ro

OU
OQ

160977

5m
1

CO

OSMANIA UNIVERSITY LIBRARY

b. it~/2
.

C//^f (j/

Accession No, ?<</*

'Author

This book should bcreturncd on or before the date last marked beUm,

GEOMETRY OF FOUR DIMENSIONS

THE MACMILLAN COMPANY

NEW YORK
-

ATLANTA

BOSTON CHICAGO DALLAS SAN FKANUSCO

CO., LIMITED BOMBAY CALCUTTA MELBOURNE

MACMILLAN &
LONDON

THE MACMILLAN

CO. OF TORONTO

CANADA,

LTD.

GEOMETRY
OF FOUR DIMENSIONS

BY

HENRY PARKER MANNING,

PH.D.

ASSOCIATE PROFESSOR OF PURE MATHEMATICS IN BROWN UNIVERSITY

Weto gork

THE MACMILLAN COMPANY

1914
All rights reserved

COPYRIGHT,

1914,

BY

THE MACMILLAN COMPANY.

Published September, 1914.

Set up and electrotyped.

J. S.

Berwick & Smith Co. Gushing Co. Norwood, Mass., U.S.A.

PREFACE
THE object and plan of this book are explained in the I had hoped to give some account Introduction (page 16).
of the recent literature, but this

that has already taken several years. list of technical terms as found in a

would have delayed work I have prepared a few of the more

out errors.

familiar writings, very incomplete, and, I fear, not withThe list may be of service, however, to those
to consult the authors referred to
;

who wish

it

will also

indicate something of the confusion that exists in a subject

whose nomenclature has not become

necessary for

fixed.

It

has been
of

me

to introduce a considerable

number

terms, but most of these have been formed in accordance

with simple or well-established principles, and no attempt has been made to distinguish them from the terms that

have already been used.

I

am

indebted to the kindness of Mr. George A. Plimp-

an opportunity to examine his copy of Rudolph's Coss referred to on page 2. I am also under many obligations to Mrs. Walter C. Bronson of Providence, to Mr. Albert A. Bennett, Instructor at Princeton University, and to my colleagues, Professors R. C. Archibald and R. G. D. Richardson, from all of whom I have received valuable criticisms and suggestions. Many of the were found references in the first four pages by Professor Archibald several of these are not given in the leading bibliographies, and the reference to Ozanam I have not
ton of
for
;

New York

seen anywhere.
PROVIDENCE,
July, 1914.

HENRY

P.

MANNING.

CONTENTS
PAGE

INTRODUCTION

CHAPTER

THE FOUNDATIONS OF FOUR-DIMENSIONAL GEOMETRY

CHAPTER
INTRODUCTORY
I.

II

PERPENDICULARITY AND SIMPLE ANGLES

LINES PERPENDICULAR TO A HYPERPLANE

II.

ABSOLUTELY PERPENDICULAR PLANES

SIMPLY PERPENDICULAR PLANES

....
.

...
.

73

74
80
85

III.

IV.

PERPENDICULAR PLANES AND HYPERPLANES

90
95

V.

HYPERPLANE ANGLES
vii

viii

CONTENTS

CHAPTER

III

ANGLES OF TWO PLANES AND ANGLES OF HIGHER ORDER

PAGE
I.

II.

III.

THE COMMON PERPENDICULAR OF Two POINT GEOMETRY THE ANGLES OF Two PLANES
POLYHEDROIDAL ANGLES PLANO-POLYHEDRAL ANGLES

LINES

105

112

114
126
133

IV.

V.

CHAPTER
I.

IV

SYMMETRY, ORDER, AND MOTION

ROTATION AND TRANSLATION
141
II.

III.

IV.

V.
VI.

SYMMETRY ORDER MOTION IN GENERAL RECTANGULAR SYSTEMS

ISOCLINE PLANES

146
153

167

179

180

CHAPTER V
HYPERPYRAMIDS, HYPERCONES, AND THE HYPER-

SPHERE
I.

PENTAHEDROIDS AND HYPERPYRAMIDS

....

199

II.

III.

HYPERCONES AND DOUBLE CONES THE HYPERSPHERE

204
207

CHAPTER

VI

EUCLIDEAN GEOMETRY. FIGURES WITH PARALLEL

ELEMENTS
THE AXIOM OF PARALLELS
I.

