Then find the product of the outer two numbers:
3 . 13 = 39 . Amazingly the same result is attained! Was

this by strange coincidence or was it a pattern that will hold
for all such strings of four consecutive Fibonacci numbers?
We can try this again with another group of four consecu-
tive Fibonacci numbers, say, 8, 13,21,34. Again, we find
the sum of the squares of the two middle numbers:
212 - 13 2 = 441 - 169 = 272. If the pattern will continue to
work, then the product of the outer two numbers must be
272. Well, 8· 34 = 272 , and so the pattern still holds. Sym-
bolically, we can write for any four consecutive numbers,
2
Fn- h Fn, Fn+h and F,,+2, this as Fn:l - Fn = Fn- 1 . Fn+ 2 • To
really convince yourself that this is always true, it is not
sufficient to merely try many groups of four consecutive
Fibonacci numbers and find that the pattern holds for each
of your tries, rather, you would have to do a proof. (See ap-
pendix B.)
11. Another curious relationship is obtained by inspecting the
products of two alternating members of the Fibonacci se-
quence. Consider some of these products:
F3 . Fs = 2 . 5 = 10, which is 1 more than the square of the

Fibonacci number 3: F4 2 + 1 = 3 2 + 1.
F4 . F 6 = 3 . 8 = 24, which is 1 less than the square of the

Fibonacci number 5: Fs2 - 1 = 52 - I.
Fs . F7 = 5 . 13 = 65, which is 1 more than the square of the

Fibonacci number 8: F6 2 + 1 = 8 2 + 1.
F6 . F8 = 8 . 21 = 168, which is 1 less than the square of the

Fibonacci number 13: F7 2 - 1 = 13 2 - 1.
By now you should begin to see a pattern emerging: the

product of two alternating Fibonacci numbers is equal to the
square of the Fibonacci number between them ±1. We still
need to determine when it is + 1 or -1. When the number to
be squared is in an even-numbered position, we add 1, and
when it is in an odd-numbered position, we subtract 1. This
can be generalized by using (_l)n, since that will do exactly

what we want: -1 to an even power is + 1, and -1 to an odd
power is -1.
Symbolically, this may be written as:
F n- IFn+ 1 == F2
n
+(_I)n
where n > 1.
Although we "convinced" ourselves that this appears to
be true, we must actually do a proof to be sure it is true for
all cases. (See appendix B.)
This relationship can be expanded. Suppose instead of
taking the product of the two Fibonacci numbers on either
side of a particular Fibonacci number as we did above, we
would take the two Fibonacci numbers that are one re-
moved in either direction. Let's see how the product com-
pares to the square of the Fibonacci number in the middle.
If we take a specific example from the Fibonacci numbers,
say, ~ == 8, the product of the two Fibonacci numbers one
removed on either side of 8 is 3· 21 == 63, and the square of
8 is 64. They differ by 1. Suppose we now use Fs == 5, then
the product of the two numbers removed on either side is
2 . 13 == 26, which differ from 52 by 1. We can write this
symbolically as:
where n > 1.
You can now try this by comparIng the product of
Fibonacci numbers two, three, four, and so on removed on
either side of a designated Fibonacci number, and you will
find the following to hold true.
By now you may begin to recognize the pattern in the
chart in figure 1-12. The difference between the square of
the selected Fibonacci number and the various products of
Fibonacci numbers, which are equidistant from the selected
Fibonacci number, is the square of another Fibonacci num-

ber. Symbolically we can write this as
Fn-k Fn+k - F2 - +17 2

n - -rk
where n > 1 and k > 1.
Number Symbolic Example for Example for Difference

(k) of Representa- between:
Fibonacci tion for the F7 = 13 F8 = 21
numbers Fn-k ·Fn+k
removed case of Fn
on either
side of and F2
n
selected
Fibonacci
number
1 F,l_I Fn+1
2 8·21 = 168 2 13·34 = 442 ±1
Fn
13 = 169 212 = 441
2 2 5·34 = 170 2 8·55 = 440 ±1

Fn_2Fn+2 Fn
13 = 169 212 = 441
3 Fn _ F n +
3 3
F2 3·55=165
13
2
= 169
5·89 = 445
212 = 441
±4
n
..,
4 F F F- 2 ·89 = 178 2 3·144 = 432 ±9
n-4 11+4
n 13 = 169 212 = 441
5 F2 1·144 = 144 2 2 233 = 466 ±25

F n _ 5 F,1+5
n 13 = 169 212 = 441
6 F F
2 1 . 233 = 233 2 1 377 = 377 ±64
n-6 n+6 Fn
13 = 169 212 = 441
k F n _ A Fn+k f:,2 +F, 2

- k
Figure 1-12
There are endless properties that one can observe with

the Fibonacci numbers. Many of them can be found by
simple inspection. Before we embark on a more structured
study of these amazing numbers, let's take another look at
the sequence (figure 1-13).
FI 1 F16 897
F2 1 F17 1,597
F3 2 FI8 2,584
F4 3 FI9 4,181
F5 5 F 20 6,765
F6 8 F21 10,946
F7 13 F22 17,711
Fg 21 F 23 28,657
F9 34 F24 46,368
Flo 55 F 25 75,025
Fll 89 F 26 121,393
F12 144 F27 196,418
F13 233 F 28 317,811
FI4 377 F 29 514,229
F 15 610 F30 832,040
Figure 1-13
12. Beginning with the first Fibonacci number, notice that

every third number is even. That is, F;, ~, Fg, F:.2' F:.s,
and F:.xare all even numbers. We could restate this by say-
ing that these numbers are all divisible by 2 or by F;.
Let's look further at this list of Fibonacci numbers. No-
tice that every fourth number is divisible by 3. Here,
~, ~, F..2' F..6' F;o' and F24 are each divisible by 3. We
24
can again restate this by saying that F4 , Fg, F:.2' F:.6'

F;o' and F;4 are each divisible by F4, or by 3.
Using patterns to see where we go from here, we would
see that the first divisibility was by 2, then by 3, then, in
good Fibonacci style, we ought to try to check for divisibility
by 5-the next number in the Fibonacci sequence. Searching
for Fibonacci numbers that are divisible by 5 is easy. They
are the following numbers: 5, 55, 610, 765, and 832,040,
which correspond (symbolically) to Fs, F:.o' F:.5' F 20 ,
24. You can check this quite easily by using the popular rule for divisibility by
3. If and only if the sum of the digits of a number is divisible by 3, then the
number 75,025, itself is divisible by 3.
F25 , and F;o. Thus, we are able to say that Fs, F:.o'
F:.5' F20 , F25 , and F;o are each divisible by 5 or by Fs.
Checking for divisibility by 8 (the next Fibonacci num-
ber), we find that ~,F:.2' F:.8' F24 , andF;oare each divisi-
ble by 8, or by ~.
Yes, every seventh Fibonacci number is divisible by 13,
or F;. You might now try to generalize this finding. You
can either say that a Fibonacci number Fnm is divisible by a
Fibonacci number Fm , where n is a positive integer, or you
may state it as the following: If p is divisible by q, then Fp
is divisible by ~. (A proof of this wonderful relationship
is in appendix B.)
13. Fibonacci relationships, as noted earlier, can also be seen
geometrically. We can consider Fibonacci squares as ar-
ranged as in figure 1-14 and figure 1-15.
,----- --- ----------- -------l

I •
I ,
! ~
13x21
21 x34
I
2x3
I 3x15 3x5, 1 1
. --'-T~ x 5x8
'I·
1x2 I'" I 2x3t ' I

8x 13 . i~ . i
13x21 1x1 1x2 i
5x8 i
8x 13
. ___._ ___ __ . _ . _ .. ____ .___ ._ ... ____. ________ J
Figure 1-14 Figure 1-15

Fibonacci square with an odd Fibonacci square with an
number of rectangles. even number of rectangles.
Figure 1-14 shows an odd number (n == 7) of rectangles

into which the Fibonacci square is divided. The rectangles
have dimensions of F; and F;+I , where i has values from 1
to n + 1. The square then has sides of length Fn+1 • The area
of the square is equal to the sum of the rectangles, which
can constitute a geometric proof. This can be written sym-

bolically as:
n+1
~ FFi-I == F2n+1
£..J I
i=2
when n is odd.
As you can see in figure 1-14, when n == 7 the sum of
the areas of the rectangles is
F2FI + F3 F 2 + F4 F 3 + FsF4 + F6 F S + F7 F 6 + FgF7

= 1 +2+6+ 15+40+ 104+273=441 =212=Fg2
On the other hand, when n is even (in figure 1-15, n == 8),

that is, when a Fibonacci square is constructed using an even
number of Fibonacci rectangles, a unit square remains that is
not used. This indicates that the Fibonacci square is 1 unit
larger than the sum of the rectangles, which gives us the
following adjusted relationship:
n+1
~ F F == F2 - 1
£..J I-II n+1
i=2
when n is even.
In figure 1-15, where n == 8, the sum of the rectangles is
~~+~~+~~+~~+~~+~~+~~
= 1 + 1 +2+6+ 15+40+ 104+273+714
= 1,156 = 342 = F 29
14. Earlier in this chapter, we mentioned the Lucas numbers.

They are: 1,3,4,7, 11, 18,29,47, .... (See appendix A for
a list of the first two hundred Lucas numbers.) We can find
the sum of the Lucas numbers much the same way we found
the sum of the Fibonacci numbers. Again, there must be a
simple way through a formula to get the sum of a specified
number of these Lucas numbers. To derive such a formula
for the sum of the first n Lucas numbers, we will use a nice
little technique that will help us generate the formula.
The basic rule (or definition) of a Lucas number is
Ln+ 2 = Ln+ 1 + Ln, where n > 1

or L n = Ln+ 2 - L n+ )
By substituting increasing values for n, we get:
L) = L3 - L2
L2 = L4 - L3
L3 = Ls - L4
L4 = L6 - Ls
Ln-) =Ln+' -Ln

Ln =Ln+2 -Ln+)
By adding these equations, you will notice that there

will be many terms on the right side that will disappear be-
cause their sum is zero. This occurs since you will be add-
ing and subtracting the same number. What will remain on
the right side will be Ln+2 - L2 = Ln+2 - 3.
On the left side, we have the sum of the first n Lucas
numbers: L) + L2 + L3 + L4 + . . . + L n , which is what we
are looking for. Therefore, we get the following equation:
L, + L2 + L3 + L4 + . .. +Ln = Ln+2 - 3 ,which says that the
sum of the first n Lucas numbers is equal to the Lucas
number two further along the sequence minus 3.
There is a shortcut notation that we can use to signify
n
the sum: L\ + L2 + L, + L4 + . . . + Ln , and that is L L/ . So
1=1
we can then write this result as:
L Li = L) + L2 + L3 + L4 +
i=)
. . . + Ln = Ln+2 - 3
or simply:
n
"'L.=[
~ n+
,-3
f
i= 1
15. Just as we found the sum of the squares of the Fibonacci

numbers, so, too, can we find the sum of the squares of the
Lucas numbers.
Here we will uncover another astonishing relationship
that continues to make the Lucas numbers special.
11 18
11 11
18
7 3 1
1
3 f- 1
3 f- 1
7 1
4 4
7 4 18
Figure 1-16
We find in figure 1-16 that, beginning with the three

small 1 x 1 squares, we can generate a series of squares the
sides of which are Lucas numbers. We can continue on this
way indefinitely. Let us now express the area of the rectan-
gle as the sum of its component squares minus the two
small squares (shaded).
['1 + ['2 + ['3 + ['4 + ['5 + ['6

= I' +3' +4' +7 ' +11 ' +18'
= 522 - 2 = 520 = 18·29
= [6 . [,
If we express the sum of the squares progressively,

which make up the smaller rectangles, we get this pattern:
12 + 32 = 3·4 - 2
12+3 2 +42=4.7-2
12 + 32 + 4 2 + 7 2 = 7 -11- 2
12+32 +42+72+112=11.18-2
12+32 +42+72+112+18 2 =18-29-2
12 + 32 + 4 2 + 7 2 + 112 + 18 2 + 29 2 = 29 -47 - 2
From this pattern we can establish a rule: the sum of the

squares of the Lucas numbers to a certain point in the se-
quence is equal to the product of the last number and the
next number in the sequence minus 2_ In other words, if, for
example, we choose to find the sum of the squares of the
following portion of the sequence of Lucas numbers: 1, 3,
4,7,11,18,29,47,76, we would get:
L21 + L22 + L23 + L24 + L25 + L26 + L27 + L28 + L29

= 12 + 32 + 4 2 + 7 2 + 112 + 18 2 + 29 2 + 47 2 + 76 2
= 9,346
However, we can apply this amazing rule that tells us

that this sum is merely the product of the last number and
the one that would come after it (sometimes known as its
immediate successor) in the Lucas sequence. That would
mean that this sum can be found by multiplying 76 by 123
and subtracting 2. Indeed, this product gives us the follow-
ing: L9 - LID = 76 ·123 = 9,348 = 9,346 + 2.
We can concisely summarize this as:
n
~ L?
.£.J I
= L n Ln+ 1-2
;=1
This will help you find the sum of the squares of the
Lucas numbers without having to first find all the terms
below it.
Imagine you would want to find the sum of the squares

of the first thirty Lucas numbers. This neat relationship
makes the task very simple. Instead of finding the squares
of each and then finding their sum-a time consuming
task-all you need to do is multiply the thirtieth by the
thirty-first Lucas number and subtract 2. (See appendix B
for a proof of this relationship.)
These proofs, although geometric, are reasonable enough to con-

vince you of the veracity of these statements. Before having you
explore more unexpected sightings of these lovely Lucas numbers,
we provide you with a summary of the properties we have noted
thus far.
A Summary of the Properties
Here is a summary of the Fibonacci (and Lucas) relationships we

have seen (let be n any natural number; n > 1):
O. Definition of the Fibonacci numbers FIl and of the Lucas

numbers Ln:
FI = 1; F2 = I; FIl + 2 = Fn + Fn + I
LI = 1; L2 = 3; Ln + 2 = Ln + Ln +
I
1. The sum of any ten consecutive Fibonacci numbers is di-

visible by 11:
1 I I (FIl + Fn + 1+ Fn + 2 + . . . + Fn + 8 + FIl + l))
2. Consecutive Fibonacci numbers are relatively prime:

That is, their greatest common divisor is 1.
3. Fibonacci numbers in a composite number position (with

the exception of the fourth Fibonacci number) are also
composite numbers.
Another way of saying this is that if n is not prime, then Fn

is not prime
(with n"# 4, since F4 = 3, which is a prime number).
4. The sum of the first n Fibonacci numbers is equal to the

Fibonacci number two further along the sequence minus 1:
n
LF;=F;+F;+F;+~+ ... +Fn =Fn + 2 -1
i=1
5. The sum of the consecutive even-positioned Fibonacci

numbers is 1 less than the Fibonacci number that follows
the last even number in the sum:
LF2i = F2 + F4 + F 6 +··· + F 2n - 2 + F 2n = F 2n + 1 - 1
;=1
6. The sum of the consecutive odd-positioned Fibonacci num-

bers is equal to the Fibonacci number that follows the last
odd number in the sum:
n
LF2i - 1 = FI + F3 + Fs + ... + F 2n - 3 + F 2n - 1 = F 2n
i=1
7. The sum of the squares of the Fibonacci numbers is equal

to the product of the last number and the next number in
the Fibonacci sequence:
n
~ F2
.£..J I
= Fn F n+ 1
i=1
8. The difforence of the squares of tvvo altemate25 Fibonacci num-

bers is equal to the Fibonacci number in the sequence whose po-
sition number is the sum oftheir position numbers:
F2n - F2n-2 = F2n-2
25. I.e., a Fibonacci number with number n and the Fibonacci number two be-
fore it.
9. The sum of the squares oftvvo consecutive Fibonacci numbers is

equal to the Fibonacci number in the sequence whose position
number is the sum oftheir position numbers:
10. For any group offour consecutive Fibonacci numbers, the

d~f(erence 0.( the squares of the middle two numbers is
equal to the product of the outer two numbers.
Symbolically, we can write this as
F2n+1 - F2 11
= Fn-l . Fn+2
11. The product of tvvo alternating Fibonacci numbers is J more or

less than the square ofthe Fibonacci number betvveen them ±J:
FIl _ I Fll + I = Fn 2 + (- 1)"
(If n is even, the product is 1 more; if n is odd, the product

is 1 less.)
The difference between the square of the selected
Fibonacci number and the various products of Fibonacci
numbers equidistant from the selected Fibonacci number is
the square of another Fibonacci number:
2
Fn_k Fn+k - Fn = + F/, where n > 1, and k > 1
12. A Fibonacci number F mn is divisible by a Fibonacci num-

ber Fm.
We can write this as
Fm I F mn' and it reads Fm divides Fmn
Or in other words: If p is divisible by q, then Fp is divisible

by Fq or using our symbols:
(in the above, m, n, p, and q are positive integers)

Here is how this looks for specific cases:

F,IFn , i.e., 11 F I , 1 I F 2, 11 F}, II F4, 1 I Fs~ I I F6, ...• I I Fn, ...
F21 F 21l , i.e., 1 I F 2, 1 1F4, 1 1F6, 1 1Ff!" 1 I FlO' 1 1F 12 • .... 1 1F211' ...
F}I F}n , i.e., 21 F}, 21 F6, 21 F9 , 21 Fib 21 F IS ' 21 Fig, ...• 2 I FJn , ...
F4 1F4n ~ i.e., 31 F4, 31 F 8 , 3 I Fib 31 F 16 , 31 F20' 3 1F 24 , ... , 31 F4n , ...
Fs I Fsn , i.e., 51 F s, 5 1FlO, 51 F IS ' 51 F20' 5 1F 2S , 5 I FlO, ... , 51 F"ill' ...
F6 1F6n , i.e., 81 F6, 81 Fib 81 F 18 , 8 I F 24 , 81 FJo, 8 1F 36 , ... , 81 F6n , ...
F7 1F7n , i.e., 13 I F7, 13 1F 14 , 13 I F 2], 131 F 2f!" 131 F 35 • 131 F42 , ... ,131 F71l • ...
13. The product of consecutive Fibonacci numbers is either the

square of a Fibonacci number or I greater that the square of a
Fibonacci number.
n+1
~
~ FF 1 1-
1 = F2n+ I' when n is odd
i=2
n+l
~ F.F 1= F2n+ 1-1, when n is even
~ / /-
i=2
14. The sum of the first n Lucas numbers is equal to the Lucas
number two further along the sequence minus 3:
L Li = Ll + L2 + L3 + L4 +
i=1
.. . + Ln = Ln + 2 - 3
15. The sum of the squares of the Lucas numbers is equal to the
product of the last number and the next number in the Lu-
cas number sequence minus 2:
n
~
~
L.2 I
= Ln Ln+ 1 - 2
i=1
Although our focus IS largely about the Fibonacci numbers, we

should not think of Fibonacci as a mathematician who is known
only for this now-famous sequence of numbers that bears his
name. Fibonacci, as mentioned, was one of the greatest mathemati-
cal influences in Western culture. He not only provided us with the
tools to do efficient mathematics (e.g., the numerals heretofore
known largely to the Eastern world), but he also introduced to the
European world a thought process that opened the way for many
future mathematical endeavors.
Notice how the origin of the Fibonacci numbers, innocently
embedded in a problem about the regeneration of rabbits, seems to
have properties far beyond what might be expected. The surprising
relationships of this sequence, within the realm of the numbers, are
truly mind-boggling! It is this phenomenon coupled with the al-
most endless applications beyond the sequence that has intrigued
mathematicians for generations. It is our intention to fascinate with
applications of these numbers as far afield as one can imagine.
Figure 1-17
A statue of Fibonacci in Pisa.
Chapter 2
The Fibonacci Numbers

in Nature
F ibonacci numbers, as we have seen briefly, have amazing

properties. We will continue to explore the wide variety of
applications and relationships that these numbers hold. Here we
will go back to the original manifestation of the Fibonacci num-
bers: in nature. We first saw them evolve in the regeneration of
rabbits. Let's see if there are other living creatures whose regen-
eration might result in an application of the Fibonacci numbers.
The Male Bee's Family Tree
Of the more than thirty thousand species of bees, the most well
known is probably the honeybee, which lives in a bee hive and has a
family. So let us take a closer look at this type of bee. Curiously
enough, an inspection of the family tree of the male bee will reveal
our famous numbers. To examine the family tree closely, we first
have to understand the peculiarities of the male bee. In a bee hive
there are three types of bees: the male bee (called a drone), who
does no work; the female bees (called worker bees), who do all the
work; and the queen bee, who produces eggs to generate more bees.
A male bee hatches from an unfertilized egg, which means it has
only a mother and no father (but does have a grandfather), whereas a
59
female bee hatches from a fertilized egg, thus requiring a mother

(queen bee) and a father (one of the drones). The female bees end up
as worker bees, unless they have taken some of the "royal jelly" that
enables them to become queen bees and begin new colonies of bees
elsewhere.
In the table (figure 2-1, where d' represents the male and ~
represents the female) we begin with the male bee whose ancestors
we are tracking. As we said, the male bee must come from an un-
fertilized egg, so only a female was needed to produce him. How-
ever, the egg-producing female must have had a mother and a fa-
ther, so the third line has both a male and a female. It then
continues with this pattern: the immediate ancestor of every male
is a single female and the ancestor of a female is a male and a fe-
male. At the right, is a summary of the number of bees in each
row. As you look down each column at the right, you will recog-
nize the familiar Fibonacci numbers. Surprised? If not, you ought
to be. This is clearly a surprising appearance of the Fibonacci
numbers. Even the biology here is quite different from the rabbit's
reproductory pattern, which generated the first recognition of the
Fibonacci numbers.
Geneological Table of a Male Bee (Drone)
Female Ma~ Tutal

Bees 8eeI Bees
,
0 0 1 1
Y 0
~ d 2
~cr I
2 3
<.? !1?
~<3 9
I
9 9
/~
<3 3 2 5
9
~ I
0 <.? ~~ ~~ I
9 5 3 8
A
~ <.~
I
<.?
A ~ 0
A0
y
I
~
A I
9 0 ~
/~
)1 0 13 5 13
/\
~ (~ 9
1\ 1\
S? () S? {~
I
?
1\ 9I 1\ 1\ ?I 91\ 91\(f 9I
« o y ('{ ~ () ~ 21 8 21
Figure 2-1
Had Fibonacci been aware of the relationship that unfolds as we

inspect the genealogy of the drone (male bee), then he could have
THE FIBONACCI NUMBERS IN NATURE 61
used it in place of what some consider as a less-than-realistic exam-

ple of the regeneration of rabbits in his Liber Abaci. However, the
rabbit problem is, in fact, not so far off from reality. After age three
months, rabbits can produce offspring and can reproduce them on a
monthly basis. So his rabbit problem was close to accurate.
The regeneration of the rabbits and the genealogy of the drone
are described with the recursive formula Fn+2 := F n + Fn+ I, which,
as we have seen, means that each number in the sequence is the
sum of its two immediate predecessors.
If we extend our theme here from the "Fibonacci numbers in
nature" to the "Fibonacci numbers in society"-which could be
stretched to be the nature of man!-then we can use the example of
the spreading of rumors in society as an example of the Fibonacci
numbers at work. We can create a model l for the spread of a
rumor, which will generate the Fibonacci numbers.
The message (i.e., the rumor) will be passed on to only one
person at a time. Specifically:
(1) If person x is told the rumor, the earliest he will pass it

on is the next day.
(2) Person x passes the rumor to only one person per day.
(3) After person x has passed the rumor twice, he is not in-
terested in it any more, so person x does not go on
spreading the rumor any further.
A chart of this stream of events should expose the Fibonacci

sequence.
1. M. Huber, U. Manz, and H. Walser, Anniiherung an den Goldenen Schnitt

(Approaching the Golden Section) ETH ZUrich, Bericht No. 93-01, 1993, p. 57.
Distribution of a Rumor
(x passes the rumor on to xa and xb)
Day 1
1
Day22
1, la =2
Day 3
1,2,lb=3,2a=4
Day 4
1,2,3,4, 2b = 5, 3a = 6, 4a = 7
DayS
1,2,3,4,5,6, 7,3b=8,4b=9,5a= 10,6a= 11, 7a= 12
Day 6
1,2,3,4,5,6,7,8,9,10,11,12, 5b = 13, 6b = 14, 7b = 15,
8a= 16,9a= 17, 10a= 18, 11a= 19, 12a=20
Day 7
1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18, 19,20,8b
= 21, 9b = 22, lOb = 23, lIb = 24, 12b = 25, 13a = 26, 14a = 27,
15a = 28, 16a = 29, 17a = 30, 18a= 31, 19a = 32, 20a = 33
Daysn 1 2 3 4 5 6 7 8 9 10 11 12
Personsp 1 2 4 7 12 20 33 54 88 143 232 376
Difference (1) 1 2 3 5 8 13 21 34 55 89 144
Figure 2-2
We find Pn = F n- 2 - 1.
The rumor would be spread so much faster, if step (2) were to
be changed to the following:
(2) At first person x passes on the news to p people, on

the second day to q people.
2. The number after each of the equal signs indicates the running total of people
with the rumor, so that the last number after the last equal sign is the total num-
ber for that day.
Fibonacci Numbers in the Plant World
Perhaps you would be most surprised to discover that these ubiq-

uitous numbers also appear in the plant world. Take, for example,
the pineapple. It might even be worth your while to get a pineapple
and see for yourself that the Fibonacci numbers are represented on
the pineapple. The hexagonal bracts on a pineapple can be seen to
form three different direction spirals. In figures 2-3, 2-4, 2-5, and
2-6 you will notice that in the three directions there are 5, 8, and 13
spirals. These are three consecutive Fibonacci numbers.
Figure 2-3
Figure 2-4 Figure 2-5 Figure 2-6

Consider the drawing of a pineapple

(figure 2-7), where we have numbered the
hexagons. The numbering has been done
by the following rule: the lowest hexagon
was assigned the number 0, the next higher
one gets a 1 (remember this continues on
the back side, not visible in the picture),
then the next higher is assigned the number
2, and so on. Notice that hexagon 42 is
slightly higher than hexagon 37. You
should be able to identify three distinct
spirals: one will have the hexagons 0, 5, 10,
... ; the second will have hexagons 0, 13,
Figure 2-7 26, . . . ; and the third spiral would include
hexagons 0, 8, 16, . . . . Now look at the
common difference between the spirals and you will find the differ-
ences to be 5, 8, and 13--once again Fibonacci numbers.
The Pinecone
There are various species of pinecones (e.g., Norway spruce; Doug-

las fir or spruce; larch). Most will have two distinct direction spirals.
Spiral arrangements are classified by the number of visible spirals
(parastichies) that they exhibit. The number of spirals in each direc-
tion will be most often two successive Fibonacci numbers. When
this occurs,3 the angle between successive leaves or botanical ele-
ments is close to the "golden angle"-about 137.5°, which is related
to the golden ratio (see chapter 4). The connection between the
golden angle and the Fibonacci numbers can be seen in the ratio
- - , W h·IC h b ecomes lmme
- '50
137 . d'late Iy c Iear w h en we re d uce th e
360 0
fraction (see chapter 5) as:
3. In spiral phyllotaxis, botanical elements grow one by one, each at a constant

divergence angle d from the previous one. This is the most common pattern, and
most often the divergence angle d is close to the golden angle. The latter case
gives rise to Fibonacci phyllotaxis.
137.5 0 55
360 0 144
The two pictures in figure 2-8 will bear this out, but you may
want to convince yourself by getting some actual pinecones and
counting the spirals yourself.
Figure 2-8
These pinecones have eight spirals in one direction and thirteen in

the other direction. Again, you will notice these are Fibonacci
numbers.
Tree Number of spirals Number of spirals
(Species) in one direction in the other direction
Norway spruce 13 8
Douglas fir (spruce) 3 5
larch 5 3
pme 5 8
More details can be found in an article by Brother Alfred

4
Brouseau.
The following designations of spiral patterns do not make any
pretense of completeness. Actual1y bracts can be lined up into se-
quences in many ways. The fol1owing are simply some of the more
obvious patterns that have been observed. (As for notation, 8-5, for
example, means that starting from a given bract and proceeding
4. Brother A. Brousseau, "On the Trail of the California Pine Spiral Patterns on
California Pines," Fibonacci Quarterly 6, no. I (1968): 76.
along two spirals, 8 bracts will be found on one spiral and 5 on the
other when going to the next intersection of the spirals.)
Pinus albicaulis (Whitebark pine), 5-3, 8-3, 8-5

Pinus flexilis (Limber pine), 8-5, 5-3, 8-3
Pinus Lamberttana (Sugar pine), 8-5, 13-5, 13-8, 3-5, 3-8,
3-13,3-21
Pinus monticola (Western white pine,
Silver pine), 3-5
Pinus monophylla (One-leaved pinon), 3-5, 3-8
Pinus edulis, 5-3
Pinus quadrifolia (Four-leaved pinon), 5-3
Pinus aristata (Bristlecone pine), 8-5, 5-3, 8-3
Pinus Balfouriana (Foxtail pine), 8-5, 5-3, 8-3
Pinus muricata (Bishop pine), 8-13,5-8
Pinus remorata (Santa Cruz Island pine), 5-8
Pinus contorta (Beach pine), 8-13
Pinus Murrayana (Lodgepole pine,
Tamarack pine), 8-5, 13-5, 13-8
Pinus Torreyana (Torrey pine), 8-5, 13-5
Pinus ponderosa (Yellow pine), 13-8, 13-5, 8-5
Pinus Jeffreyi (Jeffrey pine), 13-5, 13-8, 5-8
Pinus radiata (Monterey pine), 13-8, 8-5, 13-5
Pinus attenuata (Knobcone pine), 8-5, 13-5, 3-5, 3-8
Pinus Sabiniana (Digger pine), 13-8
Pinus Coulteri (Coulter pine), 13-8
In a similar detailed article we find spiral patterns, especially

5-8, in Aroids. The Aroids (family Araceae) are a group of attrac-
tive ornamental plants which include the very familiar Aglaonema,
Alocasia, Anthurium, Arum, Caladium, Colocasia, Dieffenbachia,
Monstera, Philodendron, Scindapsus, and Spathiphyllum species. 5
Spirals are also found on a variety of plants. Here are a few repre-
sentative examples. You may well want to search for other exam-
ples yourself.
5. T. Antony Davis and T. K. Bose, "Fibonacci System in Aroids," Fibonacci

Quarterly 9, no. 3 (1971): 253-55.

Mammilnria huitzilopochtli Mammilnria magnimamma
has 13 and 21 spirals_ has 8 and 13 spirals_

Marguerite has 21 and Gymnocalcium izozogsii
34 spirals_ has twice 5 and 8 spirals
(i.e_. 10 and 16 spirals)_

Knautia arvensis has twice This anonium has 3
2 and 5 spirals (i_e_. 4 and and 5 spirals_
10 spirals)_

Anonium_ Anonium_
The sunflower, on the other hand, has a variety of different spi-

ral numbers. The older the flowers get, the more spirals they de-
velop. In any case, however, the number of spirals will be a
Fibonacci number. They usually will take on the following pairs of
Fibonacci numbers: 13 (left-oriented spirals); 21 (right-oriented
spirals), 21 :34, 34:55, 55:89, 89:44.
~
...
. .•
. .
... ·tI· .
"......... ".. .
.......
............................
.~." ' .... ~: ~ .. ..... ~
""
""..
II • • •
..::.-
'''- .,.,
Z. a·-·, .... .
'-" .
'l'!::;...... ~
o
.... , •• ."
"'''..... ..
~. . . . '~4
• JI .. ' , t .... .t
r ••
'#~3?J.•..••:::. '. J.::;:.

. 'r..:..... . .... ... ...:. .. ".
~" '
....••••••t.' ........ ~ .
:::-"'~
.~.. ".
· , ·-
·.·
. .••
. .·.--..
. 10 t l

Sunflower 1. Sunflower 2.
-
. "-
. "

Sunflower 3. Sunflower 4.
In 2004 the principality of Liechtenstein 6 devoted an entire

postage stamp series to science, called the "Exact Sciences;
(CHF- .85) Mathematics; 2004." The stamp depicts the logarithmic
spiral, exponential growth, and the 39th Mersenne prime number, 7
2 13 .466.917 - I
One of the subjects highlighted a sunflower. 8
Figure 2-21
Based on a survey of the literature encompassing 650 species

and 12,500 specimens, R. Jean9 estimated that, among plants dis-
6. An independent state between Austria and Switzerland.

7. The 39th Mersenne prime number was found in 2001; and in 2005 we have
come as far as finding the 42nd Mersenne prime number 2 25 964 951 - I.
8. For sunflowers there exists three kinds of spirals: 82 percent are Fibonacci
spirals (8,13,21,34, ... ); 14 percent are Lucas spirals (7, II, 18,29, ... ); and
2 percent are Fibonacci bijugate spirals (10, 16,26,42, ... ). J. C. Sehoute,
"Early Binding Whorls," Rec. Trav. Bal. Neer I, no. 35 (1938): 416-558.
9. Roger V. Jean, Phyl/olaxis: A Syslemic SIudy in Plant Morphogenesis (Cam-
bridge: Cambridge University Press, 1994).
playing spiral or multijugate phyllotaxis,lo about 92 percent of

them have Fibonacci phyllotaxis.
One may wonder why this happens. Certainly mathematicians
did not influence this phenomenon. The reason seems to be that
this arrangement forms an optimal packing of the seeds so that, no
matter how large the seed range, with all the seeds being the same
size, they should be uniformly packed at any stage; this means no
crowding in the center and not too sparse near the edges. We "see"
this packing as spirals, which almost always translates into adja-
cent Fibonacci numbers.
Artificial Fibonacci-Flower Spirals

" .
........ '. .................... ........•
. .
........-.._""'- ...."
'
· . s.-.-s- ...... . '
\ ..
::' :.:' : .......................... ". ......... "\ \ \
· .
. . ".
::: :-,....' ,...............
..' , , " ' - ! :
.: .:'
a : : :
••••
a.
as
•
a.
•
. . /./ j j
. . . I '. : . : I
,: : :
~,
: : ; \ ~
•
· a.. :... • •
: : : .
:
!
.............:.::.....:~~.- ..-.-- .. /./ / } {.! \
\ ~~
....
.•... -----...............
aa • •
.... s·- ...

as a. • • • • • • • • • • • ..- ••
.. " . ......................
. . . . . .. .
".
" .
....
. . . . .. . . . . . . . .. . . . ..
. . . . . . . . .. . ...
'. .. . ....
. . . .. .. ...... . . . .. ...
· .. . . ......... . . .. .. . ...
. . . ... .a.sa..... . .. . ... :.. ..... '. '. '. '. .. ..
' ' '
. .. ....... .
: -: -.: :~-:·:i~;~l:(~;.~.-~}lWi:/:-~ -:- : .

• • • • • a.a ••••••••••••
• • • • • • •a ••••••••••••
".~.: .: :.::·::·::::~::~~~\~;:ii(.;. ;.;':' :--.....

. ..:: :
: :..:.:..:.::.~::: ::.' . ..
. .
. .. . . . .. .. .. .. .. . . . . ...
. ...... .......
. . .
.. ....... .
. . .......
.... . . . .. ..
....
· ... . ... . .. . ..... .
• • ••••• a ••••••••••
. . . .. . . . . .........
.. . . ' . .
. . . .. . .
..
Figure 2-22
10. From the Greek: Phyllo means leaf, and taxis means arrangement.
Leaf Arrangement-Phyllotaxis
Having now concentrated at the center of the sunflower, or, for that
matter, the daisy, we can now inspect the petals surrounding the
center. Again, you will discover that most plants will have the
number of such petals corresponding to a Fibonacci number. For
example, lilies and irises have 3 petals; buttercups have 5 petals;
some delphiniums have 8 petals; com marigolds have 13 petals;
some asters have 21 petals; and daisies can be found with 34, 55,
or 89 petals.
Here is a short list with some flowers arranged by the number
of petals they generally have:
3 petals: iris, snowdrop, lily (some lilies have 6 petals formed from
two sets of 3)
5 petals: buttercup, columbine (aquilegia), wild rose, larkspur,
pinks; also apple blossom, hibiscus
8 petals: delphiniums (larkspurs), Cosmos bipinnatus, II Coreopsis
tinctoria
13 petals: com marigold, cineraria, some daisies, ragwort
21 petals: aster, chicory, Helianthus annuus
34 petals: pyrethrum and other daisies
55, 89 petals: michaelmas daisies and other Compositae (asteraceae
family)
Some species, such as the buttercup, are very precise about the
number of petals they have, but others have petals whose numbers are
very near those above-with the average being a Fibonacci number!
Having now inspected the parts of flowers, we can look at the
placement of the leaves on a stem. Take a plant that has not been
pruned and locate the lowest leaf. Begin with the bottom leaf, and
count the number of rotations around the stem, each time going
11. These are species of Compositae-one of the largest families of the vascular
plants. See P. P. Majumder and A. Chakravati, "Variation in the Number of
Rays and Disc-Florets in Four Species of Compositae," Fibonacci Quarterly 14
(1976): 97-100.
through the next leaf up the stem, until you reach the next leaf
whose direction is the same as the first leaf you identified (that is,
above it and pointing in the same direction). The number of rota-
tions will be a Fibonacci number. Furthermore, the number of leaves that
you will pass along the way to reach the "final" leaf will also be a
Fibonacci number.
In figure 2-23, it took five revolutions to reach the leaf (their eighth
leaf) that is in the same direction as the first one. This phyllotaxis (i.e., leaf
arrangement) will vruy with different species, but should be a Fibonacci
number. If we refer to the rotationlleaf number ratio for this plant, it would
be about ~. One also describes the CUIVe marked in figure 2-23 as the
8
"genetic spiral of a plant."
Which is directly
above"" flrst_
•
-'",TURN'
_ 4th TURN
- 3rd TURN
_ 2nd TURN
- 1st '''-N
--------- _ _ _ .J
Figure 2-23
Here are some phyllotaxis ratios:
1/2:elm, linden, lime, some grasses

3/8:asters, cabbages, poplar, pear, hawkweed, some roses
1/3:alders, birches, sedges, beech, hazel, blackberry, some grasses
2/5:roses, oak, apricot, cherry, apple, holly, plum, common
groundsel
8/21: fir trees, spruce
5/13: pussy willow, almond
13/34: some pine trees
Different species of palms display different numbers of leaf spi-

rals, but the numbers always match with Fibonacci numbers. For
example, in the arecanut palm (Areca catechu), or the ornamental
Ptychosperma macarthurii palm, only a single foliar spiral is dis-
cernible, while in the sugar palm (Arenga saccharifera), or Arenga
pinnata, two spirals each are visible. In the palmira palm (Borassus
flabellifer), or Corypha elata, as well as in a number of other species
of palms, three clear spirals are visible. The coconut palm (Cocos
nucifera) as well as Copernicia spirals have five spirals, while the
African oil palm (Elaeis guineensis) bears eight spirals. The wild
date palm (Phoenix sylvestris) and a few other species of palms also
show eight spirals. On stout trunks of the Canary island palm
(Phoenix canariensis), thirteen spirals can be observed. Also in
some of these plants, twenty-one spirals can be found. Palms bear-
ing 4, 6, 7, 9, 10, or 12 obvious leaf spirals are not known. 12
This phyllotaxis ratio of Fibonacci numbers is not guaranteed
for every plant, but it can be observed on most plants. Why do
these arrangements occur? We can speculate that some of the cases
may be related to maximizing the space for each leaf, or the aver-
age amount of light falling on each one. Even a tiny advantage
would come to dominate, over many generations. In the case of
close-packed leaves in cabbages and succulents, such arrangements
may be crucial for availability of space.
12. T. Antony Davis, "Why Fibonacci Sequence for Palm Leaf Spirals?" Fibo-
nacci Quarterly 9 no. 3( 1971): 237-44.
Much has been written about the arrangement of leaves, petals,

and other plant aspects, but earlier works were just descriptive and
did not explain any connection between numbers and plant growth.
Rather they dealt with the geometry of the arrangements. The most
dramatic insight comes out of a recently published work by the
French mathematical physicists Stephane Douady and Yves
Couder. They developed a theory of the dynamics of plant growth
using computer models and laboratory experiments in relation to
the Fibonacci pattern. Douady and Couder 13 also found a dynamic
explanation for the golden angle (137.5°). They obtained this angle
as a consequence of simple rules of dynamics and did not postulate
it as many of their predecessors, who saw it as a consequence of a
space-saving arrangement. The spiral arrangement does not require
any special botanical explanation.
Last, but not least, we can continue our search for Fibonacci
numbers on the human body: A human being has 1 head, 2 arms,
and 3 finger joints, and 5 fingers on each arm-these are all
Fibonacci numbers. Wow! But, actually, it does not go on like this!
Moreover, the "small" Fibonacci numbers are just pure coinci-
dence, because among the eight numbers from 1 to 8, there are five
Fibonacci numbers. Therefore you have a good chance of finding a
Fibonacci number purely by coincidence. The percentage of Fibo-
nacci numbers among the rest of all other numbers changes dramati-
cally if we look at bigger intervals of numbers. On the other hand, if
you can find two different large Fibonacci numbers in one sun-
flower, you might ask, how this could be? (See page 68).
Particularly at the end of the nineteenth century, many scientists
were of the opinion that the golden section was a divine, universal
law of nature. Yet the golden section, l/J, can be seen as a limit of a
sequence, that is, made of the ratio of successive Fibonacci numbers.
(This will be covered in chapter 4 in great detail.) So, how then can
the human being, the "pride of creation," not have been designed by
the rules of the golden section?! Some see it as if it were. Leonardo
13. Stephane Douady and Yves Couder, "Phyllotaxis as a Physical Self-

Organized Growth Process," Physical Review Letters 68, no. 13 (1992):
2098-101.
da Vinci (1452-1519)14 constructed the proportion of the human

body based on the golden section, l/J (see pages 257-60). Further-
more, da Vinci partitioned the face vertically into thirds: forehead,
middle face, and chin. Plastic surgeons further divide the face hori-
zontally in fifths. In this way, we see once again the presence of the
two Fibonacci numbers, 3 and 5, come into play.
The Belgian mathematician and astronomer and founder of the
social statistics Lambert Adolphe Jacques Quetelet (1796-1874)15
and the German author, critic, playwright, poet and philosopher
Adolph Zeising (1810-1876)16 measured the human body and saw
its proportions related to the golden section, and thus greatly influ-
enced future generations.
The French architect Le Corbusier (Charles-Edouard Jeanneret,
1887-1965) assumed that human body proportions are based on
the golden section. Le Corbusier formulated the ideal proportions
in the following way: height 182 cm; navel height 113 cm; finger-
tips with arms upraised 226 cm. The ratio of height to navel height
is ~::::: 1.610619469, a very close approximation of ¢ . By another
113
measurement of the same features he got an analogous ratio of

176cm ::::: 1.6147, which is another good approximation of the ratio
109cm
of two Fibonacci numbers, 13 and 21, or ~~ = 1.615384615.

Looking further-maybe a bit too far!-the American Frank A.
Lonc l7 measured sixty-five women to substantiate the work of da
Vinci and Zeising by determining the mean ratio of a woman's
height to her navel. He then established that the ideal proportions
were those nearest 1.618, or approximately the golden section. In-
14. Italian painter, sculptor, architect, and engineer.

15."Des proportions du corps humain," Bulletin de I 'Academie Royale des sci-
ences, des lettres et des beaux-arts de Belgique, vol. 1. Bruxelles 1848-15,1,
pp. 580-93, and vol. 2.-15,2. pp. 16-27.
16. "Neue Lehre von den Proportionen des menschlichen K6rpers" (Leipzig: R.
Weigel, 1854).
17. Martin Gardner, '~About Phi," Scientific American, August 1959, pp.
128-34.
deed T. Antony Davis and Rudolf Altevogt l8 claimed that the

value of "a good-looking human body" based on Le Corbusier for
both boys and girls of a similar age is approximately the golden
ratio. Specifically, using 207 pupils in Munster, Germany, and 252
youngsters from Calcutta, India, they derived the value 1.615.
The ubiquitous appearances of the Fibonacci numbers in nature
are just another piece of evidence that these numbers have truly
phenomenal qualities.
18. "Golden Mean of the Human Body," Fibonacci Quarterly 17, no. 4 (1979):
340-44.
Chapter 3
The Fibonacci Numbers and

the Pascal Triangle
B y now you probably expect the Fibonacci numbers to appear

in the most unexpected places. "Sightings" of the Fibonacci
numbers can either be of individual members of the sequence, as in
the case of many of their appearances in nature, or as a sequence of
numbers appearing where you least expect them. We will begin
with some unusual sequences and then bring the Fibonacci num-
bers into the discussion-yes, when you least expect them to ap-
pear. Not only will they appear, but they will interact with various
other fields in a very meaningful way.
Some Sequences
So far we have seen that the Fibonacci numbers emanate from a

nicely defined sequence, originally determined by a problem deal-
ing with the regeneration of rabbits. We know they follow a
rule-that beginning with two 1s, the sum of every two consecu-
tive members of the sequence gives the next number in the se-
quence. This may not be the usual kind of sequence to which we
are accustomed. We usually like sequences such as:
1, 2, 3, 4, 5, 6, 7, ...
77
or
2, 4,6, 8, 10, 12, 14, ...
or
1,3,5,7,9,11,13, ...
We can even easily accept:
5,10,15,20,25,30,35,40, ...
or even the sequence of perfect squares:
1, 4, 9, 16, 25, 36, . . .
Suppose we didn't recognize that the above was the sequence of

perfect squares; we could check the differences between terms to
see if there is a common difference (or perhaps a common factor).
For example, if there is a common difference, then you can easily
get the next number, but if there is no common difference, then
you could try to get the next row of differences-that is, between
the terms of differences. Look at the chart below (figure 3-1) and
notice that the second differences are constant.
Original 1 4 9 16 25 36
Sequence
First 3 5 7 9 11
Difference
Second 2 2 2 2
Difference
Figure 3-1
However, when we get the sequence 1, 2, 4, 8, 16, we expect

the next number to be 32. Yet if we were told that it is 31, we
might cry out "wrong!"
Much to many people's surprise, either 32 or 31 "works" to
create a viable or meaningful sequence.
THE FIBONACCI NUMBERS AND THE PASCAL TRIANGLE 79
Yes, what may be much to your amazement, 1, 2, 4, 8, 16, 31,

... can be a legitimate sequence.
To show that this is a proper sequence (one that has a rule for
getting additional members of the sequence), we shall set up a ta-
ble of differences. Just as we might have done for the first four se-
quences above, this will enable us to determine the rule for deriv-
ing succeeding members of the sequence.
We will set up a chart (figure 3-2) showing the differences
between terms of the sequence 1, 2, 4, 8, 16, 31, ... , and then the
differences of the differences, and so on until a pattern emerges.
That is, we take the differences of the terms in the given sequence,
and if no pattern emerges, we shall try to find a pattern in the new
sequence formed by these differences. We will do this in exactly
the same way: take differences between the terms and look for a
pattern. The first row of differences leaves us without much of a
clue of a pattern, since again the last term ruins what may have de-
veloped into a pattern. The second row, likewise. However, at the
third sequence of differences, a clear pattern emerges: 1, 2, 3, ....
Original 1 2 4 8 16 31
Sequeoce
First I 2 4 8 15
Difference
Second I 2 4 7
Difference
Third I 2 3
Difference
rourth I I
dIfference
Figure 3-2
With the fourth differences forming a sequence of constants,

we can reverse the process (i.e., tum the table upside down as in
figure 3-3) and extend the third differences a few more steps, say,
up to 9.
Fourth 1 1 1 1 1 1 1 1 t
Difference
Third 1 2 3 4 5 6 7 8 9
Difference
Second 1 2 4 7 II 16 22 29 37 46
Difference
First 1 2 4 8 15 26 42 64 93 130
Difference
Onginal 1 2 4 8 16 31 57 99 163 256 386
Sequence
Figure 3-3
The bold numbers are those that were obtained by working

backward from the sequence of third differences. So the next num-
bers of the given sequence are 57, 99, 163, 256, and 386. 1
After we have just taken you from what you expected would be
a comfortable and familiar sequence, 1, 2, 4, 8, 16, 32, 64, 128,
256, ... , to a rather strange-looking one, 1, 2'; 4, 8, 16, 31, 57, 99,
163,256,386, ... , you should not think that this sequence is con-
trived and is independent of other parts of mathematics. A geomet-
ric interpretation of this unusual sequence should help convince
you of the consistency we enjoy in mathematics, and give you rea-
son to better appreciate the beauty inherent in mathematics.
Consider a sequence of circles, each with successively more
points highlighted on it. Beginning with the first circle, we will
join these points with straight line segments to form the greatest
number of regions in each of these circles. Let's count the number
of regions in each circle successively. In figure 3-4 we show five
circles, each with successively more points on it. We have omitted
the actual "first circle"-the one with only one point on it, since it
would show only one region formed-a trivial case. We marked
the regions with consecutive numbers so that the number of re-
gions can be easily counted.
1. The general tenn (that is, the nth tenn) of the sequence 1, 2, 4, 8, 16, 32, ... is
easy to find, and it is T(n) = 2 n - I. The general tenn of the sequence 1, 2, 4, 8,
16, 31, ... is a fourth-power expression, since we had to go to the third differ-
ences to get a constant.
4 3 2
. n -6n +23n -18n+24
The general tenn IS T(n) = . [But also the general tenn T(n) =
24
n+(:) + C~I) = (:) + C) + I is for all natural numbersn (wheren>O) a de-

scription of this sequence.]
8
1
2 3
1 2
5 .4
6 7
Figure 3-4
To summarize this circle partitioning we will make a chart

(figure 3-5) of the number of regions into which a circle can be
partitioned by joining the highlighted points on the circle. If you do
this independently for further cases, make sure no three lines are
concurrent (i.e., meet at one point), or else you will lose a region.
Number of points Number of regions into which

on the circle the circle is partitioned
I 1
2 2
3 4
4 8
5 16
6 31
7 57
8 99
Figure 3-5
Notice that the number of regions listed on the chart gives us

(what originally appeared to be) this unusual sequence.
Fibonacci Differences
By now you might be asking how this relates to the Fibonacci num-
bers. First of all, we know that the Fibonacci sequence does not have
a common difference between its terms. After our previous discus-
sion on sequences, this might "de legitimatize" the Fibonacci se-
quence in your mind. But, before drawing such rash conclusions,
let's take a look at the differences between the terms of the
Fibonacci sequence (figure 3-6).
Original 1 1 2 .1 ~ 8 13 21 34 55 89 144
~ucnce
Fir;t 0 1 1 2 3 5 8 13 21 34 55
DIfTerence.,
Sc=cond 0 1 I 2 3 5 8 13 21 34
D!ncrcncc~
Third 0 I 1 2 3 5 8 13
Differc=nces
rourth 0 I 1 2 3 5 IS
Dlncrem.e~
Figure 3-6
Using the same procedure for inspecting the above sequences,

we quickly notice that everywhere on this chart of differences, you
will see the Fibonacci sequence. Its appearance could overwhelm us.
It is in the first, second, third, fourth, ... differences, as well as
along the diagonals. It appears to reproduce itself in almost all di-
rections! Such a characteristic is one of the many reasons that this
sequence has these unusual properties.
If we complete the chart in figure 3-6 using the relationship
F: = F:+2 - Fn+l , which comes from the definition of the Fibonacci
numbers (i.e., Fn + Fn+1 = Fn+2 ), the list of differences would be:
Fibonacci sequence: 1, 1, 2, 3, 5, 8, 13, .. .

sequence of first differences: 0, 1, 1, 2, 3, 5, 8, 13, .. .
sequence of second differences: 1,0,1,1,2,3,5,8,13, .. .
sequence of third differences: -1,1,0,1, 1,2,3,5,8,13, ..
sequence of fourth differences: 2, -1,1,0,1, 1,2,3,5,8,13, ...
(Notice that Fo = 0, F_I = 1, F_2 = -1, ... )

Furthermore, the sequences of sums generate the Fibonacci se-
quence again, although a bit shifted over (see figure 3-7).
Sequence of Sums 10,946
4,181 6,765
1,597 2,584 4,181
610 987 1,597 2,584
233 377 610 987 1,597
89 144 233 377 610 987
34 55 89 144 233 377 610
13 21 34 55 89 144 233 377
a+b 5 8 13 21 34 55 89 144 233
a b 2 3 5 8 13 21 34 55 89 144
Fn 1 1 2 3 5 8 13 21 34 55 89
~
0 1 1 2 3 5 8 13 21 34
b-a 1 0 1 1 2 3 5 8 13
-1 1 0 1 1 2 3 5
2 -1 1 0 1 1 2
-3 2 -1 1 0 1
5 -3 2 -1 1
-8 5 -3 2
13 -8 5
-21 13
Sequence of Differences 34
Figure 3-7
Their appearance in the realm of sequence differences is quite

amazing, but how does the Fibonacci sequence relate to the other
sequences we just inspected? Or put another way: does the Fibo-
nacci sequence relate to these seemingly unrelated sequences? If

the past gives us a clue about the future, the answer should be an
unexpected "yes." Let's see if we can draw a connection between
these three unrelated sequences.
The sequence of powers of2:
1, 2, 4, 8, 16, 32, 64, 128, 256, 512, 1,024, ...
The sequence of circle partitions:
1,2, 4,8, 16, 31, 57,99, 163, 256, 386,
The Fibonacci sequence:
1,1,2,3,5,8,13,21,34,55,89,144,233,
To do this we will introduce you to the Pascal triangle, named

after the famous French mathematician Blaise Pascal (1623-1662),
who developed his "arithmetical triangle" as an aid for his work on
probability. Pascal showed his genius for mathematics as a child.
Even though mathematics books were forbidden to him as a child,
his desire and talent for mathematics could not dampen. Through his
own inventive construction, he independently proved Euclid's
thirty-second proposition, establishing the angle sum of a plane tri-
angle. In order to help his father process huge amounts of data in his
position as tax commissioner, he developed a mechanical calculator
(called "Pascaline") that by 1645 was on the market for purchase.
The Pascal Triangle
In 1653 Pascal invented his arithmetical triangle,2 although it was

not published until after his death. This famous triangular ar-
2. This triangular arrangement of numbers was first described by the Arabian

mathematician Omar Khayyam (1048-1122), but first appeared in print in a
1303 manuscript "The Valuable Mirror of the Four Elements" by the Chinese
mathematician Chu Shih-Chieh (1270-1330). Yet in the Western world Pascal is
credited with having discovered it without reference to any previous documents.
rangement of numbers (figure 3-8) is formed by beginning at the

top with 1, then the second row has 1, 1, then the third row is ob-
tained by placing 1s at the end positions and adding the two num-
bers in the second row (1 + 1 = 2) to get the 2, placed below the
interval between the 1s of the previous row. The fourth row is ob-
tained the same way: after the end 1s are placed, the 3s are gotten
from the sum of the two numbers above (to the right and left); that
is, 1 + 2 = 3 and 2 + 1 = 3. This pattern then continues to get the
succeeding rows.
The Pascal Triangle
3 3
4 6 4
5 10 10 5
6 15 20 15 6
7 21 35 35 21 7
R 28 56 70 56 28 8
9 36 84 )26 126 84 36 9 J
10 45 120 210 252 210 120 45 10
Figure 3-8
Let's look at some of the hidden gems in the Pascal triangle

(figure 3-8). Ifwe take the sum of the numbers in each row, we no-
tice .that we have arrived at the successive powers of 2 (figure 3-9).
Namely,
0
1= 1= 2 ,
1+1=2=21 ,
1+ 2 + 1 = 4 = 22 ,
1 + 3 + 3 + 1 = 8 = 23 ,
1 + 4 + 6 + 4 + 1 = 16 = 24 ,
1 +5+ 10+ 10+5+ 1 =32=2 5 ,
1 +6+ 15+20+ 15+6+ 1 =64=2 6 ,
1 +7+21 +35+35+21 +7+ 1 = 128=2 7,
1 + 8 + 28 + 56 + 70 + 56 + 28 + 8 + 1 = 256 = 2 8 ,
1 +9+36+84+ 126+ 126+84+36+9+ 1 =512=2 9 ,
1 + 10+45+ 120+210+252+210+ 120+45 + 10+ 1 = 1,024=2 10
This is the first sequence of the three we will connect through the
Pascal triangle.
Row sums
(Po..... ers of2)
2 4
3 3 R
4 6 4 16
5 10 10 5 32
6 15 20 15 6 64
7 21 35 ,5 21 7 128
R 28 56 70 56 28 R 256
9 36 ~4 126 126 84 36 () 512
JO 45 120 210 252 210 120 45 \0 1.024
Figure 3-9
Now, consider only the horizontal sums of the rows in figure

3-10 that are to the right of the bold line drawn through the Pascal
triangle. There, strangely enough, the sequence of circle partitions
emerges!
1, 2, 4, 8, 16, 31, 57, 99, 163
What a coincidence! So the Pascal triangle provides us with a con-

nection between the first two of the sequences that we wanted to
connect.
The Pascal triangle contains many unexpected number relation-
ships beyond the characteristics that Pascal intended in his original
use of the triangle. The ubiquity of this triangular arrangement of
numbers merits just a bit of further elaboration before we relate the
third of our sequences of numbers-the Fibonacci numbers-to the
others. Pascal's original use was to exhibit the coefficients of suc-
cessive terms of a binomial expansion.
Sum of the
numbers
to the right of the
line
I
4 16
5 31
IS 6 57
21 7 99
56 28 8 163
84 36 <) 256
210 120 45 10 386
Figure 3-10
That is, taking a binomial such as (a + b) to successively high-

er powers.
(a + b)o = 1
(a+b)1 =a+b
(a + b)2 = a 2 + 2ab + b 2
(a + b)3 = a 3 + 3a 2b + 3ab2+ b 3
(a + b)4 = a 4 + 4a 3b + 6a 2b 2 + 4ab 3 + b 4
(a+b)5 =a5 +5a4b+ 10a 3 b 2 + 10a2b 3 +5ab4 +b 5
(a+b)6 =a6 +6a5b+ 15a4b 2 +20a 3b 3 + 15a2b 4 +6ab5 +b6
(a + b)7 = a 7 + 7a 6b+ 21a 5b 2 + 35a4b 3 + 35a3b4+21a2b5+7ab6+b7
(a + b)8 = a 8 + 8a7b + 28a6b 2 + 56a5b 3 + 70a4b 4 + 56a 3b 5 + 28a2b 6 + 8ab7 + b 8
(a + b)9 = a 9 + 9a8b+ 36a7b 2 + 84a 6 b 3 + 126a5b 4+ 126a4b5+84a3b6+36a2b7
+9ab8+ b 9
(a + b)IO = a lO+l0a9b+ 45a 8 b 2 + 120a7b 3 + 210a6 b 4 + 252a sb 5 + 210a4b 6 + 120a 3b 7
+ 45a2b 8 + 10ab9 + b 10
Notice how the coefficients of each binomial-expansion line is

also represented as a row of the Pascal triangle. This allows us to
expand a binomial without actually multiplying it by itself many
times to get the end result. There is a pattern also among the vari-
ables' exponents: one descends while the other ascends in
value-each time keeping the sum of the exponents con-
stant-making it equal to the exponent of the original binomial
taken to a power.3
3. Some may be interested in the general term of the expansion of (a + b), so we

offer it here:
(a+br = (:) an + (;) an-I b + (;) a

n 2 2
- b + ... + L:2) 2 n 2
a b - + L:J n I n
ab - + (:) b ,
where (n)=
k
n! ,andn!=1.2.3.4
k! ·(n-k)!
... n.
We should note that we define (~) = 1.
F or example, to get the fourth entry in the eighth row, we use
n) (7) 7! 7! 4!·5·6·7 5·6·7 5·6·7

(k = = 3!·(7-3)! =
3 3!·4! = 3!·4! = ~ = 1.2.3 = 35.
By the way, the binomial coefficient (:) tells us how many ways n coin tosses
come up with k heads.

