Polyhedron Models

21:28:45 BST 2016.

ebook.jsf?bid=CBO9780511569746
 21:28:45 BST 2016.
ebook.jsf?bid=CBO9780511569746
Polyhedron Models
Magnus J . Wenninger

The right of the

University of Cambridge
to print and sell
all manner of books
was granted by
Henry VIII in 1534.
The University has printed
and published continuously
since 1584.

Cambridge University Press

Cambridge
New York Port Chester
Melbourne Sydney

21:28:45 BST 2016.

ebook.jsf?bid=CBO9780511569746
CAMBRIDGE UNIVERSITY PRESS
Cambridge, New York, Melbourne, Madrid, Cape Town, Singapore, Sao Paulo, Delhi

Cambridge University Press

The Edinburgh Building, Cambridge CB2 8RU, UK

Published in the United States of America by Cambridge University Press, New York

www.cambridge.org
Information on this title: www.cambridge.org/9780521069175
© Cambridge University Press 1971

This publication is in copyright. Subject to statutory exception

and to the provisions of relevant collective licensing agreements,
no reproduction of any part may take place without the written
permission of Cambridge University Press.

First published 1971

First paperback edition 1974

Reprinted 1975, 1976, 1978, 1979, 1981, 1984, 1985, 1987, 1988, 1989

A catalogue record for this publication is available from the British Library

Library of Congress Catalogue Card Number: 69-10200

ISBN 978-0-521-06917-5 hardback
ISBN 978-0-521-09859-5 paperback

Transferred to digital printing 2008

21:28:45 BST 2016.

ebook.jsf?bid=CBO9780511569746
 In grateful memory of T. G. who tutored me in
the Philosophy of Mathematics

21:30:37 BST 2016.

ebook.jsf?bid=CBO9780511569746
Mathematics possesses not only truth but
supreme beauty, a beauty cold and austere, like
that of sculpture, sublimely pure and capable of
a stern perfection, such as only the greatest art
can show,
Bertrand Russell

VI
21:30:37 BST 2016.
ebook.jsf?bid=CBO9780511569746
Contents

Preface to 1978 reprint viii

Preface ix Commentary on the stellation of the
Foreword xi Archimedean solids 66
Introduction: uniform polyhedra 1 Stellations of the cuboctahedron (43-46) 68
Mathematical classification 4 Commentary on the icosidodecahedron 73
Stellations of the icosidodecahedron
(47-66) 76
Section I The Convex Uniform Polyhedra:
The Platonic and Archimedean Solids Section HI Non-convex Uniform Polyhedra
General instructions for making models 12 Commentary 98
The Platonic solids (1-5) 14 General instructions for making models
20 of non-convex polyhedra 100
The Archimedean solids (6-18)
Non-convex polyhedra (67-109) 101
Section II Some Stellations and Compounds Commentary on non-convex snub poly-
Commentary on stellations and com- hedra 171
pounds of the Platonic solids 34 Non-convex snub polyhedra (110-119) 172
The stellated octahedron (19) 37 A final comment 204
Stellations of the dodecahedron (20-22) 38 Epilogue 204
Commentary on the stellated icosahedron 41 References 206
Stellations of the icosahedron (23-42) 43 List of models 207

VII
21:31:11 BST 2016.
ebook.jsf?bid=CBO9780511569746
Preface to 1978 reprint

The study of polyhedra is one area of mathe- demands that you make your own full scale
matics in which the ordinarily abstract and drawings of all facial planes from which the
speculative considerations of the subject find patterns or nets are derived.
very pleasing and attractive visual applications. For the beginner or the inexperienced this has
It is also an area in which both the amateur and great educational value. For those who already
the expert in mathematics can work with equal possess the required expertise this will avoid their
delight. being led astray by the book. On the other hand
The enthusiastic response which greeted the in all cases it must be remembered that a model
first publication of this book (1971) provided is a model. The full delightfulness of any poly-
ample evidence of this fact, as well as its appear- hedron model must ultimately be a matter of
ance in a paperback edition (1974), reprinted in a intellectual insight.
hardcover edition (1975, 1976) and now in still It may be of interest for readers to know that
another reprint (1978). a definitive enumeration of uniform polyhedra
A book which contains as many geometrical has now been made. John Skilling of the Depart-
drawings as this one, all of which demanded ment of Applied Mathematics and Theoretical
careful draughtsmanship in the originals, could Physics at the University of Cambridge has
hardly have been printed without errors in their shown that ' the list of Coxeter et al. is indeed
reproduction. Every effort was made to correct complete as regards uniform polyhedra in which
these where feasible in the reprinted editions. only two faces meet at any edge. The natural
Further corrections are being incorporated in generalization that any even number of faces
this present edition. may meet at an edge allows just one extra poly-
A word of caution is in place here for the hedron to be included in the set' (John Skilling,
dedicated model maker. Although the original ' The complete set of uniform polyhedra,' Phil.
intention of the book was to provide patterns Trans. Royal Society of London. Series A, vol.
that can be traced from the book and thus used 278, no. 1278).
directly in making the models, you may find that M.J.W.
for best results very careful workmanship still January 1978

VIII
21:31:25 BST 2016.
CBO9780511569746.001
Preface

This book presents a well-defined set of geo- them for?' Maybe the answer to this is best
metrical solids, the seventy-five (known) uniform given by a return question:' Does beauty need to
polyhedra, together with a representative set of have uses ?' Admittedly the only use a model has,
stellated forms. A description of the underlying once it has been constructed, is for display pur-
theory of polyhedra is included to bring out the poses. You can make some very attractive
relationships that exist between the various mobile models, and generally the constructions
solids. But mainly this book is simply a set of make lovely mantelpieces or centrepieces for
instructions on how to make models of these tables at a banquet on special occasions. Stars
solids. seem to go with Christmas and here you have
The sources in which you can find an account many star forms to choose from.
of the mathematical theory of this topic are given But on a more technical level you may have
at the end of the book. If in the past you found seen polyhedron forms used for space satellites.
the study of geometry a bit difficult, or if at Then again the geodesic dome is found in archi-
present you are not particularly attracted by tecture and in engineering projects. Perhaps the
geometry, you may wonder if this topic will hold polyhedral forms presented in this book have
your interest. The fact is that you really do not never been used simply because they have never
need to understand all the theoretical mathe- been widely known.
matics involved in the original discovery and I have myself constructed all the models pre-
classification of these solids. On the other hand sented here and shown in the photographs. How
you cannot avoid all the mathematics, especially long did it take me? My interest in this topic
the terminology used here and some of the sym- began in 1958 with a summer course at Columbia
bolism. Teachers College in New York. During the fol-
The objective in this book will be to set down lowing year I made my first set of models, those
an explanation of the solids, at once simple and given in section i of this book. My main source was
practical and not too speculative, one sufficient Mathematical recreations and essays by Coxeter
for the purposes of constructing the models. It and Ball. Then between 1959 and 1961 I made
is really surprising how much enlightenment will all those in Mathematical models by Cundy and
come, following the construction of the models Rollett. Next I tackled The fifty-nine icosahedra
rather than preceding it, and once you begin by Coxeter, Du Val, Flather, and Petrie. I suc-
making them you may find that your enthusiasm ceeded in working out my own nets for each of
will grow. You will soon see that each of these these. The set graced the back wall of my mathe-
solids has a beauty of form that appeals to the matics classroom, growing as I completed each
eye in much the same way that the abstract one between 1961 and 1963. The average work-
mathematics appeals to the mind of a mathe- ing time spent on each was about eight hours,
matician: plus three or four hours each to discover suitable
You may find the number of models presented nets. On the completion of this project I wrote to
here overwhelming, some of them extremely Professor Coxeter asking about Uniform poly-
complex. Why should anyone want to make hedra. He kindly sent me a complimentary copy,
them? Maybe the answer is to be found in the one of three he still had in his possession. This
reply of a mountain climber when he was asked, monograph is a detailed account of the mathe-
'Why do you want to climb the Matterhorn?' matical theory of uniform polyhedra. But for
'The mountain is there, isn't it?' There is the purposes of making the models I inspected
another question many people ask when they see the drawings, done by J. C. P. Miller and col-
these polyhedron models: 'What do you use lected together at the end of the monograph, to

IX

21:31:40 BST 2016.

CBO9780511569746.002
discover the facial planes from which I derived side makes you forget the toil of the road.
the parts. These facial planes are now being A special word of thanks is extended to Mr R.
presented in this book. A set of photographs was Buckley for his truly remarkable calculations on
also given in the monograph; these show wire the snub polyhedra and for his astoundingly
models made by M. S. Longuet-Higgins, but detailed drawings of their facial planes. Without
they sometimes represent more than one poly- his help the book could never have been com-
hedron, so they are not the same thing at all as pleted. Also a word of thanks to Dr H. Martyn
the models presented here. Cundy for his deep interest in the book at all
My working time on the non-convex uniform stages of its preparation, and to H. S. M. Coxeter,
polyhedron models varied greatly. The simpler J. C. P. Miller and M. Longuet-Higgins, who did
ones took three or four hours each, the average the original research for Uniform polyhedra, and
would be near eight or ten hours each, a few of who have provided the source of inspiration from
the complex models took twenty or thirty hours which this book springs. Each in turn provided
work. Two of the non-convex snubs required further encouragement and help to complete the
more than one hundred hours work each. Now task. Thanks are also due to Stanley Toogood for
that the work is complete, I must admit I myself the photography, to the Syndics of Cambridge
am amazed. But the Chinese proverb applies: If University Press for accepting the book for
you want to make a journey of a thousand miles, publication, and to the editorial staff of the
you begin by taking the first step. One step leads Press who in an admirable way met the challenge
to the next, and soon the beauty of the country- of producing it.

21:31:40 BST 2016.

CBO9780511569746.002
Foreword

Interest in polyhedra runs through the whole himself had been lost, presumably in the great
gamut of intellectual activity from the two-year- fire of Alexandria, which was so poignantly
old child who plays with wooden cubes to the dramatized by Bernard Shaw in Caesar and Cleo-
mature mathematician who studies the subtle- patra.) Kepler also proposed the problem of
ties of Branko Griinbaum's Convex poly topes enumerating the isozonohedra (convex polyhedra
(Wiley, New York, 1967). Some of the regular whose faces are congruent rhombi), and partially
and semi-regular solids occur in nature as solved it by discovering the (first) rhombic dode-
crystals, others as viruses (revealed by the elec- cahedron and the triacontahedron. But his most
tron microscope). Bees made hexagonal honey- important contribution to the ideas of the
combs long before man existed, and in human present book was his proposal to consider non-
history the making of flat-faced solids (such as convex polyhedra whose faces are stellated poly-
pyramids) is as ancient as any other kind of gons such as the pentagram (fig. 21). He was
sculpture. The five regular solids were studied by probably unaware of the earlier work on stel-
Theaetetus, Plato, Euclid, Hypsicles, and Pappus. lated polygons by Thomas Bradwardine (1290-
A considerable portion of the present book is 1349), who became Archbishop of Canterbury
devoted to 'uniform' polyhedra, which have the for the last month of his life.
same arrangement of regular polygons at every Salisbury Cathedral is such a magnificent
corner. (Such a polyhedron is 'regular' if the building, full of interesting relics, that many
polygons are all alike.) In any convex solid, a visitors fail to notice the tomb of Thomas
theorem of Euclid tells us that the angles at a Gorges, who died in 1610. The stone-carved
corner must add up to less than 360°. After decorations include a dodecahedron, three icosa-
making a few models for himself, the reader will hedra, and two cuboctahedra, all' skeletal' in the
soon discover that the amount by which the style of Leonardo da Vinci (1452-1519) who had
angle-sum falls short of 360° is quite consider- made skeletal models of many uniform poly-
able when there are few corners (e.g. 90° for the hedra using rods to represent the edges. A few
cube, which has eight corners) but much smaller miles to the south-west is the pleasant village of
when there are many (e.g. 12° for the snub Wimbourne St Giles, where Antony Ashley was
dodecahedron, which has sixty corners). This buried in 1627. His tomb is embellished with a
observation was fashioned into a theorem by truncated icosahedron, not skeletal but a solid
Rene Descartes (1596-1650), who proved that looking just like the author's model 9.
the angular defect, added up for all the corners, Since the time of Descartes, many other great
always makes a total of 720°. mathematicians have contributed to this subject.
At about the same time, Johann Kepler (1571- Euler discovered and proved the famous
1630) wrote an essay on The six-cornered snow- formula V-E + F=2
flake (English edition, Oxford, 1966), in which
he revealed his fondness for these figures by which connects the numbers of vertices, edges,
remarking (p. 37): 'Now among the regular and faces of any convex polyhedron. Gauss used
solids, the first, the firstborn and father of all the an irregular spherical pentagram (his penta-
rest, is the cube, and his wife, so to speak, is the gramma mirificum) to explain Napier's rules in
octahedron, which has as many corners as the spherical trigonometry. Cauchy proved that
cube has faces.' It was Kepler who first pub- every convex polyhedron with rigid faces and
lished a complete list of the thirteen Archi- hinged edges is rigid. Hamilton invented the
medean solids, giving them the names by which Icosian Game (W. W. Rouse Ball, Mathematical
they are still known. (The work of Archimedes recreations and essays, Macmillan, 1967, p. 262).

XI

21:31:55 BST 2016.

CBO9780511569746.003
Von Staudt gave a new proof for Euler's for- In his infectiously enthusiastic style, the
mula. Schlafli extended the theory to n dimen- author gives clear instructions for making
sions. Klein wrote a highly influential book models of many kinds of polyhedra. These in-
called Lectures on the icosahedron. Fedorov structions are illustrated by photographs of his
returned to Kepler's problem of isozonohedra, own collection, including what is almost certain-
discovering a strangely oblate-looking rhombic ly the only complete set ever made of the known
icosahedron; and Bilinski (as recently as 1960) uniform polyhedra. But photographs cannot
completed the enumeration by discovering a really show the models in their full splendour.
second rhombic dodecahedron which would fit The most complicated 'snub' solids are not only
snugly into a box of unit height, breadth r and extremely difficult to make but also highly decor-
length T2, where r is the number (y/5 + l)/2 which ative: a perfect instance of the connection
belongs to the celebrated ' divine proportion' or between truth and beauty.
'golden section'. H.S.M.C.

XII

21:31:55 BST 2016.

CBO9780511569746.003
Introduction: uniform polyhedra

If you are being introduced to this topic for the Once you begin to make the models described in
first time, your first question might well be this book, you will quickly learn to draw all
'What is a polyhedron?' You may recall that these figures accurately and will become ac-
geometry itself is sometimes (not too exactly) quainted with important properties belonging to
defined as the study of space or of figures in each, especially the number of degrees in each
space—two dimensional for plane geometry and interior angle. Not all the regular polygons are to
three for solid geometry. The idea of sets is per- be found in the regular solids; in fact only three
haps familiar also. If you use the language of are used. The hexahedron (six faces), commonly
sets, a plane figure may be defined as a set of line called the cube, has squares; the octahedron
segments enclosing a portion of two-dimensional (eight faces) again has equilateral triangles; the
space. Such a plane figure is called a polygon. A dodecahedron (twelve faces) has all pentagons;
polyhedron is then defined as a set of plane and finally the icosahedron (twenty faces) has
figures enclosing a portion of three-dimensional twenty equilateral triangles. Euclid's Elements
space. closes with a proof that there are only five
All the terms used in this subject are derived regular polyhedra.
from classical Greek. Plato, the famous Greek A little experimenting with cardboard figures
philosopher, left the imprint of this thought will soon lead you to see the reasoning behind a
deeply fixed in Euclid's Elements. This ancient formal proof. Just as in a polygon two sides
book, for centuries the only textbook of geo- meet at a point called a vertex of the figure, so in
metry, was concerned with 'ideal' lines and a polyhedron two faces meet at or on a line (or
'ideal' figures. The ideal lines are straight and in a line—the mode of expression is variable).
the ideal polygons are regular, that is, they have Thus each face shares one of its sides as a line in
all sides and all angles equal. The simplest common with another face. These lines are
regular polygon is the equilateral triangle. It is called the edges of the polyhedron. So each edge
the simplest because it has the least number of of a polyhedron belongs to exactly two faces and
line segments possible to enclose a portion of no more. The edges all meet at a point called a
two-dimensional space. It is an interesting fact vertex of the polyhedron.
that Euclid's Elements begins with a proposition In the tetrahedron three edges meet at each
describing how to construct an equilateral tri- vertex, or to put it another way, each vertex is
angle and ends with a study of the five regular surrounded by three triangles. It is enlightening
solids. Each of these has regular polygons of the to lay out these three triangles flat and to notice
same kind for all its faces. They are known today the sum total of the number of degrees in the
as the five Platonic solids. The tetrahedron, angles at a common vertex. Three sixties give 180
which has four equilateral triangles for its faces, degrees. If a fourth triangle is introduced the
is the three-dimensional analogue of the two- total is 240 degrees, but now you have a vertex of
dimensional equilateral triangle. It is the sim- the octahedron. Introducing a fifth triangle gives
plest polyhedron, since it has the least number of 300 degrees, and you have a vertex of the icosa-
faces possible to enclose a portion of three- hedron. A sixth triangle gives 360 degrees and
dimensional space. you can see immediately that no polyhedral
With the equilateral triangle the following vertex arises. Everything stays flat.
polygons enter the picture: the square (four Next you can try squares. A minimum of three
sides), the pentagon (five sides), the hexagon (six is required, three 90s give 270 degrees, and a
sides), the octagon (eight sides) and the decagon vertex of the cube can be formed. Adding a
(ten sides), all of course only as regular polygons. fourth square brings the total to 360 degrees and

21:32:07 BST 2016.

CBO9780511569746.004
again you are left—flat. With pentagons the a fuller treatment of these two later on in this
minimum of three will give you a vertex of the book (see pp. 25 and 26).
dodecahedron; four are too many, the total Then there are the two called the rhombi-
going beyond 360 degrees. With hexagons the cuboctahedron and the rhombicosidodecahe-
minimum of three is already too many, three dron. These two are sometimes named the small
times 120 degrees. So no regular polyhedron rhombicuboctahedron and the small rhombi-
exists with only hexagons for faces. And simi- cosidodecahedron to distinguish them from two
larly for polygons with any greater number of others called the great rhombicuboctahedron
sides. In this way you can see that only five and the great rhombicosidodecahedron. If trun-
regular solids are possible. cation is applied to the two quasi-regular solids,
There is another set of solids known as the the cuboctahedron and the icosidodecahedron,
Archimedean or semi-regular solids. These all the new faces that arise are at best rectangles and
have regular polygons as faces and all vertices thus do not come out as regular polygons. But
equal but admit a variety of such polygons in one with some modifications these rectangles can be
solid. There are thirteen such solids and they are turned into squares. Because of this some
ascribed to Archimedes because he first enumer- authors refer to the great rhombicuboctahedron
ated them, although his work on them has been and the great rhombicosidodecahedron as the
lost. References to his work on this subject are truncated cuboctahedron and the truncated
found in the writings of Pappus, a mathematician icosidodecahedron. In this book they are named
of the third century A.D. Kepler was the first of the rhombitruncated cuboctahedron and the
the moderns to formulate a complete theory rhombitruncated icosidodecahedron. The prefix
concerning them. rhombi- implies extra square faces (across edges)
The Archimedeans can be broken down into of the two quasi-regular solids. With this the
various subsets. There are first of all the five designation small may be dropped from the
derived by the process of truncation from the five names of the former two.
Platonic solids. To truncate literally means to cut Finally there are two snub versions, one of the
off. Truncation thus implies the removal of some cube and one of the dodecahedron. This snub
portion of a solid, actually the removal of a quality gives these a twisted appearance which
portion near each vertex along with the vertex makes each of them turn out in either of two
itself. This can be done to the Platonic solids in forms—with a right- or left-handed twist. These
such a way that the new faces are again regular mirror image pairs are also called enantio-
polygons while the portions of the former faces morphous pairs.
that are left also form new regular polygons. For If you are ambitious enough and systematic
example the tetrahedron can be truncated so that enough you can also prove to your own satis-
the four triangles become four hexagons and the faction that the total enumeration of thirteen is
new faces are four new triangles. Five Archi- complete, that there are no more, by using the
medeans are thus generated. They are named same approach here that you used for the five
simply: the truncated tetrahedron, the truncated Platonic solids. The appropriate theorem from
hexahedron (cube), the truncated octahedron, solid geometry that applies here states that the
the truncated dodecahedron, and the truncated sum total of the face angles of any convex poly-
icosahedron. Another subset, containing only hedral angle is less than 360 degrees. After you
two members, is that known as the quasi- have tried all possible combinations of regular
regular polyhedra. The designation 'quasi-' polygons for which this theorem remains true
implies that the solid has only two kinds of you will come up with exactly the thirteen
faces, each face of one kind entirely surrounded Archimedean solids, and two infinite families,
by that of the other kind. They are the cubocta- the prism (with square side-faces) and anti-
hedron and the icosidodecahedron. You will find prisms (with equilateral triangular side-faces).

21:32:07 BST 2016.

CBO9780511569746.004
(For further details see L. Lines, Solid geometry, vertex is: a triangle, a square, a pentagon, and
pp. 159-67.) another square. The same holds true at every
The union of these two sets, the Platonic and vertex.
the Archimedean solids, together with the two The word 'enantiomorphous' occurs fre-
infinite sets of prisms and antiprisms, yield the quently in the following pages. It simply means
set known as the convex uniform polyhedra. A the property of being right- or left-handed, as in a
polygon is convex if no interior angle is greater pair of gloves, or in mirror image pairs. In colour
than 180 degrees. Analogously a polyhedron is arrangements, if the order around a vertex of a
convex if no dihedral angle (formed by the inter- polyhedron is taken in clockwise rotation, the
section of two faces with its vertex on or in an enantiomorphous arrangement will be obtained
edge) is greater than 180 degrees. Convex is the by taking the same order of colours in counter-
opposite of concave, bending in on itself. A clockwise rotation.
polyhedron with dimples, dents or grooves in it The following abbreviations will be used for
would be non-convex or concave. The word colours: Y yellow, B blue, O orange, R red,
'uniform' implies that all faces are regular poly- G green, W white, T tan.
gons and all the vertices of the polyhedron are All this terminology and all these classifica-
alike. In a uniform polyhedron the polygons tions will undoubtedly become clearer to you
around any vertex occur in the same order in and more meaningful after you have made the
every other vertex. For example in the rhombi- models in section i, the convex uniform polyhedra.
cosidodecahedron the order going around a

21:32:07 BST 2016.

CBO9780511569746.004
Mathematical classification

This section may be omitted at a first reading.

A uniform polyhedron can be enclosed within a Mathematical recreations and essays, Regular
sphere, so that its axes of symmetry pass through poly topes, and Phil. Trans. (1954), 246A, 401.
the centre of the sphere. By central projection These ideas are easier to visualize with the aid
the edges of the polyhedron can then be made to of models. You can make your own polyhedral
generate a network of arcs of great circles de- kaleidoscope with three mirrors cut in the shape
composing the surface of the sphere into spheri- of circular sectors. The radius must be fairly
cal polygons, one for each face of the polyhedron. large, twelve inches or more; the central angles
The planes of symmetry of the solid will like- of these sectors must be as follows:
wise decompose the surface of the sphere into a for the tetrahedral kaleidoscope 54° 44', 54°
network of spherical triangles, four for each 44', 70° 32';
edge of the solid if it is a Platonic solid. These for the octahedral 35° 16', 45°, 54° 44';
spherical triangles are called Mobius triangles, for the icosahedral 20° 54', 31° 43', 37° 23'.
because it was Mobius (1849) who first ob- Interesting as it is to play with these mirrors, they
served this. He illustrated this fact by means are not always so easy to come by, nor are they
of his polyhedral kaleidoscope, consisting of completely satisfactory. So it is just as good or
three mirrors forming a trihedral angle. Given better to make models of the spherical triangles
certain dihedral angles between these mirrors using the same index card or coloured tag you
and given an object to mark a point, the use in the other models. By repeating these
images of the object (together with the object spherical triangles a sufficient number of times
itself) mark the vertices of the polyhedron. you can make a model of the entire sphere as an
Another perhaps easier illustration of Mobius intersecting set of great circles. In fact the
triangles is simply a special set of great circles colours can be worked out so that they illustrate
inscribed with chalk on a slated globe. Some the great circles but this calls for more work than
of the intersections of the great circles mark is needed in a model all of one colour.
the vertices of the polyhedron. This tessellated The tetrahedral case is the easiest to begin
network of spherical triangles covers the globe with. The parts are cut as shown in fig. 1.
once. All the triangles are congruent. Score these parts on the radial lines, then fold
In symbols one of these triangles can be des- them forming a model of a spherical triangle.
cribed by (pqr) where p, q, r are integers and the The cementing is done using only one tab as
shown. Twenty-four of these are needed and
angles of the triangle are - , - , - Here p, q, r can
they are simply cemented to each other, flat sur-
only be 2, 3, 4, or 5. But one or more of p, q, r face to flat surface, so that the tab joint dis-
may be rational; that is, certain fractions may be appears between the two surfaces. You may do
used as replacements for p, q, r, leading to the work in sections. One of these sections has
Schwarz triangles. It was Schwarz (1873) who six spherical triangles as shown in fig. 2. The
first listed the possibilities. It has been shown
angles are ^ , ^ , ^ . Four of these sections
that a set of Schwarz triangles covering the globe
more than once but still a finite number of times complete the model.
is equivalent to a set of Mobius triangles. Thus For the octahedral case you may follow the
Schwarz triangles may be classified as tetra- same procedure with another set of parts shown
hedral, octahedral or icosahedral, depending on in fig. 3. Forty-eight of these are needed, two
the Mobius triangles to which they are related. enantiomorphous sets of twenty-four in each.
(For further information see: Coxeter et al., You may make these any convenient size. In fact

21:32:18 BST 2016.

CBO9780511569746.005
 Fig. 2

Fig. 1

71
Fig. 3
11 4 X

_
-

Fig. 4
7T\
T v
/

Fig. 6
Fig. 5

21:32:18 BST 2016.

CBO9780511569746.005
 Fig. 7 Fig. 8 Fig. 9

you may make the circular band wider or nar- illustrating Mobius triangles. It amounts to
rower as you please, even leaving all the interior, making a polyhedron whose faces are plane
which of course will bring you right to the centre triangles with the same vertices as the spherical
of the sphere. The sections in this case begin to triangles. If the sides of a spherical triangle are
reveal the fact that the octahedron is the dual of p, q, r, namely p, q, r are the angles subtended at
the cube, since one of these sections may be the the centre of the sphere, the corresponding plane
eight spherical triangles arranged as shown in triangles have sides proportional to
fig. 4. The angles are =•, ^ , j . Six of these sin \p : sin \q : sin \r.
JL D T1

sections complete the model. The three cases are shown in figs. 7, 8 and 9
The icosahedral case calls for more work and of course they call for the same number
because of the greater number of parts, but the of parts respectively as the spherical triangles
procedure is still the same. It is well worth the to which they correspond. The models may also
effort it takes, because it will bring you a great be done by following the same sectional pro-
deal of enlightenment. The openness of the cedure as for the spherical cases. The numbers are
model on the interior has great advantages. One approximate measures in linear units of the sides
hundred and twenty of these parts are needed, of these triangles. If you use centimetres you can
two enantiomorphous sets of sixty in each. The get satisfactory results.
sections are pentagonal, ten spherical triangles You can also get some striking colour effects
by making one set of triangles all W and then
to a section. The angles are ^ , ^ , - . Twelve of
the others in the usual colours. The drawings
these sections (see fig. 6) complete the model. below show the respective sections and their
There is still another way to make models colour tables.

\^ w
4 ^V y
Fig. 10 Fig. 11 Fig. 12

1 2 3 1 2 3 4 (Seep. 18 for the

(1) Y B O (0) Y B O R icosahedral table.)
(2) Y R B (i) O R B Y
(3) B R O (2) R Y 0 B
(4) O R Y (The other three are
enantiomorphic to these.)

21:32:18 BST 2016.

CBO9780511569746.005
A. Tetrahedral

B. Octahedral

C. Icosahedral

21:32:18 BST 2016.

CBO9780511569746.005
 Note that the tetrahedral case has the sections derived, but there is an interchange in the
numbered (1), (2), (3), (4). The tetrahedron and number of faces and vertices. Moreover the
the truncated tetrahedron are the only uniform kinds of faces and vertices are such that an
polyhedra that do not have their vertices distri- H-sided face in the original solid yields an
buted in diametrically opposite pairs. The (0) «-edged vertex in the dual solid. The three poly-
section in the other two cases may be taken as hedrons just described thus turn out to be duals
the north polar section. Then in the octahedral of 7,15 and 16, respectively.
case (1) and (2) are cemented in place something Once you have made these models it is a good
like the faces of a cube. These are followed by exercise in spatial imagination to use them, es-
the enantiomorphous arrangement of the same pecially the spherical models, to locate the faces,
two sections, thus completing the four side-faces edges and vertices of the convex uniform poly-
of the cube. The enantiomorph of (0) then com- hedra. The diagrams here may serve as a guide.
pletes the model. The icosahedral case is the The numbers designate the vertices whose
most interesting. You may begin with the (0) images are the vertices of the polyhedron desig-
section, cementing together ten triangles alter- nated by the same number in the summary, p. 9.
nating W with one of each of the colours. Follow The snubs, 17 and 18, are not indicated below.
the icosahedral colour arrangement as shown on The vertices of these depend on a suitable point
p. 18. Then as you complete each of the other being chosen within the triangles. The exact

11
13

Fig. 14 Fig. 15

sections (1), (2), (3), (4), (5), cement it to the (0) mathematical details can be found in L. Lines,
section first and then to its neighbour in dodeca- Solid geometry, pp. 175-84. The construction
hedral fashion. The next set of six sections have there relates the snubs to the circumscribed cube
the enantiomorphous order of colours. They are and dodecahedron, showing how to find the
placed diametrically opposite their counterparts. vertices of the snubs on the faces of the cube and
You will be delighted with the pin-wheel ap- dodecahedron. Then by central projection these
pearance that turns up on all three of these cases. same points can be located in the spherical
It is most pronounced in the icosahedral case. triangles.
It is worth mentioning here that these three The summary that follows is an attempt to
models are actually Archimedean duals. Dual bring together the various aspects relating prin-
solids are those which have the same number of cipally to the symbols used for each polyhedron.
edges as the original solids from which they are You need not master this material to make the

21:32:18 BST 2016.

CBO9780511569746.005
 Summary of convex uniform polyhedra together with their symbols

Regular polygons:
triangle square pentagon
{3} {4} {5}
hexagon octagon decagon
{6} {8} {10}

Uniform polyhedra:
Platonic solids (regular solids)
1. tetrahedron {3, 3} = 3 |2 3 = 33
2. octahedron {3, 4} = 4 |2 3 = 34
3. hexahedron (cube) {4, 3} = 3 |2 4 = 43
4. icosahedron {3, 5} = 5 |2 3 = 35
5. dodecahedron {5, 3} = 3 |2 5 = 53

Archimedean solids (semi-regular solids)

6. truncated tetrahedron t {3, 3} =23 3 = 3.62
7. truncated octahedron t {3, 4} =24 3 = 4.62
8. truncated hexahedron t {4, 3} =23 4 = 3.82
2
9. truncated icosahedron t {3, 5} = 2 5 3 = 5.6
10. truncated dodecahedron t{5, 3} = 2 3 5 = 3.102
11. cuboctahedron (quasi-regular) {1} = 2 | 3 4 = ( 3 . 4 ) 2
12. icosidodecahedron (quasi-regular) {!} = 2 | 3 5 = (3.5)2
13. (small) rhombicuboctahedron r{l} = 3 4 | 2 = 3.43
14. (small) rhombicosidodecahedron r{35} = 3 5 | 2 = 3 . 4 . 5 . 4
15. rhombitruncated cuboctahedron t{J} = 2 3 4 | = 4.6.8
16. rhombitruncated icosidodecahedron t{35} = 2 3 5 | = 4.6.10
17. snub cube s{2} = | 2 3 4 = 34, 4
18. snub dodecahedron s(|) = | 2 3 5 = 34, 5

models in this book, but it is interesting to know names the polygon that appears in the vertex
that the details have been worked out. If you figure. For an explanation of the dashes, see the
should ever want to undertake further investiga-
following page. J J simply names the two kinds
tion in this field you would have to be thoroughly
acquainted with the details. of polygons found in the faces of the quasi-
The Schlafli symbol is given first, then the regular solids. It is an extension of the Schlafli
symbol with dashes, "|" a s u s e d in Uniform symbol. So too are t, r, s to mean truncated,
polyhedra, then another symbol as used in rhombic and snub respectively. Rhombic implies
Mathematical models. In the symbol {p, q}9 p the existence of extra square faces. Snub implies
names the polygon that appears in the faces, q the existence of extra triangular faces.

21:32:18 BST 2016.

CBO9780511569746.005
 The polyhedron whose vertices are the images
of D is pqr\. D is the incentre of the triangle
PQR.
pqr\ = prq\ = qrp\= etc.