221

PARALLELS

II.

THE "HYPERPLANE AT

INFINITY"

....

221

230

CONTENTS
III.

ix

HYPERPRJSMS

....
CHAPTER
VII

...
. .

PAGE

235
241

IV.

DOUBLE PRISMS
HYPERCYLINDERS
PRISM CYLINDERS AND DOUBLE CYLINDERS

V.
VI.

253

256

MEASUREMENT OF VOLUME AND HYPERVOLUME

HYPERSPACE
I.

IN

II.

VOLUME HYPERVOLUME

265

270

CHAPTER
I.

VIII

THE REGULAR POLYHEDROIDS

II.

THE FOUR SIMPLER REGULAR POLYHEDROIDS THE POLYHEDROID FORMULA

289
300

III.

RECIPROCAL POLYHEDKOIDS AND RECIPROCAL NETS OF POLYHEDROIDS

CONSTRUCTION
OF
1

303

IV.

THK REGULAR 6oo-HEDROin AND

20-11 EDKOID

THE REGULAR

317

TECHNICAL TERMS INDEX

327
335

INTRODUCTION
more than three dimensions is entirely a modern branch of mathematics, going no farther back
of

THE geometry
first

than the
of space.

however, some early

In the
first

part of the nineteenth century. There are, references to the number of dimensions

book

of the

Heaven
line

B.C.) are these sentences:

"The

of Aristotle (384-322 has magnitude in one

way, the plane in two ways, and the solid in three ways, and beyond these there is no other magnitude because the " There is no transfer into another kind, three are all," and like the transfer from length to area and from area to a
solid."*

Simplicius (sixth century, A.D.) in his

Commen-

admirable Ptolemy in his book On Distance well proved that there are not more than three distances, because of the necessity that distances should be
taries says,

"The

and that the distances defined should be taken along perpendicular lines, and because it is possible to take only three lines that are mutually perpendicular, two by which the plane is defined and a third measuring depth
defined,
;

so that

if

there were any other distance after the third

it

would be

Thus
there

entirely without measure and without definition. Aristotle seemed to conclude from induction that

is

no transfer into another magnitude, but Ptolemy

it." f

proved

* Aristoteles, De Caelo, ed. Prantl, Leipzig, 1881, 268a, 7 and 30. t Simplicii in Aristotdis De Caelo Commentaria, cd. Heiberg, Berlin, 1894, 7 ff > 33- Ptolemy lived about 150 A.D. The book on distance, irepi Smtrrdo-ew, is lost, and with it Ptolemy's " proof " except so far as it may be reproduced in the above quotation

from Simplicius.

INTRODUCTION
There
is

also in the early history of algebra a use of terms

analogous to those derived

etry,

from the plane and

solid

geom-

but applicable only to geometry of more dimensions. With the Greeks, and then in general with the mathematicians that came after them, a number was thought of as a line (of definite length), the product of two numbers
as a rectangle or plane and the product of three numbers as a parallelepiped or solid; or, if the numbers were equal, the product of two was a square and of three a cube. When
,

they began to study algebra, other terms were required

for the higher powers,

and so

in

Diophantus (third century)

we

find

square-square,

later times there

was

In square-cube, and cube-cube* a variation in the use of these terms.

the square-cube came to mean the square of the cube, or sixth power, while with Diophantus it means the square

Thus

times the cube, or fifth power. This change required the introduction of new terms for powers of prime orders, and,
in particular, for the fifth power,

which was

finally called

a sursolid .^

The

geometrical

conception

of

equations

and the geometrical forms

of their

solutions J hindered
vol. I,

* Cantor, Vorlesungen uber Geschichte der Uathematik,

3d ed

Leipzig,

1907, p. 470. f In the edition of Rudolph's Coss (algebra) revised by Stifel (Konigsberg, 1553, described by David Eugene Smith in Kara Artlhmetica, Boston, 1908, p 258) Surfifth power, Bsursohdum the seventh power, and so on (Part I, Paciuolo (about 1445-1514) in his Swnma de Anthmetica Geo~ metria Proportiom el Proportionalita, printed in 1494, uses the terms primo relato and sec undo relato (Cantor, Vorlesungen, etc vol. If, 2d ed 1000, p. 317) On the other

solidum denotes the

chap

5, fol

63)

hand, Vieta (1540-1603) follows Diophantus. He expresses all the powers above the third by compounds of quadrato and cubo, cubo-cubo-cubus being the ninth power

and elsewhere). The term geometry of Descartes (1596-1650). It is to be noted, however, that a product with Descartes always means a line of definite length derived from given lengths by proportions. Problems which lead to equations
(Francisci Vieta opera malhematica, Leyden, 1646, p. 3
sursolid occurs several times in the