THE FIBON ACCI NUMBERS AND THE PASCAL TRIANGLE 89
There are many more intriguing surprises embedded in this

wonderful triangular arrangement of numbers. Consider the Pascal
triangle in figure 3-11 with the various cells shaded.
Figure 3-11
First we notice the list of I s on both sides of the triangle:

1,1, I, I, I, I, 1, I, I, 1, I, 1, ...
The next list in the parallel direction is the list of natural
numbers:
1,2,3,4,5,6,7,8,9,10,11,12,13, ...
Continuing in this pattern (i.e., parallel lists) we get the list of
triangular numbers: 4
1,3,6, 10,15,21,28,36,45,55,66,78, ...
The next list of numbers in the same direction is the list of
tetrahedral numbers: 5
1,4, 10,20,35,56,84, 120, 165,220,286, ...
We then come to a lesser-known sequence of numbers that
exists in the fourth dimension called pentatop numbers:
1,5,15,35, 70,126,210,330,495,715, ...
Suppose you would like to add the numbers in one of these se-
quences of numbers. All you need to do is locate the number in the
cell that is immediately below and to the right (or left) of the last
4. Triangular numbers represent the number of points that can be arranged to

form an equilateral triangle.
5. Tetrahedral numbers represent the number of points that can be arranged to
form a regular tetrahedron.
number in the list of numbers you are adding. For example, con-
sider the triangular numbers (see figure 3-12):
I + 3 + 6 + 10 + 15 + 21 = 56
The same holds true for the sum of tetrahedral numbers:

I + 3 + 6 + 10 + 15 + 21 + 28 + 36 + 45 + 55 + 66 = 286
•
--.
•
..
•
»
•
,
•
"
"
"
•
•
•
~
•
'M
,.
»
M
~
'"
,.
~
N
.,
"
"
"~
"
- ~
'"
~
,~
m
nl'
~
n
" ..
Figure 3-12
Fibonacci Numbers and the Pascal Triangle
By now you must be wondering where on the Pascal triangle are the
Fibonacci numbers hiding, for, after all, isn't that what you expect
by now? They seem to crop up everywhere, even when we least ex-
pect to see them. Yes, indeed, the Fibonacci numbers are embedded
on the Pascal triangle. Look at the sums of the numbers along each
of the indicated lines (see figure 3-13). There you have the Fibo-
nacci numbers! We have then connected, with the help of the Pascal
triangle, the three sequences we sought to relate.
I
I
2
3
5
8
13
21
34
55
89
35 21 7
56 70 56 28 8
36 84 126 126 84 36 9
10 45 120 210 252 210 120 45 10
Figure 3-13
This may be easier to visualize if we left justify the triangular

arrangement of numbers as in figure 3-14:
Sum
1
1
1 2
2 3
3 1 5
4 3 8
5 6 1 13
6 10 4 21
7 15 10 1 34
8 21 20 5 55
9 28 35 15 1 89
10 36 56 35 6 144
11 45 84 70 21 233
Figure 3-14
Ron Knott, a British mathematician who has provided some

interesting insights into the Fibonacci numbers, found the Fibo-
nacci numbers appearing as sums of "rows" in Pascal's triangle.
o I 234 5 6 7 8 9 10 II 12 13 14 15 16 17 18 19 20
o I
I I
2 I 2 I
3 I 3 3
4 I 4 6 4
5 5 10 10 5
6 6 15 20 15 6
7 7 21 35 35 21 7
8 8 28 56 70 56 28 8
9 9 36 84 126 126 84 36 9
10 I 10 45 120 210 252 210 120 45 10
1I I II 55 165 330 462 462 330 165 55
12 I 12 66 220 495 792 924 792 495
13 I 13 78 286 715 1287 1716 1716
14 I 14 91 364 1001 2002 3003
15 15 105 455 1365 3003
16 16 120 560 1820
17 17 136 680
18 18 153
19 19
20
I I 235 8 13 21 34 55 89 144 233 377 610 987 1597 2584 4181 6765 10946
Figure 3-15
By drawing Pascal's triangle with all the rows moved over by

one place, we have a clearer arrangement, which shows the Fibo-
nacci numbers as sums of columns. (See the last row in figure 3-
15.)
We could also rewrite this version of the Pascal trian-
gle-moving each row over to the right by one column-and
writing each row twice. By doing this we find that the column
sums again yield the Fibonacci numbers (see figure 3-16).
2 J 4 5 6 7 X 9 10 II 12 U 14 15 16 17 18 19 20 21 22
0 1
1 1
I I
2 1 2 1
I 2
3 1 3 3
4 4 6 4
4 0 4
5 5 10 10 5
5 10 10 5
6 6 15 20 15 6
6 15 20 15 0
7 7 21 35 35 21 7
7 21 )5 35 21 7
8 8 28 56 70 56 28 8 1
g 28 56 70 56 2X 8
9 9 36 84 126 126 84 36 9
9 )0 X4 12tl 12tl X4 3tl 9
10 10 45 120 210 252 210 120 45 10
10 45 120 210 252 210 120 45 10
1I 11 55 165 330 462 462 330 165 55
II 55 165 330 462 462 330 165
12 12 66 220 495 792 924 792 495
12 66 220 495 792 924 792
13 13 78 286 715 1287 1716 1716
13 78 286 715 1287 1716
14 14 91 364 1001 2002 3003
14 91 364 1001 2002
15 15 105 455 1365 3003
15 105 455 1365
16 16 120 560 1820
16 120 560
17 17 136 680
17 130
18 18 153
H<
19 19
20
I 2 ) 5 X 13 21 34 55 89 144 233 377 610 9X7 1597 2584 4181 6765 10946 17711
Figure 3-16
If we revert once agaIn to the left-justified Pascal triangle

(figure 3-17), we can easily discover a nice group of palindromic
numbers: 6 1,001, 2,002, 3,003, 5,005, and 8,008.
6. Palindromic numbers are those that read the same in both directions, such as
3,003, or from the Pascal triangle: 1,331 and 14,641.
I I
I 2
I 3 3
I 4 6 4
I , 10 10 ,
I 6 20 6
I 7 21 " 35
"
35 21 7
I 8 28 '6 70 '6 28 8
I 9 36 84 126 126 84 36 9
I 10 120 21 0 m 210 120
III " 330 462 462 330 " 10
II
" 16'
112 66 220 49' 792 924 792
16'
49'
"
220 66 12
1287 78 13
113 78 286
'" 17 16
,,,,,
11 16 1287
'" 286
91 14
114 91 364 lOOt 2002 3432 300J 2002 100 1
""
I IS 105 455 1365 .~)(J.I
4368
.5OOS 6435
I J440
6435
12870
StMI5 XlOJ 136'
'" 10'
"
120
I 16 120
I 17 136
560 1820
680 2380
!4,()8
6188 12376 19448 243 10

11440
24.3 10
"""
19448
4368
12376
1820
6 1118
S6IJ
2380 680
"
136 17
I 18 153 8 16 3060 8568 18564 3 1824 4 3758 48620 4 3758 31824 1S564 8,.. 3060 81' m IS
I 19 17 . 969 3876 11 628 27 132 50388 75582 92378 92378 75582 5038827 132 11 628 3876 969 171 19 I
I 20 190 1140 4845 15504 38760 77520 125970 167960 184756 167%0 12597077520 3876() 15504 4845 11 40 190 20 I
Figure 3-17
In 1971 , David Singmaster- an American mathematician

living in London, who has analyzed many aspects of recreational
mathematics to secure them into the realm of serious mathemat-
ics 7--discovered 8 that the fifteenth row is the only one in which
three consecutive numbers will be in the ratio I : 2 : 3. They are:
1,001, 2,002, and 3,003. It should not be lost on the reader that
coincidentally the numbers under these three are also palindromes
in a Fibonacci ratio! (See figure 3-18.)
Notice that 1,001= 1·1,001; 2,002 = 2·1,001; 3,003 = 3·1,001;
5,005 = 5·1,001; and 8,008 = 8·1 ,001.
They all have 1,001 as their common factor.
1,001 2,002 3,003

3,003 5,005
8,008
Figure 3-18
7. Such as the Rubic ' s Cube.

8. American Mathematical Monthly 78 (1971): 385- 86.
Lest you worry about our quest for a relationship between the
Fibonacci numbers and the Pascal triangle, consider this. Using our
earlier symbols to represent the Fibonacci numbers, and con-
sidering that we can extend the Fibonacci number sequence in the
negative direction as follows: 9
. . . , 13, -8, 5, -3, 2, -1 , 1, 0, 1, 1, 2, 3, 5, 8, 13, . . . ,
we can also represent the Fibonacci numbers in terms of the

members of the Pascal triangle as shown in figure 3-19.
Fn+S= 1·Fn+8·FIl _ , + 28·Fn_2 + 56·Fn_3 + 70·Fn-4 + 56·FIl _ 5 + 28·Fn-fJ + 8·Fn_7 + l·Fn_s
Figure 3-19
At first glance, this may look artificial and contrived, but reserve
your judgment until you begin to substitute the Fibonacci values to
verify the relationship.
9. See, for instance, the sequence of fourth differences and the diagonal in figure
3-7 (from left to right).
You can see from the negatively extended Fibonacci sequence

that Fo == 0, F_ 2n == - F 2n , and F-2n+1 == F 2n - 1 for all natural numbers
n. Furthermore, the recursive definition from before (the actual
basis for the Fibonacci numbers), Fk +2 == F k+ l + Fk , still holds true
for all integer values of k.
To get a better understanding about this unusual relationship,
let's check the case where n == 5 (figure 3-20):
FI3 = 1·Fs +8·F4 +28·F3+56·F2 + 70·F 1 +56·Fo +28·F_ 1 +8·F_2 + 1·F_3
Figure 3-20
To get a better "feel" for this, we will substitute the appropriate

Fibonacci numbers in the listing shown in figure 3-20. This allows
us to calculate the "generated" Fibonacci numbers at the right in
figure 3-21.
F5 = 1·5 = 5
F6 = 1·5 + 1·3 = 8
F7 = 1·5 + 2·3 + 1·2 = 13
F8 = 1·5 + 3·3 + 3·2 + 1·1 = 21
F9 = 1·5 + 4·3 + 6·2 + 4·1 + 1·1 = 34
FlO = 1·5 + 5·3 + 10·2 + 10·1 + 5·1 + 1·0 = 55
FII = 1·5 + 6·3 + 15·2 + 20· I + 15·1 + 6·0 + 1·1 = 89
FI2 = 1·5 + 7·3 + 21·2+ 35·1 + 35·1 + 21·0+ 7·1 + 1·(-1) = 144
FI3 = 1·5 + 8·3 + 28·2 + 56·1 + 70·1 + 56·0 + 28·1 + 8·(-1) + 1·2 = 233
Figure 3-21
Lucas Numbers and the Pascal Triangle
As if this weren't enough, the Pascal triangle also enables us to con-

nect in yet another way the Fibonacci numbers to the Lucas numbers
that we mentioned in chapter 1. Recall that Edouard Lucas (1842-
1891), a French mathematician, developed the sequence of numbers
with the same recursive rule as the Fibonacci numbers, but he began
with 1 and 3 instead of 1 and 1. So he got the sequence 1, 3, 4, 7, 11,
18, ... instead of the Fibonacci numbers 1, 1, 2, 3, 5, 8, .... In fig-
ure 3-22 we list the Lucas numbers. To make things even more in-
teresting, however, we will begin the sequence one number earlier
than the traditional beginning. That is, instead of beginning with the
first Lucas number, L), we will begin with L o ' which is 2.
We will represent the nth Lucas number as Ln so that LI = 1,
L2 = 3 (and as we said earlier, Lo = 2), and Ln +2 = Ln+l + Ln.
As we begin to relate the Fibonacci and the Lucas sequences, we
should realize a direct connection: The nth Lucas number (n > 0) is
equal to the sum of the (n -1)st Fibonacci number and the (n + l)st
Fibonacci number. Symbolically, that is written as: Ln= F n-l + Fn+l.
n Ln
0 2
1 1
2 3
3 4
4 7
5 11
6 18
7 29
8 47
9 76
10 123
11 199
12 322
13 521
14 843
15 1,364
16 2,207
17 3,571
18 5,778
19 9,349
20 15,127
Figure 3-22
In figure 3-23 you can observe this relationship (Ln = F n- l +

F n+ l ) with the two sequences placed next to each other.
II 0 J 2 3 4 5 6 7 8 9 10 JJ J2 13 14 15 16 J7 18 19 20
F•• 1 1 2 3 5 8 13 2J 34 55 89 144 233 377 610 987 1,597 2,584 4.181 6.765 10.946
+
Fa_ 1 J 2 3 5 8 13 21 34 55 89 144 233 377 610 987 1597 2584 4181
-=
L. 1 3 4 , 11 18 19 4' 76 III 199 322 521 843 1,364 ~ 3,571 5,778 9,349 15,127
Figure 3-23
Since these two sequences are so closely related, it is only

fitting, and not particularly surprising, that we can also represent
the Lucas numbers in the arrangement of the Pascal triangle-just
as we did with the Fibonacci numbers (see figure 3-24). Keep in

mind that Lo = 2, L-2n = L 2n , and L-2n +1 = -L2n- 1 , for all natural
numbers n.
Ln+1 = l·L + 1·Ln_1

ll
Ln+7= l·L n+ 7·L n_1+21·Ln_2 +35·L

II _ 3 + 35·Ln-4 + 21·Ln_5 + 7·Ln-6 + 1·Ln_7
Figure 3-24
Moreover, the recursive definition, Ln+2 = Ln+l + L n, which is

the basis for the Lucas and the Fibonacci numbers, still holds true
for all integer values of n. We can see that in figure 3-25.
We are now ready to experience the beautiful Pascal triangle
relationship with actual Lucas numbers. For that we consider the
case where n = 5. To accomplish this, we will substitute the actual
Lucas numbers in the listing of figure' 3-24, so that we get the
listing shown in figure 3-26 with the Lucas numbers generated at
the right.
n Ln
0 2
-1 -1
-2 3
-3 -4
-4 7
-5 -11
-6 18
-7 -29
-8 47
-9 -76
-10 123
-11 -199
-12 322
-13 -521
-14 843
-15 -1,364
-16 2,207
-17 -3,571
Figure 3-25
Ls = I-II = II
L6 = I-II + 1-7 = 18
L7 = I-II + 2-7 + 1-4 = 29
Lg = 1·11 + 3-7 + 3·4 + 1-3 = 47
L9 = 1·11 + 4·7 + 6·4 + 4·3 -t 1·1 = 76
LIO == 1·11 + 5·7 + 10-4 + 10·3 , 5·1 + 1·2 = 123
LII = 1·11 + 6·7 + 15·4 + 20·3 ,15·1 + 6-2 -t 1-(-1 ) - 199
LI2 = I-II + 7·7 + 21·4 + 35-3 ,35·1 +21·2 -t 7-(-1) + 1-3 = 322
LI3 = 1·11 + 8·7 + 28·4 + 56·3 ,70·1 + 56·2 -t28-(-1) + 8·3 + 1·(-4) = 521
Figure 3-26
THE FIBONACCI NUMBERS AND THE PASCAL TRIANGLE 10 I
We can also get the initial Lucas numbers in the same way.
Consider the case where n = I, and substitute the values of the
appropriate Lucas numbers in figure 3-24. We then get figure 3-27,
with the Lucas numbers generated at the right.
L, = ) ·1 =
1-, = )·1 +1 -2 = 3
L, = ) · 1 +2·2+ )-(- 1) = 4
L. = ) -1 +3·2+ 3-(- 1) + ) -3 = 7
L, = )-1 +4·2+ 6-(- 1) + 4 ·3 + ).(-4) = 11
4 = )-1 +5-2+10-(- 1) + 10·3 + 5-(-4) + )-7 = 18
L, = )-1 +6-2+1S-(- I) + 20-3 + )5-(-4) + 6·7 + ) -(-11 ) = 29
L, = )-1 +7-2+2) -(- 1) +35-3 +35-(-4) +21·7 + 7 -(- 11 ) + 1- 18 = 47
L, = ) -1 +8-2+28-(-1) +56-3 +70-(-4) +56-7 + 28-(- 11) +8-18 +)-(- 29) = 76
Figure 3-27
For the sake of consistency, we would want to be able to find

the Lucas numbers on the Pascal triangle. To do so, we will modify
the original Pascal triangle by replacing the right side I s with 2s
and then carrying on the calculation as we did for the original
Pascal triangle to generate the rows. See figure 3-28.
,,
2
2
2
2 12
2 24
6 2
"
, ,
1 20 2 %
II
10
54
.. 35
156
21
112
11
294
'"
,,2
105
31.
196
336
2
2
2
192
J K4
'"
1.536
Figure 3-28
Before we search for the Lucas numbers, we should take note

of some interesting features of this type of Pascal triangle. Looking
at the diagonal rows in figure 3-28, you can see the odd numbers,
the square numbers, and the pyramidal numbers. In addition, if you
inspect the diagonal rows in the other direction, you will find a
progression followed by sequences of progressive differences, a
natural result of the Pascal triangle's construction.
As we found the Fibonacci numbers in the original Pascal
triangle, so, too, will we now locate the Lucas numbers in the
second type of Pascal triangle. To locate the Lucas numbers,
notice the sum of the numbers along each of the lines drawn in
figure 3-29. This is analogous to the way in which we located
the Fibonacci numbers on the Pascal triangle.
2
I
3
2 4
7
11
18
29
47
76
123
11 2
55 36 13 2
77 105 91 49 15 2
44 112 182 196 140 64 17 2
11 54 156 294 378 336 204 81 19 2
Figure 3-29
By writing this triangular arrangement of numbers in a left-

justified fashion (figure 3-30), we can get the column sums to
generate the Lucas numbers.
o 1 234 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
o2
1 12
2 132
3 14 5 2
4 I 5 9 7 2
5 1 6 14 16 9 2
6 I 7 20 30 25 II 2
7 8 27 50 55 36 13 2
8 1 9 35 77 105 91 49 15 2
9 10 44 112 182 196 140 64 17 2
10 1 11 54 156 294 378 336 204 81 19 2
11 12 65 210 450 672 714 540 285 100
12 13 77 275 660 1122 1386 1254 825
13 14 90 352 935 1782 2508 2640
14 15 104 442 1287 2717 4290
15 16 119 546 1729 4004
16 17 135 665 2275
17 18 152 800
18 19 170
19 20
20
2 1 347 11 18 29 47 76 123 199 322 521 843 1364 2207 3571 5778 9349 15127
Figure 3-30
If we repeat this with double entries, as we have for the Fibo-

nacci numbers on the Pascal triangle, we can again generate the Lu-
cas numbers by taking the column sums, as in figure 3-31.
0 I 2 3 4 5 6 7 8 9 10 II 12 13 14 15 16 17 18 19 20
0 2
2
1 2
I 2
2 1 3 2
I 3 2
3 1 4 5 2
I 4 5 2
4 1 5 9 7 2
5 9 7 2
5 6 14 16 9 2
I 6 14 16 9 2
6 1 7 20 30 25 11 2
7 20 30 25 II 2
7 8 27 50 55 36 13 2
8 27 50 55 36 13 2
8 9 35 77 105 91 49 15 2
9 35 77 105 91 49 15 2
9 1 10 44 112 182 196 140 64 17 2
I 10 44 112 182 196 140 64 17 2
10 1 11 54 156 294 378 336 204 81 19 2
II 54 156 294 378 336 204 81 19
II 1 12 65 210 450 672 714 540 285 100
12 65 210 450 672 714 540 285
12 1 13 77 275 660 1122 1386 1254 825
13 77 275 660 1122 1386 1254
13 14 90 352 935 1782 2508 2640
14 90 352 935 1782 2508
14 15 104 442 1287 2717 4290
15 104 442 1287 2717
15 16 119 546 1729 4004
16 119 546 1729
16 17 135 665 2275
17 135 665
17 18 152 800
18 152
18 19 170
19
19 20
20
(2) 3 4 7 II 18 29 47 76 123 199 322 521 843 1364 2207 3571 5778 9349 15127 24476
Figure 3-31
Sequences of numbers, as we have just seen, are not always

what we might expect upon a first inspection. They can be related
even if they appear completely unrelated. We have now taken one
of the most ubiquitous arrays of numbers in mathematics, the
Pascal triangle, to show how, via this array, number sequences that
are seemingly entirely unrelated can In fact have some
commonality. The Fibonacci numbers and many other familiar

sequences are embedded in the Pascal triangle. The modified
Pascal triangle relates the Lucas numbers to other well-known
sequences. It shows a very strong connection between the
Fibonacci numbers and the Lucas numbers. This is just the
beginning. We invite you to search for others in this amazing array
of numbers.
Chapter 4

the Golden Ratio
I t is time we ought to consider the geometric manifestation of the

Fibonacci numbers, the beauty of which has been appreciated in
many fields. We will show the relationship between the Fibonacci
numbers and the golden ratio, which has generally been observed
in the golden section, the golden rectangle, and the golden triangle,
as well as other related figures along the way. In any case, this ge-
ometry excursion is well worth the journey!
Fibonacci Ratios
In the previous chapter, we viewed the Fibonacci numbers as a se-

quence, rather than as individual numbers. We noticed how the
Fibonacci sequence related to other sequences of numbers that, at
the start, seemed completely unrelated to it. We will now inspect
the Fibonacci sequence of numbers by considering the relationship
of consecutive members of the sequence. Such a relationship is
best seen by taking the ratio of these consecutive numbers. We can
see that, as the numbers get larger, the ratio seems to approach a
specific number. But, as the numbers get larger, what is this num-
ber that the ratios are approaching?
107
F;_1_
- - --1
F; 1
F.1_2_
- - --2
F2 1
~ 3
-= - = 1.5
F3 2
F 5 -
2.= - =1.6
~ 3
~ 8
-= - = 1.6
F." 5
F7 13
-= - = 1.625
~ 8
F 21
_8 ==-==1.6153841
F; 13
~ = 34 = 1.619047
~ 21
F 55
== - == 1.61764705882352941
_10
~ 34
F 89 -
_11 == - == 1.618
F;o 55
F 144
-.!l.:= - : = 1.61797752808988764044943820224719101123595505
Fi.l 89
F 233 -
_13 == - == 1.61805
F;2 144
F 377
~ == - == 1.6180257510729613733905579399142 ... 2
F;3 233
1. A reminder: the bar over the digits after the decimal point indicates that these
digits repeat endlessly.
2. The period of the decimal expansion of 377 is
233
618025751072961373390557939914163090128755364806866952789699570815450643
776824034334763948497854077253218884120171673819742489270386266094420600
858369098712446351931330472103004291845493562231759656652360515021459227
4678111587982832, which then repeats endlessly. The period has length 232
[=233 -1].
THE FIBONACCI NUMBERS AND THE GOLDEN RATIO 109
We could have also considered the ratio of the numbers of the

Fibonacci sequence in the inverted order. Take a look at the two
sets of ratios in figure 4-1. Can you see anything curious about the
respective numbers they are approaching? It should begin to be-
come evident with the larger numbers
The Ratios of Consecutive Fibonacci Numbers3
Fn+1 Fn
Fn Fn+1
1 1
- = 1.000000000 - = 1.000000000
1 1
2 1
- = 2.000000000 - = 0.500000000
1 2
3 2
- = 1.500000000 - = 0.666666667
2 3
5 3
- = 1.666666667 - = 0.600000000
3 5
8 5
- = 1.600000000 - = 0.625000000
5 8
13 8
- = 1.625000000 -=0.615384615
8 13
21 13
- = 1.615384615 - = 0.619047619
13 21
34 21
- = 1.619047619 - =0.617647059
21 34
55 34
- = 1.617647059 - = 0.618181818
34 55
89 55
- = 1.618181818 - = 0.617977528
55 89
3. Rounded to 9 decimal places.

144 89
- = 1.617977528 - = 0.618055556
89 144
233 144
- = 1.618055556 -=0.618025751
144 233
377 233
- = 1.618025751 - = 0.618037135
233 377
610 377
- == 1.618037135 - = 0.618032787
377 610
987 610
- = 1.618032787 - =0.618034448
610 987
Figure 4-1
Both columns are each approaching a specific number (with

the right column lagging one ratio behind). The left column seems
to be approaching the value 1.61803 ... , while the right column
appears to be approaching 0.61803 .... As the numbers in the ra-
tio become very large, we might conclude that4 n 1
FF n
+ = F -
I
+ 1. But
n Fn
as the numbers get larger and larger, the one step lag that we ob-
served in figure 4-1 becomes negligible. We can thus say that, in
general, Fn+l ~ ~ + 1. This is called the golden ratio-when the
Fn Fn+l
Fibonacci numbers used reach their limitless "largeness."
The Golden Ratio
The limit of this ratio is perhaps one of the most famous numbers
in mathematics. Convention has it that the Greek letter ¢ (phi) is
used to represent this ratio. There is reason to believe that the letter
¢ was used because it is the first letter of the name of the cele-
brated Greek sculptor Phidias (490-430 BCE),5 who produced the
4. Even though there is a lag-by one-in the right column, when the numbers
get larger this lag becomes relatively insignificant.
5. In Greek: <PI~IAL.
THE FIBONACCI NUMBERS AND THE GOLDEN RATIO III
famous statue of Zeus in the Temple of Olympia and supervised

the construction of the Parthenon in Athens, Greece. His frequent
use of the golden ratio in this glorious building (see chapter 7) is
likely the reason for this attribution.
(In a lighthearted way, some like to relate the Fibonacci num-
bers to ¢ [phi], by saying that they are really called the Phi-bonacci
numbers or tfJ-bonacci numbers!)
A somewhat more accurate estimate of t/J would look like this:
¢ ~ 1.6180339887498948482045868343656, which is what the Fi-
bonacci number ratios in figure 4-1 seem to be getting ever closer to.
The unique characteristic that we can observe from this chart is that
t/J = ~ + 1, or ~ = t/J -1 . This way, we can calculate the value of the
reciprocal of ¢ from the value of it given above, by merely subtract-
ing 1, to get: ~= 0.6180339887498948482045868 343656 . . . . As a
t/J
matter of fact, this is the only number where such a relationship is
true. This should not be confused with the universally true mathe-
matical relationship that t/J. ~ = 1 . This, of course, is true for all num-
bers other than zero.
For those wanting a more precise value of ¢, we offer it here
to one thousand places:
~=1.618033988749894848204586834365638117720309179805762862
13544862270526046281890244970720720418939113748475408807
53868917521266338622235369317931800607667263544333890865
95939582905638322661319928290267880675208766892501711696
20703222104321626954862629631361443814975870122034080588
79544547492461856953648644492410443207713449470495658467
88509874339442212544877066478091588460749988712400765217
05751797883416625624940758906970400028121042762177111777
80531531714101170466659914669798731761356006708748071013
17952368942752194843530567830022878569978297783478458782
28911097625003026961561700250464338243776486102838312683
30372429267526311653392473167111211588186385133162038400
52221657912866752946549068113171599343235973494985090409
47621322298101726107059611645629909816290555208524790352
40602017279974717534277759277862561943208275051312181562
85512224809394712341451702237358057727861600868838295230
45926478780178899219902707769038953219681986151437803149
9741106926088674296226757560523172777520353613936
Since the relationship between if> and ~ is so spectacular, we re-

peat (for emphasis and to highlight our amazement) that the deci-
mal portion of these two values is the same.
~=.6180339887498948482045868343656381177203091798057628621
tfJ
35448622705260462818902449707207204189391137484754088075
38689175212663386222353693179318006076672635443338908659
59395829056383226613199282902678806752087668925017116962
07032221043216269548626296313614438149758701220340805887
95445474924618569536486444924104432077134494704956584678
85098743394422125448770664780915884607499887124007652170
57517978834166256249407589069704000281210427621771117778
05315317141011704666599146697987317613560067087480710131
79523689427521948435305678300228785699782977834784587822
89110976250030269615617002504643382437764861028383126833
03724292675263116533924731671112115881863851331620384005
22216579128667529465490681131715993432359734949850904094
76213222981017261070596116456299098162905552085247903524
06020172799747175342777592778625619432082750513121815628
55122248093947123414517022373580577278616008688382952304
59264787801788992199027077690389532196819861514378031499
741106926088674296226757560523172777520353613936
It is not periodic, 6 and it is also irrational. 7 The two numbers

differ by one. Therefore we get ~ = 1> - 1. To further solidify the
6. A periodic decimal, or repeating decimal, is a decimal that is finite or infinite

and has a finite block of digits that eventually repeats indefinitely.
7. An irrational number is a real number that cannot be expressed as an integer
or a quotient of integers.
connection between q> and the Fibonacci numbers, let's take a

short journey back to elementary algebra to solve the equation
I
-=</>-1.
l/J
Ifwe multiply both sides of the equation by </> we get:
1 = </>2 - </> .
Then </>2 - </> - 1 = 0 .
Applying the quadratic formula, we get: tjJ = 1±;S .

The positive value of
tjJ = 1+;S = 1.6180339887498948482045868343656 ....
Just to check our earlier presumption ( above) that the relation-
ship ~ = tjJ - 1 is really true, we can calculate
~ = -i-: = J5-1 :::: 0.61803398874989484820458683436564 ,
l/J v5+1 2
which appears to bear out our conjecture about the reciprocal rela-
tionship.
Therefore, not only does tjJ.~ = 1 (obviously!), but also
1
</>-l/J=1.
Bear in mind that </> and -~ are the roots of the equation
l/J
2
x - x-I = 0, a property we shall examine later.
Powers of the Golden Ratio
It is interesting to examine powers of </> . They will further link the

Fibonacci numbers to </>. To do so we first must find the value of </>2
in terms of </>.
4
We now use this relationship (l/J2 = l/J + 1) to inspect the succes-

sive powers of l/J by breaking them down to their component parts
in detail. It may at first appear more complicated than it really is.
You should try to follow each step (it's really not difficult-and
yet very rewarding!) and then extend it to further powers of l/J .
l/J = l/J ol/J2 = l/J(l/J + 1) = l/J2 + l/J = (l/J + I) + l/J = 2l/J + 1

1
2 2 2
l/J 4 = l/J el/J = (l/J + I) ( l/J + I) = l/J + 2l/J + 1 = (l/J + I) + 2l/J + 1 = 3l/J + 2
l/J5 = l/J3 ol/J2 = (2l/J + 1)( l/J + I) = 2l/J2 + 3l/J + 1 = 2 (l/J + I) + 3l/J + 1 = 5l/J + 3
l/J6 = l/J3 el/J3 = (2l/J + 1)(2l/J + I) = 4l/J2 + 4l/J + 1 = 4(l/J + I) + 4l/J + 1 = 8l/J + 5
l/J 7 = l/J 4 'l/JJ = (3l/J + 2) (2l/J + 1) = 6l/J2 + 7l/J + 2 = 6 (l/J + I) + 7l/> + 2 == 13l/> + 8
and so on.
By this point you should be able to see a pattern emerging. As

we take further powers of l/J, the end result of each power of l/J is
actually equal to a multiple of l/J plus a constant. Further inspection
shows that the coefficients of powers of l/J and the constants are all
Fibonacci numbers. Not only that, but they are also in their Fibo-
nacci sequence order. So you ought to be able to extend the list as
we havebegun in figure 4-2 to get the powers of l/J:
l/J6 = 8l/J + 5
l/J = ll/J + 0
l/J2= Il/J + 1 l/J7 = 13l/J + 8
8
l/J3= 2l/J + 1 l/J = 21l/J + 13
l/J4 = 3l/J + 2 l/J9 = 34l/J + 21
l/J5= 5l/J + 3 l/J = 55l/J + 34
10
Figure 4-2
Voila! Once again the Fibonacci number sequence appeared

where you may have least expected it. The Fibonacci numbers ap-
pear both as the coefficients of l/J, and as the constants as we take
powers of l/J. Furthermore, we can write all powers of l/J in a linear
form: l/Jn == a l/J + b, where a and b are special integers-the
Fibonacci numbers.
The Golden Rectangle
For centuries, artists and architects have identified what they be-
lieved to be the most perfectly shaped rectangle. This ideal rectan-
gle, often referred to as the "golden rectangle," has also proved to
be the most pleasing to the eye. The golden rectangle is one that
has the following ratio of its length and width: ; = w~l .
The desirability of this rectangle has been borne out by numer-
ous psychological experiments. For example, Gustav Fechner
(1801-1887), a German experimental psychologist, inspired by
Adolf Zeising's book Der goldene Schnitt8 began a serious inquiry
to see if the golden rectangle had a special psychological aesthetic
appeal. His findings were published in 1876. 9 Fechner made thou-
sands of measurements of commonly seen rectangles, such as
playing cards, writing pads, books, windows, etc. He found that
most had a ratio of length to width that was close to </J. He also
tested people's preferences and found most people preferred the
golden rectangle.
What Gustav Fechner actually did was to ask 228 men and 119
women which of the following rectangles is aesthetically the most
pleasing. Take a look at the rectangles in figure 4-3. Which rectan-
gle would you choose as the most pleasing to look at? Rectangle
1: 1 is too much like a square-considered by the general public as
not representative of a "rectangle." It is, after all, a square! On the
other hand, rectangle 2:5 (the other extreme) is uncomfortable to
look at since it requires the eye to scan it horizontally. Finally,
consider the rectangle 21 :34, which can be appreciated at a single
glance and is therefore more aesthetically pleasing. Fechner's
findings seem to bear this out.
8. AdolfZeising (1810-1876), a German philosopher, Neue Lehre von den Pro-

portionen des menschlichen Korpers (New theories about the proportions of the
human body) (Leipzig, Germany: R. Weigel, 1854.) The book Der goldene
Schnitt (The Golden Section) was published postumously (by the Leopoldinisch-
Carolinische Akademie: Halle, Germany, 1884).
9. Gustav Theodor Fechner, Zur experimentalen Asthetik (On Experimental
Aesthetics) (Leipzig, Germany: Breitkopf & Hartl, 1876).
D 1:1
D 2:3
D 5:6
CJ 21 : 34
D 4:5
CJ 13: 23
D 3:4
I I
1:2
D 20: 29
Figure 4-3
I 2:5
Fechner's rectangles.
Here are the results that Fechner reported:
Ratio of sides of Percent response Percent response

rectangle for best rectangle for worst rectangle
1: 1 = 1.00000 3.0 27.8
5:6 = .83333 .02 19.7
4:5 = .80000 2.0 9.4
3:4 = .75000 2.5 2.5
20:29 = .68966 7.7 1.2
2:3 = .66667 20.6 0.4
21:34 = .61765 35.0 0.0
13:23 = .56522 20.0 0.8
1:2 = .50000 7.5 2.5
2:5 = .40000 1.5 35.7
100.00 100.00
Figure 4-4
Fechner's experiment has been repeated with variations in

methodology many times and his results have been further sup-
ported. For example, in 1917 Edward Lee Thorndike (1874-1949),
an American psychologist and educator, carried out similar ex-
periments, with analogous results.
In general, the rectangle with the ratio of 21 :34 was most pre-
ferred. Do those numbers look familiar? Yes, once again the Fibo-
nacci numbers. The ratio :~ = 0.61764705882352941 approaches
the value of ~ , and gives us the so-called golden rectangle.

l/J
Consider a rectangle (figure 4-5) where the length, [, and the
width, w, are in the following proportion: ; = W~l ·
B. - - - - - - - - - -I - - - - - - - - , C
A D
Figure 4-5
By cross multiplying in this proportion we get w (w + 1) = [2,

or w 2 + wi == 12 , or w 2 + wi - 12 == O.
Ifwe let I == I, then w2 + w-I == o.
Using the quadratic formu1a,10 we get w = -1±.Js . Because we
2
are dealing with lengths, the negative value is of no interest here.
c.
Th erelore, w = -1+J5
2
J5-1 I d h ld .
- - 2 - - l/J' an t e go en ratIo agaIn
.
emerges.
10. The quadratic formula presented in the high school algebra course,
= - b±~b
2
x
- 4ac
lor th
. C:
, IS Id
e genera · equation
qua ratlc . ax 2 + bx + c = 0 .
2a
I
This is the same equation that gave us l/J above. So we now
know that the ratio of the rectangle's dimensions is
w = _,_ =~= ../5-1 or i= w+! = cjJ = ../5+1

, w+ I tfJ 2' w I 2
giving us a golden rectangle.

Let's see how this rectangle may be constructed using the tra-
ditional Euclidean tools: an unmarked straightedge and a pair of
compasses. (Another way would be to use a computer geometric-
construction program such as Geometer's Sketchpad.) With a
width of 1 unit, our objective is to get the length to be .J5+1 , so that
2
the ratio of the length to the width will be cjJ, which equals .J5+I.
2
Perhaps one of the simpler ways to construct this golden rec-

tangle is to begin with a square ABEF (see figure 4-6) with M, the
midpoint of AF. Then with radius ME and center M, draw a circle
to intersect ifF at D. The perpendicular at D intersects BE at C.
We now have ABCD (which turns out to be a golden rectangle).
8 E
C
,
J
I
r
r
I
J5
1 2 I
I
I
I
I
I
r
I
r
I
A
-- 0
M F
- J5 -}
2 2 2
Figure 4-6
Let us verify that figure 4-6 is, in fact, a golden rectangle.

Without loss of generality, we let ABEF be a unit square. There-
I
fore EF = AF = 1 and MF = 2 .
By applying the Pythagorean theorem to ~E, 11 we get

J5 J5+1
ME = 2 . Therefore, AD = 2 .
To verify that ABeD is a golden rectangle, we would want to

show that the ratio of the lengths and the widths follows the rela-
tionship mentioned earlier-namely, that CD = AD . We can do
AD CD+AD
this by substituting the above lengths into this proportion to get
15 + 1
1 2
15+1 - 1+--
15+1
2 2
which is a true equality!

Johannes Kepler (1571-1630), the revered astronomer and
mathematician, said that "geometry holds two great treasures: one
is the Pythagorean theorem and the other is the golden section. The
first we can compare to a bushel of gold and the second we can call
a priceless gem."
As we like to relate mathematical phenomena to each other, we
shall mention that Underwood Dudley12 drew a "cute" relationship
between the golden ratio and 1t. He showed that the following is
just a good approximation, but nothing more:
3.1415926535897932384 ... = 1t = 6 ¢2 = 3.1416407864998738178 ...
5
11. Applying the Pythagorean theorem gives us:

(ME)2 = (MF)2 + (EF)2
(MEf=(~r +12
( ME)2=~
4
(ME) =
'1fi4 = J52
12. Mathematical Cranks (Washington, DC: Mathematical Association of
America, 1992).
Furthermore, as we continue to connect mathematical values,

consider one of the most famous relationships in mathematics. It is
attributed to Leonhard Euler (1707-1783), one of the most prolific
mathematicians in history. The beauty of this relationship is that
the simple equation contains each of the most significant values in
mathematics.
It is: e 1ti + 1 = 0, where e is the base of natural logarithms
(Euler's number), 1t is the ratio of the circumference of a circle to
its diameter (Ludolph's number), i is the imaginary unit of the
complex numbers (square root of - 1; r-I), 1 is the unit of the
natural numbers (all numbers> 0 are produced by 1 and adding),
and 0 is the neutral element of the addition. Notice that l/J is miss-
ing. Well, we can fix that!
We know that cp = .[5+1 and I = - e1ti (from Euler's equation),
2
so we can incorporate l/J into this equation by replacing the 1 in the
. ~ eH ;
first equation with (- e1t1 ) to get: l/J = - . So now we have an
2
exact way to relate l/J to 1t, e, and i.
Constructing the Golden Section
There are many ways to construct the golden ratio with Euclidean
tools (i.e., an unmarked straightedge and a pair of compasses) be-
sides the method shown in figure 4-6. With each method we have
an opportunity to see how many lovely geometric relationships are
brought into play in our quest for the golden ratio. We shall show a
few of them here. The first, shown in figure 4-7, is attributed to
Heron of Alexandria (10-70 CE).
rl2
E
A B
Figure 4-7
You may be able to see from figure 4-7 how the construction is to
be done to get AB divided into the golden ratio. You begin with
- -
AB and construct AB ..1 BC (..1 is the symbol used to denote per-
pendicularity). Then draw the circle with center at C and radius
length CB to intersect the hypotenuse of the right triangle ABC at
D. You complete the construction by drawing the circle with center
A and radius AD to intersect the other leg at E, which is the point
on AB that divides the segment into the golden section. By using
the Pythagorean theorem on triangle ABC, we find that the hypote-
nuse,
-AC = -
r15. The result
.
IS that
AE
-
x
= -
15+1
= -- = l/>
2 BE r-x 2
~ 1.618033988. This is a start for justifying our conclusion. (Fur-
ther details and justification can be found in appendix B.)
Another Construction of the Golden Section
Another construction of the golden section along AB, using

Euclidean tools, can be seen in figure 4-8. Here the two circles,
with the line segment lengths indicated, are shown to be tangent at
- -
D and AB ..1 A C .
To do the construction, begin with the larger circle (center at C

and radius length AB) and construct a right triangle ABC where B
is on the circle. Then extend CA to meet the circle at D. Next draw
the circle with center A and radius length AD. We then can con-
clude that point E divides the line segment AB into the golden ra-
tio. That is, AE=~= .J5+ I = f/J :::; 1.618033988. (The justification
BE a-x 2
for this construction can be found in appendix B.)
a
-I
2
A E B
o
Figure 4-8
Yet Another Construction of the Golden Section
The construction shown in figure 4-9 will have two segments that
reflect the golden ratio:
f/J =.J5 + 1 and ~ = J5 - 1

2 t/J 2
They are BQ and BP .

'1*
t-
2
A p 1 B Q
2 •
Figure 4-9
The construction is rather simple. You start off with a right

triangle ABC, where AB = 2 . BC, thus making the hypotenuse
AC = .J5. Then all that remains is to construct the angle bisec-
tors of LACB and the exterior angle at C. (See appendix B.)
A Surprising Construction of the Golden Section
We can construct the golden ratio along a line and then, of

course, use the segments to construct the golden rectangle. The
construction is simple, as is the justification. Begin with an
equilateral triangle ABC inscribed in a circle. Through the mid-
point of two of its sides, construct a line that will intersect the
circle. (See figure 4-10.) For convenience, we will let the sides
of the equilateral triangle be of length 2. It is easy to establish 13
that the line joining the midpoints of the two sides of the trian-
gle is half the third side and, therefore, is of length 1.
13. To do this, draw a line segment through E parallel to AB and intersecting

Be at F. The figure formed, BDEF is a rhombus with each side of length 1.
Mt-----#-----~--__ftN
Figure 4-10
Using the relationship that the products of the segments of two

intersecting chords of a circle are equal,14 we get:
AE· EC=ME· EN
1 . 1 = (x + 1) x
r+x-l=O
.[; - 1 ]
X = , which is -.
2 t/J
So, from the previous constructions of the golden section, we

can say that the Fibonacci numbers appear even when there is no
golden rectangle.
Golden or Fibonacci Spirals
Let us continue our discussion with golden rectangle ABeD, but

now in a very curious way. We established in figure 4-6 that when
a square is constructed internally (as shown in figure 4-11), if
AF = 1 and AD = C/>, then FD = C/> - 1 = ~. We now can establish
14. When two chords of a circle intersect, the product of the segments of one
chord equals the product of the segments of the other chord.
that rectangle CDFE has dimensions FD = ~ and CD = 1. If we

inspect the ratio of length to width of rectangle CD FE , we get
:;; = -+- = l/J , and it is, therefore, also a golden rectangle.
l/J
B ...-_ _ _ _ _ _ _---.E...--_ _ _ _.... c
1
1 t/J
A F D
~----------~y~------------~
Figure 4-11
Let's continue this process of constructing an internal square in

the newly formed golden rectangle. In golden rectangle CD FE ,
square DFGH is constructed (figure 4-12). We find that
CH = 1- ~ = ;2 , so the ratio of the length to width of rectangle
CHGEis
(having multiplied both numerator and denominator by l/J2). This

thereby establishes rectangle CHGE also as a golden rectangle.
B c
N M
L
-
¢4
1
~
G
Sr T
R K
1
1 t/J
-
1 t/J
A F o
~----------------v~--------------~
Figure 4-12
Continuing this scheme, we construct square CHKJ in golden

rectangle CHGE. We find 15
1
EJ:=!... __
l := lfJ-l :=!L:=_l
lfJ lfJ2 lfJ2 lfJ2 lfJ3
We now inspect the ratio of the dimensions of rectangle EJKG.

This time the length to width ratio is
Once again, we have a new golden rectangle; this time rectangle

EJKG.
By continuing this process, we get golden rectangle GKML,
golden rectangle NMKR, golden rectangle MNST, and so on.
15. We showed earlier that t/J - ~ = 1, therefore, t/J - 1 = ~.

center E, radius EB
center G, radius GF
center K, radius KH
center M, radius MJ
center N, radius NL
center S, radius SR
The result is an approximation of a logarithmic spiral (figure

4-13).
B E
r----------------r--~..,.__~--___.
J C
G 1--.-...........""'--"------1 H
R K
A F o
Figure 4-13
The symmetric parts of this complex-looking figure are the

squares. Suppose we locate the center of each of these squares. We
can draw arcs through each of these points and then see that the
centers of these squares lie in another approximation of a logarith-
mic spiral (figure 4-14).
The spiral in figure 4-13 seems to converge (i.e., end) at a point
in rectangle ABC D. This point is at the intersection P, of
-AC and -ED (figure 4-15).
B~______~____________~E____~J ______~C
GI------'----'------+----t H
A F o
Figure 4-14
Consider once again golden rectangle ABCD (figure 4-15).

Earlier we established that square ABEF determined another
golden rectangle CEFD.
B~____________~E~______~C
G~-;----fH
A F o
Figure 4-15
Since all golden rectangles have the same shape, rectangle ABCD
is similar to rectangle CEFD. This implies that MCD is similar to
i1CDA. Therefore LCED is congruent to LDCA. And LDCA is com-
plementary to LECA. Therefore LCED is complementary to LECA.
- -
Thus LEPC must be a right angle, or AC ..1 ED.
If the width of one rectangle is the length of the other and the
rectangles are similar, then the rectangles are said to be reciprocal
rectangles. In this case, the ratio of similitude 16 is l/J.
In figure 4-15 we see that rectangle ABCD and rectangle CEFD
are reciprocal rectangles. Furthermore, we see that reciprocal rec-
tangles have corresponding diagonals that are perpendicular.
In the same way as before, we can prove that rectangles CEFD
-
and CEGH are reciprocal rectangles. Their diagonals ED and CG
-
are perpendicular at P. This may be extended to each pair of con-
secutive golden rectangles shown in figure 4-16. Clearly P ought
to be the limiting point of the spiral.
B E J
r-------------........--~_"....__-_____"
C
~---~-----I H
A F o
Figure 4-16
We can use this relationship of the diagonals to construct con-

secutive golden rectangles. We could simply begin with golden
rectangle ABCD and construct a perpendicular from D to AC and
from its intersection E with BC construct a perpendicular to AD
to complete the second golden rectangle. This process can be re-
peated indefinitely.
Let's take another look at the spiral we generated by drawing
quarter circles--one that approximates the golden spiral (see figure 4-
17). We have a golden rectangle ABCD, with sides of length a and b
16. The ratio of similitude is the ratio of the corresponding sides of the two
similar figures, in this case, rectangles.
(a > b), where it follows that a = l/J. band b = l/J -1 · a. We shall then
construct the spiral with the quarter circles as before. This will allow
us to get the length of the spiral that approximates the golden spiral.
D c
e.-==
E:i - - .-- - --- - ----+-:3I....
A a
Figure 4-17
The actual golden spiral does not evolve from these quarter cir-
cles. This diagram merely gives us a good and easily understood
approximation. The table in figure 4-18 provides a progressive cal-
culation of the length of the spiral.
Length Width (quarter circle radius)

J5-1 IJ5-1
a =ao b = bo = l/J-I· a = 2
·a = 2
·a
3-J5 IJ5-3
a l = bo b l = ao - b o =l/J-I' al = 2
·a = 2
·a
2J5-4 2J5-4
a 2 =b l b2 = a l - bl = l/J -I . a 2 = 2
·a -
2
·a
7-3J5 3J5-7
a3 = b2 b3 = a2 - b2 = l/J -1 • a 3 = 2
·a -- 2
·a
SJ5-11 SJ5-11
a4 = b 3 b4 = a 3 - b3 = l/J -I . a 4 = 2 ·a = 2
·a
lS-SJ5 SJ5-1S
a 5 = b4 b 5 =a4 -b4 -- l/J -I • a 5 = 2
·a = 2
·a
13J5-29 13J5-29
a 6= b 5 b6 = a 5 - b 5 = l/J -I . a6 = 2
·a -
2
·a
47-21J5 21J5-47
a7 = b6 b 7 = a6 - h6 = l/J -I . a 7 = 2 ·a - 2
·a
•••
Figure 4-18
Amazingly, as you inspect the results, you will see the Fibo-
nacci numbers, Fn (1, 1,2,3,5,8, 13,21, ... ), as the product with
.J5, as well as the constants highlighting the Lucas numbers, Ln
(1, 3, 4, 7, 11, 18, 29, 47, ... ). This could justify calling this spiral
the Fibonacci-Lucas spiral.
The real golden spiral, also called a logarithmic spiral, looks
something like that in figure 4-19.
The real golden (logarithmic) spiral cuts the sides of the
squares at very small angles. The apparent spiral (which consists of
quarter circles) touches it. So the sides of the golden rectangles
aren't tangents to this golden spiral (as in the case of the approxi-
mation); they are each cut twice.
D~ _ _ _ _~_ _~_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _~C
A B
Figure 4-19
It is sometimes called an "equiangular spiral," since the radius

keeps a constant angle with the curve. It was named such by Rene
Descartes (1596-1650), the French mathematician, who is respon-
sible for the field of analytic geometry that is done on a "Cartesian
plane"-named for its founder. Descartes mentioned this spiral in
1638 correspondences with another French mathematician Marin
Mersenne (1588-1648), who is famous for his work with prime
numbers. The Swiss mathematician Jacob Bernoulli (1655-1705)
referred to this spiral as the logarithmic spiral. He was so en-
chanted with it and its properties that he requested it as his epitaph
with the words: "Eadem mutata resurgo" (It changed me, and yet
remain unchanged). 17
In nature the nautilus shell exhibits such a golden spiral (figure 4-20).
Figure 4-20
Figure 4-21 shows a curve that crosses the x-axis at the

Fibonacci numbers.
The spiral part crosses at 1, 2, 5, 13, and so on, on the positive
axis, and 0, 1,3,8, and so on, on the negative axis.
The oscillatory part crosses at 0, 1, 1,2,3,5,8, 13, etc. on the
positive axis.
,/ ,,
,
I ,
\
\
--
'/
--
\
•
Figure 4-21
17. The sculptor who then created the epitaph chiseled not a logarithmic spiral.
but rather an Archimedean one!
THE FIBONACCI NUMBERS AND THE GOLDEN RA no 133
The curve is strangely reminiscent of the shells of nautilus and

snails (figure 4-22). This is not surprising, since the curve tends to
a logarithmic spiral as it expands.
Figure 4·22
A Surprising Sighting of the Fibonacci Numbers
We typically call the area between two concentric circles a ring. In

figure 4-23, you will notice that as the area of the ring gets smaller,
the area of the ellipse that is tangent to each of the circles gets
larger. At some point the ellipse will be equal in area to the ring.
Strangely enough, the ellipse will be equal in area to the ring lS
when the ratio of the radii of the two circles is 0.618 ... or ~.
4>
18. The justification for this claim is as follows: The area of a circle of radius r
is,,? The area of an ellipse with "radii" a and b (as shown above) is "ab (Note
how when a ~ b in the ellipse, it becomes a circle and the two formula are the
same.) So the outer circle has radius a, the inner circle radius b, and the area of
the ring between them is therefore: " (b' - a'). This is equal to the area of the
ellipse when" (b' - a') ~ "ab, and b' - a 2 - ab ~ O.
If we let the ratio oj the two circles' radii ~ alb, be R, say, then dividing the
equation by a 2 , we have R' - R - I ~ 0, which means R is 4>. The equation of an
ellipse is (x/b)' + (y/a)' ~ I. When a ~ b, we have the equation of a circle of ra-
dius a(~b): (x/a)' + (y/a)2 ~ I.
Again, when you least expect it, if! or ~¢ , or for that matter, the
Fibonacci numbers, come up to greet you.
Figure 4-23
Let's look at the right triangle and consider when the tangent of
an angle would be equal to the cosine of the same angle- that is,
when is tan LA = cos LA ?
Consider the right triangle ABC (figure 4-24) with AC = I and
BC = Q . Then by the Pythagorean theorem, we get:
AB = Ja 2 +1 .
e-__________________________ ~~A
C
I
Figure 4-24
a
tan LA =-
1
1
cos LA = ~ 2
a +1
We want to have tan LA = cos LA :
a I
1- .Ja 2
+1
Solving this equation for a :
a~a2 + 1 = 1
a (a + 1) = 1
2 2
4
+ a 2 -1 = 0
a
We shall let p = a 2, so that we then get the equation p2 + P - 1 = 0,
at which point you should recognize this equation.
Then p =
15 - 1 = -,1 or a 2 1
= -.
l/J l/J 2
Let's use these values for the sides of right triangle ABC (see fig-
ure 4-25).
c ~----------------------------~A
1
Figure 4-25
So you can see that when a triangle has sides of lengths:
1, .Ja 2
+! = ~i+! =.,fo, and a = ~, the tangent of an acute angle of
the right triangle will equal the cosine. Again, the golden ratio-or, if
you wish, the Fibonacci numbers-appears when you least expect it.
Another Emergence of the Fibonacci Numbers

in Geometry
An engaging problem that was posed by 1. A. H. Hunter 19 asks us

to determine the points on any rectangle ABeD that will leave
three triangles of equal areas when the center triangle (triangle 4 in
figure 4-26) is removed from it.
A p q B
2
r
r+s
D p+q
c
Figure 4-26
You can guess by now that this will in some way result in the
Fibonacci numbers-another sighting in geometry of this wonder-
ful sequence. We begin (using the markings in figure 4-26) by set-
ting the areas of the three triangles equal to one another.
p(r + s) qr Area~3 = s(p + q)

Area~l = , Area~2 =-
2 2 ' 2
Since Area~l = Area~2 = Area~3,
p(r + s) qr p(r + s) s(p + q)
- and - , or
2 2 ' 2 2
p(r+s)=qr, and p( r + s) = s( p + q)
qr
p= and pr=sq
r+s
19. "Triangle Inscribed in a Rectangle," Fibonacci Quarterly 1 (1963): 66.