E and F apply only to the non-convex uniform

Fig. 16
polyhedra. They are given here to complete the
summary. PE is the external bisector of the angle
The meaning of the dashes, NV "', may be briefly
at P. F is the excentre.
summarized as follows: A spherical triangle
Suppose the angle Q 'PR is -,; then - + — = 1. If
PQR whose angles are - , - , -may also be named
the angle PRQ is %n9 then the polyhedron
(pqr). In terms of the polyhedral kaleidoscope:
whose vertices are the images of E is 2q\p'.
The polyhedron whose vertices are the images
of P is p\q r (or p\rq).
p\q2=p\2q= {q,p} 2q\p' = t'{p,q}.
q\p2 = q\2p= {p,q} The polyhedron whose vertices are the images
-(jj. of F is 2 p'q\.

The polyhedron whose vertices are the images

of C is qr\p (or rq\p). C is the point of inter- The polyhedron whose vertices are the images
section of the bisector PC of the angle QPR of a suitable point within PQR is | pqr.
with the opposite side QR.

The symbols r' and t' stand for quasi-rhombic

2q\p = t and quasi-truncated respectively.

10
21:32:18 BST 2016.
CBO9780511569746.005
The Convex Uniform Polyhedra

The Platonic and Archimedean Solids

11
21:32:36 BST 2016.
ebook.jsf?bid=CBO9780511569746
General instructions for making models

The first thing you must do to make a model of through at each vertex, using the template for
any polyhedron is to make an accurate drawing this purpose as a guide. You may draw pencil
for the required parts. For the convex polyhedra lines around the edges of the template before
these are simply polygons of 3, 4, 5, 6, 8, and 10 you move it to its next position on the sheets of
sides. But you must remember that in any one card. In this way you can prick as many patterns
polyhedron all the edges must be the same length. as you need or wish to do at one time. Your next
Hence the polygons belonging to one poly- step is to cut with scissors along the pencil lines
hedron must have sides of the same length. As left by outlining the template while the sheets of
you can see from a drawing, the decagon, for card are still held together by the staples. The
example, is very large compared to a triangle sheets may have a tendency to move slightly
with the same edge length. You must keep this in while you are cutting, but this is not too serious
mind when making the models and choose a because the quarter inch border all around gives
suitable scale. This will be determined by how you a little leeway. But once the pieces have been
you want to use the polyhedron and where you cut in the way just described you will get best
intend to display it. In the following descriptions results from here on by handling each part in-
of the individual models a value is given for the dividually. With a sharp pointed instrument such
circumradius, that is, the radius of a circum- as the point on geometry compasses, you must
scribing sphere, in terms of a polyhedral edge now score the card using a set square or ruler for
length of 2 units. This will help you to determine a straight edge. If you plan to do a lot of model
how big the completed model will turn out to be. making it will pay you to take the filler from a
Doubling the radius gives you the diameter and ball-point pen and replace it with the steel point
this can be taken as an approximate value for from the compasses. A mimeograph stylus also
the height of the completed model. serves the purpose very well. You must draw the
Once you have carefully drawn the parts, scoring lines to connect the needle holes. Pencil
namely the required polygons, it is best to make lines are not needed, since the process of scoring
a template. This is done by placing the drawing sufficiently outlines the shape of each part. More
of the polygon over a piece of card or stiff paper. accurate trimming with scissors must now be
Index card stock or coloured tag is recom- done. As mentioned before you will get best
mended. Then prick through at each vertex with results by handling each part separately. Cut
a probing needle. The kind found in a biology directly through or into the needle holes and out
dissecting kit serves the purpose very well. You again so that the quarter inch border is left as a
may then draw pencil lines from hole to hole and tab. Then fold the tabs down. The scored lines
trim the card with scissors leaving about a make this a simple and accurate operation.
quarter inch border all around outside the These tabs will be used for cementing the parts
pencil lines. This is your template. together. Where the parts have more acute
It is now a simple and easy matter to multiply angles, you must trim the tabs again after the
copies of the parts any number of times. This is folding. If done before, you will find the folding
done by placing the template on top of a number more difficult. Experience will teach you how
of sheets of card. It is best to staple the sheets much trimming to do and how accurate it must
together. Usually four, five, or six parts are your be for best results. The rule is to leave as much as
requirements at any one time and this will then possible for the cement to hold and to remove
be the number of sheets of card, say one of each as much as necessary so no jamming occurs at
colour needed, that you may staple together. the vertices on the interior of the model.
Now again using the probing needle prick A good household acetone cement provides

12
21:33:00 BST 2016.
CBO9780511569746.006
the best adhesive since it is quick drying and then the circumradius for the edge length of two
adheres very firmly. The procedure is to apply units and finally the vertex figure. The symbol is
the cement all along one tab, then to join the tab not too important for the purposes of making the
from another part to it, to move these parts back models. It belongs more to the mathematical
and forth slightly to help spread the cement analysis and classification of these geometrical
evenly on both parts, then to manoeuvre the solids (see p. 9). The number of faces and the
parts into accurate positions before the cement kind of polygons appearing as faces are given by
becomes too stiff. You will find a pair of tweezers 4{3} + 4{6}; that is, the polyhedron has four
helpful at times, especially as the work pro- triangles and four hexagons for its faces.
gresses and the model begins to take shape. This information will help you to prepare the
Clamps are also helpful and even necessary on right number and the right kind of polygons in
more intricate models. You can make your own each case. In the instructions given for making
clamps by taking clothes pegs of the coiled wire individual models the word net is frequently
spring variety and turning them inside out, used. In its context it simply means any part or
namely separating the parts to get the two parts needed in the construction of a poly-
wooden prongs reversed and then replacing the hedron. Thus the word, as it is used in this book,
spring between them. will mean the drawing of the part or parts
You will find that the method of assembly for needed for a template. The vertex figure is also
polyhedron models suggested here will generally given because it is very helpful in giving you
give you fairly rigid results, since the tabs serve information about the order in which the poly-
as interior structural ribs along all the edges of gons surround each vertex of the polyhedron.
the model. For this reason it is best to follow the You may think of the vertex figure as the base of
general rule of leaving tabs all around on every a pyramid, all of whose slant edges are of unit
part. It is only occasionally and in fact only length. Or to put it another way, you may choose
rarely that you may have to depart from this any vertex of a polyhedron and take note of the
procedure. This occurs only in the more complex edges meeting at that vertex. Then the points on
models described later on in this book. For all these edges each a unit length from the vertex of
the convex polyhedra it is best to leave all the the polyhedron will be the vertices of the vertex
tabs. figure. Every uniform polyhedron is character-
The convex uniform polyhedra are presented ized by its vertex figure which is a cyclic polygon
first. They are the easiest to make and you will (cf.Coxeter, 1954, p. 404).
find it best to begin with them. In each case the For colour arrangements the map colouring
symbol designating the polyhedron is set down, principle generally gives the most striking effects;
followed by the kind and number of polygons namely, polygons sharing a common side (meet-
comprising the facial planes of each polyhedron, ing at an edge) must be of different colours.

13
21:33:00 BST 2016.
CBO9780511569746.006
1 The tetrahedron
The simplest of all polyhedra is the tetrahedron.
It has four equilateral triangles for faces. This is
the least number possible to enclose a portion of
three-dimensional space. Certain properties im-
mediately appear in it which are characteristic of
the entire set of uniform polyhedra. All its faces
are regular and each face shares its edges with
just one other face. Also all its vertices are alike.
A model of the tetrahedron can be made by
using one net for the entire solid as shown. How-
ever by doing it this way, you will have all faces
the same colour. So too all the convex polyhedra
can be made of one net and thus of one colour.
(See Cundy & Rollett, Mathematical models.)
But if you want each face of the tetrahedron, and
more generally each face of any polyhedron, to 3|2 3 = {3,3}
be a different colour, then you should prepare 4{3}
individual nets for each face that is a different Vi
polygon. For the tetrahedron all you need is one
net, an equilateral triangle. Prepare four parts,
each one with tabs all around as shown and each
of a different colour, say Y, B, O, R. Then
cement these four triangles together in the same
position as shown. Now bring the remaining tabs
together, cementing one pair first and letting
this set firmly. Then apply cement to both re-
maining tabs and close the triangle down as you
would close the lid on a box. The model exerts
its own pressure and your fingers can do the rest
along the edges until the cement is set.

14
21:33:16 BST 2016.
CBO9780511569746.007
 The octahedron

This is the polyhedron, whose faces are com-

posed of eight equilateral triangles. Here
opposite faces lie on parallel planes, so four
colours serve very well. You may begin the
construction of a model of this polyhedron by
cementing four triangles as shown. When the
remaining tabs between triangles 1 and 4 are
cemented, you will have a square pyramid with
triangular slant side faces but without the square
base. The other tabs remain at the edges of this
base. This section completes half the model.
The other half is enantiomorphous to the first
half. Actually it is simpler to continue your work
by cementing four triangles, one at a time, to the
four tabs around the edges of the open square
4|2 3 = {3,4} base. It is easy to watch the opposite faces to get
8 {3}
the right colours. Then the tabs between adjacent
V2 triangles may be cemented and the last triangle
again closed down like a lid. You may now ob-
serve that the square which showed its edges on
the completion of the first section is actually only
one of three such squares in the completed
model. The three squares have edges on three
mutually perpendicular equatorial planes. This
fact is utilized in one of the non-convex uniform
polyhedra 67 to be described later.

12 3 4
Y B O R

15
21:33:16 BST 2016.
CBO9780511569746.007
 The hexahedron (cube)

The most commonly known and most widely

used of all polyhedra is undoubtedly the cube, or
to give it a fancier name, the hexahedron. Its six
faces are all squares meeting two at each edge
and three at each vertex. Since opposite faces are
parallel, a simple colour arrangement is possible
using three colours. You may begin the con-
struction of a model for this polyhedron by
choosing one square and then cementing four
others around it as shown. Then you may
cement the tabs between adjacent squares to
form the four vertical edges of the cube, again
with all tabs forming ribs on the interior. Finally
you may add the last square, and in this case it
really does answer the description which calls
it the lid on a box. 3|2 4 = {4,3}
6 {4}
V3

1 2 3 4 5 6
Y B O B O Y

The cube may not be a very exciting poly-

hedron in its own right, but it has some wonder-
ful properties in relation to the other Platonic
solids, as well as with some of the Archimedeans.
A compound of five cubes can be enclosed in a
dodecahedron and this makes a beautiful
model. (See Cundy & Rollett, Mathematical
models, pp. 135-6.)

16
21:33:16 BST 2016.
CBO9780511569746.007
 The icosahedron

The icosahedron is one of the five Platonic

solids, next in simplicity to the tetrahedron and
octahedron. It shares with these the fact that all
its faces are equilateral triangles. In making a
model of this polyhedron there are two effective
ways of arranging five colours. They can be
arranged so that each of the five appears around
each vertex but then opposite faces will not be of
the same colour. The second arrangement has
opposite faces the same but one colour repeats
itself around each vertex except the two polar
vertices. Both of these arrangements are very
useful, because many of the uniform polyhedra
to be described later in this book have icosa-
hedral symmetry. So you will find it profitable to
5|2 3 = {3,5} have two models of the icosahedron illustrating
20 {3} both colour arrangements for future reference.
5i r i You may begin both with the same initial ar-
rangement of five equilateral triangles as
shown. These form a low pentagonal pyramid
without a base. The next set of five triangles may
then be cemented to the pentagonal edge as set
out in the colour table. Between these the third
set of five easily find their position because only
one more triangle is needed to complete the ring
of five at each vertex. One more set of five then
completes the model.

First colour arrange- Second colour arrange-

ment ment
1 2 3 4 5 1 2 3 4 5
Y B O R G Y B O R G
R G Y B O B O R G Y
O R G Y B G Y B O R
Y B O R G R G Y B O

A few comments will be helpful to you in the

use of the colour tables listed above. The first
line may be thought of as the set of five triangles
surrounding the north polar vertex of the icosa-
hedron. The next two lines actually form an
equatorial band of ten triangles, alternating with
each other. The fourth line is the set of five
surrounding the south polar vertex. You can

17
21:33:16 BST 2016.
CBO9780511569746.007
also read the order of colours surrounding the cyclic permutation of colours is easier to follow.
other ten vertices by following the rotation of The alternative colour tables are:
colours, starting with two adjacent colours in the
first row, then proceeding into the second row First colour arrange- Second colour arrange
and down to the third, then back to the second ment ment
and finally ending in the first where you started. (0) Y B O R G (0) Y B O R G
For example (1) B Y R O G (1) Y B G O B
(2) O B G R Y (2) B O Y R O
Y -> B B (3) R O Y G B (3) O R B G R
t i f (4) G R B Y O (4) R G O Y G
R G or R G (5) Y G O B R (5) G Y R B Y

Only six vertices are listed, the (0) vertex being

This suggests an alternative way of setting out the north polar vertex, but in both cases the dia-
the colour tables, a way which will be very useful metrically opposite vertices have the enantio-
for later models. This is done by listing the morphous arrangement. You can get this by
colours surrounding each vertex and numbering reading the colour table in reverse, that is, from
the vertices. In doing this each triangle of the right to left. A little experiment with these ideas
icosahedron gets named three times, but the will soon make it clear to you.

18
21:33:16 BST 2016.
CBO9780511569746.007
 The dodecahedron

The dodecahedron is in some ways the most

attractive of the five Platonic solids, although
the icosahedron is a very close second, if second
at all. The relationship of the dodecahedron to
its three stellated forms, to be described later,
probably gives it the advantage for first place.
A model of this polyhedron can be done in
four colours in two different ways. With six
colours opposite faces can be the same colour,
and this arrangement carries over very well into
the stellated forms mentioned above. So this
arrangement is described here.
You may begin by cementing five pentagons,
one of each of five colours, Y, B, O, R, G
around a central pentagon which may be white,
3|2 5 = {5,3}
W. Once the tabs between these five coloured
12{5}
pentagons are cemented, half the model is com-
pleted. This half-completed dodecahedron will be
mentioned again later in connection with stel-
lated forms of the icosahedron, where it can
serve as a construction cradle for other models.
In such a case the tabs would be turned out-
ward, but here you have them on the interior as
usual. The other faces of the dodecahedron now
easily take their positions so that opposite faces
are paired according to colour.
The arrangement of four colours for the
dodecahedron is shown here. It can also be done
enantiomorphously. This four colour arrange-
ment is sometimes more suitable, especially for
other models having dodecahedral symmetry.
Therefore it is given here for future reference.

19
21:33:16 BST 2016.
CBO9780511569746.007
6 The truncated tetrahedron

For this polyhedron, you may obtain a very

effective colour arrangement by using the same
four colours for the hexagons that you used for
the triangles in the tetrahedron. Then the tri-
angles here may all be of a fifth colour. Or since
these triangles are on planes opposite and
parallel to the hexagon planes you may make
these parallel planes the same colour. This
arrangement is shown here. If you cement the
parts in the order shown you will get them in
their proper places. Then the remaining tabs are
cemented, a pair at a time, as described before

12 3 4 5 6 7 8
Y B O R O R Y B
3|3 = t{3,3}
4 {3} +4(6}

for the tetrahedron. Another way to make the

model is simply to make a tetrahedral cup to
begin with as shown below. This has a triangular
bottom and three hexagons for sides. The tabs
form ribs on the inside of the cup. Then you
continue by cementing triangles and hexagons as
needed. You will find it best to use a triangle as
the last part and to cement one of the tabs of this
triangle, first letting it set up firmly, and then
closing the hole as you would close the lid on a
box. This is a general procedure to follow for all
models.

20
21:33:29 BST 2016.
CBO9780511569746.008
 The truncated octahedron

The way to construct a model of this polyhedron

is by now becoming familiar to you, if you have
done the previous ones. The colour arrangement
of the hexagons follows that of the triangular
faces of the octahedron, namely four pairs of
opposite faces using four different colours and
all the squares using a fifth colour. So you may
begin here by surrounding a hexagon alternately
with squares and other hexagons as shown. Then
the tabs between these are cemented to form a
cup, completing half the model.
Once this is done it is easy to continue cementing
the other parts and watching the opposite hexa-
gons to get the colours right. A square is added
last of all. You will undoubtedly observe that
= t{3,4} complete rigidity is not achieved until the last
8{6}+6{4} edges have been cemented. But once this has
been done, all the convex polyhedra make very
rigid models.

1 2 3 4 5 6 7
Y G B G O G R

21
21:33:29 BST 2016.
CBO9780511569746.008
8 The truncated hexahedron (cube)
This polyhedron is a truncated cube, again not
very exciting as a model, but belonging never-
theless to the set of uniform polyhedra. The
colour arrangement for the octagons may follow
that of the squares in a cube, leaving a fourth
colour for all the triangles. You may again begin
your work on this model by surrounding an
octagon with triangles and other octagons as
shown. The tabs on the surrounding octagons
will then be cemented to each other, thus sur-
rounding the triangles. The tabs on the triangles
can be cemented as tabs on the lids of triangular
holes. This is not difficult as long as you can
work on the inside of the model while it is still
open and incomplete. The last octagon is Y and
finally four R triangles close the corners. You 3|4 = t{4,3}
8{3}+6{8}
will see that a little more skill is being called for
here to get accuracy in your work, but undoubt-
edly as you proceed with the work of making
polyhedron models you are developing this skill.

1 2 3 4 5 6 7 8 9
Y R B R O R B R O

22
21:33:29 BST 2016.
CBO9780511569746.008
 9 The truncated icosahedron
As a truncated form of the icosahedron, a model
of this polyhedron may well follow the five-
colour icosahedral arrangement for the hexagons
and a sixth colour for all pentagons. You should
have no difficulty in cementing the parts cor-
rectly if you follow the icosahedral colour table.
Thus you may begin with a W pentagon and
surround it with a set of five coloured hexagons
Y, B, O, R, G. Then if you keep your attention
on each ring of hexagons, adding the W penta-
gon each time at the centre of the ring, you can
easily complete the next set of five rings. Each
hexagon of course belongs to three rings. The
completed model is very attractive with its com-
bination of hexagon and pentagon faces.
2 5|3 = t{3,5}
20{6} + 12{5}

f /29 + V 5

23
21:33:29 BST 2016.
CBO9780511569746.008
10 The truncated dodecahedron
This is the polyhedron whose facial planes are
triangles and decagons. You may again use the
four-colour dodecahedral arrangement for the
decagons and make all the triangles a fifth
colour. Around a R decagon cement in order a
Y, B, G, B, G decagon alternating with O tri-
angles. The next set of decagons is Y, R, Y, B,
R; the first Y of this set adjoining the G that
comes between the two B decagons in the first
set. The rest of the O triangles are then cemented
in place.
This polyhedron does not have a particularly
pleasing shape, perhaps because the area of the
decagons is very large compared with that of the
triangles. For the same reason a model of this
polyhedron must be made with the decagons 2 31 5 = t{5, 3}
reinforced or stiffened on the inside, say with 20{3}+12(10}
double thickness card, otherwise these faces have /37 + 1 V 5
a tendency to sag. On the other hand if you keep 2
the model small, this reinforcing is not necessary.

10 10

24
21:33:29 BST 2016.
CBO9780511569746.008
 11 The cuboctahedron
The name of this polyhedron suggests a close
relationship to the cube and the octahedron, and
indeed this is so. The six squares are on the facial
planes of a cube and the eight triangles on the
facial planes of an octahedron. You may later
wish to make a compound of these two Platonic
solids and then you will observe that the cubocta-
hedron is the portion of space common to the
two.
To make a model of this polyhedron the three
colours used for the cube may serve here for the
squares and a fourth colour for all eight tri-
angles. You may begin with a triangle and
cement a square to each of its edges as shown.
Then three more triangles between these squares
2|3 4 = {|} will complete half the model, a kind of cup with
8{3}+6{4} a triangular bottom and squares and triangles
2 alternating for sides. Once this section is com-
pleted you may now easily continue and get the
arrangement of colours right by observing the
opposite squares for the correct colour.
An important property of this polyhedron is
the fact that it has two types of faces, each kind
entirely surrounded by that of the other. As such
it is designated quasi-regular.

1 2
Y B

25
21:33:29 BST 2016.
CBO9780511569746.008
12 The icosidodecahedron
This polyhedron is a combinatorial solid, quasi-
regular, in the same way as the cuboctahedron.
It is the interior part common to the compound
of an icosahedron and dodecahedron. If you
limit yourself to five colours, a suitable arrange-
ment can be worked out for a model of this
polyhedron by making all the triangles Y and
using the other four colours for the pentagons.
This follows the four-colour arrangement for
the dodecahedron.
Thus you may begin with a B pentagon and
cement five Y triangles to its tabs. Next five
more pentagons are cemented so each shares
two of its edges with two adjacent triangles al-
ready in place around the B pentagon. The
colours should be O, R, G, R, G. Another set of 2|3 5 = {?}
five triangles will then complete half the model, 20{3}+12{5}
leaving a ring of tabs in the form of an equatorial 2r
decagon. In doing the second half of the model
you may proceed by adding alternately triangles
and pentagons to the equatorial edges as just
described. Place an O pentagon so its vertex
coincides with the vertex of the G pentagon that
appears between the two R pentagons. The
order, repeating the O in naming the five colours,
is then: O, B, O, R, B. The last G pentagon is
added as soon as some of the last five triangles
are in place. The remaining triangles then com-
plete the model. You will now notice five other
sets of equatorial edges. This property is used
for some of the other non-convex uniform poly-
hedra.

26
21:33:29 BST 2016.
CBO9780511569746.008
 13 The rhombicuboctahedron
The name of this polyhedron again indicates its
nature. The set of squares forming its faces
break up into two subsets and thus the colour
arrangement for these may well do the same.
The triangles may then all be of another colour.
To make a model of this polyhedron you may
begin by making a section which forms a
shallow cup having an octahedral upper edge as
shown.
Next a square is cemented to each tab at the
octahedral upper edge of the cup. These squares
alternate in colour as set out below. In following
the map colouring principle you will easily see
that each R square must share an edge with a B
triangle, and each Y square must share an edge
with a R square. The rest of the model is then
8{3} + (12 + 6){4} easy to do, one part at a time, and continuing
with the alternating colour arrangement for the
squares. This turns out to be a rather attractive
model even though it is composed of only tri-
angles and squares.
It is worth mentioning that a pseudo-rhombi-
cuboctahedron can be formed by rotating an
octagonal 'cap' of the rhombicuboctahedron
through an angle of 45° relative to the rest of the
solid. There ,then arises a solid with all vertices
alike but not Archimedean. This is because the
cubic and rhombic squares get mixed up.

1 2 3 4 5 6 7 8 9
Y B R B R B R B R
R Y R Y R Y R Y

A A 9
7

1 5

V V 3

27
21:33:29 BST 2016.
CBO9780511569746.008
14 The rhombicosidodecahedron
This polyhedron is in some ways the most at-
tractive of the Archimedean solids. The simplest
and most suitable arrangement of colours for a
model of this polyhedron is obtained by making
each of the three different kinds of faces a differ-
ent colour, say all triangles Y, all squares B, and
all pentagons O. Then you may work around
each pentagon and complete the rings with alter-
nate triangles and squares in such a way that
adjacent rings share two triangles and a square
in common. You will find variations of this poly-
hedron turning up in the non-convex uniform
polyhedra described later in this book. However,
other arrangements of colour will be suggested
there. They could of course also be used here,
quite effectively. | {|}
20 {3}+ 30 {4}+ 12 {5}

28
21:33:29 BST 2016.
CBO9780511569746.008
 15 The rhombitruncated
cuboctahedron

This polyhedron, also known as a truncated

cuboctahedron, again lends itself to a simple
colour arrangement. Three colours will serve in
pairs for opposite octagonal faces, then all hexa-
gons share a fourth colour and all squares a fifth
colour. Thus to make a model of this polyhedron
you may begin as usual with a cup shaped
section as shown. Once this is completed four
more octagons are cemented in place as set out
in the colour table. You will then have no
further difficulty in completing the model. As
• these models now become more intricate they
also become more interesting and attractive.

2 3 4| = t{J}
8 {6}+ 12 {4}+ 6 {8}
VC13 + 6V2) 1 2 3 4 5 6 7 8 9
Y B O B O B O B O
R G R G

29
21:33:29 BST 2016.
CBO9780511569746.008
16 The rhombitruncated
icosidodecahedron

This polyhedron is also called the truncated

icosidodecahedron. Here too the simplest colour
arrangement turns out to be the best, that is,
three colours, say Y for the decagons, B for the
hexagons and O for the squares. You may also
use the same procedure for constructing this
model as you used for the last one, working
around each decagon to make a ring of alter-
nating hexagons and squares. Again each ring
shares two hexagons and a square in common
with adjacent rings. This polyhedron also has its
analogous forms among the non-convex uni-
form polyhedra to be described later. Again
where the decagon plane appears in this model
you must be sure that it is stiff enough so that it 2 3 5|= t{l}
will not sag. So here too if you keep the model 20{6} + 30{4} + 12{10}
small you will automatically get a better result.

10

30
21:33:29 BST 2016.
CBO9780511569746.008
 17 The snub cube
This polyhedron can be inscribed in a regular
cube in such a way that its six square faces will be
coplanar with those of the cube but will be in a
slightly twisted position (cf. L. Lines, Solid
geometry, p. 76). Each square is entirely sur-
rounded by triangles accounting for twenty four
of these. Then eight more triangles close the re-
maining spaces to complete the solid. This suggests
the following colour arrangement. The squares
may have three colours in opposite pairs. Each of
these square faces will be entirely surrounded by
triangles of the same colour. Thus the same three
colours will occur in the triangles but shifted to
maintain the map colouring principle. Finally
the other eight triangles share a fourth colour.
12 3 4 = s {2}
To make a model of this polyhedron you may
(8 + 24){3} + 6{4} follow the colour table set out below, showing
2-68742 67475 the arrangement for the first three sections. These
sections must now be cemented together using R
triangles as connectors between the other tri-
angles. These three sections complete half the
model. You can do the other half in the same way
provided you get the squares in colour pairs
opposite each other.

1 2 3 4 5 6
Y B B B B R
B O O O O R
O Y Y Y Y R

31
21:33:29 BST 2016.
CBO9780511569746.008
18 The snub dodecahedron
This polyhedron has the same relation to the
regular dodecahedron that the snub cube has to
the regular hexahedron. To get a suitable colour
arrangement in a model of this polyhedron you
may make all the pentagons Y. Then each of
these is to be totally surrounded by triangles,
each pentagon getting five triangles of the same
colour. The sets of triangles however may be
done in four colours. These parts are then as-
sembled in the four-colour dodecahedral ar-
rangement using Y triangles as connectors. The
colour table shown below will help you.
This is the last of the set of convex uniform poly-
hedra. The non-convex uniform polyhedra are
given after the set of stellations and compounds
which follows. |2 3 5 = s{|}
(20 + 60){3} + 12{5}
4-31167 47491

32
21:33:29 BST 2016.
CBO9780511569746.008
Some Stellations and Compounds

33
21:33:55 BST 2016.
ebook.jsf?bid=CBO9780511569746
Commentary on stellations and
compounds of the Platonic solids

The word 'stellation' comes from the Latin turns out to be the well-known five-pointed star,
word 'stella' which means 'star'. There are star also called the pentagram. See fig. 19. It was
polygons as well as star polyhedra. Exactly what known to the ancients, as is evident from the fact
this means is best understood by drawings and that it was used by the Pythagorean brotherhood
models rather than by abstract definitions. You as a symbol of health. Similarly a hexagon leads
can thus begin again with the simplest polygon, to a six-pointed star or hexagram (really not a
the equilateral triangle, and see what happens if single polygon but a compound of two equi-
you produce each line segment forming its sides. lateral triangles). An octagon leads to an eight-
You find that no new portions of two dimen- pointed star or octagram; a decagon to a ten-
sional space can be enclosed. See fig. 17. The pointed star or decagram. The pentagram, octa-
lines will forever get farther apart. The same gram, and decagram can still be considered as
thing happens if you try producing the sides of a single polygons of five, eight, and ten sides re-
square. The lines are in parallel pairs and will spectively, since you can trace out their sides in a
never meet to enclose any portion of the plane continuous movement going around the centre
other than the interior of the original square. of the figure twice in the case of the pentagram
See fig. 18. With the pentagon something more instead of once as in the pentagon. In the figures
interesting begins to happen. The sides of the follow the order of the numerals. In the octa-
pentagon when produced will meet and enclose gram and decagram a continuous movement
more space exterior to the original pentagon. It will take you three times around the centre. Note

Fig. 17 Fig. 18 Fig. 19

1
IS A
> /
v \ y
0 / A\ / \
Fig. 20
8

5
K 2

Fig. 22

34
21:34:34 BST 2016.
CBO9780511569746.009
that the internal points of intersection are dis- But now a closer examination will reveal to
regarded. These facts are expressed symbolically you that this polyhedron actually turns out to
in the fractions used for naming these star poly- be, not one polyhedron, but a compound of two
gons: f, f, -3-. These stars can assume other —two larger tetrahedra interpenetrating and
shapes as well, but only these are mentioned here sharing a common centre, the octahedron's
because only these appear again later in this centre of symmetry. Kepler discovered this solid
book. (See Coxeter, Introduction to Geometry, (1619) and called it the 'Stella octangula'. It also
p. 36.) has the property that its eight vertices can be
If you turn your attention now to the process made to coincide with the eight vertices of a cube
applied analogously to three-dimensional space while its edges are diagonals of the square faces
you can begin again with the simplest poly- of the cube.
hedron, namely the tetrahedron. Instead of Further extension of the facial planes of the
producing line segments you must here think of octahedron will not enclose any further space, no
extending the facial planes indefinitely. The four more cells are formed, so the stellation process
planes of the tetrahedron enclose only that terminates with only one stellated form for the
portion of three-dimensional space which be- octahedron.
longs already to the original tetrahedron. The six If you turn now to the dodecahedron and
planes of the cube come in parallel pairs, mutu- produce each facial plane you will find that it
ally perpendicular, something like the two pairs leads to the formation of three distinct types of
for the square, the two-dimensional analogue. cells inside the intersecting planes. Besides the
But no new portions of space are enclosed. The dodecahedron itself, there will be twelve pentago-
eight facial planes of the octahedron, however, nal pyramids. These convert the dodecahedron
lead to something more interesting. These will into the small stellated dodecahedron. Then
enclose not only the original octahedron, but there will also be thirty sphenoids or wedge-
also other portions of space exterior to this octa- shaped pieces which convert the small stellated
hedron. You will discover that there is actually a dodecahedron into the great dodecahedron.
set of eight small tetrahedra, like cells, each Finally there will be twenty triangular di-
sharing one of its faces with a face of the original pyramids which convert the great dodecahedron
octahedron. If you now imagine these tetra- into the great stellated dodecahedron, which
hedra added to the octahedron so that the faces more literally might be called the stellated great
they share internally melt away, leaving all the dodecahedron. Here the stellation process stops.
interior hollow, you have a non-convex poly- Thus the dodecahedron leads to three stellated
hedron. But you can equally well imagine it as a forms. Two of these were discovered by Kepler
set of intersecting triangular faces, larger tri- (1619), the third by Poinsot (1809).
angles than those belonging to the faces of the Even more interesting now is the fact that these
small tetrahedra. These larger triangles still keep polyhedra are not compounds as in the case of
the original property of a convex polyhedron, the octahedron but distinct new polyhedra. In
namely each edge belongs only to two faces. The fact they are regular polyhedra, since in two of
edges of course also intersect each other, but the them the faces are in each case a set of twelve
interior points of intersection on these line seg- intersecting pentagrams, in the third a set of
ments are disregarded, just as in the case of the twelve intersecting pentagons. It was the mathe-
two-dimensional stars. Each side of the penta- matician Cauchy (1811) who pointed out that
gram for example crosses two other sides but these are in fact stellations of the dodecahedron
these points are disregarded in counting the sides. and that these three together with the great
So in the stellated octahedron, you still have only icosahedron which is a stellation of the regular
eight faces, and the end points of these edges are icosahedron are the only regular stellated forms
the vertices of the polyhedron. possible. So to the five regular solids of the

35
21:34:34 BST 2016.
CBO9780511569746.009
ancient world modern mathematics has added inner triangle is a face of the original octa-
the four regular star polyhedra, whose facial hedron, the outer one, of the Stella octangula.
planes are regular polygons or star polygons. For the dodecahedron a star polygon within a
These faces still meet at the edges by twos but star polygon will give the pattern. See fig. 25.
they intersect each other before they do so. The numbering reveals which parts form the
Note that internal lines of intersection are dis- exterior portions of the facial planes. From
regarded. The models will clearly demonstrate these you can derive the nets required for the
these facts. construction of the models.
For the purposes of constructing these models In the following pages the lightly shaded por-
it is best to become acquainted with the stellation tion of each facial plane indicates the portion
pattern as it is found in one of the facial planes— which is visible on the top side of the polyhedron,
any one, because it is the same in all. For the and the darker shading, the portion visible on the
octahedron this is a triangle within a triangle, underside of the facial plane. It is from the shaded
the inner one with its vertices at the midpoints portions that the nets are derived for constructing
of the sides of the outer one. See fig. 24. The the models.