"one degree more complicated than conies." Conies were called by the Greeks solid loci, and these more complicated curves were called by Descartes sursolid loci (La Gtomttrie. See pp. 20 and 29 of the edition published by Hermann, Paris, 1886). I Such solutions are given in the second and sixth books of Euclid's Elements.
of the fifth or sixth degrees require for their geometrical solution curves

^See Heath's edition, Cambridge University Press, 1908, vol.

I, p.

383.

EARLY REFERENCES TO DIMENSIONS

the progress of algebra with the ancients. Higher equations than the third were avoided as unreal,* and when the study
of higher equations forced itself upon mathematicians, it meant an impossible extension of geometrical notions, which met with many protests, and only in later times

gave way to a purely numerical conception of the nature

of algebraic quantities.

Thus

Stifel

(1486 ?-is67), in the

to

Algebra of

Rudolph

already referred

(footnote,

pre-

ceding page), speaks of "going beyond the cube just as if there were more than three dimensions," "which is," he adds, "against nature." f John Wallis (1616-1703) in his Algebra objects to the "ungeometrical" names

given to the higher powers.

He

calls

one of them a

"Monster
Centaure."

in

Nature, possible than a Chimaera or He says: "Length, Breadth and Thickness,

less

take up the whole of Space. Nor can Fansie imagine how there should be a Fourth Local Dimension beyond these

Three.

"J

Ozanam
two

(1640-1717),

after

speaking of the
of

product of

letters as a rectangle

and the product

three as a rectangular parallelepiped, says that a product of more than three letters will be a magnitude of "as many

dimensions as there are

letters,

but

it

will

because in nature we do not

know

of

only be imaginary any quantity which

has more than three dimensions."

Again, we find in the writings of some philosophers references to a space of four dimensions. Thus Henry More (1614-1687), an English philosopher, in a book published
in 1671, says that spirits

have four dimensions,^ and Kant

(1724-1804) refers in several places to the number of dimensions of space.

||

1896, pp. 544

* Matthicssen, Grundzugc der antikcn und modernen Algebra, 2d ed., Leipzig, and 921. f Part I, chap, i, fol. Q recto. J London, 1685, p. 126.
Dictionairc mathematiquc,

Amsterdam, 1691,
Pt.
I,

p. 62.
7, p.

T Enchiridion metaphysicum,
|(

chap. 28,

384.
if

For example, he says

in the

Critique of

Pure RMSJ.I, ''For

the intuition

INTRODUCTION

Finally, there is a suggestion made by certain writers that mechanics can be considered a geometry of four dimensions with time as the fourth dimension (see below,
p.

n).

This idea

is

1813),

who advanced

usually credited to Lagrange (1736it in his Theorie des fonctions analy-

* It is expressed, however, in published in 1797. " an article on Dimension" published in 17 54 by d'Alembert (1717-1783) in the Encyclopedic edited by Diderot and
tiques, first

himself.

D'Alembert attributes the

d'esprit de

suggestion to

"un

homme

ma

connaissance."f

These are the only ways in which we have found our

subject referred to before 1827.

In the period beginning with 1827 we may distinguish those writings which deal with the higher synthetic geometry from those whose point of view is that of analysis.
In synthetic geometry our attention
is

confined at

first

chiefly to the case of four dimensions, while in analysis

we

are ready for

n variables by the time we have confirst

sidered two and three.