From the second equation we can set up the following propor-

tion: p = q , which tells us that the two sides of the rectangle were
s r
divided proportionally. But what is that proportion? (Can you

guess by now?)
Let us replace p from the second equation, above, with the
value of p shown in the first equation:
qr )r = sq
( r+s
2
r =s(r+s)
If we divide both sides of the equation by S2 , we will get:
r2 s(r+s)
2 2
S S
r2 sr s2
-2 = -2 + -
2
S S S
r2 r
-=-+1
S2 s
Putting this in a more recognizable form (remember, ¢ and!

t/J
2
are the roots of the equation x - x - 1 = 0):
We can see (because r > S, q > p) that the solution of this

. . r _ 1+15 _ A\ ( _ q)
equatIon IS - - - - - 'Y - - •
s 2 p
Therefore, the points on the sides of the rectangle that will de-
termine the triangles that must be removed to leave the three trian-
gle of equal areas must be placed so that they partition the sides in
the golden ratio ( ¢ ). Again, the Fibonacci numbers are involved!
The Diagonal of the Golden Rectangle
We have done quite a bit with the golden rectangle, yet there never
seems to be an end to what you can do. For example, the golden
rectangle-the relationship of whose sides are in the golden ra-
tio-allows us a neat way to find the point along the diagonal that
cuts it into a ratio related to the golden ratio. It is only because of
the unique properties of this special rectangle that we can do this
so easily.
Consider the golden rectangle ABCD, whose sides AB == a and
Be = b, so that : = 1>. As shown in figure 4-27, two semicircles are
drawn on the sides AB and BC to intersect at S. If we now draw
segments SA, SB, and SC , we find that LASB and LBSC are right
angles (since they are each inscribed in a semicircle). Therefore,
AC is a straight line, namely, the diagonal. We can now show,
rather elegantly, that point S divides the diagonal in the golden ratio.
D c
Figure 4-27
From the mean proportional ratios (obtained from similar tri-

angles MBC, MSB, and MSC), we get the following:
We use the segment length markings in figure 4-27, for each of
the right triangles:
~SB:
2
a == e(e+ f)
2
M3SC: b == f(e+ f)
2
a e
Therefore, 2
-
b f
2
a a
But since - = ¢, then -2 = ¢2 = ¢+ 1
b b
1
[Recall: ¢ = - + 1 and the powers of ¢.]
l/J
Thus the point S divides the diagonal of the golden rectangle
in a ratio involving the golden section (1/l,2 or I/l;') and, in turn, the
Fibonacci numbers.
Another Curiosity That Will Generate

the Golden Ratio
Consider the semicircle in figure 4-28 with the three congruent cir-
cles inscribed so that the tangency points are as shown.
Figure 4-28
We seek to find the ratio of the radii of the large semicircle to

one of the smaller circles. In figure 4-29, AB = 2R and AM = R .
Each of the congruent small circles has a radius r. Consider right
triangle CKM with legs rand 2r, where the hypotenuse then has
length r.J5 . So we now have MK = r.J5 and PK = r, and therefore,
MP = r(.J5 + 1) = R.
Put another way, R =

r
J5
+ 1 = 2l/J .
Out of this seemingly unrelated situation with three congruent
circles inscribed in a semicircle, we find the ratio of their radii is
related to the golden ratio.
Figure 4-29
Fibonacci Nuntbers and a Curious Dilemma
This discussion of the role of the Fibonacci numbers in geometry

can also take on other entertaining aspects. Toward this end, we
will examine a rather curious problem. It was made popular by the
English mathematician Charles Lutwidge Dodgson (1832-1898),
who, under the pen name of Lewis Carroll, wrote The Adventures
of Alice in Wonderland. 20 He posed the following problem: the
square on the left side of figure 4-30 has an area of 64 square units
and is partitioned as shown.
20. See Stuart Dodgson Collingwood, ed., Diversions and Digressions of Lewis
Carrol (New York: Dover, 1961), pp. 316-1 7.
5 3 '8 5
!
3 3
••••• 1 t··· • _ ... ~ • • • • • • • • • • • • _ . . . .0 • ••••••• •• _!.. ..... . ...._._ ..~.............""..._... _. n'
5 5 5
i
•••••• 1-. "

!
• •_ •••• '(0'0' .0
!
+ .0
3
• •0
: 5
5
i 5 3 :
Figure 4-30
These four parts are then reassembled to form the rectangle at

the right of figure 4-30. This rectangle has an area of 13 . 5 = 65
square units. Where did this additional square unit come from?
Think about it before reading further.
All right, we'll relieve you of the suspense. The "error" lies in
the assumption that the figures will all line up along the drawn di-
agonal. This, it turns out, is not so. In fact, a "narrow" parallelo-
gram is embedded here, and it has an area of one square unit (see
figure 4-31).
2 :
l 5
I
1
: 1{ A ' i
______ 2_______ .. _______ ~____ ... ______ ;
3 3
!
5
3
i.
+I
51
Figure 4-31
We can discover where the error lies by taking the tangent

function of a and f3 so that we can discover the measure of these
angles. Remember, they ought to be equal if they lie on the diagona1. 21
21. Alternate-interior angles of parallel lines are congruent.

Since tan a =~ , then a ~ 20.6°.

8
Since tan J3 =.:. , then (3 ~ 21.8°.

5
The difference, (3 - a, is merely 1.2°, yet enough to show that
they are not on the diagonal.
You will notice that the segments above were 2,3,5,8, and 13
-all Fibonacci numbers. Moreover, we already discovered22 that
Fn_1Fn+ I = Fn 2 + (-1 ) n, where n > 1. The rectangle has dimensions
5 and 13, and the square has a side length 8. These are the fifty,
sixth, and senenth Fibonacci numbers: ~,~, and F, .
This relationship tells us that
5 ·3= 82 + 1
65 = 64 + 1
This puzzle can then be done with any three consecutive

Fibonacci numbers as long as the middle number is an even-
numbered member of the Fibonacci sequence (i.e., in an even po-
sition). If we use larger Fibonacci numbers, the parallelogram will
be even less noticeable. But if we use smaller Fibonacci numbers,
then our eyes cannot be deceived as in figure 4-32:
3' i 3
-- ·-----r-·- .; . .. 'j . _.. _.. ..
3 3 3
t . ··2
5 3
3 j
! -
!
3 ' I i
i ---
• I
,- , -- "-- - :H' --- ~ , .. -- 'j '" '-':' , - I' - "j- ---

Figure 4-32
22. Chapter 1, page 55, item 11.

Here is the general fonn of the rectangle (figure 4-33).
To do this properly, without the missing area, the only parti-

tioning-amazingly enough-is with the golden ratio, l/J as is seen
in figure 4-34.
</}-1
1
1
</}-1
1 1
(/>-1
</}-l
1
cp (/>2 = (/>+ 1
Figure 4-34
The areas of the rectangle and the square are equal here (figure
4-34), as we shall show as follows:
The area of the square
= ¢. ¢ = </i = ¢ + 1 = J; + 3 = 2.6180339887 ...

2
The area of the rectangle
= (l/J + 1) . 1 = l/J + 1 =
J; +3 = 2.6180339887 ...
2
Thus the areas of the square and the rectangle under this parti-
tion are equal.
The Golden Triangle
Now that we have thoroughly investigated the famous golden rec-

tangle, we are ready to consider the golden ratio as it pertains to a
triangle-the golden triangle. As you would expect, much like the
golden rectangle, which has the Fibonacci numbers embedded
within it, so, too, does this golden triangle exhibit the Fibonacci
numbers, and consequently, the golden ratio. Let's consider a tri-
angle that contains this golden ratio. We will begin with an attempt
to place an isosceles triangle into another similar isosceles triangle
in a somewhat analogous way to the way we embedded our similar
golden rectangles earlier. To do this, we may get a configuration as
in figure 4-35. The sum of the measures of the angles of triangle
ABC is a + a + a + 2a = 5a = 180° , and therefore a = 36° .
A
A
~o x
\
x
B x c
Aside from placing the two similar triangles the way we did
above, we could also have simply begun with an isosceles triangle
whose vertex angle measures 36°. Construct the bisector
BD of LABC (figure 4-36).
Since isosceles triangles with congruent vertex angles are
similar, then MBC is similar to MCD. We shall now let AD = x
and A B = I. However since ~A DB and ~D B C are isosceles,

BC=BD =AD =x.
From the similarity above, we get a (hopefully by now) famil-
. . I x
Iar equatIon: -x = -I-x .
As before, this gives us: x 2 + x-I =0 and x = ~-1 . (The
negative root cannot be used for the length of AD.)
Remember that
15-1 I
2 = q, ·
. side 1
In MBC the ratIo of - = - = l/> .
base x
We therefore call this a golden triangle. One easy way to con-
struct a golden triangle is to first construct the golden section (done
earlier in this chapter; for example, in figure 4-9: A B = 2 and
Be = 1, angle bisectors provide the points P and Q with BP = ~
and BQ = l/». We draw a circle around 0 with radius 1. On this
circle we choose a point A and draw around this center a second
circle with radius x = ~ . The intersection point(s) of the two cir-
l/J
cles, as shown in figure 4-37, helps to determine a golden triangle
(compare also with figure 4-36).
Figure 4-37
By taking consecutive angle bisectors BD, CE, DF,

- -
EG, and FH of a base angle of each newly fonned 36°, 72°, 72°
triangle, we get a series of golden triangles (see figure 4-38). These
golden triangles (36°, 72°, 72°) are: MBC, MCD, 8.CDE, MJEF,
MFG, and M'GH. Obviously, had space permitted, we could have
continued to draw angle bisectors and thereby generate more
golden triangles. Our inspection of the golden triangle will parallel
that of the golden rectangle. We proceed in pursuit of the
Fibonacci numbers.
B c
Figure 4-38
Let us begin by having HG = 1 (figure 4-38). Since the ratio of

side of a golden triangle is l/J, we find that for golden M'GH:
b ase
GF l/J GF l/J
-=-, or - = - and GF= l/J.
HG 1 1 1'
Similarly for golden MFG: FE = if> , but GF = l/J, so FE = l/J 2.

GF 1
ED l/> 2 3
In golden WEF: - = - , but FE = l/J , therefore ED = l/J .
FE 1
.
AgaIn, for 8.CDE: -
DC
= -l/J , but ED = l/J 3 ,therefore DC = l/J 4 .
ED 1
For MCD: CB = q, , but DC = ¢ \ therefore CB = ¢ 5.

DC 1
Finally for MBC: BA = q, , but CB = ¢ 5, therefore BA = ¢ 6.

CB 1
This can be summarized by using our knowledge of powers of

¢ (developed earlier) as follows (this time we point out the
Fibonacci numbers):
0
HG = l/J = Ol/J + 1 = Fol/J + F -I
GF = l/J' = ll/J + 0 = Fl/J

I
+ F0
FE = l/J2 = ll/J + 1 = F l/J + F
2 1
ED = l/J3 = 2l/J + 1 = F l/J + F3 2
4
DC = l/J = 3l/J + 2 = F4 l/J + F3
CB = l/J5 = 5l/J + 3 = F5l/J + F4
BA = l/J6 = 8l/J + 5 = F l/J + F 6 5
As we did with the golden rectangle, we can generate an ap-

proximation of a logarithmic spiral by drawing arcs to join the vertex
angle vertices of consecutive golden
A
triangles (see figure 4-39).
That is, we draw circular arcs as
follows:
AB (circle center at D)
Be (circle center at E)
CD (circle center at F)
DE (circle center at G)
EF (circle center at H)
FG (circle center at J)
Figure 4-39
There are many other truly fascinating relationships emanating

from the golden ratio. After you have been exposed to the golden
triangle, the next logical place to tum for more applications is the
regular pentagon 23 and the regular pentagram (the five-pointed
star), since these are essentially composed of many golden trian-
gles. You will then see that the golden ratio abounds through-
out-as do the Fibonacci numbers.
The Golden Angle
Let us first take a side step and look at the golden angle, one that
divides a complete circle, 360 0 , into the golden ratio. Notice in
figure 4-40 the angles 'V and <p are in the ratio that approaches the
golden ratio:
'" = 3600 _ 360° = 137.5077640 ...0 _ 137.50.

l/J
q> = 360° = 222.4922359 ... ° "" 222.5°.
l/J
q> = 360° "" 222.50

l/J
-_.- A
'" = 3600 _ 360° _ 137.50
I
I
l/J
Figure 4-40
23. A regular pentagon is one whose sides have the same length and whose an-
gles are all the same measure.
= 360 ~ larger angle ~

0
The ratio of circle 1.618 and 1.618.
larger angle 222.5 0 , smaller angle
In each case, the golden ratio is approximated, as it can be, by
the quotient of successive Fibonacci numbers.
The Pentagon and the Pentagram
We now come to the beautiful geometric shape that sums up much

of the golden ratio in one configuration.
--~
The golden triangle is embedded many times in the regular

pentagram, which just happened to be the symbol of the Pythago-
reans. According to Pythagoras, all geometric shapes could be de-
scribed in terms of integers. So it would have come as a great dis-
appointment to him when one of his followers, Hippasus of
Metapontum (ca. 450 BeE), showed that the ratio of a diagonal of
the regular pentagon (i.e., a side of the regular pentagram) to the
side length of the pentagon could not be expressed as a fraction of
integers, in other words, this ratio is not rational! This, then, car-
ried over to their symbol: the pentagram. This secret society was a
bit troubled by this- which today can be seen as the very begin-
ning of our concept of irrational numbers. That is, these are num-
bers that cannot be repressed as a ratio of two whole numbers;
hence the name irrational. In the regular pentagon, the ratio of the
side to the diagonal is irrational. But which irrational number did
he find? Yes, you guessed it! It was the golden ratio, 1/1.
To show that this length relationship is actually irrational, we

use the relationship that in a regular pentagon every diagonal is
parallel to the sides it does not intersect. In figure 4-43 the trian-
gles AED and BTC have therefore parallel sides, so they are similar
to each other.
Figure 4-43
Therefore AD=BC. But BT == BD - TD == BD -AE. In the

AE BT
regular pentagon, therefore, the following ratio holds:
diagonal: side == side: (diagonal - side).
Symbolically, we can write this as
d a d 1
or
a d-a d
a
--1
a
(with d as the length of the diagonal and a the length of the side).
If we now let x == d we get the equation x == _1_, which then

a x-I
can be converted to the quadratic equation x 2 - x-I == 0, of which
d is a positive root and just happens to be the irrational number
a
¢ = .[5+1 . (Remember:
2
.Js is irrational!)
This is what we claimed at the outset: the ratio of the diagonal

to the side of a regular pentagon is irrational. As the irrational
number 1t = 3.1415926535897932384 ... is connected inseparably
with the circle, so, too, the irrational number l/J = 1.6180339887
498948482 ... is connected inseparably with the regular pentagon!
The regular pentagon ABCDE (figure 4-43) is a fascinating
figure with lots of useful properties. Here are some for you to ap-
preciate and perhaps ponder over. You might look for other such
properties.
For figure 4-43 the regular pentagon ABCE has the following
properties:
(a) The size of every interior angle is 108°:
LEAB =LABC =LBCD =LCDE =LDEA = 108°
(b) The size of the angles of a golden rectangle are:
LBEA=LCAB=LDBC=LECD=LADE=3~
LPEB =LQAC =LRBD =LSCE =LTDA= 36°
LCDA =LDEB =LEAC =LABD =LBCE = 72°
(c) The following triangles are all isosceles:
WAC, MBD, MCE, MDA, and ~CEA

MEA, ~CAB, WBC, MCD, and MDE
MEB, ~QAC, MBD, MCE, and ~TDA
(d) The triangles 8DAC and 8QCD are similar (as are many
others figure 4-43).
( e) All diagonals of the pentagon have the same length.
(f) Every side of the pentagon is parallel to the diagonal "fac-

ing" it.
(g) As an example, AD : DC = CQ : QD.
(h) The intersection point of two diagonals partitions both di-

agonals in the golden section.
(i) PQRST is a regular pentagon.
Which of the triangles are golden triangles in the regular pen-

tagon?
At this point we can bring in our Fibonacci numbers. We already

established that the various line segments of a regular pentagon
cannot be measured in integers. We say the segment relationships
are incommensurable. Yet we know that the ratio of adjacent
Fibonacci numbers approaches the golden ratio. With the obvious
loss of accuracy, we can display the regular pentagon and penta-
gram in terms of the Fibonacci numbers (see figure 4-44).
Figure 4-44
These conditions can be approximately interpreted as follows:

If the side and the diagonal of the little pentagon were ap-
proximately of lengths 34 and 55 mm, the side length of the larger
pentagon (89 mm) might have been obtained from the sum of these
values (34 + 55). Similarly, we can generate the other lengths as

shown in figure 4-44. There again, we have the Fibonacci num-
bers!
Remember, the irrationality Hippasus of Metapontum encoun-
tered, which made the Pythagoreans a bit disturbed, was that, for
example, 144 and 55 have a value of about 1.618. This is only an
89 34
approximation of the actual irrational value. Had the line segments
been exactly 34,55, 89, and 144 (mm), then all the fame Hippasus
of Metapontum garnered would have been ignored. So he remains
the "originator" to have stumbled on irrational numbers. Yet all
this fame did not come without his attackers. A remark by the
Greek philosopher Plato (427-348/347 BeE) shows how terribly
the discovery of Hippasus moved the Greek:
I thought such an ignorance befits not people but rather a

herd's pigs, and I am not only ashamed of me but for all
Greeks. 24
The pentagram (five-comer star) inscribed in a regular penta-

gon abounds with the golden ratio l/>, since it is composed of lots
of golden triangles. In figure 4-45 the regular pentagon has side
length 1.
A
E~--------.-----~--------~B
o c
Figure 4-45
24. Plato, Laws for an Ideal State.

Since the golden triangle is everywhere to be seen in the figure,

we have for each golden triangle: bside = if!, therefore AD = if! . With
ase DC
DC = I , we get AD = if! . In the golden MEH we get
side = if! A£
base AH
Therefore AH = ~ .
q,
. I I 25
Smce EH=DC= I, then FH= EH - EF = 1- - = , .
q, q,
The golden ratio is also involved with various area compari-
sons in figure 4-45.
The ratio of the area of the larger pentagon ABCDE to the area
4
of the smaller pentagon is ~.
The ratio of the area of the larger pentagon ABCDE to the area
3
of the pentagram is q,2 .
You can see the continued pattern in pentagrams and pentagons

(figure 4-46) harboring the golden ratio and, consequently, the
Fibonacci numbers.
E c
Figure 4-46
25. This comes from the now-familiar equation: q,' - ¢ - I = 0, and then we
I I
divide both sides by 1// to get: 1- - = - . See also figure 4-12.
¢ ¢'
Constructing a Regular Pentagon
The construction of a regular pentagon is more complicated than

most other constructible regular polygons. The regular hexagon is
easily constructible. One need only draw circles around the cir-
cumference of a given circle with the same radius as the given cir-
cle. As you can see in figure 4-47, only four such circles are
needed. Joining the intersection points gives us a regular hexagon.
Figure 4-47
Were we to try to construct a regular pentagon in a similar

way, we would find ourselves in a dilemma. Perhaps the most im-
portant artist Germany has contributed to Western culture is Al-
brecht Durer (1471-1528). A much forgotten work by this artist (in
1525) is a geometric construction (using a straightedge and a pair
of compasses) of a regular pentagon, which he knew was an ap-
proximation of a regular pentagon but extremely close to perfect. It
is so nearly perfect that its inaccuracy is not visually detectable. He
offered this very easy construction to the mathematical community
as an alternative method to get a "regular" pentagon, with the full
knowledge that is was off by about half a degree. 26 Its deviation
from a perfect regular pentagon is minuscule, but that cannot be
26. C. J. Scriba and P. Schreiber, 5000 Jahre Geometrie. Geschichte, Kulturen,

Menschen (Berlin: Springer, 2000), pp. 259, 289-90.
ignored. Until recently, engineering books still provided it as a

method for constructing a regular pentagon. We shall provide an
explanation of the method here, despite its being slightly inaccu-
rate, since it is instructive and has been used for many years.
In figure 4-48 we begin with a segment AB. Five circles of ra-
dius AB are constructed as follows:
1. Circles with centers at A and B are drawn and intersect at

QandN.
2. Then the circle with center Q is drawn to intersect circles
with centers at A and B at points Rand S, respectively.
3. QN intersects circle Q with center at at P.
4. SP and RP intersect circles with centers at A and B at
points E and C, respectively.
5. Draw the circles with centers at E and C, with radius AB
to intersect at D.27
6. The polygon ABCDE is (approximately) a regular pentagon.
-'. -'.
--.0.
:Q
'. "
'.
E C
:
."
..
Figure 4-48
27. Notice we only need one of the points of intersection (D) of these two
circles.
Joining the points in order (figure 4-49), we get the pentagon

ABCDE.
.................. -...............
.....
/........ ...., •••.~....... ......,..\...)
'.
, ........
...
/
....
Figure 4-49
Although the pentagon "looks" regular, it's mLABC is about

22 of a degree too large. In other words, for ABCDE to be a regular
60
pentagon, each angle must be 108°. Instead we will show that
mLABC ~ 108.3661202° .
•• ••••••••••••••••••••·•·•••• •••••••
n••••• ••••••·••••• .......................
...... ....
..•• " " "
./ \ '\
E f N \ C \
: ;
". ./
Figure 4-50
The rhombus in figure 4-50, ABQR, has LARQ = 60° and

BR = AB,{j since BR is actually twice the length of an altitude of
equilateral MRQ. Since MRQ is an isosceles right triangle,

mLPRQ = 45° ,and then mLBRC = 15° .
We shall apply the law of sines to MCR: . BR _ BC ·
sin LBCR sin LBRC '
·
that IS, . ABJi
LBC'D = sin.AB15 '
· LBCfn
or SIn r:. 150
.n = ",3 SIn .
sin .n
0
Therefore, mLBCR ~ 26.63387984 .

In MCR,
mLRBC = 180° - mLBRC - mLBCR
~ 180° - 15° - 26.63387984°
~
138.3661202°
Thus, since mLABR = 30°,
mLRBC - mLABR ~ 138.661202° - 30° ~ 108.3661202°
and not 108°, as it should be in order for it to be a regular penta-
gon!
The results of Diirer's construction:
mLABC= mLBAE ~ 108.37°, mLBCD = mLAED
~ 107.94°, and mLEDC ~ 107.38°.
To construct a proper regular pentagon, we would first con-

struct a golden triangle and then simply mark its base off along a
given circle as shown in figure 4-51.
Figure 4-51
We can even create the golden ratio by paper-folding methods.

Take a strip of paper, making sure the sides are parallel, and fold it
into a regular knot. Then carefully pull it taut so that it looks like
picture 3 in figure 4-52. You can tear off the end flops as in picture
4, if you wish. You have now formed a regular pentagon. If you
hold this pentagon up to the light and look through the knot, you
should see the pentagram inside the pentagon that you formed.
Figure 4-52
We have shown the many geometric sightings where-when

you least expect it-the Fibonacci numbers appear and sometimes
in the form of the golden ratio. We leave you in this chapter with a
series of regular pentagons that are placed so the vertex of each
divides the side of another in the golden ratio (see figure 4-53).
Figure 4-53
Ifwe join specific points (figure 4-54). we can generate a series

of spirals reminiscent of those we encountered in chapter 2.
Figure 4-54
In figure 4-55 the sides of neighboring (or adjacent) pentagons

are in the ratio equal to t/J: I. Once again, we can use the
Fibonacci numbers to render the appropriate approximation.
Figure 4-55
There are many more such appearances in geometry that we

will leave the reader to discover.
Chapter 5
The Fibonacci Numbers

and Continued Fractions
W e have seen the Fibonacci numbers as they appear in nature,

as they manifest themselves among other numbers and se-
quences, and as they can be seen in geometry. They also can be
seen in the study of continued fractions. A continued fraction is
just another form of a fraction that allows us to look at the nature
of numbers in a different way. We will introduce you to continued
fractions. Once again, after you are comfortable with continued
fractions, we will investigate their relationship to the Fibonacci
numbers.
Continued Fractions
We will begin with a brief introduction to continued fractions. A

continued fraction is a fraction in which the denominator has a
mixed number (a whole number and a proper fraction) in it. We
can take an improper fraction such as ~ and express it as a mixed
7
number:
6 6
1-=1+-
7 7
Without changing the value, we could then write this as
161
6 1
1+-=1+-
7 7
6
which in tum could be written (again, without any value change)

as:
1
1+----
1
1+-
6
This is a continued fraction. We could have continued this process,

but when we reach a unit fraction (i.e., a fraction in which the nu-
merator is 1 and the denominator is a positive integer; in this case,
the unit fraction is ~), we are essentially finished.
6
So that you can get a better grasp of this technique, we will

create another continued fraction. We will convert g to contin-
7
ued fraction form. Notice that at each stage, when a proper frac-
tion is reached, we take the reciprocal of the reciprocal (e.g.,
change ~ to _1 ,as we will do in the example that follows), which
5 5
2
does not change its value:
12 5 1 1 1 1
-=1+-=1+ =1+ =1+ =1+
7 7 7 2 1 1
1 +- 1+- 1+
5 5 5 1
2+-
2 2
If we break up a continued fraction into its component parts

(called convergents),1 we get closer and closer to the actual value
of the original fraction.
1. This is done by considering the value of each portion of the continued fraction
up to each plus sign, successively.
THE FIBONACCI NUMBERS AND CONTINUED FRACTIONS 163
F Irst convergent 0 f
o 12
- == 1
7
12 1
Second convergent of - == 1+ - == 2
7 1
1 2 2 5
Third convergent of g ==1+ ==1+-==1-==-
7 1+~ 3 3 3
2
12 1 12
Fourth Convergent of - == 1+ - - - -
717
1+--
1
2+-
2
The above examples are all finite continued fractions, which

are equivalent to rational numbers (those that can be expressed as
simple fractions). It would then follow that an irrational number
would result in an infinite continued fraction. That is exactly the
case. A simple example of an infinite continued fraction is that of
J2 . Although we show it here, we will actually generate it just a
bit further on.
J2==I+---------I--------
1
2 + ----------
1
2+--------
1
2+-------
1
2+-----
1
2+----
1
2+--
2 + ...
We have a short way to write a long (in this case infinitely

J,
long!) continued fraction: [1;2,2,2,2,2,2,2, ... or when there are
these endless repetitions, we can even write it in a shorter form as
[ 1;2J, where the bar over the 2 indicates that the 2 repeats end-
lessly.
In general, we can represent a continued fraction as:
1
1
1
a 3 + ... .
1
1
a n-I +-
an
where ai are real numbers and ai ~ 0 for i > O. We can write this in
a shorter fashion as: [aO ;al'a2 ,a3 , • • • ,an_pan] .
As we said before, we will generate a continued fraction equal
to 12.
Begin with the identity: J2 + 2 = J2 + 2
Factor the left side and split the 2 on the right side:
J2 (1 + J2) = 1+ J2 + 1
Divide both sides by 1+ J2 to get:

J2=I+ 1 =[I;I,J2]
1+J2
Replace J2 with .J2 = 1 + 1 r;; and simplify the terms:

l+v2
J2 = 1+ 1 1 = 1+ 1 = [1; 2, I, J2 ]
1
1+(1+ 1+J2) 2+ 1+ J2
Continue this process. The pattern now becomes clear.

J2 == 1 + ___1-1-- = [1; 2, 2, 1,.J2], and so on.

2+ 1
2 + -------::=
1+J2
Eventually, we conclude with the following:
J2 == 1 + _ _ _1 _ _ = [1; 2, 2, 2, ... ]
1
2+-----
1
2+---
2+ ...
Thus we have a periodic continued fraction for .J2 :

(.J2 = [ 1; 2, 2, 2, ... ] = [1; 2])
There are continued fractions equal to some famous numbers

such as Euler's e(e:= 2.7182818284590452353 . . . )2 and the
famous 1t (1t:= 3.1415926535897932384 ... ):
1
e = 2 + -------------
1
1+ - - - - - - - - - - - -
1
2 + ----------
1
1+ ---------
1
1+--------
1
4+------
1
1+-----
1
1+---
6+ ...
= [2; 1,2, 1, 1,4, 1, 1,6, 1, 1,8,1,1,10, ... ] = [2; 1, 2n, 1]
2. The number e is the base of the system of natural logarithms. When n in-
creases without limit, then e is the limit of the sequence (.+:f. The symbol e
was introduced by the Swiss mathematician Leonhard Euler (1707-1 783) in
1748. In 1761 the German mathematician Johann Heinrich Lambert
(1728-1777) showed that e is irrational, and in 1873 the French mathematician
Charles Hermite (1822-1901) proved e is a transcendental number.
A transcendental number is a number that is not the root ofany integer poly-
nomial equation, meaning that it is not an algebraic number of any degree. This
definition guarantees that every transcendental number must also be irrational.
Here are two ways that 1C can be expressed as a continued

fraction. 3
4
1t = - - - - - - - - -
_1C =1+ _ _ _ _1_ _ __
}2 2 1·2
1 + - - - -2 - - - 1+-------
3 2·3
2+------ 1+-----
52 3·4
1+----
2+--- 2
- 4·5
7 1+---
2+--- 2 1 + ...
9
2+-
Sometimes we have continued fractions representing these

famous numbers that do not seem to have a distinctive pattern:
1C = 3 + -----------------
7 + ---------------
15+----------------
1+ - - - - - - - - - - - -
292 + - - - - - - - - - -
1+--------
1+-------
1+------
2+----
1+---
3 + ...
1C= [3; 7,15,1,292,1,1,1,2, 1,3,1,14,2,I,I,2,2,2,2,1,84,2, ... J.
We have now set the stage for the golden ratio. 4 Can we
express this Fibonacci-related ratio as a continued fraction?
Let's try to use these continued-fraction techniques with </J.
3 For more on the various representations of 1C see A. S. Posamentier and

I. Lehmann, 1C: A Biography of the World's Most Mysterious Number
(Amherst, NY: Prometheus Books, 2004).
4. Remember </J = 1.61803 39887498948482 ....
Actually, we will use our now well-known relationship for l/J to

produce one of the nicest continued fractions_ Begin with this
relationship: t/J = 1 + ~ ·
We can now replace the l/J in the denominator on the right side
with its equal: 1 + ~.
- . 1
Th IS gIves us: l/J = 1 + - 1 .
1+-
l/J
Continuing this process will give us the following:
1
l/J= 1 + -= [1; l/J]
l/J
1
l/J= 1 + 1 = [1; 1,l/J]
1+-
l/J
1
l/J=I+ =[ 1; 1, 1,l/J]
1
1+
1
1 +-
l/J
1
l/J=I+ = [1; 1, 1, 1,l/J]
I
1+
1
1+
I
1 +-
l/J
1
l/J=I+ = [1; 1,1,1, 1,l/J]
1
1+
1
1+
1
1+
1
1 +-
l/J
1
41=1+ = [1; 1, 1, 1, 1, 1,41]
1
1+
1
1+
1
1+
1
1+
1
1 +-
41
1
41=1+ = [ 1; 1, 1, 1, 1,1,1,41]
1
1+
1
1+
1
1+
1
1+
1
1+
1
1+-
41
1
41=1+ =[ 1; 1, 1, 1, I, 1, ... ]=[1]
1
1+
1
1+
1
1+
1
1+
1
1+
1
1+
1 + ...
and so on.
I
Thus we now have a continued fraction equal to l/J from our
val ue for l/J :
1 1 -
- - - - - - = [0; 1, 1, I, 1, 1, I, ... ] = [0; 1]
1
1+----
1
1+---
1
1+-
This has to be the nicest continued fraction of all, since the golden
section continued fraction (l/> = J5 + I ::::: l.6180398874989) and its
2
reciprocal (~ = ~-I ::::: 0.61803398874989) are comprised of all Is.
On the other hand, these continued fractions, though elegant,

are very slow to approach their true value. You will need many
terms to get a good approximation. We will do this with the help of
the Fibonacci numbers:
I Fa 2 F2
lP) = I + - = 1 + - = - = - = 2
1 F., I Fi
I 1 1 Fi 3 F3
lP2 = 1+ - - = 1 +-= 1 +-= 1 + - = - = - = 1.5
1+ ~ lP, 2 F2 2 F2
1
1 1 2 F 5F -
lP3 = I + = 1+- = 1 + - = I + - 2 = - = - 4 = 1. 6
1 + _I_ lP2 3 F3 3 F3
1
1 +-
1
1 1 3 F38Fs
lP4 = 1 + - - - - = 1 +-= 1 +-= I + - = - = - = 1.6
1 + __1__ lP1 5 F4 5 F4
1
1+--
1
1 +-
I
I 5 F4 13 F6
lPs = 1 + - - - - - = I +-= 1 +-= 1+-=-=-= 1.625
1+---- ~ 8 ~ 8 ~
1+---
1
1+--
1
1 +-
1
1 1 8 Fs 21 F7
l/1 6 = 1 + - - - - - - - = 1 +-= 1 +-= 1 + - = - = -
1+ ___ 1___ l/1 s 13 F6 13 F6
1
1+-----
1
1+---
1
1+--
1
1+-
1
= 1. 615384
1 I 13 F6 34 Fg
l/17 = 1 + - - - - - - - - - = 1 +-= 1 +-= 1 + - = - = -
1 + _ _ _ _1____ l/16 21 F7 21 F7
1
1+------
1
1+-----
1
1+---
1
1+--
1
1 +-
1
= 1. 619047
The nth case is then:

1 1 F F
l/1n = 1 + - - - - - - - - - = 1 +--= I +~=2±.!..
n . 1+------ l/1n-1 ~, Fn
1+-----
1+----
1+---
1+--
1+ ...
We then see that the limiting value of the continued fraction for
5
<Pn is:
. F J5+1
11m ~ = = <P (See chapter 4.)
n~oo Fn 2
5. This reads "the limiting value of Fn+l , as n gets larger and larger and
Fn
approaches 00,
.J5+1
equals - A\
- = 'f' ."
2
In 1968 Joseph S. Madachy6 introduced a new constant J..l, the

continued fraction
1
J..l = 1 + - - - - - - - - - - - -
3
2 + -----------
8
5+ ----------
21
13 + - - - - - - - -
55
34+-----
144
89+----
377
233+-
= 1.3941865502287836729028896495777209667374096430683 ...
We see that the terms of this continued fraction are successive

Fibonacci numbers. For this constant J..l is l/> < J..l < .J2 a rough
estimation (inequality).
The famous Scottish mathematician Robert Simson (1687- 1768),
who took Euclid's Elements and wrote an English-language book on
it, which is largely responsible for the development of the foundation
of the high school geometry course taught in the United States, was
F
the first to recognize that the ratio ---1!±..!.. of two consecutive Fibonacci
Fn
numbers will approach the value l/> of the golden ratio.

Consider the following list of fractions (figure 5-1) to see the
value of l/> being approached.
($ 0) $1 $2 C/>3 $4 $5 $6 $7 ... $
Numerator (J) 2 3 5 8 13 21 34
Denominator (I) ] 2 3 5 8 13 21
J 2 .3 4 56 7
Figure 5-1
6. HRecreational Mathematics," Fibonacci Quarterly 6, no. 6 (1968): 385-92.

In order to see how the shape of the resulting rectangle gets to

look more and more like a golden rectangle, we provide you with a
series of drawings. The width b remains constant, while the length
a varies. Each time the rectangle takes on the dimensions of the
successive Fibonacci numbers. At the far right, you have the true
golden rectangle for comparison. So with the numerator and de-
nominator of lP; (with i == 1, 2, 3, . . . , 7), you can see the
development of the golden rectangle (figure 5-2).
a-.1·b a-.2·b a-.".6 a-.".b II-.s·b a- ••. b a-.7·b ... a- •. b
J
'2.b 8 13 21 34 .;5+1
a=
a= s·b a-,.b a-"'i3 b a-ll·b ... a--2- b
Figure 5-2
You will notice that </15 ==.!1. == 1.625 is a good approximation for
8
</1 == 1.6180339887498948482 .... While for </17 == 34 == 1. 619047 , we
21
can hardly see the difference between the rectangle formed and the
actual golden rectangle. In figure 5-3 the left-side rectangle is the
approximation, and the right-side rectangle is the golden one.
(/).21 -33.98
21 21
Figure 5-3
A Nest of Radicals
Interestingly enough, the 1s play another role in the value of l/J •

Consider the "nest of radicals"
and let's see how we might find the value of it.

This is an analogous method to the one we used to evaluate an
infinite continued fraction, since this nest also goes on infinitely
long. So we begin by setting this value equal to x :
We then square both sides of this equation to get:
Since
we can replace the nest of radicals in the above equation with x to

get x 2 = 1+ x, which can be written as x 2 - x-I = o.
The positive value of x is then 1+; = cf>. So we then have
Aha! Another surprising appearance of the golden ratio!

Fibonacci and Lucas Numbers
Let us return now to the Lucas numbers. We would expect to find a

relationship between the ratios of consecutive Lucas numbers and
the golden ratio, just as we have for the Fibonacci numbers. In
figure 5-4, you will notice that both sequences 7 approach the
golden ratio.
With that being the case, we ought to be able to use continued
fractions with the Lucas numbers also to generate the approach to
the golden ratio. Consider the following ratios of consecutive Lu-
cas numbers:
3 2 1
-=1+-=1+-
1 1 1
2
4 1 1 1
-=1+-=1+ =1+--
3 3 1+2 1+_1
1
2
731 1
-= 1 + -= 1 + - = 1 + - - - -
4 4 4 1
1+ 1
3
1+-
1
2
1
This, in a short version, can be written as [1; 1, 1, 1, ... , -].
2
The fact that the last term here is ~ ,rather than 1 as it is with the
2
Fibonacci numbers, becomes relatively insignificant as the contin-

ued fraction increases in terms.
The study of continued fractions holds many other fascinating
surprises. But for now we have had the pleasure of using this tech-
nique to further explore the Fibonacci and Lucas numbers.
7. Rounded to nine decimal places.

Fn+1 L n+1
Fn Ln
1 3
- = 1.000000000 - = 3.000000000
1 1
2 4
- = 2.000000000 - = 1.333333333
1 3
3 7
- = 1.500000000 - = 1.750000000
2 4
5 11
- = 1.666666667 - = 1.571428571
3 7
8 18
- = 1.600000000 - = 1.636363636
5 11
13 29
- = 1.625000000 - = 1.611111111
8 18
21 47
- = 1.615384615 - = 1.620689655
13 29
34 76
- = 1.619047619 - = 1.617021277
21 47
55 123
- = 1.617647059 - = 1.618421053
34 76
89 199
- = 1.6182181618 - = 1.617886179
55 123
144 322
- = 1.617977528 - = 1.618090452
89 199
233 521
- = 1.618055556 - = 1.618012422
144 322
377 843
- = 1.618025751 - = 1.618042226
233 521
610 1 364
- = 1.618037135 , = 1.618030842
377 843 Figure 5-4
987 2,207
- = 1.618032787 = 1.618035191
610 1,364
Chapter 6
A Potpourri of Fibonacci
Number Applications
W e will now take a brief tour through some rather unusual

and sundry applications and sightings of the ubiquitous Fi-
bonacci numbers. Some will be serious and others lighthearted.
These examples are merely a sampling of the wide variety of uses
and appearances in which the Fibonacci numbers have been found
to manifest themselves. This is by no means meant to be a com-
plete collection. Such a collection would be limitless!
Business Applications
We have seen evidence of the Fibonacci numbers popping up in

the most unexpected places. Now, we go searching for them in the
volatile world of the stock market.
All human activities have three distinctive features, pat-

tern, time and ratio, all of which observe the Fibonacci
summation series
-R. N. Elliott
Investor confidence was at an all-time low after the stock mar-

ket's Great Crash of 1929, whereupon many Americans viewed the
stock market as a costly roll of the dice. Nonetheless, a little-
177
known, albeit successful, accountant and engineer, Ralph Nelson

Elliott (1871 - 1948), decided to pore over decades of stock per-
formance charts and trace trader movement in an attempt to make
sense of the crash. As his research progressed, distinct and repeti-
tive zigzagging patterns began to emerge as Elliott studied the ups
and downs, or rather, ebb and flow, of market behavior. He dubbed
these patterns "waves" and categorized them as either "impulsive"
or "corrective" waves that could be measured and used to forecast
market behavior.
Figure 6-1
Ralph Nelson Elliott.
In a letter to Charles Collins, the publisher of a national market

newsletter of the day with an extensive following, which included
R. N. Elliott, the enthusiastic Elliott wrote that there was "a much
needed complement to Dow theory" that he called his "wave the-
ory." This correspondence began in November 1934 when Elliott
wrote to Collins about his "discoveries" with the hope that he
would find support from Collins. Collins was impressed by the ac-
curacy of Elliott's analysis and invited him to Detroit to explain
the process in greater detail. Although Elliott's insistence that all
market decisions should be based on wave theory prevented
Collins from directly employing Elliott, he did help him to estab-
A POTPOURRI OF FIBONACCI NUMBER ApPLICATIONS 179
lish an office on Wall Street. Later, in 1938, Collins wrote a book-

let, under Elliott's name, titled The Wave Principle. Elliott went on
to further publicize his theory through various letters and magazine
articles, including publication in the Financial World magazine.
Then, in 1946, Elliott expanded on The Wave Principle with a
book called Nature's Law-The Secret of the Universe.
Elliott's famous, influential booklet, The Wave Principle, is
based on the belief that market behavior should not be seen as ran-
dom or chaotic but rather as a natural reflection of investor confi-
dence, or lack thereof, (impulsive waves) and self-sustaining mar-
ket mechanisms (corrective waves). In layman's terms, Elliott felt
the market, like "other things in the universe," moves in predict-
able cycles once its patterns of behavior have been established and
made visible to the trained eye. At this point, it should come as no
surprise that the Fibonacci numbers should be visible within those
patterns.
What may be surprising is just how pervasive the Fibonacci
numbers are in Elliott's analysis of the stock market. Tracing most
bear markets, he found that they move in a series of 2 impulsive
waves and 1 corrective wave, for a total of 3 waves (sound famil-
iar?). On the other hand, his research showed that a bull market
usually jumps upward in 3 impulsive waves and 2 corrective
waves, for a total of 5 waves (recognize these numbers?). A com-
plete cycle would be the total of these moves, or 8 waves. The re-
lationship of the Fibonacci sequence to the wave principle does not
end there.
Elliott's wave principle also states that each "major" wave can
be further subdivided into "minor" and "intermediate" waves. A
regular bear market has 13 intermediate waves. Yes (as you might
have suspected), a bull market has 21 intermediate waves, for a
total of 34. Continuing on with the sequence, there are 55 minor
waves in a bear market and 89 in a bull market, for a total of 144.
Elliott himself was not at all surprised to find the Fibonacci se-
quence while tracing these market waves because he had always
believed that "the stock market is a creation of man and therefore
reflects human idiosyncrasy." However, it was only when Elliott
began to study the relationships between the size of the waves for
his second work, Nature's Law-The Secret of the Universe, that
he realized his new discoveries were based on "a law of nature
known to the designers of the Great Pyramid 'Gizeh,' which may
have been constructed 5,000 years ago." (See chapter 7.)
Of course, he was talking about the golden ratio. Elliott be-
lieved that this most famous of ratios, along with other Fibonacci-
generated ratios, could be used to predict stock prices with as-
tounding accuracy. In order to understand how he reached this
conclusion, it is important to investigate what these Fibonacci ra-
tios are. Consider the results when dividing the Fibonacci numbers
by their two immediate successors; can you see a pattern emerge
after the first several calculations? (See figure 6-2.)
It is clear that after the first several columns, the results in each
of the succeeding columns approximate .2360, .3820, and the
golden ratio, .6180. Written as percentages, these numbers trans-
late to: 23.6 percent, 38.2 percent, and 61.8 percent and are called
Fibonacci percentages. It is interesting to notice a very curious
property: the sum of the first two percentages equals the third per-
centage (.2360 + .3820 = .6180); and the last two percentages
(.3820 and .6180) happen to have a sum of 100 percent, a fact that
can be proved algebraically.
Elliott's premise, or the Fibonacci indicator, is that the ratio,
or the proportional relationship, between two waves could be
used to indicate stock prices. Elliott found that an initial wave up
in price was usually followed by a second wave downward, or
what market analysts call a "retracement" of the initial surge.
Upon closer inspection, these retracements seemed, more often
than not, to be Fibonacci percentages of the initial price surge,
with, 61.8 percent as the highest retracement. Once again, the
golden ratio reigns supreme!
n Fn
-
F" -F"- -F"-
F,,+I F,,+2 F,,+3
1 1 I 1 1
- = 1.00000000 - = .500000000 - ::::; .33333333
1 2 3
2 1 1 1 1
-= .500000000 - ::::;.33333333 - = .20000000
2 3 5
3 2 2 2 2
- ::::: .666666667 -=.40000000 - =.25000000
3 5 8
4 3 3 3 3
- = .600000000 - = .37500000 - ::::; .230769231
5 8 13
5 5
5 5 5
- = 0.625000000 - ::::; .384615385 - = .238095238
8 13 21
6 8
8 8 8
- = 0.615384615 - ::::; .380952381 - = .235294118
13 21 34
7 13
13 13 13
- = .619047619 - = .382352941 - = .236363636
21 34 55
8 21
21 21 21
- = .617647059 - = .381818182 - ::::; .235955056
34 55 89
9 34
34 34 34
- = .618181818 - = .382022471 - = .236111111
55 89 144
10 55
55 55 55
- ::::; .617977528 - = .381944444 - = .236051502
89 144 233
11 89
89 89 89
-::::; .618055555 - = .381974249 - = .236074271
144 233 377
12 144
144 144 144
- ::::; .618025751 - = .381962865 - = .236065574
233 377 610
13 233
233 233 233
- ::::; .618037135 - = .381967213 - ::::; .236068896
377 610 987
Figure 6-2
While many of today's market analysts incorporate wave pat-

terns and retracements to forecast the market, some analysts take
the Fibonacci numbers a step further. They scan previous dates and
results looking for the Fibonacci numbers. There might be 34
months or 55 months between a major high or low, or as few as 21
or 13 days between minor fluctuations. A trader might look for re-
petitive patterns and try to establish a Fibonacci relationship be-
tween dates to help forecast the future market.
Before you scoff at the importance of the Fibonacci numbers in
the stock market, consider that Elliott, at the age of sixty-seven and
without the aid of any computers, forecast the end of a bear market
decline from 1933-1935 to the exact day_
By the way, while considering the applications of the Fibonacci
numbers in the financial world, even one of the main vehicles by
which many of these trades are made, the credit card, measures
very close to the golden rectangle-55 mm by 86 mm is actually
only 3 mm short in the length from fitting a Fibonacci ratio ap-
proximation.
A book by Robert Fischer, Fibonacci Applications and Strate-
gies for Traders,l expounds on the use of the Fibonacci numbers in
investment strategies. It is a complete, hands-on resource that
shows how to measure price and time signals quickly and with ac-
curacy, by using the logarithmic spiral. He explains how to act on
these analyses. This innovative trading guide first takes a fresh
look at the classic principles and applications of the Elliott wave
theory, then introduces the Fibonacci sequence, exploring its ap-
pearance in many of the fields discussed here, and then in stock
and commodity trading.
Fischer then explains how the Fibonacci sequence is used to
measure equity and commodity price swings and to forecast
short- and long-term correction targets. You can learn how to
accurately analyze price targets on market extensions, and how,
by using the author's logarithmic spirals, to develop price and
time analyses with some precision. The book claims to enable
you to calculate and predict key turning points in the commodity
1. (New York: Wiley, 1993).

market, analyze business and economic cycles, and discover en-

try and exit rules that make disciplined trading possible and
profitable. We are not suggesting that this is a foolproof method
for becoming rich. But the Fibonacci numbers continue to fasci-
nate some in the investment field.
The Vending Machine
A completely different type of application would be setting up a

vending machine to dispense a variety of candies. Y ou are told that
the cost of each of the candies is to be a multiple of 25 cents. Each
type of candy will be a different arrangement (order) of coins de-
posited, so that the amount of money deposited into the machine
and the arrangement of the order in which the coins have been de-
posited will determine which of the candies has been selected. We
need to calculate how many different types of candy the machine
can accommodate. Put another way, we need to determine the
number of different ways each multiple of 25 cents can be arrived
at using the machine's coin slot, which accepts only quarters and
half-dollars. For example, for a 75-cent candy selection there are
three possible ways to pay: three quarters (QQQ), a quarter and a
half-dollar (QH), or a half-dollar and a quarter (HQ). Each of these
payments will result in a different type of candy. When we make a
chart (figure 6-3) of the various amounts of money that the ma-
chine can accept, a curious set of numbers appears. Can you guess
what that might be?
Once again, seemingly out of nowhere, the Fibonacci numbers
appear in the right-side column, indicating the number of ways by
which the various amounts of money can be deposited. This would
help the manufacturer of the vending machine predict the number
of ways that one could deposit $3.00. The answer would be 12
times 25 cents, or the thirteenth number in the Fibonacci sequence,
namely, 233.
Cost of Number of List of ways to pay Number

the multiples of of ways
candies 25 cents to pay
1
$.25 1 Q
2
$.50 2 QQ,H
3
$.75 3 QQQ, HQ~ QH
5
$1.00 4 QQQQ,HH~QQH,HQQ,QHQ
QQQQQ, HHQ, QHH, HQH~ QQQH~ 8

$1.25 5
HQQQ, QHQQ, QQHQ
$1.50 6 QQQQQQ, QQQQH, QQQHQ~ QQHQQ, 13

QHQQQ, HQQQQ~ QQHH, HHQQ~
HQHQ,QHQH,HQQH,QHHQ~HHH
$1.75 QQQQQQQ, QQQQQH, QQQQHQ~ 21

7
QQQHQQ, QQHQQQ, QHQQQQ~
HQQQQQ, QQQHH, QQHHQ, QHHQQ~
HHQQQ, QHQHQ, HQHQQ, QQHQH~
HQQQH, HQQHQ, QHQQH, HHHQ,

HHQH, HQHH, QHHH
QQQQQQQQ, QQQQQQH, QQQQQHQ~ 34

$2.00 8
QQQQHQQ~ QQQHQQQ, QQHQQQQ,
QHQQQQQ, HQQQQQQ, QQQQHH~
QQQHHQ~ QQHHQQ~ QHHQQQ,

HHQQQQ~ HQHQQQ, QHQHQQ~
QQHQHQ, QQQQHH, HQQHQQ~
QHQQHQ, QQHQQH, HQQQHQ~
QHQQQH~ HQQQQH, QQHHH, HQQHH,

HHQQH~ HHHQQ~ QHQHH, HQHQH,
HHQHQ, QHHQH, HQHHQ, QHHHQ~
HHHH
Figure 6-3
Climbing a Staircase
An analogous situation arises with the following problem. Suppose

you are about to use a staircase with n stairs and you can either take
one step at a time or two steps at a time. The number of different
ways of climbing the stairs ( en ) will be a Fibonacci number-much
the same way as the previous situation with the vending machine.
Suppose n = 1, which means there is only one stair to climb.

Then the answer is easy: there is only one way to climb the stair.
If n = 2, and there are two stairs, then there are two different
ways to climb the steps: two single steps or one double step.
If n = 3, and there are three steps, then there are three different
ways to climb the steps: 1 step + 1step + 1step, 1 step + 2 steps, or 2
steps + 1 step. This can be generalized (now that you are experts at
counting these stair-climbing possibilities). For the staircase of n
steps (where n > 2), the number of ways of climbing them is C n •
After you take the first step up on a single stair, there are n - 1 stairs
left to climb. So the number of ways to climb the remainder of the
staircase is Cn _ l . If you begin your climb with a double step (i.e., 2
steps have been climbed), then there are C n - 2 ways to climb the rest
of the staircase. Thus the number of ways to climb the staircase of n
steps is the sum of these: C n == C n _ 1 + C n - 2 • Does this remind you of
the Fibonacci recursive relationship? The values of C n will follow
the Fibonacci pattern, with Cn == Fn + 1 • You can check this with
smaller numbers of stairs and see that it works, or you can compare
this to the previous vending machine problem and see that it is the
equivalent of that-with the 25-cent units analogous to the stairs.
On the other hand, we can inspect a rather large example. The
Empire State Building is 1,453 feet 8 and 9/16th inches tall from
street level to its highest point-a lightning rod. There are annual
races up the building's 1,860 stairs. In how many ways can an Em-
pire State Building stairs runner reach the topmost step?
It is an unbelievably large number:

C n = C1.860 = FJ.861 = 37,714,947,112,431,814,322,507,744,749,931,
049,632,797,687,008,623,480,871,351,609,764,568, 156, 193,373,
680,151,232,412,945,298,517, 190,425,833,936,823,942,275,395,
680,820,896,518,732,120,268,852,036,861,867,624, 728,128,920,
239,509,015,217,615,431,571,741,968,260,146,431,90 1,232,750,
464,530,968,296,717,544,866,475 ,402,917,320,392,352,090,243,
657,224,327,657, 131,325,954,780,580,843,850,283,683,054,714,
131, 136,328,674,469,916,443,464,802,738,976,662,616,325, 164,
306,656,544,521,133,547,290,540,333,738,912, 142,760,761 ~ 3.7
71494711 · 10388
Painting a House Creatively
The stories of an n-story house are to be painted blue and yellow,

with the stipulation that no two adjacent stories are blue. However,
both of them may be yellow. We will let an represent the number
of possible colorings for a house with n stories (n > I).
I story: 2 stories: 3 stories:

2 possibilities 3 possibilities 5 possibilities
4 stories:
8 possibilities
Figure 6-4
In figure 6-4, you will see the possible ways to paint a house of
various story heights. Now suppose we have a building with five sto-
ries. The fifth story can be painted either yellow or blue. If it gets a
yellow color, then the remaining stories can be painted like a four-
story house. If it is painted blue, the fourth story must get yellow and
the first three stories can be painted like a three-story house. Therefore,
we find that a5 = a4 + a3 = 8 + 5 = I3 = F6 + F5 = F 7. Do you recognize
the recursive pattern here? It should be reminiscent of the pattern used
to generate the Fibonacci numbers.
Ordered Sum of Ones and Twos
Another nice application of the Fibonacci numbers can be seen

from the following challenge:
In how many ways can a natural number, n, be written as
an ordered sum of ones and twos?
Let's inspect the chart in figure 6-5 and count the ways in which
we can fulfill this task.
n 1 2 3 4 5 6 7
Fn 1 1 2 3 5 8 13
Fn+1 1 2 3 5 8 13 21
ordered 1 1+1 1+1+1 1+1+1+1 1+1+1+1+1 1+1+1+1+1+1 1+1+1+1+1+1+1
sum
of ones 2 1+2 1+1+2 1+1+1+2 1+1+1+1+2 2+1+1+1+1+1
and twos 2+1 1+2+1 1+1+2+1 1+1+1+2+1 1+2+1+1+1+1
2+1+1 1+2+1+1 1+1+2+1+1 1+1+2+1+1+1
2+1+1+1 1+2+1+1+1 1+1+1+2+1+1
2+2 2+1+1+1+1 1+1+1+1+2+1
1+2+2 1+1+1+1+1+2
2+1+2 1+1+2+2
2+2+1 1+2+2+1 2+2+1+1+1
2+2+1+1 2+1+2+1+1
2+1+2+1 2+1+1+2+1
1+2+1+2 2+1+1+1+2
2+1+1+2 1+2+2+1+1
1+2+1+2+1
2+2+2 1+2+1+1+2
1+1+2+2+1
1+1+2+1+2
1+1+1+2+2
2+2+2+1
2+2+1+2
2+1+2+2
1+2+2+2
Figure 6-5
We might conclude from the pattern that seems to be evolving

in the chart (figure 6-5) that the natural number n can be written in
Fn+l different ways as an ordered sum of ones and twos, where Fn
is the nth Fibonacci number.
For all n > 2, the representation of n ends with 1 or 2. The pre-
vious terms add up to n - 1 or n - 2. This yields Fn+l = Fn + F n- 1;
n > 2.
Representing Natural Numbers as the Sum of

Fibonacci Numbers
Suppose we take any natural number,2 say, 27, and try to represent
this number as the sum of some Fibonacci numbers. We could do it
as 27 = 21 + 5 + 1, or as 27 = 13 + 8 + 3+ 2 + 1, or even as
27 = 13 + 8 + 5 + 1, and with other combinations as well. This will
be possible for all natural numbers. 3 You might want to try it for
other randomly selected natural numbers. What you should find
(with enough examples) is that there will be only one way of repre-
senting any natural number with nonconsecutive Fibonacci num-
bers.4 In the above example, the first illustration, 27 = 21 + 5 + 1,
is the only representation of 27 with nonconsecutive Fibonacci
numbers. Here are some examples of the unique representations of
natural numbers with nonconsecutive Fibonacci numbers:
1 = F2 7 = F3 + Fs= 2 + 5
2 = F3 8 = F6
3 =F4 9 = F2 + F6= 1 + 8
4 = F2 + F4 = 1 + 3 10 = F2 + F6= 2 + 8
5 =Fs 11 = F4 + F6 = 3 + 8
6 = F2 + Fs = 1 + 5 12 = F6 + F4 + F2 = 8 + 3 + 1
Covering a Checkerboard
Strangely enough, the Fibonacci numbers can describe the number

of ways to tile a 2 by (n-l) checkerboard with 2 by 1 dominoes;
that is, each domino will cover exactly two adjacent squares on the
checkerboard. Let's take a look at how that can be done for the
first few examples of these.
2. A natural number is one of the counting numbers: 1, 2, 3, 4, 5, ....