Fig. 24

36
21:34:34 BST 2016.
CBO9780511569746.009
 19 The stellated octahedron
(Kepler's Stella octangula)

The octahedron has only one stellated form. It

turns out to be a compound of two tetrahedra.
To make a model of this polyhedron all you need
for a net is an equilateral triangle. The colour
arrangement for the first four trihedral pyramids
is given below. These parts each have triangular
edges and they are cemented to each other in the
same way as the octahedron itself is assembled.
The pyramids here may be handled as if they
were faces. You must see to it that each of the
facial planes keeps its own colour. You will also
find parallel planes having the same colour.
The other four pyramids are the enantiomorphs.
You may get these by interchanging columns 1
and 3.
Simple as this polyhedron is, it is yet very
attractive.

1 2 3
(1) O B Y
(2) Y O R
(3) R B Y
(4) B R O

37
21:34:46 BST 2016.
CBO9780511569746.010
20 The small stellated
dodecahedron

This polyhedron is one of the four Kepler-

Poinsot solids. All you need for a net is one
isosceles triangle with angles of 72, 72 and 36
degrees. This is the triangle found at one star
arm of the pentagram, the five-pointed star. Five
of these triangles are cemented together as
shown below for each vertex part and the colour
arrangement is given in the colour table.
The five W triangles must each be cemented to
the (0) vertex part. If you remember that this
colour arrangement gives you each star plane of
the same colour, you will see the W star com-
pleted and two arms of each of the other star
planes taking shape. You will find it more inter-
esting to cement the parts as you complete 5 2 I = {I 5}
them. The next six vertex parts are enantio- 12{f}
morphs and again it is best to place each one in
position as you complete it. Each of these is
placed diametrically opposite its counterpart.
The method of construction described here is
one which gives you a completely hollow model.
This may cause it not to be completely rigid.
Each vertex part is slightly deformable because
it is in the shape of a pentagonal pyramid with-
out the pentagonal base. You could therefore
cement the pyramids to the faces of a dodeca-
hedron, but you will find that this procedure
does not give you a neat and satisfactory result.
You can get a good result with a small hollow
model. Also if you apply the cement carefully
along the full length of the tabs, then add
another drop of cement at the concave (false)
vertices around the base of the vertex parts, you
will find the result to be satisfactory. On the
other hand your own ingenuity may also suggest
other suitable ways of obtaining the desired
rigidity.

1 2 3 4 5
(0) Y B O R G
(1) W G O R B
(2) W Y R G O
(3) W B G Y R
(4) W O Y B G
(5) W R B O Y

38
21:34:58 BST 2016.
CBO9780511569746.011
 21 The great dodecahedron
This polyhedron is composed of twelve inter-
secting pentagon planes. When the model is
made in six colours, it readily gives the appear-
ance of a solid star embossed on a pentagon
plane, but each star shares each of its arms with
an adjacent star. The net for making a model is
simply one isosceles triangle with angles of 36,
36, and 108 degrees. The simplest procedure for
assembly is to make twenty trihedral dimples
and to cement them together, very similar to the
way the twenty triangles of an icosahedron are
joined. The arrangement and colour table for
the parts are set out below.
Triangle 5 is cemented to triangle 2. This com-
pletes half the model. The rest of the parts are
5 = { the enantiomorphs, and they are placed dia-
12(5} metrically opposite their counterparts.

/ 1 V 3
\
1 2 3 4 5 6 / / V \
Y W G (6) G O Y 2
(1) i
(2) B W Y CO Y R B
(3) O w B (8) B O G
R w R
(4)
V / o (9)
G w R (10)
o
R B G
Y R
^ 5
(5)
\ 6 4 /

39
21:34:58 BST 2016.
CBO9780511569746.011
 The great stellated
dodecahedron

This is the final stellation of the regular dodeca-

hedron. A model of this polyhedron can be
made by cementing triangular pyramids to the
faces of an icosahedron, but this is not recom-
mended since it does not give a neat and satis-
factory result. It is not too difficult to make the
model completely hollow inside and still rigid,
since the triangular pyramids are not easily de-
formable even with their bases missing. The net
here is simply the 36, 72 isosceles triangle, one of
the star arms. Cementing is done as shown
below together with the colour table.
The first five parts, 1 2 3, are joined in a ring
with 1 on the outer edge, which becomes pen-
tagonal. Then the W of the other parts, 4 5 6, is
cemented to 1. You will notice in doing this that 3 | 2 f = {f,3}
the star arms take their positions so that facial 12{f}
3*7"-!
planes are the same colour. The remaining parts
are enantiomorphous and are placed diametric-
ally opposite their counterparts. This model has
great decorative possibilities.

1 2 3 4 5 6
Y G B W G B
B Y O W Y O
O B R wB R
R O G wO G
G R Y wR Y

40
21:34:58 BST 2016.
CBO9780511569746.011
Commentary on the stellated icosahedron

The meaning of stellation should now be making After these and some others were discovered the
itself apparent to you. Some stellations are com- question naturally presented itself: how many
pounds. So far you have seen only one case of stellated forms are possible? In 1900 M. Bruck-
this, the stellated octahedron. But more are now ner published a classic work on polyhedra
found in the case of the icosahedron. In the case entitled: Vielecke und Vielflache, in which he
of the dodecahedron all three of its stellations presented a number of new stellations of the
turn out to be genuinely new polyhedra, in fact icosahedron. Several more are due to A. H.
they all classify as regular. Wheeler (1924). In 1938 H. S. M. Coxeter in
The icosahedron has twenty faces and if all conjunction with P. Du Val, H. T. Flather and
twenty facial planes are extended indefinitely J. F. Petrie gave the question a systematic in-
you may well imagine, or more likely you may vestigation. By applying a few restrictive rules
well fail to imagine, the multitude of cells suggested by J. C. P. Miller to determine what
enclosed within these intersecting planes. It is a forms shall be considered properly significant
fact that stellations of the icosahedron may all be and distinctive, Coxeter arrived at a total enum-
derived from the cells enclosed within these eration of fifty-nine; thirty-two with full icosa-
planes. Besides the icosahedron itself you will hedral symmetry and twenty-seven enantio-
find morphous forms with an attractively twisted
20 + 30 + 60 + 20 + 60+120+12 + 30 + 60 + 60 appearance. Coxeter's work, The fifty-nine
cells of ten different shapes and sizes. The great icosahedra is available from the University of
icosahedron is composed of all but the last sixty Toronto Press.
pieces. In making models of these stellated The stellation pattern for the icosahedron is
forms, and this holds for the octahedral and very interesting. It is most easily obtained by
dodecahedral as well, you can make these cells drawing one large equilateral triangle, one of the
first, once you have worked out their nets, and faces of the great icosahedron. On each side of
then you can cement the cells to a polyhedron this triangle you may locate two points dividing
base or to each other. But in practice this does the sides of the triangle in the golden ratio. The
not give a satisfactory result and would be symbol sometimes used is r. r = 1-618 approxi-
extremely tedious. However, acquaintance with mately, or r = i(V5 +1). The Fibonacci series is
the cell forms is very helpful in working out nets very useful here:
that are practical. In the following descriptions 0 1 1 2 3 5 8 13 21 3 4 . . . .
you will find these nets given for you. Once you The ratio of two consecutive members of this series
have done some of these, or certainly by the approaches the golden ratio as a limit. With a ruler
time you have done all of them, you will be able marked in sixteenths of an inch the measures
to find many more on your own. The nets given IT, -£-, T , I are very useful. Lines radiating
here are not necessarily the only possible ones or from these points will give you the pattern. See
the best ones. They are merely those actually fig. 26.
used in the construction of the models pictured The colour arrangement shown in fig. 27 can
in the photographs. be used to great advantage in every one of the
Several compounds occur among the stella- stellations of the icosahedron. Fundamentally it
tions of the icosahedron. There is a compound of is the icosahedral arrangement which uses five
five octahedra, a compound of five tetrahedra in colours in such a way that each of the five is
two forms, enantiomorphs, and a compound of found at each vertex but in a different order
ten tetrahedra. Surely this would have greatly from one vertex to the next. Six vertices are laid
delighted the mind of Plato had he known of it. out and numbered. The other six have the enan-

41
21:35:15 BST 2016.
CBO9780511569746.012
tiomorphous arrangement. This figure amounts which you may find surprising. The explanation
to a colour table and for that reason it can be lies in the principle of duality. The icosahedron
used again and again for all polyhedra having the and the dodecahedron form a dual pair; so too,
icosahedral arrangement and symmetry. It is the the octahedron and the cube. The tetrahedron is
same as the first arrangement, given on p. 18. self reciprocal, namely its dual is another tetra-
Many of the icosahedral stellations have a hedron. (See Cundy & Rollett, Mathematical
very marked dodecahedral symmetry, a fact models, p. 116.)

42 Fig. 26
21:35:15 BST 2016.
CBO9780511569746.012
 23 Compound of five octahedra
This polyhedron has two equilateral triangles on
one facial plane, as shown in the first figure. To
construct the model you may first make thirty
copies of the net shown, six of each of the five
colours. First assemble each of these in the form
of a pyramid but without the rhombic base.
Each of these will be a vertex of some octa-
hedron of the compound. Then take a set of five,
one of each colour, and cement them in the form
of a ring, following the (0) arrangement of
colours. Between the extending arms of this ring
a second set of five vertices is cemented, but their
orientation is such that the short slant edges of
each pyramid continue on a line with the
grooved edge between vertices of the central
ring. This means that the grooved edge and the
short slant edge form a straight line, part of the
edge of an octahedron of the compound. If you
remember to keep the basic octahedral shapes in
mind you will see them begin to develop, and the
colour will then help you to proceed correctly.
You can in fact find the rest of the icosahedral
rings beginning to appear, so this may also help
you. Once you have done this much the rest is
not hard to follow. This hollow model is not
completely rigid, but if built on the scale
indicated by the net, it will prove satisfactory.
Certainly it is better than trying to add parts to a
basic octahedron.

43
21:35:25 BST 2016.
CBO9780511569746.013
24 Compound of five tetrahedra
This polyhedron is unusually attractive because
of its twisted appearance. To make a model of
this compound all you need is twenty copies of
the net shown below, four of each of the five
colours. First make trihedral vertices with the
bottom edges looking rather jagged. If you begin
by making a ring of five vertices cemented
together with the edges that are marked 'A' on
one coinciding with the same edges of another
you will find the points at the jagged end forming
a dimple in the centre of the ring. Once you have
built this much of the model you will easily be
able to find the right positions for the other
vertex parts. The colour arrangement here
makes each tetrahedron entirely of one colour.
The centre points of each dimple are actually the
vertex points of the interior icosahedron, which
of course is not being constructed. But the
arrangement of colours in each dimple is the
icosahedral arrangement. The method of assem-
bly suggested here is perhaps a bit difficult to
execute, because all the jagged edges and points
fit into three different and adjacent dimples. The
secret is to give your attention to cementing the
tabs at one edge at a time, always beginning
with the long edge and then working out into
the dimples. The last vertex part will call for
considerable skill and patience. You may find it
better not to cement this part as a pyramid, but
to leave one section open till last. This makes a
very rigid model. It is well worth the trouble it
takes to make it as suggested here. The photo-
graph proves it can be done.

44
21:35:25 BST 2016.
CBO9780511569746.013
 25 Compound of ten tetrahedra

This polyhedron is a combination of the two

enantiomorphous forms of the compound of five
tetrahedra. A model may be made by using the
nets shown. If you cut the leftside arm to the
centre, the triangular wing can be turned down.
Then another wing as shown, but of another
colour, can be cemented in place, enantio-
morphous to the one you folded down. The one
tab on this part which is left at its shortest side
can be folded up and cemented to the under
surface of the other wing. These two parts form
half the grooved or cupped portion found
between the sections: one of which is shown
below. Their remaining tabs are used to cement
these sections together. Colours for the (0)
section are shown. Twelve sections are needed
for the complete model. This makes a very rigid
and very attractive model.

45
21:35:25 BST 2016.
CBO9780511569746.013
26 Triakis icosahedron

This model is the first stellated form of the

icosahedron. It can be assembled from twenty
parts like the net shown. These are low pyra-
mids without their triangular base. You can get
a suitable arrangement of colours using the
icosahedral colour table (fig. 27, p. 42), but if you
make each part as one colour you will not have the
facial planes the same colour. To get this each
triangle of thefigurewould have to be done as a
separate part. You may wish to work this out
for yourself.

46
21:35:25 BST 2016.
CBO9780511569746.013
 27 Second stellation of the
icosahedron

This is a very beautiful model with twelve long

spikes of pentahedral symmetry radiating from
the dimples of the compound of ten tetrahedra.
In the nets shown here you can again work out a
symmetrical colour pattern that is very suitable
without having facial planes the same colour.
The figure shows one face of a pentahedral
spike, the two smaller triangles attached to it
being folded up and cemented to form a groove.
Five of these make one spike with the grooves
radiating away from it at its base. The other net
is like the part used in the compound of ten
tetrahedra except that here it is all of one colour.
It serves as a connector for the spikes and is
cemented to the grooves, namely between them.
Again you may wish to see what colour arrange-
ments you can discover for yourself from the
icosahedral colour table.

47
21:35:25 BST 2016.
CBO9780511569746.013
28 Third stellation of the
icosahedron

This very simple polyhedron is a deltahedron.

There is a whole family of deltahedra, whose
common feature is that they all have faces which
are equilateral triangles. (See Cundy and Rollett,
Mathematical models, pp. 142-4.) Here three
equilateral triangles are to be found in each
facial plane. It has the edges and even the
appearance of a dodecahedron, but it is actually
one of the stellations of the icosahedron. It may
be imagined as a dodecahedron with the penta-
gons removed, and then pentahedral dimples
whose faces are equilateral triangles replacing
the pentagons. This suggests a simple method
for making a model of this polyhedron. You
may follow the icosahedral colour arrangement
exactly as shown on p. 42. The pentahedral
dimples each form a section, and these sections
are then cemented together in dodecahedral
fashion. The colour arrangement suggested here
will give you the three equilateral triangles on
each facial plane the same colour. You may there-
fore consider this polyhedron as a three-dimen-
sional analogue of the colour table itself.
Many of the other stellated icosahedra exhibit
portions of these equilateral triangles in their
facial planes. This fact makes this model par-
ticularly useful for its relationship to the other
forms. In some of the models that now follow,
you will see how a section of this one can serve
as a structural mould or cradle. Such a section is
shown here. The centre can be depressed or
elevated before cementing, depending on how it
is to be used.

48
21:35:25 BST 2016.
CBO9780511569746.013
 29 Fourth stellation of the
icosahedron

The stellation process applied to the icosahedron

leads to ten different types of basic cells, as
mentioned above. One of these, whose net is
shown here, can be built into a polyhedron
which has these cells only vertex connected. It is
equivalent to a stick model of the regular
dodecahedron, but the cells here are used in
place of the sticks. The method of assembly is as
follows. First make a cradle or mould to hold
the cells. This is half the deltahedron 28, the
interior serving as the mould. You will find it
best to cut holes into the vertices of the cradle
so the cells will not adhere to it when the
cementing is done. Once you have this cradle
ready, make five cells, one of each colour, and
place them in a ring in the bottom of the cradle.
The cells have only their vertices touching each
other. A drop of cement on these vertices will do
the trick. Let this ring set for a time. Meanwhile
prepare more cells. When these are ready they
can be placed with one vertex of each resting on
the two vertices already cemented, while the long
edges follow the slant edges of the cradle. Again
apply a drop of cement. But now this must be
allowed to set up firmly. After about two hours
it can be moved and turned in the cradle so that
other cells can be placed in the bottom to be
cemented. In this way the rings of cells are com-
pleted one after the other. A little patience and a
steady hand should give you a very attractive
model. The colours can be worked out so that
diametrically opposite cells are the same colour.
You can then observe that six cells of one colour
have the long edges on the faces of a cube, if you
imagine a cube enclosing the dodecahedral form.
This is the first example showing dark and light
shading to illustrate the top and the underside of
the same facial plane.

49
21:35:25 BST 2016.
CBO9780511569746.013
30 Fifth stellation of the
icosahedron

Many of the stellations of the icosahedron

appear outwardly very similar to the great icosa-
hedron to be presented later. One of these whose
nets are shown here makes a particularly
pleasing model, because it is again an example
of a polyhedron whose parts or cells are only
vertex connected. The triangles that appear in
the nets are the same as those of the great icosa-
hedron. For that reason it may be better for you
to turn to the great icosahedron first and to
make that model before attempting this one.
The same technique of cementing the vertex
parts holds here and the same colour table is
used here. In both cases it will give you a model
with each facial plane the same colour. This
polyhedron differs from the great icosahedron,
however, in that a vertex part turns out here to
be a complete polyhedron, actually an enclosed
set of cells, in the form of an intricate star
pyramid with convex and concave side faces and
a pentagram shaped base coming to a blunt
point at its centre. You should easily recognize
this base as equivalent to the dimples in the
compound of ten tetrahedra. Twelve of these
star pyramids make the complete model. They
are vertex connected at the five vertices sur-
rounding the base in such a way that small
chinks appear through which the interior can be
seen. The bases of the star pyramids form the
interior. These star pyramids are easy to as-
semble. First you must make the five that
correspond to the numbered vertex parts (1),
(2), (3), (4), (5). The (0) part is done last in this
case. When these five pyramids are made they are
laid out in a ring on a pallet shown on p. 51.
This pallet may be made out of a heavy piece of
cardboard. It is the regular pentagram but it is
best to cut holes in the card at the places indi-
cated by the small circles. It is at these points that
the cementing of the sections is to be done. They
are laid down so two edges coincide with the two
sides of a star arm. When this is done you will
find two vertices coming in contact between each

50
21:35:25 BST 2016.
CBO9780511569746.013
 section and adjacent sections, one on the plane
of the pallet and the other directly above it. A
drop of cement at these vertex contact points
will do the trick. After about 15 minutes another
drop of cement will help to give added strength.
After an hour or two you can carefully move the
whole ring off the pallet. The surprise is that it
holds together so that you can now turn it over.
Then if you have done a careful job the (0)
section will fit nicely on top with the five corners
at its base making good contact with the double
vertices already cemented. So on this (0) section
you can apply the cement to all five corners at
once. This completes half the model. The
second half is enantiomorphous. Getting the
sections properly oriented is a bit tricky. You
may have to puzzle over this for a while, but if
you remember that the arrangement is such that
each facial plane is of the same colour, passing
from one section to another, you should get it
after a little trial and error. Once the second
ring is done the first half can immediately be
cemented to it before turning the second ring
over. Then one more section completes the
model.

51
21:35:25 BST 2016.
CBO9780511569746.013
31 Sixth stellation of the
icosahedron

The parts shown here are the nets for another

stellated form of the icosahedron. These are easy
to recognize as the twelve long spikes, radiating
this time from the dimples of the deltahedron, 28.
You may begin by assembling a ring of the lower
parts in the usual icosahedral arrangement of
colours. Then assemble a spike in the same
arrangement. It is then a simple matter to insert
the base of the spike into the hole at the middle
of the ring and to cement the tabs there one at a
time with the aid of clamps to hold them till the
cement is set. The twelve sections are then as-
sembled in the usual dodecahedral manner. A
little experimentation will soon show you the
correct orientation of parts so as to get the
facial planes each in its own colour. This makes
a simple and sturdy model which at the same
time is also very attractive.

52
21:35:25 BST 2016.
CBO9780511569746.013
 32 Seventh stellation of the
icosahedron

The net shown below may be used to make

twenty parts, each a hexahedral short spike.
These parts can then be cemented together to
form another stellated icosahedron. Four of each
of the five colours may be used to make a suit-
able model, although in this way it will not have
facial planes of one colour. You may begin with
a ring of five parts, cemented so as to have the
acute angles at the bottom of the parts all point-
ing inward to the centre of the ring. Once this is
done the rest is not hard to complete by doing
each ring in the icosahedral arrangement of
colours.
This model is even more attractive when it is
made showing each facial plane in its own colour.
See if you can work out this colour arrangement
for yourself.

53
21:35:25 BST 2016.
CBO9780511569746.013
33 Eighth stellation of the
icosahedron

The net shown here can be used to make a model

of a stellated icosahedron that is very similar to
the great icosahedron. In fact you may imagine
this polyhedron as being formed by the removal
of the wedge-shaped cells found around the base
of the vertex parts of the great icosahedron.
That is why the net shown here is slightly differ-
ent at the lower end of the triangle from the
triangle used for the great icosahedron. Also only
one such triangle is shown here because the one
net can serve for all the parts, sixty as shown and
another sixty enantiomorphous to these. You
may then follow the same paired arrangement of
parts as you find given in the colour table for the
great icosahedron. The sections are a bit more
difficult to cement to one another because of the
deeper depressions at the bases of these parts.
But as you near the end of your work you can
also change your method of assembly, doing
these concave edged first and then cementing
the convex parts where you can more easily
pinch the parts together from the outside.

54
21:35:25 BST 2016.
CBO9780511569746.013
 34 Ninth stellation of the
icosahedron

The net shown here is again easy to recognize as

part of the long spikes, only here it is slightly
longer because the model is made up of twelve
such spikes and nothing else. So the spikes are to
be assembled in the usual pentahedral form
following the usual icosahedral arrangement of
the colours. Then these parts are cemented to
one another along the tabs at the base. This
model can be made so each facial plane has its
own colour but you may not find this arrange-
ment immediately. A little searching should
reveal it to you. Here also the last spike is a bit
difficult to cement in place since you cannot
reach the last tabs from the outside. But patience
will give you a lovely model.

55
21:35:25 BST 2016.
CBO9780511569746.013
35 Tenth stellation of the
icosahedron

The spike shown opposite is the net for one of

the basic cells arising from the stellation process
applied to the icosahedron. It is the only cell that
comes in two enantiomorphous forms, sixty of
one form and sixty of the other, and it is simply
a short tetrahedral spike. But the interesting
thing about it is that a set of sixty can actually be
built into a fantastically delicate yet surprisingly
stable model. Basically these spikes form all the
edges of the compound of five tetrahedra, meet-
ing by three's at each sharp pointed vertex and
by two's at one of the base vertices. At first sight
it looks quite impossible to make a model of this
polyhedron which shows mostly empty space.
But with other parts, which are later discarded,
serving as structural supports, it becomes possible the cementing. The core is made with all tabs
to make a model. These structural parts are inside as usual for any convex polyhedron. You
shown opposite. Actually each of these nets can will no doubt recognize this part as the spike
be used to construct two other stellated icosa- found in other stellations of the icosahedron.
hedra in their own right. These will be described Once all these parts are ready you may proceed
later. But the following description will help you as follows. Take three spikes of the same colour
assemble what may well be called the most un- and insert them into the casing so that they fit
believable of all models. You may begin by snugly into the three corners. If you now insert
making sixty spikes, twelve of each of the five the core so that it is centrally located inside the
colours, using the spike as a net. The colour casing and then adjust the spikes so that their
arrangement in the completed model will simply blunt ends are in the same facial planes as the
be that of the compound of five tetrahedra. You facial planes on the blunt end of the casing and
will then need only five casings and five cores; core, you will be able to close the little trap doors
their colour is not important since they are only of the casing so that they fit perfectly. If the tabs
structural and their names here indicate the pur- on the trap doors have been turned outward then
pose they serve. Turn all the tabs out on the three clamps will hold everything securely while
casing, but cement only the two longer ones to you apply a drop of cement at the point where
form the trihedral angle, congruent to a vertex the vertices of the spikes coincide. Repeat this
angle of the regular tetrahedron. The three with the next set of three spikes of a second
quadrilaterals at the bottom should also have colour in another casing and with another core,
tabs but no cementing is done; these quadri- and so on until you have five sets of three. After
laterals serve as little trap doors to be opened about an hour you will find that the spikes may
and closed when inserting and removing three be removed from the casing and the one drop of
spikes for cementing and a core to keep the cement will prove to be strong enough to hold
spikes in position. You will notice that the the spikes to the form given to them by the
figures for the casing and the core have an arc casing and the core. Now you must place these
mark at the vertices. This means that you must five sets of triplet-spiked parts on a special plat-
cut these parts away so as not to interfere with form or cradle. This is made of one section of

56
21:35:25 BST 2016.
CBO9780511569746.013
the deltahedron 28 turned over so the dimple cement to the points of contact. After two hours
becomes a low pentagonal pyramid to which or so this whole section or ring will be sufficiently
at the lower portions of its slant sides five of the rigid so that it may be moved and turned. As
29 parts are cemented. The five parts may be you complete other triplets they are placed on
cemented because they need never be removed, the platform and each ring is thus completed
but it is advisable to cut away the obtuse vertices while the model is turned each time to accom-
of 29 where cementing of cells is to be done. This modate the new parts. Admittedly you will have
is your platform or cradle. You will now find that to exercise a lot of patience and a very steady
the five sets of triplets may be placed on this hand. Also some side supports are helpful but
platform so that the vertices point down and not absolutely necessary—the dodecahedron
coincide with the vertices of the pentagonal base shell is adequate and you may be able to devise
of the platform while two other vertices at the other ingenious ways to get pressure in one
blunt ends of two adjoining parts will just touch. direction or another, to achieve contact at the
A little experimenting with the parts will make proper points. The photograph shows that it can
this more clear than words can do. Once the be done, and it has been done in the manner
parts are so placed you can again add a drop of described above.

The casing

The spike

57
21:35:25 BST 2016.
CBO9780511569746.013
36 Eleventh stellation of the
icosahedron

A model of 36 can be made using the same plat-

form and technique as for 35. The net to use is the
figure called' the casing' in 35. You will notice of
course that here the cementing of the parts is
considerably easier.

58
21:35:25 BST 2016.
CBO9780511569746.013
 37 Twelfth stellation of the
icosahedron

The figure shown here is the net for one part of a

platform needed to support the cells whose net
was given previously under the caption, 'the
core'. Here the core becomes a polyhedral cell,
so its vertex must not be removed as was done
when it served only as a structural piece. The
platform part has wings that belong to the facial
planes of 32. So here the bottom edge has the
length of the wedges 29, but not their sides. The
part makes a low triangular pyramid without a
base. Five of these pyramids are joined in a ring
to make the platform. Then if the tabs at the
pentahedral lower edges are turned up they will
serve as a groove. The cells may now be placed
on the platform, but again it is advisable to cut
away the vertex of the pyramid as shown in the
figure by the arc mark, because at these points
the cells will have their vertices coming in con-
tact and thus at these points the cementing is
done. Once a ring of cells has been done it must
be allowed to set up rigidly. Then it can be
moved and again each ring of cells will be com-
pleted in a similar fashion by moving the model
in various stages of completion so that the new
cells always lie on the platform. A dodecahedral
shell will again make a good support for stabi-
lizing the model as it is given its various posi-
tions, but it may also be dispensed with. This
means if you use it that the platform may be
placed in the bottom of a dodecahedral cradle
with one pentagon removed to make it possible
for you to work on the parts. This model is
strikingly pleasing in its simplicity and openness.

59
21:35:25 BST 2016.
CBO9780511569746.013
38 Thirteenth stellation of the
icosahedron

The net shown here can be used to make a

beautiful polyhedron with twelve long penta-
hedral spikes each of which is surrounded at its
base by five shorter trihedral spikes. You will
easily recognize both of these, as they have ap-
peared in previous models. The icosahedral
arrangement of colours will give you the longer
spikes with facial planes the same colour, but
the shorter spikes will be like those described
for 35. The technique for assembly here is to do
the short spike first. Cut the card so no tab
remains on the right side of the long part and
continue this cut to the first interior line as
shown. The two triangles on either side may now
be folded down and their remaining edges
brought into contact to form the short spike.
Finally the small triangle at the bottom is folded
down so the two acute vertices coincide. Now
five of these parts are joined to make the long
spike. This will turn out to be a complete poly-
hedron in itself since the bottom of the spike
becomes entirely enclosed. But now the five
parts whose icosahedral colour arrangements are
(1), (2), (3), (4), (5) can be laid out on the same
star-patterned pallet used for 30. The obtuse
vertices around the base of the spikes will then
be found to be coinciding, one blunt vertex of
one spike touching one on an adjacent spike.
Theoretically the small spikes should also have
acute vertices in contact, but this is almost im-
possible to achieve in practice on all the small
spikes at once. So it is advisable to cement the
obtuse vertices forming a ring of parts and as
many small spikes as can conveniently be done.
The rest of the procedure is the same as for 30.
Once the whole model is finished you can gently
apply pressure to the smaller spikes and adjust
them, as needed.

60
21:35:25 BST 2016.
CBO9780511569746.013
 39 icosahedron
Fourteenth stellation of the

The net shown here leads to an attractively

twisted polyhedron. It has some of the usual
dodecahedral symmetry, like the deltahedron
28, but it also has pentagonal holes leaving the
interior hollow and open to view. The method of
assembly is to follow the usual icosahedral
colour arrangement, the same as in the delta-
hedron 28. You may begin with the (0) section.
This will be the usual dimple but with a hole in
the middle. The tabs are shown on the net to
indicate how one edge is broken; the upper tab
of this edge serves to join a ring of five parts
forming the dimple, and the lower tab serves to
join edges forming a trihedral angle on the
interior. One of these trihedral angles is not
completely formed until three sections are
assembled. You will need a bit of skill to reach
the interior as the model progresses, but you can
apply the cement with a probing needle, working
it in between the tabs, and then applying pressure
with your fingers along the interior edges till the
cement is set. This model calls for patience and
accurate workmanship, since the assembly is
very intricate. But it makes a very interesting
model.

61
21:35:25 BST 2016.
CBO9780511569746.013
40 Fifteenth stellation of the
icosahedron

You can use the net shown here to make another

polyhedron with the same attractively twisted
appearance as the previous one. This one has the
sixty short trihedral spikes arranged so that the
interior of the polyhedron remains hollow and is
visible only through narrow chinks. You may
begin by making five parts as shown. The tabs
are indicated and a special cut as well; the other
lines are folds in the usual manner. This may
now be formed into one of the short spikes.
When you have five of these ready they can be
assembled to form the usual pentahedral
dimple from which the spikes radiate. The cut
will receive a tab which has been turned outward
from a neighbouring part and the parts may thus
be cemented with the help of some pressure
exerted by the fingers or with tweezers from the
outside. This completes one section. Twelve
sections are needed. They are assembled with the
aid of the same deltahedral platform used before.
A dodecahedral shell may be used so that the
interior of the shell provides side supports as the
sections are being cemented. The cementing is
done at the obtuse vertices which come in con-
tact. It is not always possible in practice to
cement all the sharp vertices of the small spikes
at once. But as the model is completed some
adjustments become possible after the cement is
set. Careful workmanship will give good results.

62
21:35:25 BST 2016.
CBO9780511569746.013
 41 Great icosahedron
Of all the polyhedra so far described perhaps the
most beautiful and attractive is the great icosa-
hedron, the last of the four regular (Kepler-
Poinsot) star polyhedra. The vertex figure here is
a regular pentagram. In this respect it is like the
great dodecahedron. These two solids stand
alone as the only regular star-vertexed poly-
hedra. You will see many with star faces among
the uniform polyhedra to be presented later in
this book, but none of them is star vertexed. A
model of the great icosahedron is not hard to
make. The nets are simple and when the model is
assembled as described here it is also very sturdy
and rigid although it is completely hollow inside.
Doing it in five colours takes a little more time,
f|2 3 = { but it is well worth the effort. The paired
20{3} arrangement of parts and the colour table
are given below.
A set of five pairs makes the fan-like form
shown. You must now see to it that the folding
is done so that it is down between each member
of the pairs and up between the pairs, accordion
fashion. Then by cementing the remaining edges
a vertex part for the great icosahedron begins to

12 34 56 78 9X
(0) Y G BY OB RO GR
(1) BG YB RY OR GO
(2) OY BO GB RB YR
(3) RB OR YO GY BG
(4) GO RG BR YB OY
(5) YR GY OG BO RB

63
21:35:25 BST 2016.
CBO9780511569746.013
take shape. The smaller isosceles triangles are
then cemented in place to form a pentagonal
dimple from which the vertex part rises. Twelve
of these are needed, the first set of six as set out
in the colour table and the second set of six in
the enantiomorphous order. These vertex parts
are then joined following the icosahedral
arrangement, enantiomorphous pairs being dia-
metrically opposite each other. In cementing the
vertex parts together, give your attention to only
one edge at a time. Clamps can readily be used
because the dihedral angle between two adjacent
facial planes along the pentagonal edges of a
vertex part is very acute. Even the last vertex part
goes on in the same way. On the last edges the
cement is applied carefully to the crack and then
worked down between the tabs with the probing
needle before setting the clamps. If you find that
there are openings or small holes left at the
corners of the pentagonal edges, do not be dis-
mayed. You can close these after the model is
completed by adding a drop of cement to each
hole, working it in with the probing needle and
then applying a little pressure when the cement
begins to harden. You will find that this closes
the holes successfully and at the same time adds
rigidity to the model. This model, when it is well
made, is always a delight to behold.