So far as we know, the

contribution to the synthetic

made by Mobius, who points geometry out that symmetrical figures could be made to coincide if there were a space of four dimensions.! In 1846 Cayley
of four dimensions is
were a concept gained a posteriori
that, so far as hitherto observed,

we should not be able to say any more than no space has yet been found having more than three
.
.

dimensions" (translation by F Max Muller, 2d ed revised, Macmillan, 1905, p. 19). C. H. Hinton finds in four-dimensional space illustration and interpretation of the ideas of Plato, Aristotle, and other Greek philosophers (sec Fourth Dimension

London, 1904, chap. iv). *p. 223; (Euvres, vol. IX, Paris, 1881, p. 337. t See paper by R. C. Archibald, "Time as a Fourth Dimension," Bulletin of the American Mathematical Society, vol. 20, 1914, pp. 409-412. t He states very clearly the analogy with symmetrical figures in a plane and symmetrical groups of points on a line. Reasoning from this analogy, he says that the coincidence of two symmetrical figures in space would require that we should be able to let one of them make a rotation in space of four dimensions Then he adds, "Da aber ein soldier Raum nicht gedacht werden kann, so ist auch die Coin-

BEGINNINGS OF THIS GEOMETRY

makes use
of

geometry of four dimensions to investigate

method that is systematically developed by Veronese.* Cayley had al" in the Anthe a with title Chapters ready published paper
certain configurations of points, suggesting a
alytical

Dimensions/' f but as this paper contains no actual reference to such a geometry, we may

Geometry

of

(ri)

think of the paper of 1846 as the beginning of his published Some of the most interesting writings on this subject.

examples of the direct study of these geometries were given by Sylvester. In 1851, in a paper on homogeneous functions,! he discusses tangent and polar forms in ^-dimensional

geometry; in 1859, in some lectures on partitions, he and in 1863, in a makes an application of hyperspace memoir "On the Centre of Gravity of a Truncated Triangular Pyramid,"^ he takes up the corresponding
;

figures in four

for all of these figures, using analytic

and n dimensions and proves his theorems methods to some

Clifextent, but appealing freely to synthetic conceptions. ford also, about this time, makes a very interesting applica-

tion of the higher geometry to a problem in probability.

cidenz
in

||

diesem Falle unmoglich

la

"

(Der

barycentrische

Calcul, Leipzig,

1827,

140, p. 184).

*"Sur quelques th6oremes de

31,

pp

213-226

(in

particular,

pp.

geometric dc position," Crclle's Journal, vol. 217-218); Collected Mathematical Papers,

Cambridge, vol. I, 1889, No. 50. See also Veronese, Fondamcnti, etc. (the full title given below on p. 9), p. 690 of the German translation, and Veronese's memoir In introducing this method of reasoning, Cayley says: (mentioned on p. 8). "On peut en effet, sans recourir a aucune notion metaphysique a 1'egard de la possibilite de Tespace a quatre dimensions, raisonner comme suit (tout cela pourra aussi e"tre traduit facilement en langue purement analytique)."
is
. . .

t Cambridge Mathematical Journal, vol. 4, 1844; Math. Papers, vol. I, No. n. J Cambridge and Dublin Mathematical Journal, vol. 6, p. i ; Collected Mathematical Papers, Cambridge, vol. I, 1904, No 30. Outlines of these lectures are published in the Proceedings of the London Mathematical Society, vol. 28, 1896, p. 33; Mathematical Papers, vol. II, 1908, No. 26. Math^[ Philosophical Magazine, fourth series, vol. 26, Sept., 1863, pp. 167-183 ematical Papers vol. II, No. 65.
; t

\\Educational Times, Jan., 1866; Mathematical Reprints, vol. Mathematical Papers, Macmillan, 1882, p. 601.

6,

pp. 83-87;

INTRODUCTION

Quite independently of this beginning of its synthetic development, we find a notion of a higher geometry springCertain geometriing out of the applications of analysis. cal problems lead to equations which can be expressed with any number of variables as well as with two or
three. Thus, in 1833, Green reduces the problem of the attraction of ellipsoids to analysis, and then solves it for any number of variables, saying, "It is no longer confined

were to the three dimensions of space.' * Other writers make the same kind of generalization, though not
1

as

it

always pointing out so directly its geometrical significance, f It was but a step farther to apply the language of geometry
to
all

the forms and processes of algebra and analysis.

is clearly announced by Cauchy in 1847, memoir on analytical loci, where he says, "We shall call a set of n variables an analytical point, an equation or system of equations an analytical locus/' etc.f The most important paper of this period is that of Riemann, "On the Hypotheses which Lie at the FoundaIn this paper Riemann builds tions of Geometry." notion the of multiply-extended manifolds and their up

This principle

in a

measure-relations.