3. There is a proof of this in appendix B.
4. This was proved by the Belgian amateur mathematician Edouard Zeckendorf
(1901-1983), and the theorem is known as the Zeckendorf theorem.
There is only one way to tile a 2 x I board:
Figure 6-6
There are two ways to tile a 2 x 2 board:
n c
U ::J
Figure 6-7
There are three ways to tile a 2 x 3 board:
Figure 6-8
There are five ways to tile a 2 x 4 board:
n c n :::J
U ::J C U
c n c C
:J U ::J :J
Figure 6-9
There are eight ways to tile a 2 x 5 board:
n n n n n :::J n c n
u u U C U :::J
n :::J n :::J ::J n n
C U c: C :::J U
n
:::J C
:J c c n
:::J ::J
Figure 6-10
There are thirteen ways to tile a 2 x 6 board:
n n n nne n U n C ::IU
u u u U U ::J
n c n n c: c n ::J n
U :::J U U ::J :::J c u U
n ::I :::J c n n
C C U :J U U
c n c
::J U ::J
c ::JnC :::JU C l: n
::J ::J U
c c C
::J :::J ::J
Figure 6-11
There are twenty-one ways to tile a 2 x 7 board:
n u n u n u n nu nU nc ::J nne
U U :::J
n
nn::Jn n n :J :J n c n n
u c u U C C U :J U
ncn::J
U ::J C
ncr:::
U :J :J
n n:::Jnn
C U U
n ::J C n n ::J ::J n
c U :J c c U
n:J:J:J C n n n C n n :J
c c c ::J U U ::J U C
c n c n C n ::J n C n ::J ::J
::J U ::J ::J c U ::J C C
c c n n
::J ::J U
r::: :::J r::: :::JnC :J C r:: c
:::J :::J :::J
Figure 6-12
While we are on the topic of checkerboards, there is a cute

problem with a very useful problem-solving message. Suppose you
have an 8 by 8 checkerboard with a pair of opposite-comer squares
missing. (See figure 6-13 .) Can you show how to cover it with 31
dominoes, each of which is capable of covering exactly two adja-
cent squares of the checkerboard?
Figure 6-13
As soon as the question is posed, most people try various ar-

rangements of square covering. This may be done with actual tiles
or by drawing a graph grid on paper and then shading adjacent
squares two at a time. But before long, frustration begins to set in,
since this approach cannot be successful.
Here the issue is to go back to the original question. A careful
reading of the question reveals that it does not say to do this tile
covering; it asks only if it can be done. Yet, because of the way we
have been trained, the question is often misread and interpreted as
"do it."
A bit of clever insight helps. Ask yourself this question: When
a domino tile is placed on the checkerboard, what kind of squares
are covered? A black square and a white square must be covered
by each domino placed on the checkerboard. Are there an equal
number of black and white squares on the truncated checkerboard?
No! There are two fewer black squares than white squares. There-
fore, it is impossible to cover the truncated checkerboard with the
31 domino tiles, since there must be an equal number of black and
white squares. Asking the right questions as you begin to solve a
problem and inspecting the problem posed are essential aspects of
being successful in problem solving in mathematics.
Fibonacci Numbers and Pythagorean Triangles-

Generating Pythagorean Triples
Perhaps the one mathematical relationship that most adults can re-
call is the famous Pythagorean theorem, which established that the
relationship among the sides of a right triangle, whose sides have
lengths a, b, and c, is a 2 + b 2 = c 2 • We may further recall that there
are groups of three numbers, such as 3, 4, 5 and 5, 12, 13 as well as
8, 15, 17, that satisfy this relationship, since 3 2 + 4 2 = 9 + 16 = 25
= 52, 52 + 122 = 25 + 144 = 169 = 13 2, and 8 2 + 15 2 = 64 + 225
= 289 = 17 2 • Such groups of three numbers that satisfy the Py-
thagorean theorem are called Pythagorean triples. When the three
numbers have no common factor,5 such as the three examples
above, then we call them "primitive Pythagorean triples." An ex-
ample of a nonprimitive Pythagorean triple is 6, 8, 10, since all the
members of this group have a common factor of 2.
From the start, we can state that the Fibonacci numbers have
nothing in common with the Pythagorean theorem. They were dis-
covered independently and have no subsequent connection. Yet
much to our amazement, we can use the Fibonacci numbers to
generate Pythagorean triples. Let's see how this can be done. To
make a Pythagorean triple from the Fibonacci numbers, we take
any four consecutive numbers in this sequence, such as 3, 5, 8, and
13. We now follow these rules:
1. Multiply the middle two numbers and double the result.

Here the product of 5 and 8 is 40 then we double this to get
80. (This is one member of the Pythagorean triple.)
2. Multiply the two outer numbers.
Here the product of 3 and 13 is 39. (This is another member
of the Pythagorean triple.)
3. Add the squares of the inner two numbers to get the third
member of the Pythagorean triple.
Here 52 + 8 2 = 25 + 64 = 89.
5. Here and for future references to common factor, we will not consider the
number 1, which is always a common factor.
So we have found a Pythagorean triple: 39, 80, 89. We can

verify that this is, in fact, a Pythagorean triple by showing that
39 2 + 80 2 == 1,521 + 6,400 == 7,921 == 89 2 • (A proof of this delightful
and surprising procedure can be found in appendix B.)
Another Connection between the Fibonacci Numbers

and Pythagorean Triples
There is an established method for generating integer sides for a

Pythagorean triangle-that is, a right triangle whose sides satisfy the
Pythagorean theorem: a 2 + b 2 == c 2 • This method uses the following
set of equations to generate the appropriate values of a, b, and c:
a-=m 2 -n, 2
b -= 2mn, and
c-=m 2 +n, 2
where m and n are natural numbers.

For example, if m -= 5 and n -= 2, then we can get a Pythagorean
triple as follows:
a == m 2 - n 2 == 25 - 4 == 21
b == 2mn == 2 ·5· 2 = 20, and
c == m 2 + n 2 == 25 + 4 = -
29'
which yields the Pythagorean triple (20, 21, 29).

These three equations, sometimes referred to as Babylonian
formulae, were already known to have helped the Babylonians
generate such Pythagorean triples as (3,456; 3,367; 4,825), and
( 12,709; 13,500; 18,541).
Remember, a Pythagorean triple (a, b, c) is called primitive, if
a, b, and c are relatively prime, that is, there is no common factor
among a, b, and c. This occurs when m and n are two relatively
prime natural numbers, with m > n, and of different parity.6
6. Two (integer) numbers have the same parity if they are both even or both odd.
By now you must be wondering where the Fibonacci num-

bers come in. Well, suppose we let m and n take on the values of
Fibonacci numbers sequentially, beginning with m = 2 and n =
1. You will find to your amazement that in each case C turns out
to also be a Fibonacci number, as you can see in Figure 6-14.
k m n a=m 2 -n2 b=2mn c = m 2 + n2

1 2 1 3 4 5
2 3 2 5 12 13
3 5 3 16 30 34
4 8 5 39 80 89
5 13 8 105 208 233
6 21 13 272 546 610
7 34 21 715 1,428 1,597
8 55 34 1,869 3,740 4,181
9 89 55 4,896 9,790 10,946
10 144 89 12,815 25,632 28,657
11 233 144 33,553 67,104 75,025
12 377 233 87,840 175,682 196,418
Figure 6-14
We can easily justify this phenomenon. Consider the general

case for the two consecutive Fibonacci numbers:
Using the formulas from above, we get the following:
2 2
ak =m - n = ~2+2 - Fl+1 ,
bk = 2mn = 2Fk + 2 F k + I, and
Ck = m + n = ~2+2 + ~2+1 F k + 2 + F k +)2,
2 2 2
which according to relationship 9 in chapter 1 (page 43) is always

equal to another Fibonacci number, F;k+3.
Can a Triangle Have Sides with Lengths Represented

by Fibonacci Numbers?
Clearly, you can have an equilateral triangle each of whose sides

have length 5, a Fibonacci number, or you can have an isosceles
triangle with sides 13, 13, and 5-also made up of Fibonacci
lengths. However, if the numbers we select to represent the sides
of a triangle are all different Fibonacci numbers, then the result is
quite different. No such triangle can exist! We can show that a tri-
angle can never have all three sides with lengths represented by
three distinct Fibonacci numbers.
This can be easily established7 by invoking the triangle ine-
quality-namely, for a triangle to exist, the sum of any two sides
must be greater than the third side: hence, for a triangle with side
lengths a, b, and c, the following relationships would be true:
a + b > c, b + c > a, and c + a > b
For any three consecutive Fibonacci numbers, Fn + Fn+l = F n+2 ,

and so there can be no triangle with sides having measures F n,
F n + 1, and F n +2 •
In general, consider Fibonacci numbers, F r, F s, and Fr, where
Fr < F.\. -I and F.., +1 < Ft. Since F S - 1 + Fs = Fs+l and Fr < F s- 1, we
have F,. + F.., < F s+1, and since Fs+l < Fr, we have Fr + Fs < Ft.
Therefore, there can be no triangle with sides having measure F r ,
F s , and Ft.
A Multiplication Algorithm Using the Fibonacci

Numbers
It is said that the Russian peasants used a rather strange, perhaps

even primitive, method to multiply two numbers. 8 It is actually
quite simple, yet somewhat cumbersome. Let's take a look at it.
7. V. E. Hogatt Jr., Fibonacci and Lucas Numbers (Boston: Houghton Mifflin,

1969), p. 85.
8. There is also evidence that this algorithm was used by the ancient Egyptians.
We shall begin with a simple case. If we take the product of two

integers, where one of the factors is a power of 2, then the process
becomes trivial as you will see from the example of 65 . 32.9 We
can see that the product of two integers remains constant if one
factor is doubled and the other factor is halved. We do this in the
following table:
65·32 = 65 32
130 16
260 8
520 4
1.040 2
2,080
6S· 32 = 2,080
Figure 6-15
So products in which a factor is a power of 2 are particularly

simple to calculate, since we can read off figure 6-15 that I· 2,080
= 2,080. Thus we have found the product of 65 and 32.
We shall now consider the problem of finding the product of
43·92 . This time we are not using a factor that is a power of 2.
Let's work this multiplication together. We begin by setting up
a chart of two columns with the two members of the product in the
first row (figure 6-16). We place the numbers 43 and 92 at the head
of the columns. One column will be formed by doubling each
number to get the next, while the other column will take half the
number and drop the remainder. For convenience, our first column
(the left-side column) will be the doubling column, and the second
column will be the halving column. Notice that by halving an odd
number such as 23 (the third number in the right-side column) we
get II with a remainder of I and we simply drop the I. The rest of
this halving process should now be clear.
9. Remember 32 is a power of2: 2 s = 32.

43 92
86 46
172 23
344 11
688 5
1,376 2
2,752 1
Figure 6-16
Find the odd numbers in the halving column (the right col-
umn), then get the sum of the partner numbers in the doubling col-
umn (the left column). These are highlighted in bold type. This
sum gives you the originally required product of 43 and 92. In
other words, with the Russian peasant's method, we get
43 ·92 == 172 + 344 + 688 + 2,752 == 3,956
In the example above, we chose to have the first column be the

doubling column and the second column be the halving column.
We could also have done this Russian peasant's method by halving
the numbers in the first column and doubling those in the second.
See figure 6-17.
43 92
21 184
10 368
5 736
2 1,472
1 2,944
Figure 6-17
To complete the multiplication, we find the odd numbers in the
halving column (in bold type), and then get the sum of their partner
numbers in the second column (now the doubling column). This
gives us 43·92 = 92 + 184 + 736 + 2,944 = 3,956.
Obviously, you are not expected to do your multiplication in
this high-tech era by copying the Russian peasant's method. How-
ever, it should be fun to observe how this primitive system of
arithmetic actually does work. Explorations of this kind are not

only instructive but also entertaining.
Here you see what was done in the above multiplication algo-
rithm. 1o
43 . 92 = (21 . 2 + 1)· 92 = 21· 2 . 92 + 1· 92 = 21· 184 + 92 = 3,956
21 . 184 = (10 . 2 + 1)· 184 = 10 . 2 . 184 + 1· 184 = 10· 368 + 184 = 3,864
10 ·368 = (5 . 2 + 0)- 368 = 5 - 2 ·368 + O· 368 = 5· 736 + 0=3,680
5 . 736 = ( 2· 2 + 1)· 736 = 2 . 2 . 736 + 1· 736 = 2· 1,472 + 736 = 3,680
2 . 1,472 = (1 - 2 + 0)- 1,472 = 1 - 2· 1,472 + O· 1472 = 1- 2,944 + 0 = 2,944
1 . 2,944 = (0 - 2 + 1)· 2,944 = 0 - 2 ·2,944 + 1- 2944 = 0 + 2,944 = 2,944
3,956
For those familiar with the binary system (i.e., base 2), one can
also explain this Russian peasant's method with the following rep-
resentation.
43·92 = (1· 2 5 + 0.2 4 + 1· 2 3 + 0.2 2 + 1· 21 + 1· 2°).92
1 3
= 2° ·92 + 2 .92 + 2 ·92 + 2 ·92
5
= 92 + 184 + 736 + 2,944

= 3,956
Whether or not you have a complete understanding of the dis-
cussion of the Russian peasant's method of multiplication, you
should at least now have a deeper appreciation for the multiplica-
tion algorithm you learned in school, even though most people to-
day multiply with a calculator. There are many other multiplication
algorithms, yet the one shown here is perhaps one of the strangest.
It is through this strangeness that we can appreciate the powerful
consistency of mathematics and that allows us to conjure up such
an algorithm. Now let us bring in the Fibonacci numbers. With the
aid of these numbers, we can develop an analogous multiplication
algorithm. The Fibonacci multiplication algorithm will use addi-
10. We have defined an algorithm as a step-by-step problem-solving procedure, espe-

cially an established, recursive computational procedure for solving a problem in a finite
number of steps.
tion instead of doubling. Let us use the same two numbers that we
used for the multiplication example above: 43· 92 .
This time we will place the 92 at the top of the right-side column
(figure 6-18). On the left side we will begin with the second 1 of the
Fibonacci sequence and list the members of the sequence until we
reach the Fibonacci number that is just less than the other number to
be multiplied-in this case, 43. So we list the Fibonacci numbers to
34, since the next Fibonacci number, 55, would exceed 43.
In the right-side column, we double 92 to get the next number,
and then add every pair of numbers to get the succeeding one (just
as we do with the Fibonacci numbers). We now select the numbers
in the left-side column that will give us a sum of 43 (the original
multiplier). We have selected 1 + 8 + 34 = 43. The sum of their
partners in the right-side column is our desired product:
92 + 736 + 3,128 = 3,956
1 92
2 184 = 92+ 92
3 276 = 184 + 92
5 460 = 276 + 184
8 736 = 460+ 276
13 1,196 = 736 + 460
21 1,932 = 1,196 + 736
34 3,128 = 1,932 + 1,196
Figure 6-18
This may not be the algorithm of choice, but it is fascinating to

observe how our Fibonacci numbers can even enable us to do
multiplication.
Using the Fibonacci Numbers to Convert Kilometers

to, and from, Miles
Most of the world uses kilometers to measure distance, while the

United States still holds on to the mile. This requires a conversion
of units when one travels in a country in which the measure of
distance is not the one we are accustomed to. Such conversions can
be done with specially designed calculators or by some "trick"
method. This is where the Fibonacci numbers come in. Before we
discuss the conversion process between these two units of measure,
let's look at their origin.
The mile derives its name from the Latin for thousand: mille. It
represented the distance that a Roman legion could march in 1,000
paces (which is 2,000 steps). One of these paces was about 5 feet,
so the Roman mile was about 5,000 feet. The Romans marked off
these miles with stones along the many roads they built in
Europe-hence the name, "milestones"! The name statute mile
goes back to Queen Elizabeth I of England, who redefined the mile
from 5,000 feet to 8 furlongsll (5,280 feet) by statute in 1593.
The metric system dates back to 1790 when the French
National Assembly (during the French Revolution) requested that
the French Academy of Sciences establish a standard of measure
based on the decimal system, which they did. The unit of length
they called a "meter," derived from the Greek word, metron, which
means measure. Its length was determined to be one ten-millionth
of the distance from the North Pole to the equator along a meridian
going near Dunkirk, France, and Barcelona, Spain.12 Clearly, the
11. A furlong is a measure of distance within Imperial units and US customary

units. Although its definition has varied historically, in modem terms it equals
660 feet, and is therefore equal to 201.168 meters. There are eight furlongs in a
mile. The name furlong derives from the Old English wordsfurh (furrow) and
lang (long). It originally referred to the length of the furrow in one acre of a
ploughed open field (a medieval communal field which was divided into strips).
The term is used today for distances horses run at a racetrack.
12. There are three fascinating books on this subject: Dava Sobel, Longitude
(New York: Walker & Co., 1995); Umberto Eco, The Island of the Day Before
(New York: Harcourt Brace, 1994); and Thomas Pynchon, Mason & Dixon
(New York: Henry Holt, 1997).
metric system is better suited for scientific use than is the

American system of measure. By an act of Congress in 1866 it
became "lawful throughout the United States of America to
employ the weights and measures of the metric system in all
contracts, dealings, or court proceedings" (though it has not been
used that often). And there is no such law establishing the use of
our mile system.
Now to convert miles to and from kilometers, we need to see
how one mile relates to the kilometer. The statute mile (our usual
measure of distance in the United States today) is exactly
1,609.344 meters long. This translated into kilometers is 1.609344
kilometers. One the other hand, one kilometer is .621371192 mile
long. The nature of these two numbers (reciprocals that differ by
almost 1) might remind us of the golden ratio, which is approxi-
mately 1.618 and has a reciprocal of approximately .618. Remem-
ber, it is the only number whose reciprocal differs from it by ex-
actly 1. This would tell us that the Fibonacci numbers, the ratio of
whose consecutive members approaches the golden ratio, might
come into play here.
Let's see what the equivalent of five miles would be in kilo-
meters:
5 . 1.609344 == 8.04672 ::::: 8
We could also check to see what the equivalent of eight kilo-

meters would be in miles:
8 .. 621371192 == 4.970969536 ::::: 5
This allows us to conclude that approximately five miles is equal

to eight kilometers. Here we have two of our Fibonacci numbers.
The ratio of a Fibonacci number to the one before it is ap-
proximately l/J. Therefore, since the relationship between miles
and kilometers is very close to the golden ratio, they appear to be
almost in the relationship of consecutive Fibonacci numbers. Using
this relationship, we would be able to approximately convert thir-
teen kilometers to miles. Replace 13 by the previous Fibonacci
number, 8, and thirteen kilometers is about eight miles. Similarly,
five kilometers is about three miles and two kilometers is about

one mile. The higher Fibonacci numbers will give us a more accu-
rate estimate, since the ratio of these larger consecutive numbers
gets closer to tP.
Now suppose you want to convert twenty kilometers to miles.
We have selected 20, since it is not a Fibonacci number. We can
express 20 as a sum of Fibonacci numbers,l3 convert each num-
ber separately, and then add them. Thus, 20 kilometers == 13 kilo-
meters + 5 kilometers + 2 kilometers, which, by replacing 13 by 8,
5 by 3, and 2 by 1, is approximately equal to twelve miles.
To use this process to achieve the reverse, that is, to convert
miles to kilometers, we write the number of miles as a sum of Fibo-
nacci numbers and then replace each by the next larger Fibonacci
number. Let's convert twenty miles to kilometers. Thus, 20 miles
== 13 miles + 5 miles + 2 miles. Now, replacing each of the Fibonacci
numbers with the next larger in the sequence, we get 20 miles ~ 21
kilometers + 8 kilometers + 3 kilometers == 32 kilometers.
13. Representing Numbers as Sums of Fibonacci Numbers: It is not a trivial

matter for us to conclude that every natural number can be expressed as the sum
of other Fibonacci numbers without repeating anyone of them in the sum. Let's
take the first few Fibonacci numbers to demonstrate this property:
n The Sum of Fibonacci Numbers equal to n

I I
2 2
3 3
4 1+3
5 5
6 1+5
7 2+5
8 8
9 1+8
lO 2+X
II 3+X
12 1+3+8
13 13
14 1 + 13
15 2 + 13
16 3 + 13
You should begin to see patterns and also note that we used the fewest number
of Fibonacci numbers in each sum in the table above. For example, we could
also have represented 13 as the sum of 2 + 3 + 8, or as 5 + 8. Try to express
larger natural numbers as the sum of Fibonacci numbers. Each time ask yourself
if you have used the fewest number of Fibonacci numbers in the sum. It will be
fun to see the patterns that develop. By the way, Edouard Zeckendorf
(1901-1983) proved that each natural number is a (unique) sum of nonconsecu-
tive Fibonacci numbers.
There is no need to use the Fibonacci representation of a number

that uses the fewest Fibonacci numbers. You can use any combination
of numbers whose sum is the number you are converting. For in-
stance, forty kilometers is two times twenty kilometers, and we have
just seen that twenty kilometers is twelve miles. So forty kilometers is
two times twelve miles, or twenty-four miles (approximately).
Fibonacci Numbers in Physics
The field of optics provides us with a nice application of the

Fibonacci numbers. 14 Suppose you place two glass plates face to
face (figures 6-19 and 6-20) and wish to count the number of pos-
sible reflections. This might be achieved by intermittently covering
a surface, say, at the back side. For convenience, we shall number
the various surfaces of reflection as in figure 6-20.
The first case, where there are no reflections, is where the light
source passes right through both plates of glass (figure 6-21).
There is one path the light ray takes.
Figure 6-21
14. This problem was posed by L. Moser and M Wyman, "Problem B-6," Fibo-
nacci Quarterly 1, no. 1 (February 1963): 74, and solved by L. Moser and J. L.
Brown, "Some Reflections," Fibonacci Quarterly I, no. 1 (December 1963): 75.
The next case is where there is one reflection. There can be

two possible paths that the light can take in this case. It can reflect
off the two surfaces shown in figure 6-22.
Figure 6-22
Suppose, now, that the ray of light is reflected twice within this
set of two plates of glass. Then there will be three possible reflec-
tions (figure 6-23).
Figure 6-23
If the light ray is reflected three times, then there will be five
possible paths that the ray can take-as seen in figure 6-24.
Figure 6-24
By now you will surely anticipate that the next case, that of the
light ray taking four reflections, will result in eight possible
paths-as seen in figure 6-25.
Figure 6-25
You may be tempted to generalize this situation to the

Fibonacci numbers. And, if you do, you are, in fact, right! How-
ever, you may want to be able to generalize this with a mathemati-
cal argument. Let's try to justify this in a mathematical fashion.
1 23 4
Figure 6-26
Suppose the last reflection occurs at face 1 or 3, as it did with

the even number of reflections. The previous reflection had to be
off faces 2 or 4. (figure 6-26).
If the last reflection, of n reflections, was off face 1, then there
were n - 1 reflections prior to this one, or n - 1 possible paths (see
figure 6-26). If the last reflection (of n reflections) was off face 3,
then the previous reflection-the (n - 1)st-must have occurred
off face 4, and must have had n - 2 reflections prior to this one.
And so there were n - 2 possible paths. Therefore, since both re-
flections-off face 1 or 3-can be considered the last reflection,
we get the number of paths as (n - 1) + (n - 2), which would indi-
cate that F" is the number of paths. This began with zero reflec-
tions giving us 1 path, then one reflection, which gave us 2 paths,

and three reflections, which gave us 3 paths. This implies that Fn+2
indicates the number of paths for n reflections. Again the Fibo-
nacci numbers appear!
Last-Digit Patterns
We mentioned in chapter 1 that there is a pattern among the final

digits of the Fibonacci numbers. Namely, they have a period of 60.
That is, the first 60 Fibonacci numbers, ~ - F60 , have the same
last digits (or sometimes called terminal digits), respectively, as the
second 60 Fibonacci numbers, ~l - ~20' the third 60 Fibonacci
numbers, F;21 - F;80 ' the fourth 60 Fibonacci numbers, F;81 - F;40'
and so on for each group of 60 Fibonacci numbers. (You can check
this with the list of the first 500 Fibonacci numbers provided for
you in appendix A.)
Suppose we now consider the last two digits of each of the
Fibonacci numbers (which are below in bold).
00,01,01,02,03, 05,08,13,21,34,55,89,144,233,377,610,987,
1,597,2,584,4,181,6,765, 10,946, 17,711,28,657,46,368,75,025,
121,393, ...
If we look at the 300th Fibonacci number in appendix A, we

find the last two digits repeating beginning with the 301 st
Fibonacci number. Hence, this period is 300 numbers long. Yes, if
we were to look at ~OO' we would find the pattern continuing. You
might be speculating about the period of repetition of the last three
digits of the Fibonacci numbers. Well, we won't make you struggle
with this computation. For the next few groups of terminal digits,
we can conclude the following:
• The period of the last three digits of the Fibonacci numbers

repeating is 1,500.
• The period of the last four digits of the Fibonacci numbers
repeating is 15,000.
• The period of the last five digits of the Fibonacci numbers
repeating is 150,000.
If you do have the time, you can verify this. Or else you can
take our word for it.
The First Digits of Fibonacci Numbers
Now that we have inspected the relationship among the terminal digits
of the Fibonacci numbers, let's turn to the initial digits. One would
assume that there would be an even distribution of the nine numerals
among the initial digits of the Fibonacci numbers. If you take the time
to inspect the first five hundred Fibonacci numbers in appendix A, you
will find that the most frequent numeral is 1, and the least frequent is
9. Figure 6-27 provides a summary of the frequency of the initial digits
of the first one hundred Fibonacci numbers.
Initial Frequency
di2it (Percent)
1 30
2 18
3 13
4 9
5 8
6 6
7 5
8 7
9 4
Figure 6-27
Quite surprising! The reason lies in the fact that Fibonacci num-
bers (as you will see in chapter 9 when we investigate the Binet
formula) are expressible as powers of J5. This, in part, also ac-
counts for the uneven distribution of the Fibonacci numbers
throughout the list of natural numbers. What does this actually

mean? Within the first one hundred natural numbers (i.e., the num-
bers from 1 to 100), there are eleven Fibonacci numbers. Amazingly
enough, among the next hundred natural numbers (from 101 to 200)
there is only one Fibonacci number (144). As we look further down
the list of natural numbers, we find that among the next three hun-
dred natural numbers there are only two Fibonacci numbers, and
further there are only two Fibonacci numbers between 500 and
1,000 (i.e., 610 and 987). This paucity continues, and you can see it
clearly in appendix A. Who would have expected this strange distri-
bution! This is even made more dramatic when seen through a
graph. The visualization shows how fast (and steeply) the values
grow. The eighteenth Fibonacci number (F I8 == 2,584) still can be
read on the graph in figure 6-28 the nineteenth Fibonacci number
(F 19 == 6,765) is far off the scale shown in figure 6-28.
2500 2500
2000 2000
1500 1500
1000 1000
500 500
...
20 20
Figure 6-28
A Fibonacci Curiosity of Digits 15
Here is a strange "coincidence" (or is it one?) that seems to occur

for some of the Fibonacci numbers. And it is that the sum of the
digits of these Fibonacci numbers, F n , is equal to n. The table in
figure 6-29 lists the Fibonacci number where this relationship
holds true.
n Fn Sum of the Digits
0 0 0
1 1 1
5 5 5
10 55 10
31 1,346,269 31
35 9,227,465 35
62 4,052,739,537,881 62
72 498,454,011,879,264 72
Figure 6-29
This is just a curiosity to further entice you to appreciate these

intriguing numbers.
Examining Relationships between Successive

Fibonacci Numbers
While we are examining the characteristics of the members of the

Fibonacci sequence, we might look at the relationship between two
15. Leon Bankoff, "A Fibonacci Curiosity," Fibonacci Quarterly 14, no. 1
(1976): 17.
consecutive Fibonacci numbers. Of course, by this point, you know

that the ratio of two consecutive Fibonacci numbers gives us a
good approximation of the golden ratio. However, there are other
relationships worth noting. A quick look at appendix A will verify
that no two consecutive Fibonacci numbers are even. Another way
of saying this is that they do not have a common factor of 2. As a
matter of fact, you will notice that every third Fibonacci number is
even, or that there are two odd Fibonacci numbers between every
two consecutive even Fibonacci numbers.
Looking at the Fibonacci numbers that have a factor of 3, you
will find they are ~,~,F;2,F;6' ... . Every fourth Fibonacci
number has a factor of 3. Although we already noted this in chap-
ter 1, we can enhance this by stating that no two consecutive
Fibonacci numbers have a common factor. (We call two numbers
that have no common factor "relatively prime numbers.") We can
do a simple proof of this fact here.
• If x and y have a common factor, then it must also be a fac-

tor ofx + y.
• If x and y have a common factor, then it is also a factor of
y-x.
• If x and y have no common factor, then y and x + y also
have no common factor. This is because if y and x + y had a
common factor, then their difference, y - (x + y), which is
just equal to -x, would also have a common factor, but that
is absurd!
• Therefore, if the first two members of a sequence are rela-
tively prime, then all pairs of consecutive members of the
sequence are relatively prime.
• We know that F:. and F2 are relatively prime (since they
have no common factor).
• Therefore, all the members of the sequence Fn and Fn+1 are
relatively prime.
This little proof is a fine example of how a mathematical truth is

established.
A Cute Curiosity!
Consider the four consecutive Fibonacci numbers F 7 , F g , F 9 , and

FlO, or specifically, 13,21,34, and 55. Now let's look at these
numbers in another form. We will break each one down into its
prime factors: (13), (7,3), (2,17), and (5,11). Ifwe rearrange this
list of factors in ascending order, we get: 2, 3, 5, 7, 11, 13, and 17,
which are the first seven prime numbers. The product of these
numbers is 510, which just happens to be the Library Dewey
Decimal classification number l6 for mathematics! This is a coinci-
dental fun fact!
Special Fibonacci Numbers
Those numbers representing dots that can be arranged to form an

equilateral triangle are called triangular numbers. As shown below,
the first few triangular numbers are 1,3,6, 10, 15,21, and 28.
28
15 •
21
••
•
10 • •• •••
3 •
6 •
••
•
•••
• •••
••••
••••
•••••
• ••• ••••• •••• ••• ••••••••• •••••
••••••
••••••
•••••••
Figure 6-30
A triangular number IS one that is an integer of the form

n(n+l) _
2 ' where n - 1, 2, 3, ....
16. American librarian Melvil Dewey, born Melville Louis Kossuth Dewey, in
Adams Center, New York, on December 10, 1851, is best known as the inventor
of what came to be called the "Dewey Decimal System of Classification," which
is used in most local and school libraries to catalog books. Devised in 1876 as a
system for small libraries, it has the advantage of a limited number of general
categories and short call numbers. The system is based on ten classes of subject
(000-999), which are then further subdivided. Dewey also promoted the use of
the metric system, and helped found the American Library Association in 1876.
When Dewey created Columbia University's School of Library Economy in
1887, he began the field of library science in the United States.
However, there are only four Fibonacci numbers that are trian-
gular numbers. They are 1, 3, 21, and 55.
There are only three Lucas numbers that are triangular num-
bers. They are 1, 3, and 5,778. So even in the realm of triangular
numbers, the Fibonacci numbers make an appearance.
The thirty-sixth triangular number 17 is 666, a number that has
captured the popular imagination for a long time. (Many associate
it with the devil.) The sum of the squares of the first seven prime
numbers is 666. That is, 22 + 32 + 52 + 7 2 + 112 + 132 + 172 == 666.
This unusual number also offers some nice number relation-
ships, such as:
666 == 16 - 2 6 + 36
666 == 6 + 6 + 6 + 6 3 + 6 3 + 6 3
666 == 2 . 3 . 3 . 37, and 6 + 6 + 6 == 2 + 3 + 3 + 3 + 7
And just for the sake of entertainment on the number 666, con-
sider the first six Roman numerals written in order: DCLXVI,
which just happens to equal 666.
Of course, this number, 666, is also related to the Fibonacci
numbers.
And the analog for the subscripts:
1 - 9 + 11 + 15 = 6 + 6 + 6
Similarly for the cubes of the Fibonacci numbers:
FI 3+ F2 3 + F4 3 + F5 3 + F6 3 = 1 + 1 + 27 + 125 + 512 = 666
And the analog for the subscripts:
1+2+4+5+6=6+6+6
17. The thirty-sixth triangular number = (36)(36 + 1) = 18·37 = 666.

2
Now let's have this curious number 666 connect to the Fibo-
nacci numbers with the golden ratio-as is fitting to do in this
book:
-2sin( 666°) = 1.6180339887498948482045868343656381177203091798057 ...
a very close approximation of the golden ratio, which is

l/> = 1.6180339887498948482045868343656381177203091798057 ...
Finally, in 1994 Steve C. Wang published the following myste-

rious equation in the Journal of Recreational Mathematics 18-with
the title "The Sign of the Devil":
l/> = - [sin (666°) + cos (6 . 6 . 6 0)]
While earlier we had a fine approximation of the golden ratio,

this equation, then, places the number 666 in an unquestionable
connection with the golden ratio!
The Curiosity of the Eleventh Fibonacci Number
The eleventh Fibonacci number, F..I' is 89. This number has many
unusual characteristics. For example, if you take any number and
find the sum of the squares of the digits, and continuously repeat
this process, you will either end up with 1 or 89. 19 We will show
how this works in a few cases:
18. 26, no. 3 (1994): 201-205.

19. If you take this further you will be led to the number 4, but this is via the
number 89. This would continue as:
8 2 + 9 2 = 145
12 + 4 2+ 52 = 42
4 2 +22=20
22 + 0 2 = 4 ,
which then takes you into a loop back to 89:
4 2 = 16
12 + 6 2 = 37
3 2 + 7 2 = 58
52 + 8 2 = 89
For the number 23:
22 + 32 == 13
12 + 32 == 10
12+0==1
For the number 54:

52 + 4 2 == 41
42+12==17
12+72=50
52 + 0 == 25
22 + 52 == 29
22 + 9 2 == 85
8 2 + 52 == 89
For the number 64:

6 2 +42==52
52 + 22 = 29
22 + 92 = 85
8 2 + 52 == 89
To convince yourself of this peculiarity, try this with some

other numbers. If you are still piqued by this Fibonacci peculiarity,
consider the reciprocal of 89:
1
= 0.01123595505617977528089887640449438202247191 01123595505
89
61797752808988764044943820224719 011235955056179775280898
87640449438202247191 011235955056179775280898876404494382
02247191 ...
= 0.01123595505617977528089887640449438202247191
You will notice it repeats after forty-four decimal places, or as

we say, it has a period of 44. Curiously, at the very middle of this
period of forty-four digits you will notice there is the number 89.
0.01123595505617977528089887640449438202247191
Its unusual Fibonacci property is even more surprising. Notice

the first few digits of the period, 0, 1, 1, 2, 3, and 5, are the first
few Fibonacci numbers. So what happens after that? To see this
continue, consider the following:
1 0 1 1 2 3 5 8 13
- = - 1+ -2+ - + - 4+ - + - + - +7
- + ...
89 10 10 10' 10 10 5 106 10 108
IFF F F F F F F
_=_0 +_1 +_2 +_3 +_4 +_5 +_6 +_7 + ...
89 10 1 10 2 10 3 10 4 105 106 10 7 lOX
This may be seen, perhaps, more concretely when written as

decimals:
o
+.01
+.001
+.0002
+.00003
+.000005
+.0000008
+.00000013
+.000000021
+.0000000034
+.00000000055
+.000000000089
+.0000000000144
+.00000000000233
+.000000000000377
+.0000000000000610
+.00000000000000987
.0112359550561 7977 ...
This will continue to generate the above period of forty-four digits

if we take the decimal approximations further.
But the question might be asked: Why does this happen? We
can do some numerical manipulations that will show how the
Fibonacci numbers tie in with this remarkable number 89-itself

the eleventh Fibonacci number.
We begin with the following equality:
10 2 = 89 + 10 + 1 (I)
We will multiply both sides of (I) by 10:
10 3 = 89 . 10 + 10 2 + 10
and then substitute (I) in the resulting equality to get:
10 3 = 89·10 + 89 + 10+ 1
10 3 = 89·10 + 89 + 2·10 + 1 (II)
Continuing this process, we will multiply both sides of (II) by 10:
104 = 89 . 10 2 + 89 . 10 + 2 . 10 2 + 1 . 10
104 =89.10 2 +89.10+2(89+10+1)+1. 10
and then substitute (I) again, where appropriate to get:
104 = 89 . 10 2 + 89 . 10 + 89 . 2 + 3 . 10 + 2 (III)
Continuing this process gives us the following:
10 5 =89.10 3 +89.10 2 +89.2.10+3.10 2 +2.10

10 5 =89.10 3 +89.10 2 +89.2.10+3(89+10+1)+2.10
105 = 89 . 10 3 + 89 . 10 2 + 89 . 2 . 10 + 89 . 3 + 5 . 10 + 3
105 =89(1.10 3 +1.10 2 +2.10+3)+5.10+3 (IV)
and
106 =89(1.10 4 +1.10 3 +2.10 2 +3.10+5)+8.10+5 (V)

and this can be generalized to:
for all positive integers n.
We then divide both sides by IOn + 1 to get:
lOF +F
We can establish that lim n+l n = 0, the proof of which
n~oo 10n
+
1
can be found in appendix B.
As we conclude our brief discussion of the eleventh Fibo-

nacci number, let us take note of its successor: the twelfth Fibo-
nacci number, F:.2 = 144 = 122 . Aside from the first two numbers
of the Fibonacci sequence, F:. = 1, and F2 = 1 , it is the only Fibo-
nacci number that is also a square number.
Displaying a Watch
What might the Fibonacci numbers have to do with the way a

watch is displayed in a showcase or in an advertisement? Well,
here we are using the geometric form of the Fibonacci numbers,
namely, the golden rectangle. You may notice that advertise-
ments and window displays often show a watch with the hands
set at 10: 1o.
Figure 6-31
The hands make an angle of 19 ~ minutes,20 which is the
equivalent of 155 degrees (figure 6-31). Now consider the rectangle
formed by placing its vertices at points at which the hands indicate
10: 10, that is, at minute markers 2 and 10 ~ . Then fonn a rectangle
(figure 6-32) with these two points as two adjacent vertices and with
the point of intersection of its diagonals at the center of the watch
face. This rectangle is very close to the golden rectangle, whose di-
agonals meet at an angle of about 116.6 degrees (figure 6-33).
A"'____---r'_ _._n__-7iB
•
,
D c
20. Since IO minutes is .!. of an hour, the hour hand moved'!' of the distance
6 6
I
between 10 and 1 1, from 10:00 to 10: I O. Therefore, the hour hand moved - of 5
6
minute markers, or .? of a minute marker. Thus the angle at 10: 10 is 19.!. min-
6 6
ute markers. To find the degree measure of that angle, we simply find the part of
I 115
19- -
60 that 19 .!. is- ...Q.=_6_ =~ which is 115°.
6 60 60 360'
This can be determined by using the golden rectangle (figure

6-33) and finding the measure of the angle whose tangent function is
¢
tanx = FB = 2" = I/J

FE I
2
That angle. x, has measure 58.28 degrees. Therefore,

mL.AEB = 116.56 0 , which is very close to the 115 0 that is felt by
many exhibitors to be the ideal placement of the hands of a
watch. Some might have suspected that the 10: I 0 time place-
ment was to allow the observer to see, in unobstructed fashion,
the brand name of the watch. Not necessarily so. Consider an
advertisement at the Prague airport (figure 6-34), where there
was no need to highlight a brand name. Once again. the time
10: lOis shown.
Prague '''enna Brattslava Wtl'5aw

I , ,, ,' , , ' , 1 ,I,
''V'' "V'
, t·"
,,' ,, . "V'
, ,'
,' ---/ ',---/,
, t .' , ,'
("'1''''\~.~
""'l>In... t_·c
One Central European Exhibition Centre ~'\
Figure 6-34
Take, for another example, a recent United States postage

stamp that exhibits the "American Clock."
Figure 6-35
Here the time shown is about 8:21 Y2, where the hands form an an-
gie of about 121.75°,21 which is not exactly the angle ofthe diagonals
of an ideal golden rectangle. But it is close enough for the naked eye
to appreciate. The golden rectangle seems to have again determined
the ideal!
Some may recall that when they were taught to drive a car,
they were instructed to hold the steering wheel at the positions of
the 10 and 2 of a clock. Perhaps those points present some sort of
special balance. Something to ponder!
2 I. To find the angle formed, we will do it in two parts. First, the angle formed
between the "6," or 30-minute marker, and the 21.5-minute marker L (30 -
21 Y,) = 8.5 . 6° = 5 I 0. Second, the angle formed by the 30-minute marker and
the hour hand at 8:21.5. In 21.5 minutes the hour hand moved 21.5 of the 5 min-
60
ute markers between the 8 and 9 on the clock, or 2~5 of 30°, which is 10.75°.
The angle between the hour hand and the 30-minute marker is 10.75° + 60 ° =
70.75°. The angle formed by the hands of the clock is, then, 51 ° + 70.75° =
121.75°.
Fibonacci Helps Us Determine Seatings
There are often interesting mathematical phenomena that arise

from problems that stem from seating arrangements. We will con-
sider one here in which, once again, our Fibonacci numbers will
emerge. Suppose that from a school full of boys and girls, you will
be placing a number of chairs on a stage and you want to know
how many ways to seat the boys and girls so that no two boys sit
together. (This is not meant to be an indication that if two boys sit
together, they will misbehave!) The chart in figure 6-36 will sum-
marize the various seatings.
Number Arrangements Number of

of Chairs Ways to Seat
Students
1 B,G 2
2 BG, GB, GG, gg 3
3 GGB, ~, BGB, GBG. GGG, 5
ggg,BGG,~,
4 GGGG, ~, QQCC, GCQQ, 8
BGBG, GBGB, CQQC; BGGB,
QQQC, GGGB, CQQQ. BGGG.
QQCQ,GGBG, QCQQ,GBGG
5 GGGGB, QQQQQ, QQCGB, 13
CCQQQ, BGBGB, CQCQQ,
CQQCQ, QCCQQ. QQQCQ,
GCCQQ, CQQQQ, BGGGB,
QQCQQ, GGBGB, QCQQQ.
GBGGB GGGGG, QQQQC,
QQCCC, CCQQC, BGBGG,
GBGBG, CQQCC, BGGBG,
QQQCC,GGGBG,CQQQG,
BGGGG,QQCQC,GGBGG,
QGQQC,GBGGG
Figure 6-36
As you can clearly see, in each case the number of seating pos-
sibilities is a successive Fibonacci number, beginning with 2.
Let us now change the restrictions and insist that no boy or girl
sits without someone from the same sex next to him/her and that
the first position must be a girl. Figure 6-37 displays the various
acceptable arrangements and shows us that the Fibonacci numbers
give the number of ways this seating can be done. With six chairs,
there are five possible seating arrangements.
Number Arrangements Number of

of Chairs Ways to Seat
Students
1 ~ 0
2 Q~, ~Q,GG, gg 1
3 ~,~, BGg,~, GGG, ggg, I
~,~,
4 GGGG, QQQQ, QQ~~,
GGBB, 2
QCQ~, ~QGQ, ~QQ~; QCCR,
QQQC, GcCQ, CQQQ, QCCC,
QQCQ,CCQ~,QCQQ,CQC~
5 ~~CCR, QQQQQ, QQGGQ, GGBBB, 3
QCQCQ, CQCQQ, GQQCQ. QGGQQ,
QQQGQ, GGGBB, GQQQQ, QGGGQ,
QQCQQ,GGQCQ,QCQQQ,GQc~Q
GGGGG, QQQQG, QQCCC,
CCQQC, QCQGG,CQCQC,
CQQCG, QCCQC,QQQGC,
cCCQC, CQQQG, Reccc,
QQCQC,CGQCC,RCQQC,CQCGC
Figure 6-37
Fish in a Hatchery
Consider a fish hatchery that is partitioned into sixteen congruent

regular hexagons-arranged in two rows (figure 6-38)-with
pathways between each adjacent hexagon. 22 A fish begins a jour-
ney at the upper-left side of the double row of hexagons and ends
up at the lower-right hexagon. Our problem is to determine the
number of paths the fish can take to complete his journey to reach
hexagon K, if he can only move to the right.
x z B D F H J
y A E G I K
Figure 6-38
22. This example may not be too realistic, but it provides us with a nice applica-
tion of the Fibonacci numbers.
As the fish begins his travels, there is only one path he can take
to hexagon W (figure 6-39), since if he were to go horizontally first,
he would no longer be traveling to the right to get to hexagon W.
Figure 6-39
To get to hexagon X, there are two paths, one directly to the

right, and one through W (figure 6-40).
Figure 6-40
For the fish to get to hexagon Y, there are three paths: hexa-
gons W-X-Y, X-V, and W-Y (figure 6-41).
"
y
Figure 6-41
Perhaps you will anticipate the next number of possible hexa-

gon paths that the fish can travel when there are five hexagons
available. Yes, there are five paths, as shown in figure 6-42.
They are: W-Y-Z, W-X-Y-Z, X-Z, W-X-Z, and X-Y-Z.
Figure 6-42
It would be correct to assume that this pattern will continue as

we increase the number of hexagons for the fish to traverse. To
summarize, we provide the following chart:
Number of hexagons 1 2 3 4 5 6 7 ... n

for the fish's journey
Number of different 1 1 2 3 5 8 13
paths
Figure 6-43
So we might conclude that for the fish to reach the sixteenth

hexagon, K, we would need to find F 16 , which is 987. Thus, there
are 987 possible paths for the fish to travel from the first hexagon
to the sixteenth hexagon, K-always traveling only to the right.
The Game of Fibonacci Nim
At this point in the chapter, perhaps we shall have the enjoyment

of playing a game where the winning strategy is dependent on the
Fibonacci numbers. The rules of the game are as follows.
The Rules:
• There is one pile of pennies. The first player may remove

any positive number of pennies, but not the whole pile.
• Thereafter, each player may remove at most twice the num-
ber of pennies his opponent took on the previous move.
• The player who removes the last penny wins.
• You move first: you can win from the initial position, but
don't make any mistakes!
The questions before us, as we begin to play the game, are:

Will the number of pennies at the beginning determine which
player will win the game? What strategy should be used in playing
the game? And what might this have to do with the Fibonacci
numbers?
The strategy for this game, as you might have expected, is
based on the Fibonacci numbers. Just as we can use binary notation
to represent every positive integer as a sum of distinct powers of 2,
we can use a similar process to write integers as a sum of Fibo-
nacci numbers (see page 188). The winning strategy for this game
involves removing the smallest of the Fibonacci summands. 23 Here
is a method for doing this: pick the largest Fibonacci number not
exceeding the given integer, then subtract that Fibonacci number
and continue the process with the difference. For example, to write
20 in terms of Fibonacci numbers, subtract 13 from it to get 7.
Subtract 5 from 7 to get 2, which itself is a Fibonacci number.
Thus 20 == 1-13 + 0-8 + 1-5 + 0-3 + 1-2 + 0·1 == 101010. 24
23. A summand is one member of a sum.

24. We are merely representing the number by showing all the Fibonacci num-
bers: those for inclusion in the sum are multiplied by 1 and those not included in
We can express 20 = 13 + 5 + 2, and thus the winning move in-

volves taking two pennies. This is equivalent to taking the right-
most 1 in the Fibonacci representation of the number of pennies in
the pile.
Number Fibonacci repre- The small- Number of Fibonacci

of pen- sentation of the est of the pennies that representation
ntes In number of pen- Fibonacci remain after of the number
pile nies summands the move of pennies that
remain
20 101010 2 18 101000
19 101001 1 18 101000
18 101000 5 13 100000
17 100101 1 16 100100
16 100100 3 13 100000
15 100010 2 13 100000
14 100001 1 13 100000
13 100000 13 0 0
12 10101 1 I1 10100
11 10100 3 8 10000
10 10010 2 8 10000
9 10001 1 8 10000
8 10000 8 0 0
7 1010 2 5 1000
6 1001 1 5 1000
5 1000 5 0 0
4 101 I 3 100
3 100 3 0 0
2 10 2 0 0
1 1 1 0 0
Figure 6-44
Compare the Fibonacci representations in the second column

with those in the fifth column in figure 6-44. Note that the winning
move in this game is equivalent to taking the rightmost 1 in the
Fibonacci representation of the number of pennies in the pile. Less
obvious is that the second player cannot make a similar move.
Try this strategy with a friend. Are there initial-pile sizes for
which the first player cannot be assured of a win with this strategy?
the sum are multiplied by O. Note: We do not use the first I of the sequence and
begin with the second 1.
When Is a Fibonacci Number Equal to

a Lucas Number?
What we seek to determine is when might a Fibonacci number, Fn ,

be equal to a Lucas number, Lm? Let's consider some of the rela-
tionships we have established earlier, which compare these two
numbers.
For m > 3
In the first case, we have that Lm > Fm+ I ' and in the second case
we get: L m < F m+ 2.
For m > 3, we have Fm+l < Lm < F m+ 2 •
Therefore, Fn == Lm only when m < 3.
The only solutions are
F==L==l
I I
F==L==l
2 I
F 4 ==L2 ==3
By this point in the book, you are probably expecting the

Fibonacci numbers to appear, even where they may be least ex-
pected. The applications or sightings of these fabulous Fibonacci
numbers seem to be boundless. Besides these kinds of appear-
ances, the Fibonacci numbers can also surprise us by the many in-
terrelationships they have, as well as those they have with the
Lucas numbers. We will close this chapter with a sample listing of
some of these amazing relationships. Convince yourself that they
hold true by applying them to specific cases, or for the more ambi-
tious reader, try to prove them true.
A Plethora of Fibonacci Relationships
Let k, m, n be any natural numbers; where k, m, n > 1:
LI = 1; L2 = 3; L n+2 = L n + L 11+1
(Fn , F n+l ) = 1. The greatest common divisor of Fn and Fn+1

is 1 (they are relatively prime)
n
L F; = F; + F; + F; + ~ +
i=l
... + ~1 = Fn+2 - 1
n
L F21 = F2 + F4 + F6 + ... + F 21l -2 + F 2n = F21l+1 -
1=1
I
n
L F 2i-
i=1
1 = F) + F3 + Fs + ... + F 2n - 3 + F 2n - 1 = F 2n
n
~
~
F2 =Fn F n+ I
I
1=1
F2 - F2 == F
11 11-2 2n-2
F 11- k F n+ k - F n2 = + Fk2, where n > 1 and k > 1

n+1
~ FF
£...J / /- 1 = F2
n+ I' when n is odd
i=2
n+1
LF;F;-l = F}+I -1, when n is even
i=2
n
L,=1 Li = LI + L2 + L3 + L4 + ... + Ln = Ln+2 - 3
n
~ L2/
£...J = LnLn+ 1 - 2
i=1
F22k = F 2k- 1 F 2k+1 - 1
F=FF
n m n+l-m +Fm-I Fn-m
Fn-I +Fn+1 =Ln
Fn+2 -Fn-2 =Ln
Fn +Ln =2Fn+ 1
F 2n =FL
n n
Fn+l Ln+1 -FL=F
n n 2n+1
Fn+m +(_I)m Fn-m = LmFn

Fn+m - (_I)m Fn-m = Fm Ln
FnLm-LnFm =(-I)m2Fn-m
F n+1 L n =F2n+l -1
3Fn + Ln = 2Fn+ 2
5Fn +3Ln = 2Ln+ 2
Fn+ ILn = F2 n+ 1+1
L n+m + (_I)m Ln-m = L mLn
L2n + 2( _1)n = L2n
2
L4n - 2 = 5F2 n
L4n + 2 = L~n
Ln_J+ Ln+J = 5Fn
LmFn +LnFm = 2Fn+m
Ln+m -(-I)mLn-m =5FF
m n
L2n -2L2n = -5F2n
L2n -2(-I)n=5F2 n
Ln = Fn+2 + 2Fn_1
Ln = Fn+2 - Fn_2
Ln = LJFn + LOFn_J
Ln-J L n+J + Fn-I Fn+1 = 6Fn2
F3n+1 + F3n - F3n-J = 3F3n
Chapter 7
The Fibonacci Numbers Found

in Art and Architecture
T he golden ratio seems to be ever present in both art and architec-

ture. Why this is, we can only speculate. We have already estab-
lished that it has been proclaimed the most beautiful ratio of all
time-both numerically and geometrically. Here, we will provide a
brief survey of its apparent sightings in these two aesthetic fields. And
we will look at the visual art in both two and three dimensions.
During the Renaissance, the study of proportions in architec-
ture, sculpture, and painting as an aesthetic mathematical device
was based on the notion that harmony and beauty in art can best be
determined through particular numbers, often seen in their rela-
tionships with other numbers. This is how the golden section came
into prominence. There are countless articles and books that at-
tempt to show that outstanding art and architecture derived from,
in one form or another, the principle of the golden section.
The golden section, whose ratio is also considered beautiful,
was already known in antiquity and by the Pythagoreans when they
constructed a regular pentagon and a regular dodecahedron. Many
architects through the ages, intuitively or deliberately, used the
golden section in their sketches and construction plans, either for
the entire work or for the apportionment of parts of the work. In
particular, this aesthetic bias frequently manifested itself as the
231
golden rectangle. Since the golden ratio, tjJ = J5 +1 , is an irrational

2
number, architects were known to use lengths of the sides of a

rectangle whose ratio approximated l/> . And, as you now know, the
most common approximations for l/> are the Fibonacci numbers.
Moreover, the larger these numbers, the closer the ratio is to l/>.
Fibonacci Numbers in Architecture
For years, architects have used the golden ratio, l/>, in their draw-
ings, sometimes as the golden rectangle, other times as a partition
indicator. To circumvent the difficulty of using the irrational num-
ber (l/», the golden rectangles that were constructed typically in-
volved dimensions that were Fibonacci numbers. The golden ra-
tio-an ideal-was very nearly achieved, since, as we already
know, the ratio of two consecutive Fibonacci numbers approaches
the golden ratio, with the greater the magnitude of the Fibonacci
numbers, the closer the approximation of l/J.
Perhaps the most famous structure that exhibits the golden
rectangle is the Parthenon on the Acropolis in Athens. The artistic
concept for building this magnificent structure came from Phidias
(460-430 BeE). It was erected to thank Pericles (500-429 BeE)
for saving Athens during the Persian War (447-432 BeE). The
purpose of this temple was to house the statue of the goddess
Athena. It is believed that the current designation for the golden
ratio, l/>, came from the first letter of Phidias's name (in Greek:
<I>~IAL). We must stress at this point that we have no evidence
today that Phidias was aware of the golden ratio; we can only as-
sume it from the shape of the structure. As you can see in figure
7 -1, the Parthenon fits nicely into a golden rectangle. Further-
more, in figure 7 -1, you will notice a number of golden rectan-
gles, generated from the structure of the building.
THE FIBONACCI NUMBERS FOUND IN ART AND ARCHITECTURE 233
Figure 7-1
4>
........
1111111111111111111111111111 -
i i
Figure 7-2
If one takes measurements of various ornaments above the col-

umns, the golden ratio keeps cropping up. Figure 7-2 exhibits the
golden ratio throughout the Parthenon. To what extent ancient ar-
chitects consciously used the golden rectangle is still largely a
mystery. Some believe that we moderns are just looking for ways
to superimpose a golden rectangle over a specific structure-

though no one can deny its presence.
Another example of the golden section can be seen on
Hadrian's Arch on the Propylaeum temple, also situated on the
Acropolis. The arrangement of snails and other ornaments at the
top of the arch further exhibits the golden ratio through a regular
pentagon that can be superimposed as in figure 7-3 . (Notice the M
represents the major portion of the golden section, while the m rep-
resents the minor portion.)
E
~
....
~
~
I~
.,
~
E co
0
Figure 7-3
Figure 7-4
(Co pyright Wolfgang Randt. Used with permission.)
According to the Greek historian Herodot (ca. 490-425 BCE),

the Khufu (Cheops) Pyramid at Giza was constructed in such a
way that the square of the height is equal to the area of one of the
lateral sides. Using our superimposed triangle in figure 7-4 (and
the Pythagorean theorem), we get the following:
2
2 a 2
hI!,. =-+ hp
4
The area of one of the lateral triangles is A = a . hI!,..