64
21:35:25 BST 2016.
CBO9780511569746.013
 42 Final stellation of the
icosahedron

This is the final stellation of the icosahedron. It

is a spiny-looking polyhedron, the spines falling
into fairly well-defined clusters of five. Twelve
of these clusters complete the solid. A net for
making a model of this polyhedron is shown
below. It will not give the facial planes each its
own colour, but then the work is reduced by using
the net as shown. Sixty of these trihedral spikes
are needed. A set of five joined in a ring makes
one section or the cluster mentioned above with
the bottom forming the edges of a pentagon. So
twelve of these are joined in the usual dodeca-
hedral fashion. It is best here to reduce the scale
of the model, unless you want to show its actual
size in relation to the great icosahedron. If the
latter is about 10 inches high, the final stellation
on the same scale is nearly 24 inches high. If the
model is done as suggested here, completely
hollow inside, the smaller scale gives a better,
more stable, result. The sixty spikes radiating
from the central mass of this polyhedron make
it look like rays of light emanating from the sun.

If you are interested in making more of these

stellated icosahedra you can work out your own
nets by consulting The fifty-nine icosahedra, by
Coxeter, Flather, Du Val and Petrie. In this
booklet all the cases are illustrated with excellent
drawings of both the solids and their facial planes.
Each one presents its own challenge and will give
you a sense of satisfaction upon completion.

65
21:35:25 BST 2016.
CBO9780511569746.013
Commentary on the stellation of the Archimedean solids

In the previous pages you saw the stellation is: Are any of these stellated Archimedeans
process applied to the Platonic solids. You may regular or uniform polyhedra? Before attempt-
now be wondering whether the Archimedean ing any answer to this problem, you will find it
solids can also be stellated. The answer is that enlightening to see the stellation process applied
they can. The procedure is the same: each facial to at least two Archimedeans.
plane must be extended indefinitely to generate The cuboctahedron and the icosidodeca-
cells exterior to the original solid. Using these hedron are given here because of their close
cells as building blocks, you can form many new relationship to the dual Platonic pairs. Also as
solids, theoretically at least. In practice however, quasi-regular solids they should prove to be the
for the purpose of making models, the stellation most interesting or the most likely ones to gener-
pattern is more useful, although you will find it ate further regular or uniform polyhedra.
helpful to have some acquaintance with the cells How is the stellation pattern arrived at? If you
as well. Nets for these cells can easily be found look back at the case of the octahedron, you will
from the stellation pattern. But you may be notice that the pattern is actually a set of six
asking yourself: why should anyone want to lines. They can easily be counted as they come in
stellate the Archimedean solids ? Isn't it a lot of parallel pairs (see fig. 28). The inner triangle is
work? Yes, the work involved here does begin to one face of the original octahedron. If you set a
look overwhelming. It really calls for team work. model of the octahedron on the drawing so that
It is hard to find much published material on one of its faces exactly covers this inner triangle,
this topic, and undoubtedly it is because of the the other lines are easily seen to be the inter-
great number of possible forms that come sections of the other facial planes with the plane
crowding up for consideration. A complete of this base triangle. Since the triangle on top is
enumeration of all the possible stellations is a directly opposite the base triangle, it is on a
mathematical question that has yet to be in- plane parallel to the plane of the paper and so it
vestigated. No doubt some restrictive rules, such generates no line on the stellation pattern. Thus
as J. C. P. Miller designed for the case of the the eight faces of the octahedron are all ac-
icosahedron, would have to be applied here also. counted for.
These forms do not always turn out to be par- If you turn now to the dodecahedron, its
ticularly attractive or aesthetically pleasing, but twelve faces ought to give a stellation pattern of
then again many of them do. The final stellation ten lines. This is indeed the case; five parallel
usually seems to be of more than ordinary pairs appear (see fig. 29). If you set a model of
interest. the dodecahedron on this drawing, so that one of
A more important question mathematically its faces coincides exactly with the innermost

Fig. 28 Fig. 29

66
21:35:37 BST 2016.
CBO9780511569746.014
 Fig. 30 Fig. 31

pentagon, you can move your eye into the plane six square pyramids with equilateral triangles
of the other faces produced and observe how for side faces, eight triangular pyramids with
they cut the plane of the paper precisely on the triangular bases and isosceles right triangles for
other lines of the drawing. side faces, twenty-four dipyramids with equi-
The icosahedron gives a similar result. Its lateral and isosceles right triangles for faces,
twenty facial planes generate a stellation pattern twenty-four pyramids like the six before, and
of eighteen lines, nine parallel pairs (see fig. 30). finally twenty-four rhombic pyramids with
You can now see what principle is at work in rhombic bases and equilateral and isosceles
the stellation process. Applying it to the Archi- right triangles for side faces. How many solids
medean solids simply means that two or more can be formed from these cells ? That all depends
stellation patterns turn up, one for each face that on what restrictive rules you wish to apply. Some
is different. may be only vertex connected or they may have
The cuboctahedron has eight triangular faces holes leading to the interior, like some of the
and six square faces, a total of fourteen. So it stellated icosahedra. Miller's rules for the icosa-
will have two stellation patterns, each with hedron invoke chiefly symmetry requirements
twelve lines. For the triangular face there are and accessibility from the outside. The models
three sets of parallel lines, four to a set; for the 43, 44, 45 and 46 are examples of only four
square face there are four sets of parallel lines, possible cases.
but they come in one set of two, then a set of As before, the shaded portions of each facial
four, and then a repetition of these, a set of two plane show what is visible from the outside of
and a set of four (see fig. 31). The cells are easily the polyhedron. It is also from these shaded
enumerated. There are only four different kinds. portions that the nets are derived for the con-
Besides the cuboctahedron itself you will find struction of the models.

67
21:35:37 BST 2016.
CBO9780511569746.014
43 Compound of a cube and
octahedron

The first stellation of the cuboctahedron is a

compound of the cube and the octahedron. This
makes a very interesting model when done in
colour. The cube may be done in three colours
and the octahedron in two more. The nets are
simply the two different types of triangles shown
in the facial planes below. Four of the equi-
lateral triangles are cemented with alternate Y and
B parts to make a square pyramid without its
base. Three of the isosceles right triangles make
a triangular pyramid, again without its base.
The colours O, R, G are arranged so the oppos-
ite-feeial: planes of the cube have the same colour.
Six of the first kind of pyramid and eight of the
second will complete the model. This hollow
model gives a neater result than can be obtained
by cementing pyramids of one kind to a basic
cube or octahedron. A net with one long tab
crossing a weak point, as shown here, can also
be used to make a very rigid model.

68
21:36:02 BST 2016.
CBO9780511569746.015
 44 Second stellation of the
cuboctahedron

This polyhedron arises as the second stellation

of the cuboctahedron. Twenty-four dipyramids
have here been added to the compound of the
cube and the octahedron. A model of this poly-
hedron, however, is more easily made using the
same nets as before. Here it is best to make a
ring of four parts using the net with one long tab
crossing a weak point. These four parts form a
square section prism, the long tabs being at the
square bottom and the jagged sides coming right
down to the bottom at the centre point. The top
of this prism is then closed with eight equi-
lateral triangles. This completes one section of
the model. Six of these sections are needed
altogether. The colour arrrangements of the
facial planes may follow those of the previous
model as they extend themselves to this case.
The final result is easily recognizable as a trun-
cated form of the Stella octangula.

69
21:36:02 BST 2016.
CBO9780511569746.015
45 Third stellation of the
cuboctahedron

This polyhedron is very interesting in more than

one way. First of all the square faces stand out
very plainly falling as they do into three sets of
pairs. The members of each pair are parallel to
each other and the sets are perpendicular to each
other. Secondly the polyhedron is a sort of com-
pound, composed of six square pyramids, the
squares mentioned previously serving as the
bases of these pyramids while the triangular side
faces quickly disappear into the interior of the
solid, their vertices coinciding with the middle
point of the opposite depression. All this is more
readily seen in a model than it is in words or
description.
To construct this model, begin with four
squares and cement a pair of equilateral triangles
to each of them. These parts are then cemented
together by joining the triangles, the squares
remaining to the outside of the ring. Six of these
sections are required, and again they are joined
by the equilateral triangles. Finally, trihedral
dimples whose faces are the small isosceles right
triangles, close the holes formed by three of the
connecting triangles.

70
21:36:02 BST 2016.
CBO9780511569746.015
 46 Final stellation of the
cuboctahedron

The final stellation of the cuboctahedron is par-

ticularly attractive since it is a compound of two
tetrahedra, Kepler's Stella octangula being the
final stellation of the octahedron, and of three
perpendicular prisms of square cross-section
whose common portion on the interior is the
original cube. The outer or end faces of these
prisms are four rhombic faces forming deep
tetrahedral dimples. Two very simple nets serve
here for the parts needed to make a model of this
polyhedron. The method of assembling in
sections is the best to use. Four of the chevron-
shaped parts are cemented in a ring to form a
prism open at both ends. Four rhombic parts
are then cemented to form the deep tetrahedral
dimple and this is used to close the upper end of
the prism. The cementing is best done one edge
at a time and clamps can easily be used. Six of
these sections are needed and they are joined by
the vertex parts of the stellated octahedron.
These vertex parts also have rhombic faces, so
the same net serves for them as for the dimples.
In doing the last section it is best to cement it in

71
21:36:02 BST 2016.
CBO9780511569746.015
place before closing its end with the rhombic is that the result be a genuinely new polyhedron,
dimple. In this way you can still work on the tabs not a compound. You can see that this is not
through the open end of the prism which can fulfilled. The first model, 43, is a compound of a
easily be closed last of all. cube and an octahedron. Model 44 has one
You can see that none of these stellated cubocta- facial plane regular, the octagram, but it is com-
hedra is regular or uniform. But maybe some bined with a truncated triangle which is not a
others not shown here would turn out to be so. regular polygon; so that fails. The third one,
To find them you would have to investigate the model 45, also comes out with one face regular,
stellation pattern and find regular polygons the square, but it again is combined with a
whose line segments coincide with the lines of truncated triangle. The final stellation has
the pattern. The first model given, 43, fulfills neither of its faces regular. But it does bear some
this requirement. However a second requirement resemblance to 92 of the uniform polyhedra.

72
21:36:02 BST 2016.
CBO9780511569746.015
Commentary on the icosidodecahedron

The icosidodecahedron has twelve pentagons cells. In other words these latter cells are further
and twenty triangles as faces, a total of thirty- split up and subdivided by the extended facial
two. This looks a formidable number to investi- planes of the icosidodecahedron.
gate. It will mean studying two stellation The following polyhedra exhibit only a
patterns, each composed of thirty lines. representative sub-set of stellated forms in this
As a first step to drawing the patterns it is set. You will quickly notice that some of them
worth noting that an icosidodecahedron is the are compounds or variations of the three stel-
solid that is common to a compound of a lated dodecahedra and/or some stellated icosa-
dodecahedron and an icosahedron. These pat- hedron. This gives many of them the same beauty
terns are already known. Those to be done now of form. But—the big question—are any uni-
must therefore bear some resemblance and ought form polyhedra to be found ? The answer seems
to serve as guides. So proceeding as for the to be—no. Striking resemblances turn up, but
stellated cuboctahedra you should be able to none of them satisfies the definition of a uniform
verify the two patterns given on the following polyhedron.
pages. As for finding regular polyhedra it has already
These two stellation patterns lead to forty been mentioned that the great mathematician
different kinds of cells. You would have to be Cauchy in 1811 proved that the four Kepler-
exceptionally ambitious to verify this, so no Poinsot solids taken together with the five
more will be said about them here, except for Platonic solids exhaust the list of regular poly-
this one brief remark. Just as the icosidodeca- hedra. So if you continue to look, you are
hedral patterns include as a subset the lines of merely joining the ranks of angle trisectors or
the dodecahedral and icosahedral patterns, so circle squarers or cube duplicators.
too the icosidodecahedral cells are building For identification the models are numbered
blocks for the dodecahedral and icosahedral 47 to 66.

73
21:36:23 BST 2016.
CBO9780511569746.016
Fig. 32

74

21:36:23 BST 2016.

CBO9780511569746.016
 Fig. 33

75

21:36:23 BST 2016.

CBO9780511569746.016
47 First stellation of the
icosidodecahedron

This polyhedron is a compound of two Platonic

solids, the dodecahedron and the icosahedron.
It is the first stellation of the icosidodecahedron.
This begins what is called the main line, namely
the polyhedra derived successively from previous
ones by adding cells to cover completely all the
'outside' surface area. Thus twelve low pentag-
onal pyramids and twenty small triangular
pyramids cover completely the interior icosi-
dodecahedron. A model, however, can be made
completely hollow inside. You may make the
pentagonal pyramids following the icosahedral
arrangement of colours, but it is best to make all
the small triangular pyramids the same colour,
say W. These pyramids, that is both types, are
left without bases. They are then cemented to
one another in the same way that you cemented
the parts for any convex polyhedron. This
method of construction gives a very neat result.

76
21:36:38 BST 2016.
CBO9780511569746.017
 48 Second stellation of the
icosidodecahedron

This is the second stellation of the icosidodeca-

hedron in the main line. It is related to the
small stellated dodecahedron, being in fact a
special type of truncation of that regular star
polyhedron. This suggests a method of con-
struction.
A net of five parts is shown with the tabs
illustrated also. The triangles may be arranged in
the icosahedral colour scheme, ten of them being
required to close the opening once the ring of
five parts has been formed. Then these sections
are cemented as usual in the case of a dodeca-
hedron. The tabs give it good rigidity since they
span a crucial weak point. This polyhedron is
not particularly attractive—merely a stepping
stone to other forms.

77
21:36:38 BST 2016.
CBO9780511569746.017
49 Third stellation of the
icosidodecahedron

This is the third stellation in the main line of the

icosidodecahedron. It is interesting to see a
pentagon as one face, but the other is not a
regular polygon. In making a model of this poly-
hedron it is best to make sets of trihedral parts
which are vertices of the polyhedron. These are
formed by using one rhombus and two triangles
for each part, as shown. Five of these form a
ring. The rings are cemented together using the
small dimples and triangles as one connector
between three rings. This may not be a very
attractive model, but it helps to illustrate the cell
formation of polyhedra. You may follow the
colour arrangement through from the first two
models, making all the pentagon parts W and
using the five-colour icosahedral arrangement
for the rest.

78
21:36:38 BST 2016.
CBO9780511569746.017
 50 Fourth stellation of the
icosidodecahedron

This polyhedron is a compound of the small

stellated dodecahedron and the triakis icosa-
hedron, both the first stellations of the dodeca-
hedron and the icosahedron respectively. The
facial planes shown above make this relationship
evident. An attractive model results from following
the usual colour arrangement used in the original
separate polyhedra. Thus you may make the vertex
parts of the small stellated dodecahedron in the
form of pentahedral angles and the vertex parts of
the triakis icosahedron as trihedral mounds whose
faces are three kites. These mounds then become
the connectors for the other vertex parts. The
photograph will make the arrangement evident.
This polyhedron is no longer in the main line
of stellation. It is more interesting to follow out
various combinations of compounds. Since the
dodecahedron has three stellated forms and the
icosahedron has fifty or more to choose from,
the results can become something like musical
'variations on a theme'.

79
21:36:38 BST 2016.
CBO9780511569746.017
51 Fifth stellation of the
icosidodecahedron

In this polyhedron the small stellated dodeca-

hedron penetrates the compound of five octa-
hedra. To construct a model you may proceed as
before. First make the pentahedral vertex parts
of the small stellated dodecahedron in the usual
colour arrangement. But notice that the shape
is slightly different toward the bottom. The net
for the vertex parts of the compound of five
octahedra is simply a set of four quadrilaterals
derived from the appropriate facial plane shown
above. These again form mounds, slightly higher
than in the previous model. These mounds again
serve as connectors for the pentahedral vertex
parts. However they are also connected to each
other in three's. This will be evident to you from
the photograph. The octahedra here can easily
escape notice. They only become apparent when
the model is examined at close range.

80
21:36:38 BST 2016.
CBO9780511569746.017
 52 Sixth stellation of the
icosidodecahedron

This polyhedron has two pentagrams coinciding

with the twelve facial planes of a dodecahedron,
one of the pentagrams (the larger one) still
belonging to the small stellated dodecahedron.
It can be derived from the previous polyhedron
by the removal of two different types of cells,
which alter the shape of the compound of five
octahedra to the resulting form it assumes here.
The same technique of construction that was
used in the two previous models can be used
here. The twelve vertex parts of the small stel-
lated dodecahedron are done in the usual way
and a hexahedral mound serves as a connector.
The faces of this mound are the following: a kite
from the smaller pentagram, then two triangles
which are enantiomorphous pairs from the icosa-
hedral plane, then another repetition of these—a
kite and a pair of triangles. The drawings and
the photograph together should make this clear.

81
21:36:38 BST 2016.
CBO9780511569746.017
53 Seventh stellation of the
icosidodecahedron

This polyhedron is a compound of the great

dodecahedron, the second stellation of the
dodecahedron, and 32, one of the stellations of
the icosahedron. To build a model of this poly-
hedron the best procedure is to make the dimples
of the great dodecahedron with a large hex-
agonal hole in each, following the usual colour
arrangement. Next prepare the hexahedral
spikes, whose faces come from the icosahedral
planes, according to their colour arrangement.
These spikes are then cemented to the dimples,
closing the holes. These sections are then as-
sembled as for the great dodecahedron. The
single tab crossing the weak point gives the
rigidity needed for a beautiful model.

82
21:36:38 BST 2016.
CBO9780511569746.017
 54 Eighth stellation of the
icosidodecahedron

This polyhedron is easy to recognize as the

compound of five tetrahedra penetrated by the
great dodecahedron. The vertices of the latter
appear as small rosettes at the bottom of the
dimples of the former. If you were successful in
making a model of the compound of five tetra-
hedra, you will undoubtedly want to attempt
this one also. The procedure in making this
model is practically the same. You may begin
with the trihedral vertices of the compound of
five tetrahedra and join them in a ring, as before.
In this case, of course, a decagonal hole is left
in the centre of the ring. Fill this hole with a
vertex part of the great dodecahedron, the
rosette mentioned above. It is easy to see, from
the pentagon plane, that this has ten triangular
faces. The secret in cementing these rosettes is
to do them one tab at a time. You will need con-
siderable skill and patience toward the end of
the work. But it is worth the effort, because it
makes a very beautiful model. The colour
arrangements are the usual ones used in the
separate polyhedra.

83
21:36:38 BST 2016.
CBO9780511569746.017
b b Ninth stellation of the
icosidodecahedron

This polyhedron is the compound of ten tetra-

hedra, but the ghost of a great dodecahedron
leaves the traces of its facial planes in the holes
at the bottom of the dimples and on the interior
which remains visible through the holes. The
rosettes used in the previous case would exactly
fill these holes, but this makes a very interesting
model as it is. However, it calls for a slightly
different method of assembly. Two of the butter-
fly shapes may be joined with two pairs of the
smaller triangles, the grooves between them
acting as connectors. The Y-cut at the bottom,
that is, under the groove, then has two of the
irregular pentagons as faces. These faces
eventually form the interior surfaces. This
assembly forms one section. Thirty of these
sections will complete the model.

84
21:36:38 BST 2016.
CBO9780511569746.017
 56 Tenth stellation of the
icosidodecahedron

This simple polyhedron has the external appear-

ance of the deltahedron 28, one of the stella-
tions of the icosahedron. Here the ghost of a
small stellated dodecahedron leaves its traces in
the hole at the bottom of the dimple and on the
interior. The vertex parts of the small stellated
dodecahedron would fill these holes. This means
that the same thing could be done to the com-
pound of five tetrahedra and the compound of
ten tetrahedra; the adjustments to be made to
the nets are easily seen from the stellation
patterns.
The method of assembly for this polyhedron
had best be like that used in the previous model.
One section has two truncated triangles joined
along the longest side and two isosceles trapezia
which eventually become the interior of the
model. These sections are rather flexible but
once they are joined rigidity is achieved. Thirty
sections complete the model.

YV

85
21:36:38 BST 2016.
CBO9780511569746.017
57 Eleventh stellation of the
icosidodecahedron

This polyhedron is truly remarkable because it

bears such a close resemblance to one of the
uniform polyhedra. As the drawings here show
it has one facial plane an equilateral triangle and
the other a decagram, almost regular but not
quite. It is actually a truncated form of the great
stellated dodecahedron, the truncation occur-
ring very low, near the base of a vertex part. This
fact suggests a method of assembly for a model.
You may join three of the parts shown below.
Keeping one tab long strengthens the weak
point. Three kites form a dimple and are used to
close the opening of the truncated pyramid part.
These parts are then cemented in the same way
as was done in the great stellated dodecahedron.
The usual dodecahedral and icosahedral colour
arrangements can be used very effectively.

21:36:38 BST 2016.

CBO9780511569746.017
 58 Twelfth stellation of the
icosidodecahedron

This polyhedron is very attractive. A truncated

form of the great stellated dodecahedron is here
making its appearance, penetrating a delta-
hedron, like the one found among the stellated
forms of the icosahedron, 28. The icosahedral
faces are composed of three equilateral triangles,
slightly larger than those in 28, so that their
common portion turns out to be the triangular
faces of the interior icosidodecahedron. The
dodecahedral faces are decagrams, almost
regular in appearance but not quite so. It is the
icosahedral faces which have truncated the great
stellated dodecahedron. As you will see once you
have constructed a model of this polyhedron, it
also gives the appearance of five long cells
tapered at their outer ends all placed neatly in a
ring in each dimple of the deltahedron, 28. This
fact suggests a very simple way to construct a
model of this solid. Begin by making a set of five
long cells with the four parts as shown (note that
they are not closed cells). These cells are then
cemented in a ring, adjoining each other and
radiating outward from their blunt ends.
Portions of the triangular planes are cemented
between these cells, giving a section reminiscent
of those used in other stellations of the icosa-
hedron, in particular 28 and the great icosa-
hedron 41. Twelve of these sections will com-
plete the model.

Portion of triangular plane The four parts for a

long cell (not closed)

87
21:36:38 BST 2016.
CBO9780511569746.017
 In the drawings given on the previous page dodecahedral pattern. This also means that the
only some of the lines from the stellation pat- vertex parts of the solid are dissected into
terns are shown. You can easily see for yourself numerous cells. Various selections of these cells
how this is so. You will then also notice how the could lead to many different truncated forms of
vertices of the great stellated dodecahedron just the great stellated dodecahedron.
manage to make it to the exterior lines of the

59 Thirteenth stellation of the

icosidodecahedron

In this polyhedron the great stellated dodeca-

hedron penetrates 34, one of the stellations of the
icosahedron. Twenty plus twelve vertex parts are
thus seen radiating from the central mass. If you
have succeeded in making models of the two
separate polyhedra, this compound should not
give you too much trouble. The vertex parts of
both types are first assembled separately and
then joined to one another. You may use the
usual colour arrangements and begin by sur-
rounding an icosahedral vertex part with a ring
of five dodecahedral vertex parts. Once this is
completed the rest becomes evident. The secret
in doing the last part is this: cement the three
kites of a dodecahedral vertex part separately,
not as one unit. In this way the three longer
sides of these kites are cemented last of all and
can be pinched with the fingers from the outside.
This makes a very attractive model.

88
21:36:38 BST 2016.
CBO9780511569746.017
 60 Fourteenth stellation of the
icosidodecahedron

In this polyhedron the great stellated dodeca-

hedron penetrates a truncated form of the great
icosahedron. The truncation is effected by
removal of some of the stellation cells from the
latter. It certainly gives a spiny result resembling
a sea urchin. The twelve vertex parts of 34 are
still seen here surrounded by a ring of five lower
vertex parts. The vertex parts of the great stel-
lated dodecahedron exhibit rhombic faces.
A good method for constructing the model is
to assemble the three parts needed for the lower
vertex part, then to assemble a ring of five. This
ring will have a pentagonal hole at the bottom of
a depression, its edges formed by the small
isosceles triangle with base angles 72°. This
pentagonal hole is closed with the long spiked
vertex part of the stellated icosahedron 34. This
is one section of the model. Twelve of these are
needed and they are joined together using the
dodecahedral vertex parts as connectors. The
last part may be completed as in the previous
model.

89
21:36:38 BST 2016.
CBO9780511569746.017
61 great
Compound of the
stellated dodecahedron
and the great icosahedron

The two regular star polyhedra, the great stel-

lated dodecahedron and the great icosahedron
are probably the most attractive polyhedra of all.
Here they are to be found together in one com-
pound, different from the one given by Cundy
and Rollett, Mathematical models, pp. 132-3.
Here they appear together as a stellated form of
the icosidodecahedron. To make a model of this
polyhedron you may use the same technique as
that used for the great icosahedron. The nets are
shown below. These vertex parts are not as stable
as the former ones, since some of the lower por-
tions are missing. But these get their rigidity back
again when the dodecahedral vertex parts are
added in rings of five. You will find that it will
take a great deal of patience to do this model well.
But then it is also well worth the effort.

90
21:36:38 BST 2016.
CBO9780511569746.017
 62 Fifteenth stellation of the
icosidodecahedron

This is a remarkable polyhedron because it so

closely resembles number 95 among the uni-
form polyhedra given later in this book. It is not
itself uniform because the hexagons in its facial
planes are not regular and the pentagons are
incomplete or broken at the vertices. The draw-
ings of the facial planes shown below reveal this
clearly. But it is not difficult to see what adjust-
ments are required to achieve uniformity. This
polyhedron is a truncated version of the great
icosahedron. The pentagonal planes of the
icosidodecahedron are here effecting the trunca-
tion. Number 95 is also a truncated great
icosahedron, but the truncation is more simply
effected by a plane parallel to a plane of the
vertex figure in such a way that the triangles are
converted to regular hexagons.
A model of this polyhedron can be assembled
the same way as that used in the great icosa-
hedron. The nets for the icosahedral planes are
shown below; they are merely truncated versions
of the other nets. After you have assembled a
ring of these a star-shaped hole is left. This is
closed with a very elaborately pitted star form.
Its centre is a vertex part of the small stellated
dodecahedron turned inside out to form a cup.
A set of five trihedral dimples forming the star
arms are cemented to the edges of the cup. This
completes the pitted star. It is cemented into the
icosahedral part one edge at a time with the aid
of clamps. Twelve icosahedral parts complete
the model.

91
21:36:38 BST 2016.
CBO9780511569746.017
63 Sixteenth stellation of the
icosidodecahedron

Some of the stellated icosahedra exhibit cells or

combinations of cells which are connected only
by the vertices. This gives these polyhedra an
open, airy quality. The stellation cells of the
icosidodecahedron can be selected in the same
way, giving the same result. Three of these are
presented here, this model and the following
two. You may imagine this one as resulting
from the removal of the stellated'icosahedron 32
from the compound polyhedron 53 (see pp. 53
and 82). The vertex parts of the great dodeca-
hedron are left, all alone in the form of beautiful
solid stars, a set of twelve, joined by the vertices.
The nets for one star arm are shown below. Five
of these pairs are needed for one solid star. To
join them at their vertices you will need a con-
struction cradle. This cradle is one section of the
great dodecahedron, containing one complete
pentagonal plane, inverted of course so that it
becomes a cradle. It is best to cut holes in the
edges of the cradle at the points where the
cementing is done. Patience and a steady hand
will give you the beautiful model shown in the
photograph.

92
21:36:38 BST 2016.
CBO9780511569746.017
 O ^r Seventeenth stellation of
the icosidodecahedron

This polyhedron and the next one are closely

related to the compound of the great stellated
dodecahedron and the great icosahedron 61.
You may imagine in this case that the great
icosahedron has disappeared leaving the traces
of its facial planes on the interior of this one.
But to get the remaining cells of the great stel-
lated dodecahedron to be vertex connected the
rhombic faces of these cells must be completed.
This is evident from the drawings below when
you compare them with the drawings of the com-
pound. Completing the faces means adding cells
to bring the vertices into contact. You can make a
model of this polyhedron by using the net shown
below. Three of these make one vertex part for
the great stellated dodecahedron, a completely
enclosed cell trifurcated at its base. Here again a
construction cradle is needed to join the cells.
This is simply one ring of five vertex parts of the
regular great stellated dodecahedron. Again it is
best to cut holes in the edges at the points where
the cementing is done. You must also leave three
face triangles of the cradle uncemented, other-
wise the model cannot be removed and turned in
the cradle as the work proceeds. The completed
model holds together surprisingly well, a little
springy but well able to support its own weight,
much like the models 35 or 29.

93
21:36:38 BST 2016.
CBO9780511569746.017
65 Eighteenth stellation of the
icosidodecahedron

In this polyhedron the ghost of a great stellated

dodecahedron leaves its traces in the holes and
in the interior of the great icosahedron. The
decagram, really the truncated pentagram,
which appears so clearly in the drawing below is
virtually lost in the polyhedron because the eye
is arrested by the exterior triangular planes.
Only close inspection will reveal the true nature
of the interior surfaces. A model of this poly-
hedron is easily made in the same way that you
assembled one of the stellated icosahedra, a
vertex-connected model like this one. In fact if
you make it the same size as 30 you can lay out
the cells on the same pallet. A net for one part of
such a cell is shown here. Ten of these are
needed, five enantiomorphous pairs, for one
vertex part The same colour arrangement as that
used for the great icosahedron serves very well
here. As you can easily see, twelve of these cells
or vertex parts will complete the model. You
will also see that the contact points are different
here. This makes the model slightly more difficult
to construct, but it still comes out quite rigid,
surprisingly enough.

94
21:36:38 BST 2016.
CBO9780511569746.017
 66 Final stellation of the
icosidodecahedron

The stellation patterns for this model are given

on pp. 74 and 75.
The final stellation of any polyhedron is usually
of more than ordinary interest. Here is the final
stellation of the icosidodecahedron. It gives the
appearance of twelve bursting sprays, like fire-
works in a night sky, emanating from a central
mass, but here all the sprays have mathematical
precision. The final stellation of both the dodeca-
hedron and the icosahedron are evident in this
polyhedron. The great stellated dodecahedron
just manages to reach the exterior as small tri-
hedral cells, almost lost, like blades of grass at
the foot of giant oaks. The final stellation of the
icosahedron is itself a set of twelve clusters of
five long spikes to a cluster. Here five thin cells
fill some of the space between these spikes,
giving the whole section the spray-like effect
mentioned above. This suggests a method for
constructing a model of this polyhedron.
Begin by making a cup or tapered prism, open
at both ends, one end very jagged and the other

95
21:36:38 BST 2016.
CBO9780511569746.017
in the form of a regular pentagon. Five of the to leave all the tabs on these parts and simply
parts marked A are used for this. Then five pairs apply cement to all six tabs of one part and
of the parts marked B, ten trapezia, are cemented press it in place between three of the cups. The
together, being joined at the lower, short, blunt shaded portion of the net A shows the area that
ends. These parts fold up, accordion fashion, is covered by these parts. This model calls for
enantiomorphous pairs facing each other. Next careful workmanship. The final result can be
cement five rhombic parts so their lower acute very attractive in the usual icosahedral colour
vertices go between the openings of the accordion arrangement with W for all the dodecahedral
folds. Once this is done you have the deeply planes.
pitted interior of a cup or tapered prism. This
can now be cemented into the cup, one edge at a You should now be able to discover other stel-
time. It is best to do the rhombic edges first lated forms of the cuboctahedron or the icosi-
since clamps can easily hold at these edges, the dodecahedron by yourself. This can be done if
dihedral angles here being very acute. The other you are acquainted with the stellation cells and
edges along the centre spike on each of the five see how their faces are found in the stellation
sides of the cup can then be cemented, and held patterns. In fact you should now also be able to
in place with your fingers until the cement is set. stellate other Archimedean solids by yourself.
This completes one section. Twelve sections are You do not need complete stellation patterns to
needed and these are joined together in dodeca- begin. These are put together as you proceed, by
hedral fashion. Finally the small vertex parts of trial and error if by no other way, something like
the great stellated dodecahedron are cemented a crossword puzzle or, even more so, like a
in place after all the rest is complete. It is easiest three-dimensional jig-saw puzzle.