He
ds
is

discusses the

nature of the

line-

element ds when the manifold

variables.

When

is expressed by means of n to the square root of the sum equal

* Mathematical Papers of George Green, edited by N. M.

p. 188.

Ferrers, Macmillan, 1871,

" t C. G. J. Jacobi, De binis quibuslibet functionibus homogeneis," etc., Crelle's Journal, vol. 12, 1834, p. i; Cayley, two papers published in the Cambridge Mathematical Journal, vol. 3, 1841 ; Mathematical Papers, vol. I, Nos. 2 and 3 ; Schlafli,

"Ueber das Minimum des Integrals /(V</*i 2 + dx<? n Journal, vol. 43, 1852, pp. 23-36 "On the Integral ! dx dy
.

-M#n2),"
.
.

etc., Crette's

dz," etc., Quarterly

Journal, vois. 2 and 3, 1858-1860. % "Memoir sur les lieux analytiques," Comptes Rendus, vol. 24, p. 885. "Ueber die Hypothesen, welche der Geometric zu Grunde liegen," presented to the philosophical faculty at Gdttingen in 1854, but not published till 1866;

Gesammelte Werke, Leipzig, 1892, No. xiii, pp. 272-287; translated by Clifford in Nature, vol. 8, 1873, pp, 14 and 36; Mathematical Papers, No. 9, pp. 55-69.

RIEMANN, GRASSMANN

of the squares of the quantities dx, as in the ordinary plane and space, the manifold is fiat. In general there is a deviation

from

flatness, or

curvature;

and the simplest cases

is

are those in which the curvature

constant.

Riemann

points out that space

infinite

that, in fact,

may be unbounded without being it cannot be infinite if it has a con-

stant positive curvature differing at all from zero. We therefore attribute to Riemann the Elliptic Non- Euclidean

Geometry, which from this time on takes its place beside that other discovered by Bolyai and Lobachevsky. His paper has a bearing on our subject in two ways: in the
first

n dimensions is a space of n and dimensions, geometrical conceptions are clearly before mind the throughout the discussion and then the notion
place, his manifold of
;

of a curvature of space suggests at once a space of four

dimensions in which the curved three-dimensional space may lie. Soon after, it was shown by Beltrami that the
planimetry of Lobachevsky could be represented upon real
surfaces of constant negative curvature just as the Elliptic Two-dimensional Geometry is represented upon the sphere,

and the way was

fully opened for the study of spaces of constant curvature and of curvature in general.* Another work that has an important influence on recent

developments of hypergeometry, especially in its application to physical theories, is the Ausdehnungslehre of Grass-

mann,

first

published in 1844, though

di interpretazione della
6,

little

noticed at the

* Beltrami, " Saggio

matematiche, vol.
cost ante,"

geometria non-euclidea," Giornale di 1868; Opere, Milan, vol. I, 1902, pp. 374-405. Another memoir by Beltrami, "Teoria fondamentale degli spazii di curvatura

Opere, vol.

Annali di matematica pura ed applicata, Ser. 2, vol. 2, 1868-1869; I, pp. 406-429, develops and explains much in Riemann's paper that is difficult to understand. There are French translations of both memoirs by Houel, Anndes Sdentifiques de VEcole Normale Superieure, vol. 6, 1869.
Beltrami considers the representations of the three-dimensional geometries upon curved spaces as only analytic, while the representations of the two-dimensional See Opere, vol. I, p. 396 geometries upon surfaces of constant curvature are real.

and

p. 427.

8
time.

INTRODUCTION

His theory of extensive magnitudes is a vector analysis, and the applications which he makes to plane

geometry and to geometry of three dimensions can be made in the same way to geometry of any number of dimensions. The number of memoirs and books relating to geometry of four or more dimensions has increased enormously in In 1870, recent years. We can mention only a few. " his on Abstract Memoir Geometry/' Cayley published in which he lays down the general principles of w-dimensional geometry.* Another important contribution to the science was an unfinished paper "On the Classification of Loci" by Clifford. f An important paper by Nother on was published in iSyo.J birational transformations Other papers were published by Halphen in 1873 and by
Jordan in 1875,
tion of metrical
nates.