2
According to the construction principle introduced by Herodot
(above), where the square of the height is equal to the area of the
lateral triangle, we get:
2
2 a 2 a
h p = hI!,. --=A
4 2 hI!,. .
= -.
2
Ifwe divide both sides of the equation a hI!,. = h~ -~ by a hI!,.'
2 4 2
we get:
a
_ hI!,. 2
1----
a
-
h I!,.
2
By making the following substitution of h"

a
= x by first changing
2
a
- 1 1
its form to ...L = -, we get the equation: 1 = x - -, which just hap-
ht! x x
pens to generate the equation x 2 - x - 1 = 0, whose solution is al-
ready known to us as XI = </J and x 2 = -.!. . By now you have realized

l/J
that x 2 is negative and holds no real meaning for us geometrically; so
we won't consider it here. Using today's measurement capabilities,
this great pyramid has the following dimensions.
Cheops Length of Height of Pyramid h/),. C

pyramid the side of lateral height - --
the base: triangle: hp a 2hp
-
a hI). 2
measurements 230.56 m 186.54 m 146.65 m 1.61813471 3.144357313
Figure 7-5
Lo and behold, the ratio of the height of the lateral triangle to

h
half its base is 2.
a
= 1.61813471.
2
Was this done intentionally by the design of an ingenious ar-
chitect? This is a question that you'll have to answer for yourself.
We can only point out what has been found by measurement and
historical clues.
We assume that the Egyptians did their measurements with cu-
bits, which was defined as the length of a man's forearm (i.e., from
fingertip to elbow) and measured as a length of 52.25 cm. A Brit-
ish Egyptologist, W. M. F. Petrie (1853-1942), established meas-
urements for this famous pyramid as follows:
; = 220 cubits (= 114.95m)[i.e. a=440cubits=229.90m]

hp = 280 cubits ( ::::: 146.30m)
h/),. = 356 cubits (::::: 186.01m)
The angle formed by the lateral height ( h/),. ) and the base yields
a cosine measure, which is a quite-surprising result: cosL( ~, h/),.)
2
= ~~~ = :~ . You would almost think that there is some sort of "fix"
in this analysis. However, we are just reporting the facts as they
have been recorded.
From the above, we can speculate that the base of the pyramid,
whose side is 230.56 m, has a perimeter of 922.24 m. If we divide
this perimeter by twice the height of the pyramid, we get, curiously
enough, an approximation for one of the most famous numbers in

mathematics, n. 1
Spectators have noted the golden section on Mexican pyra-
mids, Japanese pagodas, and even Stonehenge (ca. 2800 BCE) in
England. There is, moreover, speculation that the Arch of Triumph
in ancient Rome was built on the basis of the golden ratio.
The oldest stone structure in Germany, stemming back to the
post-Roman period (770 CE), is the "Konigshalle" in the town of
Lorsch (figure 7-6). This magnificent example of architecture from
the early Middle Ages has an impressive open-air ground floor.
The interior space of this building exhibits practically a perfect
golden rectangle. We still don't know today what purpose this
building had and how it was used. Perhaps if one day the building
plans could be found, we might be able to see if the dimensions
might have intentionally been Fibonacci numbers.
Figure 7-6
The Cathedral of Chartres, France (built from 1194-1260),

clearly exhibits the golden rectangle. In figure 7-8, you can see the
front portal, and figure 7-7 shows a window from the Cathedral of
Chartres exhibiting this famous rectangle.
I. For more on It, see Alfred S. Posamentier and Ingmar Lehmann, 11:: A Biogra-
phyofthe World's Most Mysterious Number (Amherst, NY: Prometheus Books,
2004).
Figure 7-9
Figure 7-10
Throughout the Renaissance, the designs of many construction

projects used the Fibonacci numbers or the golden ratio. They can
be found in the plans of the dome of the Santa Maria del Fiore
Cathedral in Florence (see figure 7-9). The rough sketch by Gio-
vanni di Gherardo da Prato (1426) exhibits the Fibonacci numbers,
55, 89, and 144, as well as 17 (which is half the Fibonacci number
34) and 72 (which is half the Fibonacci number 144). The dome
was actually constructed in 1434 by Filippo Brunelleschi
(1337-1446) and has a height of 91 m and a diameter of 45.52 m,
which (sadly) does not give us the golden ratio.
The remainder of the Cathedral in Florence further demon-
strates the Fibonacci numbers. But the most dramatic proportions
are those shown in figure 7 -11, where the ratio of 89: 5 5
(== 1.6181818 ... ) has an extremely close approximation to the
golden ratio 1.618033988 ....
"
144 •
II
~------------~
Tf,7 .
i89
-*..
Figure 7-11
Let ' s move to more modern times. Le Corbusier (1887-

1965),2 was a French-Swiss architect who between 1946 and 1952
designed the unites d'habitation in Marseille, France (see figure
7-12), which again exhibits the golden ratio. Here it is exhibited in
the tower, which divides the remainder of the building into the
golden ratio. This was not done by chance, but rather by specific
design! He displays the golden section throughout.
Figure 7-12
2. His actual name was Charles-Edouard Jeanneret.

Furthermore, all of the measurements in the apartments are

based on a theory of the proportions that Le Corbusier developed
in 1948, and published in a book titled The Modulor: A Harmoni-
ous Measure to the Human Scale Universally Applicable to Ar-
chitecture and Mechanics. 3 This stringent rule caused somewhat
of a furor in the art world. Le Corbusier developed his scheme on
the basis of the golden section. This included his decision that the
height of a door was to have a measure of 2.26 m (7.4 feet),4 so
that a person of height 1.83 m (6 feet) can touch the top with out-
stretched arms. Le Corbusier supported his ideas with the fol-
lowing: "A man with a raised arm provides the main points of
space displacement-the foot, solar plexus, head, and fingertip of
the raised arm-three intervals which yield a number of golden
sections that are determined by Fibonacci."
The Modulor (see figure 7-13), a model of how the human pro-
portions should be standardized for use by architects and engi-
neers, consists of two scales or bands, which are marked with two
sequences of natural numbers-the sequence:
(*) 6,9,15,24,39,63,102,165,267,432,698,1130,1829
and an intermediate sequence:
(**) 11,18,30,48,78,126,204,330,534,863,1397,2260
Usually, at second glance, we can see the relationship to the

golden section, l/J.
3. Le Corbusier, The Modulor: A Harmonious Measure to the Human Scale

Universally Applicable to Architecture and Mechanics and Modulor, 2 vols.
(Basel: Birkhauser, 2000).
4. 1 meter = 3.28 feet.
1.829
1.130 2.260
698 1.397
432 863
267 534
165 330
102 204
63 126
39 78
24 48
15 30
9 18
6 II
(*) ( •• )
Figure 7-13
If we divide the members of the sequence (*) by 3 and those of

the sequence (**) by 6, we get:
(*), 2,3,5,8,13,21,34,55,89, 144,232 ~ , 376f, 609 ~
and
- - - -
(**)' 1.83, 3, 5, 8, 13,21, 34, 55, 89, 143.83, 232.83, 376. 83
If we round off the fractional members of each sequence, we

are then dealing exclusively with the Fibonacci numbers.
This relationship to the Fibonacci numbers results from the fact
that Le Corbusier's sequences almost approximate a geometric
progression with the common ratio of the golden section, rp. In his
Modular, Le Corbusier claimed to have found a scale that adapted
exceptionally well to the visual arts.
Le Corbusier was substantially influenced by the ideas of the
Romanian lawyer, engineer, and diplomat Matila Costiescu Ghyka
(1881 - 1965),5 whose writing about the golden ratio connected the
religious aspects of Luca Pacioli (1445- 1517)6 with the aesthetic
writings of Adolf Zeising (1810- 1876).7 Ghyka interpreted the
golden section as a fundamental secret of the universe and sup-
ported this position with examples from nature. This perspective
further inspired Le Corbusier, whose aim was to create a harmonic
design over a specific living space. He seized on the same philoso-
phy in creating resident apartments.
Figure 7-14
In 1947 Le Corbusier was a member of the architectural commis-

sion that planned the building of the UN headquarters in New York
(see figure 7-14). He was ultimately responsible for the basic concept
of the plans for the thirty-nine-story Secretariat building. As you
might expect, the height and the width of the building are close to the
ratio of 1.618: I. Yes, once again the golden ratio stares us in the face!
Last, but not least, an obvious exhibit of the golden ratio can be
found in an aerial view of the Pentagon in Washington, DC (see
5. Matila Ghyka, Esthetique des proportions dans la nature et dans les arts (Paris,
1927); Matila Ghyka, The Geometry oj Art and Life (New York: Sheed and
Ward, 1946; reprint, New York: Dover Science Books, 1977).
6. Fra Luca Pacioli, Divina Proporzione- A Study oj the Golden Section First
Published in Venice in 1509 (New York: Abaris Books, 2005).
7. Adolf Zeising. Neue Lehre von den Proportionen des menschlichen K6rpers
(Leipzig: R. Weigel, 1854); the book Der goldene Schnitt was printed posthu-
mously at the expense of the Leopoldinisch-Carolinischen Akademie: Halle,
1884.
figure 7-15). By now you might (rightly) expect that the intersec-
tion point of two diagonals divides each of the diagonals in the
golden ratio, as is the case with any regular pentagon.
Figure 7-15
In the field of architecture, there are countless illustrations of

the use of the golden ratio, which, of course, employ the Fibonacci
numbers. The lingering question persists: Are these coincidences
or are these golden sections marked off deliberately? In the case of
some buildings such as those designed by Le Corbusier, we have
enough evidence to know that the use of the golden ratio was in-
tentional. As for the ancient architecture, we can sometimes
speculate or assume that if an architect had a connection with a
mathematician, then he might have used it by design. Still, there is
always the notion that this ratio delivers the most beautifying par-
tition in geometry and therefore the "trained eye" might have sim-
ply landed on the golden ratio through artistic prowess.
Fibonacci Numbers and Sculptures
The golden section, 1/>, endows a work of art with a particularly

pleasing and "beautiful" shape. The golden section, 1/>, also
serves as the bridge to the Fibonacci numbers. Recall that the
ratio of consecutive members of the Fibonacci sequence ap-
proximates the golden section:
5. 1.666... , ~ = 1.6, .!2 = 1.625, ~ = 1.615 ...

3 5 8 13
The roughest approximation occurs with the smallest Fibonacci

numbers, and yet even then its greatest deviation from the actual
value l/J is less than _1 ,or 5 percent.
20
In 1509 the Italian Renaissance scholar Fra Luca Pacioli (ca.
1445-1517), in his book on regular solids, titled De divina propor-
tione,8 stressed the importance of the golden section and enthusi-
astically called it "De Divina Proportio" (i.e., divine proportion).
The use of the adjective divina shows the level of aesthetic signifi-
cance he attached to the golden section.
In the first part of the twentieth century, two important books
on the subject were published: in 1914 The Curves of Life 9 writ-
ten by Sir Theodore Andrea Cook, and in 1926 The Elements of
Dynamic Symmetry by Jay Hambidge. 1o Both books had a great
effect on artists, in particular the latter work. Hambidge' s primary
discovery was that there were two types of symmetry-static and
dynamic. His thorough study of Greek art and architecture, par-
ticularly of the Parthenon, convinced him that the secret of the
beauty of Greek design was in the conscious use of dynamic sym-
metry-the law of natural design based upon the symmetry of
growth in man and in plants.
The scholar Otto Hagenmaier 11 pointed out that the Apollo
Belvedere (a Roman copy of Leochares; Vatican Museum,
Rome)12 is an example of a famous sculpture with proportions that
reflect the golden section. In figure 7 -16 you can see how the
sculpture measures up in its golden ratio proportions.
8. Luca Pacioli, De divina proportione (Venezia, 1509; reprinted in Milan in

1896, and Gardner Pelican, 1961).
9. (London, 1914; reprint, New York: Dover, 1979).
10. (New York, Dover, 1967). The dynamic symmetry is based on the laws of
thermodynamics.
II. Otto Hagenmaier, Der Goldene Schnitt-Ein Harmoniegesetz und seine
Anwendung (Grafeling, Germany: Moos & Partner, 1958); 2nd ed. (Mi.inchen:
Moos Verlag, 1977).
12. Attic sculptor of the fourth century BeE; the Roman copy was found at ex-
cavations in Rome toward the end of the fifteenth century.

Roman copy of Apollo Aphrodite of Melos.
Belvedere.
The great mathematician Felix Klein (1849- 1925) could not

rest until he tried to explain the beauty of this magnificent sculp-
ture mathematically and beyond the golden ratio. He marked the
significant points of a Gaussian curveD on the face of Apollo in
order to discover the secret of its beauty. Is this really what we ex-
pect from mathematics? Maybe not, but it opens our eyes wider to
the mysteries of aesthetics.
The Aphrodite of Melos (also called Venus de Milo) in the
Louvre museum in Paris was created at about 125 BeE (see figure
7 -17). Her navel partitions the entire figure into the golden ratio.
Was this by chance or by design? We will never know for certain.
13. The Gaussian curve is the nonnal curve.

The artists' association Section d'Or was founded in 1912 by

Jacques Villon (1875-1963)14 in Paris, during the late Cubism pe-
riod. The association's name itself already highlights its belief in
the golden section. Its stated goal was to create absolute painting
purely from numerical ratios in which the golden section was not
just nominally used, but rather was the theoretical starting point.
The French-Hungarian sculptor and painter Etienne Beothy
(1897-1961) did not confine himself to the two-dimensional
painting that the Section d'Or community did, but expanded his
work to the three-dimensional realm. Beothy supported his theo-
retical groundwork further by providing a detailed construction
plan for all his sculptures. He relied in large measure on the
"golden series"-apparently a reference to the Fibonacci num-
bers-for the distances, curves, and other proportions in his works.
In 1926 Beothy wrote in the preface to the Hungarian version
of his book La Serie d'Or, 15 "It [the golden series] has the same
significance for the visual arts as harmony has for music .... What
works for music will also be achieved for the visual arts through
the golden series." At the end of this book, Beothy concludes that
the best sequence of numbers that can represent the golden series is
the additive sequence:
. . . 0, 1, 1, 2, 3, 5, 8, 13, 2 1, 34, 55, 89, 144, . . . 16
14. Whose actual name was Gaston Duchamp, French painter and graphic artist.
15. Uwe Ruth, Etienne Beothy, and Helga Muller-Hofstede, Etienne Beothy: Ein
Klassiker der Bildhauerei-Retrospektive (Sculpture Museum, Marl, Germany,
Exhibition 1979, ed. Dr. Uwe Ruth).
16. Etienne Beothy, La Serie d'Or (Paris: Edition Chanth, 1939).
248 THE FABULOUS FIBONACCI N UMBERS
Figure 7-18
(Used wilh permission of Uwe RUth .)
Beothy's wood sculpture, Essor 11 (op. 77), composed in 1936

(Sculpture Museum, Marl, Germany), appears to grow from a nar-
row footing and twists upward (see figure 7-18). It appears as if it
were a flame. Yet behind this sculpture lie exact plans based on the
Fibonacci numbers. It is out of this mathematical basis that the
sculptor derived a work of great beauty.
Through the minimalist art movement in the 1960s the use of
the Fibonacci numbers became fashionable. Some works by Timm
Ulrichs (b. 1940) rely entirely on the golden section. For a
sculpture (figure 7-19), he cut bread and vegetables in the propor-
tion of this number, ¢. He also sliced up a book with the title The
Golden Section by Otto Hagenmaier 17 into the golden section!
17. Der Goldene Schnitt: Ein Harmoniegeselz und seine Anwendung, 2nd ed.
(Grafe ling, Germany: Moos & Partner, 1958).
Figure 7-19
Timm Ulrichs, Der goldene Schnitt (Application example),
1969.
(Photo by Timm Ulrichs. Used with permission.)
In 1969 Ulrich wrote "I took several food products-bread,

sausage, and pickles that one usually divides with a knife- and
deliberately partitioned them into the golden section, and then em-
blazoned it in gold."
Jo Niemeyer (b. 1946), a German graphic artist and architect,
created works according to plans that were based on the golden sec-
tion. He even called his works Geometric Compositions. Niemeyer
is still moving on-in the true sense of the word! His work, 20 Steps
around the Globe ls was a breathtaking "country-art" project that he
completed in 1997. Once again, the golden section is the key ele-
ment of the project. It involved the installation of twenty high-grade
18. Ulrich Grevsmiihl, 20 Punkte- 20 Steps. Ein Land-Art Projekt von Jo

Niemeyer. See 20 Steps around the Globe. http://www.jo.niemeyer.com.
steel columns once around the earth. The columns were placed at
carefully calculated points along a great circle. 19 These twenty
points were determined by dividing the great circle around the earth,
whose circumference was given as 40,023 km. Yet the golden ratio
played a crucial role in determining the position of the points .
• • • • •
1 2 3 4 5
Figure 7-20
The distance from point 1 to point 2 was determined to be

.458 m (see figure 7-20).
Then the distance of point 2 to point 3 was determined by
l/J .. 458 m = .741 m. Point 3 is then 1.198 m from point 1.
The distance from point 3 to point 4 was determined by
l/J. 1.198 m = 1.939 m. That means that point 4 is 3.137 m from
point 1.
The distance from point 4 to point 5 is obtained by l/J . 3.137 m
= 5.077 m. Then point 5 is 8.214m from point 1.
By continuing in this way, he found that point 20 would
coincide with point 1 (with all points along a great circle).
The specific positions of the columns were more exactly car-
ried out with the help of the golden angle 20 (see figure 7-21).
The first column was placed in Lapland (Finland). This is near
the Swedish border and north of the polar circle. The eighth col-
umn can be seen on the photo of the starting point (figure 7-22).
The complete route (figure 7-21) is described with all necessary
data in the project 20 Steps around the Globe 21 at the Web site
http://www.jo.niemeyer.com. Jo Niemeyer believes that only the
proportions are essential for the perception and understanding of this
feat, and says, "There are no measures--only the proportions exist."
19. The great circle of a sphere is the largest circle that can be drawn on a sphere
and has its center at the center of the sphere.
20. The golden angle (see p. 148) = 137.507 764050 ... 0.
21. The geodesic draft was carried out under the scientific advice of Dr. Ulrich
Grevsmilhl (Hinterzarten); the computer program was by Jorg Pfeiffer. See 20
Steps around the Globe, http://www.jo.niemeyer.com.
,.
......
•
I •
•.. -
) D
•
Positions
1-12 Ropinsalmi Lapland Finland

0-18 km
13 Kautokeino District Norway
47 km
14 Anarjokka Norway
124 km
15 Pitkajarvi/Nickel Russia
325 km
16 The Barents Sea Int. Water
851 km
17 Noryj Urengoj Siberia Russia
2,229 km
18 Baotou [nner/Mongolia China
5,837 km
19 Westernport Bay Australia
[5,282 km
20 = 1 Ropinsalmi Lapland Finland
Figure 7-21
Figure 7-22
The starting point: RopinsalmiILaplandiFinland,
Position 68°40'06"N 21°36'21"E.
(Reprinted with permission of Jo Niemeyer.)
The Italian artist Mario Merz (1925-2003) set up a monument

to the Fibonacci numbers in several works of art. He is the doyen
of the arte povera (the poor art), which distinguishes itself by the
meager materials and simple gestures it employs. It is the Euro-
pean answer to the American minimalist art. For him the Fibonacci
numbers were a guiding light as well as an inspiration.
His works Fibonacci Napoli (1971) and Animali da 1 a 55
(I 997- 2000) are his best known. Yet, as an example of his work in
a black-and-white format, his drawing Ivy may more graphically
illustrate his respect for the simple elegance of the Fibonacci num-
bers (see figure 7-23).
Figure 7-23
By reaching out to the Fibonacci numbers, Merz relegates to

nature a reason for its power. "The Fibonacci numbers are in a
rapid expansion ... .The numbers have the power, to ease walls."
This is how Mario Merz describes the peculiarity and effect of this
historical design system. 22
-
Figure 7-24
(Reprinted with permission of Kunstmuseum Wolfsburg.)
On eight marble plates, Merz assembled his sculpture Igloo

Fibonacci (1970) with the help of brass pipes and steel hinges; on
this he mounted adhesive pieces of tape with (white) numbers. The
eight arms of the igloo are connected to each other with joints at the
distances that reflect the Fibonacci proportions (see figure 7-24.)
Obviously Merz was quite fascinated with the Fibonacci num-
bers, as is further evidenced by his having decorated the great hall
in the Zurich, Switzerland, central railway station with the Fibo-
nacci numbers. He also used the Fibonacci numbers to decorate
smokestacks: the 52 m smokestack at the Center for International
Light Art in Unna, Germany, in 2001 (figure 7-25), and a smoke-
stack in Turku, Finland, in 1994 (figure 7-26).
22. Mario Merz, Die Fibonacci-Zahlen und die Kunst, in Mario Merz's
Exhibition catalogue of the Folkwang-Museum, Essen, Germany, 1979, p. 75.
The Austrian artist Hellmut Bruch (b. 1936) was also greatly
influenced by the Fibonacci numbers as evidenced by his many
designs that clearly reflect the Fibonacci numbers in the relative
lengths of their components (see figures 7-27, 7-28, and 7-29.) It is
said that these lengths enabled him to create the most astounding
shapes of tangible and evident harmony.
The columns, made of stainless steel, exhibit through their
lengths a portion of the Fibonacci sequence (89 cm, 144 cm, 233
cm, 377 cm, 610 cm). The name Hommage it Fibonacci is an
eloquent testimoniaL
Employing the Fibonacci sequence in nearly all of his work,
Bruch has said that through these numbers he tries to show the
fullness of forms and the variety of nature and tries to understand
them. Bruch is not only interested in demonstrating the Fibonacci
series but, more important, in seeing what these numbers can do to
enhance his artistic possibilities.

(Reprinted with permission of Hellmut Bruch.)
The German sculptor Claus Bury (b. 1946) heralds the theme
of his sculpture merely with its name: Fibonacci's Temple. The
structure, built in 1984 and located in Cologne, Germany, is made
of raw spruce wood and is 15 m long, is 6.3 m wide, and 5.70 m
high at the center (figure 7-30) .
. ,
::t ~ "\
Figure 7-30
(Reprinted with permission of C laus Bury.)
The preparation for this structure reveals that Bury was already
fascinated with the Fibonacci numbers and in particular with the
theoretical writings of Le Corbusier in Modular (see pages
241-42). An inspection of Bury ' s plans demonstrates that he
clearly constructed this work according to the Fibonacci numbers
(see figures 7-31 and 7-32).
• •
L
1
• T--
•
[.
-.. ......
,... . --
--. . ... .. ... .
-
Figure 7-31
(Reprinted with permission of C laus Bury.)
2·3·~~
• •: • 6·'S~ ·l -0<>
~ ~s
2-'
..
., Ii
S
• ~ .2 3 "• e ~ ..3'-2' -00
> •
~fZ
-'"
• 5 32 J z -00
Z
• • • • • 1 .! , • • 5 _'"
Figure 7-32
(Reprinted with permission of Claus Bury.)
The Fibonacci numbers clearly have inspired modem artists in

their sculptures. Yet we still remain uncertain about the conscious
influence that these fabulous numbers have had on the ancient art-
ists. Were they aware of these numbers or of the golden ratio, or
did they do these works intuitively and they just happened to fit the
golden ratio? This leads us to look for further answers in the two-
dimensional medium of painting.
Fibonacci Numbers and Paintings
Leonardo da Vinci (1452- 1519i 3 illustrated the book De divina

proportione24 by the Franciscan monk Fra Luca Pacioli (ca.
1445- 1517) with an anatomical srudy of the "Vitruvian" man. This
drawing supported Pacioli's discussion of the Roman architect Vi-
truvius (ca. 84 BCE- 27 BCE)25 whose documents about architec-
rure suggest that he believed the proportions of the human body are
a basis for architecture. The ratio of the square side to the circle
radius in this well-known picture (figure 7-33) corresponds to the
golden section with a deviation of 1.7 percent .
. ,. .
.. •
Figure 7-33
23. Famous Italian painter, sculptor, architect. natural scientist, and engineer.
24. Luca Pacioli, De divina proportione (Venezia, 1509; reprinted in Milan in
1896, Gardner Pelican, 1961).
25. Actually Vitruvius Pollio, Roman armed forces technician and engineer. He
published ten books about architecture and civil engineering and developed a
theory of proportions.
The respected German painter and graphic artist Albrecht

Durer (1471 - 1528) also relied heavily on Vitruvius's work when
he wrote his Four Books on Human Proportions in 1523. Here he
refines the earlier works and expresses them as a system of pro-
portions, where the unit of measure is the human body with the
parts expressed as fractions thereof.
In Durer's books on geometry, fortress building, and human
proportions, he supported all of his ideas with theoretical (mathe-
matical) arguments, as opposed to aesthetic ones. As a matter of
fact, his works in descriptive geometry influenced some of the
greatest thinkers of the Renaissance- Johannes Kepler (1571-
1630) and Galileo Galilei (1564- 1642) among them.
In his famous self-portrait (ca. 1500), Durer drew himself with
a head of wavy hair, the outlines of which create an equilateral tri-
angle. This can be seen in figure 7-34, where Durer actually su-
perimposes the triangle and several other guidelines over the self-
portrait. The base of the equilateral triangle divides a height of the
entire picture into the golden ratio. The chin also divides the height
of the entire picture into the golden ratio (in the other direction).
Figure 7-34
It remains unclear why Durer never connected the golden sec-

tion to his drawing of the regular pentagon,26 since at age twenty-
nine he had constructed the golden section in his self-portrait.
Another classic work of art is Sandro Botticelli's painting The
Birth of Venus (1477). Sir Theodore A. Cook (1867- 1928) ana-
lyzed it with the aid of the golden section. He superimposed a
number scale on the figure of Venus, which had the first seven
powers of ¢ (see figure 7-35).
As earlier, in our review of architectual and scultpural applica-
tions of the Fibonacci numbers, we are faced here with a similar
problem to determine whether conscious use was made of the
golden section or the Fibonacci numbers. Many scholars collected
information in the nineteenth and twentieth centuries to search for
applications of the golden section. These applications need to be
critically and sensitively inspected. Oftentimes, we stumble on
fractions that approach the golden section and immediately assume
that the golden section was intended. We must be careful not to
draw conclusions where they aren't necessarily warranted.
Figure 7-35
26. See page 155ff for a discussion of DOrer' s construction of the " regular pen-
tagon," one that looks regular but is just slightly inaccurate.
Let us analyze a few samples where the golden section appears,

though you could unearth countless examples from the past several
centuries in which you could find similar "sightings." We will be-
gin with, perhaps, the most famous painting in Western civiliza-
tion, Leonardo da Vinci's Mona Lisa (figure 7-36), which was
painted between 1503 and 1506 and is on exhibit in the Louvre
museum in Paris. King Frant;ois I of France paid 15.3 kilograms of
gold for this painting. Let's look at this masterpiece from the point
of view of the golden ratio. First, one can draw a rectangle around
Mona Lisa's face and find that this rectangle is a golden rectangle.
In figure 7-37, you will notice several triangles; the two largest
ones are golden triangles. Furthermore, in Figure 7-38, you will
observe specific points on the body of Mona Lisa as golden sec-
tions. Since da Vinci illustrated Pacioli's book Da divina propor-
tione, which thoroughly discussed the golden section, it can be as-
sumed that he was consciously guided by this magnificent ratio.
Another great painter Raphael (1483- 1520) also exhibits the

golden section in his Sistine Madonna (1513). The horizontal line in-
dicated in figure 7-39, which emanates from the eyes of Pope Sixtus II
and the head of Saint Barbara, divides the height of the picture into the
golden section and also partitions the figure of the Madonna into two
equal parts. In figure 7-40, you will notice that the lines emanating
from very selected points in the figure aid in detennining the golden
ratio from another point of view; here, too, the Madonna's figure is
divided by the golden section. You will also observe that the isosceles
triangle, the ratio of whose base and altitudes closely approximates the
golden ration, superimposed on the picture in figure 7-40 exactly en-
cases the four heads. Whether Raphael consciously selected these di-
mensions with the golden section in mind, or whether these were
merely the product of an artistic eye, remains the master's secret.
In figure 7-41 a regular pentagram has been superimposed over

Raphael's Madonna Alba (1511-1513). This geometric fonn is
clearly visible along appropriate linear parts. This once again dem-
onstrates Raphael's penchant for the golden section-which we
already know is embedded in the regular pentagram.
Figure 7-41
(Natio nal Gallery, Washington. DC.)
Let's jump a few hundred years ahead to the French painter

Georges Seurat (J 859-1891), who was known for having a strict
geometric structure to his works. Although he used colors to a dra-
matic extent, Seurat wanted to highlight his use of geometric space
as the main attraction, which thereby helped set the groundwork for
modem art.
Figure 7-42
THE FIBONACCI N UMBERS FOUND IN ART AND ARCHITECTURE 263
The painting Circus Parade (1888, New York, Metropolitan

Museum of Art) by Seurat (figure 7-42) is replete with evidence of
the golden section. Some observers see the golden section produced
by the horizontal line at the top edge of the banister, and in the verti-
cal direction by the line that is just to the right of the main figure.
Others find that a golden rectangle is formed by these two lines and
the horizontal line just below the nine lights at the top of the picture.
There is also speculation that the golden spiral is embedded in this
picture. 27 The 8 x 3 unit rectangle (at the upper-left side) gives fur-
ther evidence to the possibility that the artist was aware of the rela-
tionship of the Fibonacci numbers to the golden section. In Seurat's
Bathers at Asnieres (1883, Tate Gallery, London), the golden ratio
can be seen as well as several golden rectangles (figure 7-43).
Figure 7-43
Jo Niemeyer (b. 1946), the German graphic artist and architect

mentioned earlier, created works according to plans that were
based on the golden section. He even calls his works Geometric
Compositions. In figure 7-44 you will see that his work titled Uts-
joki displays several golden sections.
27. H. Walser. Der Goldene Schnitt (Leipzig: EAG.LE. 2004). p. 135.

I
•
Figure 7-44
(Used with permission of Jo Niemeyer.)
For the painting Variation VI, 1987 (figure 7-45), we have

clear evidence that Niemeyer does not intuitively create his works,
but rather he sketches them from a mathematical basis. Notice that
he even marked the right side of the sketch (figure 7-45) with the
golden ratio (.61803), should our visual assessment of his sketch
not suffice.

(Used with permission of Jo Niemeyer.)
Niemeyer does not restrict himself to only two-dimensional art;

he also works in the third dimension. Again with the sketch of his
Modulon (figure 7-47) we are given clear evidence of the mathe-
matical underpinnings of his work. Modulon is an art object (a
cube) divided into sixteen building blocks according to the golden

ratio (see figure 7-48).
The possible variations that can be achieved by placing, order-
ing, grouping, and filling spaces are practically unlimited. The col-
oring is restricted, however, to the primary colors: blue, red, and
yellow, as well as black and white. The Modulon has been in the
collection of the Museum of Modem Art in New York since 1984.
Figure 7-47
1
2
AS
8 E A
Major Minor
I=1F : FR
EB : J:lB
v
,, "
.",-
\
--.- \
/'
./
.r .r \'
...
..r
./
",.
"
'~~-"x\
.. ~
~
,
. ..
_ "
,JO
".
,a'"~ .... ~ ~.
Figure 7-48
The Icelandic artist Hreinn Fridfinnsson (b. 1943) clearly based

his art on the golden rectangle and the golden spiral. His work called
Untitled (\988, Kunstmuseum Liechtenstein, Vaduz) can be seen in
figure 7-49. This is art based on the golden ratio throughout.
Figure 7-49
(Used with permission of Hreinn Fridfinnsson.)
Another example of an artist's conscious effort to use the

Fibonacci numbers as the basic structure of her work is that of the
German artist Rune Mields (b. 1935). In her Evolution: Progres-
sion and Symmetry 111 and IV (figure 7-50), she exhibits her
graphics as representing the Fibonacci numbers. The artist stresses
in the accompanying text of the catalog that "[i]n an ascending
line, a progression of triangles is generated, with the help of the
famous mathematical series of the Leonardo Pisano, called
Fibonacci, (0, I, I, 2, 3, 5, 8, 13, 21 ... ), who developed it at the
beginning of 13th century. It was an assumption that the arising
form should be respectively subject to the laws of the symmetry.,,28
28. Catalog on the occasion of the exhibitions: Rune Mields- SANCTA

RA no, ed. Kunstverein Lingen Kunsthalle [April 3- May 29, 2005,
Kunstverein Lingen, Germany , Kunsthalle, Lingen (Ems, Germany); August
I7- Novem: Buxus Verlag, 2005), p. 45.
.'"
Figure 7-50
(Used wilh permission of Rune Mields.)
There is some evidence that the golden section has been em-
bedded in the composition of the following works of art. Whether
it was intentional or intuitive is for you to decide. Many of these
works can be found in books, on the Internet, or for the world trav-
eler: in person!
• The relief Dionysius' Procession (Villa Albani, Rome)

• The fresco St. Francis Preaching to the Birds by Giotto
(1266- 1337)29 - (Basilica San Francesco, Assisi, Italy)
• The fresco Trinity by Masaccio (l401-1429)3o--{Santa
Maria Novella, Florence, Italy)
29. Actually Giallo di Bondone, Italian painter and master builder.

30. Actually Tammaso di Giovanni di Simone Gujdi , Italian painter who is re-
garded as a founder of Renaissance painting. The Trinity fresco advanced the art
of mural painting and the development of the altar because of the central per-
spective.
• The Deposition from the Cross by Rogier van der Weyden

(ca. 1400-1464)31 -(altar piece, Prado, Madrid, Spain)
• The Baptism of Christ by Piero della Francesca
(1415/1420-1492)32-(London, National Gallery)
• Madonna and Child by Pietro Perugino (ca. 1455-1523)33
-(Vatican Museum, Rome)
• The painting The Girl with the Ermine by Leonardo da
Vinci (National Museum, Krakow, Poland)
• The wall painting The Last Supper by Leonardo da Vinci
(Refectory of Santa Maria delle Grazie, Milan, Italy)
• The round painting of Madonna Doni by Michelangelo
(1475-1564)34 -(Uffizi Gallery, Florence, Italy)
• The painting Crucifixion (with the Virgin Mary, Saint
Jerome, Mary Magdalene, and John the Baptist) by Raphael
(National Gallery, London, England)
• The panels Adam and Eve by Albrecht Durer (Prado Mu-
seum, Madrid, Spain)
• The painting The School ofAthens by Raphael (Vatican
Museum, Rome)
• The fresco The Triumph of Galatea by Raphael (Villa Far-
nesina, Rome, Italy)
• The copper engraving Adam and Eve by Marcantonio Rai-
mondi 35
• The painting A Self-portrait by Rembrandt Harmensz van
Rijn (1606-1669)-(N ational Gallery, London, England)
• Gelmeroda by Lyonel Feininger (1871-1956)-( Gelmer-
oda VIII, Whitney Museum, New York; and Gelmeroda
XII, Metropolitan Museum of Art, New York)
• Half a Giant Cup Suspended with an Inexplicable Append-
age Five Meters Long by Salvador Dali (1904-1989)
31. Also called Roger de Le Pasture, Flemish painter.

32. Actually Pietro di Benedetto dei Franceschi, also Pietro Borgliese, Italian
painter.
33. Actually Pietro Vannucci.
34. Actually Michelangelo Buonarroti; Michelangelo was, however, sculptor,
painter, architect, and poet. In 1508 Michelangelo also designed the uniforms of
the Swiss Guard for the Vatican.
35. Italian copperplate engraver (ca. 1480-ca. 1534).
The following is a partial list of artists who are often men-

tioned in connection with the golden section:
• Paul Signac (1863-1935)

• Paul Serusier (1864-1927)
• Piet Mondrian (1872-1944)36
• Juan Gris (1887-1927)37
• Otto Pankok (1893-1966)38
Paul Serusier not only knew about the golden section, but he
also indicated it in his sketches. In contrast, Gris, Mondrian, and
Pankok have flatly denied using it.
Many admirers of l/J and Fibonacci would be disappointed to
learn that the careful examination by the art historian Marguerite
Neveux at the end of the twentieth century removed many pictures
from the "golden section list." She analyzed x-ray pictures ofvari-
ous canvases and came to the conclusion that most of the artists
divided their canvases into eighths before starting their work.
There are many ways to use these fractional partitions; yet more
often than not, ~ is selected. So if this art historian eliminated
8
some art from the "golden section list," we can still claim them to
the "almost golden section" since they used two Fibonacci num-
bers that can generate a rough approximation of the golden section.
It is truly fascinating how this "magical" golden ratio has in-
spired artists over the centuries-sometimes deliberately and at
other times intuitively. Whether the use of the golden section was
conscious or not, no one can deny its prevalence in many of the
great masterpieces of the Western world.
36. Painting I and Composition with Colored Areas and Gray Lines 1 (1918),
for example, are often quoted, as well as Composition with Gray and Light
Brown (1918; Museum of Fine Arts, Houston, Texas) and Composition with Red
Yellow Blue (1928).
37. Actually Jose Victoriano Gonzalez Perez, Spanish painter and graphic artist.
38. German painter, graphic artist, and wood engraver.
Chapter 8

Musical Form
Fibonacci Data on the Internet
C:~e:~ya~:~::ot~:t r:i;:'1e:ho~na;~e7:t~~:tt:~:~tO~~, :1~~:~::~

sequence and lIs relationship to musIc?] Some of the Web sites one
encounters are most fascinating. but what you generally find is often
confusing or msipu,l. Unfortunately, most of the inronnation that
shows up on your screen is not meaningful or even accurate. This
may be the result of some well-Intentioned authors who were writ-
ing for elementary school kids and tried to spoon-feed them some-
thlllg about F1bonacci and music that IS easily digestible. Well, thIS
stuff now shows up all over the place, and It docs not tell you any-
thmg generally sigmficant aboul the relationship. While 11 IS true
that the violin IS a beaullful example oflhe application of the golden
~ect]()n to eighteenth-eenrury italian Instrument making, the discus-
271
sions about eight-note scales are completely erroneous. The diatonic

scales we all know and love have seven notes, not eight, because
number 8 is a repeat of number 1 and continues the scalar motion
into the next octave. The fact that the black notes on the piano are
divided into twos and threes has nothing to do with the Fibonacci
sequence, since it is impossible to tell which one comes first, two or
three, because they relate to nothing except the white notes that
separate them. A true Fibonacci keyboard would have 2, 3, 5, and 8
black notes in every octave. Now that would be something!
It is also misleading to suggest that certain pleasing frequency ra-
tios, such as 5:3 (a major sixth) and 8:5 (a minor sixth), are related to
anything like the Fibonacci sequence because they are only a few
among a multitude of ratios that contribute to the rich intervallic (har-
monic) essence of the music we know and love. What about the major
third (5:4) and minor third (6:5) that are the triadic foundation of all
the popular music you have stored in your iPod? And last, if you listen
to the pieces that are generated by Fibonacci applications to pitch
and/or rhythm, you get to hear music that is perhaps interesting on first
hearing, but not worth a second play. So let us talk about Fibonacci
applications that really matter.
The golden section has been successfully employed by composers
mainly in two areas of the compositional process. The first relates to
the location of the climax and the second relates to form. Let us talk
about the climax first because that is an easier concept with which to
deal, especially if you were not forced to take piano lessons.
The Chopin Preludes
One of the great collections of nineteenth-century piano pieces is the

Preludes by Frederic Chopin (1810-1849). His book contains twenty-
four of the most extraordinary musical miniatures, each one a world
unto itself The fust of these is based on an interesting game that Cho-
pin is playing with himself In figure 8-1 we see the essential melodic
activity of the right hand, which is simplified for purposes of illustra-
tion. Each measure (except the final six) contains two notes, a step
THE FIBONACCI NUMBERS AND MUSICAL FORM 273
apart, one of which harmonizes with the left hand accompaniment (the
whole note) and one that does not (the black note). This piece, which
lasts only about a half a minute, is constructed of two dramatic arches
of differing sizes. The melody begins on the notes G-A and rises to the
E-D in measure 5, where it remains for three measures and then de-
scends back to the G-A in measure 9. From here the melody rises even
farther and climaxes on the D-C in measure 21. It then begins its de-
scent to the G-A in measure 25 at which point it leaps twice to E-D
before coming to rest on five Cs with A-G pairs below them. The
wonderful climax of this miniature masterpiece comes exactly in
measure 21 at the golden section of its 34 measures. (Remember these
numbers? Yes, they are true members ofthe Fibonacci sequence! Also
recall that 34· .618", 21.)
The same accurate placement of the climax at the golden section
occurs in Prelude No. 9 in E major. This piece is twelve measures
long and contains forty-eight beats. The climax occurs exactly on beat
29 ( 48· .618 '" 29) at the beginning of the eighth measure. Sometimes
it happens and sometimes it doesn't. In any case, the exact location of
the climax does not always have to conform to a prescribed mathe-
matical formula. Still, in a large number of cases it occurs in close
proximity to the golden mean. Most of the Preludes are not in a golden
proportion, and they work perfectly well in their less-than-perfect con-
dition. Apparently, Chopin found that it was not necessary to use l/J to
guarantee musical success.
f . - I.. - I .. - I- "I- .. I- "I- "I" - I .. -

10
d • d •
f I.. - I- " 1 1-" I .. I - " I - .. I •
<>
-=- • ft - ..
$., g.
,t
28
- .. 1"- .. 1"-.. 1"- .. 1"- .. 1" I..

Figure 8-1
Chopin: Prelude No.1 in C major.
Figure 8-2 is the graphic equivalent of figure 8-1 and shows the
placement of the golden climax in the sequence of pitches. It was de-
signed for those of you who are not familiar with musical notation.
21
•
• •• ••
• ••• • • ••
• ••• • • 34
••• ••• • ••••••
•
TIME
Figure 8-2
Chopin: Prelude No.1 in C major.
Binary Form
Now let us talk about musical form. This may prove to be more
taxing, but rest assured, it will be worth it.
Music whose structure contains two parts is said to be in binary
form. There are two kinds of binary form: equal binary form is di-
visible into two parts of relatively equal size, while unequal binary
form is divided into two parts of which the second is significantly
larger than the first. For a very long time composers have been per-
fectly happy with the results of using equal binary form-both
halves seem to balance each other very nicely. When they use une-
qual binary form, which may be equally successful, the question
arises as to the relative proportions of both parts, and this is where
the golden section enters the equation.
We often think of great composers as highly inspired creative
types who sit alone in garrets by candle light drawing their inspi-
ration from a distant muse. We imagine their creations to be the
result of some reverie when, in fact, most of what they do tends to

be quite rational and deliberate. While it is true that there are mo-
ments in the creative process when they listen to an inner voice
that tells them to do something out of the ordinary, most of the
time they are drawing upon years of training and experience that
has fostered a technical craft that allows them to make music that
"works." By "works" we mean that it conforms to a general game
plan that is self-limiting and keeps the piece from becoming a
hodgepodge of this, that, and the other thing. The difficult job of
the composer is to control the elements of the piece so that it
maintains an integral logic but still sounds like it was just made up
on the spur of the moment. Most important, the music must present
a smooth, continuous narrative that entertains its audience. With
very few exceptions, most composers are very bright people who
have worked hard to hone their skills and refine their craft. When
they put pen to paper (or make music by moving a mouse), they
are making very deliberate choices about the sequence of sounds
and silences that tells their musical tale. Quite often, the process of
composition takes on gamelike qualities. Some of the rules of this
game are derived from its style and some are unique to this specific
piece. Composers are not only tunesmiths and harmonists; they are
often game players.
The Mozart Piano Sonatas
One of the interesting games that have been played by a number

of famous composers relates to the use of the golden section in
determining the form of a piece. Wolfgang Amadeus Mozart
(1756-1791), who loved numbers and all kinds of games, was
especially fond of this application. Apparently, when he sat down
to write his piano sonatas, he always had the same plan in
mind-to try to use the proportions of the golden section to create
formal elegance and balance. In Mozart's time, the solo keyboard
sonata was comprised of three movements, or separate pieces.
The first was spirited and energetic; the second was slow and
lyrical; and the third was the fastest of all and brought the sonata
to an exciting conclusion.
Before we go any further, be warned that music theory can be a
fascinating line of study for those who wish to spend a lifetime in
pursuit of truth and beauty, since both can be tantalizingly elusive
when it comes to understanding the likes of Mozart. Even the most
adept theorists never feel completely satisfied when trying to dis-
cover and report on the internal workings of creations like Mozart
sonatas. We aim for complete understanding and always settle for
something less. Music theory is as complicated as physics and
mathematics, but it also has the aesthetic element that seems to take
it into the realm of the mystical or magical. Quite often, Mozart
seems more the magician than the musician. Such is the case when
we study the form of the first movements of his piano sonatas.
One of the most significant contributions to the music of the
late eighteenth century was the development of what we call the
sonata-allegro form. This name is derived from the fact that it was
used almost exclusively in the first movements of all major instru-
mental forms, all of which may be considered a type of sonata. 2 A
symphony is essentially a sonata for an orchestra while a string
quartet is a sonata for two violins, a viola, and a cello, and a con-
certo is a sonata for a soloist and an orchestra. As used in the so-
natas, it is a form that has two basic parts, each of which is re-
peated. The first part is the exposition where the musical materials
of the piece are presented. The section is repeated so you can hear
all that stuff over again since you probably did not catch it all the
first time because it went by so fast. When the repeat is accom-
plished, we move to the second part that contains the development
and the recapitulation. The development does what it says-it is
where the materials of the exposition are distorted, chopped up,
and tossed about. It is often quite tumultuous and exciting. It is
here that the increasing level of tension brings us to the climax of
the piece. As the tempest subsides, we return to the beginning of
2. A composition for one or more solo instruments, one of which is usually a

keyboard instrument, usually consisting of three or four independent movements
varying in key, mood, and tempo.
the piece. That is what recapitulation means-a return to the head.

This is where a lot of the sleight of hand takes place, because if we
are not paying close attention, we think we are hearing a repeat of
the exposition, but we are not. Much has been changed, but it is
often quite subtle and goes unnoticed.
Now you have learned more about sonata-allegro form than
you wanted to know, but a little extra knowledge won't hurt you.
And the next time you hear a symphony or a sonata you will have
some idea of how the first movement is put together. As we have
seen, the first movement is in binary form (two parts) although the
parts are not of equal size. That is the significant factor. The ques-
tion for Mozart and his contemporaries was how to get these two
parts in balance even though they are not the same size. Enter the
Fibonacci sequence and the golden mean whose proportions of
.6180339 provide a possible solution. 3
.3819661
It should be pointed out that in the baroque period (1600-1750)
a great deal of music, especially dance forms, was in equal binary
form-both halves were approximately the same size, contained
similar music, and were usually repeated. Along come the com-
posers of the classical period (1750-1825), and they expanded this
form into what we now know as the sonata-allegro form. This is
what we call a rounded binary form because it features a return of
the opening section in the latter part of the second. It looks some-
thing like this: II: A :11: AI A :11
Mozart wrote eighteen piano sonatas and all but one employ
sonata-allegro form in the first movement (the remaining one uses
theme and variations form). As may be seen in figure 8-3, of the
seventeen, six (35 percent) are exactly divisible by the golden sec-
tion and are identified as such in the Measures column by the word
"golden." Eight (47 percent) are very close and are identified in the
Measures column by numbers ranging from -3 to +4. These quan-
tities represent the displacement of the golden section. And three
3. This fraction is built by the quotient *~ 0.618033988 ... and ~: ~ 16,~~ ~

0.3819661865 ....
(18 percent) are really not close enough for serious consideration
because the expositions are six, eight, and twelve measures too
long. Statistically, that certainly leaves one with the impression
that the use of the golden section was important to Mozart.
Mozart Sonata Key Length Exposition Proportion Accuracy

No. I. K. 279 C major 100 38 0.38 golden
No.2. K. 280 Fmajor 144 56 0.389 golden
No.3. K. 281 Bb major 109 40 0.367 ·2
No.4. K. 282 Eb major 36 15 0.417 I
No.5, K. 283 G major 120 53 0.442
No.6, K. 284 o major 127 51 OA02
NO.7. K. 309 C major IS6 59 0.378 golden
No.8, K. 310 A minor J3J 49 0.368 ·1
No.9. K. 311 o major 112 39 0.348 ·3
No. 10. K. 330 C major 149 57 0.383 golden
No. II, K. 331 A major 135 55 Theme & Var.
No. 12, K. 332 F major 229 93 0.406
No. 13. K.333 Bb major 170 63 0.371 ·1
No. 14. K.457 Cminor 185 74 0.4 4
No. 15. K.S45 C major 73 28 0.384 golden
No. 16. K.S70 Bb major 209 79 0.378 golden
No. 17, K. 576 o major 160 S8 0.363 ·2
No. 18. K. 533 F major 240 103 0.429 12
Figure 8-3
Mozart Piano Sonatas.
Of the six that are right on the money, there is one, Sonata No.
(K.279), that is exactly 100 measures in length and the exposi-
tion ends in measure 38. It just does not get more obvious than
that, and because it is the first, it serves as a kind of declaration of
purpose. Here is what it looks like:
Exposition : I I: Development Recapitulatiori :II
38 MEASURES 62 MEASURES
Figure 8-4
Those of you who know something about the music of the late
eighteenth century might be asking yourself whether Franz Joseph
Haydn (1732-1809) was just as keen to use the golden section in
his piano sonatas. After all, he is the other giant of the classical
period and it would be interesting to see his approach to sonata-

allegro form. As it turns out, his adherence to cP did not match Mo-
zart's. In analyzing an equal number of randomly chosen piano
sonatas, we find that only 18 percent (~) of his sonata-allegro
forms are golden, while 53 percent (~) are close, and 29 percent
(;) are out of consideration. Make of it what you will.
Haydn Sonata Key Length Exposition Proportion Accuracy

No. 14. 1767 E major 84 30 0.357 -2
No. 15. 1767 o major 110 36 0.327 -6
No. 16. 1767 Bb major 116 38 0.327 -6
No. 17. 1767 o major 103 42 0.408
No. 19, 1773 C major 150 57 0.38 golden
No. 21. 1773 Fmajor 127 46 0.362 -3
No. 25, 1776 (J major 143 57 0399
No. 26, 1776 Eb major 141 52 0.369 -2
No. 27,1776 Fmajor 90 31 0.344 -4
No. 31.1778 o major 195 69 0.353 -6
;\10.32, 1778 I: minor 127 45 0.354 -4
No. 33, 1780 Cmajor 172 68 0.395
No. 34, 1780 C# minor 100 33 0.33 -5
~o. 35, 1780 Dmajor 103 40 0.388 golden
No. 42. 1786 G minor 77 30 0.39 golden
'No. 43, 1786 Abmajor 112 38 0.339 -5
No. 49. )793 Eb major 116 43 0.371 -I
Figure 8-5
Haydn Piano Sonatas.
So, are the Mozart movements better than those of Haydn? Can
a statistical analysis help us arrive at a reasonable conclusion? The
proportions of the average Mozart movement are .389 as compared
.611
to Haydn's .364. However, Mozart's placement of the double bar

.636
relative to cp ranged from -3 to +12, while Haydn's only ranged
from -6 to +2. Perhaps there are other ways of using the data to
help us. But it is also possible that numbers may have nothing to
do with the quality of the music. Here is a suggestion: listen to all
of the thirty-four movements under consideration in pairs-one
Mozart followed by one Haydn-and see how it goes. Even though
you may be as confused at the end of the test as you were at the
beginning, at least you will have heard a lot ofterrific music.
Beethoven's Fifth Symphony
The opening five measures of the first movement of this symphony

forms probably the most universally recognized musical statement
in classical music. Few of the millions who adore this piece realize
how extraordinary and revolutionary it is in the history of Western
classical music. There are so many aspects to this work that are
worthy of discussion that countless authors have written tomes
about it. Since this is a book about the Fibonacci sequence, let us try
to contain our enthusiasm and stick to the topic (it may not be easy).
125/ 182/ 248/ '268/ 374/ 47f1/ em/

249 :J06 372 392 498 602 626
Exposition : II Developmenl Recapitulation Coda Codetta
I Oboe caden7.a
IVlsas!lemhl)
Figure 8-6
Beethoven, Symphony No. 5,jirst movement.
Before you get dazzled by the data, you need to understand

how the form of this movement differs from those in the sonatas
and symphonies of Mozart and Haydn. Ludwig van Beethoven
(1770-\827) was born into the generation of musicians that fol-
lowed these two pillars of Austro-Hungarian musical culture and
proved to be the most revolutionary and influential composer of
the Romantic period. Surprisingly, the exposition, development,
and recapitulation in this astounding work are all essentially the
same size, thus preventing a golden section of these three sections.
Instead of bringing the first movement to a conclusion, the reca-
pitulation rushes headlong into a coda4 that turns out to be another

development section-something his predecessors never contem-
plated. Compounding that is the addition of a codetta (little tail
section) to the coda which, when taken together, add a significant
and unprecedented fifth section. This tail is big enough to wag the
dog! Therefore, we end up with five sections, not four, all of which
are between 124 and 128 measures. This is a new kind of sonata-
allegro form.
As may be seen in figure 8-6, this newly expanded sonata-
allegro form contains three golden sections that may have been part
of a deliberate attempt to apply the principle in new ways. First, the
arrival of the recapitulation in measure 372 is the golden section of
the entire piece without the shocking second coda. Without the co-
detta, the piece would be 602 measures in length: 602· .618 = 372 .
Right on the mark! The rest of the evidence hits the target-not ex-
actly in the bull's eye, but close enough for music.
The end of the repeat of the exposition (measure 124 . 2 =
measure 248) is almost a golden section of the entire movement
248 =.396) .
( 626
If we remove the two measures of rest at the end of the
exposition (measures 123-24) the proportion is even closer

244 =.389).
( 626
In a generous mood we could credit Beethoven for that
one too.
Two very special compositional events occur at points of critical
proportion. The first occurs in the development, where the four-note
motive begins to break into two notes and then into one. This disas-
sembly of the four-note motive, that has remained intact so far, oc-
curs at measure 306. This is a golden section of the movement up to
the end of the recapitulation in measure 498 (306=.614). One of the
498
most inspired moments in the entire symphony occurs in the reca-

pitulation when, in measure 392, the entire orchestra stops, as it did
in the analogous spot in the exposition, except for the oboe. The lit-
4. A coda is a section of music added at the end of a composition, introduced to

bring it to a satisfactory close.
tle cadenza, which the oboe plays, has no precedent and is dramati-
cally shocking and perplexing to those familiar with sonata-allegro
form. This astounding solo occurs only six measures from the
golden section of the entire movement (626· .618 = measure 386).
"Close, but no cigar," you say. Well, we will never know if Beetho-
ven had intended this special moment to occur at a golden section,
but it is tantalizingly close, and it is possible that, in the heat of
creation, measures were added and subtracted with the final result as
we see it now. The process of continually reediting his music was
Beethoven's usual practice, for he wrestled mightily until his almost
indecipherable scores were finally finished and could be sent to be-
fuddled copyists and publishers.
Wagner's Prelude to Tristan und Isolde
When, in the late 1850s, Richard Wagner (1813-1883) sat down to

write a mythic opera about the delays and disappointments of
tragic love, he had a plan to force a gigantic leap in the art of musi-
cal composition. So revolutionary is this work that today theorists
are still arguing about its construction. We will bypass all that ex-
traordinary stuff, however, and get to the I/J at the heart of the
prelude that opens this masterpiece.
Wagner, like Mozart, was a composer who liked to play games.
Perhaps the most extraordinary evidence of Wagner's game play-
ing is his deliberate use of the golden section to delineate signifi-
cant keys in the Tristan prelude. By "key" we mean what scale, or
collection of notes, is being used as the primary material of a piece
of music. C major is a scale that contains the notes C, 0, E, F, G,
A, B, and C. Beginning pianists like this scale because it is com-
prised of only white notes, which are easier to play. Anyway, a
piece that is in C major uses the notes of that scale as the building
blocks of its composition. At the beginning of every line of music
there is what we call a key signature. This is a collection of flats or
sharps that composers put there so they do not have to notate the
flats and sharps next to every note. It is, in effect, a form of musi-
cal short hand. For example, if you see a sharp (#) on the F line of
the staff in the key signature, it means that the performer should
play only F sharps (black notes), not F naturals (white notes).
It should be mentioned that European intellectuals in the 1850s
were familiar with the principle of the golden section, since it had
come back into vogue around this time. Wagner certainly knew
about it, but would not have let slip that he was using this princi-
ple. He, like so many composers, did not like to tell his audience
how he did his tricks of musical magic. These composers assumed
that trained musicians who were curious would dig into their
scores in order to unearth the hidden secrets. Until recently, most
Wagner theorists did not realize that the golden section applica-
tions were at work in this piece. They lay undiscovered for more
than 140 years.
An examination of the score of the prelude reveals something
very strange. The key signature of the piece has no flats or sharps,
indicating that the scale being used is either C major or A minor. In
this case, it is A minor. In measure 43, the key signature changes to
A major (three sharps) even though the key is not A major. In
measure 71 the key signature returns to A minor. What is strange
about this is the fact that the beautiful, but revolutionary, music
you hear is never actually in any key-the tonal orientation moves
frequently from key to key. So the question is, Why would Wagner
go to the trouble of changing key signatures twice in a piece that
really isn't in any key? The extreme fluctuation of keys does not
seem to call for a key signature, and both changes take place in the
middle of phrases. If you chart this phenomenon, it looks like this:
1--- 42 measures --- II --- 28 measures --- II --- 41 measures ---I
Figure 8-7
Look familiar? Have you done the math? The double bar at the
end of measure 70 divides the III measures by a section that is
extremely close to the golden section (~=630),
III
and the double bar
at the end of measure 42 does the same thing for the first 70 meas-
ures (~=.6). This is only one of the many fascinating ways this
piece is organized in terms of key structure. There are applications
of the golden sections all over the place, as you can see by looking
at the diagram in figure 8-8.
Figure 8-8
Wagner: Prelude to Tristan und Isolde.
The harmonic events that take place at each of these points of de-
marcation may not mean much to you unless you are a scholar of
Wagner's chromaticism and shifting tonal centers. But rest as-
sured, these events are critical to the construction of the piece.
Wagner knew what he was doing. If you are wondering where
Wagner put his climax, it is not at the golden section (measure 68),
but in measure 83, three-quarters of the way through the
piece-very late in the game by traditional standards.
The following information is only for musicians, the foolishly
curious, or mathematical sticklers. The golden mean of a piece 111
measures long is calculated to be measure 68.
(obtained by: 111· .618 = 68.6). The question is: Should we expect
Wagner to do something really special at the golden mean? The an-
swer is: Sure, why not? While it is true that he placed the double
bars at the end of measure 70, measures 68-69 are significant be-
cause they contain the last of the dominants 5 in A minor (E 9 ) that

began in measure 63, and this final dominant is preceded by a su-
pertonic Bo. This is the only true ii-V progression in the tonic kel
(Am) in the entire piece! The strategic importance of measures
69-71 is reinforced not only by the return to the opening key signa-
ture but also by the shift away from E9 to G 9 , the dominant of C
major, as well as the arrival at the apex-the only consecutive return
of the opening A-F dyad 7 in either the melody or the bass line!
Bartok's Music for Strings, Percussion, and Celesta
When, in 1936, Bela Bartok (1881-1945) was commissioned by