96
21:36:38 BST 2016.
CBO9780511569746.017
Non-convex Uniform Polyhedra

97
21:36:57 BST 2016.
ebook.jsf?bid=CBO9780511569746
Commentary

You have now seen the stellation process applied plainly. The great stellated dodecahedron is a
to the Platonic solids and two of the Archi- faceted dodecahedron as well as a stellated one.
medean solids. You have also seen that it leads If you imagine straight lines joining each vertex
to very few uniform polyhedra. In fact only the to three adjacent ones, the whole set of these line
three dodecahedral stellations and one icosa- segments forms the edges of a regular dodeca-
hedral stellation turn out to be uniform. You hedron. Thus the vertices of the great stellated
may recall that a polyhedron is uniform if all its dodecahedron coincide with those of a dodeca-
faces are regular polygons (these now include hedron encasing it. The great icosahedron is a
the regular stars) and all its vertices are alike. faceted icosahedron as well as a stellated one,
The list so far contains the five Platonic solids, for the same reason. Many of the models now to
the thirteen Archimedean solids and the four be presented will amply illustrate this principle
Kepler-Poinsot solids. Are there any more uni- of faceting.
form polyhedra? It may surprise you to learn Badoureau was mentioned above. Other in-
that there are at least fifty-three more! How were vestigators include Hess (1878), who discovered
they ever discovered? Thirty-seven of them are two new uniform polyhedra. (Notice the earlier
due to Badoureau (1881) who systematically date.) Pitsch (1881) working independently dis-
considered each of the Platonic and Archi- covered eighteen, some of them not contained in
medean solids in turn with a view to finding the list by Badoureau. Then between 1930 and
regular polygons or regular stars on their facial 1932 Coxeter and Miller discovered twelve other
planes or cutting through the interior of these uniform polyhedra not previously known, but
solids. This is a different approach from that of publication was put off in the hope of obtaining
stellation. If such a polygon is found, it is evident a mathematical proof that there are no more.
that its vertices coincide with some of the ver- Independently M. S. Longuet-Higgins and H. C.
tices of the related convex polyhedron. The Longuet-Higgins between 1942 and 1944 re-
planes of these polygons may intersect. If por- discovered eleven of these twelve. These two
tions of the solid are removed symmetrically, teams learned of each others work in 1952.
another uniform polyhedron may result. This Meanwhile Lesavre and Mercier (1947) re-
process is called faceting, a sort of reverse of discovered five of the twelve. In Uniform poly-
stellating. Stellating implies the addition of cells hedra, published 1954, from which these facts
to a basic polyhedron which serves as a core. have been culled, the total now stands at
Faceting implies the removal of cells, so that the seventy-five uniform polyhedra. But here it is
basic polyhedron may still be imagined as a case admitted: 'it is the authors' belief that the
or enclosing web for the new one. If you examine enumeration is complete, although a rigorous
Kepler-Poinsot solids from this point of view proof has still to be given' (p. 402).
you will see that the small stellated dodeca- The method used by these recent investigators
hedron and the great dodecahedron can both be differs from that of the previous ones. It is based
derived by faceting an icosahedron. The vertices on a systerhatic investigation of all possible
of the former and the edges of the latter coincide Schwarz triangles as they apply to the poly-
respectively with the vertices and edges of an hedral kaleidoscope. Schwarz triangles are
icosahedron imagined as a case enclosing them. related to the Mobius triangles mentioned
If you examine the models you will see this very previously (see pp. 4-6).

98
21:37:12 BST 2016.
CBO9780511569746.018
 Examples of how metrical properties are found in
non-convex uniform polyhedra
Polygons found as faces: {3}, {4}, {5}, {6}, {8}, {10},
to which you must now add the three star polygons: {f}, {f}, ft-}

K3-V5)

>
[/
A A
\ (V5-2) one
unit
cf. 76

K3-V5)
\ /
Fig. 34

V
7 \ / / \\ i|(2-V2)
j K3V2-4)

> \ 7
X \ /
(3-2V2)
M3V2-4)
one
unit
cf. 77

\ /
A K2-V2)

(V5-2)
: M7-3V5)

X >4 "(V5-2)
; K7-3V5)
one
unit
cf. 81

• (V5-2)

Fig. 36
99
21:37:12 BST 2016.
CBO9780511569746.018
General instructions for making models of
non-convex uniform polyhedra

The non-convex uniform polyhedra are des- in the more intricate models that some further
cribed, each showing the facial planes required points on the edges will be needed, but again you
and the pattern of parts arising from the inter- will find that the golden ratio turns up once
section of the facial planes. No specific dimen- more in these smaller segments. Thus a careful
sions or measurements are given for any of these study of the drawings will enable you to make
drawings because there is a very simple relation- the models of any desired size.
ship which will give you the key to all of them. The facial planes of the non-convex uniform
It lies in the fact that the pentagram or five- polyhedra are not always entirely visible.
pointed star and the decagram or ten-pointed Sometimes some portion of the plane is hidden
star both exhibit the golden ratio in their dimen- in the interior of the solid, or a portion is visible
sions. Thus whenever these stars are found along as the upper part of the surface while another is
with other regular polygons these polygons have visible as the lower or reverse side of the same
their edges divided according to the golden ratio surface. The light shading is used to show the
r = 1-618 approximately. This is illustrated by upper surface, the dark the lower surface, while
way of the examples on p.99. The octagram or the invisible portion of the facial plane is left
eight-pointed star exhibits the famous \/2 in its unshaded. When the entire polygon is exterior
metrical properties, -y/2 = 1-414 approximately. it is left unshaded. The nets for constructing the
Thus once you have drawn these three stars models are derived from these visible and in-
accurately you have all the measurements in the visible portions.
line segments of the stars themselves. It is only

100
21:37:56 BST 2016.
CBO9780511569746.019
 67 Tetrahemihexahedron
This simple polyhedron is easily recognizable as
a faceted form of the octahedron. Topologically
it is the famous one-sided heptahedron, homo-
morphic with the one-sided surface named after
Steiner. (See Cundy and Rollett, p. 193.) In this
polyhedron three equatorial squares lie in three
perpendicular planes sharing their edges with
four triangles.
To construct a model of this polyhedron four
colours may be used. The equilateral triangles
may all be the same colour, say R. Cement the
isosceles right triangles to the edges of the equi-
lateral triangle as shown, and make four of these
units, all with the same colour arrangement.
These parts must then be given the form of tri-
angular pyramids, the R triangle serving as the
4{3} + 3{4} base and the Y, B, O triangles as slant sides.
V2 Now a special cementing technique must be em-
ployed. Some of the tabs must be turned out-
ward and cemented to form a tongue running
along the slant edges of the pyramid, while other
tabs are turned inward as usual but left un-
cemented to form a groove into which the
tongue tab of another part can be inserted. If
you remember while assembling the parts that
each square in the completed model must be the
same colour, you can join two appropriate
pyramids by applying cement on both sides of a
tongue tab before inserting it into the groove of
the other part. When you have done this you
should have two half-squares, whose planes

101
21:37:39 BST 2016.
CBO9780511569746.020
bisect each other at right angles along the line of tabs to join the cups together so that an edge of
the tongue and groove edge. The third pyramid one cup may be made to coincide with that of
may now similarly be cemented in place, and another. The four R triangles are then added
finally the fourth pyramid. You must exercise last, cementing one edge at a time and then
your own judgement on which tabs to use as closing it like a lid in the usual manner. You will
tongues and which as grooves. Once you see the find that the acute dihedral angles at the edges
model taking shape this will not be hard to do. make it an easy matter to cement these last
An alternative method of construction, also equilateral triangles, although your work on the
useful for other models, is to make four tri- cementing of the cups must be very accurate to
hedral cups as shown. All the tabs are turned make these last triangles fit well. From this point
outward to form ribs on the outside of the cup, of view the first method of construction is prob-
which is actually a triangular pyramid without ably better or easier to execute. Your own
its base. These ribs can then be properly trimmed experience will tell you which one you may
and manoeuvred to serve as double thickness prefer.

102
21:37:39 BST 2016.
CBO9780511569746.020
 68 Octahemioctahedron

This polyhedron is a faceted cuboctahedron,

also called an octatetrahedron. Four equatorial
hexagons share their edges with eight triangles.
Here again two methods of construction are
possible. You can make eight tetrahedra, the
first four having the colour arrangement set out
below.
12 3 4
B Y O G
B O Y R
B G O R
B Y G R

The other four also will all have a B triangle as

I 3|3 base, but they will have the enantiomorphous
8{3}+4{6} arrangement of side faces. Actually this amounts
2 to reversing only two colours, namely 2 and 3
become 3 and 2. These tetrahedra are then
joined to one another by the tongue and groove
technique. Again you must exercise your own
judgement to decide which tabs to turn in and
which to turn out. The completed model will
have all outer triangles B and the hexagon planes
Y, O, R, G. If you keep this in mind while you are
cementing the parts you should have no diffi-
culty in arranging them in the proper positions.
The alternative method of construction is to
make six tetrahedral cups, the first three in the
colour arrangement shown.

12 3 4
Y G R O
Y R O G
Y R G O

The other three are again enantiomorphous to

the first set of three. As before in this method of
construction all the tabs will be exterior to the
cups. Thus they can serve as double thickness
tabs, and be suitably trimmed and disposed so
that they can be cemented to join the cups along
common edges. Then all the B triangles are
added last of all.

103
21:37:39 BST 2016.
CBO9780511569746.020
69 Small cubicuboctahedron

This polyhedron is a faceted version of the

rhombicuboctahedron. The squares lie in the
facial planes of a cube, the octagons lie on
parallel planes below the squares, and the tri-
angles are the same as those of the rhombi-
cuboctahedron. Since the cube needs only three
colours, a very effective colour arrangement can
be obtained here by using the other two colours
for the triangles. Use the square, rectangle and
two triangles shown in the facial planes for nets.
Begin the construction of this model by making
four triangular pyramids as set out opposite.

\ 4|4
8{3}+6{4} + 6{8}

104
21:37:39 BST 2016.
CBO9780511569746.020
 12 3 4
R Y O B
G Y B O
R Y O B
G Y B O

Cement all the tabs outward on the slant edges to

form tongues. Next construct an open prism as
shown below. This is the upper part of the
model, so the B and O rectangles must be turned
down in the first part, and the other rectangles
treated in a similar fashion. The four pyramids
can now be cemented at the four corners of the
prism, the tabs between the B and O rectangles
forming the groove to receive the tongue tabs of
the pyramids. You must of course orient them
according to their appropriate colours, keeping
the Y triangles below. Now add a ring of four
more prisms. They are arranged as set out in the
second and third line of the colour table that
follows, two of each.

4
1 2 3 4 5
Y O B O B 5 1 3
O Y B Y B
B Y O Y O
2

The remaining four pyramids and three prisms

are identical in colour arrangement and the tech-
nique of assembly is the same. Actually you will
discover that the rectangles from the prisms can
simply be turned into place so that the tabs on
their edges easily make contact with the tongue
tabs of the pyramids without worrying about
the groove arrangement. The end result is equi-
valent to the groove but it is achieved in a
different way. This will become clear as the
model takes shape. This turns out to be a very
sturdy and attractive model.

105
21:37:39 BST 2016.
CBO9780511569746.020
70 Small ditrigonal
icosidodecahedron

This polyhedron has twelve pentagrams in the

same facial planes as the dodecahedron and
twenty triangles in those of the icosahedron. As
is evident from the vertex figure they meet in
alternate sets of three around each vertex of the
polyhedron. So it can be called a ditrigonal
icosidodecahedron. The pentagrams may be
done in six colour pairs, but to preserve the map
colouring principle you will have to use the
second or alternative icosahedral arrangement
for the triangular planes. This can be achieved
most easily by working around each star,
cementing the smaller triangular pairs as di-
hedral grooves between the star arms. The
arrangement and colour table are set out opposite. 3|3f
As soon as you have completed the grooves 20{3} + 12{f}
around the W star, you can immediately add the
next five coloured stars. To get these correctly
placed, put the G star opposite the Y triangle,
and so on around with Y, B, O, R, G.

106
21:37:39 BST 2016.
CBO9780511569746.020
 Star 01 23 45 61 89
W GY RG OR BO YB
Y GB OG GO RG BR
B YO RY YR GY OG
O BR GB BG YB RY
R OG YO OY BO GB
G RY BR RB OR YO

In the colour table it is evident that each colour

pair is named twice. This makes it simpler to
follow, because you can actually work in that
order, surrounding each star with the dihedral
grooves. Undoubtedly you have by now ob-
served the cyclic permutation of colours. This is
also evident in the colour table if, disregarding
the first line, you read down each column
successively.
Once this much of the model has been com-
pleted you should have no more difficulty. The
remaining pentagrams should be placed in
opposite colour pairs and the grooves are all
determined by watching the triangular planes to
see that they are kept in their colour sequence.
You will find it easier to cement these triangle
pairs together first, and then to cement them
between the star arms. All the remaining tri-
angles can be completed in this way. Thus a W
star is cemented last of all.
This is best done in stages; cement only one
tab first and let this set up firmly, then work on
one edge at a time. As the openings at the edges
begin to narrow, apply the cement carefully and
work it in with the probing needle. Deft fingers
and a little patience will do the rest.

107
21:37:39 BST 2016.
CBO9780511569746.020
71 Small icosicosidodecahedron
In this polyhedron the twenty triangles are on
facial planes above and parallel to twenty hexa-
gons. This means that the first icosahedral
colour arrangement will do very well here with
its five colours, parallel planes being the same
colour, leaving white for all the stars. The tri-
angles have edges in common with the hexagons,
and vertices in common with the stars. This
leaves grooves between the star arms, each
groove being formed of two trapezia coming
from intersecting hexagon planes which suggests
the following method of assembly. Make five
grooves as illustrated below following the (0) line
in the colour table. These are cemented so that
they radiate outward and downward between the
W star arms. Triangles are then cemented between 3f|3
the grooves, each colour being determined by the 20{3}+20{6} + 12{f}
hexagon plane below it. In the colour table each V5
groove is mentioned twice, but again this is easier

12 12 12 12 12
(0) Y G BY OB RO GR
(1) YG RY BR OB GO
(2) BY GB OG RO YR
(3) OB YO RY GR BG
(4) RO BR GB BG OY
(5) GR OG YO BY RB

21:37:39 BST 2016.

CBO9780511569746.020
to follow from the point of view of construction. plete, since each star with its ten edges has ten
If you work systematically around each star the trapezia joined to it, that is, five grooves. The
colour arrangements are easy to follow and you secret is not to try to do too much at once. One
can cement the grooves in place as you deter- edge cemented at a time is the best rule to
mine their colour and position. The grooves and follow. If you keep five clamps handy, you can
triangles thus help to determine each other as far keep moving them around and by the time you
as colour is concerned. You will find that opposite finish cementing the tabs at a fifth edge you can
grooves on the model are enantiomorphs, but remove the clamp from the first edge. Also it is
in cementing the trapezia pairs it does not generally easier to cement the longer edges of the
matter too much. Turning the pairs end for end trapezia first to the triangles which they sur-
will reverse the colours as required. round and then the shorter edges or ends of the
This model requires a bit of patience to com- grooves to the star arms.

109
21:37:39 BST 2016.
CBO9780511569746.020
72 dodecahedron
Small dodecicosi-

This polyhedron is easily recognizable as a

faceted rhombicosidodecahedron. The six colour
dodecahedral arrangement can be used for the
pentagons and the decagons which lie on
parallel planes one above the other. Thus the
procedure for constructing this model is to begin
with a W pentagon to which five trapezia are
cemented as shown. Turn the tabs on these
trapezia inward but leave them uncemented to
serve as grooves into which the tongue tabs of
the small triangular pyramids must be inserted.
The colour tables for both these parts are set out
below.
You must now follow the second icosahedral
arrangement to place the triangular pyramids in f 5|5
their correct positions. The colour tables below 20{3}+ 12(5}+ 12(10}
give only half the required parts, but again the V(H+V5)
rest are all enantiomorphous to these and they
each take their positions diametrically opposite
their counterparts. In this way the map colouring
principle is preserved.

0 12 3 4 5
(0) W Y B O R G
(1) Y W G O R B
(2) B W Y R G O
(3) O W B G Y R
(4) R W O Y B G
(5) G W R B O Y

1 2 3
(1) Y W O R
(2) B W R G
(3) O W G Y
(4) R W Y B
(5) G W B O
(6) Y GOB
(7) B Y R O
(8) O B G R
(9) R O Y G
(10) G R B Y

110
21:37:39 BST 2016.
CBO9780511569746.020
 111
21:37:39 BST 2016.
CBO9780511569746.020
73 Dodecadodecahedron

This polyhedron has twelve stars in the same

facial planes as the dodecahedron, but on parallel
planes below these stars are twelve pentagons
each sharing edges with five surrounding stars
and intersecting each other. To construct a
model of this polyhedron, begin by cementing a
set of five trihedral dimples between the arms of
a W star. The arrangement is shown in the
colour table.
The first set of five dimples are cemented so that
the rhombus 1 is below the W star, forming a
white pentagon plane parallel with the star. The
other colours begin the next five intersecting
pentagon planes so that two rhombi appear with
the same colour along the straight line between
alternate star arms. The next five stars may now 211 5 - g)
be cemented in place, their colours being deter-
mined by the two rhombi just mentioned. The 12{f} + 12{5}
next set of five dimples are then added, (6) below 2
(1), and so on for the rest. Enantiomorphism
applies for the remaining parts.

1 2 3
(1) W Y G
(2) w B Y
(3) w O B
(4) w R O
(5) w G R
(6) o G Y
CO R Y B
(8) G B O
(9) Y O R
10) B R G

112
21:37:39 BST 2016.
CBO9780511569746.020
 74 Small rhombidodecahedron
This polyhedron is another version of the
rhombicosidodecahedron. Here the pentagons
are removed giving place to shallow pentagonal
cups whose bottoms, also pentagons, belong to
the decagon planes. The triangles, as well, are
removed leaving shallow dimples whose faces
are also part of the decagon planes. The squares
however, are retained. The dodecahedral colour
arrangement works well for the decagon planes
and suggests the following method of assembly
Begin with a W pentagon and cement five
trapezia to it forming a shallow pentagonal cup
as shown. You can follow the same colour table
as that used for the small stellated dodeca-
hedron. All twelve cups can be cemented to-
25f gether to form an interior dodecahedron This
30{4} +12(10} leaves the spaces between the trapezia to be
V01+V5) filled alternately with squares and shallow tri-
hedral dimples. The colour tables set out below
give a remarkable result. Cement the Y square
between the W and R pentagons, with the other
squares following round in order, the B square
between the W and G pentagon, and so on. This
gives a ring of squares at the top in the usual

Pentagons Squares
(0) W Y B O R G
(1) R Y B G W O
(2) G B O Y W R
(3) Y O R B WG
(4) B R G O WY
(5) O G Y R WB

113
21:37:39 BST 2016.
CBO9780511569746.020
You will note that the rings have squares in
common, so each square is listed twice in the
table. Again enantiomorphism applies to the
remaining rings of squares. The remarkable
result mentioned above can now be observed.
The five W squares alternate with the other five
coloured squares as an equatorial skew band
when the polyhedron is held so that the two W
decagons are at the poles. The same relationship
holds for each of the six colours.
The trihedral dimples have the same colour
arrangement and even the same shape as those
of the great dodecahedron. However, because of
the shallowness it is advisable to eliminate the
tabs altogether by cementing the tab of one tri-
angle directly to the under surface of another as
shown. Their positions are not hard to locate on
the model.

114
21:37:39 BST 2016.
CBO9780511569746.020
 75 Truncated great
dodecahedron

The same colour arrangement may be used here

as for 21. The stars and decagons being on
parallel planes should be the same colour. The
best method for constructing this model is
simply to make the trihedral dimples following
the colour table for the great dodecahedron.
Cement these dimples together along their re-
maining long edges and add stars at the short
edges as required. The arrangement for one
dimple is shown.

2 f | 5 = t{5, f}
12{f} +12(10}

r-
/17 + 5V5

10

115
21:37:39 BST 2016.
CBO9780511569746.020
76 Rhombidodecadodecahedron
This lovely polyhedron is almost as spherical as
a beach ball and with the arrangement suggested
here is equally as colourful. Its name suggests
its relation to earlier models. This model has a
great number of parts, a total of 312, to cut,
trim, and cement. The usual technique of pro-
viding all the parts with tabs all around will
produce a fairly good model, if it is not too
large. The squares can be arranged as in 74, but
here the planes intersect each other so that the
skew band is all the more delightful. An arrange-
ment of parts and a colour table are set out
opposite to help you get started.
With these parts cemented you will have no
trouble finding the right colours for the smallest
parts belonging to the pentagon planes. These are
cemented in place at once. Continue to work on
12(1}+ 30(4}+ 12(5}
the rest of the model and complete it except for
the small triangular holes that will be left. These
holes are closed with small shallow trihedral
dimples whose three triangles come from the
square planes. Again the colours for these are
now not hard to determine. One extra word of
advice. It is usually easier to cement concave
parts together first, then to cement them to the
model as you near the end of your work. The
triangles 11-15 thus become folded rhombi in two
colours and serve as connectors between sections.

116
21:37:39 BST 2016.
CBO9780511569746.020
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
W Y G B Y O B R O G R O R G Y B
Y W B G W O G R O B R O R B W G
B W O Y W R Y G R O G R G O W Y
O W R B W G B Y G R Y G Y R W B
R W G O W Y O B Y G B Y B G W O
G W Y R W B R O B Y O B O Y W R

117
21:37:39 BST 2016.
CBO9780511569746.020
77 Great cubicuboctahedron
This polyhedron is a faceted cube. Each octa-
gram lies on the face of a cube which you may
imagine as enclosing the polyhedron. Then each
corner gets a tetrahedral dimple and each edge
gets a dihedral groove. The dimples and grooves
alternate between the star arms. Since the cube
can be done with three colours the octagrams
can follow the same arrangement. Six stars are
thus paired and six squares get the same colours
because they are parallel to the stars and below
them. The triangles can then use the other two
colours alternately.
Begin constructing this model with a Y octa-
gram. Then make four dimples as shown, also
four triangle pairs. Note that these triangles are
slightly larger than those used in the dimples. 3 4|f
Cement the dimples and grooves alternately 8{3} + 6{4}+6{f}
between the star arms, seeing that the colours
run on their respective planes.
With this done you should have no further
difficulty. The colours are easily determined for
the stars, the second Y star being cemented last,
one edge at a time.

0 1 2 3 1 2
G Y B O R G
R Y O B G R
G Y B O R G
R Y O B G R

/ \

\ /
\ /

118
21:37:39 BST 2016.
CBO9780511569746.020
 119
21:37:39 BST 2016.
CBO9780511569746.020
78 Cubohemioctahedron

This polyhedron is another faceted version of

the cuboctahedron and thus the same construct-
ion methods are applicable here as in 68. Prepare
your nets from the facial planes. You may make
six pyramids with square bases, three as set out
below and the other three their enantiomorphs.
Assembly of these pyramids is by the tongue and
groove technique.
The alternative method of assembly has eight
trihedral cups, the first four arranged as shown
and the other four as their enantiomorphs. As
usual all tabs are turned outward, then the
double tab of one cup is cemented to the under-
surface of another to make the edges coincide.
The squares are added last to complete the
model.

0 1 2 3 4
B Y O R G
B Y R O G
B Y R G O

1 2 3
R Y O
O G Y 1/ 2 \ 3
R G O
Y G R

120
21:37:39 BST 2016.
CBO9780511569746.020
 79 Cuboctatruncated
cuboctahedron

This polyhedron is a faceted octahedron. The

dihedral grooves between the star arms are parts
of hexagon planes. The relationship to the octa-
hedron suggests a very effective colour arrange-
ment. The eight hexagons can be parallel pairs
in four colours, and all six octagrams are a
fifth colour. The stars may all be white or
the same colour as the octagrams since they are
on planes parallel to and above the octagram
planes. To construct a model of this polyhedron,
begin by surrounding triangles, the central part
of the hexagon, with the parts coming from the
octagon planes as shown. The shortest edges of
these jsurrounding parts should then be cemented
3 f 4| to form a shallow cup. Next the grooves are
8{6} + 6{8} + 6{f} made as shown and these two parts cemented
alternately around an octagram.

0 1 2 3
Y G G G
B G G G
O G G G
R G G G R O

The next four octagrams can be added immedi-

ately. After this four more grooves form an
equatorial band connecting these octagrams.
Their colours are

4 5
O Y
R B
Y O
B R

You will have noted that the last two of these

pairs are enantiomorphs of the first two. They
appear of course diametrically opposite on the
model. The next set of four shallow cups and
their four connecting grooves may then be

y
W 121
21:37:39 BST 2016.
CBO9780511569746.020
cemented in place. The cups are each opposite or left because turning this part end for end will
their respective colours and the grooves opposite give you the desired order. A final octagram then
their enantiomorphs. In cementing the pairs that completes the model. This makes an attractive
form the grooves you need not worry about right and rigid polyhedron.

122
21:37:39 BST 2016.
CBO9780511569746.020
 80 Ditrigonal dodecahedron

This polyhedron is especially interesting because

of its close relationship to the great stellated
dodecahedron. It may be thought of as this
latter solid with the embossed stars of the great
dodecahedron removed from their pentagonal
planes, then turned over and set down to make
the vertices of these embossed stars coincide
with the points between five vertex parts of the
great stellated dodecahedron 22, the edges
being shared in common. This sounds a bit com-
plicated in words, but once the model is com-
pleted you will readily see it for yourself. This
model can also be thought of as a faceted version
of 70. The dihedral triangular grooves of 70 are
here removed to give place to deep holes between
3|f5 the star arms. These holes are formed by the
|} + 12{5} intersecting pentagon planes, and thus their faces
V3 are the familiar 72°, 36° and 36° 108° isosceles
triangles. The relationship of this polyhedron to
the two stellations of the dodecahedron means
that the same colour arrangement is suitable
here.
To make a model of this polyhedron, prepare
twelve pentagrams, two of each of the six
colours. Next prepare a set of five hexahedral
cups, the holes described above, whose parts
are shown on p. 124. A complete colour table
for two more sets of five is set out on the same
page.
Turn the tabs to the outside of these cups, which
are then cemented between the star arms. The
first set of five cups will surround the W penta-
gram. After these are in place add the five
coloured pentagrams. The Y star must lie on a
plane above the Y pentagon, which shows only
two parts of its area in the cups. These are the

123
21:37:39 BST 2016.
CBO9780511569746.020
triangles 1 or 2. Once you locate the correct 1 2 3 4 5 6
position for this first star the others follow (1) B G O R O R
around in order as Y, B, O, R, G. Then the (2) O Y R G R G
second set of five cups can be prepared and used (3) R B G Y G Y
as connectors. Cup (6) goes between the G and (4) G O Y B Y B
Y star, (7) between the Y and B star, and so on (5) Y R B O B O
around in order. The third set of five cups will (6) W O R B R B
complete half the model. Their positions are (7) W R G O G O
easily found by watching the pentagon planes to (8) W G Y R Y R
see that they keep their respective colours. The (9) W Y B G B G
second half of the model is enantiomorphous to (10) W B O Y O Y
the first, the parts being diametrically opposite (11) G R B W B W
their counterparts. The W star is cemented last. (12) Y G O W O W
This makes a very sturdy model, but as you (13) B Y R W R W
approach the end of your work it requires very (14) O B G W G W
careful cementing to make the stars fit well. (15) R O Y W Y W

124
21:37:39 BST 2016.
CBO9780511569746.020
 81 Great ditrigonal dodec-
icosidodecahedron

This polyhedron is a member of the icosidodeca-

hedral family. This suggests the usual colour
arrangement, six colours in opposite pairs for
the decagrams and the pentagon planes parallel
to them, and the second five-colour icosahedral
arrangement for the triangles, making opposite
triangular planes the same colour.
* *
The simplest method of assembly for a model
of this polyhedron is to make the dimples and
grooves and cement them alternately between
the star arms. Start with a W decagram cement-
ing the parts as shown, and continue in the
same way to complete the model.

1 2 3 4 5 6
Y W o R B Y
B W R G O B
17-3V5 O Y
wG R O
R wY B G R
G wB O Y G

125
21:37:39 BST 2016.
CBO9780511569746.020
82 Small ditrigonal dodec-
icosidodecahedron

This is a most remarkable polyhedron because it

was discovered in our own century and pub-
lished for the first time in 1954 (see Coxeter, op.
cit.), yet it is closely related to the stellated
dodecahedra. From this point of view it is
strange that it should have been missed by the
earlier investigators. But then the stellated
dodecahedra themselves were missed until
Kepler's time. This polyhedron can be thought
of as a truncated form of the great stellated
dodecahedron with the embossed or solid stars
of the great dodecahedron taken from their
planes and turned over here to fill the spaces
between the stumps remaining after truncation.
This relationship is helpful in making a model of
this polyhedron.
3 115
20(3} + 12(1} + 12(10}
Make twenty truncated pyramids as shown.
/17 + V 5
The side faces follow the colour arrangement of
2
the great stellated dodecahedron. Leave the tabs
at the slant edges uncemented but turned in to
serve as grooves into which tongue tabs are in-
serted. The triangles follow the first icosahedral
arrangement.

1 2 3 4
Y Y G B
B B Y O
O O B R
R R O G
G G R Y

You will at once see that the map colouring

principle is abandoned here, but only in the first
ring of five parts. Also once you have cemented
these parts as a ring, you will see the position at
their centre for the first solid star part. These
solid stars are made following the colour ar-
rangement of the great dodecahedron. Cement
one short edge of the 36°, 108° triangles between
the star arms of the pentagram, but turn the tabs
on the longer edges out to form tongue tabs.
The other short edges are turned in and cemented
to each other, thus forming a sort of star pyramid

126
21:37:39 BST 2016.
CBO9780511569746.020
with concave edges as usual, but with ribs on the
outside at the convex edges. These solid stars
can now be easily cemented into their respective
places by inserting the five tongue tabs into the
five grooves made by the tabs of the trapezia 2,
3, 4 shown opposite. The W pentagram, call it
the base of the solid star pyramid, should lie on a
plane above and parallel to the W decagon,
whose parts appear alternately with upper and
then lower surface visible. The procedure will
become apparent as the work proceeds. This
makes a very sturdy, rigid, and interesting
model.

127
21:37:39 BST 2016.
CBO9780511569746.020
83 Icosidodecadodecahedron

This polyhedron is very much like 76 but the

squares there are replaced with hexagons here,
giving it a more fascinatingly dimpled character
and also adding greater rigidity and beauty.
Once you begin to make this model you will be
surprised to see how easily it can be assembled
despite its complicated structure.
To simplify construction, the procedure has
been broken down into four steps or parts.
Solid stars, called part i, are built first. Begin
with a regular pentagram and fill the spaces
between the star arms with triangles belonging
to the hexagon plane to make a low inverted star
pyramid. The longest tabs on these triangles are
turned out to form ribs along the slant edges
leading from the star points to the central vertex. |5|3
20{6} 12{f} + 12(5}
V7

128
21:37:39 BST 2016.
CBO9780511569746.020
The figure shows a plan of this part and a colour
table for six such parts.
Part i

Star 1 2 3 4 5 6 7 8 9 10
(0) W Y G B Y O B R O G R
(1) Y B G Y B R Y O R G O
(2) B O Y B O G B R G Y R
(3) O R B O R Y O G Y B G
(4) R G O R G B R Y B O Y
(5)G Y R G Y O G B O R B

Part II is a sort of wedge or open trihedral scoop.

The sides belong to the hexagon planes and the
central triangle comes from the pentagon plane.

The colour table follows: Part II

1 2 1 2 1 2 1 2 1 2 3
(0) G O Y R B G O Y R B W
(1) Y G R B O Y G R B O Y
(2) B Y G O R B Y G O R B
( 3 ) O B Y R G O B Y R G O
( 4 ) R O B G Y R O B G Y R
( 5 ) G R O Y B G R O Y B G
Part II is cemented so that it shares its acute
Part HI Part iv
vertex with the central point of the solid star and
its longer edges run along the ribs of the solid
star filling the spaces between the star arms. In
each case the first-named colour pair of part n
goes between sections 7, 8 of part i and the rest
follow round in order. If you use clamps, part n to experiment on your own with more colours to
can easily be kept in place. In effect you have get a better effect, although the arrangement
here the tongue and groove technique but described above is very satisfactory. Only a
achieved in a different manner. It is this different trained eye would notice the colour defect, if it
method of assembly which makes this model may be so called. This model takes a great deal of
surprisingly easy. Parts i and n make a section. time to assemble because of its many parts.
Twelve sections complete the model. These Again you may wish to experiment on your own
sections are joined together using part in as con- by making other nets combining more of the
nectors. These are the same colour pairs as in 76, parts to save labour. You will have to be the
and their position is now easy to determine. judge of whether the result is more pleasing.
Finally twenty more small trihedral dimples, Undoubtedly this model is one of the most satis-
part iv, are needed to close the small triangular factory of the whole set of uniform polyhedra
openings at the base of part n. Their colour from the point of view of being easy to make yet
arrangement also is easily determined from the being complex in appearance. It is easy to make
hexagon planes. from the point of view of assembly, not of time
You will notice that the map colouring prin- involved. It turns out also to be a very rigid
ciple has again been abandoned. You may wish model.