Paul Sacher (1906-1999) to write a work for chamber orchestra, he
decided to provide the wealthy conductor with something very
special. What he produced was the Music for Strings, Percussion,
and Celesta, one of the seminal works of the twentieth century. It
is a piece that uses no woodwinds or brass instruments in its en-
semble. What we hear is a small string orchestra in combination
with a variety of percussion instruments: side drum, snare drum,
cymbals, tam-tam, bass drum, tympani, xylophone, celesta, harp,
and piano. This instrumentation had never been used before, or has
been since. Its sound is totally unique. It is quintessential Bartok.
This masterpiece is comprised of four movements, the first of
which is a fugue. Just in case you forgot what you learned in Music
10 I, a fugue is a contrapuntal (two or more melodies at the same
time), additive musical process that begins with a solo statement of
the fugue subject (melody). When the statement ends (it can be
anywhere from a few notes to a protracted melody), it is imitated
in another voice (or instrument) at a different pitch level (higher or
lower). At this point you hear the fugue subject in the second
voice, accompanied by a countermelody in the continuation of the
5. A dominant is the fifth note ofa scale, or the chord built on that note.
6. A tonic is the first note of a scale.
7. A dyad is any two notes or pitches (consecutively or simultaneously), also
known as an interval, but dyad sounds better when the two pitches in question
are important.
first voice. Most fugues have three or four voices, and each of
them gets to state the subject in the opening section we call the ex-
position. When each of the voices has had a chance to present the
subject, and we have three or four countermelodies all playing to-
gether, we move on to the development section. It is here that just
about anything can happen. The subject may be played by one or
more voices in inversion (upside down), backward; it may be
rhythmically altered or chopped into pieces. There may also be a
lot of overlapping of subject statements, something we call stretto.
There is no prescription for what should be done. That is left to the
composer's imagination and invention. It is a form that was quite
popular in the baroque period, and Johann Sebastian Bach
(1685-1750) was undoubtedly its master. Composers since Bach
have tried their hand at this complicated contrapuntal challenge,
but fugues were relatively rare in the nineteenth and twentieth
centuries, and are so far in the twenty-first century. It is not an easy
thing to write because it is the musical equivalent of a crossword
puzzle that makes sense both vertically and horizontally at the
same time.
Two centuries after Bach, Bartok provides us with one hum-
dinger of a fugue in the first movement of his Music. What he did
was take an old idea and dress it up in new clothes. It is, in effect, a
super fugue for the twentieth century. The most striking update is
the presentation of the fugue subject on all twelve steps of the
chromatic scale (all the black and white notes within one octave)
beginning with A. The arrival on the twelfth pitch (E flat) in meas-
ure 56 is also the climax of the piece and is very close to golden in
its proportion (~=63). From this point to the end of the piece, the
statements work their way back through all twelve pitches until the
final statement on A. Figure 8-9 shows how the subject entrances
are related by pitch. The first statement begins on A and the second
is a fifth higher on E. The third entrance is a fifth lower than A (D)
and is followed by an entrance a fifth high than E (B). The pattern
of fifths continues until measure 56, where we find all the strings
playing in unison E flats. The scheme works backward from there

until the end:
EI
G# AI
c# c#
F# F#
B B
E E
A 56 A
D D
G
G
C
C F
F BI
BI
EI
Figure 8-9
If you take a good look at the diagram in figure 8-10, you will
notice that it bears an uncanny resemblance to figure 8-8. The major
difference is the placement of the climaxes, with Bartok being more
of a traditionalist than was Wagner seventy-seven years earlier. This
diagram (figure 8-10) displays the significant golden section and its
relationship to subject statements or entrances of new instrumental
colors. It is obvious from the pitch scheme and the use of the golden
section that this piece is one enormously complicated game. There
are other games at work in this score, but they require more training
and patience on the part of the reader than may be appropriate here.
But, lest you feel cheated, one or two more delicious morsels
couldn't hurt. The fugue subject is comprised of four little phrases,
each of which rises and falls like a miniature version of the grand
scheme of the piece. When we come to the last phrase of the fugue,
we find the second phrase of the subject played by the first violin,
while the second violin plays its inversion. This simultaneous, mir-
rored bow form (A-E flat-A) looks and sounds like a miniature ver-
sion of the diagram you see below.
tlml'd"j
Figure 8-10
Bartok: Music for Strings, Percussion, and Celesta.
Good Math, Bad Music
A possible reason that much contemporary music has never been

well received may be that too many composers got lost in the
number games of twentieth-century compositional practice. While
it is true that all composers play mind games with themselves in
the process of creating music, many luminaries in the generation
that included the likes of Pierre Boulez (b. 1925) and Milton Bab-
bitt (b. 1916) lost sight of the fact that people are mostly emotional
beings, not numerical calculators (sacred hearts, not sacred brains).
Too many of the works from that period are brilliantly calculated
but are devoid of the color, mood, passion, or musical narrative
that can move an audience to tears or rapture.
When Arnold Schonberg (1874-1951) applied his twelve-tone
method (formulated in 1924), he imbued his soundscapes with the
nineteenth-century Romanticism of his youth. He was always the
Romantic. Here was a brilliant mind that was searching for a re-
placement for the diatonic tonal system and, in the process, created
exquisitely crafted works that few people wanted to hear. His music
had a new technology clothed in an out-of-date aesthetic. It was his
student, Anton von Webem (1883-1945), who seemed to apply an

aesthetic that was truly modem, more in tune with the times and the
new note-picking system. For a long time performers of his music
did not realize that he, too, was a Romantic, and their performances
reflected that ignorance-they were insipid. He just disguised his
romanticism better than Schonberg. His sparse textures must be per-
formed with the same passion that one would apply to the music of
Johannes Brahms (1833-1897) and Gustav Mahler (1860-1911),
only the attention is focused on far fewer notes.
This misunderstanding of W ebem' s underlying Romanticism
led a lot of young composers that followed him to believe they
could play all those wonderful numerical and structural mind
games and everything would work out just fine. Well, it did not.
Over many decades the musically adventurous among us have at-
tended too many concerts where minds were dazzled, but hearts
were unmoved. Often they went home unable to remember a single
musical gesture. Too often they heard a vast amount of musical
data without hearing anything that seemed to be meaningful. What
this music often lacked was modesty. Mozart played as many mind
games as anyone who ever lived, only he hid his calculations
where only the refined ear and eye could find them. His out-front,
tuneful lyricism disguised his underlying machinery. His music
invites us to sing along, and, once we are seduced by his magic, we
spend a lifetime trying to figure out how he did it.
Music is a communication from heart to heart and mind to
mind-it must tell us what it means to be a human being, not a
UNIV AC. 8 The most difficult part of being a composer is trying to
balance the heart and the mind, and it does not take much to tilt
slightly in the wrong direction with results that are brilliantly in-
sipid and vacuous. There are those, like minimalist composer Tom
Johnson (b. 1939), who believe otherwise. Here is a fellow who
has employed a great deal of mathematical calculations in his
compositional process with questionable results. Many art forms in
8. UNIVAC is an acronym, standing for UNIVersal Automatic Computer.

UNIV AC was one of the eight major computer companies throughout most of
the 1960s.
the twentieth century have had practitioners who eschew their own
humanity. For them, emotion plays no part in the artistic process.
Johnson admits this denial of the personal element as follows:
1 have often tried to explain that my music is a reaction

against the romantic and expressionistic musical past,
and that 1 am seeking something more objective, some-
thing that doesn't express my emotions, something that
doesn't try to manipulate the emotions of the listener
either, something outside myself. Sometimes 1 explain
that my reasons for being a minimalist, for wanting to
work with a minimum of musical materials, is because it
also helps me to minimize arbitrary self-expression.
Sometimes 1 say, "I want to find the music, not to com-
pose it.,,9
This is in tune with the philosophy originally promoted by John

Cage (1912-1992), a highly influential thinker whose music has
never found wide acceptance with a paying audience. Many who
have heard Johnson's music feel that he fulfills his minimalist in-
tentions most effectively. Pieces such as his Narayana's Cows, a
composition that employs the Fibonacci series as a pitch generator,
are pleasing to some, but have not found wide acceptance. Many
people, when listening to his music, come away disappointed that
they have not been touched by the personal revelations of a sen-
tient being. They experience the musical equivalent of eating a
vegetarian snack-while at the concert they are entertained by a
skillful array of sounds, but an hour later they are hungry for
something more profound.
It's too often the case that good math doesn't make good music.
The Coda
If you are still reading, it means that you are of stout heart and firm
mind because, for a book that was supposed to be about math, you
9. http://www.ChronicleoftheNonPopRevolution,Kalvos.org/johness4.html.
THE FIBONACCI NUMBERS AND M USICAL FORM 291
have been inundated by an awful lot of music theory. But, if you

got this far, you can see that though the golden mean plays a sig-
nificant role in music, it is by no means the whole story. Be low
you will find a little reminder of Fibonacci in music: a picture ofa
violin with its proportions re lated to 4>. It is. unquestionably, a
lovely shape to behold and, when put in the right hands, can be an
extraordinary conveyor of passionate expression. Antonio Stradi-
varius ( 1644- 1737) is, undoubted ly, the most famous maker of
violins. His instruments set a standard that is still being fo llowed
today. The proportions. components, and assembly of his instru-
ments have been studied and imitated by a ll those attempting to
replicate an instrument that sings freely and projects to the farthe st
reaches of the concert hall. Today, hi s instruments cost severa l
mill ion dollars on those rare occasions when they come on the
market. If you can ' t make the payments on a "Strad" and want to
go out to the shop to see what you can paste together, here are the
Fibonacci proportions to get you started. Good luck!
Figure 8-11
Chapter 9
The Famous Binet Formula for

Finding a Particular
Fibonacci Number
U ntil now, we usually inspected the Fibonacci numbers by

their location in their sequence. If we wanted to find the
tenth Fibonacci number, we would write the sequence to 10 num-
bers, and then we would have the tenth number. In other words, we
would simply list them as 1, 1, 2, 3, 5, 8, 13, 21, 34, 55 and count
to the tenth number in the sequence. This would give us 55 as the
tenth Fibonacci number. Such a procedure can be somewhat cum-
bersome, however, if we seek, say, the fiftieth Fibonacci number.
After all, listing fifty Fibonacci numbers is no easy matter if you
don't have the list handy that we after you in appendix A.
Well, the French mathematician Jacques-Philippe-Marie Binet
(1786-1856) provided us with a formula for finding any Fibonacci
number without actually listing the sequence as we just did. To de-
velop this formula, we will gently take you through some elemen-
tary algebra, explaining all of our steps along the way. But before
we embark on this somewhat tedious task, we need to understand
what it means to derive such a formula. We will, thus, examine a
simple and very well-known kind of number-the square num-
ber-to determine what it means to work with a recursive sequence
of numbers such as the Fibonacci numbers.
293
294 THE F ABULOUS FIBONACCI NUMBERS
Figure 9-1
Jacques-Philippe-Marie Binet.
The recursive nature o f the Fibonacci numbers comes from the

defin itive relati onsh ip: Fhl = F~ + F"~I' w here we know that
F; = I and Fl = I. We are accustomed to finding any Fibonacci
number in the sequence as long as we know its two preceding
numbers.
Suppose we seek to find the s ixth square number (remember,
the square numbers are 1, 4,9, 16,25,36,49,64 , 81, and so on),
then wc simpl y square 6 to get 36. If we are call ed on to find the
11 8th square number, then we square 118 to get
11 8 · 11 8 = 13,924. This is an explicit d efin ition of the square
numbers. Let 's try 10 set up a recursive defin ition of the square
numbers.
We begin with the defin ition of S• • which wi ll represent the
nth square number. We know that SI = I, and we wi ll see that
5.,1 = s. + (2n + I) . Using this definition, we would get the fol -
10wing: 1
I. We ean a lso usc that.so .. O. the Oth square number.

THE FAMOUS BINET FORMULA 295
5, = Su + (2 · 0+ J) = 0 + J = I
S, = S, + (2. I + I) = I + 3 = 4
SJ = Sz+(2· 2 + 1) = 4 + 5 = 9
S. = S}+(2· 3 + 1) = 9+ 7 = 16
S~= S. + (2 · 4+ 1) = 16+9=25
This tells us that we can get a square number by adding to its

predecessor-square number a number that is one greater than twice
the position number of that square. Of course, you may be thinking
that this is a silly way to look at a rather simple problem- when you
can just square the number and that 's it. However, we are trying to
show that there is a recursive relati onship from which one can also
generate square numbers-even though it is ridiculously more com-
pl icated.
From this we can a lso conclude that the sum o f consecutive
odd numbers (beginning with 1) is always a square number, which
leads us to the recursive re lationship as d isplayed below.
Since 1 + 3 + 5 + ... + (2n - I) + (2n + I) = (n + 1)2,
and (n + 1)2 = n 2 + (2n + I), for any natural number n, we get the
recursive relationship we used above.
Even tho ugh we can easily prove this by the technique known
as mathematica l induction, we can see this in a geometric repre-
sentation (figure 9-2), which can serve a similar purpose
,- ,
4 ~ ! +J
~ ~ T'3+S
1 6 ~ I~J"5+7
25 - 1. 3 "'S"7"'~
)6 ~ I ')+ 5 +7+9+ 11
4~ . ' +J+5+7 ... 9+ II + 13

64 ~ I +3+5+7+ 9+ 11 + IJ+ 15
BI ~ I · J +5+7· 9.1I + IJ + 15 +!7
100 - I +)+5 +7 +9 + 11 + 1l'" 15 + 17+ 19
F igure 9-2
We have now seen the square numbers from the explicit defini-
tion (our more common one) and the recursive definition (the much
more cumbersome one). Until now, we have been viewing the
Fibonacci numbers from the recursive relationship. In 1843 Jacques-
Philippe-Marie Binet (1786-1856)2 developed an explicit definition
for the Fibonacci numbers. As is often the case in mathematics when
a formula is named after a mathematician, controversies arise as to
who was actually the first to discover it. Even today when a
mathematician comes up with what appears to be a new idea, others
are usually hesitant to attribute the work to that person. They often
say something like: "It looks original but how do we know it wasn't
done by someone else earlier?" Such is the case with the Binet
formula. When he publicized his work, there were no challenges to
Binet, but in the course of time, there have been claims that
Abraham de Moivre (1667-1754) was aware of it in 1718, Nicolaus
Bernoulli I (1687-1759) knew it in 1728, and his cousin Daniel
Bernoulli (1700-1782)3 also seems to have known the formula
before Binet. The prolific mathematician Leonhard Euler
(1707-1783) is also said to have known it in 1765. Nevertheless, it
is known today as the Binetformula.
Let us now slowly try to develop this relationship. We will begin
by recalling the golden ratio (~=~)
x x+!
that led us to the equation:
2
x - x-I = O. The roots of this equation are ¢ and -.!. , where ¢ =
.J5+1 IS
~-
. t h e announce d go Id en ratIO
. an d th erelore
C' •
IS -1 =
'" ~
2 .J5-1 .
= ~-
2 '" ,,5+1 2
By now you may be familiar with the sum and difference of

these terms as:
,;, 1_.J5+1 .J5-1_

'P- - - - - - - - - - - 1
¢ 2 2
2. " Memoire sur I'integration des equations lineaires aux diffeences finies d'un
ordre quelconque, a coefficients variables," Comples rendus de I 'academie des
sciences de Paris, vol. 17, 1843, p. 563.
3. The Bernoulli family is like a clan (of 8 mathematicians in 3 genera-
tions)-famous and estranged!
As a next step, we will take successive powers of this sum and

difference. Figure 9-3 shows this.
n Sum Difference
1 I = 1...[5 I =1
¢+ - ¢- -
¢ ¢
2 ¢2+ -'-- =3 ¢2 _ ¢12 = 1...[5
¢2
3 ¢3+ ~ = 2...[5
¢3 - ;3
=4
¢3
4 ¢4 + -
I =7 ¢4 _ ¢14 = 3...[5
¢4
5 ¢s + -
I = 5...[5 ¢s_¢ls =11
¢s
6 ¢6 + - I = 18 ¢6 _ ¢16 = 8...[5
¢6
7 ¢7 + -
I = 13...[5 ¢ 7 _ ¢\
= 29
¢7
8 ¢s + -
I = 47 ¢ s _ ¢Is = 21...[5
¢s
9 ¢9+ ~ = 34...[5 ¢9 _ ¢19 = 76
¢9
10 ¢ 10 I = 123 ¢ 10 _ _1_ = 55...[5
+10 ¢IO
¢
Figure 9-3
An inspection of the values generated for the sum and differ-

ence suggests that we can represent them in terms of Fibonacci
numbers ( Fn ) and Lucas numbers ( Ln)' as in figure 9-4.
n Difference
I
l/J+ -
, Sum
=1j5=F;j5
,
l/J- - = 1 = LI
41 41
2 l/J 2 + -
, = 3 = L2 = Ij5 = F2 j5
2 l/J 2 - 41'2
41
3
l/J3+ ~
41
3 =2j5=F;j5 l/J3 - ;3 =4 = L3
4 l/J4 + -
, = 7 = L4 = 3j5 =F.j5
414 l/J4 - 41'4
5 l/J 5 + -
, =5j5=~j5 = II = Ls
5 l/J 5 - 41'5
41
6 l/J6 + -
1 = 18 = L6 =8j5 = F6 j5
6 l/J6 - 4116
41
7
l/J7+ ~ =13j5 =F;j5 l/J 7 - 41\ = 29 = L7
417
8 l/J S + -
1 = 47 = Ls = 21j5 =~ j5
8 l/JS _ 41ls
41
9
l/J9+ ~9 =34j5 =F9j5 l/J9 - 41~ = 76 = L9
41
10 l/J 10 +\0
1 = 123 = L IO = 55j5 = F;o j5
l/JIO- -'-
4110
41
Figure 9-4
Focusing on the Fibonacci numbers, we notice that they appear
alternately as coefficients in the sums and differences ofthe powers of
j5+, 1 j5-1
l/J = - - and - = - - . For even-number powers they appear as the
2 41 2
difference, and for odd-number powers they appear as the sum. This
can be handled by using -1 to various powers; since when -1 is taken
to an odd power, the result is negative, and when it is taken to an even
power, the result is positive. This can be summarized with the
following expression:
1
1/>" -(-1)";= ( - 2 -I .J~>+ J" -(-I)" (.J5
- 2--I J"
=( .J52+1} -C- .J5} 2
= 1/>" - ( ; }
Remember, in figure 9-4 each of the Fibonacci numbers was

multiplied by .J5, so in order for us to have an expression equal to the
Fibonacci numbers, we need to divide our result by .J5 . Therefore,
A summary of this division is seen in figure 9-5.
Js[¢n -( -;J]
n Fn
~n -*J
-(
I ~- (-;)
= 1.J5 =1
Js[¢-( -;)]
-;I) = 1.J5
2 =1
~2 -
( 2
Js[¢' -( -;)]
3
~3_(_;)3 = 2.J5 =2
Js [¢' -(-; )']
-;I} Js [¢' -(-; J]
4 = 3.J5 =3
~4 - (
-;I) Js [¢' -(-; J]

5 = 5.J5 =5
~ 5- (
-;I} Js [¢' -(-; J]

6 = 8.J5 =8
~6 - (
7 =13
~7_(_*)7
=13.J5
Js[¢'-(-;)']
8 =21
~8_(_*)8 Js[¢' -(-; J]
= 21.J5
~9_(_*)9 Js [¢' -(-; J]

9 = 34.J5 =34
Js[¢'O -( -;r]
= 55.J5
~'O_ (_*)10
\0 = 55
Figure 9-5
The test to see if we, in fact, have developed a fonnula for gen-
erating any Fibonacci number is for us to apply the fonnula to our
now-familiar recursive relationship Fn+2 = Fn + F n+ l .
This would require a proof by mathematical induction. (For
this we refer you to appendix B.)
By substituting the value of cP in [",n -( -ir1, we get the Binet
formula in terms of real numbers.
F. ~ Js[-' -Hl} Js[( ~2+1 J-[ -+- n

~ Js[C\~J _(1-2~rJ
So now we shall use this formula. Let's try using it to find a
Fibonacci number we would ordinarily not find by using the
recursive definition of the Fibonacci numbers-that is, writing out
the Fibonacci sequence until we get to the 128th number. Applying
the Binet fonnula for n = 128 gives us:
F,28=JsCPI28 __ 12H 1=Js [( 1 \15 )128 _(1 - 15 )128]

[ ( ~) 5
2
5
= 251,728,825,683,549,488, 150,424,261
(You might want to check this with the listing in appendix A.)
Thus we have the Binet formula:
which will give us any Fibonacci number for any natural number n.
Let's stop and marvel at this wonderful result. For any natural
number n the irrational numbers in the fonn of 15 seem to disap-
pear in the calculation and a Fibonacci number appears. In other

words, the Binet formula delivers the possibility of obtaining any
Fibonacci number with the aid of the golden ratio <p.
At this point you may be wondering about the practicality of
the Binet formula. Well, you are right if you surmised that we
would likely obtain these larger Fibonacci numbers with the aid of
a computer. However, there is an intrinsic value to knowing that
the Binet formula exists and that it is possible to get the Fibonacci
numbers in two ways: explicitly and recursively-just as we saw
earlier was the case with the square numbers.
The Binet Formula for Lucas Numbers
It would be only fitting that there be an analogous formula to find a

Lucas number without listing the Lucas sequence up to the number
sought. Using for Lucas numbers a similar argument as the one we
used for the Fibonacci numbers, we can get the Binet formula from
figure 9-4. There we see that when n is odd, the Lucas number
appears in the difference expression <p n - ~,
rpn and when n is even,
the Lucas number appears in the sum expression <p n + ~.
rpn We can
state this mathematically (as we did before) with the following:
L =,t,n+(_I)n~
n ~ ¢n
That means (see the Binet formula):
( 1+ vis ( 1 _ vis
Ln = <pn + (-1) .
n (l)n
1> = <p n + (l)n
-1> = l-2 - )n + l-2 - )"
(for all n E N).4
To check this "inductively," we can test it for a few values of

n. Let us try one odd and one even number.
4. This means for all natural numbers. Yet, in this case, we will also include O.
For n = 3, we get:
3 1
=¢ - - = 4
¢3
(We can use the chart in figure 9-4, rather than doing the
computation again.)
For n = 6, we get:
Just to get a better feeling about how to evaluate these beyond

the chart, we shall evaluate the Binet formula for the eleventh
Lucas number, LII .
L =,/,11 +(-1)11 .(~)II I

=¢II _ _
II 'I'
= [ .J5 + I
= 199
2
r-[ r¢
.J5 -1
2
¢"
89.J5 + 199
2
89.J5 -199
2
which is the eleventh Lucas number. So you can see how we can
get any Lucas number without generating the sequence to that
point-which would be potentially a very tedious task! Thus we
have a Binet formula for generating Lucas numbers as well as the
Fibonacci numbers.
Finding Individual Fibonacci Numbers
Suppose you have just found the twenty-fifth Fibonacci number

and you would like the twenty-sixth Fibonacci number-without
knowing the twenty-fourth Fibonacci number. We could use the
Binet formula, but there is another way that might be simpler.
Since we know that the ratio of two consecutive Fibonacci num-
bers is approximately the golden ratio, l/J = 1.618. That is,
F
~ = 1.6 1803399
F 25
Therefore, we get F26 = 1.61803399· F25 , and then we round off the
answer. So in order for us to find the twenty-sixth Fibonacci num-
ber, we just multiply the twenty-fifth Fibonacci number by
1.61803399 and then round off the answer: 75,025 . 1.61803399
"" 121,393.0001, which when rounded off is 121,393, which is the
twenty-sixth Fibonacci number. You can easily see-by try-
ing-that this method of finding a Fibonacci number, given only
the one before it, can be used for the Fibonacci numbers after the
first one-that is, for finding F", where n > 1.
Another Way to Find a Specific Fibonacci Number

(Using a Calculator or a Computer)
Finding a specific Fibonacci number can be done in several ways,

as we have shown so far in this chapter. However, the use of a
computer or a calculator presents us with yet another option. If we
are asked to find the thousandth Fibonacci number, F;ooo, then by
using the definition-even with the use of technology-it would
be a laborious task. We would have to use the definition:
Fa = 0; F; = 1; and Fn = Fn_, + Fn_2 ' where n > 1, and then find all
999 numbers that precede F;ooo.
Consider the following two relationships:5

2
F2n - 1 = F}_l + Fn and F2n = Fn (2Fn_1 + Fn)
With these two relationships, we need only Fn and Fn_1 to
compute both F2n and F2n - 1 • So here is how we would go about
finding F;ooo' using these relationships (with technological assis-
tance naturally):
To find F; 000' we must first find F,oo and F.99 .

To find F,oo and F.99' we must first find F"o and F249 ·
To find F"o and F249 , we must first find F;24 and F;2S'
To find F;25 and F;24' we must first find P"" P"2' and F63 ·
To find P"" P"2' and P"3' we must first find F,o' F", and F,2'
To find F,o' F", and F,2 we must first find F;4' F;" and F;6'
To find F;4' F;s' and F;6 we must first find P", F" and F".
To find P" , F" and F" we must first find F2 , F" and F. .
To find F" F" and F. we must first find F'.;, F;, and F2 •
Remember, F; = F2 = 1, and F" = O.
Although we had to compute 22 Fibonacci numbers, it is still a
lot less work than having to compute 999 Fibonacci numbers to get
the thousandth Fibonacci number.
Testing for Fibonacci Numbers
Now let's look at the situation in reverse. Suppose we were given a

number and we wish to determine if this number is a Fibonacci
number. There is (curiously enough) a method for testing if a num-
ber is a member of the Fibonacci sequence. The test goes this way:
5. The first of these relationships was considered in chapter I, item 9 on page

43. The second relationship is new to us but also holds true. For a proof of this,
see appendix B.
A number n is a Fibonacci number if (and only it)

5n 2 + 4 or 5n 2 - 4 is a perfect square. 6
We do this without a proof, but we can test this with some of

our familiar numbers, say, at first, n = 5. We can then calculate that
5.5 2 - 4 = 125 - 4 = 121, which is a perfect square, namely, 112.
Therefore, 5 is a Fibonacci number, since it passes this (square
root) test. Indeed, we know 5 = Fs. Now we choose n = 8. We can
then calculate that 5.8 2 + 4 = 320 + 4 = 324, which is a perfect
square, namely, 18 2 • Therefore, 8 is a Fibonacci number since it
passes this "square root" test. Indeed, we know that 8 = F6 . We
should also note that since 4 is not a Fibonacci number, it will fail
this square root test. That is, 5n 2 ± 4 is not a perfect square, when
n=4: 5.4 2 +4=84, and 5.4 2 -4=76. You might want to try a
few more examples to convince yourself that this powerful test
really works.
We have now seen that the Fibonacci numbers~which are so
closely tied to the sequence from which they emanate---can also be
treated as individual entities. We can now identify specific
Fibonacci numbers and can test to see if a given number is, in fact,
a member of the Fibonacci number sequence.
6. This means the following: if 5n 2 + 4 or 5n 2 - 4 is a perfect square, then n is a

Fibonacci number. And conversely, if n is a Fibonacci number, then 5n 2 + 4 or
5n 2 - 4 is a perfect square.
Chapter 10

Fractals
U nless you are young enough to have experienced recent re-

forms in school mathematics curricula, your idea of geome-
try is probably the study of ideal shapes such as lines, circles,
squares, and rectangles. More precisely, you have probably studied
the objects and mathematical relationships compiled by the Greek
mathematician Euclid (ca. 365-300 BeE), from more than two
thousand years ago. The so-called Euclidean geometry studies
points, lines, planes, and objects such as circles and poly-
gons-objects that for many of us precisely denote the term geo-
metrical figures. We usually think of a geometrical figure when we
recognize a circle on the surface of a wheel, or a rectangle on a ta-
bletop. After all, it is not so difficult to find approximations of
these figures in objects made by people. However, how often do
we find smooth planes, lines, or curves in nature?
Natural objects are often irregular. Their. surfaces and lines are
usually rough and fragmented. For example, compare the smooth
feel of a finished wooden table to the roughness of a tree bark. Or
compare the prickly crown of a real-life conifer, such as a pine
tree, to the solid shape that names it, the cone. Or, moreover, think
of the appearance of ragged coastlines, with all their projections
This chapter was written by Dr. Ana Lucia B. Dias, professor of mathematics at
Central Michigan University.
307
308 T HE FABULOUS FIBONACC t NUMBERS
and indentations, in comparison to that of unifonn curves such as

the c ircle, in whic h all points in its circumference lie regularl y at a
same distance from the center. We can see that although Euclidean
geometry is useful in human creations, it is definite ly not the best set
of tool s to describe and understand natural objects or phenomena.
In the late nineteenth and early twentieth centuries mathemati-
c ians such as Gaston Jul ia (J 893- 1978), Pierre Fatou ( 1878- 1929),
and Georg Cantor (184 5- 19 18) experimented with curves that
have revealed the mselves to be much more suited to represent and
model natural phenomena. Although the ir contemporaries were not
very enthusiastic abo ut the new objects, referring to them as mon-
strous and patho logica l. mathematicians and nonmathemat icians
a like for decades now have been fa scinated by the beauty of im-
ages of fracta ls made possible by computer graphics (figure 10-1 ).
Figu.-e 10-1
Some images of fractals.
With the aid of the computer, mathematician Benoit Mandel-

brot (b. 1924) showed in 1980 that pictures of the mathematical
objects created by Gaston Julia were quite beautiful , not mon-
stro us. And more important, he showed that, rather than patho logi-
THE FIBONACCI N UMBERS AND FRACTALS 309
cal, the ragged outlines and the repeating patterns of those figures
were o ft en found in nature (figure 10-2). He used the Latin word
fractus, meaning broken or fractured, to coin a word to denote the
new mathematical objects:fracrals.
Figure 10-2
In fi gure 10-2 the images to the left are pictures of real-life

scenes. Those to the right are fracta l models.
Interestingly, the Fibonacci numbers a lso arise in the analysis
of some fractal s_ In thi s chapter we are going to describe two in-
stances in whi ch Fibonacci numbers appear in two different frac-
tals, even tho ugh the construction of the Fibonacci sequence was
by no means embedded in the procedures for the creation of suc h
fracta ls- which makes the find ing of such numbers a ll the more
fascinating!
The characteristic feature of fractals is self-similarity: geomet-

ric patterns seen in the big picture of a fractal are repeated in its
parts in smaller and smaller scales.
Making fractals involves the repeated application of a geomet-
ric rule or transformation to an original figure or set of points,
which we will refer to as the seed of the fractal.
Once we determine what the generative procedure and the seed
of a fractal will consist of, we can begin constructing the fractal by
repeatedly applying the generative procedure-first to the seed,
then once again to the resulting output, and so on. Therein lies an-
other definitive aspect of fractal construction: fractals are made up
of consecutive phases called iterations. An iteration is the act of
applying one algorithm or procedure one time through in a repeti-
tive process.
When constructing a fractal, the iterations of the generative
procedure are done recursively; that is, the input of each iteration is
the output of the previous one-with the exception of the first it-
eration, which is applied to a seed. In some cases this means that
each subsequent iteration will be more cumbersome than the pre-
vious one! In those cases, programmable technology is definitely a
great help.
A fractal ideally entails the iteration of a procedure an infinite
number of times, although in practice we can iterate a procedure
only a finite number of times. We can use computers or calculators
to help us perform as many iterations as we want, which would
give us different stages in the construction of a fractal. Or we can
use mathematics to deduce what would be the result of performing
that infinite process.
Before we move on to fractals that yield the Fibonacci num-
bers, we will illustrate the generative process described above and
the appropriate terminology through a classical example, the Koch
snowflake (l904)'(figure 10-3).
For the construction of this fractal, the seed will be an equilat-
eral triangle. Because the generation of the fractal happens by suc-
1. Named after the Swedish mathematician He\ge von Koch (1870-1924).

THE FIBONACCI NUMBERS AND FRACTALS 311
cessive iterations, we will call the result of each iteration a stage in

the fractal construction.
The generative procedure will consist of erasing the middle
third of every line segment in a stage and replacing it with two line
segments of the same length (one-third of the original segment) at
an angle of 60°; this will form cusps (which look like partial equi-
lateral triangles), where before there were segments. We can see
this procedure illustrated in figure 10-3.
Figure 10-3
Generative procedure for the Koch snowflake.
Each iteration will consist of applying the fractal-construction

procedure to each line segment in a stage of the fractal, which will
create the next. stage. Figure 10-4 shows the first two iterations in
the construction of the Koch snowflake. But now we do this on
three sides of an equilateral triangle.
Stage 0 Stage-l Stage 2
Figure 10-4
Construction of Koch snowflake.
You may want to sketch stage 3 of the Koch snowflake on a

separate piece of paper, or use a computer drawing program. Re-
member this iteration will require applying the generative procedure
to every segment in stage 2. It involves quite a bit more work than
the previous iteration. While the first iteration consisted of applying
the generative procedure to three line segments, in the second itera-

tion that number increased to twelve. In the third iteration you will
need to work on forty-eight line segments. That increase in com-
plexity is displayed in figure 10-5. With each iteration, each line
segment at a stage will be replaced by four line segments, forming a
cusp, in the next stage. So if we know how many segments there are
at a stage, we can find out the number of segments at the next stage
by multiplying that number by four. This relation can be written al-
gebraically and recursively as: Sn = 4· Sn-I (which just happens to
equal 3· 4 n ), where Sn is the number of segments at stage n, and Sn-l
is the number of segments at the previous stage.
Stage (n) Number of Line

Segments (Sn)
o 3
12 = 4 . 3
2 48 = 4 . 12
3 192 = 4 . 48
n Sn = 4· Sn_1
Figure 10-5
The replacement of each segment by a visual spike, and the ac-

celerating increase in the number of segments in this fractal, is what
gives it the main features of a fractal: the jagged appearance and the
property of self-similarity. If you zoom in on any spike, you will
find smaller and smaller copies of it. MagnifYing fractals reveals in
them small-scale details similar to the large-scale characteristics.
Another popular fractal is the Sierpinski 2 gasket (1915) (figure
10-6). Its seed is also an equilateral triangle. Each iteration consists
of splitting a triangle into four smaller equilateral triangles by us-
ing the midpoints of the three sides of the original triangle as the
new vertices, then deleting the middle triangle from further action,
2. Named after Waclaw Sierpinski (1882-1969), a Polish mathematician.

that is, a quarter of the area is removed. The construction of the

fractal continues by iterating this procedure over and over: from
every triangle formed, a triangle is removed from its interior. This
results in not only a rough, fragmented surface, but also in self-
similarity-the two main features of fractals.
Stag.. 0 Stag.. 1 Stage 2 Stage 3
Figure 10-6
Construction of the Sierpinski gasket.
We are now ready to encounter our protagonist, the Fibonacci

numbers. In fractals, as well as elsewhere, these numbers make
intriguing incidental appearances.
One of such unexpected sightings of the Fibonacci sequence oc-
curred in the examination of a recently created fractal, the Grossman
Truss. 3 Let us see how this fractal is constructed: we start with an
isosceles right triangle as the seed. An isosceles right triangle has two
sides of equal length forming a right angle (figure 10-7).
Figure 10-7
Stage 0 of the construction of the Grossman truss.
3. The creation of this fractal and a mathematical analysis of its properties were
reported by George W. Grossman, in "Construction of Fractals by Orthogonal
Projection Using Fibonacci sequence," Fibonacci Quarterly 35, no. 3 (August
1997): 206-24.
3 14 1)iE FABULOUS FIBONACCI NUMBERS
We will now drop a perpendicular from its right angle until it in-
tersects the opposite side. We then bounce that perpendicular clock-
wise, until it meets the opposite side, again orthogonally or perpen-
dicularly. Finally, we will delete the triangle created at the interior of
the figure. This completes the first iteration of the generati ve proce-
dure. The result is a new figure, in which the original isosceles right
triangle appears split into two other triangles, with a triangular gap
between them. If you examine this figure carefully, you will see that
all triangles formed are also isosceles right triangles-Ihat is, they are
similar to the original seed, but they appear in different positions
(figure 10-8).
Figure 10-8
Stage t of the construction of the Grossman truss.
It 's time to move o n to the second iteration. In this fractal.

each iteration consists of repeating the procedure described above
in the largest triangles at each stage. Take a look at the triangles
still existing on Stage I (remember the middle triangle has been
deleted and is no longer part of the figure-a blank rema ins in its
place). You can see that there are two triangles of different size.
Therefore, th e next iteration will consist of applying the genera-
tive procedure to the largest and only to the largest triangle. The
result is shown in figure 10-9.
THE F I BONACCI NUMBERS A ND FR ACTALS 3 15
Figure 10-9
Now Ihat o ur fractal has another gap in it. what can we say
about the remaining triangles? Are they all equa lly sized? Carefu l
inspection of the shaded triangles in fi gure 10-9 reveals thai there
are two sizes of triang les at this stage: two large ones and one
small one. So for the next iteration, the generative procedure will
need to be perfonned on the two large triang les.
The construct ion o f this frac tal proceeds in thi s way, always
apply ing the generative procedure to the largest triang les at a stage.
The result is a nice-looking fracta l. in which the basic shape shown
in figure 10-9 can be repeatedl y seen in smaller scales and at dif-
ferent positions throughout the fra clal. Figure 10-10 shows the
fractal after eight iterations.
Figure 10- 10
3 16 TH E FABULOUS FIBONACCI NUMBEI<S
In figure 10-10, as well, we can see that the remaining triangles

come in two sizes. Can you tell how many triangles there are of the
larger and how many of the smaller sizes? Maybe it will help to look at
the fractal after fewer iterations. Figure 10-11 shows the Grossman
truss after four iterations. Again, there are shaded triangles of two sizes.
They have been filled with different shades of gray to help USCOWlt.
Figure 10·11
The Grossman truss after 4 iterations.
Ifwe loo k at the shaded triangles in figure 10- 11 . we see five

larger triangles and three smaller triang les. These numbers happen
to be the fifth and fourth terms in the Fibonacci sequence.
What happens after one more iteration? The number of
larger triang les is now e ig ht. and the re are five smaller triangles
(figure 10- 12).
Figure 10-12
Stage 5 of the construction of the Grossman truss, showing the
eight larger triangles in the dark shade of gray and the five
smaller triangles in the lighter shade.
TH E FIBONACCI N U MBERS AN D FRACTALS 3 17
Wilh trained eyes, you should be able to find that in figure

10·10 there are twenty·one smaller triangles and thirty·four larger
ones. This pattern continues indefi nitely- the number of triangles
of a like size in the Grossman truss is a lways a Fibonacci number!
There are more places where the Fibonacci numbers appear.
We can also shift our gaze to the triang les that have been deleted
from the fractal - the empty triangular gaps. How many s izes are
there at a determined stage, and how many of each size?
Figure 10-13 shows three cop ies of Stage 3 of the Grossman
truss. How many different-sized triangular gaps are there in Stage
3? There is one gap of the largest size; one gap is the second-
largest; and there are two of the smaller gaps. Do you see a pattern
being fonned?
Figure 10·13
Counting gaps in Stage 3 of the Grossman truss.
Can you predict how many different-sized gaps there will be af-
ter one more iteration? Remember, all the gaps at a stage will still be
there in subsequent iterations, since the gaps are removed from fur-
ther action. But the procedures applied to the areas of the fracla l thai
3 18 THE FABULOUS FIBONACCI NUMBERS
are sti ll filled will create more gaps, of smaller and smaller sizes.
How many new gaps will be created with each iteration?
The pattern suggested by figure 10-13 becomes even more
striking if we look at furthe r iterations. Figure 10- 14 shows copies
of Stage 5 of the frac tal, and a count of the number of gaps of each
different size.
F igu re 10- 14
Count ing gaps in Stage 5 of the G r ossma n truss.
Again, it is the Fibonacci sequence! And whereas counting tri-

angles of the Grossman truss gives us two Fibonacci numbers at
THE FIBONACCI NUMBERS ANI) FRACTALS 319
every stage, counting gaps can give us the entire sequence in a sin-
gle stage. More precisely, counting gaps with the Grossman truss
after n iterations will give us the n firs t tenns of the Fibonacci se-
quence!
What is most amazi ng about this fracta l is that the Fibonacci
sequence emerges in it even though it was not directly involved in
the procedure that generated it.
The way in which the Grossman truss was created did not hint
at a possible participation of the Fibonacci numbers in the resulting
fractal. This makes all the more mysterious and amazing the ways
in wh ich this sequence appea rs both in nature and in human -
created objects- such as fractals. It makes us wonder about the
true nature of mathematics: even though mathematical objects are
human creations, the re lationships within them seem to follow by
necessity. They emerge even though we did not put them there.
Robert Devanel has shown another amazing way of finding
Fibonacci num bers in a fraC lal. The Fibonacci sequence has made
its way into one of the most well-known fractals: the MandelbrOI
set (fi gure 10- 15).
Figu re 10·15
4. A mathemat ical ellposit ion of these concepts is reponed in Roben L. Deva-

ney, "The Fractal Geometry of the Ma ndclbrot Set," in Fractals. Graphics. and
Mathematics Edllcation, cd. Mic hae l Frame and Ben oit Mandelbrot (Wash ing-
ton, DC: Mathematica l Association of America, 2(02 ), pp.6 1-68
But first, let's see what the Mandelbrot set is. Its image is so
popular it could earn the title of the "emblem of fractal geometry."
Its strange beauty mesmerizes laymen and experts alike. But what
does that image represent? As with the other fractals we have ex-
amined, some elements are involved in its construction: a seed, a
rule or transformation, and an infinite number of iterations. But
unlike our previous examples-which were mainly geometri-
cal-the Mandelbrot set is a set of numbers. The image you see in
figure 10-15 is just a plot, in the complex plane,s of the numbers
that belong to the set.
How do we know whether a number is or is not in the Mandel-
brot set? We have to test each number to find out. This infinitely
large task can only be done with the aid of a computer, and only a
finite number of times, although a very large number of times. In
fact, it was only under the right conditions, in which Benoit Man-
delbrot's vision and intellect was combined with the environment of
IBM's Watson Research Center, that a revival of the work on this
set that had been initiated by Julia in the I 920s was made possible.
So the construction of the image of the Mandelbrot set requires
one more element besides the seed, the rule, and the iterations that
the fractals, previously discussed, also had: it involves a test of
numbers. Let us say the number we are testing is c.
The seed for this fractal is the number 0; not a triangle or a
segment, but a number, because this fractal is numerical in nature.
The rule or transformation is: square the input and add c, which
can be expressed algebraically as x 2 + c .
Suppose we want to test the number c = I. Our transformation
becomes: x 2 + I .
Let us see the result of a few iterations, starting with the seed 0
as the input, and then using the output of each iteration as the input
for the next:
5. The complex plane is the two-dimensional representation of complex num-

bers with a real axis and an imaginary axis.
2
0 +I = I
I' + I = 2
2' +1 = 5
52 + 1= 26
26' + I = 677
677' + I = 458,330
We can see that the more iterations, the greater the result will
be. The terms of the sequence of numbers will increase without
bound. We say that "it goes to infinity."
Let us test for another number, c = O. With this value for c, our
rule becomes: x 2 + 0 .
Starting with the same seed 0, a few iterations will show that
the sequence will be fixed at 0:
First iteration: 02 + 0 = 0
Second iteration: 02 + 0 = 0
For each value of c, the "test" (repeatedly iterating the rule)

will tell us whether the result will go to infinity, or if it will not.
Values of c that will result in an escape to infinity are not in the
set; all the others are in the set. The image of the Mandelbrot set
is actually a record of the fate of each number, c, under this
test. 6 The key to understanding the image is to unveil the code
used. The most frequently used code for plotting the results of
these tests is to use the color black to represent those points in
the plane that are in the Mandelbrot set and to color the others
according to their "escape speed," that is, using different colors
6. In fact, figure 10-15 is only an approximation of the Mandelbrot set. In actu-

ality, we cannot know for sure whether a number c lies in the Mandelbrot set,
because to determine that with absolute certainty we would need to iterate the
"test" an infinite number of times. But even with computers, we can obviously
iterate anything only a finite number of times. But it so happens that the se-
quence formed by iterating the rule to a certain value of c may behave differ-
ently only after a very large number of iterations. So we can make our approxi-
mation better by iterating a great number of times. Still, this will not lead to
absolute accuracy.
to represent the numbe r of iterations that va lue takes to reach a

certain distance from the origin. Another traditional way of
plolting the Mandelbrot set is just to use black for points that are
in the set and white fo r those that are not.
We w ill now loo k at the image of the MandelbrOI set wilh a
categorical eye. At the core of the image, we can see a heart-
shaped figure, the main cardioid.? We can also note many round
decorations, or bulbs (figure 10- 16). We call any bu lb that is
directly attached 10 the main cardio id a primary bulb. The pri-
mary bu lbs have in turn many sma ller decorations attached to
them. Among them, we can identify what appear to be antennas
(figure 10-17).
Figure 10-16
The main cardioid and bulbs in
the Mandelbrot set.
7. A cardioid is a heart-shaped c urve generated by a fixed point on a circle as it

...,lls around another circle of equa l radius.
TilE FIBONACCI N U MBERS AND FR ACTALS 323
F igure 10-17
Detail of decorations in the Ma ndelbrot set.
We will call the longest of these amennas the main antenna.

Finally, the main amennas show several "spokes" (figure 10- 18).
Note that the number of spokes in a main antenna varies from
decoration to decoration. We will ca ll this number the period of
that b ulb or decorat ion. T o determine the pe riod of that bu lb, just
count the number o f spokes on an antenna. Do n 't forget to count
the spoke e manating from the primary decoration 10 the main
junction po im. Figure 10- 19 displays various primary b ulbs and
their periods.
Figure 10-18
Main antennas and their "spokes."
324 Til E FAB ULO US FIBONACCI NUMBERS
. ;.".
F igure 10· 19
Determining the period of bulbs by counting the "spokes" in
Iheir m a in anlenna.
How can the F ibonacci sequence be seen in the Mandelbrot

set? We will consider the period of the main cardioid to be I.
Then, by coun ting the spokes in the main antenna of the largest
primary bulbs, we will determ ine their period. The results of this
counting, that is, the period of the main cardioid and of some of the
largest primary bulbs, are registered in fi gure 10-20.
Figur e 10·20
The F ibonacci numbers in the Mandelbrot set.
You may be surprised to see from inspection of figure 10-IS

that the largest bulb between the bulb of period 1 and the bulb of
period 2 is a bulb with period 3. The largest bulb between a period-
2 bulb and a period-3 bulb is a period-5 bulb. And the largest bulb
between a period-5 bulb and a period-3 bulb is a period-S bulb.
Voila! Once again the Fibonacci numbers appear. Why is that so?
There are no obvious explanations. The Fibonacci numbers are not
directly related to the way in which the periods of primary bulbs
are calculated. Once again, the Fibonacci sequence makes a myste-
rious and remarkable appearance.
So where else might the Fibonacci sequence emerge? Appar-
ently, if we develop the habit of counting-counting the most di-
verse sets of objects we can think of: seeds, spikes, triangles-we
are bound to find the Fibonacci sequence in many of them.
Should you develop the habit of counting various collections of
things, you will probably make your own amazing sightings of the
Fibonacci sequence. It is mostly everywhere-all we have to do is
to look carefully.
Epilogue
N ow that you have experienced the wonders of what might be

considered the most famous sequence of numbers in all of
mathematics, you will undoubtedly be keen to search for other
manifestations of them. You should be aware that there are con-
stantly new discoveries made where the Fibonacci numbers amaz-
ingly appear. Even if some might argue that a few of our examples
might be a bit forced or contrived, no one can deny how ubiquitous
the presence of this sequence of numbers is.
We found the Fibonacci numbers in nature and art. They grew
to prominence in the seminal work by Leonardo of
Pisa-Fibonacci, who brought the ten numerals to the attention of
Western culture. The Fibonacci numbers have been found to hold
structures together in the most beautiful way as an integral part (or
generator) of the golden ratio, and they have tied together more
diverse branches of mathematics than perhaps any other entity in
mathematics. They even manifest themselves in the way we ma-
nipulate investments on the stock market.
Not only do the Fibonacci numbers demonstrate what most
mathematicians vehemently argue is the beauty in mathematics, but
they also seem to be either an integral part of what we in our society
consider beautiful or they help determine that beauty. This can be
seen in the many art works, both old and new, that exhibit the golden
ratio---itself determinable from the Fibonacci numbers. Though for
some artists and composers the golden ratio plays less of a role than
for others-and, of course, music and art are more than a mathemati-
cal formula. Its significance in the arts is undeniable. Whether it is the
327
328 EPILOGUE
beauty of these numbers or the challenge they continuously lay before

us to discover new manifestations of their appearance, the Fibonacci
numbers never cease to amaze us. Mathematicians, through the Fibo-
nacci Association, established in 1963, continue to find applications
of these numbers both inside and outside the field of mathematics and
publish them four times a year.
The Fibonacci numbers are very easy to remember, but if you
cannot commit, say, the first twelve to memory (remember F" just
happens to be equal to 122 = 144), we have the additive process to
rely on, and for very large Fibonacci numbers, the convenient Bi-
net formula will bring us there immediately. As you look for the
Fibonacci numbers in any field, remember, care must be taken not
to overstate their application. Not all beauty is consistent with the
Fibonacci numbers-as we have found in music-and not all art
relies on the golden ratio for its beauty. Yet there is a role for the
Fibonacci numbers in both of these fields as long as we observe it
objectively and not force their appearance. There are plenty of
natural illustrations, so we do not have to artificially claim the
numbers' presence artificially.
We hope, by reading this book, you will not only have been
motivated to search for new appearances of the Fibonacci numbers,
but will also have developed anew, or renewed, love for mathe-
matics and a greater appreciation of the beauty of this most im-
portant science-referred to by the famous German mathematician
Carl Friedrich Gauss (1777-1855) as "the queen of science."
Afterword
By
Herbert A. Hauptman
T he Fibonacci numbers have fascinated me for decades. Yet

when I was asked to provide my thoughts about them, I de-
cided to see if I could discover patterns independent of what others
may have found earlier. One never knows if his discoveries are
new, for when one asks mathematicians if some of these discover-
ies are new, one often gets the answer: "Well, I haven't seen this
before, but that doesn't mean that it hasn't been published some-
where." Since 1963 the Fibonacci Association (mentioned else-
where in this book) has published a quarterly journal on the seem-
ingly endless relationships involving the Fibonacci numbers. I
offer here some of my "discoveries" in the hope that they will pro-
vide further insights and motivate the reader to seek others.
1. Some Plausible Speculations
We begin with what must surely be one of the most fundamental

concepts of all: the idea of unity, or the number one. We continue
in this vein with what is, just as certainly, the most primitive ar-
ithmetical operation: the operation of addition. We obtain in this
way the subject matter of this book, a construct of great beauty, the
wonderful sequence of numbers known as the Fibonacci numbers,
1, 1,2,3,5,8, 13, .... As the reader already knows, each number
in this sequence, after the first two, is obtained by addition of the
329
330 AFTERWORD
preceding two. Can one imagine a simpler construction? Yet this

sequence of numbers has the most remarkable properties, many of
which are described in this book. In this afterword, adopting a
somewhat different point of view, we give a brief overview of the
divisibility properties of the Fibonacci numbers and hint at the ex-
citing further developments, which the reader is invited to explore. I
1.1. Notation
Denote by Fn the nth Fibonacci number, so that FI = 1, F2 = 1,

F3=2,F4=3, Fs=5, ....
1.2. The Even Fibonacci Numbers
We ask a simple question. Which of the Fibonacci numbers are

divisible by 2? Consulting the table of Fibonacci numbers, we ob-
serve that F 3, F 6, F 9, F 12 , F 1S , •.• are all divisible by 2 and, in fact,
these are the only even Fibonacci numbers. We are tempted there-
fore to conjecture that the even Fibonacci numbers consist of the
set F 3, F 6, F9, F 12 , F 1S , •••• We note in addition that the subscripts
3,6,9,12, ... of the even Fibonacci numbers themselves form an
arithmetic progression with initial term and common difference
equal to 3; in other words, it consists of all the numbers divisible
by 3.
Encouraged by this discovery, we ask next which Fibonacci
numbers are divisible by 3.
1.3. The Fibonacci Numbers That Are Divisible by 3
Again consulting the table, we observe that those Fibonacci num-

bers that are divisible by 3 are F4 = 3, Fs = 21, FI2 = 144, FI6 = 987,
F 20 = 6,765, .... The subscripts 4,8, 12, 16,20, ... again form an
arithmetic progression, this time with the initial term and common
difference equal to 4, and therefore consist of all numbers that are
1. This was briefly mentioned in chapter 1 (pages 47--48, item 12)