129
21:37:39 BST 2016.
CBO9780511569746.020
84 Icosidodecatruncated
icosidodecahedron

This polyhedron is to the icosahedral group,

what 79 is to the octahedral. The construction
technique remains the same. To make this
model, build shallow triangular cups and di-
hedral grooves and cement these alternately
between the star arms of the decagrams. A suit-
able colour scheme is given opposite.
The first five of these cups surround the W deca-
gram, the paired dihedral grooves connecting
adjacent cups.
The first set of five grooves extends radially
downward from between the star arms. The next
set of five can now be cemented along their
edges, the Y star above the Y decagon plane
which at this stage of construction shows only 3 1 5[
two of its five parts. The other stars follow 20{6} +12(10} + 12{J-3a}
around in the usual order. Now the third set of 4
shallow cups may be added, (6) below (1), and so
on. The colours of the second set of grooves are
determined by the triangles at the bottom of the
cups. So too with the third set of grooves. Again
the map colouring principle has been over-
10
looked, but this must happen when the dodeca-
hedral and icosahedral colour arrangements are
both used in one model, while keeping to only
six colours. Here the second icosahedral arrange-
ment is used. To complete the model make the
remaining parts enantiomorphous to those tabu-
lated. The numerous edges around each deca-
gram require perseverance in cementing the tabs.
Also these decagrams may need some backing, to
make them stiffer, when the scale of the model is
fairly large, in which case the parts are also
easier to cement.

130
21:37:39 BST 2016.
CBO9780511569746.020
 1 2 3 4

A/A
(1) Y W O R / \
(2)
(3)
B
O
W R G
wG Y
\
// \\
A
(4)
(5)
R
G
wY B
wB o
/ // \\j
\
(6) B
0
Y R O
B G R
\ 2 >J
(7)
(8) R O Y G
(9) G R B Y
(10) Y G O B

1 2 f
Y G / 1 \

z
(1)
(2)
(3)
(4)
(5)
B
O
R
G
Y
B
O
R
\\ 2 /
\
(6) O B A
(7) R O / \
(8) G R
(9) Y G / \ .
(10) B Y / \ <
R B
(11)
(12) G O
/
/
\
(13) Y R
\
(14) B G
/ \
(15) O Y /
/ \
\

131
21:37:39 BST 2016.
CBO9780511569746.020
85 Quasirhombicuboctahedron

This polyhedron is very similar to 77. The octa-

grams of that polyhedron are removed here with
only the edges retained making it deeply
dimpled and cupped. It has two different sets of
intersecting squares. If you follow the colour
arrangement of 77 for the triangles and corner
dimples used there, you will again get an effective
arrangement. However, with more squares in
this model, you cannot preserve the map colour-
ing principle with only five or six colours. In fact
the many parts required here call for a great deal
of perseverance to complete a model in colour.
You may therefore find it more satisfactory to
make a model in one colour first, to see what
colour possibilities you can work out for your-
self. | 4 | 2 = r'{|}
Suitable nets, combining many parts in one, 8{3} + (12 + 6){4}
for such a model are given opposite. The letters V(5-2V2)
identify the individual nets shown separately at
full working scale. You will need six of part i,
the central portion of the faceted octagram.
Leave tabs all round. The broken lines should be
scored on the reverse side and the folding along
these lines is up instead of down. Cut twenty-four
star arms, part n. You will also need eight of
part in, the dimpled corner portions, and twelve
of part iv. Assemble parts i and n first and then
use parts in and iv as connectors.

X X
\
R / \ X
•IK \ Hill
R / X
\ /
X X

132
21:37:39 BST 2016.
CBO9780511569746.020
Remove

Remove

Part
133
21:37:39 BST 2016.
CBO9780511569746.020
 86 Small rhombihexahedron

This polyhedron is another faceted rhombi-

cuboctahedron very much like 69. The triangles
found there are here removed and the set of
squares found there, are here replaced with a
different set of squares. You can use the same
set of open prisms in constructing this model as
you used in 69. The triangular pyramids there
will now give their sides to trihedral dimples
with the same colour arrangement. The square
prisms are cemented first to form an interior
cube; then the set of squares, which may all be
one colour either R or G, are cemented along
opposite sides to the edges of the prisms.
Finally the dimples are cemented in place. You
must of course see to it that the double tabs
below the dimples are properly trimmed and ad- 2 4
justed so they do not jam. This makes a very
12{4} + 6{8}
rigid model.

!'£S%"*"l%"£§?$ i i •§•||vC"ll•!l•l^^•••••••••••••^^^v.^^^v.^^^^^^^v.^^^^^^^^^^^^^^•.•.•.•.•.-.•.-.•.•.
' ' • " ^ • ' ' • ' '

134
21:37:39 BST 2016.
CBO9780511569746.020
 87 Great ditrigonal
icosidodecahedron

This polyhedron is another ditrigonal icosido-

decahedron like 70. It differs from 70 in that here
the pentagrams are removed and pentagons
replace them on a parallel plane closer to the
centre of the solid. The twenty triangular planes
have exterior parts made up of two sizes of
equilateral triangles, and the twelve pentagon
planes have the usual 36°, 72° and 36°, 108°
isosceles triangles. The colour arrangement can
also follow that used in 70.
To make a model of this polyhedron begin by
making a pentahedral cup or inverted pyramid,
the same as that used for the small stellated
dodecahedron, but turned inwards so the tabs
f|3 5 form ribs on the outside of the cup. The same
20{3} + 12{5} colour scheme works here but turning the parts
V3 inside out, you may note, is equivalent to build-
ing the enantiomorphous arrangement. Next it
will be best to prepare a set of five trihedral
dimples, following the colour table on p. 136.
These dimples are cemented to the edges or lip of
the cup. The structure is a bit unstable at this
stage, but if you press this whole section as com-
pleted so far against a pane of glass and look
through from the other side you will see a per-
fectly faceted five pointed star. Pairs of equi-
lateral triangles must next be cemented between
the star arms, just as in 70. These begin to give the
model some rigidity, but it will be advisable to
add more backing inside the star along the edges
from point to point. If you make the model
without this, you may find these edges are
slightly out of line.
The colour arrangement for six faceted stars is
shown in the figure and table on p. 136.
The procedure is to continue making the faceted
stars and to use the triangle pairs as connectors,
as in 70. Again, the second half of the model is
made enantiomorphously. The fact that the map
colouring principle is violated is not too notice-
able in the completed model. Also it may be
easier to follow the model itself as it grows
rather than the colour table.

135
21:37:39 BST 2016.
CBO9780511569746.020
 Cups
1 2 3 4 5
(0) Y B O R G
(1) W G O R B
(2) W Y R G O
(3) W B G Y R
Cup (4) W O Y B G
(5) W R B O Y

Star arm dimples

6 7 8 6 7 8
(0) B Y G (3) B Y R
O B Y G R W
R O B Y Y B
G R O R G G
Y G R W O Y
Star arm dimple Triangle pair (1) Q R B (4) O B G
O B W Y G W
R R G B B O
B O O G Y Y
W Y R W R B
(2) Y G O (5) R O Y
R O W B Y W
G G Y O O R
O R R Y B B
W B G W G 0

136
21:37:39 BST 2016.
CBO9780511569746.020
 88 Great icosicosidodecahedron
This polyhedron is related to 81, the difference
being that here the decagrams are gone and the
edges alone are retained, while the addition of
hexagons introduces multifaceted decagrams in
place of plane decagrams. And multifaceted is
literally true! To make a model of this poly-
hedron you will have to prepare 76 parts for
each faceted decagram alone, not to mention the
other parts which serve as connectors. It may
interest you to know that the total number of
individual small segments of surface area gener-
ated by all the intersections of the three regular
polygons belonging to the facial planes of this
polyhedron reaches the imposing figure of 1232.
This is a real challenge to the perseverance of
|5|3 any model maker! Because some of the parts are
20(3}+ 20(6}+ 12(5} so small, the model must be on a scale suffi-
/17-V5 ciently large to enable you to handle them. Also
~ 2

21:37:39 BST 2016.

CBO9780511569746.020
some stiffening is required along the internal 2 and 6 of part in are cemented to 1 and 2 of
decagram edges to keep them straight. The part ii. A set of small trihedral dimples will now
various parts are set out in the figures and a close the spaces between the star arm pairs and
complete colour table is given. You will find it the small triangle mentioned in part I. This last
best to keep all pentagon planes W and to use set of dimples is part iv. This completes one
the five colours Y, B, O, R, G for the hexagons faceted decagram. Twelve of these are needed
and triangles. and are joined together with the same dimples
Begin with the cup shown in part I. This shows and grooves as in 81, except that here the penta-
the small triangles needed between 1 and 2, 2 and gon planes are all W.
3, etc. Part n is a deep cup with a pair of very The rest of the model again follows enantio-
small triangles at the bottom, which is cemented morphically. All you need is about 30 hours.
so the edge of the 0 portion is placed with the Once you get working systematically each face-
edge of 1 2 3 4 5 in part I. The star-arm dimples ted decagram may take about 2 hours, a total of
are shown as part in in enantiomorphic forms 24 hours for all, and then another 6 hours to get
since their colour arrangement is not enantio- all parts joined with the paired triangular
morphous when they are adjacent. The edges of grooves and corner dimples.

Part i Part II

138
21:37:39 BST 2016.
CBO9780511569746.020
 Part I Part II Part in Part iv
0 1 2 3 4 5 0 1 2 3 4 1 2 3 4 5 6 7 8 1 2 3
(0) W Y B O R G W G Y Y B W Y R O W G Y O O G Y
W Y B B O W B G R WY B R R Y B
w B O O R W O Y G wB O G G B O
w o R R G W R B Y wO R Y Y O R
w R G G Y W G O B wR G B B R G
(1) W B Y R O G w G B B Y W B O R wG B R R G B
w Y O O B W O R G wY O B G Y O
w B R R O W R G Y wB R Y Y B R
w 0 G G R wG B Y wO G B B O G
w R Y Y G wY B O wR Y O O R Y
(2) W O B G R Y w Y O O B wO R G wY O G G Y O
w B R R O wR G Y wB R Y Y B R
w O G G R wG Y B wO G B B O G
w R Y Y G wY B O wR Y O O R Y
w G B B Y wB O R wG B R R G B

(3) W R O Y G B w B R R O wR G Y wB R Y Y B R
w O G G R wG Y B wO G B B O G
w R Y Y G wY B O wR Y O O R Y
w G B B Y wB O R wG B R R G B
w Y O O B wO R G wY O G G Y O

(4) W G R B Y O w O G G R wG Y B wO G B B O B
w R Y Y G wY B O wR Y O O R Y
w G B B Y wB O R wG B R R G B
w Y O O B wO R G wY O G B Y O
w B R R O wR G Y wB R Y Y B R

(5) w Y G O B R w R Y Y G wY B O wR Y O O R Y
w G B B Y wB O R wG B R R G B
w Y O 0 B wO R G wY O G B Y O
w B R R O wR G Y wB R Y Y B R
w O G G R wG Y B wO G B B O G

Part in Part iv

139
21:37:39 BST 2016.
CBO9780511569746.020
89 Small icosihemidodecahedron
This polyhedron and 91 are both faceted versions
of the icosidodecahedron. The decagons cut
right through the centre of the solid on equatorial
planes. This polyhedron has deep pentahedral
cups or inverted pyramids all of which have
their apex at the centre of the solid. As for colour
arrangements, each of the decagons can have a
colour of its own W, Y, B, O, R, G. Then the
triangles should follow the usual icosahedral
order, namely the second alternative. Again you
have a choice of two techniques for construction;
the tongue and groove arrangement for twenty
triangular pyramids, each of which has the
colour arrangement of the great stellated dodeca-
hedron; or the double tabs turned outward as
ribs on twelve pentahedral pyramids, as in the |3|5
small stellated dodecahedron. The triangles are 20{3} + 6{10}
added last of all in this latter method, which is IT
probably the easier to execute. If you try both
you can see for yourself which gives the best
result. This makes a very rigid model.

140
21:37:39 BST 2016.
CBO9780511569746.020
 90 Small dodecicosahedron

This polyhedron forms a pair with 82 in that it

was also published for the first time in 1954.
(See Coxeter, op. cit.) Faceted stars are again
found here, formed by the intersection of only
two types of planes, hexagons and decagons. The
first icosahedral arrangement of colours may be
used effectively for the twenty hexagon planes
and the dodecahedral arrangement for the deca-
gons. The method of assembly for this model is
to make the faceted stars whose central facets
follow the order of the six colours used in 20, but
forming cups or inverted pyramids. The star
arms have dimples made of small equilateral
triangles belonging to the hexagons and two
isosceles triangles belonging to the decagons.
You can get these dimples arranged correctly by
3 5
following the colours from the cups outward to
20{6} +12(10} the star arms, only you must give the equilateral

J'-/17 + 3V5 triangles a turn counterclockwise two places to

begin with, as shown below. Pairs of trapezia
form dihedral grooves between the star arms
with the usual arrangement. The shallow tri-
angular cups have their vertices in common with
the vertices of the star arms and their side faces
continue the facial planes of the side faces of the
star arm dimples, so the colours are not hard to
match. In the next set of five faceted stars you
will find the Y hexagon sharing one of its edges
with a Y decagon, and so on around for the
other colours. This defect in the map colouring
principle again does not seriously affect the end
result.
The rest are done in the usual cyclic permutation
of colours, with opposite parts enantiomor-
phous. You should now be able to complete the
model without further directions. The faceted
stars are not entirely rigid and so some internal
stiffening may be required. However, this is not
needed if the model is small.

1 2 3 4 5 6 7 8 9 10
(0) Y B O R G R G Y B O
(1) W Y R G O R Y O B G
(2) (etc., in cyclic order down each column.)

141
21:37:39 BST 2016.
CBO9780511569746.020
142
21:37:39 BST 2016.
CBO9780511569746.020
 91 dodecahedron
Small dodecahemi-

The relation of this polyhedron to 89 and to the

icosidodecahedron has already been mentioned.
The cups or inverted pyramids are trihedral
holes with vertices at the centre of the polyhedron.
The double tab or ribbed technique is the best to
follow in making the model. Twenty triangular
pyramids, identical to those of the great stellated
dodecahedron 22 must be made with all double
tabs turned outward to form ribs. The first ring
of five pyramids can then be cemented together
and a W pentagon added to serve as a base. This
is actually the exterior surface of the completed
solid. The pentagons are on planes parallel to
the equatorial decagons and so should be the
|5|5 same colour. If you keep 89 in front of you as
12{5}+6{10} you work and remember that the triangles there
IT are replaced here by the pentagons you should
find it easy to complete the model. Again this
makes a very rigid model.

143
21:37:39 BST 2016.
CBO9780511569746.020
92 Quasitruncated hexahedron
This polyhedron is a quasitruncated cube. Six
octagrams lie on the facial planes of an interior
cube sharing their edges with the eight triangles
whose planes intersect the cube. By using three
colours for the octagram pairs and the other two
colours for the triangles you can achieve a very
suitable arrangement. To make a model of this
polyhedron assemble the parts as shown.
Part i forms a cup with four pointed side faces
and a square bottom. Part n forms a four-sided
box, open at both ends, the bottom end straight
and the top end jagged. Part I is cemented as a
dimple into the jagged end of part n to form one
section. Six of these sections complete the solid,
the three described below in the colour table and
their three enantiomorphs. You must be sure 2 3|f = t'{4,3}
that part I is properly orientated before cement- 8{3} + 6{f}
ing it into part n, so the colours are correctly
arranged. The sharp dihedral angles at these
edges make cementing an easy process. This
makes an attractive and rigid model.

Parti
1 2 3 4 5
O R G R G
B R G R G
Y R G R G

Part II
12 3 4
Y B Y B
Y O Y O
O B O B

144
21:37:39 BST 2016.
CBO9780511569746.020
 93 cuboctahedron
Quasitruncated

This polyhedron has six octagrams in the same

facial planes as a regular octahedron. The set of
twelve squares intersect each other by threes in
such a way that a set of eight small triangular
holes provide openings that penetrate deeply
into the interior of the solid. The sides and
bottoms of these holes belong to the facial planes
of the intersecting hexagons. Thus the sides of
these holes are in fact a truncated version of the
Stella octangula and the bottoms coincide with
the interior regular octahedron. This suggests
the following method of assembly.
Begin by making the truncated pyramids
shown, but turn all the tabs out, because this
2 3 fl = t' {4} part is one of the holes. Since in an ordinary
8{6} + l32{4}+6{|} sized model the holes admit very little light, all
the parts may as well be W. Double tabs are
needed at the bases of part I, and are used to
cement the parts together as in the regular octa-
hedron. All these should be completed before
proceeding to part n. Part n contains the central
portion of the square planes, which can be given
the colours Y, B, O, R and the two side wings
which are W because they come from the hexa-
gon planes. This part can then be cemented at its
short edges to the edges of the small triangular
holes of part i. Next, complete another set of
these as colour pairs to the first four. A third set
of four forms an equatorial band of squares.
Finally pairs of triangles which belong to the
corners of the square planes are cemented

Part i

Part II

145
21:37:39 BST 2016.
CBO9780511569746.020
between alternate spaces around the octagram. handle, provided again that you finish an edge at
If you watch the model as you complete it up to a time. This makes a very sturdy model, because
this point, it will not be difficult to determine the it already has an elaborate interior structure, but
correct colours for these. This octagram, with its it also takes very careful work on this interior to
set of triangular grooves, can now very easily be get the exterior parts to fit well. It is interesting
cemented in place. The very acute overhang that to note that the holes in this model generally go
occurs in this model makes it a simple matter to unnoticed by the inexperienced observer.

21:37:39 BST 2016.

CBO9780511569746.020
 94 Great icosidodecahedron

This polyhedron is called the great icosidodeca-

hedron because it has twenty triangles and
twelve pentagrams in the same facial planes as
the regular icosidodecahedron, yet it is not a
stellation of the latter. A suitable colour scheme
is to make all the pentagrams W and use the five
colours for the triangles in the usual icosahedral
order.
To construct a model of this polyhedron begin
with the cups whose sides belong to five inter-
secting triangles and whose bottom is the small
pentagon from the pentagram plane.
Next it is best to prepare the W star arms in sets
of three, forming shallow trihedral dimples with
the star arms pointing outwards as shown. These
213 1 - then serve as connectors between pentahedral
dimples or cups. Enantiomorphism again applies
20{3} + 12{f}
27" 1
to the second half of the model. This makes a
very attractive model and proves to be very
rigid.

0 1 2 3 4 5
(0) W Y B O R G
(1) W Y B G O R
(2) wB O Y R G
(3) wO R B G Y
(4) wR G O Y B
(5) wG Y R B O

21:37:39 BST 2016.

CBO9780511569746.020
95 Truncated great icosahedron
This polyhedron is a truncated version of the
great icosahedron. It may therefore follow the
same colour arrangement. The hexagonal planes
here take the place of the triangular planes. The
method of assembly is therefore very similar. All
the pentagrams may be W.
Begin with a W star and cement the parts
shown below between the star arms. Follow the
same paired arrangement as in the colour table
for the great icosahedron 41. When the isos-
celes triangles have been cemented you will have
completed one section of the model, a penta-
hedral dimple serving as a sort of tray for a
raised star. The next five sections are done in the
same way, each with a W star. These sections are
then cemented to each other in the same way as 2 f | 3 = t{3,f}
the vertex parts of the great icosahedron. 20{6} + 12{f}
Equally you may choose to do all the stars in the /29-V5
dodecahedral arrangement using opposite stars
of the same colour and requiring six colours in
all. But then you cannot avoid having each star
share one of its edges with a hexagon of the same
colour. However this does not detract from the
beauty of this model, because the two planes
make a sharp angle, almost a 90° turn, at these
edges.

21:37:39 BST 2016.

CBO9780511569746.020
 96 Rhombicosahedron

This polyhedron is closely related to 76 and 83.

They all have the same set of thirty intersecting
skew squares forming equatorial bands. The
hexagons take the place of the pentagons and
thus the star planes become faceted stars. Also a
set of thirty shallow cups appear here directly
over the middle section of each square, the
bottom of each cup being part of the square
plane, and its four sides, each deeply recessed,
being part of four hexagon planes. It would be
more tedious than necessary to make a model of
this polyhedron while maintaining the same
colours for each plane. The method of assembly
suggested here will maintain different colours
for the squares and the upper surface of the
hexagon planes. This will still give a very attrac-
2 3I tive result and only an informed observer would
20{6}+30{4} notice the colour discrepancies since they occur
only on the deeply overhanging underside of
the hexagons and thus are hardly visible. You
may find even the suggested method of assembly
tedious enough because of the large number of
parts involved, so again your perseverance will be
put to the test if you wish to complete the entire
model.
Begin by assembling a faceted star as shown.
Each of these has a central cup made of five
equilateral triangles in the first icosahedral
arrangement. The isosceles triangles in the star
arms are next cemented to the edges of the cen-
tral cup. These are parts of the square planes. A
colour table for six parts follows.

1 2 3 4 5 6 7 8 9 10
(0) Y B O R G R G Y B O
(1) Y R O G B G O W Y R
(2) B G R Y O Y R W B G
(3) O Y G B R B G W O Y
(4) R B Y O G O Y W R B
(5) G O B R Y R B W G O

149
21:37:39 BST 2016.
CBO9780511569746.020
 When you have assembled these ten parts of a these, five of each of the colours W, Y, B, O, R, G.
faceted star, you will have no difficulty cementing Part in forms a shallow cup with deeply recessed
the pairs of obtuse triangles that finish the sides which are almost invisible when viewed
dimples in the star arms. These pairs lie in the straight on. Cement five of these cups between the
planes of the equilateral triangles in the central pairs in part n. The colour sequence is determined
cup and so the colours are determined by follow- by the corners of the squares formed by part n.
ing the plane out to the star arm. Again these Continue making faceted stars, each with a set
faceted stars lack rigidity, so it is better to add of five of part n between the star arms and
some backing. The pairs shown below, as part n, cement them together using the cups of part in
are cemented between the star arms. They have as connectors. Finally the small triangular holes
the same colour arrangement as in 76, but here left at the blunt ends of three of part n are closed
the shape is different. You will notice a small last of all with the small trihedral dimples shown
triangle at the blunted vertex of this part. It is a in part iv, the faces of which belong to the
tiny portion of the underside of the hexagon square planes. The dimples are best cut without
planes, so the colour will not have to be the tabs as they will then easily fit into their posi-
same as the upper side of these planes. Assembly tions. The colour arrangement is followed by
is greatly simplified if it is cut in one piece with the watching the order of colours in the intersecting
quadrilateral to which it is attached. Part in is squares. This makes a fairly rigid model and is
also best done as one net. You will need thirty of certainly noteworthy for its complexity.

Part II Part iv

150
21:37:39 BST 2016.
CBO9780511569746.020
 97 Quasitruncated small
stellated dodecahedron

This polyhedron is a quasitruncated version of

the small stellated dodecahedron. A model can
easily be assembled in twelve sections whose
base edges are equivalent to the edges of the
dodecahedron. The colour arrangement is the
usual six-colour dodecahedral one, set out in the
table below.
The top edges of 6, 7, 8, 9, 10 are cemented to
1, 2, 3, 4, 5. The section then has a pentagonal
edge at the bottom. The tabs at these edges are
cemented to each other as if the sections were
faces of a dodecahedron. Notice that 1 is on a
plane parallel to 6, 2 to 7, and so on, making
decagram and pentagram planes the same
colour. This makes a very attractive and rigid
model.
17-5V5
2"
0 1 2 3 4 5 6 7 8 9 10
(0) W Y B O R G Y B O R G
(1) Y W G 0 R B W G O R B
(2) B W Y R G O W Y R G O
(3) O W B G Y R W B G Y R
(4) R W O Y B B W O Y B G
(5) G W R B O Y W R B O Y

21:37:39 BST 2016.

CBO9780511569746.020
98 Quasitruncated dodecahedron

This polyhedron is to the great stellated dodeca-

hedron what 88 is to the stellated octahedron.
Here the decagrams replace the octagrams. A set
of thirty squares intersecting by threes appears
here, just as twelve squares appeared in 88,
forming twenty small triangular holes which
penetrate deeply into the interior. The facial
planes of these holes belong to the intersecting
decagons.
The simplest way to make this model is to use
only three colours, one for each type of polygon.
Begin as in 88 where you built the interior trun-
cated version of the stellated octahedron. Here
you build the interior truncated version of the
great stellated dodecahedron to form the sides of
the holes, but the bottoms are equivalent to the
surface of the great dodecahedron. Since this
will all be one colour, one net serves for all
twenty parts (part i). Remember to turn the tabs
out so these parts can be cemented together. All
twenty are best completed before continuing
with the rest of the model.
Next, prepare the central parts of the squares
with their adjoining wings, the squares of a
second colour, the wings of the same colour as
part i. The figure shows the arrangement of nets
for part n. Part n is then cemented to straddle
from hole to hole. Next the decagrams are pre-
pared with the triangle pairs from the square
plane corners cemented between every other star
arm. This is analogous to the procedure followed
in 88. The deep recess of part n makes it easy to
cement these decagram parts, although the large
number of edges will take some time. With per-
severance you will finish a sturdy and pleasing
model.

152
21:37:39 BST 2016.
CBO9780511569746.020
Part i

Part II

153
21:37:39 BST 2016.
CBO9780511569746.020
99 Great dodecicosido-
decahedron

This polyhedron is very attractive when done in

colour because it readily appears as a set of
highly embossed and richly designed solid stars
on decagram planes. It is easy to make this
model by the usual sectional method of assem-
bly. Part i is shown with its chevron-shaped faces.
These form trihedral dimples with the same
colour arrangement as the great dodecahedron
21, so you may refer back to it for your
requirements here. Part n is made up of star arm
pairs and these serve as connectors for part i.
Each of these is cemented so the star arm
colours are on planes above and parallel to the
same colour of the chevron-shaped faces of part i.
The five rhombi from the triangle planes, part
in, form a rosette. Part in follows the first icosa- 20(3} + 12{f}
hedral arrangement of colours. As soon as you V01-V5)
have cemented a ring of five of each of part i and
part II alternating, the first rosette can be
cemented to fill the hole at the centre. Once this
has been done the colours for the other pairs of
part II are easy to determine because the parallel
planes take the same colour. In cementing the
rosettes keep an eye on the triangular planes.

Part i

Part II

154
21:37:39 BST 2016.
CBO9780511569746.020
 100 Small dodecahemi-
cosahedron

This polyhedron has twelve pentagrams on the

facial planes of a dodecahedron and ten equa-
torial hexagons whose centre points all coincide
with the centre of the solid. It is related to 73
whose dimples are here replaced with deep hexa-
hedral holes or inverted pyramids. These holes
may be thought of as the twelve vertex parts of
the great icosahedron, all inverted and turned
inwards. The colour arrangement suggested be-
low will show very effectively an important feature
of polyhedral symmetry, namely the generation of
spherical lunes.
Turn all the tabs outward to form ribs and
then use these to cement the parts to each other,
the tabs of one under the surface of the other so
10{6} + 12{f} the edges coincide. W pentagrams are cemented
2 as needed. The usual enantiomorphism applies,
and you should have no difficulty completing
the model. This makes a very rigid model, but
care must be exercised to make the pentagrams
accurately so they fit well. The final result also
depends on the accuracy of the inverted pyra-
mids.

1 2 3 4 5 6
(1) Y R G Y B G
(2) B G Y B O Y
(3) O Y B O R B
(4) R B 0 R G 0
(5) G O R G Y R
(6) R R B B G Y
(7) G G O O Y B
(8) Y Y R R B O
(9) B B G G O R
(10) O O Y Y R G

155
21:37:39 BST 2016.
CBO9780511569746.020
101 Great dodecicosahedron
This polyhedron is like 81 except that the
dimples and grooves there are here replaced
with deeper holes and cups, nonahedral and
tetrahedral. The usual dodecahedral and icosa-
hedral arrangement of colours applies again.
The model can be assembled by making the parts
as shown. These alternate between the star arms
and serve as connectors.
You can easily see which position each part n
should take if you keep the intersecting hexa-
gons, each its own colour. You will also find the
cups sharing edges internally but the tabs on
these edges need only be adjusted, not cemented.

3 i
i
L2-I
17-3V5

Parti Part II
1 2 3 4 5 6 7 8 9 1 2 3 4
(1) B Y R B B G Y G Y (1) Y G Y R
(2) O B G O O Y B Y B (2) B Y B G
(3) R O Y R R B O B O (3) O B O Y
(4) G R B G G O R O R (4) R O R B
(5) Y G O Y Y R G R G (5) G R G O
(6) B O Y B G O R B R (6) Y G Y 0
(7) O R B O Y R G O G (7) B Y B R
(8) R G O R B G Y R Y (8) O B O G
(9) G Y R G O Y B G B (9) R O R Y
(10) Y B G Y R B O Y 0 (10) G R G B
(11) B G O Y
. —1 (12) O Y R B
p
L YJ
7 6 \
\
(13)
(14)
R
G
B
O
G
Y
O
R
^ \ < 2 (15) Y R B G
8 / N
7
\

\ 9\
K\ ;/
4
J
5

"7 71 /

Part i Part II

156
21:37:39 BST 2016.
CBO9780511569746.020
 157
21:37:39 BST 2016.
CBO9780511569746.020
102 Great
dodecahemicosahedron

This polyhedron is a faceted version of the

dodecadodecahedron. The facial planes of the
faceted stars are a combination of the inter-
secting pentagons and hexagons, whose parts
are easily recognizable. The hexagons are in the
same planes as those of 100. To make a model of
this polyhedron make the faceted stars using the
colour table given below.
The figure is merely a plan and does not show
the exact shape of each part, only its relative
position. The 0 and 6 7 8 9 10 parts belong to
the pentagon, the rest to the hexagon. These
faceted stars are then joined by using the
dimples between star arms as in 73. This makes
a very attractive model and is very rigid without I 5|3
any further internal support. 10{6} + 12{5}
2

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
(0) W Y B O R G Y B O R G B O R G Y G Y B O R
(1) Y Y B G B G W G O R B B G B G Y G Y B G Y
(2) B B O Y O Y W Y R G O O Y O Y B Y B O Y B
(3) O O R B R B W B G Y R R B R B O B O R B O
(4) R R G O G O W O Y B G G O G O R O R G O R
158 (5) G G Y R Y R W R B O Y Y R Y R G R G Y R G
21:37:39 BST 2016.
CBO9780511569746.020
 103 Great rhombihexahedron
This polyhedron is closely related to 77. The
tetrahedral dimples and dihedral grooves of that
solid are here replaced with deeper nonahedral
cups and tetrahedral dimples. Three colours can
again be used for the octagrams, making op-
posite pairs the same colour. All the other planes
are squares intersecting internally. Since there
are twelve of these, six colours will serve as six
pairs. This means of course that the Y squares
must meet the Y stars along at least one edge,
and so on for each of the other colours, but the
sharp angle at these edges helps greatly to make
this violation of the map colouring principle
almost unnoticable.
The best method of construction is to make
the cups and dimples first, then to cement them
alternately between the star arms. You will find
12(4}+ 6{f}

159
21:37:39 BST 2016.
CBO9780511569746.020
 Parti
1 2 3 4 5 6 7 8 9
(1) Y O B G W R G W R
(2) B W R Y O G Y O G
(3) R o G B wY B wY
(4) G wY R o B R o B

Part ii
1 2 3 4
(1) Y O w R
(2) B W o G
(3) R o wY
(4) G w o B

that adjoining cups and dimples share common

edges internally but the double tabs at these
edges need only be adjusted, not cemented.
If you watch the parts of the square planes to see
that each keeps its own colour you should have
no trouble finding the correct position for each
part. All the parts listed above are cemented to a
Y octagram. Once this is done you are ready to
cement the next four octagrams, O, B, O, B.
These will then be completely surrounded by
cups and dimples whose colour arrangement is
enantiomorphous to the first set and diagonally
opposite their counterparts on the completed
model. Another Y octagram is added last, one
edge cemented at a time in the usual manner.
The deeper cups and dimples make this model
even more attractive than 77.

160
21:37:39 BST 2016.
CBO9780511569746.020
 104 Quasitruncated great
stellated dodecahedron

This polyhedron is a quasitruncated version of

the great stellated dodecahedron. The triangle
planes cut the vertices of the latter almost at
their bases, so that the pentagrams are trans-
formed into decagrams. Trihedral dimples close
the cut ends. Begin this polyhedron by making
the truncated pyramids, one of whose side faces
is shown below. The colour arrangement for
these is the same as that for the great stellated
dodecahedron 22, so you may simply refer
back to it for the colour table to be used here.
The trihedral dimples have the icosahedral
arrangement of colours, as given in the table.
You must watch carefully how you orient these
i 3 | | + t' {f,3} dimples before cementing them into their
20{3} + 12{^} positions closing the jagged end of the trun-
/37-15V5 cated pyramids. (1) must have its Y arm cemented
between the B and G of the pyramid, and so on
round. Once you have completed the first ring
of truncated pyramids and their dimples, the
other parts are easier to locate. Since opposite
dimples are not enantiomorphs the complete
colour table is given.

1 2 3 1 2 3
(1) Y G O (11) B Y G
(2) B Y R (12) O B Y
(3) O B G (13) R O B
(4) R O Y (14) G R O
(5) G R B (15) Y G R

4 5 6 4 5 6
(6) R G Y (16) Y B R
(7) G Y B (17) B O G
(8) Y B O (18) O R Y
(9) B O R (19) R G B
(10) O R G (20) G Y O

161
21:37:39 BST 2016.
CBO9780511569746.020
105 Quasirhombicosi-
dodecahedron

This polyhedron is very similar to 99, except

that here the decagram planes give way to a
fantastically intricate network of intersecting
planes generating a very interestingly faceted
structure. Fortunately it is not too difficult to
make and only requires plenty of perseverance to
cut and join the numerous parts. The arrange-
ment of colours for triangles and pentagrams
used in 99 can be retained here, but the squares
are best in white. This vastly simplifies the
assembly.
Begin with a dimpled rosette of five colours as
in 99. Surround this with five parts, all W and of
one net as shown for part i. Cement a set of five
star arms to the side parts. The Y star arm
shares its vertex with the Y rhombus of the
rosette, and so on for the other colours. Then 20{3} + 30{4} + 12
make a set of paired wedges, as shown for part n, V01-V5)
whose triangular faces are very small parts of
the square and triangle faces. The tabs on the
shortest sides of these triangles are cemented to
the tabs at the small cut-out of part i. The best
way to handle these wedges is to cement the W
first, one tab at a time, then both coloured tabs
at once using tweezers, since the parts are so
small in a normal sized model. The last parts to
be cemented for this section can now be done,
one edge at a time. These are the ten pieces, shown
as part ia, belonging to one facial plane of the
pentagram. Once these are added you should
have a fairly rigid structure whose projection on
the plane of the pentagram is shown opposite.
This completes one section of the model.
A total of twelve are required altogether. The
first or (0) section has the first set of five sections
surrounding it. These are cemented by one of

162
21:37:39 BST 2016.
CBO9780511569746.020
Part ia

the blunt star arms to the blunt star arms of the

(0) section, and these five all join blunt star
arms with each other, the intersecting penta-
grams giving full length to the star arm. You
will then see between these sections open holes
whose edges make a skew hexagon. These should
be filled with the cupped structure shown as
part in.
Part II Three pairs of W triangles are cemented to the
edges of triangles 2, 3, 4. Be careful to dis-
1 2 3 4 tinguish between the upper and lower edges of
(1) Y O B G these W triangles. Part in may now be cemented
(2) B R O Y to fill the skew hexagonal holes. If you watch the
(3) O G R B triangles 2, 3, 4 so that their colours agree with
(4) R Y G O those in the rosette you can get part in correctly
(5) G B Y R placed. The triangle 1 of part in belongs to the
(6) Y R B O central portion of the triangle plane.
(7) B G O R This turns out to be a very beautiful model, so
(8) O Y R G it is well worth the time it takes to assemble it.
(9) R B G Y You may expect to spend at least 30 hours to
(10) G O Y B complete it.
The rest are the enantiomorphs.

163
21:37:39 BST 2016.
CBO9780511569746.020
 106 Great icosihemi-
dodecahedron

This polyhedron is closely related to 94. The

pentahedral dimples or rosettes formed by the
triangle planes are still here, but the pentagrams
have disappeared giving way to equatorial
decagrams whose facial planes form shallow
cups where the star arms used to be. The bottom
of each cup is a trihedral dimple. This suggests
a method of assembly. Make the cups first, each
with double tabs turned out to form ribs in the
usual manner. Use these ribs as tabs a second
time to cement a set of five cups in a ring. The
colour arrangement for these follows:
1 2 3 4 5 6
(1) B R G G Y B
(2) O G Y Y B O
(3) R Y B B O R
(4) G B O O R G 2T-1
(5) Y O R R G Y
(6) G R Y O W G
(7) Y G B R W Y
(8) B Y O G W B
(9) O B R Y W O
(10) R O G B W R
The rest are the enantiomorphs.
The same icosahedral arrangement as in 94 is
also used here for the rosettes. These may be
cemented in place as each ring of cups is com-
pleted. Keep your attention on the facial planes,
both for the cups and the rosettes to see that each
decagram plane and each triangle plane keeps its
own colour.

164
21:37:39 BST 2016.
CBO9780511569746.020
 I vJ / Great dodecahemi-
dodecahedron

This polyhedron is closely related to 94 and to

106. It is simple and very rigid in its structure,
with equatorial decagrams on planes parallel to
and midway between pairs of opposite penta-
grams. This is beautifully brought out in the
colour arrangement. To make a model of this
polyhedron begin with the pentagonal cups as
listed in the colour table below.
Cement each cup of part i using the double tabs
as connecting tabs, turned slightly and brought
under the surface of its neighbour, to make the
slant edges of the cups coincide. Then make
twenty star arm dimples, the trihedral arrange-
ment being the same as the colour table used for
fill the great dodecahedron. These are used to close
{f} + 6 the holes between the cups, matching the same
IT-1 planes for colour. This model is very attractive.

0 1 2 3 4 5
(0) W Y B O R G
(1) Y W G O R B
(2) B W Y R G O
(3) O W B G Y R
(4) R W O Y B G
(5) G W R B O Y
The rest are the enantiomorphs. Part i

165
21:37:39 BST 2016.
CBO9780511569746.020
108 icosidodecahedron
Great quasitruncated

This polyhedron has very many parts, and in a

model of ordinary scale the faceted star parts are
very small. For this reason a full colour model
would entail a very tedious amount of work. The
method of assembly described here limits the
colours to three, one for each of the three types
of polygons that appear as facial planes. A good
colour combination is obtained by using Y for
the hexagons, R for the squares, and B for the
decagons. Begin by making a set of five parts of
each of the pairs shown, parts i (a) and (b). These
ten parts are cemented alternately in a ring,
leaving a star shaped hole in the centre. This
hole is now filled with a tiny faceted star or
pentagram. They are shown full scale for a 2341 =
model of 8 inches on an edge. Make the central
pentahedral dimple first. Then cement the small
triangles to its edges. Next add the pairs of tri-
angles completing the star arm dimples and
finally add the other paired triangles between
the star arms. This dimpled or faceted star can
now be cemented to close the star shaped hole
left in the ring of part I parts. You will have to
make a total of twelve rings and twelve faceted
stars for the complete model. These twelve sec-
tions are joined with three types of connectors,
a nonahedral cup, of six side faces and three
bottom faces (part in a), a dihedral groove (part
in b)9 and a tetrahedral dimple (part in c). These
are shown in full scale as part in a, b, c. These
connecting parts help to give some rigidity to the
model, but for best results some further interior
supports are needed, especially under the edges
of part i.

166
21:37:39 BST 2016.
CBO9780511569746.020
 Part me
167
21:37:39 BST 2016.
CBO9780511569746.020
109 Great rhombidodecahedron

This polyhedron has the same decagrams as 99,

but the pentagrams and triangles there are
replaced by squares here. The intersection of
these squares with one another introduces deep
holes at the places where the twin star arms and
triangle rosettes occurred in 99. In place of the
twin star arms you will find hexahedral cups
whose four sides belong to the facial planes of
four squares and whose bottom is a pair of small
isosceles triangles of the 36°, 72° variety belong-
ing to the facial planes of two decagrams. In
place of the rosettes you will find a fascinatingly
complex structure of cavities which fortunately
is not too difficult to construct. As for colour
arrangements you undoubtedly will agree that it
would be much too tedious to use six colours for 211
the squares. So the suggestion here is to make all
the squares W and then distribute the six colours V01-V5)
Y, B, O, R, G, and W according to the usual
dodecahedral arrangement as in 99. In this way
the chevron-shaped parts are used once more
here. Thus you may begin the construction of
this model with a set of five trihedral chevron-
shaped dimples as in 99. These are part i. The
other parts are also set out full scale for a model
whose edges will be 8 inches.
Part ii shows the hexahedral cup, the sides
being one net for parts all W. The small triangles
must be done in colour pairs determined by their
position relative to part i, so you should need no
specific directions to get them right if you work
systematically. Since they are so small it is
easiest to cement them as pairs, then each pair is
cemented as a bottom to the W cups. Five of
part i and five of part n then form a ring with a
hole in the centre. This hole is in the shape of a
skew decagon. Next you must make five of part
in as shown. All have the paired triangles in W as
one net with the other part from the decagram
planes in the usual dodecahedral order of
colours. When completed part in forms a tri-
hedral cup pointed at the bottom, the upper
edges having the shape of a skew quadrilateral.
These parts are then immediately cemented so

168
21:37:39 BST 2016.
CBO9780511569746.020
 Part iv

Part m

Part v

169
21:37:39 BST 2016.
CBO9780511569746.020
the angle formed by the W edges is vertically takes the colours of the decagram planes since it
opposite the angle formed by the two sides of is the central portion of those planes. The tabs
the part n cups. You will now see the decagram on this multifaced cup are of course all turned
planes being filled in, so you can get the parts outward as ribs on the outside of the cup, which
placed correctly by watching this. When these is then gently forced down the remaining central
parts have all been cemented you will still find a hole left by the five parts of part in and cemented
hole in the centre, still a skew decagon but an edge at a time. The reason for exerting
deeper down. This hole is now again filled with 'gentle force' is that this cup has vertical sides
another cup shown as part iv. A net for the sides, which just fit into the hole so you must adjust the
all W, is shown p. 169, part v. You must be careful outer ribs while you are placing it in its position.
to distinguish between the upper and lower ends Clamps are helpful in cementing twenty sets of
of the series of ten trapezia, because they are not tabs, and with a little skill and perseverance you
the same, but very nearly so. The series makes a should succeed.
ring in the form of a ten sided tube or prism. It is You may actually complete the whole model
all W since all these trapezia belong to the facial as a shell without filling any of the central holes.
planes of ten intersecting squares. The ring of This has the advantage of giving you the overall
irregular pentagons shown as part iv is also one colour scheme, especially for parts n and iv.
net all W since they belong to the facial planes of However, complete rigidity is not achieved until
five more squares. You can see now why it would all the central holes are closed. Once this is done
be tedious to make all of these square planes a you have a very attractive model, very interesting
different colour. The ring of irregular pentagons because of its intricate structure. Since all the
may now be cemented to the lower end of the squares are W, their facial planes are not too
ten sided prism. The regular pentagon then evident, but with the model in your hands they
forms the bottom and closes off that end. It are not hard to locate.

170
21:37:39 BST 2016.
CBO9780511569746.020
Commentary on non-convex snub polyhedra

There are two convex snub polyhedra, the snub taining all the numerical data by programming
cube and the snub dodecahedron. Among the a computer to work out the analytic equations.
non-convex polyhedra there are at least nine, ten Briefly stated, his method involves a system of
if you count 119, which is rather different from spherical coordinates, the polar axis being along
the others. The snub quality manifests itself in an axis of rotational symmetry and the zero
the two convex cases, in the twisted way in which meridian through a vertex of the polyhedron.
the squares and pentagons are related to a cir- Then with right and left specified, polar co-
cumscribing cube and dodecahedron respective- ordinates for all vertices can be calculated. These
ly. The twist introduces dextro and Iaevo, right- are translated into the usual Cartesian co-
and left-handed varieties in each case, and also a ordinates to obtain equations for the facial
special set of triangular faces, the snub triangles. planes. Then the line of intersection of a pair of
The same twisted characteristic is found in the facial planes is readily obtained by solution of a
non-convex cases and also sets of snub triangles. system of equations. The computer supplied
Model 119 is special in that the diametral numerical results correct to six significant
squares may be considered as the snub faces. figures. For the purposes of model-making this
The patterns on the facial planes of these non- degree of accuracy is unnecessary.
convex snubs are tantalizingly irregular. None of The drawings given here were derived from
the usual symmetry manifests itself, except, sur- large-scale drawings, edge length 20 cm (for 117
prisingly enough, in 110 which is very simple and 118, 20 inches) supplied by Mr Buckley. Only
and in 118 which is very complex. Because of some of the principal lines of intersection are
this lack of facial symmetry (the solids as such given. The exterior portions of each facial plane
have rotational symmetry), the intersections of are shaded as usual, light grey for the upper side
the facial planes determining the pattern on and dark grey for the under-side. Numerical data
these faces can only be found by calculation. has been reduced to two figures. All the models
This involves the use of analytic or coordinate shown in the photographs are less than 12 inches
geometry. The use of the usual Euclidean tools, in height, except for 117 and 118 which are nearly
ruler and compasses alone, will not suffice 24 inches tall for an edge length of 20 inches. The
Mr R. Buckley of Windsor, Berks, England, nets included in the instructions will give you
has recently performed these calculations, ob- models on this scale.

171
21:38:34 BST 2016.
CBO9780511569746.021
110 Small snub icosicosi-
dodecahedron

This is the first of the non-convex snub poly-

hedra, and it is the simplest to make. Twenty
pairs of equilateral triangles, forty in all, share
icosahedral facial planes, giving this polyhedron
the appearance of having twenty hexagram
faces. But these hexagrams are not quite regular,
although the triangles from which they are
formed are equilateral, as they must be in a
uniform polyhedron. The twelve pentagrams are
completely surrounded by another set of tri-
angles, sixty in all. This suggests a simple method
of construction.
Begin by cementing the scalene triangular
parts as dihedral grooves between pentagram
star arms, following the icosahedral arrange- |33f
ment of colours. The pentagrams are all the (40 + 60){3} + 12{f
same colour, W. This is part I of the model. The 2-91638 06615
first section is shown below, the (0) section. Five
hexagrams (part n) are then cemented around
part i. These may be a seventh colour or the
usual Y, B, O, R, G arranged to maintain the
map-colouring principle. The next five sections
of part i are then cemented in place. You will
now see openings between the hexagrams.
Close these with pairs of small equilateral tri-
angles, part in, each the colour appropriate to
the facial plane to which it belongs. Cement the
pairs together first and then cement them to
close the openings between the hexagrams.

Edge length = 6-6 cm

a = 2\
b = 2-3
c = 2-5

Parti
Part II

Part HI
172
21:38:55 BST 2016.
CBO9780511569746.022
Part HI

173
21:38:55 BST 2016.
CBO9780511569746.022
111 Snub dodecadodecahedron
This polyhedron has twelve pentagrams on
facial planes that lie above and parallel to
twelve pentagons, very much like 73, only here
the pentagrams share edges with sixty equi-
lateral triangles that give it the snub quality. As
you can see from the drawings of the facial
planes the triangles and pentagons intersect in
such a way that tiny slivers appear on one side of
the triangles and on all five sides of the penta-
gons. You will therefore need a great deal of
patience to make this model properly.
Begin the same way as in 110, surrounding W
pentagrams with the appropriate scalene tri-
angular dihedral grooves between the star arms,
following the same icosahedral arrangement of
colours. This is part I of your work. Part n con- 1215
sists of a rather complex assembly of parts that 12{|}+ 60(3}+ 12(5}
eventually turns out in the shape of a skew-sided 2-54887 97641
equilateral triangle having three grooves running
from its vertices toward a point near the incentre
and three slivers radiating from near the in-
centre out toward the sides. The figure opposite
will give you some idea of its appearance.

174
21:38:55 BST 2016.
CBO9780511569746.022
 Assembly of part n can be done in steps. First
cement a triangular sliver of the appropriate
colour to the triangular part also of the appro-
priate colour, as shown here. Then cement a
pentagon sliver to its triangular part. Tabs are
Triangular part
shown for these slivers because trimming is
important here. Now join these parts and form
the sliver ridge between them. Clamps or
tweezers are helpful. Three of these assemblages
are needed for one skew triangular part, and a
Pentagon sliver total of twenty skew triangular parts are needed.
With a little patience you should be able to get
Triangle sliver them together. Only work on one edge at a time
and do the best you can with the small tabs on
the blunt ends of the slivers, working them into
Triangular part place with needle-point tweezers. If you trim the
tabs on the triangular parts as shown you can
now join them into one tab enclosing the pointed

Part II

Pentagon part

175
21:38:55 BST 2016.
CBO9780511569746.022
end of the sliver. This gives added strength and Edge length = 6-7 cm
rigidity to this edge. Once you have completed
five of part n, they may be cemented in a ring
around the (0) section of part i. A colour table
follows for all twenty skew triangular parts. The
subscript colour is that of the colour sliver. Since
pentagon parts are all W they are not in the table.

1 2 3 1 2 3
YG Go O Y O R RB Bo
BY Y R RB RG G o O R
O B B G Go G Y Y R RG
Ro O Y Y R Y B BG G Y
G R RB B G Bo O Y Y B
O G G Y Yo BR Ro O B
R Y Y B BR OG G R Ro
G B Bo O G RY Y G GR
Yo O R R Y GB BY Y G
BR RG G B Yo O B B Y

A simpler approximate model of this poly-

hedron can be made with the twenty skew tri-
angular parts omitting the slivers. The figure
should give you the idea. These can be used as
part II and cemented to part i in the way ex-
plained above. Approximate facial planes are
shown below. The scale of this approximate
model may be much smaller, in fact on a small
scale it gives good results. See the photograph.

176
21:38:55 BST 2016.
CBO9780511569746.022
 112 dodecahedron
Snub icosidodeca-

This polyhedron, like 111, has pentagrams on

planes above and parallel to pentagons, but here
the pentagrams are twisted with relation to the
pentagons, the twist making room for twenty
triangles in addition to the sixty triangles
sharing edges with the pentagrams.
An alternative method of assembly is best
used in making a model of this polyhedron, by
cementing the pentagrams last because of the
intricate understructure. The work may there-
fore be done as follows. Use the icosahedral
arrangement of colours for the sixty triangles
associated with the pentagrams, all of which are
W. The other twenty triangles had best be a
seventh colour, say T (for tan), because a small
12{f}+(20 + 60){3} central portion of these triangles appears in the
2-25379 58256 bottom of a trihedral dimple whose faces are the
corner portions of three different pentagon
planes all of which are W. Single polyhedral
vertex parts may be constructed in the arrange-
ment shown in the figure. The colour arrange-
ment for the (0) section alone is set out on p. 178.
The other sections are in the usual cyclic per-
mutation of colours.

177
21:38:55 BST 2016.
CBO9780511569746.022
 Clamps can easily be used in doing this since
there is an acute overhanging edge to work on.
If the pentagrams tend to sag do not be dis-
mayed, gentle pressure can be exerted to
1 2 3 4 5 6 7 straighten them after the cement is well set or
R Y T Y W R Y after the model is completed. Twelve of these
G B T B W G B sections complete the model. Admittedly these
Y O T O W Y O sections have many jagged edges, but you will be
B R T R W B R amazed at the way they all fit neatly together.
O G T G W O G The alternative method of construction is
reminiscent of that used in 83. If you use this
Triangles 1 and 7 must be turned up and over 2 method begin by first cementing the triangles 1
and 6, and then the tabs between 1 and 2 and and 7 between the star arms, then fold them
between 7 and 6 are used again as double thick- under and turn the other long tabs outward to
ness tabs and cemented to form a vertex part, form ribs radiating from the centre out to the
the centre point being raised in doing this. Tri- star points. Next the triangle pair 2 and 6 form
angles 1 and 7 will later become the under sur- grooves between the solid star arms, the ribs
face of one star arm. Five of these vertex parts serving as cementing tabs for these triangles.
are joined in a ring to form one section. The tab Finally the remaining parts 3, 4, 5 complete the
at the blunt end of triangle 7 is cemented to the vertex parts around the star points and complete
upper surface of triangle 1 whose tab is best one section as before. This second method of
removed to eliminate the ribs here which would assembly assures a better fit for the pentagrams,
interfere with the pentagram. This pentagram is because in this method there is less tendency for
now added as a cap completing this section. them to sag.

21:38:55 BST 2016.

CBO9780511569746.022
 113 Great inverted snub
icosidodecahedron

This polyhedron is another snub that is simpler

in construction than most of the others in this
set. The reason is that it does not have a crossed
vertex figure, so its vertex parts are thereby
simplified. Thus a model can easily be made by
simply joining the sixty vertex parts following
the arrangement set out below, showing one
vertex part. Make five such parts and then join
them in a ring to form one section. Twelve of
these sections complete the model. The central
dimples in these sections will remind you of the
analogous dimples in the compound of five
tetrahedra. Dimples like these appear also in 115
and 116. The colour arrangement may again be
icosahedral. The colours for only one section,
(20 + 60){3} + 12{f} the (0) section, are set out below, because the rest
1-63216 13496 follow the usual permutation pattern. All the
pentagrams are W.
These sections have very jagged edges and
several very small tabs, so you will need patience
and care in building this model.

1 2 3 4 5
W R G Y G
W G Y B Y
W Y B O B
O R O
W O R G R

179
21:38:55 BST 2016.
CBO9780511569746.022
114 Inverted snub
dodecadodecahedron

This polyhedron has a very unusual feature in

that the star arms of the pentagrams are slightly
nicked by the facial planes of the pentagons and
one set of triangles. This introduces a very com-
plex structure near the polyhedral vertices. The
pentagons are also on planes parallel to the
pentagrams but very deep down near the equa-
torial planes of the polyhedron. The under-
structure of the pentagrams is reminiscent of
that found in 83 and in the snub 112. This fact
suggests a method of constructing a model of
this polyhedron, like the alternative method
mentioned for 112.
Begin by cementing triangles 1 and 2 to the star
arms of the pentagram (see opposite), then turn |2f5
them under and cement the other long tabs 12{f} +60(3}+ 12(5}
turned out as ribs. It is not necessary to cut the 1-70326 04562
nicks into the star arms, although after the
cementing is completed on triangles 1 and 2 and
the solid star has been constructed, the nick in
triangle 2 should be cut and removed without
severing the rib tab. This will make it easier for
the nick in the pentagon corner portion, part 5,
to fit later without further cementing, as will be
explained shortly. The usual icosahedral colour
arrangement is used, with all the pentagrams W.
Your next task is to prepare the parts needed
between the star arms. The arrangement of these
parts is shown opposite. Here too the usual icosa-
hedral colour arrangement is used, and so only
the (0) section is given. The usual permutations
apply to the rest.
Tabs are shown because trimming is import-
ant.
Edge length = 14 cm
0=20
6 = 0-7
c=l-4
</=0-7
£ = 5-3
h = 3-6
A:=3-8
7=4-2

y=0-l
z = 0-4

180
21:38:55 BST 2016.
CBO9780511569746.022
 First cement the colour pairs, parts 3 and 4. Then
cement part 5 which is W to part 6 which is of an
icosahedral colour. Next cement the tabs a, b, c
which join these paired parts. A sharp pointed
ridge or dihedral angle is thus formed between
parts 5 and 6 and the whole assemblage takes on
some rigidity. Your final task is to cement these
assemblages between the solid star arms.
The ribs under the star take the tabs e and f.
m
Apply the cement to these tabs. Then manoeuvre
the assemblage into place, and clamp before the
cement sets. The small nicked triangle of part 3
dangles at the star point but being attached by

(0) 1 2 3 4 5 6
Y Y Y O W Y
B B B R W B
O 0 O G w 0
R R R Y w R
G G G B w G
181
21:38:55 BST 2016.
CBO9780511569746.022
one continuous tab to the larger portion it is easy first the tab at g from part 6 of one section to the
to handle. It will get more attention later, but tab at h from part 4 of another section. The tabs
now you must give special attention to the nick at d are joined next. Only when the entire model
in the pentagon corner portion, part 5. Gently has been completed should you attempt to get
ease the upper point over the star arm while the the final tabs at k and m adjusted. It is useless to
lower point stays below. You need not worry attempt it earlier. You will need tweezers and a
about cementing these since the little nicks have probing needle to manoeuvre the parts gently
no tabs. The upper point will get further atten- into place, then add a drop of cement and fix
tion later. Once you have placed five of these them with a clamp. The edge here is just acute
assemblages between the star arms one section enough to take a clamp until the cement is set. A
of the work is completed. Twelve sections are good-quality paper 'heals' itself after the clamps
needed to complete the model. are removed or you can help by smoothing it
Now it is rather a complex task to get the with the probing needle. This model calls for
sections joined. The secret is always to cement much patience, so good luck to you on this one!

182
21:38:55 BST 2016.
CBO9780511569746.022
 115 Great snub
dodecicosidodecahedron

This polyhedron has the special feature of paired

pentagrams on common facial planes, found also
in 119. Here the twisted arrangement in the
twelve dimples is found, as it is in 113 and again
in 116. However the crossed lines of the vertex
figure in this case give this polyhedron a more
complex structure for the vertex parts than that
found in 113. The crossed lines appear again in
116.
To make a model of this polyhedron it will be
best to depart from the dodecahedral sections
used before and to use an icosahedral assembly
technique instead. In this way the crossed facial
planes under one star arm can be more easily
executed. As for colour the crossed triangular
planes (parts 1-5) can well be all one colour, say
V2 T (for tan), reserving W for the paired penta-
grams and the usual Y, B, O, R, G for the penta-
hedral dimples, the usual icosahedral arrange-
ment and permutation of colours applying.
Notice that in the parts used for assembly pur-
poses triangles 1 and 2 are interchanged from the
positions they occupy on the facial planes in the
polyhedron itself. This is because these parts will
be folded back to form the understructure of one
star arm. Join three of the parts 1, 3 to form a

183
21:38:55 BST 2016.
CBO9780511569746.022
sort of three-bladed propeller. Leave tabs all
around, except at x. This tab may be removed
because no cementing is done along this edge.
Triangle 1 folds neatly under a star arm, as
explained later.
Next prepare three pairs of the pentagram
parts shown as part 7. Fold the lower star arm
shown in fig. (a) up and then cement the cross-
hatched part shown in fig. (b) to the under-
surface of the upper star arm allowing the lower
part to protrude as shown in fig. (c).
The shaded area in fig. (b) and (c) will later be
hidden, so it need not be marked in any way. In
fact it is best not to score the paper for this area
but only for the tabs. Although the tab at h is
cut it will serve as a single tab, as will now be
explained.
Three of these pentagram parts must now be
cemented around the three-bladed propeller.
First join the tabs h, allowing the lower star arm
of fig. (a) to dangle forward, that is, above the dangling star arm. The sharp overhanging edge
propeller blades. As soon as the cement has set along the star arm makes clamping possible
along these edges give the pentagram parts a while the cement is setting. When you have done
sharp crease downward. You will then be able this on all three star arms you should have a very
to bring the tabs k into contact, so they may be rigid assembly, the central portion of an icosa-
cemented. Now the tab m of triangle 1 on the hedral section which already shows the vertex
propeller blades is ready to be cemented to the points of three adjacent polyhedral vertices.

184
21:38:55 BST 2016.
CBO9780511569746.022
 The next step is to cement three of parts 2, 4 are Y. Since a total of twenty such assemblages
to the appropriate edges of the icosahedral are needed to complete the model, the colours
assemblage you have just completed. Join the Y, B, O, R, G serve in turn for part 6, three of
tabs n. Once the cement is set, triangle 2 is the same colour to each assemblage.
folded under the star arm, to bring the tabs p You will find it best to cement the sections to
into contact so that they can be cemented. You each other as you complete them, first a ring of
will now see the intersection of the crossed tri- five with the (0) arrangement of colours in the
angular planes clearly formed where the fold dimple in the centre of the ring. Always cement
between parts 1, 3 and between 2, 4 occur. Tabs the longer tabs first, beginning with tab t of part
k and p hold these planes in rigid position so no 5 from one assemblage and cementing it to tab t
further joining at the folds is needed. No tab is of part 6 from an adjacent assemblage. The
required at x of triangle 2, since the same other tabs then fall readily into place. In doing
situation occurs here as at x of triangle 1. You the last, the twentieth section, leave the tabs s
will now also see how the shaded area in fig. (b) without cement, so these parts on the sixtieth
is hidden. vertex part can be temporarily folded back
The last step of the work is now simple to giving you room to work with a probing needle
execute. Add the parts 5 and 6 to complete each or tweezers on the other tabs. Cement the tabs s
of the three vertex parts of the section, matching last of all.
the tabs q, r and s. Part 5 is T in colour, like the This should be a very successful model. The
parts 1, 3, since it belongs to the same facial method of construction suggested here can lead
planes, but part 6 in this first assemblage should to remarkably good results. If you work alone it
be Y. In fact all three of part 6, completing each will take you about 40 or 50 hours to complete
of the three vertex parts of this first assemblage, the task!

(c)

185
21:38:55 BST 2016.
CBO9780511569746.022
 I IU Great snub
icosidodecahedron

This polyhedron has some very small areas

which must technically be considered as the
exterior portions of its facial planes, but they are
so small that the drawings shown below do not
reveal them. Some features of 115 repeat them-
selves here, the twelve pentahedral dimples, the
sixty polyhedral vertices, but instead of paired
pentagrams single pentagrams intersect each
other in a similar icosahedral fashion. This
suggests the use of the same assembly technique
in making a model of this polyhedron as that
used in 115. A few compromises will be intro-
duced in the construction work for the sake of
practicality, compromises which barely betray
themselves in the finished model. 12 3 f
Some enlargements are set out opposite to (20 + 60){3} + 12{f
reveal the detailed parts of the facial planes and 1-29004 04746
the parts numbered for reference.

Edge length = 20 cm
0 = 40
= 4-9
c=0-3
d=01
e=H
/=40
g=\0

186
21:38:55 BST 2016.
CBO9780511569746.022
 The parts for constructing this polyhedron are
set out below, each part identified by a number
and each tab by a letter. These parts form an
icosahedral section or assembly, like that in 115,
but this one is slightly more complex. The colour
arrangement for one section follows; the rest are
derived by the usual permutations of colours.
The use of the same colour for parts 3 and 5
cannot be avoided since these belong to the snub
triangle faces.
1 2 3 4 5 6 7 8
Y B B Y B T T W
Y R R Y R T T W
Y G G Y G T T w
Leave tabs around all parts except at edges
marked x, where no tabs are needed. Begin by
cementing the pairs 1 and 3, and 4 and 2. Then
cement part 7 to part 1, at edges marked p.
Part 8 is ingeniously contrived to get a small
spiked wedge to raise its ridge above the star arm
or pentagram plane. Both sides of the wedge are
small triangles, one of which actually belongs to
the snub triangle plane and technically should be
one of the five colours. Technically it also cuts
slightly into part 4. However, the compromise
of making both triangles W and letting the
wedge touch part 4 without cutting it simplifies
the work of construction, and you will see later
that this cannot readily be seen because it is

21:38:55 BST 2016.

CBO9780511569746.022
well hidden under a star arm in the completed successful you should now see the central portion
model. Also, although tab a of part 8 is cut of this first icosahedral section well formed.
to shape the wedge its two segments are Three vertex points become evident, which will
treated as one tab and cemented to part 4 as eventually belong to three adjacent polyhedral
shown. The next step is to fold up parts 1 and 7 vertices, and a sort of three-bladed pin wheel
and to cement tab b of part 3 to tab b of part 7. formed in the centre of the section by the three
This forms a small but deep trihedral cup, the intersecting edges of three star arms. You should
bottom of which should technically be closed now also see the ridges of the little wedges at the
with a tiny triangle coming from the snub tri- base of the star arms fairly well in line with
angle plane and a tiny quadrilateral from the the edges just cemented.
pentagram plane. However, omitting these again The next step is much simpler. Cement
simplifies the work. So simply cement the tiny together tabs f of parts 1 and part 8, making sure
tabs at the bottom of part 1, 3, and 7 as best you that the colour of part 3 corresponds to that of
can and let the trihedral cup have a roughly part 2, B with B, R with R, G with G, around
pointed bottom. This will never be seen in the the pin wheel section. As soon as the cement is
completed model. set give these edges a good crease down. This
Parts 2 and 8 are now similarly folded up and should bring the tabs, already cemented, be-
tab c of part 2 is cemented to tab c of part 8. The tween parts 1 and 3 into contact with the tabs
two very tiny shaded areas, one on part 2 and between parts 2 and 4 behind or below the star
the other on part 8 need not be cut or marked. arms. These tabs now serve as double-thickness
You will see after folding that the wedge butts tabs and are cemented to form the line of inter-
against the shaded area of part 2 and the edge at section of the snub triangle planes which cross
x of part 2 crosses into the shaded area of part 8 behind the star arm.
but no cementing is needed at these places. This The final step is the simplest of all. Cement
folding also forms a trihedral cup, this one both parts 4 and 6 at the tabs marked h and join parts
technically and actually pointed at the bottom or 5 and 6 at the tabs marked k. The tab k of part 5
base of part 8. The cup formation gives both is broken into two segments; the smaller tab
these partially completed sections a fair amount will be cemented later on to its mate on part 7
of rigidity. when the sections are joined. When this has been
The next step calls for some very intricate done the three vertex parts of one section are
work, almost impossible to describe in words complete. You may find that the central part of
without an illustration but not impossible to the pin wheel is not entirely rigid. This is due to
execute. You will have to determine for your- the fact that no cementing was done here. As
self, by trial and error, the best way to manipulate the model develops and the sections are added
the parts mentioned in the following description. one after the other, they exert their own pressure
Take the B, Y, W section of parts 2, 4, 8 and inwards and the final result is satisfactory. In
join the R, Y, W section of the same parts 2, 4, joining the sections, always cement the longer
8 to it, matching the tabs d of parts 2 and 8. Let tabs first. Five sections form a ring with the
the cement set and give the edge between the B pentahedral dimple in the (0) arrangement of
and W a good crease downwards. Then join the colours in the centre. The remaining tabs require
third, the G, Y, W section to the R, Y, W some care and patience, some being so small
section in the same way, tab d to tab d. Let the that you may find they do not need cement. The
cement set and give the edge between the R and neatness of the finished model depends on the
W a good crease downwards. Now manoeuvre attention you give to each detail. A total of
the final G and W edges into position and twenty sections are needed to complete the
cement the tabs d, using clamps if necessary at model. If you do all the work alone you may
the sharp overhanging edge. If you have been expect to spend about 50 hours on this one!