AFfERWORD 331
divisible by 4. We conjecture therefore that the collection of the

subscripts of all the Fibonacci numbers that are divisible by 3
themselves consist of all numbers divisible by 4.
1.4. The Fibonacci Numbers Divisible by 4
The pattern is now clear. We now find the Fibonacci numbers di-
visible by 4 to be F6 = 8, Fl2 = 144, Fl8 = 2,584, F24 = 46,368, ... ,
the subscripts of which form an arithmetic progression with initial
term and common difference equal to 6. Hence the subscripts of all
Fibonacci numbers that are divisible by 4 are themselves divisible
by 6 and in fact consist of all the multiples of 6.
Proceed as in 1.2, 1.3, and 1.4 to deduce the following: the Fi-
bonacci numbers Fn that are divisible by 5 (or 6, or 7, respectively)
have subscripts, n, that are divisible by 5 (or 12, or 8, respectively)
and, in fact, consist of all the multiples of 5 (or 12, or 8, respec-
tively).
2. The Minor Modulus m(n)
In this section we formulate in greater generality what we have

learned in the previous section. Let n be any positive integer. In
view of our previous work, we make the fundamental assumption
that there exist infinitely many positive integers x such that the
Fibonacci number Fx is divisible by n. The smallest such number x
clearly depends on n. We call it the minor modulus of n and denote
it by m(n). Thus m(n) is the smallest positive integer x such that Fx
is divisible by n. Referring to our previous work, we deduce, for
example, that m(2) = 3, m(3) = 4, m(4) = 6, m(5) = 5, m(6) = 12,
and m(7) = 8. A more extensive table of values of the minor
modulus m(n) is readily derived as explained in the previous sec-
tion, and is shown in table 1. It is assumed that the reader has ex-
tended this table to n = 200 at least so he can confirm to his satis-
faction the several conjectures made here.
332 AFfERWORD
Table of m(n), 1 :s n :s 100
n men) n men) n men) n men) n men)
1 1 21 8 41 20 61 15 81 108
2 3 22 30 42 24 62 30 82 60
3 4 23 24 43 44 63 24 83 84
4 6 24 12 44 30 64 48 84 24
5 5 25 25 45 60 65 35 85 45
6 12 26 21 46 24 66 60 86 132
7 8 27 36 47 16 67 68 87 28
8 6 28 24 48 12 68 18 88 30
9 12 29 14 49 56 69 24 89 11
10 15 30 60 50 75 70 120 90 60
11 10 31 30 51 36 71 70 91 56
12 12 32 24 52 42 72 12 92 24
13 7 33 20 53 27 73 37 93 60
14 24 34 9 54 36 74 57 94 48
15 20 35 40 55 10 75 100 95 90
16 12 36 12 56 24 76 18 96 24
17 9 37 19 57 36 77 40 97 49
18 12 38 18 58 42 78 84 98 168
19 18 39 28 59 58 79 78 99 60
20 30 40 30 60 60 80 60 100 150
Table 1
We summarize our major conjecture as follows: Let n be any

positive integer; denote by m = men) the minor modulus of n. Then
the Fibonacci number Fx is divisible by n if and only if x is divisible
bym.
AFfERWORD 333
The minor modulus men) itself has a number of interesting prop-

erties that inspection of table I makes plausible and which we sum-
marize shortly. First however we digress to give a brief account of
the prime numbers that playa special role in this development.
3. The Primes
Definition: Every number (> 1) is divisible by itself and unity. If

these are its only divisors, it is said to be prime. Thus, 17, which is
divisible by no numbers other than 1 and 17, is prime. Similarly, 2,
3, 4, 7, I 1, and 13 are also primes. A famous theorem attributed to
Euclid states that there are infinitely many primes. Numbers like
6 = 2· 3 or 8 = 2 3 are, on the other hand, not primes since 6 is di-
visible by 2 (in addition to 1,3, and 6) and 8 is divisible by 2 (in
addition to 1, 4, and 8). They are said to be composite. The reader
may readily confirm that the first few composite numbers are 4, 6,
8,9, 10, and 12. Clearly all primes, other than 2, are odd.
4. m(pk) Where p Is Prime and k Is a Positive Integer
The minor modulus ofa prime power is of particular importance in

the sequel. Reference to table 1 makes plausible the following
conjectures:
4.1.
m(21) = m(2) = 3, k = 1
m(22) = m(4) = 6, k = 2
m(2k) = 3 . 2k-2, if k > 2
4.2.
4.3.
If p is an odd prime, m(pk) = pk-I m(p)
334 AFTERWORD
For example, to confirm 4.1, observe that
when k = 1, F3 = 2, which is divisible by 2, and x = 3 is the

smallest value of x for which Fx is divisible by 2' = 2, whence, by
definition, m(2) = 3;
when k = 2, F6 = 8, which is divisible by 2 2 , and x = 6 is the
smallest value of x for which Fx is divisible by 22 = 4, whence
m(22) = 6;
when k = 3, F6 = 8, which is divisible by 2 3 , and x = 6 is the
smallest value of x for which Fx is divisible by 2 3 = 8, whence
m(2 3 ) = 6;
4
when k = 4, F'2 = 144, which is divisible by 2 , and x = 12 is
the smallest value for x for which Fx is divisible by 24 = 16,
whence m(24) = 12;
when k = 5, F24 = 46,368, which is divisible by 2 5 , and x = 24
is the smallest value of x for which Fx is divisible by 2 5 = 32,
whence m(2 5 ) = 24;
when k = 6, F48 = 4,807,526,976, which is divisible by 2 6 , and
x = 48 is the smallest value of x for which Fx is divisible by 2 6 =
64, whence m (2 6 ) = 48;
and so on.
Next, to confirm 4.2, observe that when k = 1, F5 = 5, which is

divisible by 5, and x = 5 is the smallest value of x for which Fx is
divisible by 5.
Whence, by definition, m(5) = 5, and when k = 2, 52 = 25,
F 25 = 52 ·3,001, which is divisible by 52, and x = 25 is the smallest
value of x for which Fx is divisible by 52.
Whence, again by definition, m(5 2 ) = 52, and when k = 3,
53 = 125, FI25 = 5 3 .3,001. 158,414,167,964,045,700,001, which is
divisible by 53, and x = 125 is the smallest value of x for which Fx
is divisible by 53.
Whence, by definition, m(5 3 ) = 53, and when k = 4, 54 = 625,
F 625 = 54. P, where P is a product of 5 primes, which is divisible
by 54.
Whence, by definition again, m(5 4 ) = 54, and so on.
AFTERWORD 335
Finally, we shall confirm 4.3 for odd primes, p = 3, 5, 7 and

k = 1, 2, 3, and leave to the reader the pleasure of confirming 4.3
for primes 11, 13, 17, 19andk= 1,2,3.
Thus when p = 3, k = 2 we have from table 1 and the Fibonacci

table:
m(3) = 4, m(3 2) = 12 = 3' . m(3) = 3·4
when p = 3, k = 3,
m(3) = 4, m(3 3) = 36 = 32 . m(3) = 9·4
whenp = 3, k = 4,
m(3)=4, m(3 4 ) = 108 = 3 3 . m(3) = 27·4
when p = 5, k = 2,
m(5) = 5, m(5 2) = 25 = 5' . m(5) = 5·5
when p = 5, k = 3,
m(5) = 5, m(5 3) = 125 = 52 . m(5) = 25·5
when p = 5, k = 4,
m(5)=5, m(5 4 ) = 625 = 53 ·m(5) = 125· 5

when p = 7, k = 2,
m(7) = 8, m(72) = 56 = 7' . m(7 ) = 7·8
whenp= 7, k= 3,
m(7) = 8, m(73) = 7 2 m(7) = 49·8 = 392
and we readily confirm from the table that Fx is divisible by

7 3 = 343 when x = 392 but for no smaller value of x.
336 AFTERWORD
5. Special Primes p = 10n ± 1, q = 10n ± 3 If n Is a

Positive Integer
If, upon division by 10, the prime p leaves the remainder 1 or 9, it is

said to be a prime IOn ± 1, because in this case there exists an inte-
ger n such that p = IOn ± 1. For example, the first few primes IOn ±
1 are:p= 11, 19,29,31,41,59,61,71, and 79, since 11 = 10·1 + 1,
19 = 10· 2 - 1, 29 = 10· 3 - 1, 31 = 10· 3 + 1, and so on.
Note that the last digit of a prime 10 ± 1 is either 1 or 9.
If, on the other hand, upon division by 10 the prime q leaves the
remainder 3 or 7, it is said to be a prime IOn ± 3 because in this case
there exists an integer n such that q = IOn ± 3. For example, the first
few primes IOn ± 3 are q = 3, 7, 13, 17,23,37,43,47, and 53.
Note that the last digit of a prime IOn ± 3 is either 3 or 7.
Every odd prime other than 5 is either p = IOn ± 1 or
q = IOn ± 3. The special importance of primes p = IOn ± 1 and
primes q = IOn ± 3 is due to the following, 5.1 and 5.2, which the
reader is invited to confirm.
5.1. Ifp Is a Prime IOn ± 1 Then m(p) Is a Divisor ofp - 1.
Consult table 1 to confirm for the primes p = 11, 19, 29, 31,41, 59,
61,71,79, and 89.
5.2. If q Is a Prime IOn ± 3 Then m(q) Is a Divisor of q + 1.
Consult table 1 to confirm for the primes q = 3, 7, 13, 17,23,37,

43,47,53,67, 73, 83, and 97.
6. A Simple Exercise
For each fixed positive integer n = 2, 3, 4, ... consult table 1 to

calculate the sequence n, men), m(m(n», m(m(m(n»), ....
What conclusion do you draw?
AFTERWORD 337
7. Greatest Common Divisor
We conclude this afterword with the remarkable formula described

below in section 8. We shall need a short preliminary digression.
The greatest of all common divisors of two specified positive
integers rand s is denoted by g = (r, s).
It has the property that any common divisor of rand s is also a
divisor of g.
Thus, if r = 30, and s = 75, then g = (30, 75) = 15. We see that
15 is the greatest of all common divisors of 30 and 75. Observe
that the common divisor 5 of 30 and 75 is also a divisor of 15, but
not the greatest one.
If rand s have no common divisor other than 1, we say that r
and s are relatively prime and write (r, s) = 1.
8. A Remarkable Formula
We conclude with four remarkable properties of the Fibonacci

numbers. In number 4 above, we described properties of m(2k) and
m(pk), where p is an odd prime, which, together with table 1, fa-
cilitates the calculation of men), when n is an arbitrary power of a
prime. Here we give a formula which, again by means of table 1,
facilitates the calculation of men) for arbitrary integers n. This for-
mula expresses m(rs), where rand s are relatively prime, in terms
of mer) and m(s). The reader may readily confirm that if (r, s) = 1
then
m(rs) = m(r)m(s)
(m(r), m(s))
In other words, if rand s are relatively prime, that is, they have
no common factor other than unity, then m(rs) is equal to the prod-
uct of m(r)and m(s) divided by the greatest common divisor of mer)
and m(s). Confirm this formula by consulting the extended table 1.
Clearly this formula facilitates the calculation of men) once the val-
ues of m(2k) and m(pk), where p is an odd prime, have been found.
338 AFTERWORD
Let me again summarize the ideas involving the minor

modulus men).
Let n be any positive integer. Then there exist infinitely many
positive integers x such that the Fibonacci number F, is divisible
by n. The smallest such number x clearly depends on n. We call it
the minor modulus of n and denote it by men). Then men) is the
smallest positive integer x, such that Fx is divisible by n.
Let n be any positive integer. Then F, is divisible by n if and
only if x is divisible by men). Alternatively, all solutions of the
congruence
Fx =0 (mod n)
are given by
x =0 (mod men»~
We also have for m(pk), where p is prime, the following:
1. m(2) = 3, m(4) = 6, if k > 2, m(2k) = 3. 2 k- 2

2. m(5 k ) = 5k
3. Ifp is an odd prime, m(pk) = mk-1m(p)
4. If P is a prime IOn ± 1 , then m(p) divides p - 1
5. If q is a prime IOn ± 3, then m(q) divides q + 1
. m(r)m(s)
Furthermore, tf(r, s) =1, then m(rs) = ---'--'--"":"'-'---
(m(r), m(s»
We now consider the primitive factors. For the equation

m(x) = x, there are infinitely many solutions. They are all given by
k
x = 1,5,25, 125,625, ... , 5 , k = 1,2,3, ...
x =12, 60, 300,1500,7500, ... , 12·5 k , k = 0,1,2,3, ...
We call these solutions of the equation m(x) =x, the primitive

numbers.
Consider the sequence men), m(m(n», m(m(m(n» .... If we
let n be any positive integer, then the sequence men), m(m(n»,
AFTERWORD 339
m(m(m(n») ... tenninates in a primitive number, the second re-

markable property of the Fibonacci numbers.
If one represents this function of the minor modulus m(n)
graphically, astonishing pictures arise.
So let n be a (given arbitrary) number. Then the smallest index
of x E N with n I Fx (where Fx is a Fibonacci number) is said to be
the minor modulus m(n) of n.
At first, table I seems quite "chaotic" and also in its graphic
presentation (figure I).
m(n)
Figure 1
It rather looks like "fly droppings"-concentrated in the lower

half of the quadrant, which lies under this one of diagonals going
from the lower left to the upper right. If the points are connected
(see figure 2), we recognize hardly any pattern.
340 AFTERWORD
m(n)
Figure 2
If we were to continue table I to about n = 3,200, then radial

parts can be recognized in the graph-truly an amazing result,
which gives usthe third remarkable property of the Fibonacci num-
bers (see figure 3).
men)
14000
12000
10000
8000
6000
-------~-
;;~::~:;;£,j::::~;. .-'
4000
2000
o
o 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000
Figure 3
So you can see, even if a pattern does not immediately emerge,

you may have to look a bit further until one appears. There are
AFTERWORD 341
boundless beauties embedded in the Fibonacci numbers. I leave it

to the reader to discover others.
The Ultimate Word
In this afterword we have discussed some of the divisibility prop-

erties of the Fibonacci numbers, and to this end have introduced
the concept of the minor modulus m(n). This work naturally raises
the question: What can we say about those Fibonacci numbers that,
upon division by the specified number n, leave the fixed remainder
r* O? Readers are invited to explore this question for themselves
and to make some of the exciting discoveries that await them. The
work is not easy but will be well rewarded.
As I read this delightfully informative book, I was motivated to
search for new mathematical relationships regarding the Fibonacci
numbers, some of which I just presented. I felt my search was in-
complete until I could relate these ubiquitous numbers to the field
that has occupied a large portion of my professional
life-crystallography-and for which some of my work was ac-
knowledged by the Nobel prize. My search turned up the follow-
ing: a crystal may be defined mathematically as a triply periodic
array of (or a triply periodic function, the electron density func-
tion) points (atoms). As such, it may have certain elements of
symmetry, for example, a center of symmetry, a mirror plane, a
two-fold rotation axis, and so on. Certain elements of symmetry,
consistent with its triply periodic structure, are, on the other hand,
forbidden-a simple example of forbidden symmetry is the five-
fold symmetry element.
When, nevertheless, a crystal, the aluminum-manganese alloy,
AI6Mn, was observed in 1984 to have the forbidden fivefold sym-
metry, the crystallographers were presented with a dilemma. The
dilemma was resolved with the recognition that this alloy consti-
tutes a new state of matter, since called a quasi-crystal, having
some properties of crystals and others of noncrystalline matter,
such as glass.
342 AFTERWORD
Quasi-crystals may be described by means of a quasi-lattice;

for example, in the one-dimensional case, by the Fibonacci lattice
generated by the two basis vectors of lengths a and b defined by
successive application of the rule Sn+1 = Sn_ISn' n = I, 2, 3 ... ,
where
So = a, SI = b, a = ¢b, where ¢
.f5 + 1
= - 2 - = 1.618034 ...
Then, as is readily confirmed,
So = a, S, = b ,S, = ab ,SJ = ab' ,S, = a' b J

,S, = a' b' ,S6 = as b B , •••
In the Fibonacci lattice, the relationship of which to the sequence

of Fibonacci numbers is clear, we have the fourth remarkable
property of the Fibonacci numbers.
This example illustrates, once again, how often, and in what
unexpected ways, the most abstract mathematical construct may
throw light on a real-world phenomenon.
For some readers, my musings in the realm of the Fibonacci
numbers may be a bit challenging. Yet I hope that each reader will
see this fine book as a springboard to further investigations into the
Fibonacci numbers and perhaps their relationship to the reader's
field of interest. You have here a source for boundless entertain-
ment and an entree into the beautiful world of mathematics.
Appendix A
List of the First 500 Fibonacci

Numbers, with the First 200
Fibonacci Numbers Factored
n Fn Number Factors for the first 200
of digits
(I)
,I)
(I) pnme
(I) pnme
, (I)
(I)
pnme
2'
, lJ
21
(2)
(2)
pnme
37
9 34 (1) 217
10
II " (2)
(2J
511
pnme
Z'3!
12
lJ ''''
23J
(3)
(3) pnme
1329
"
15
16
377
610
(3)
(3) 2561
11 '"
1597
25114
(JJ
(4)
(4)
3747
pnme
2'1719
"
19
20
41S1
6765
(4)
(4)
37113
351141
21 '0946 (5) 213421
22 17711 (5) 119199
2J 28657 (5) pnme
24 463611 (5) 253'723
25 15025 (5) 5' 3001
12139) (6) 233521
"
27 196418
3171111
(6)
(6)
21753109
313292111
"
29
)0
514229
1132040
(6)
(6)
pnme
2'5113161
31 1)46269 (7) 5572417
l2 2178309 (7) )7472207
J3 3524578 (1) 289191101
34 570211117 (7) 15973571
JS 9227465 (7) 513141961
l6 14930352 (II) 2'3'17·19107
31 241571117 (II) 73·1492221
"
J9
40
390118169
632459116
102334155
(8)
(8)
(9)
371139349
2233135721
35711412161
"
42
4J
165Sl!OI41
267914296
433494437
(9)
(9)
(9)
2711959369
2' \3 29211421
pnme
'"45 701408733
IIJ49()3170
(9)
(10)
34389199307
251761 10944\
46 11136)11903 (10) 139461211657
41 2971215073 (\O) pnme
41107526976 (10) 253' 7 23471103
"
49
'0
77711742049
125116169025
(10)
(1l)
13 97 61611709
5' 11 \01 1513001
20365011014 215976376021
""
(II)
329512110099 (II) 3233 52! 904111
II 53)16291173 (II) 95355945741
343
344 ApPENDIX A
n F. Number Factors for the first 200

of digits
54 116261571272 (11) 2'171'1531095779
"" 139511311(,2445
225K514)3717
(121
(12)
SI!9MI474541
3" U 29 2M1 1450_~
"
58
59
365435296162
5912116729879
956722026041
(12)
(12)
(12)
2171137975483)
5'11941!9514229
35] 2710260697
'0
61
1541:1001:1755920
250473071:1.961
(13)
(13)
2'3'5113141612521
4513 55.5003497
62 405273953711111 (13) ~57 2417 3010149
65574703191142 (13) 21317421352}96111
10610209857723 J 7471010220744111
"
65
66
171676110177565
277771:1900352t111
(14)
(14)
(14)
.5 233 14736206161
2' 1191999901191101
449455702121153 (14) 26911611491429913
727234602411141 (14) 3671597357163443
117669030460994 2 1l7IQ9IKQ77211657
"
(lS)
70 190392490709135 (15) 511112971911141961
71 30)1061521170129 (15) 607)41>165)71073
72 4911454011!l79264 (IS) 2~ 3'71719 n 107 1036~1
7J 806515533049393 (15) 9375IC!986020717
74 1304969544928657 (16) 73149222154018521
"
76
77
2111485077978050
3416454622906707
5527939700884757
(16)
(16)
(16)
25' 613001230686501
337111934929134601
1,1:<991111611141132521
1:1944394323791464 (16) 2' 79233 521 1159135721
"
79 14472334024676221
234167283484676115
(17)
(17)
1~79211:10471494753
.' ~ 7 II 41 47 1601 2161 3041
3711119062373143906 (17) 217531092269437319441
"
82 61305790721611591 (17) 2711959)69370248451
"
K4
99194853094755497
160500(4)1I163670KK
(17)
(III)
(18)
prnl\e
2'3' 1129113211 2111 421 1427
~ 15979521 3415914041
"" 25969549691 1 1225K5
420196140727489673 (18) 6709144481433494437
"" 67989163763861225K
1100087778366101931
17799794160047141~9
(18)
(19)
(19)
21735142293821263937
3743 K9199 263 307 881967
"
90
91
2K800671943701116120
4660046610375530309
(19)
(19)
10091665011113218004111
2' ~ II 17 1911 61 IKI 541109441
I J' 23~ 741469 159607993
92 75401131104746346429 (19) 3 1J9 4614969211657275449
93 1220016041512111767311 (20) 255724174531100550901
94 197402742198M!223167 (20) 2971215073 6643113111179
31940434634990099905 (20) 537 IU 7(,129641 67735001
5161W70M154115l1323072 (20) 23'72347769110322073107
"97
OK
1136211434119K4l1422977
1353011152344706746049
(20)
(21)
1933119 3(j1l49K9 361040209
112997616117095997116069
99 211:1922995834555169026 (21) 21789197 191!01 (85468051),
100 3542241!4817926)915075 (21) 35' II 41 101 1514013001570601
101 573147l14401311170ll4101 (21) 74351937777011579711613
102 927372692193078999176 (21) 2' 91'11597 3469 35716376021
103 150052053020611960113277 (22) 519121 5644193512119709
104 2427IN3221!39997501!245, (22) J 7 103 233 521 904111 102193207
10' 39284137646061171165730 (22) 251361421 141961112811K2341!1
100 (,35630699300684624111113 (22) 9535594574111921111!51371
107 10211471075761371741391.' (23) 1247!03K242065050061761
10< 16641027750620563662090 (23) 2' J' 17 1953107 1095779111211427
109 2692574115011234281076009 (23) 11277211777325296754111:1417
110 43566776251:11:154844731:1105 (23) 5 II' K9 199 331 661 39161474541
III 7049252476701191251114114 (2') 2731492221 1459000305513721
112 1140593010259439705522)9 (24) 3713294721:111450310J450ll1141!1
113 1845511:125793033096366JD (24) 677 272602401466814027129
114 291161 I 126818977066911!552 (24) 2'."1132297979349541<3395419
115 41131629526120101632K411115 (24) 5131112116572441738K87'16391!1
"'
117
118
71117740794309117230203437
12649370320429973934H!022
20467111114739114(2)691759
(24)
(25)
(25)
,59.1471941:19 514229 127001!3~113
2 17233297(71)5721 395l1961!5093
35,709119093364192710260697
119 331 I64KI435169!12017IHOOIII (25) 13 1597 15951293911151!557HII121
120 535!13592549909666401!711140 (25) 253'5711233141612412101252120641
121 11670007391:1507948651:1051921 (25) 119 \l7415!1134663814455960II'I
122 1402H36665349K91529K9237(,1 (26) 4513 555003497560074112931<01
123 2269113740520068639569756!12 (26) 22711959369611541957733949701
124 367267407055057792551199443 (26) 3557241730103493020733700601
125 59425114757512643212H75125 (26) 5' 3001 ISK414167964045700001
12' 96151855463018422461177456!1 (26) 2' IJ 17 1929211421 100911249352396111
127 155576<J702205310656KI64969J (27) 21941 556!105304H227732210073
128 25 I 72H1I2568354948l1 I 50424261 (27) '747127 10872207 44HI 11I61!122011641
12. 4073057959040805531<32073954 (27) 2257541711513396391193433494437
130 659034621 58763004198249H215 (27) 5111112335212011124571147.16200161
131 10663404174917105951114572169 (21!) pnmc
132 1725375039079340637797070384 12K) 2' 3' 43 H9 1993079901 191!01 26U99601
133 2791715456571051233611642553 (2~) n37113345742293351101301942501
134 45170904956503911171401!712937 (2K) 2694021 116!1491429913 24994118449
135 730111105952221443]050203554'10 (2H) 2517 53 61 109109441 11147(,9<.154367361
13' 111I251!964471!71!1349764290611427 (29) 3767159735716344323230657239121
137 19]34702400093278081449423917 (29) prune
13' 3096059111!47965113057871!492344 (29) 2'137 139461691 H29 11!0772H657 14H5571
139 500'l53012411051!391139327916261 (29) 277 21145375011155267229061!9093
"0 H105 590009602 3 504 19720640H60 5 (29) 35 1113294171211191114196112317523121
"I 1311512013440111895336534324H66 (30) 2 1O!l21191435097 142017737 2971215073
212207101440105399533740733471 (30) 067146(65)7I07361!I!Mf,50251111399
'"
I"
3433511302711411172948702750~K337
555565404224292694404015791HOI!
(10)
(30)
H9 233 H51!1 192951141517561150496021
25"7171923471071103103611110749957121
lI9K923707001:1479989274290R50145 (30) 551422'13496199969.107.170791190201
'"
146
I"
1454411911123277268367830664(95)
23534121111124125267295259749209K
(31)
(31)
15154\1 9)75H29 116020717 111199937029
2119729342135296168709347502052673
I" 3110790 19294 7402535663090413405 I (31) 171149n2111911754018521K1(4)477963
"9 616131414771527110295113501626149 (31) 1105~7 162709400094'1 115607646594.'i77
ISO 996921667711193033110214405760200 (31) 2'~' II 3161101151300112301 11!451 23061!6501
151 161305314249045HI4157979073l1634\1 (32) 57021111666624525HII646469915H77
LIST OF FIBONACCI NUMBERS 345
n Number Factors for the first 200

of digits
'52 2('U9974KI020'HII1I4X02012)1,146549 (32) 3737 113 9349 29134601 109134639691:10401
153 4223027952699114662171(102205321<91( (32) 217' 1597 637f>021 7175323114950564593
15' 6K).lO()216290923510191112533679447 (2) 132911919922971"199886111411325219321929
15~ II()Sf>0307156090Hl72J7632754212345 (B) 5557241721701123705331:18161182771:1621
150 17I1!1903J4711511(llbK2574552H11191792 (33) 2' 3' 79 233 52185990481 135721 1221:10217041
21<'1450641'141273911549501111042104137 (11) ~ I ~ 11617 7636411 I 10424204306491346737
461040976726457153752543329995929 (33) 15792111047149475332361122672259149
'50
100
7577916111667131139247631372100066
122111.1259539411<11293000174702095995
(3,)
(14)
2 l17 953 55945741 97639037 229602768949
35·7114147160121612207304123725145626561
IlIl69l2116576126061077S5058997065597
'"
162
19K)924214061919432247!106U74196061
J2I()()~o!W94561071252479110776292056
(34)
(34) 2' 17 1953 10922693079437357791944162650261
"J SI9}9KIOB5IX027151495711685041!1I111 (34) 977411926093336551939332566223208133
'165
04 X4()4037KJ2974114IU~27437676267KOI73
1359110 I KK5M921620402N554477261:!2'IQ
(34)
(35)
3 1632789593698004833502075693702411451
2561 !!9661 19801116461474541518101900241
'60 220020566119466296922983322104048463 (35) 3576131116202401259991941153094755497
356000755459584~X96322211765l! 1316753 (35) 111104700793 1966344318693345608565721
'"
'" 5760213223542475SIH1620619H6115365216 (35) 253'7'1323298316721121:11421142714503657405113
932022077KI3K3214K494290752666t11969 (35) 233 337 119909104600155609126213229732669
'"
170
17'
150X04340016HQ79707356352739520471115
24400t.547791119111155115004349218729154
(6)
(36)
5 II 159735719521 11511551 127600313415914041
2 1737 113 797 68415483357414617608791\44361
172 .194l1I011117l1I4999156320699623170776339 (36) 36709 1444111 433494437 31319571151657K281
17) 63111117435(,13190341905763972389505493 (36) 16393437115721389678749007629271532733
17' 1033621112342l1111949822646159556021\11'O2 (37) 2' 59 173 349 194119514229947104099·3821263937
175 1672445759041379114013222756794971\7325 (37) 5' 137013(01141961172312037)0201189308301
270607401Q469569331B511691163510069157 3743471191992633071:181967930582415624111561
'" 437H5I'1114151094917.'149091117]1459~56482
(37)
23532191261110513406112970276812710260697
'"
(37)
17" 70ll459392.l'l~0511151611496091\94969925639 (37) 179 1069 16650111132 I 1:l004K 1 22235502640988369
'79 114631137t>54914676953405211(,26429782121 (38) 214111 1560119341811166409031:191:1929534613769
'"0 IlI~4770768947191162121901311521399707760
300101121454'.1(,.145)907530667147829411911111
(311)
(311)
2'3'511 1719314161 107 IK1541 2521 109441 107833420lli
8611942245311175789237231:1547574551461093
'"'
'"
'"
411551152'11444354401197201105669229197641
71156935tl5q9391111'14027251472l1I705l16117522
(311)
(311)
13' 292335217414691596079936119667151970161
2109745135550034971429734797197575711001133
'84 127127I1N74311.1433414tl97227l14116211711115163 (39) 3713946149692865725336727544995063721931163
'"
'"
2056'172303412332281742237513033465726115
13211251100117067562321196029711963445711411
(3'1)
(39)
5731492221170294551319130555690709761KI61
2' 557 2417 63799 3010349 355107494531100550901
53R522.14043030079049541971110929111030533 (39) 119373 1597 10157807)05963434099105034917037
'" 11713474505173(,11352.'1166151110.'1.'1261541111381 (39) 3563564129712150736643831111794632894751907
'""
'" 14091169790947669143.112035591975596511:1914 (40) 2131753 109421 31193335239681955921950316735037
''91
90 22111217241405037490121165140211511212007295
3691011 7032412 70663 94406116994.'1331101:1526209
(40)
(40)
5 1137113 I'll 761 93492964141611677350011173112901
41170723671313757810806256989128439975793
'92
I9J
597230427311777441355693 }1I39769202 05 3 3504
96633'113062904507750100253925251:129059713
(40)
(40)
2' 3' 7 23 47 76910117110322073167 44KI 1111625752411703
9465271192910209304320323269339768260011497
19' 15635695511016l1194910579363790217l149593217 (4)) 19338932993084911936104020956678557502141579
,9> 252990H6l1l164~H64561155119lll9I.'12743678652930 (41) 2561233 135721 1473620616111899925083749911776111
196 409347112466(26)1405961687529729615211246147 (41) 313 2997 211158111 6168709 59978606961025309469041
197 662331169353011541162l1I751:1142155705206899077 (41) 15761257959692271502656977171851071251186549
(42) 2' 17 1989 197 199991 21799901 19801 1513909185461105133
'"
'99
1071611651111971232611779261195128666735145224
173402521172797l1I31596l150372114371942044301 (42) 397436782169201002048261171378550055269633
200 28U571172'N25101400'7611'H241.'OJII677189525 (42) -'5'71141101151401216130015706019125201
201 45397.1694165307953197296969697410619233826
202 7345448671578180932349089021 10449296423351
203 1 18851856 lJ2J 1260464322058718078599 15657177
204 19230634284809441396671 14773918309212080528
205 31 1 1581989804070186099J20645726169127737705
206 5034645418285014325766435419644478339818233
207 8146227408089084511865756065370647467555938
208 13180872826374098837632191485015125807374171
209 2 I 327100234463183349497947550385773274930109
210 34507973060837282187130139035400899082304280
211 55835073295300465536628086585786672357234389
212 90343046356137747723758225621187571439538669
213 146178119651438213260386312206974243796773058
214 236521166007575960984144537828161815236311727
2 15 3826992856590 14 174244530850035136059033084785
216 619220451666590135228675387863297874269396512
217 100 I 91973 7 325604109473206237898433933302481297
218 1621140188992194444701881625761731807571877809
219 2623059926317798754175087863660165740874359 J06
220 4244200115309993198876969489421897548446236915
221 686726(X)4 1627791953052057353082063289320596021
222 11111460156937785151929026842503960837766832936
223 17978720198565577 10498 1084 195586024127087428957
224 29090 18035550336225691 0 1 11038089984964854261893
225 47068900554068939361891195233676009091941690850
226 76159080909572301618801306271765994056795952743
227 123227981463641240980692501505442003148737643593
228 199387062373213542599493807777207997205533596336
229 3226 I 5043836854783580 186309282650000354271239929
230 52200210621 006R326179680 1 17059857997559804836265
231 844617150046923 109759866426342507997914076076194
232 J 366619256256991435939546543402365995473880912459
233 2211236406303914545699412969744873993387956988653
234 357785566256090598163895951314723998886183790 1112
235 5789092068864820527338372482892113982249794889765
236 936694773142572650897733199603935397111 1632790877
237 151560398002905470363 I 5704478931467953361427680642
238 245229875317 I 6273545293036474970821924473060471519
239 396790273J2006820581608740953902289877834488 152161
240 6420201486372309412690 1777428873111802307548623680
241 1038810421957299147085!O5 18382775401680142036775841
242 1680830570594530088354 I 2295811648513482449585399521
243 271964099255182923543922814194423915162591622175362
244 44004715631463593237933511000607242864504 1207574883
245 712011255569818855923257924200496343807632829750245
346 ApPENDIX A
n F.
246 1152058411884454788302593034206568772452674037325128
247 1864069667454273644225850958407065116260306867075373
248 3016 J2807933872843252844399261363388871298090440050 I
249 488019774679300207675429495 102069900497328777 ! 475874
250 7896325826131730509282738943634332893686268675876375
251 12776523572924732586037033894655031898659556447352249
252 20672849399056463095319772838289364792345825123228624
253 33449372971981195681356806732944396691005381570580873
254 5412222237103 765877667657957123376 J 483351206693809497
255 8757 J 5953430 18854458033386304178158174356588264390370
256 141693817714056513234709965875411919657707794958199867
257 229265413057075367692743352179590077832064383222590237
258 37095923077113188092745331805500 1997489772178 180790104
259 600224643828207248620196670234592075321836561403380341
260 97118387459933912954764998828959407281 1608739584170445
261 1571408518427546378167846658524186148133445300987550786
262 2542592393026885507715496646813780220945054040571721231
263 4114000911454431885883343305337966369078499341559272017
264 6656593304481317393598839952151746590023553382 I 30993248
265 10770594215935749279482183257489712959102052723690265265
266 17427187520417066673081023209641459549 I 256061 0582 12585 I 3
267 28197781736352815952563206467131172508227658829511523778
268 4562496925676988262564422967677263205735326493533278229 I
269 73822750993122698578207436143903804565580923764844306069
270 119447720249892581203851665820676436622934188700177088360
271 193270471243015279782059101964580241188515 I 12465021394429
272 31271819149290786098591076778525667781 144930 I 165198482789
273 5059886627359231407679698697498369189999644 I 36302198772 I 8
274 81870685422883100 17538806375350935968 I 1413714795418360007
275 1324695516964754142521850507284930515811378128425638237225
276 2143402371193585144275731144820024112622791843221056597232
277 346809788815833928679758165210495462843416997 I 646694834457
278 5611500259351924431073312796924978741056961814867751431689
279 907959814751026371787089444902993336949 113 I 7865 14446266 146
280 146910984068621881489442072459549 I 21 1054809360 I 382197697835
281 23770696554372451866815101694984845480039225387896643963981
282 384617949612346400 1575930894093975759058731898927884166 I 8 16
283 6223249151560709188257441 0635924603070626544377175485625797
284 100694286476841731898333719576864360661213863366454327287613
285 16292677799244882378090813021278896373 I8404077436298129 I 3410
286 26362106446929055567924 I 8497896533243930542711 10084140201023
287 426547842461739379460 149980002442288124894678853713953114433
288 690 168906931 029935139391829792095612517948949963798093315456
289 11167167493927693145995418097945379006428436288 17512046429889
290 18068856563237992497389336395866335 I 3 160792578781310139745345
291 2923602405716568564338475449381171413803636207598822186175234
292 4730488062040367814077409088967804926964428786380 132325920579
293 765409046 7756936378415884538348976340768064993978954512095813
294 12384578529797304192493293627316781267732493780359086838016392
295 20038668997554240570909178165665757608500558774338041350112205
296 32423247527351544763402471792982538876233052554697128188128597
297 5246191652490578533431164995864829648473361 1329035169538240802
298 84885164052257330097714121751630835360966663883732297726369399
299 13734708057716311543202577 1710279 131845700275212767467264610201
300 222232244629420445529739893461909967206666939096499764990979600
301 3595793252065835609617656651721890990523672 14309267232255589801
302 581811569836004006491505558634099066259034 I 53405766997246569401
303 9413908950425875674532712238062881653 I 1401367715034229502159202
304 152320246487859157394477678244038723157043552112080 1226748728603
305 2464593359921 179141398048006246675396881836888835835456250887805
306 398779582479977071534282478868706262845227240995663668299% 16408
307 6452389184720949856740872794933738025334 109298792472139250504213
308 10440 18500952072057208369758362080065378638 I 7087491 08822250 120621
309 16892574194241670428824570378554538679120491007541580961500624834
310 27332759203762391 000908267962175339332906872716290689783750745455
31 I 442253333980040614297328383407298780 120273637238322707452513 70289
312 7155809260176645243064 I 1063029052 I 7344934236440 122960529002115744
313 I 15783425999770513860373944643635095356961600163955231274253486033
314 1873415 I 860 153696629101 505094654031270189583660407819180325560 1777
315 30312494460 1307480 151388995590 17540805885743676803342307750908781 0
316 490466463202844446442404046536715720760753273372111614880764689587
317 793591407804151926593793042126891 128819610710140 145037958273777397
318 1284057871 006996373036197088663606849580363983512256652839038466984
319 207764927881114829962999013079049797839997469365240 1690797312244381
320 336 I 707 I 49818 144672666 I 872194541 04827980338677 I 6465834363635071 1365
321 54393564286292929722961773502446028063803133708 17060034433662955746
322 880 10635784474376449623645696987076343606520479817183780700 13667111
323 142404200070767306172585419199433 I 0440740965418798778412503676622857
324 230414835855241682622209064896420 1807510161 7466780496790573690289968
325 37281903592600898879479448409585328515842582885579275203077366912825
326 60323387178125067141700354899227346590944200352359771993651057202793
327 97605290770725966021179803308812675106786783237939047 196728424 1 15618
328 157928677948851033162880 158208040021697730983590298819190379481318411
329 2555339687195769991840599615168526968045177668282378663871 07905434029
330 4134626466684280323469401197248927185022487504 I 8536685577487386752440
331 66899661538800503153100008 1241745415306766517246774551964595292186469
332 10824592620564330638779402009666381338090 1526766531 1237542082678938909
333 1751455877444438095408940282208383549115781784912085789506677971125378
334 283391513950087115928688048317502 1682924797052577397027048760650064287
335 4585371 0 16945309254695820765383405232040578837489482816555438621189665
336 741928615644618041 398270 1248558426914965375890066879843604199271253952
337 12004657 1733914896686785220 13941832147005954727556362660159637892443617
338 1942394332983767008266122326250025906 I 971 3306 I 7623242503763837163697569
339 31428600503229159751339745276442091208977285345179605163923475056141186
340 508525438330668298340009685389423502709486 I 596280284 766 7687312219838755
LIST OF FIBONACCI NUMBERS 347
n F.
341 822all443362959895853407] 381538444147992590 I J079824521B 1610787275979941
342 I J]! 336881 693628! 9419341682354326791750874517270785300499298099495818696
343 21541483250565880900468239616971123323080041 8578767753330908886771798637
344 3485485206750216284240240785240380249816749358495530538302069862676 J7333
345 563963353180680437428706474693749258212475354428320807161115873039415970
346 91251 1873855702065852730553217787283194150290277873860991322859307033303
347 147647522703638250328143702791 1536541406625644706194668152438732346449273
348 2388987100892084569134167581129323824600775934984068529143761591653482576
349 386546232792846707241560460904086036600740 1579690263197296200323999931849
350 6254449428820551641549772190 170 184190608177514674331726439961915653414425
351 101199117567490 18713965376799211044556615579094364594923736162239653346274
352 16374361185569570355515148989381228747223756609038926650 176124155306760699
353 26494272942318589069480525788592273303839335703403521573912286394960 106973
354 42868634127888159424995674777973502051 06309231244244822408841 0550266867672
355 693629070702067484944762005665657753549024280 15845969798000696945226974645
356 112231541 198094907919471875344539277405965520328288418022089107495493842317
357 18159444826830 165641394807591 1 105052760867948344134387820089804440720816962
358 293825989466396564333419951255644330 16683346867242280584217891 1936214659279
359 4754204377346982207473680271667493829277014 17016557193662268716376935476241
360 769246427201 094 785080787978422393713094534885688979999504447628313150 135520
361 124466686493579300582815600558914309602223630270553719316671634469008561 1761
362 201 39132921368877909089439840115368091 16771188394517192671163973003235747281
363 32585801570726807967370999896006 79905139007491100054385837880317693321359042
364 5272493449209568587646043973612216714255778679494571578509044290696557 106323
365 85310736062822493843831439632 12896619394786170594625964346924608389878465365
366 13803567055491817972029187936825113333650564850089 197542855968899086435571688
367 2233464066177406 7356412331900038009953045351020683823507202893507476314037053
368 361382077172658853284415198368631 23286695915870773021 05005886240656274960874 1
369 5847284837903995268485385173690 1133239741266891456844557261755914039063645794
370 9461 10560963058380 1329537157376425652643718276222986560732061832060 1813254535
371 15308390447534579069814922331066538976617844965368671 0 1645823 74234640876900329
372 24 769496057165162871 1444594884429646292615632415916575771902992555242690 154864
373 4007788650469974 19409593818195095036058794082069603285936485366789883567055 193
374 64847382561864904812 1038413079524682351409714485519861708388359345126257210057
375 I 049252690665646467530632231274619718410203796555123147644873726135009824265250
376 16977265162842955156516706443541444007616135 1 1040643009353262085480136081475307
377 2746979206949941983182302875628764119171817307595766156998135811615145905740557
378 44447057232342374988339735 19982908519933430818636409166351397897095281987215864
379 71916849301841794820162763956116726391 0524812623217 5323349533708710427892956421
380 1 16363906534 18416980850249915594581159038678944868584489700931605805709880172285
381 18828075583602596462866526311206253798143927071100759813050465314516137773128706
382 304644662370210 134437167762268008349571826060 1596934430275139692032184 7653300991
383 492925418206236099065833025380070887553265330870701041 15801862234837985426429697
384 79757008057644623350300078764807923712509139103039448418553259155159833079730688
385 1290495498782682332568833813028150 12467835672190 109552534355121389997818506160385
386 20880655793591285660718346(X)6762293618034481 1293149000952908380545157651585891073
387 3378561078141 81 08986406684137043794864818048348325855348726350 19351554 70092051458
388 54666266575009394647125030 1438060884828525294776407554440171882480313 121677942531
389 8845187735642750363353171428084988334 76705778259666107927435384415468591769993989
390 1431 181439314368982806567444246559718305231 073036073662367607266895781713447936520
391 23157002128786440 1914188458705505855 1781936851295739770295042651311250305217930509
392 374688165219301300 194845203130 16182700871679243318134326626499182070320 18665867029
393 6062581865071657021 0903366183566768218691 04775627553202957692569518282323883797538
394 980946351726467002303878864965829509195627269995936663562034248772531434254966456 7
395 15R72045382J36327044 1291252680 149719 I 38253774755869198385780350572435966664334621 05
396 25681 508899600997067167913917673267005781650 175546286474198377544968911008983126672
397 4155.155428 1937324111297039185688238919607027651133206312776412602212507675416588777
398 6723506318153832 1 1784649531033615059253886778266794927869747901471814 18684399715449
399 1087886 1746H75645289761992289049744844995705477812699099751202749393926359816304226
400 1760236806450 1396646K22694539241 1250770384.1833044921918867259928965753450442160 19675
40 I 2848122981 0848961 175798893768146099561538008878230489098647719564596927140403232390 I
402 460835978753503578226215883073872246385764472086797082873203188542544616448248343576
40.1 74564827686199.118998420482075533324200 1 14456086910197.1859680384188513887852280667477
404 1206484255615496 7682 1042070382920548838690903295589905673288357273 10585043005290 I 1053
405 19521325324774899581946255245845387.1038805.159382500 I 030592563956919572392152809678530
406 .1158616788092986 72640504622841374421877496262678090008732544 7529650630896453338689583
407 511074932057047668459967175299828294916301622060590 1117918011486570203288606148368113
408 826936610866.146.1411004717981412027 16793797884738680 1205243459016220834 185059487057696
409 13.1801154292..1.19400956043897.14410310117100995067992702323161470502791037473665635425809
410 21649481 5.178974035066091 077158223372850389739153795035284049295190 1187 1658725122483505
411 35029596967131343602213497450132647402139968983.1722058515664000218029091323907579093 14
412 566790785050287471088226051660549846871789428987517093799713295408147807911158803928 19
413 IJ 1708675472 16(X)907 1 10361026162876.1208931891 18821239152315.17729562617689923506638302133
414 I 4838775W771 R88.17819858707782342616 776497854 78087562461 1509059103432470714622518694952
415 240096429449.148928530894810398630248865816766662999539843046 788666050160638 129156997085
416 .18848418342653776635075.15 18180972865642314621443875164454555847769482631352751675692037
417 6285806128758R669488 16483285796031 145081313881 {}68747042976026.164.155.12791990880832689 122
418 1() 170647963024244h 12.1240 1846 7605759RO 1504460095507498687521584842050 15423.14363250838 1 159
419 1645645409178.111 1561 14050175340 179094658577.11J7657624573049761 120640548215334513341070281
420 266271 02()54807356 I 7346452022 I oo75507480902340720R3 74441801919604845563638678145849451440
421 4.108.15561 4659()46 77346050219744093416946760080486599901485 16807254861 I 18540 1265919052 1721
422 6971 065820 13978231J08069542195416892442766242120743734566536(X)33033 1675492690805039973161
423 I 12794214.1479882916426745641698262341.17442250 I 694037247150528 10558 17787346703464230494882
424 18250487254'.).\861 15550744 !O636524.1 I 26580208492290 14745928158881386149462839394269270468043
425 29529908689737440719341867053506936071 7650742459551 I 8W9664 I 6244 I 967250 186097733500962925
426 477803959446760522744 162776900.11248729785923474969864327823043828 I 16713025492002771430968
427 7731030463441.149299375814474353818480 15509977209249827274872062700839632 1 1589736272393893
428 1250907005790895452681744224]3569433531 336921195894847055310250098200676237081739043824861
429 20240 100521 35030382619325671771 07618332887918916819829782797456368284639448671475316218754
430 327491705792592583530 1069896106770518642248401 12714676838107706466485315685753214360043615
431 52989271 10060956217920395567877846701971 12759029534506620905162834769955134424689676262369
432 8573844167986882053221465463984617220613375991422491834590 1286930 1255270820 177904036305984
433 138727712780478:~8271 14186 J 03186246392258450358171783690079918032136025225954602593712568353
434 22446615446034720324363326495847081 143 197879573140.1287353893090 1437280496774780497748874337
435 36319386 724082558595505 1875277095450657823831548581656361884893.157330572272938309146 1442690
348 ApPENDIX A
n ~
436 58766002170 117278919868514023556626208980262727998494371577798350 I 0586219504163589210317027
437 95085388894 J9983751537370 15512661712747626458828566600077662876858389194223354668067 1759717
438 15385139106431711643524221557482279748374290861 08551 543793440860359447816! 73771 02698R2076744
4)9 248936779958516953950615917126088968758505554493711814387I 1037372178370103971256950553836461
440 402788171022834O70385858! 3270091 ! 7662422484631 0456696876645445975772848265708967220435913205
441 65 J 72495098135 10243364740498270007350007540 1759827878315356483347951218369680224170989749666
442 105451 312200418509472233218252791250 12430024807028457519200 1929323724066635389191391425662871
44~ 170623807298553611905880623235491323624375649830112453507358412671675285005069415562415412537
444 276075119498972121378113841488282573748675897900397028699360341995399351640458606953841075408
445 446698926 79752573328399446472377389737305154 7730509482206718754667074636645528022516256487945
446 7227740462964978546621 0830621205647112172744563090651 0906079096662473988285986629470097563353
447 1169472973094023587946102770935830368494778993361415993112797851329548624931514651986354051298
448 189224701939052144260821 10771478868396165064389923225040 1887694799202261321750 1281456451614651
449 30617199924845450305543138480837172081 112854323537384971316747993215712381490 15933442805665949
450 495396701 187506647316252492523160404772779187134606100 1 150551747313593851366517214899257280600
451 80 1568700435961 1503716838773315321255839077303699799498282226546635 165089515533148342062946549
452 129696540 16234677976879363698546925303566869175045860499432778293948758940882050363241 320227149
453 20985341 0205942894805962024718622465594059464787456599977150048405839240303975835 1 1583383173698
454 3395499503682896745747556617040917 1862972815653791520497147783134532682971279633874824703400847
455 5494033605742325693807176864227141842237876213253 718049486278797511660700 1677217386408086574545
456 8889533109425222439554733481 26805902853515777863287009920 10571 109649289972956851261232789975392
457 1438356671516754813336191 03454952008707730339918865881486873359084765896974634068647640876549937
458 2327309982459277057291664382676325989930819177051945824 78883930 194415186947590919908873666525329
459 376566665397603 18706278554 17225846077008 1225762406046396575728927918 1083922224988556514543075266
460 6092976636435308927919519799902172066938941753292550464446412194 73596270869815908465388209600595
461 98586432904113407985473752171280 181439470643295331551 041 0398508752777354792040897021902752675861
462 1595161992684664972646689501 703019021 08860060828257055(8550)9728226373625661 856805487290962276456
463 258102632172579905250 142702341582083548330704 1235886067265438236979150980453897702509193714952317
464 4176188314410464025148116525118839856571907649518456624120477965205524606115754507996484677228773
465 6757214636136263077649543548534660692055214690754342691385916202 184675586569652210505678392181090
466 10933402950546 7271 0279766007365350054862712234027279931550639416 7390200 1926854067 18502 163{)69409863
467 17690617586682990 180447203622188161240682337031 02714200689231 036957487577925505892900784 1461590953
468 286240205372297172832448636958416617893094593 7129994132239870453696507597194046564 75 10004531000816
469 4631463812391270746369206 73180298230299917964023270833292910 14906539951751 1955245765 17845992591769
470 749386586611424247469369310 1387148481930 1255773627024651689719443505027723135990224027850523592585
471 121253296785055132210628998331901307849293052175954107980980734350044979474331514800545696516184354
472 196191955446197556957565929345772792668594307949581 132632670453793550007197467505024573547039776939
473 317445252231252689168194927677674100517887360125535240613651188143594986671799019825119243555961293
474 51363 72076774502461257608570234468931864816680751 163732463216419371449938692665248496927<.X)595738232
475 83108245990870293529395578470 1120993704369028200651613859972830080739980541 065544674812034 151699525
476 1344719667586153 1814197166417245678868908506962757679871062944720 1788497441 0332069524504824747437757
477 2175802127494856116713672426425688880595219724476419600966267302098624954951397614199316858899137282
478 3520521795081 009298 1333890681502567674860704207521875880725617741 16509929361729683723821683646575039
479 5696323922575865414847061494575945648081290145228607189038829076215134884313127297923138542545712321
480 9216845717656874712980450562726202415567360565980794777111390850331644813674856981646960226192287360
481 14913169640232740 12782751205730214806364865071120940 1966150219926546779697987984279570098768737999681
482 241300 153578896148408079626200283504792160 11277190 1967432616]077687842451]66284 126 1217058994930287041
483 3904318499812235496863547467733049854286466 198839959870941 1830703425204209650825540787157763668286722
484 631732003560 1 1969809443437297358849022080673265589795452673441480303628721313666802004216758598573763
485 10221638535413432477807891 1974689347564945335253989394 162085272183728832930964492342791374522266860485
486 1653895857 10 1462945875223492720481965870260085195791896147587 13664032461652278159144795591280865434248
487 267605971 06428061936560 126124673754415197134377356858377684398584776 1294583242651487586965803132294733
488 43299555677442691395312361051878574073899735229314777339 16026995 1 1793756235520810632382557083997728981
489 70060 15278387075333 18724871765523284890968696066 7]63571684466853595550508187634621]9969522887130023714
490 1 13359708461313444 7271848482284309025629966048359864130560049384871348807054284272752352079971 127752695
491 ]83419861245 1841980590573354049832310520934 744426580487728496070230903857873047734872321602858257776409
492 2967795697064976427862421836334 14 ]3361509007927864446]8288545455] 02252664927332007624673682829385529104
493 4801994309516818408452995190383973646671835537213025 1060 17041525333 156522800379742496995285687643305513
494 7769790006581794836315417026 7 1811498282273632999946972430558698043540918772771 1750 12166R96R517028834617
495 12571784316098613244768412217 102088629494571867212494830322628505768565710528091492618664254204672140 130
496 2034 157432268040808108382924382020361231730819721 1964554628215486203974898255803242740333222721700974747
497 32913358638779021325852241460922292241811880064424459384950843991972540608783894735358997476926373114877
498 5325493296]459429406936070704 742495854129188261636423939579059478]765]5507039697978099330699648074089624
499 8616829160023845073278831 216566478809594] 068326060883324529903470 1490561]582359271 3458328176574447204501
500 139423224561697880 1397243828704072839500702565876973072641 08962948325571622863290691557658876222521294125
Appendix B
Proofs of Fibonacci
Relationships
For C h a pter 1
Some of the fo llowing proofs involve a technique called mathe-

matical induction. A brief explanation of this procedure may be in
order here.
Introduction to M a them atical Induction
Consider a set of domi noes with an endless number of tiles set up

as illustrated below.
Figure 8 - 1
If we were asked to knock down all the tiles, we could consider

two methods:
349
350 ApPENDIX B
(1) we could knock down each tile separately; or

(2) we could knock down just the first tile, if we were sure that
any tile knocked down would automatically knock down
the tile after it.
The first method would not only be inefficient but would also
never assure us of knocking down all the tiles (since the end may
never be reached). The second method would guarantee that all the
tiles are knocked down. If we knocked down the first tile, we
would be assured that it, and any knocked down tile, would knock
down its successor tile-that is, the first tile knocks down the sec-
ond, which knocks down the third, which knocks down the fourth,
and so on. All would be knocked down.
This second method is directly analogous to the axiom of
mathematical induction:
A proposition involving the natural number n is true for all
natural numbers:
(a) if the proposition is correct when n = I; and

(b) if k is any value of n for which the proposition is correct, then
the proposition is also correct for n = k + I.
Now for the proof of some of the relationships that were ac-
cepted intuitively in the chapters.
1. (p. 33) Listing the remainders of the first few Fibonacci

numbers upon dividing by 11, we have
1, 1,2,3,5,8,2,10, 1,0, 1, 1, 2, 3, 5, 8, 2, 10, 1,0, ....
So we see that the remainders repeat in cycles of

length 10. Since it is the remainder upon dividing a num-
ber by 11 that determines its divisibility by 11, all we
have to do is check that in adding any ten consecutive
numbers in the sequence
PROOFS OF FIBONACCI RELATIONSHIPS 351
1,1,2,3,5,8,2,10,1,0,1, 1,2,3,5,8,2,10,1,0, ...
we get a sum divisible by 11. And we check this as follows.

Since the cycle of this sequence is of length exactly 10,
adding any ten consecutive numbers in this sequence will
always come out to adding the ten numbers-I, 1,2,3,5,
8, 2, 10, 1, O-in a cycle.
1
y---.-- 1
1/ \2
: I
10~ ~3
2 ----- .---- 5
8
Figure B-2
(Imagine these ten numbers arranged clockwise in that

order around a circle, with the sequence above obtained by
traveling around the circle over and over. Then you can see
that any numbers missed at the beginning of a cycle-if the
sum is started somewhere in the interior of the cycle-are
regained from the next cycle; for example, consider the
sum 5 + 8 + 2 + 10 + 1 + 0 + 1 + 1 + 2 + 3. This is because
no matter where you start counting on the circle, counting
ten numbers clockwise around the circle will amount to
counting all ten numbers, because that is exactly how many
numbers there are.)
These ten numbers have a sum of 33, which is indeed
divisible by 11.
2. (p. 33) We proceed by induction. Fl = 1 and F2 = 1 are, of
course, relatively prime. Now, assume that Fk and Fk+ 1 are
relatively prime (our induction hypothesis). If Fk + 1 and Fk+2
352 ApPENDIX B
had a common factor other than 1, say the factor b, then

since Fk = Fk+2 - Fk+l' Fk would have the factor b too. But
then Fk and Fk+l would both have the factor b, and that
would contradict our assumption that Fk and Fk+l are rela-
tively prime (do not have any common factor other than 1).
So Fk+l and Fk+2 cannot have any common factors other
than 1, which means they are relatively prime as well. This
completes our induction.
3. (p. 35) We have to prove that the Fibonacci numbers in
composite-number positions-with the exception of the
fourth Fibonacci number, F4 = 3-are composite numbers.
We will show that, in general, if m is divisible by n, then
Fm is divisible by Fn, or, in other words, that Fnk is divisi-
ble by Fn for any nand k.
We proceed by induction on k. For k = 1, we need to
check that Fnl is divisible by Fn; and this is, of course,
true. Now assume Fnp is divisible by Fn (this is our induc-
tion hypothesis). We want to prove the statement for
k = p + 1, that is, that Fn(p+l) is divisible by F". Using the
lemma l in number 8 (following later)
we get:
which is divisible by F" because both Fnp _1 Fn and FnpFn+l

are divisible by Fn (by the induction hypothesis,
Fnp is divisible by Fn , and so is F"pFn+1 ). This completes the
induction.
I. A subsidiary proposition is used to prove another proposition; a helping theo-

rem.
So, based on all this, if n is a composite nwnber, that is,

if we can write n = ab, where a and b are both different from
I, then we know that Fn is divisible by Fa (and also by Fb ,
for that matter). Now, for n greater than or equal to 2, Fn is
greater than any Fibonacci number that comes before it (the
Fibonacci sequence is increasing), so since n = 4 (which is
greater than 2) is the first composite number, certainly Fa is
different from Fn' Now we also need to make sure Fa is dif-
ferent from I to show that Fn is composite (a composite
number must have some divisor other than I and itself).
Now, Fa = 1 can only happen if a = I or a = 2. But a (and b)
was assumed not to be equal to I (because n is composite),
so the only possibility is a = 2. The only problem, then, can
occur when a = 2 and b = 2, that is, when n = 4. Indeed, for n
= 4, F" = 3 , which is not composite.
5. (p. 37) We proceed by induction. For n = I, we need to

check that F2 = F3 -1 ; and this is, indeed, true (1 = 2 - I).
Now assume the statement is true for n = k, that is, assume:
Then we need to check the case when n = k + I:
(F2 +··· +F2k)+F2k+2

= F2k +1 -1 + F2k+2
= (F2k+l + F2k+2)-1
= F2 k+3- 1
= F2(k+l)+1 -1
which is exactly the statement for n = k + I. So our proof by

induction is complete.
6. (p. 38) We again we proceed by induction. For n = I, we
need to check that F1 = F2 ; and this is, of course, true
354 ApPENDIX B
(1 = 1). Now assume the statement is true for n = k, that is,

assume:
Then we must check to see if this holds true for n = k + 1:
(F; + ... + F 2k - 1 ) + F 2k + 1
= F2k + F2k+l
= F 2k + 2
= F 2 (k+l)
which is exactly the statement for n = k + 1. So our proof by

induction is complete.
7. (p. 39) We again proceed by induction. For n = 1, we need
to check that F/ = F;F2 ; and this is, of course, true
(12 = 1 . 1). Now assume the statement is true for n = k, that
is, assume:
Then we check the case for n = k + 1:
(F;2 + ... + F/) + F/+ 1

= FkF;,.+1 + F/+ 1
=F;,.+I(F;,.+Fk+l)
= Fk+1F;,.+2
which is exactly the statement for n = k + 1. So our proof

by induction is complete.
8. (p. 42) We have, by factoring (difference of squares),
F/ - F/"_2
=(Fk -Fk- 2 )(Fk +Fk - 2 )
= Fk - I (~+ Fk - 2 )
= Fk_1Fk + ~_IFk-2
= ~_IFk-2 + Fk~_1
First we'll prove the following lemma, which will come in

useful and later on.
Proof of Lemma
We proceed by induction on n. (Actually, this will be a

form of induction that is called "strong induction," where
we will be using the statements for n = k - I and for n = k
to prove the statement for n = k + I. This will also mean
that we will need two base cases rather than just one; so we
will have to check the statement for n = I and for n = 2.)
For n = I, we have to check that Fm+' = Fm-,F; + FmF2 , or,
since F2 = 1 and F2 = 1, we have to check that
Fm+' = Fm_, + Fm; this is, of course, true because it is the
very relation by which we define the Fibonacci numbers.
For n = 2, we have to check that Fm+2 = Fm- 1F2 + Fm~ , or,
since F2 = 1 and F3 = 2, we have to check that
Fm+2 = Fm_, + 2Fm; and this is true by the following se-
quence of equalities:
Now assume the statement is true for n = k - I and n = k,

that is, assume
356 ApPENDIX B
F m+k - I = Fm_1Fk _ 1 + Fm~

and
(This is our induction hypothesis.) Then

Fm_1Fk+1 + FmFk+2
= Fm_1(Fk_1 + Fk )+ Fm(Fk + Fk+l)
= Fm_1Fk_1 + Fm_1Fk + FmFk + FmFk+l
= Fm_1Fk + FmFk+1 + Fm_1Fk_1 + FmFk
= Fm+k + Fm+k- 1
= Fm + k + 1
That is, Fm_1Fk+l + F,nFk+2 = Fm+k+1 , which is exactly the

statement for n = k + 1. So our induction is complete.
Now using this lemma, we see (by substituting m = k
and n = k - 2) that
Hence F/ - F/_2 = F2k - 2 , and the proof is complete.

9. (p. 43) F} + F}+l = F2n + 1 ; that is, the sum of the squares of
the Fibonacci numbers in positions nand n + 1 (consecu-
tive positions) is the Fibonacci number in place 2n + 1.
Proof
Using the lemma from item 8 (above), we prove our origi-

nal statement. In the lemma, let m = n + 1 and n = n. Then
we getF2n +1 = FnF;, + Fn+IFn+l , or, in other words,
2 2
F2n+1 = Fn + Fn +1' which is the equation we wanted.
10. (p. 43) A proof that Fn2+1 - Fn2 = Fn_1 . F;,+2

It follows immediately, because of the binomial theo-
rem [a 2 - b2 = (a + b)( a - b) ] and the definition of the Fi-
bonacci numbers, that
11. (p. 44) A proof of Fn-1F"+1 =Fn2 +(-1)", wheren~1.

By mathematical induction:
For n = 1:
P(1): Fo F,=F;'+(-lr;
0·1=1'-1=0
For n = k:
For n = k + 1:
P(k + 1): to check if it is true for k + 1:
Fk2+2 - Fk2+1 = ( ~ + Fk+1)2 - Fk2+1

= Fk + 2Fk+1Fk + ~2+1
2
- F/+ 1
= Fk(Fk +2Fk+l )
= Fk [ ( Fk + Fk+ I) + Fk I]+
= Fk (Fk+2 + Fk+,)
= Fk • ~+3
And we will complete the proof by mathematical induc-

tion if we in the following use the assumption for n = k
in the form F k - I· Fk+ 1 - F/= (_l)k+1.
358 ApPENDIX B
• • 2 ( )k+ I
P(k + 1): To check If P(k + 1) IS true: F kF k+2 = F k+1+ -1
F kF k+2 - F/+ 1 = Fk(Fk+1+ F k )- Fk2+1

= FkFk+1+ F/ - F/+ 1
= F/ + Fk+1(Fk - F k+1)
= F/ + Fk+1(-Fk_1)
= F/ - Fk+1Fk_1
2
= (-1)(Fk+1Fk_1- Fk )
= (-l)(-l)k
= (_l)k+l
= (_ l)k
12. (p. 47) We have to show that Fmn is divisible by Fm for any
nand m. We proceed by induction on n.
For n = 1, we need to check that F,m is divisible by Fm;
and this is, of course, true. Now assume Fmp is divisible by
Fm (this is our induction hypothesis-the statement for
m = p). We want to prove the statement for m = p + 1, that
is, that Fm(P+l) is divisible by Fm. Using the Lemma in
number 8, we get:
which is divisible by Fm because both summands (addends)

Fmp-1Fmand FmpFm+l are divisible by Fm (by the induction
hypothesis, Fmp is divisible by Fm), so it is also Fm(p+l).
This completes the induction.
14. (p. 49) We again proceed by induction. For n = 1, we need
to check that L 1 = L 3 - 3; and this is, of course, true
(I = 4 - 3). Now assume the statement is true for n = k,

that is, assume:
Then we must check to see if this holds true for n = k + I:
We have
Ll + L z + ... Lk + L k+1
= L k +2 - 3+ L k + 1
= L k+1 + L k+2 - 3
= L k+ 3 - 3
which is exactly the statement for n = k + I. This completes

our proof by induction.
15. (p. 51) We again proceed by induction. For n = I, we need
to check that L~ = L1L2 - 2; and this is of course true
(12 = 1·3 - 2 = I). Now assume the statement is true for
n = k, that is, assume:
Then we must check to see if this holds true for n = k+ 1:
L~ + L~ + ... + L~ + L~+l = L k+1L k+2 - 2
L~ + L~ + ... + L~ + L~+l
= LkLk+l - 2 + L~+l
= L k+1(Lk + L k+1) - 2
= L k + 1L k + 2 - 2
which is exactly the statement for n = k + 1. This completes

our proof by induction.
360 ApPENDIX B
For Chapter 4
Heron of Alexandria's Method for Constructing the Golden

Section (figure 4-7) (p. 121)
AB = r, BC = CD =!..., and AD = AE = x
2
Applying the Pythagorean theorem to MBC:
r2 2 r2 2 2
x 2 + rx + - = r + -, or x + rx - r = 0
4 4
X 2=--±
rF22r~r2rr
-+r
r
=--± -=--±-.J5=--(l±.J5)
'. 24 2422 2
Since x must be positive, we disregard x = -!... (1 +.J5); there-

2
r C .,[;-1 r I r r
fore x=--(1 --.;5) = - - . - = - ' - , or X"" .618033988· -.
2 2 2 I/J 2 2
.,[; - 1 r .,[; - 1 r
-- - -- - (.,[; -I)'~
AE x 2 2 2 2 4
BE r-x .,[; - 1 4r .,[; - 1
(4 _.,[; + I). r
r---·- -- -
2 2 4 2 2 4
= t/J = 1.618033988
r .J5-1 r
AB =r AE = BE = BC = CD =- and AD = AE =x=--·-
, 2' 2 2
Construction of the Golden Section as in Figure 4-8

(p.122)
AB=a , AC =.':2' BC = CD , and AD = AE =x
Applying the Pythagorean theorem to MBC: BC 2 = AB2 + AC 2

Therefore,
BC=CD=~5=2 =a ~,
x = AE = AD = CD - AC =a-
J5 - a J5-1
- = a· - - - and
2 2 2'
BE=a-x=a---
3-J5
2
a
.J5 -\
2 .J5-1
AE: BE = x: (a - x) = -----='__;=_-
a
3-.J5 - 3-.J5
2
.J5-1 3+.J5 3.J5+5-3-.J5 2.J5+2 2(.J5+1) .J5+1

=3-.J5· 3+.J5= 9-5 4 4 2
= l/J "" 1.618033988
a .J5-1
(AB=a, AC=-, BC = CD, and AD = AE =x=a·--)
2 2
Construction of the Golden Section as in Figure 4-9 (p. 123)
Begin by constructing a right MBC with AB = 2 cm, BC = 1 cm.

By the Pythagorean theorem, we get A C =.J5 cm.
362 ApPENDIX B
Using the theorem that states that the angle bisector of a trian-
gle divides the side to which it is drawn proportionally to the other
two sides, we find that
AP J5
PB
and angle bisector CQ divides side AB (externally) proportionally

to the sides of the angle to give us
AQ .J5
QB
Our task now is to show that point P divides the line segment
AB into the golden ratio (or its reciprocal).
From the relationship established above, we get:
1 PB PB PB
J5 - AP = AB-PB = 2-PB
PBJ5 = 2- PB
PB=_2_= J5-1
J5 +1 2
which is the reciprocal of the golden ratio (PB =i).

We can show that the point Q divides the line segment AB
externally into the golden ratio.
From the second relationship above we get:
1 QB QB QB
J5 = AQ = AB+QB = 2+QB
QBJ5 =2+QB
QB=_2_= J5+1
J5 -1 2
which is the golden ratio (1)).

You might notice that in right MCQ, CB is the mean propor-

tional between PB and QB so that
CB QB
PB CB
/5+1
I _2_
/5-1 = I
2
which is certainly true!
For Chapter 6
Sums of Distinct Fibonacci Numbers (p. 188)
Every positive integer n can be expressed as a finite sum of distinct

Fibonacci numbers.
Proof
Let Fk be the largest Fibonacci number ::;; n. Then n = F;. + n],

where n] ::;; Fk . Let Fk, be the largest Fibonacci number::;; n] . Then
n= Fk + Fk , + n 2 , n ;:: Fk > F;., .
Continuing like this, we get n = Fk + F., + F. + ... , where
n;:: Fk > F., > F., .... Since this sequence of decreasing positive
integers must terminate, the result follows.
Proof of Fibonacci Generation of Pythagorean Triples (p. 192)
Suppose a, b, c, and d form a Fibonacci sequence. Then we get

c = a + b, d = c + b = a + b + b = a + 2b. That is, a, b, a + b, and
a + 2b form a Fibonacci sequence.
364 ApPENDIX B
Thus, a, b, a + b, b + a + b is by definition a Fibonacci se-

quence.
The operations indicated in the procedure are:
A = 2bc = 2b(a + b) = 2ab +2b 2

B = ad= a(b + a + b) = a 2 +2ab
C = b 2 + c 2 = b 2 + (a + b)2 = a 2 + 2ab + 2b 2
Applying the Pythagorean theorem to see if these triples fit the

theorem:
A2 = (2ab + 2b 2)2 = 4a 2b 2 + 8ab 3 + 4b 4

S2 = (a 2 + 2ab)2 = a 4 + 4a 3 b +4a 2b 2
C 2 = (a 2 +2ab + 2b 2 i = a 4 + 4a 3 b + 8a 2b 2+ 8ab 3 + 4b 4
More generally, to find a square which remains a square when in-

creased or decreased by d, we have to find three squares that form
an arithmetical progression with a common difference d.
The squares of the numbers are:
(a 2 _ f=a
2ab - b 2 4
- 4a'b + 2a 2 b' + 4ab' + b 4
(a +b f = a
2 2 4 4
+ 2a'b' +b
(a' + 2ab- b f = a
2 4
+ 4a 3b+ 2a'b' - 4ab 3 +b 4
They form an arithmetic progression with difference
d = 4a b- 4ab' = 4ab(a
3 2
-b')
If a = 5 and b = 4, then d = 720, and we have three squares

(a 2 + b 2)2 = 41 2, 412 - 720 = 31 2 , and 412 + 720 = 49 2.
Dividing by 12 2 , we have Fibonacci's solution.

From anyone solution we can form infinitely many other solu-

tions. Therefore, if we take a = 412 = 1,681 and b = 720, we have
another solution, with
d= 5(24.41. 49.31)2 = 11,170,580,662,080

and
(a 2+b 2)2 = 11,183,412,793,921
The three squares in arithmetic progression are then x - 5, x,

and x + 5, with
x = (11,183,412,793,921)2
2,234,116,132,416
Proof on the Property of 89 (pp. 216-17)
Refer to the development on pages 216-17.
We have already shown that (VI) is true for n = 1:
10' +, = 89·(F,·IO' -I) + IOF 1 +, + FI = 89· (1. I) + 10·1 + 1 = 89 + 10 + I

= 100,
that is, 102 = 89 + 10 + 1 (I)
Suppose that for some fixed integer k ~ 1,
Then multiplying by 10 and again substituting (I), we get:

366 ApPENDIX B
10'+2 = 89·(F,·IO' + F 2·10'-' + ... + F, ·10) + 10 2F,+, + IOF,
= 89-(F,'IO' + F 2·1O'-' + ... + F,-IO) + (89 + 10 + I)Fk+' + IOF,
= 89·(F,·IO' + F,.IO'-' + ... + F,·1O + F,+,) + IOF,+, + F,+, + IOF,
= 89·(F,·IO' + F 2·1O'-' + .•. + F,'IO + F,+,) + IO(F,+, + F,) + F, +,
= 89·(F,·IO' + F 2·1O'-' + ... + F,·IO + F,+,) + IOF'+2+ F,+,.
This shows that (VI) is true in the (k + I )st case, if it is true in

the kst case.
And since we already have that it is true for n = I, it is true for
every positive integer n.
Dividing both sides of (VI) by 89·1 On+l, we obtain
~=~+~+~+
2
+ F,,-, +~+ IOF,.. , + F"
89 10 10' 10' 10" 10"" 10""
=~+~+~+
10' 10' 10'
· IOF +F
where 11m "" " = O.
n~oo 10""
Chapter 9
Proof of the Binet Formula for Finding a Fibonacci Number

(pp. 299-300)
We will prove that F" = Js{p" -(-iJ1

By mathematical induction, we have shown in figure 9-5 that

for n = I the following is true:
For n = k, therefore, from figure 9-5 we can see that

Fk + 2 = Fk + Fk +! (our induction hypothesis).
Therefore, we must show that for n = k + 1 the following is
also true:
~+3 = Fk+! + ~+2

We begin with:
[ Recall that ¢! + 1 = ¢2 and ( -i) + 1= ( -i J]

= Js[¢k+!¢2 -( -i J+! (-i J]
Js[¢k+3_( -iJ+
3
= ]
= Fk+3, which was to be shown.

368 ApPENDIX B
We could have also done this with our alternate version of the
Binet formula-using real numbers:
= ~[( ~+I)k+1 -C-~y+I]+ ~[( ~+ly+2 _C-~)k+2]
= ~[( ~+ly+3 _C-~y+3]
again showing what was required to complete the induction proof.

Thus we have the Binet formula:
which will give us any Fibonacci number for any natural number n.
Another Way to Find a Specific Fibonacci Number (Using a

Calculator or a Computer) (p. 303)
We will use the following:
(This lemma was also used in the proof for chapter I, item 8.)
We substitute m = n, then we get on the one side of the equation

and on the other side
also
F2n = Fn (F,,-l + Fn+ 1 )

Since Fn+l = F,,-l + F", we get:
References
Alfred, Brother U. An Introduction to Fibonacci Discovery. San

Jose, CA: Fibonacci Association, 1965.
Beutelspacher, Albrecht, and Bernhard Petri. Der Goldene Sc hnitt.
Mannheim, BI-Wiss.-Verl., 1995.
Bicknell, Marjorie, and Verner E. Hoggath Jr., eds. A Primer for
the Fibonacci Numbers. San Jose, CA: Fibonacci Association,
1973.
Blecke, Nathan. Finding Fibonacci in Fractals. MA thesis, Central
Michigan University, 2001.
Boyer, Carl B. A History of Mathematics. New York: Wiley, 1991.
Burton, David M. The History of Mathematics. 3rd ed. New York:
McGraw Hill, 1997.
Crownover, Richard M. Introduction to Fractals and Chaos.
Boston, MA: Jones and Bartlett, 1995.
Dunlap, Richard A. The Golden Ratio and Fibonacci Numbers.
River Edge, NJ: World Scientific Publishing, 1997.
Eatwell, John, Murray Milgate, and Peter Newman. The New
Palgrave: A Dictionary of Economics. 4 vol. London: Basing-
stoke, 1987.
Eves, Howard. An Introduction to the History of Mathematics.
Philadelphia: Saunders College Publishing, 1990.
Falconer, K. J. The Geometry of Fractal Sets. New York:
Cambridge University Press, 1985.
Garland, Trudi Hammel. Fascinating Fibonaccis: Mystery and
Magic in Numbers. Palo Alto, CA: Dale Seymour, 1987.
371
372 REFERENCES
Ghyka, Matila. The Geometry of Art and Life. New York: Dover,
1977.
Heath, Sir Thomas L. Manual of Greek Mathematics. Oxford:
Clarendon Press, 193 1.
Herz-Fischler, Roger. A Mathematical History of the Golden
Number. New York: Dover, 1998.
Hoggatt, Verner E., Jr. Fibonacci and Lucas Numbers. Boston:
Houghton Mifflin, 1969.
Huntley, H. E. The Divine Proportion. New York: Dover, 1970.
Koshy, Thomas. Fibonacci and Lucas Numbers with Applications.
New York: Wiley, 2001.
Mandelbrot, Benoit B. The Fractal Geometry of Nature. San
Francisico: W. H. Freeman, 1982.
- - - . "A Multifractal Walk Down Wall Street." Scientific
American, February 1999.
Posamentier, Alfred S. Advanced Euclidean Geometry. Emeryville,
CA: Key College Publishing, 2002.
- - - . Math Charmers: Tantalizing Tidbits for the Mind.
Amherst, NY: Prometheus Books, 2003.
Posamentier, Alfred S., and Ingmar Lehmann. n: A Biography of
the World's Most Mysterious Number. Amherst, NY: Prom-
etheus Books, 2004.
Prechter, Robert R., Jr. R. N. Elliott's Masterworks: The Definitive
Collection. Gainesville, GA: New Classics Library, 1994.
Rasmussen, Steen Eiler. Experiencing Architecture. Cambridge,
MA: MIT Press, 1964.
Rowley Kevin. "Fractals and Their Dimension." Plan B paper,
Central Michigan University, 1996.
Runion, Garth E. The Golden Section and Related Curiosa.
Glenview, IL: Scott Foresman, 1972.
Shishikura, M. "The Boundary of the Mandelbrot Set Has
Hausdorff Dimension Two." ASf£~risque 222, no. 7 (1999):
389--405.
Sigler, Laurence. The Book of Squares. New York: Academic
Press, 1987.
REFERENCES 373
- - - . Fibonacci's Liber Abaci. New York; Springer-Verlag,

2002.
Vorobiev, Nicolai N. Fibonacci Numbers. New York: Blaisdell,
1961. Basel, Switzerland: Birkhauser Verlag, 2002.
Walser, Hans. The Golden Section. Washington, DC: Mathe-
matical Association of America, 2001.
Index
e. See Euler's number algebra, 12, 19-20, 19n2

1/1. See golden ratio (section) algebraic number, 165n2
i. See Ludolph's number algorithms, 12, 198, 198n 10, 310
o (the number), 37, 50, 111,205-206 and ancient Egyptians, 195n8
2 (the number) multiplication algorithm and Fibo-
Fibonacci numbers divisible by, 47, nacci numbers, 195-99
210,330 al-Khowarizmi, 19
power of, 196, 196n9 alternating numbers, 42, 42n23, 44-45,
3 (the number) 54, 54n25, 55, 142
Fibonacci numbers divisible by, 47, Altevogt, Rudolf, 76
210,330-31 American Library Association, 211 n 16
rule for divisibility by, 47n24 angle
4 (the number) angle sum of a triangle, 84
Fibonacci numbers divisible by, bisectors, 123, 145, 146,362
330-31 divergence angle, 64n3
relationship to 89, 213nl9 and fractals, 311, 313, 314
5 (the number), Fibonacci numbers di- golden angle, 64, 64n3, 74, 148-49,
visible by, 47 250, 250n20
6 (the number), Fibonacci numbers di- and isosceles triangles, 144
visible by, 331 and pentagons, 151, 157
8 (the number), Fibonacci numbers di- and right triangles, 128, 134, 135
visible by, 48, 331 used in watch displays, 217-20,
II (the number), Fibonacci numbers 220n21
divisible by, 33, 331,350-51 vertex angle, 144, 147
12 (the number), Fibonacci numbers Animali da I a 55 (Merz), 252
divisible by, 331 antennas (in fractals), 322-25
20 Steps around Globe (created by Nie- Aphrodite of Melos (statue), 246-47
meyer), 249-52, 250n21 Apollo Belvedere (statue), 245-46
89 (the number), 213-17, 213n19, Arabic numerals. See Hindu numerals
365-66 architecture, Fibonacci numbers in,
144 (the number), 208, 217, 328 232-44
666 (the number), 212-13 arte povera, 252
arts and Fibonacci numbers, 231-69
abacus, 12, 19n3
Acropolis (Athens), 232, 234 Babbitt, Milton, 288
Adam and Eve (Di.irer), 15, 268 Babylonian formulae, 193
Adam and Eve (Raimondi), 268 Bach, Johann Sebastian, 286
375
376 INDEX
Baptism o/Christ, The (Piero della Brouseau, Alfred, 65

Francesca), 268 Bruch, Hellmut, 254-55
Barbara, Saint, 260-61 Brunelleschi, Filippo, 239
baroque period, 277, 286 bulbs (in fractals), 322-25
bar over digits, meaning of, 14, 14n3, bull markets, 179
30, 108nl Bury, Claus, 255-56
Bartok, Bela, 285-88 business applications of Fibonacci num-
base-sixty numbering system, 21, 21 n6 bers, 177-83
base-ten numbering system, 11
base-two numbering system, 198 Cage, John, 290
Bathers at Asneres (Seurat), 263 calculator or computer used to find a
bear markets, 179, 182 Fibonacci number, 303-304,
bees, male, 13, 59-61 368-69
Beethoven, Ludwig van, 280-82 Cantor, Georg, 308
Beothy, Etienne, 247-48 cardiods (in fractals), 322-25, 322n7
Bernoulli, Daniel, 296 Carmen (Bizet), 11
Bernoulli, Jacob, 131-32 "Cartesian plane," 131
Bernoulli, Nicolaus, I, 296 Cassai, Tommaso. See Masaccio
Bernoulli family, 296n3 Cathedral of Chartres (France), 237-38
"Bigollo," See Fibonacci, Leonardo Center for International Light Art (Ger-
bijugate spirals, 69n8 many), 253-54
binary fonns of music, 274-75, 277 checkerboard, covering a, 188-91
binary numbering system, 198 Cheops (Khufu) Pyramid (Giza). See
Binet, Jacques-Philippe-Marie, 28, Great Pyramid (Giza)
293-94,296 Chinese remainder theorem, 24
Binet fonnula, 207 Chopin, Frederic, 272-74
for finding a particular Fibonacci chords ofa circle, 124, 124nl4
number, 293-301, 328, 366-68 chromatic scale, 286
for finding a particular Lucas num- Chu Shih-Chieh, 84n2
ber, 301-302 circles
binomial coefficient, 88n3 chords of, 124, 124n 14
binomial expansion, 87-88, 88n3 concentric circles, 133
binomial theorem, 356 congruent circles, 139-40
Birth o/Venus, The (Botticelli), 259 and constructing a pentagon, 156-58
bisectors of an angle, 123, 145, 146, 362 great circle, 250, 250n 19
Bizet, George, 1 1 sequence of, 80-81, 84, 87
Bonacci, Guilielmo (William), 17-18 used to construct golden ratio,
Bonacci, son of. See Fibonacci, Leon- 121-22,124,360--61
ardo Circus Parade (Seurat), 263
Boncompagni, Baldassarre, 19n4 Claus, M. See Lucas, Franl'ois-Edouard-
Bondone, Giotto di. See Giotto (di Bon- Anatole (Edouard)
done) coda in music, 281, 281n4
"Book of Squares." See Liber quadrato- codetta in music, 281
rum (Fibonacci) coins and vending machines, 183-84
"book on calculation." See Liber Abaci Collins, Charles, 178, 179
(Fibonacci) Cologne (Gennany), 255-56
Borgliese, Pietro. See Piero della Fran- Columbia University, School of Library
cesca Economy,21lnl6
Botticelli, Sandro, 259 common differences, 64, 78, 82, 330,
Boulez, Pierre, 288 331,364
bracts, counting of, 13, 27, 63--64, common factors, 33-34, 33n20, 94, 192,
65-66 192n5,351-52
Braun, A., 27 See also relati vely prime numbers
INDEX 377
common fractions, 20n5, 24n12, 30 Der goldene Schnitt (scu[pture by U[-

complex numbers, 120, 320 richs),248-49
complex plane, 320, 320n5 Der goldene Schnitt (Zeising), [15,
composite numbers, 35, 35n21, 53-54, I [5n8
352-53 Der goldene Schnitt. Ein Harmoniege-
Composition with Colored Areas and setz und seine Anwendung
Gray Lines I (Mondrian), 269n36 (Hagenmaier), 248
Composition with Gray and Light Descartes, Rene, \3 I
Brown (Mondrian), 269n36 Devaney, Robert,3[9,3[9n4
Composition with Red Yellow Blue development in music, 276, 280, 281,
(Mondrian),269n36 286
compound interest, 24 devil, sign of, 212, 213
computer or calculator used to find a Fibo- Dewey, Me[vi[, 2[ In16, 271nl
nacci number, 303-304, 368-70 Dewey Decima[ classification system,
computers and music, 289 211, 211n16, 271nl
concentric circles, 133 diagonal of the golden rectangle, \38-39
congruent circles, 139-40 diatonic scales, 272
consecutive numbers, 43-44, 55, 56 di fferences
four consecutive Fibonacci numbers, common differences, 64, 78, 82, 330,
211 331,364
odd numbers, 295 Fibonacci differences, 78-80, 82, 83,
ratios of, 109-10, 210 95n9
construction in Lucas numbers, 30 I
constructing a pentagon, ISS-58 pattern in differences of squares,
constructing fractals, 310-17 43-44,45-46,54,55,357,364
constructing the golden ratio, sequences of progressive differences,
120-24,362-65 102
continued fractions and Fibonacci num- sum and difference, 296-98
bers, 161-75, 162nl and sum of successive powers,
finite continued fractions, 163 297-98
golden ratio as a continued fraction, digits
166-72 curiosity of, 208
infinite continued fractions, 163-64 first-digit patterns, 207-208
Cook, Theodore Andrea, 245 last-digit patterns, 206-207
corrective waves, 178, 179 See also integers; numbers
Cossa[i, Pietro, [7n [ Di minor guisa (Fibonacci), 20
Couder, Yves, 74 Dionysius' Procession (re[ief at Villa
credit cards, measurements of, 182 A[bani),267
Crucifixion (Raphae[), 268 distances, conversion of measures of,
crystallography, 342 200-203
cubits (as a measurement), 236 divergence angle, 64n3
curves, 72, 13[, [32-33,247,308,322 "divine proportion," 245
Gaussian curve, 246, 246n 13 division
Curves of Life, The (Cook), 245 common divisor, 337
denominator, 23
Da[i, Salvador, 268 divisibility of Fibonacci numbers,
Davis, T. Antony, 76 47-48,55-56,330,341,350,
decimal expansion, [08n2 358
De divina proportione (Pacio[i), 245, divisors of composite numbers, 353
257,260 numerator, 23
denominator, 23 sequences of remainders, 31-32
Depositionfrom the Cross (Weyden), dodecahedron, 231
268 Dodgson, Charles Lutwidge, 140
378 INDEX
dominant in music, 285, 285n5 Feininger, Lyonel, 268

dominoes Fibonacci, Leonardo, II, 17-22,
covering a checkerboard, 188-91 23nn8-9,56-57,266,327
knocking down as example of Fibonacci Applications and Strategies
mathematical induction, 349-50 for Traders (Fischer), 182
door, height of, 241 Fibonacci Association, 17, 28, 329
Douady, Stephane, 74 Fibonacci multiplication algorithm,
"Dow theory" of investing, 178 198-99
drones, 13, 59~1 Fibonacci Napoli (Merz), 252
Duchamp, Gaston, 247, 247nl4 Fibonacci Nim, 225-26
Dudley, Underwood, 119 Fibonacci numbers, mathematical as-
DUrer, Albrecht, 15, 155, 158,258-59, pects
259n26,268 definition of, 53
dyad in music, 285, 285n7 introduction to, 26-32
dynamic symmetry, 245, 245nl0 list of first 500, 343--48
and Lucas numbers, 15, 97-98,
Egyptian pyramids, 180,234-37 104-105,227,228-30,297-98
Elements (Euclid), 20, 171 and the Pascal triangle, 90-98, 99,
Elements of Dynamic Symmetry, The 102,103,104-105
(Hambidge),245 proofs of Fibonacci relationships,
Elliott, Ralph Nelson, 177, 178-83 349-58,360~9
ellipse, area of, 133, 133n 18 properties of, 33-56, 337--41
Empire State Building, climbing stairs and Pythagorean triples, 192-94
of, 185 testing to determine if it is, 304-305,
equal binary form of music, 274-75, 277 305n6
equations See also golden angle; golden ratio
integer polynomial equation, 165n2 (section); golden rectangle;
linear equations, 23 golden spirals
polynomial equation, 165n2 Fibonacci numbers, use of
"equiangular spiral," 131 in art and architecture, 231-69
Essor II (sculpture), 248 business applications of, 177-83
Euclid, 12, 19,20,84, l71, 307 climbing a staircase, 184-85
Euler, Leonhard, 120, I 65n2, 296 converting miles and kilometers,
Euler's number, 120, 165, 165n2 200-203
even-positioned Fibonacci numbers, 47, determining path of fish in a hatch-
210,330 ery,222-24
sum of, 37-38, 54, 353 determining seating at events,
event seating and Fibonacci numbers, 221-22
221-22 and fractals, 307-25
Evolution: Progression and Symmetry in geometry, 136--43
III (Mields), 266-67 in music, 271-91
Evolution: Progression and Symmetry in nature, 12, 13,25-27,57,59-76,
W(Mields),266-67 132,133
expansion, binomial, 87-88, 88n3 in optics, 203-206
exposition in music, 280 and painting a house, I 86
and physics, 203-206
factors, common, 33-34, 33n20, 192, and watch displays, 217-20
192n5,351-52 Fibonacci phyllotaxis, 64n3, 70--74
"factors of a multiplication," 23 Fibonacci ratios, 94, 107-10, 180, 182
"factors of a number," 23 See also golden ratio (section)
Fatou, Pierre, 308 Fibonacci spirals. See golden spirals
Fechner, Gustav, 115-17 Fibonacci's Temple (Bury), 255-56
feet (as a measurement), 241, 241n4 Fifth Symphony (Beethoven), 280-82
INDEX 379
finding a particular Fibonacci number Girl with the Ermine, The (da Vinci),
in a fractal, 319, 319n4 268
using a calculator or a computer, Giza, Great Pyramid at, 180,234-37
303-304,368-369 golden angle, 148-49,250, 250n20
using Binet formula, 293-301, and plants, 64, 64n3, 74
366--69 golden ratio (section), 13-15, 13n2,
using golden ratio, 303 74-76,107--60,180,263,296,327
finite continued fractions, 163 in art and architecture, 23 1-69
first differences. See differences and congruent circles, 139-40
first digits of Fibonacci numbers, and consecutive Fibonacci numbers,
207-208 210
Fischer, Robert, 182 constructing, 120-24,360-63
fish, determining path in a hatchery, as a continued fraction, 166-72
222-24 as an irrational number, 149-51
fivefold symmetry element, 342 and Lucas numbers, 174-75
fixed integer, 365 and measurements, 201
Florence (Italy), 20, 238-40 in music, 271-91
Flos (Fibonacci), 20, 22 paper-folding and, 159
forbidden symmetry, 342 powers of, 113-14
Four Books on Human Proportions precise value of, 111-12, 166n4
(Diirer), 258 golden rectangle, 115-20, 123, 125-27,
fourth di fferences. See di fferences 128-29, 172
fractals, 307-25 in art and architecture, 232-33, 237,
fractions 259,263
common fractions, 20n5, 24n 12, 30 and credit cards, 182
continued fractions and Fibonacci diagonal of, 138-39
numbers, 161-75, 162nl and watch displays, 217-20
improper fractions, 161-62 Golden Section, The (Hagenmaier), 248
infinite continued fractions, 163--64 golden spirals, 124-33,263,266
proper fractions, 161, 162, 173 See also spiral patterns
unit fractions, 162 golden triangle, 107, 144-48,260
use of horizontal bar in, 25 in pentagrams, 149-54, 158
French Academy of Sciences, 200 great circle, 250, 250n 19
French National Assembly, 200 Great Crash of 1929, 177
Fridfinnsson, Hreinn, 266 Great Pyramid (Giza), 180,234-37
fugues in music, 285-86, 287 Grevsmiihl, 250n21
furlongs (as a measurement), 200, Grimaldi, Giovanni Gabriello, 17n I
200nll Gris, Juan, 269, 269n37
Grossman, George W., 313n3
Galileo Galilei, 258 Grossman truss, 313-19, 313n3
game of Fibonacci Nim, 225-26
Gauss, Carl Friedrich, 328 Hadrian's Arch, 234
Gaussian curve, 246n 13 Hagenmaier, Otto, 245, 248
Gelmeroda (Feininger), 268 Half a Giant Cup Suspended with an
generating fractals, 3 10-17 Inexplicable Appendage Five Me-
geodesic draft, 250n21 ters Long (Dali), 268
Geometric Compositions (Niemeyer), Hambidge, Jay, 245
249,263--64 Hansel und Gretel (Humperdinck),
geometry, Fibonacci numbers in, 11-12
136-43 harmonics in music, 283
Gherardo, Giovanni di (da Prato), 239 Haydn, Franz Joseph, 278-80
Ghyka, Matil Costiescu, 243 Hermite, Charles, 165n2
Giotto (di Bondone), 17,267, 267n29 Herodot, 235
380 INDEX
Heron of Alexandria, 120-21,362 Journal of Recreational Mathematics,

hexagon, 155 213
Hindu numerals, II, II n I, 19, 23, 23n8 Julia, Gaston, 308, 320
Hippasus of Metapontum, 149, 153
Hisdb al-jabr w'almuqabdlah (al- Kepler, Johannes, 27, 119,258
Khowarizmi), 19 Khayyam, Omar, 20, 84n2
Hispanus, Dominicus, 20 Khufu (Cheops) Pyramid (Giza). See
Hommage a Fibonacci (Bruch), 254-55 Great Pyramid (Giza)
horizontal fraction bar, 25 kilometers, conversion to miles,
human body and Fibonacci numbers, 74, 200-203
75-76,241-42,257,258 Knott, Ron, 91
Humperdinck, Engelbert, 11-12 Koch, Helge von, 310n I
Hunter, J. A. H., 136 Koch snowflake, 3 I 0-12
IBM, Watson Research Center, 320 Lambert, Johann Heinrich, 165n2

Igloo Fibonacci (Merz), 253 Lame, Gabriel, 29-30
immediate successor, 52 Lame numbers, 30
improper fractions, 161-62 La Serie d'Or (Beothy), 247-48
impulsive waves, 178, 179 Last digit of Fibonacci numbers,
Indian figures. See Hindu numerals 206-207
induction and proofs of Fibonacci num- Last Supper, The (da Vinci), 268
bers, 349-69 leaf arrangements, 71-74
infinite continued fractions, 163-64 Leaning Tower of Pis a, 18,23
integer polynomial equation, 165n2 Le Corbusier, 75, 76, 240-43, 244, 256
integers, 20n5, 21, 24n12, 33n20, Leochares,245-46,246nI2
112n7, 149, 196,211 Leonardo da Vinci, 74-75, 75n 14, 257,
fixed integers, 193 260,268
and parity, 193, 193n6 Leonardo of Pisa (Leonardo Pisano). See
and pentagons, 152 Fibonacci, Leonardo
and perfect squares, 21 n 7 Liber Abaci (Fibonacci), 11, 12, 19,20,
positive integers, 48, 55, 162,217, 22-27
225,331-32,333,336-37,338, See also Fibonacci numbers; rabbits,
339,363,366 regeneration of
and Pythagorean triangle, 193 Liber quadratorum (Fibonacci), 20
special integers, I 14 linear equations, 23
See also digits; numbers logarithms
investments and Fibonacci numbers, and Euler's number, 120, 165n2
178-83 logarithmic spiral, 69,127,131-32,
irrational numbers, 20, 20n5, 112, 133,147,182
112n7, 149-51, 153, 163, 165n2, Lonc, Frank A., 75-76
232,300-301 Lucas, Franl'ois-Edouard-Anatole
isosceles triangles, 144-48, 151, 158, (Edouard), 15, 27-28, 27n 15,97
195,261,313,314 Lucas numbers, 15, 27-28
iterations of fractals, 310, 31 I, 320-22, Binet formula for finding a particular
321n6 number, 301-302
in Grossman truss, 313-18 definition of, 53
Ivy (Merz), 252 and Fibonacci numbers, 15, 97-98,
104-105,227,228-30,297-98
Japanese pagodas, 237 and golden ratio, 174-75
Jeanneret, Charles-Edouard. See Le and golden spirals, 13 I
Corbusier Lucas spirals, 69n8
Johannes of Palermo, 20 and Pascal triangle, 97-104
Johnson, Tom, 289-90 proofs for, 360-61
INDEX 381
properties of, 49-53 music, Fibonacci numbers in, 271-91

as triangular numbers, 212 Music for Strings, Percussion, and Ce-
Ludolph's number, 120 lesta (Bartok), 285-88
Madachy, Joseph S., 171 Narayan's Cows (Johnson), 290

Madonna Alba (Raphael), 261-62 natural numbers, 80n I, 120, 188n2,
Madonna and Child (Perugino), 268 202n13, 207-208,30In4, 350
Madonna Doni (Michelangelo), 268 ordered sum of ones and twos, 187
main antenna (in fractals), 323 prime natural numbers, 193
main cardioid (in fractals), 322-25 sequences of, 241
Mandelbrot, Benoit, 308-309, 320 as sum of Fibonacci numbers, 188
Mandelbrot set, 319-25, 321 n6 nature, Fibonacci numbers in, 12, 13,
Masaccio, 267n30 25-27,57,59-76, 132, 133,
mathematical induction and proofs of 222-24
Fibonacci numbers, 349-<i9 Nature's Law-The Secret of the Uni-
measurements verse (Elliott), 179, 180
cubits, 236 nautilus shells, 132
feet, 241, 241 n4 "nest of radicals," 173
furlongs, 200, 200n II Neveux, Marguerite, 269
meters, 241 , 241 n4 Niemeyer, Jo, 249-52, 263-<i5
miles and kilometers, 200--203 nonconsecutive Fibonacci numbers, 188,
Mersenne, Marin, 131 202n13
Mersenne prime numbers, 69, 69n7 non prime numbers, 35, 35n21, 54
Merz, Mario, 252-54 notation
Mexican pyramids, 237 binary notation, 225
Michelangelo Buonarroti, 268, 268n34 and Fibonacci numbers, 37, 41,
Mields, Rune, 266-67 330
miles, conversion to kilometers, and Lucas numbers, 50
200--203 musical notation, 274
minor modulus, 331-36, 337-41 numbering systems
Modulon (Niemeyer), 264-65 base-IO, II
Modulor: A Harmonious Measure to the base-60, 21, 21 n6
Human Scale Universally Applica- binary system (base-2), 198
ble to Architecture and Mechanics, Hindu numerals, II, 19,23
The (Le Cobrusier), 241, 242, 256 Roman numerals, 12
Moivre, Abraham de, 296 numbers
Mona Lisa (da Vinci), 260 algebraic number, 165n2
Mondrian, Piet, 269, 269n36 alternating numbers, 42, 42n23,
Mozart, Wolfgang Amadeus, 275-80, 44-45, 54, 54n25, 55, 142
289 complex numbers, 320, 320n5
multijugate phyllotaxis, 70 composite numbers, 35, 35n21,
multiplication, 23, 216, 225-26n24, 299, 53-54,352-53
303,312,365-66 consecutive numbers, 43-44, 55, 56,
finding a sum by multiplying, 41, 52, 109-10,210,211,295
53 irrational numbers, 20, 20n5, 112,
and golden ratio, 114, 125 112n7, 149-51, 153, 163,
and multiples of 6, 331 165n2,232,300-301
multiplication algorithm and Fibo- natural numbers, 80n I, 120, 187-88,
nacci numbers, 195-99 188n2,193,202nI3,207-208,
products of Fibonacci numbers, 241,30In4,350
44-47,56 nonconsecutive Fibonacci numbers,
and Pythagorean triples, 192 187,202n13
writing quotients, 30n 19 nonprime numbers, 35, 35n21
382 INDEX
odd prime numbers, 208, 328, patterns

333-36,337-41 in differences of squares, 43-44,
palindromic numbers, 93-94, 93n6 45-46,54,55,357,366
pentatop numbers, 89 of even Fibonacci numbers, 47
perfect numbers, 24, 24n II first-digit patterns, 207-208
prime natural numbers, 193 last-digit patterns, 206-207
prime numbers, 33-34, 33n20, ordered sum of ones and twos, 187
35n22, 54, 69n7, 131,211,212, of spirals in plants, 66-70
331-37 in squares of Fibonacci numbers, 42,
primitive numbers, 339 43-44,45-46,355-59
rational numbers, 24, 24n 12, 30, 163 symbol to show repetition ofnum-
reciprocals, III, 113,162,169,201, bers, 14, 14n3, 30, 108nl
214,362 "wave theory" of investing, 178-83
relatively prime numbers, 33-34, See also sequences
33n20, 53, 193,210,228,337, pentagon, 148, 148n23, 149-54,231,
338,351-52 259,259n26
successive Fibonacci numbers, constructing, 155-58
209-11 paper-folding and, 159
symbol to show repetition ofnum- Pentagon (Washington, DC), 244
bers, 14, 14n3, 30, I08nl pentagram, 148, 149-54
tetrahedral numbers, 89, 89n5, 90 paper-folding and, 159
transcendental numbers, 165n2 pentatop numbers, 89
triangular numbers, 89, 89n4, 90, Perez, Jose Victoriano Gonzalez. See
211-13 Oris, Juan
See also digits; integers; squaring perfect numbers, 24, 24n II
numbers perfect square, 21-22, 21n7, 78, 305,
numerator, 23 305n6
Pericles, 232
odd-positioned Fibonacci numbers, sum "periodicity," 14, 14n3, 30, I08nl, 350
of, 38-39, 54, 295, 355-56 periodic decimals, 112n6
odd prime numbers, 208, 328, 333-36, period of bulb or decoration (in fractals),
337-41 323-24
optics and Fibonacci numbers, 203-206 Persian War, 232
ordered sum of ones and twos, 187 Perugino, Pietro, 268, 268n33
petal arrangements, 71, 74
Pacioli, Luca, 243, 245, 257, 257n23, Petrie, W. M. F., 236
260 Pfeiffer, Jorg, 250n21
pagodas, 237 Phidias, 110, 232
painting a house and Fibonacci numbers, phyllotaxis, 64n3, 70-74, 70n I 0
186 physics and Fibonacci numbers,
Painting [ (Mondrian), 269n36 203-206
paintings, Fibonacci numbers in, Physicus, Theodorus, 20
257-69 ~,119-20, 165-66, 166n5,237
palindromic numbers, 93-94, 93n6 Piano Sonatas (Haydn), 278-80
Pankok, Otto, 269, 269n38 Piano Sonatas (Mozart), 277-78
paper-folding and pentagon, 159 Piero della Francesca, 268, 268n32
parastichies, 64-66 Pietro di Benedetto dei Franceschi (Pie-
parity, 193, 193n6 tro Borgliese). See Piero della
Parthenon (Greece), 15, 111,232-33, Francesca
245 pineapples, 13, 63-64
Pascal, Blaise, 84, 84n2 pinecones, 13,27,64-66
"Pascaline" (calculator), 84 Pisa (Italy), 18, 20, 22
Pascal triangle, 84-105, 84n2 Leaning Tower of Pisa, 18, 23
INDEX 383
pitch generator, 290 ratios

plants, Fibonacci numbers in, 63-76 Fibonacci ratios. See Fibonacci ratios
Plato, 153 golden ratio (section). See golden
Pollio, Vitruvius, 257-58, 257n24 ratio (section)
Polo, Marco, 17 and irrational numbers, 149
polynomial equation, 165n2 ratio of similitude, 128, 128n 16
poor art, 252 recapitulation in music, 276, 277, 280,
positive integers, 55, 162,217,225, 281
331-32,333,336-37,339,363, reciprocal rectangles, 128, 129
366 reciprocals, II I, 113, 169, 20 I, 214, 362
Praclica geomelriae (Fibonacci), 19 recreational mathematics, 24, 25, 27n 15,
Prato (Giovanni di Gherardo da). See 94, 94n7, 213
Gherardo, Giovanni di (da Prato) rectangles
Prelude No.1 in C major (Chopin), 274 created using Fibonacci numbers, 40,
Prelude No.9 in E major (Chopin), 273 48-49
Preludes (Chopin), 272-74 created using Lucas numbers, 51
primary bulb (in fractals), 322, 324 golden rectangle, 115-20, 123, 125-
prime numbers, 35n22, 54, 131, 211, 27,128-29,138-39,172,182,
212,331-37 217-20,232-33,237,259,263
Mersenne prime numbers, 69, 69n7 puzzle concerning area of, 140-43
odd prime numbers, 208, 328, reciprocal rectangles, 128, 129
333-36,337-41 and triangles of equal area, 136-37
prime natural numbers, 193 recursive sequence, 26, 27, 27n 13,
relatively prime numbers, 33-34, 31-32,96-97,185,186,293-94,
33n20, 53,193,210,228,337, 295,296,300
338,351-52 in fractals, 310, 312
special primes, 336-37 and Lucas numbers, 97, 99
primitive numbers, 339 reflections in glass and Fibonacci num-
products of multiplication, 44-47, 56 bers, 203-206
progressive differences, 102 relatively prime numbers, 33-34, 33n20,
proper fractions, 161, 162, 173 53, 193,210,228,337,338,
Propylaeum temple (Greece), 234 351-52
pyramids, 180, 234-37 remainders, sequences of, 31-32
Pythagoras and Pythagoreans, 149, 153, Rembrandt Harmenszoon van Rijnlnl,
231 268
Pythagorean theorem, 119, 119n 11, 134, Renaissance, 17,231,239,245,258,
235,360 267n30
using Fibonacci numbers to generate ring, area of, 133
Pythagorean triples, 192-94, Roger de Le Pasture. See Weyden,
363-65 Rogier van der
Roman mile (as a measurement), 200
quadratic formula, 113, II?, Il?nlO, Roman numerals, 12, 20
150 Romantic period, 280, 288
"queen of science," mathematics as, 328 Rome, 237
Quetelet, Lambert Adolphe Jacques, 75 Rubie's Cube, 94n7
quotients, writing of, 30n 19 rule for fractals, 320
rumors, distribution of, 61-62, 62n2
rabbits, regeneration of, 12,25-27,57, Russian peasant's method of multi plica-
61 tion, 195-99
radicals, nest of, 173
Raimondi, Marcantonio, 268, 268n35 Sacher, Paul, 285
Raphael Santi, 260-62, 268 SI. Francis Preaching 10 Ihe Birds
rational numbers, 24, 24n 12, 30, 163 (Giotto),267
384 INDEX
Salinger, J. D., 12 golden spirals, 124-33,263

Santa Maria del Fiore Cathedral (Italy), logarithmic spiral, 127, 131-32, 133
238-40 and pentagons, I 59-60
scales in music, 271-72, 282-83, spokes (in fractals), 323
285nn5-6,286 square roots, 305
Schimper, C. F., 27 squares
Schonberg, Arnold, 288-89 created using Fibonacci numbers, 40,
School of Athens, The (Raphael), 268 48-49
Scotus, Michael, 20 created using Lucas numbers, 51
sculpture, Fibonacci numbers in, 244-56 squaring numbers, 293-96
seating at events and Fibonacci num- alternating numbers and, 44-47, 55,
bers, 221-22 142
second differences. See differences pattern in differences of squares,
Section d'Or (artist association), 247 43-44,45-46,54,55,355,364
seed of the fractal, 310-11, 320 patterns in squares of Fibonacci
Self-Portrait, A (Rembrandt), 268 numbers, 42, 43-44, 45-46
sequences, 80n I perfect square, 21-22, 21n7, 78, 305,
of circles, 80-81, 84, 87 305n6
and Fibonacci numbers, 77-84, 114, subtracting squares of Fibonacci
174-75 numbers, 42-43, 355-59
and Lucas numbers, 174-75 sum of squares of Fibonacci num-
of natural numbers, 241 bers, 39-42, 43, 54,353-54
and Pascal triangle, 87-105 sum of squares of Lucas numbers,
of progressi ve di fferences, 102 51-53,56
recursive sequence, 27, 27n13, transforming fractals, 320-21
31-32,96-97,293-94 See also numbers
of remainders, 31-32 staircases, Fibonacci numbers and
See also patterns climbing, 184-85
Serusier, Paul, 269 static symmetry, 245
Seurat, Georges, 262-63 statute mile (as a measurement.), 200
sexagesimal number. See base-sixty stock market and Fibonacci numbers,
numbering system 178-83
Sierpinski, Waclaw, 312n2 Stonehenge (Great Britain), 237
Sierpinski gasket, 312-13 Stradivarius, Antonio, 291
Sigler, Laurence E., 19n4 subtracting squares of Fibonacci num-
Signac, Paul, 269 bers, 42-43, 357-61
"Sign of the Devil, The" (Wang), 213 successive Fibonacci numbers, 209-11
similitude, ratio of, 128, 128n 16 sum
Simson, Robert, 171 angle sum of a triangle, 84
Singmaster, David, 94 and difference of successive powers,
Sistine Madonna (Raphael), 260-61 297-98
Sixtus II (pope), 260-61 of even-positioned Fibonacci num-
snails, 133 bers, 37-38, 54, 353
sonata-allegro form of music, 275-76, of Fibonacci numbers and distance
281 conversion, 202, 202n 13
sonatas as a musical form, 275-80, Fibonacci numbers in Pascal trian-
276n2 gle,91-94
special integer, I 14 finding a sum by multiplying, 41, 52,
special primes, 336-37 53
spiral patterns, 65-74, 69n8, 131 formulas for getting sum of Fibo-
artificial Fibonacci-flower spirals, 70 nacci numbers, 36-42, 53, 54
bijugate spirals, 69n8 formulas for getting sum of Lucas
"equiangular spiral," 131 numbers, 49-51,56
INDEX 385
natural numbers as sum of Fibonacci triples, Pythagorean, 192-94,365-67

numbers, 188, 365 Tristan und Isolde (Wagner), 282-85
of odd-positioned Fibonacci num- Triumph a/Galatea, The (Raphael), 268
bers, 38-39, 54, 295, 353-54 twelve-tone method in music, 288-89
ordered sum of ones and twos, 187 20 Steps around Globe (created by Nie-
in Pascal triangle, 85-87 meyer), 249-52, 250n21
sequences of sums of Fibonacci
numbers, 82-83 Ulrichs, Timm, 248-49
sum and difference, 296-98 unequal binary form of music, 274-75
sum of squares of Fibonacci num- United Nations headquarters (New York
bers, 39-42, 43, 54, 353-54 City), 243
sum of squares of Lucas numbers, unites d'habitation (France), 240
51-53,56 unit fractions, 162
summand, 225, 225n23 UNIV AC (computer), 289, 289n8
sunflowers, 68-69, 69n8 Unna (Germany), 253-54
Swiss Guard, uniforms for, 268n34 Untitled (Fridfinnsson), 266
symbol to show repetition of numbers, Utsjoki (Niemeyer), 263-64
14, 14n3, 30, 108nl
symmetry, 245, 245n10, 266, 342 "Valuable Mirror of the Four Elements,
forbidden symmetry, 342 The" (Khayyam), 84n2
Vannucci, Pietro. See Perugino, Pietro
Temple of Olympia, III Variation VI (Niemeyer), 264
testing numbers to find Fibonacci num- vending machines and Fibonacci num-
bers, 304-305, 305n6 bers, 183-84
test of numbers in fractals, 320, 321, Venus de Milo (statue). See Aphrodite
321n6 of Mel os (statue)
tetrahedral numbers, 89, 89n5, 90 vertex angle, 144, 147
thermodynamics, laws of, 245nl0 Villon, Jacques, 247, 247n 14
39th Mersenne prime number, 69, 69n7 violin as example of golden ratio, 271,
Thorndike, Edward Lee, I 17 291
tonic in music, 285, 285n6 "Vitruvian" man, 257
Tower of Hanoi puzzle, 27nl5 Vitruvius,257-58
transcendental numbers, 165n2
triangles Wagner, Richard, 282-85, 287
angle sum of a triangle, 84 Wang, Steve C, 213
and constructing the Grossman truss, watches, displaying of, 217-20
314-19 Watson Research Center, IBM, 320
Fibonacci numbers as lengths of Wave Principle. The (Elliott), 179
sides, 195 "wave theory" of investing, 178-93
golden triangle, 144-48, 149-54 Webern, Anton von, 289
isosceles triangles, 144-48,314 Weyden, Rogier van der, 268, 268n31
Omar Khayyam and, 84n2 worker bees, 59
Pascal triangle, 84-105
Pythagorean triangles and Pythago- Zeckendor~Edouard, 188n4,202n13
rean triples, 192-94 Zeckendorf theorem, 188n4
triangular numbers, 89, 89n4, 90, Zeising, Adolph, 75, 115, 115n8, 243
211-13,212nI7 zephyr (zero), II, 19, 23
Trinity (Masaccio), 267, 267n30 Zeus (statue of), I I I

The Fabulous Fibonacci Numbers TQW - Darksiderg

Uploaded by

Copyright:

Available Formats

The Fabulous Fibonacci Numbers TQW - Darksiderg

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

The Fabulous Fibonacci Numbers TQW - Darksiderg

Uploaded by

Copyright:

Available Formats

The (Fabulous)

The Fabulous Fibonacci Numbers. Copyright © 2007 by Alfred S. Posamentier

Inquiries should be addressed to

Printed in the United States of America on acid-free paper

To my wife and life partner, Sabine, without whose support

Chapter 1: A History and Introduction to the

Chapter 2: The Fibonacci Numbers in Nature 59

Chapter 3: The Fibonacci Numbers and the

Chapter 9: The Famous Binet Formula for Finding

Chapter 10: The Fibonacci Numbers and

Afterword by Herbert A. Hauptman 329

Appendix A: List of the First 500 Fibonacci Numbers,

Appendix B: Proofs of Fibonacci Relationships 349

T he authors wish to thank Professor Stephen Jablonsky of the

The success of a book is usually dependent on two types of

The Fabulous Fibonacci Numbers

I n a remote section of the Austrian Alps, there is a long-

The nine Indian 1 figures are: 9 8 7 6 5 4 3 2 1.

This constitutes the first formal mention of our base-ten numeral

Gretel, or J. D. Salinger (b. 1919) for The Catcher in the Rye, we

this sequence of numbers can go on infinitely, as it begins with two

l/J = 1.6180339887498948482045868343656 ...

Then consider the quotient of the somewhat larger pair of con-

and the quotient of an even larger pair of consecutive Fibonacci

Notice how these increasingly larger quotients seem to sur-

165 580 141

Compare this last quotient to the value of the golden ratio:

A History and Introduction to

W ith the dawn of the thirteenth century, Europe began to

Bonacci," or family of Bonacci. He was born to Guilielmo (Wil-

In 1192 Guilielmo Bonacci became a public clerk in the cus-

calculate monies due them. This meant determining currency

2. The name "algebra" comes from the title of this book.

Here Fibonacci uses algebraic methods to solve geometric problems

5. Irrational numbers are those that cannot be expressed as a common fraction

However, with today's computer algebra system we can get the

2~3, 930 352 2~3, 930 352 2

Another of the problems with which he was challenged is one

We have then shown that 41

6. A number in base 60.

tract 1, 2, 3, or 4 instead of 5, the problem could not have been

Three people are to share an amount of money in the

second person then returns ! of what he took, and the

third person returns ! of what he took. When the total of

Although none of Fibonacci's competitors could solve any of

The Book Liber Abaci

Although Fibonacci wrote several books, the one we will focus on

The nine Indian 8 figures are: 9 8 7 6 5 4 3 2 1.9

Despite their relative facility, these numerals were not widely

The second section of Liber Abaci includes a large collection

There are also problems involving perfect numbers, II problems

because it casually includes a recreational mathematics problem in

The Rabbit Problem

To see how this problem would look on a monthly basis, con-

rabbits matures in one month to become offspring-producing adults

This problem generated the sequence of numbers:

The Fibonacci sequence is the oldest known (recursive) recurrent

Introducing the Fibonacci Numbers

These numbers were not identified as anything special during the