188
21:38:55 BST 2016.
CBO9780511569746.022
 117 Great inverted retrosnub
icosidodecahedron

This polyhedron is truly remarkable in its com-

plexity. Deep pentahedral cups display deeply
recessed decagrammic rosettes which close off
their tapered central portions. These cups have
steep outside facial planes that are fantastically
intricate. The vertex figure shows how two tri-
angle faces meet just beyond the central portion
of a star arm. This introduces very slender spiked
wedges whose vertices coincide with the vertices
of the pentagrams and which then continue
down toward the central portions of the penta-
gram faces but are cut into two more segments
before disappearing into the interior of the solid.
You will see this better from the model once you
|2f I start making it than from any description of it in
(20 + 60){3} + 12{f} words. So if you want to attempt this one, here
1-16000 30093 is an assembly technique.

Pentagram
fl=12-5
= 13 5
= 40

Snub triangle
Edge length = 50 cm
a=\25 «=1 0
6 = 4-0 />=8-5
c = 13-5 ?=8-5
</=8-5 r=3-0
e=\0 5=12
/=3-2 /=3-8
£=3-8 «=7-8
h = 10 17 = 2-1
y=30 w=2-7
k = 70 x=3-5
/=3-8 >>=40
m = 3-2 z=13-4

189
21:38:55 BST 2016.
CBO9780511569746.022
 Begin work with the interior parts of one Part i
pentahedral cup. Throughout the following des-
cription it is assumed that you will be able to
recognize the parts by their shape and to see
where they are found in their respective facial
planes. As for colour, all pentagram planes are
W, one set of triangles are all T, and the other
set of triangles take the icosahedral arrangement
of five colours. Only one section of part I is
shown. First prepare the rosette parts and cement
them to their respective larger parts at tabs
marked a. Fold the rosette parts alternately up
and down, up between the central Y and T parts
whose tabs band c then fit the corresponding tabs
on the larger Y part. Tab d then matches tab d.
Prepare five of these sections in the (0) arrange-
ment of the colours, the rosettes following their
own cyclic permutation. The colour O is shown,
so repeating this with the others the order is:
O R G Y B, while the larger parts follow the
order: Y B O R G respectively. The five sections
are cemented in a ring, the tab at x from one is
cemented to the tab at y from the neighbouring
part, and so on around. This completes part i:
the pentahedral cup with the decagrammic rosette
at the bottom.
One of the five sections of part n is laid out
opposite. The W parts belong to one star arm, the
T and B parts between them form a slender spiked
wedge cutting through the face of the star arm.
First cement these four upper parts. Notice that
the W part on the left has a shaded area joining
a small attached triangle. This should not be cut.
Later the second segment of the spiked wedge
will be cemented at x to cover this area. But
before doing this it is easier to cement the paired
triangles of colour B into place along the tabs
marked a, b, d, f. Next the second segment of
the spiked wedge is cemented in place. When
this is done you will see how this spiked wedge
seems to penetrate the paired triangles through
their shaded areas which should also be left un-
cut. The paired triangles meet at a very acute
angle along their common edge and they are
held in a rigid position by the V-cut near the base
of the star arm. Next cement the Y triangle at
tabs marked 1 and m. When this is done turn the

190
21:38:55 BST 2016.
CBO9780511569746.022
Part II

191
21:38:55 BST 2016.
CBO9780511569746.022
Part in

21:38:55 BST 2016.

CBO9780511569746.022
 Y triangle up and cement the next set of tabs, v This completes one section, namely one ring
and w, of Y to B. You will then see how the end of five polyhedral vertices. Twelve such sections
part of the second segment of the spiked wedge are needed to complete the model. To join these
covers the shaded area of the Y triangle, so this cement the tab h in part m of one section to
should also be left uncut. Repeat this section five tab g of another section. A third section added
times, each in its permutation of colours. The to the first two will show that this portion of the
five sections are then cemented to part i, first by pentagram planes near the tabs h and g forms a
the tabs p and when these are set, by the tabs q. shallow trihedral dimple, occurring twenty
You will see that part n has the effect of filling times on the completed model. Your patience
the spaces between the steeply projecting points and perseverance will have to hold out for more
of part i, revealing more clearly but still in an than 100 hours if you want a complete model of
unfinished way a ring of five polyhedral vertices. your own.
These polyhedral vertices will be finished or However, you can introduce some com-
completed by the addition of part in, now to be promises to simplify and shorten the work. You
described. might for example design the parts to omit the
One of the five sections of part in is laid out decagrammic rosette and all the wedges. Your
opposite. All the shaded areas should be left un- model will then have the same vertices as the
cut because various wedges will cover and so original, but only some of the major portions of
hide these areas. Also full tabs are shown for the the facial planes will be left. A set of simplified
W part because these will be needed for joining parts is shown below, together with a photograph
sections and it is important to have them done of this simplified model.
correctly. Notice that the largest part, B in
colour, has a smaller B part attached to it. This
means it can be cut as one piece of paper, but a
T part comes between the two small B parts, as
shown. The W part will give some rigidity to the
T and B upper parts, shaping them into a
shallow groove. Once you have cemented these
parts, the two lower wedges are assembled and
cemented as shown, the shaded areas marked y
are cemented to each other and then the shaded
areas marked z. These wedges can be done as
small irregular polyhedra in their own right, but
generally their end faces need not be closed. The
T and Y wedge is the third segment of the
slender spiked wedge lying on the face of the star
arm, the paired O parts will eventually be the
continuation of the paired O triangles, like the
paired B triangles in part n. This section is
repeated five times using the appropriate permu-
tation of colours. Then each is cemented so tabs
s and r of part m go with those at s and r of
part II. A pair of very tiny triangles forming a
very small wedge is shown in part in, but this
need not be added until further sections are
completed. In fact it is so small that for practical
reasons it could be omitted and never be missed.

193
21:38:55 BST 2016.
CBO9780511569746.022
118 Small inverted retrosnub
icosicosidodecahedron

This complex polyhedron has one thing in com-

mon with 110, namely it has facial symmetry as
you can easily see from the drawings of the facial
planes. But here you will notice a great deal more
complexity. Hence for this model, as for the
previous one, 117, you will need unusual
patience and perseverance to complete a model.
The assembly method outlined here will make
use of only three colours, Y, R, W, one for each
of the three kinds of facial planes.

Iff!
(40 + 60){3} + 12{f}
1-16138 96003

Edge length = 20 cm
a = 4-0
b=\0
c = 0-5
d=0-5
e = 0-4
/=7-0

194
21:38:55 BST 2016.
CBO9780511569746.022
 Begin with three sets of the parts shown as
part i, cementing the tab a of one part to the
tab b of another, then c from one to d from
another. After the third set has been added, you
will have a deep cup whose trihedral bottom is
composed of three W rhombi and its sides of six
Y quadrilaterals. Twenty of these will be needed
to complete the model.
A good way to proceed with the construction
of part II is to begin with the larger parts shown as
(a) on p. 196; the layout of pieces is shown on
p. 197. The R quadrilateral forms a dihedral
groove with the large Y piece. The small R and W
pieces form a small wedge and when this is
cemented in place at the corresponding tabs g, h,
k, it gives some rigidity to the R quadrilateral.
The pieces at the bottom of (a) will eventually fold
up, but before this is done it is better to assemble
all the small pieces in (b). These turn out to be a
sort of butterfly embossed on a shallow dihedral
groove. The plan for this is shown in (c) where the
dihedral groove is cross-hatched. The quadri-
lateral and the Y triangle at the top of part n (b)
can be left uncut as one piece of paper. This Y
portion should be given a very sharp fold down-
wards to make cementing easier or even un-
necessary.
Once you have completed the unit shown in

Part i

195
21:38:55 BST 2016.
CBO9780511569746.022
196
 part II (b) it is cemented first by tab p of part n (a).
Turn the Y parts of part n (b) sharply down to
form a very acute dihedral angle, and cement the
tabs at y. You will then find that the tabs v and w
can more easily be joined or may not need any
(a) (b) cement. The tabs r can be cemented next and then
the very small W triangle positioned to close the
bottom of the long narrow trihedral hole by
matching the tabs marked s. For practical reasons
this triangle may be omitted, since it is almost
impossible to see from the outside.
The bottom half of part n (a) may now be
folded up and the appropriate tabs at p are
again cemented first, then those at y. Finally the
small hole at the bottom is closed with the kite
shaped W part, shown in n (a), tab q going with
tab q at the bottom of n (b). All of part n will thus
give you a large butterfly section. Three of part II
are cemented around each part I joined by the tabs
marked i. Since the model calls for a total of
twenty of part i, you will need sixty of part II.
And part m is still to come. You can of course
(c) proceed to part m at once and join parts as they
are completed.
One section of part m is set out on p. 198. The
Part II pieces of part in (b) form a dihedral wedge and
this is cemented by tabs a, b, c, d, f to the larger
triple R pieces of part m (a). It is easiest to begin
with tab b, then the R parts with tabs d and f are
folded up to form a deep groove surrounding
the W part of part in (b). The pieces shown in
part in (c) and (d) are wedges. It is best to
(a) assemble these as shown, leaving the shaded
parts x and y uncut. You can then spread
cement over the shaded areas of these wedges
and cement them in place over the corresponding
areas of part in (a). The bottom part of part in
(a) may now be folded up. The wedges will then
be deeply embedded in the narrow groove that is
formed.
Repeat the instructions for part in (a), (b), (c),
(d) to build a second section. Then these two are
joined at the tabs marked j , using the piece
shown in part in (e) as a connector. Part HI is
(e) now complete. It forms a sort of wedge-shaped
sandwich, the 'bread' being two equilateral tri-
Part in angles of colour R, joined at their bases and

197
21:38:55 BST 2016.
CBO9780511569746.022
 Part in

198
O p
diverging slightly toward their vertices to show colours for only the first cup. The others would
the 'meat', an assortment of smaller wedges. be derived from the usual permutations. You
You will need a total of thirty of these wedge- could work out the colours for parts n and in by
shaped sandwiches to complete the model. Once following the facial planes through from part i as
you have completed three of them you can of the work progresses.
course cement them around the completed parts Admittedly this polyhedron will take a long
i and II. You will then have completed three time to assemble. As in the previous model you
polyhedral vertices and have the beginnings of can expect to spend more than 100 hours on this
three others. As the model grows it is not hard to one. However, some simplifications are possible
see how parts i, n, and in are fitted together. here also. They are as follows: in part in omit
Parts II and in alternate in a ring, their edges the small wedges of part in (c) and (d), and
forming a set of ten lines sloping gently inwards simply make in (b) pointed at the bottom near tab
and meeting at the central point of a dodeca- y of part m (a); in part II omit the embossed
hedral section. This would suggest an alternative butterfly assemblage of part n (c) and make n (c)
assembly technique. The long edges of all the a simple shallow dihedral groove pointed at the
parts make any order of assembly easy. bottom filling in the area near q of n (a). Part I
An alternative colour arrangement is sug- remains the same. This simplified model will
gested by the twenty cups of part i, namely the have the same vertices as the original and only
usual icosahedral arrangement. The six quadri- some small, scarcely noticeable portions of the
laterals of part i could be YB, YG YR, to name facial planes will be missing.

199
21:38:55 BST 2016.
CBO9780511569746.022
119 dodecahedron
Great dirhombicosi-

This polyhedron is remarkable in more ways

than one. First the symbol | f -J 3 f shows
that it differs from every other uniform poly-
hedron, in that all the others have a symbol
made up of only three numbers, either integers
or fractions or both. These three numbers relate
them to Schwarz triangles on a spherical surface.
The existence of this polyhedron indicates that
there is in general no reason for the restriction to
triangles. Does this mean that possibly other
uniform polyhedra may still be discoverable ? A
quotation from the paper by Coxeter and others,
Uniform polyhedra (p. 427), will answer this
question: ' We can only say that higher spherical
polygons would have to satisfy various con- Iff 31
ditions which are almost always incompatible.' 40{3} + 60{4} + 24{f}
So the most that can be said is that the list is V2
probably complete with this as the seventy-fifth
of the set of uniform polyhedra.

200
21:38:55 BST 2016.
CBO9780511569746.022
 Another way in which this polyhedron differs
from all the others is that it is the only known
polyhedron that has more than six faces at each
vertex. In fact it has eight. These faces occur in
coplanar pairs, twelve pairs of pentagrams,
twenty pairs of triangles and thirty pairs of dia-
metral squares. The vertex figure shows how
they are ordered in rotation at each vertex; the
squares occurring alternately with the other
faces.
In making a model of this polyhedron a
method of assembly by dodecahedral sections is
suggested. The central dimples of these sections
are reminiscent of similar dimples that occur in
the compound of ten tetrahedra, one of the
icosahedral stellations. Begin by cementing a
ring of five of part 1 in the icosahedral (0)
arrangement of colours. You should have no
difficulty in identifying this part and the others
referred to, from the drawings of the full facial
planes. When the ring is completed you will have a
dimpled five-pointed star.
It will be simplest to make all the squares one
colour, T. So the next step is to cement the small
V-shaped pieces, part 2, between the dimpled star
arms, first by the tab a and when this is well set, by
the tab b. You should now have a lip all around
the dimpled star, like the lip on a slip-cap cover.
Part 3 is joined next, by the tabs at c and d. The
colours of part 3 should correspond to those of
part 1. They are easy to get right because you
can see them as the continuation of the facial
planes of part 1, half a dimpled star arm forming
a dihedral wedge apparently protruding through
these planes. When this is done cement tabs t
and you will see that the spaces between the
dimpled star arms have been filled; the outer
edges at tabs e and f form a skew decagon. This
completes the dimpled star, the central portion
of one dodecahedral section.
Continue now with the assembly of parts 4, 5,
6, 7 and 8. Parts 4, 5, 7 all belong to the facial
planes of the squares, so their colour is T.
Dotted lines are scored on the reverse side and
the fold is made upwards. Part 6 belongs to the
triangle planes and hence it will require colour

201
21:38:55 BST 2016.
CBO9780511569746.022
 m

pairs. Although the drawing shows the 6 joined sort of butterfly section, deeply grooved between
to 5, and 7 to 8, these parts must actually be parts 4 and 5, with 6 and 7 forming a wedge
prepared separately because they are different apparently penetrating the surface of 4 through
colours. The colours for 6 are not hard to get the shaded area. Part 8 is now added, by joining
right once you see the permutation pattern. For the tabs at j and k. Part 8 is W, since it belongs
the first set of five, the (0) set, they go like this: to one of the paired pentagram planes. All these
YR BG OY RB GO. The order is Y, B, O, R, G will be the same colour, W.
for the left member of each pair, starting at the When you have completed five of these butter-
beginning of the line. Then, Y, B, O, R, G for the fly sections cement them around the dimpled
right member but starting at the third pair and star, tabs e and f of part 5 to e and f of part 3;
continuing the cyclic order from the end of the namely at the skew decagon edges. Cement one
line to the beginning. The other sections later set of tabs at a time. Say those at e first, let them
follow the same pattern. For example the (1) set up well, then give the butterfly section a good
section is: BO YG RB OY GR where the cyclic crease downwards and cement the tabs at f. You
order is B, Y, R, O, G. And so on for the rest. can now iuse clamps at these edges because there
In constructing this section, leave tabs all is an acute dihedral angle.
round for all parts. The shaded area of part 4 As this is completed you will see the outer
need not be scored or marked in any way and it edges forming a new skew decagon at the tabs 1
needs no tab. The next step is, as in the case of and m. You will also see that the fold between
116, difficult to describe in words without actual parts 4 and 5 from one butterfly section come
demonstration, but some trial and error on your into contact, or nearly so, with the corresponding
part should lead to success. Start with the left- fold in an adjacent section. At this stage you
hand side. Give parts 6 and 7 a good crease might want to join these parts along the line of
downwards to form a triangular wedge or di- contact, but this is not really necessary. Star
hedral angle. Spread a drop or two of cement on arm sections must still be added, as will be
the tabs at g and h. Then fold 4 and 5 up and explained presently, and then a third line of
manoeuvre the tabs g and h on to the shaded area contact will appear in the same place. Since all
of 4. Clamp with tweezers and allow the cement this will eventually be hidden inside the model
to set. Then perform the same operation on the and since the completed model will have suffi-
right. If you have succeeded you will now have a cient rigidity without joining these parts along

202
21:38:55 BST 2016.
CBO9780511569746.022
this line of contact, it simplifies the work to sections. This central hole is closed with the
proceed to the next step. section composed of parts 12 and 13. The six
The star arm sections, parts 9 and 10, belong quadrilaterals, part 12 (all T) belong to six dif-
to the paired pentagram planes and part 11 to ferent paired square planes, forming the sides of a
the paired squares. So 9 and 10 are both W and cup which is pointing directly toward the centre of
11 is one piece, colour T, scored along the the polyhedron. It is cut off, not very far down,
dotted line. Leave tabs all around as usual. The by the isosceles hexagon, part 13 (colours Y, B,
tabs at n and p are of special design. The pro- O, R, G), which belongs to the exact centre
truding segments will later join one section to portion of the paired triangle planes.
the next to form a single tab with its neighbour. The best procedure here is to make five of
The fold in 11 forms the third line of contact these cups (twenty in all will be needed) and to
referred to above. add them immediately to the (0) dodecahedral
This star arm section is very simple. First it section just completed. You can determine the
will be best to cement the tabs q and r to form a colour of 13 by aiming the point z of part 13 at
deep dihedral groove. Then cement this section the point z of part 1, and thus getting the
by tabs 1 and m to tabs 1 and m of the two parts 4. colours to correspond.
Cement 1 first and, when this is set, cement m. Twelve dodecahedral sections are needed for
Once you have done this all around the skew the complete model. Always fill in the V-cuts left
decagon edges, the protruding tabs at n and p by part 9 around these sections as each part is
are joined across the pointed end of part 8. This cemented in place. The assembly will remind you
completes one dodecahedral section. The joined of the way in which the regular dodecahedron is
tabs, n and p, now give this section distinctly assembled, but what a difference! Where the
pentagonal edges, except that the corners have regular dodecahedron has twelve faces, here
V-cuts left from part 9. If you were to complete there are twelve multifaceted sections of a most
three dodecahedral sections, joining them along intricate design. Truly, a remarkable poly-
the tabs n and p, you would see that a hexagonal hedron ! It will take you at least 50 hours work
hole would be left centrally between these to complete this model.

21:38:55 BST 2016.

CBO9780511569746.022
A final comment

Even if you have now made only a few of the You may be wondering why the stellation pro-
non-convex uniform polyhedra, you can see cess was treated so thoroughly in Section n. This
from the models what properties belong to the was done, first of all, because in some ways it is
set as a whole. The most interesting fact is that breaking new ground. Secondly, it is intrinsic-
all of them are derived from Schwarz triangles— ally a simple process, although it may indeed
except one, model 119. It is exceptional in lead to polyhedral forms almost too numerous to
another way. It is the only known polyhedron detail. With enough perseverance you can dis-
that has more than six polygons surrounding cover any number of these forms by yourself.
each vertex, four squares alternating with two Lastly, it should help you to understand another
triangles and two pentagrams. All the squares kind of stellation, namely edge stellation. Edge
are on planes through the solid's centre of stellation is that in which the edges of a polyhe-
symmetry. It is classified as a snub polyhedron dron are produced to generate the edges of a new
because here the squares may be regarded as polyhedron. A simple example is found in the
snub faces instead of the usual triangles as in dodecahedron, whose edges if produced generate
other cases. The existence of this polyhedron the edges of the small stellated dodecahedron.
indicates that spherical polygons as well as Stick models can show this very plainly. Many
spherical triangles may give rise to other uniform uniform polyhedra are edge stellations of other
polyhedra. However it is a complex task to in- uniform polyhedra. But it must be left for you to
vestigate the possibilities. It still remains to be pursue this matter further on your own.
done.

Epilogue

This book has presented only some polyhedral beyond the models presented here, there are
forms. For anyone acquainted with the field more, and more and more! The object of an
there are obvious omissions, the two infinite sets investigator would not be to multiply forms but
of prisms and antiprisms, all the Archimedean to arrive at the underlying mathematical theory
duals (except for the three given on pp. 6-8), that unifies and systematizes whole sets of poly-
and many other polyhedral forms. Among the hedral forms.
Archimedean duals two are especially note- From this point of view the mathematical in-
worthy, the rhombic dodecahedron and the vestigation takes its origin from an inductive
rhombic triacontahedron. The former is given in process, akin to the scientific method; namely, to
Cundy and Rollett along with the stellated observe individual instances of any phenomenon,
forms worked out by Dorman Luke. Stellated then to classify and systematize in order to
forms of the latter are presented in summary arrive at general principles which serve as the
fashion without drawings except for the stella- basis of a deductive process. Many people are
tion pattern by J. D. Ede in the Mathematical not aware of this aspect of mathematics, but the
Gazette, XLII (1958). All the Archimedean duals history of mathematics is full of instances bear-
can be stellated, as indeed any polyhedron can. ing this out. (See G. Polya, Mathematics and
In the light of what you have now learned, you plausible reasoning.)
can discover the stellation patterns by yourself, So, to end on the same metaphor as that used
and thus make models of all these polyhedra in the preface, the road still stretches on before
using the methods and techniques described. So you. Why don't you continue your journey?

204
21:39:16 BST 2016.
CBO9780511569746.023
A final comment

Even if you have now made only a few of the You may be wondering why the stellation pro-
non-convex uniform polyhedra, you can see cess was treated so thoroughly in Section n. This
from the models what properties belong to the was done, first of all, because in some ways it is
set as a whole. The most interesting fact is that breaking new ground. Secondly, it is intrinsic-
all of them are derived from Schwarz triangles— ally a simple process, although it may indeed
except one, model 119. It is exceptional in lead to polyhedral forms almost too numerous to
another way. It is the only known polyhedron detail. With enough perseverance you can dis-
that has more than six polygons surrounding cover any number of these forms by yourself.
each vertex, four squares alternating with two Lastly, it should help you to understand another
triangles and two pentagrams. All the squares kind of stellation, namely edge stellation. Edge
are on planes through the solid's centre of stellation is that in which the edges of a polyhe-
symmetry. It is classified as a snub polyhedron dron are produced to generate the edges of a new
because here the squares may be regarded as polyhedron. A simple example is found in the
snub faces instead of the usual triangles as in dodecahedron, whose edges if produced generate
other cases. The existence of this polyhedron the edges of the small stellated dodecahedron.
indicates that spherical polygons as well as Stick models can show this very plainly. Many
spherical triangles may give rise to other uniform uniform polyhedra are edge stellations of other
polyhedra. However it is a complex task to in- uniform polyhedra. But it must be left for you to
vestigate the possibilities. It still remains to be pursue this matter further on your own.
done.

Epilogue

This book has presented only some polyhedral beyond the models presented here, there are
forms. For anyone acquainted with the field more, and more and more! The object of an
there are obvious omissions, the two infinite sets investigator would not be to multiply forms but
of prisms and antiprisms, all the Archimedean to arrive at the underlying mathematical theory
duals (except for the three given on pp. 6-8), that unifies and systematizes whole sets of poly-
and many other polyhedral forms. Among the hedral forms.
Archimedean duals two are especially note- From this point of view the mathematical in-
worthy, the rhombic dodecahedron and the vestigation takes its origin from an inductive
rhombic triacontahedron. The former is given in process, akin to the scientific method; namely, to
Cundy and Rollett along with the stellated observe individual instances of any phenomenon,
forms worked out by Dorman Luke. Stellated then to classify and systematize in order to
forms of the latter are presented in summary arrive at general principles which serve as the
fashion without drawings except for the stella- basis of a deductive process. Many people are
tion pattern by J. D. Ede in the Mathematical not aware of this aspect of mathematics, but the
Gazette, XLII (1958). All the Archimedean duals history of mathematics is full of instances bear-
can be stellated, as indeed any polyhedron can. ing this out. (See G. Polya, Mathematics and
In the light of what you have now learned, you plausible reasoning.)
can discover the stellation patterns by yourself, So, to end on the same metaphor as that used
and thus make models of all these polyhedra in the preface, the road still stretches on before
using the methods and techniques described. So you. Why don't you continue your journey?

204
21:39:33 BST 2016.
CBO9780511569746.024
 A faceted form of the small stellated dodecahedron
(from Bruckner: VIII, 14)

A stick model of the icosahedron

A stick model of the dodecahedron

205
21:39:33 BST 2016.
CBO9780511569746.024
References

Bruckner, M. Vielecke und Vielflache. Tuebner, 1900. Ede, J. D. Mathematical Gazette (1958), XLII.
Coxeter, H. S. M. Introduction to geometry. John Heath, T. L. Euclid's elements, vol. 3. Dover, 1956.
Wiley and Sons, 1961. Hilbert, D. and Cohn-Vossen, S. Geometry and the
Coxeter, H. S. M. Regular polytopes, 2nd ed. imagination. Chelsea, 1956.
Macmillan, 1963. Lines, L. Solid geometry. Dover, 1965.
Coxeter, H. S. M., 'Polyhedra', chapter 5 in Ball, Lysternick, L. A. Convex figures and polyhedra.
W. W. R. Mathematical recreations and essays. Dover, 1963.
Macmillan, 1965. Polya, G. Mathematics and plausible reasoning.
Coxeter, H. S. M., Du Val, P., Flather, H. T. and Vol. I. Induction and analogy in mathematics.
Petrie, J. F. Thefifty-nineicosahedra. University Vol. II. Patterns of plausible inference. Oxford,
of Toronto, 1951. 1955.
Coxeter, H. S. M., Longuet-Higgins, M. S. and Steinhaus, H. Mathematical snapshots. Oxford, 1960.
Miller, J. C. P. 'Uniform polyhedra'. Phil. Wenninger, M. J. Mathematical Gazette (1968), LII.
Trans. 1954, 246A, 401-50. Wenninger, M. J. Polyhedron models for the class-
Cundy, H. M., and Rollett, A. P., Mathematical room, N.C.T.M. Publication, 1966.
models, 2nd ed. Oxford, 1961.

206
21:40:03 BST 2016.
CBO9780511569746.025
List of Models

The figures in parentheses refer to the number of the 44 Second stellation of the cuboctahedron 69
model in the paper by Coxeter et al. 'Uniform 45 Third stellation of the cuboctahedron 70
Polyhedra'. 46 Final stellation of the cuboctahedron 71
47 First stellation of the icosidodecahedron 76
1 Tetrahedron (15) Page 14 48 Second stellation of the icosidodeca-
2 Octahedron (17) 15 hedron 77
3 Hexahedron (18) 16 49 Third stellation of the icosidodecahedron 78
4 Icosahedron (25) 17 50 Fourth stellation of the icosidodeca-
5 Dodecahedron (26) 19 hedron 79
6 Truncated tetrahedron (16) 20 51 Fifth stellation of the icosidodecahedron 80
7 Truncated octahedron (20) 21 52 Sixth stellation of the icosidodecahedron 81
8 Truncated hexahedron (21) 22 53 Seventh stellation of the icosidodeca-
9 Truncated icosahedron (27) 23 hedron 82
10 Truncated dodecahedron (29) 24 54 Eighth stellation of the icosidodeca-
11 Cuboctahedron (19) 25 hedron 83
12 Icosidodecahedron (28) 26 55 Ninth stellation of the icosidodeca-
13 Rhombicuboctahedron (22) 27 hedron 84
14 Rhombicosidodecahedron (30) 28 56 Tenth stellation of the icosidodeca-
15 Rhombitruncated cuboctahedron (23) 29 hedron 85
16 Rhombitruncated icosidodecahedron (31) 30 57 Eleventh stellation of the icosidodeca-
17 Snub cube (24) 31 hedron 86
18 Snub dodecahedron (32) 32 58 Twelfth stellation of the icosidodeca-
19 Stellated octahedron or ' Stella oct- hedron 87
angula' 37 59 Thirteenth stellation of the icosidodeca-
20 Small stellated dodecahedron (43) 38 hedron 88
21 Great dodecahedron (44) 39 60 Fourteenth stellation of the icosidodeca-
22 Great stellated dodecahedron (68) 40 hedron 89
23 Compound of five octahedra 43 61 Compound of the great stellated dodeca-
24 Compound of five tetrahedra 44 hedron and the great icosahedron 90
25 Compound of ten tetrahedra 45 62 Fifteenth stellation of the icosidodeca-
26 Triakis icosahedron—First stellation of hedron 91
the icosahedron 46 63 Sixteenth stellation of the icosidodeca-
27 Second stellation of the icosahedron 47 hedron 92
28 Third stellation of the icosahedron—A 64 Seventeenth stellation of the icosidodeca-
deltahedron 48 hedron 93
29 Fourth stellation of the icosahedron 49 65 Eighteenth stellation of the icosidodeca-
30 Fifth stellation of the icosahedron 50 hedron 94
31 Sixth stellation of the icosahedron 52 66 Final stellation of the icosidodecahedron 95
32 Seventh stellation of the icosahedron 53 67 Tetrahemihexahedron (36) 101
33 Eighth stellation of the icosahedron 54 68 Octahemioctahedron (37) 103
34 Ninth stellation of the icosahedron 55 69 Small cubicuboctahedron (38) 104
35 Tenth stellation of the icosahedron 56 70 Small ditrigonal icosidodecahedron (39) 106
36 Eleventh stellation of the icosahedron 58 71 Small icosicosidodecahedron (40) 108
37 Twelfth stellation of the icosahedron 59 72 Small dodecicosidodecahedron (42) 110
38 Thirteenth stellation of the icosahedron 60 73 Dodecadodecahedron (45) 112
39 Fourteenth stellation of the icosahedron 61 74 Small rhombidodecahedron (46) 113
40 Fifteenth stellation of the icosahedron 62 75 Truncated great dodecahedron (47) 115
41 Great icosahedron (69) 63 76 Rhombidodecadodecahedron (48) 116
42 Final stellation of the icosahedron 65 77 Great cubicuboctahedron (50) 118
43 First stellation of the cuboctahedron— 78 Cubohemioctahedron (51) 120
Compound of a cube and an octa- 79 Cuboctatruncated cuboctahedron (52) 121
hedron 68 80 Ditrigonal dodecahedron (53) 123

207
21:40:16 BST 2016.
ebook.jsf?bid=CBO9780511569746
81 Great ditrigonal dodecicosidodecahedron 101 Great dodecicosahedron (79) 156
(54) 125 102 Great dodecahemicosahedron (81) 158
82 Small ditrigonal dodecicosidodeca- 103 Great rhombihexahedron (82) 159
hedron (55) 126 104 Quasitruncated great stellated dodeca-
83 Icosidodecadodecahedron (56) 128 hedron (83) 161
84 Icosidodecatruncated icosidodecahedron 105 Quasirhombicosidodecahedron (84) 162
(57) 130 106 Great icosihemidodecahedron (85) 164
85 Quasirhombicuboctahedron (59) 132 107 Great dodecahemidodecahedron (86) 165
86 Small rhombihexahedron (60) 134 108 Quasitruncated icosidodecahedron (87) 166
87 Great ditrigonal icosidodecahedron (61) 135 109 Great rhombidodecahedron (89) 168
88 Great icosicosidodecahedron (62) 137 110 Small snub icosicosidodecahedron (41) 172
89 Small icosihemidodecahedron (63) 140 111 Snub dodecadodecahedron (49) 174
90 Small dodecicosahedron (64) 141 112 Snub icosidodecadodecahedron (58) 177
91 Small dodecahemidodecahedron (65) 143 113 Great inverted snub icosidodecahedron
92 Quasitruncated hexahedron (66) 144 (73) 179
93 Quasitruncated cuboctahedron (67) 145 114 Inverted snub dodecadodecahedron (76) 180
94 Great icosidodecahedron (70) 147 115 Great snub dodecicosidodecahedron
95 Truncated great icosahedron (71) 148 (80) 183
96 Rhombicosahedron (72) 149 116 Great snub icosidodecahedron (88) 186
97 Quasitruncated small stellated dodeca- 117 Great inverted retrosnub icosidodeca-
hedron (74) 151 hedron (90) 189
98 Quasitruncated dodecahedron (75) 152 118 Small inverted retrosnub icosicosidode-
99 Great dodecicosidodecahedron (77) 154 cahedron (91) 194
100 Small dodecahemicosahedron (78) 155 119 Great dirhombicosidodecahedron (92) 200

208
21:40:16 BST 2016.
ebook.jsf?bid=CBO9780511569746