the latter giving a methodical generalizageometry by means of Cartesian coordi-

Perhaps the most important of all was a memoir Veronese published in 1882,^ in which he takes up a by
study of the properties of configurations, the quadratic
in

any number

of variables, the characteristics of curves,

:

he employed synthetic, correspondence of spaces, etc. not analytic methods, and inaugurated a purely synthetic method of studying these geometries. Veronese's Fondamenti di geometria contains an elementary synthetic treatment of the geometry of four dimensions and the geometry
of

n dimensions; and the Mehrdimensionale Geometric

* Philosophical Transactions,
\

of

160 Mathematical Papers, vol. VI, 1893, No 413. Philosophical Transactions, vol. 169, 1878; Mathematical Papers, No. 33, pp.
vol.
;

305-331t "ZurTheoriedes eindeutigen Entsprechens algebraischcr Gebildevon beliebig vielen Dimensionen," Mathematische Annalen, vol. 2, pp. 293-316.

Halphen, "Recherches de geometric a n dimensions," Bulletin de la Soci&U Mathematique de France, vol. 2, pp. 34-52; Jordan, "Essai sur la ge'ome'trie a n
dimensions," id. vol. 3, pp. 103-174. If "Behandlung der projectivischen Verhaltnisse der

Raume von verschiedenen Dimensionen durch das Princip des Projicirens und Schneidens," Mathematische
19, pp.

Annalen, vol.

161-234.

SLOW RECOGNITION

Schoute, employing a variety of methods, makes these * A bibliography with subjects very clear and interesting.
nearly six hundred
metriche. f
titles,
il

up

to 1907, is to be

found in

Loria's // passato ed

The

presente delle principali teorie geolatest bibliography is that of Sommerville, J

which contains 1832 references on n dimensions up to 1911: about one-third of these are Italian, one-third

German, and the

rest

mostly French, English, and Dutch.

We
sions

see that the geometries of more than three dimenwere slow in gaining recognition. The general

notion that geometry is concerned only with objective external space made the existence of any kind of geometry

seem

to

space. hesitated to use the higher geometry,^ although the workrcttilinee esposti in
;

depend upon the existence of the same kind of Consequently some of our leading mathematicians

* Veronese, Fondamenti di geomclria a piu dimensioni ed a piu spczie di unita forma elementare, Padua, 1891 German translation by Schepp, Grundziige der Geometric von mehreren Dimensionem, etc Leipzig, 1894. Schoute,
,

Mehrdimensionale Geometric, Sammlung Schubert, and XXXVI, Leipzig, 1902 and 1905. Another elementary treatment of the subject is by Jouffret,
Geometric a quatre dimensions, Paris, 1903

XXXV

t 3d
t

ed.,

Turin, 1907.

Bibliography of Non-Euclidean Geometry, Including the Theory of Parallels t the Foundations of Geometry, and Space of n Dimensions, University of St. Andrews, Scotland, 1911. There is now a considerable popular interest in the four-dimensional geometry,

many curious things about it, and because of attempts which have been made to explain certain mysterious phenomena by means of it. This interest has produced numerous articles and books written to describe the fourth dimension In 1908 a prize of #500 was offered through the in a non-mathematical way.
because of the

American for the best non-mathematical essay on the fourth dimension. hundred and forty-five essays were submitted in this competition. Some of these have been published in a book, whose Introduction, by the present writer,
Scientific

Two

gives quite a full discussion of the various questions connected with the subject

(The Fourth Dimension Simply Explained, Munn and Company, New York, 1910). II Thus Darboux, in a memoir presented in 1869 at the Academy of Sciences and
published in 1873, speaks of a lacune in geometry of space as compared with plane geometry, for certain plane curves can be studied with advantage as projections from space, but "Comme on n'a pas d'espace a quatre dimensions, les me"thodes de procourbes
jection ne s'ctendent pas a la ge'eme'trie dc 1'espace" (Sur itne classe remarquabk de et de surfaces algebriques, Paris, p. 164). Even in 1903, in his Report at

10

INTRODUCTION
little diffi-

ing out of its details presented comparatively

culty to them.

This objection has led some writers to

emphasize those applications of four-dimensional geometry that can be made in three-dimensional space, interpreting it as a geometry four-dimensional in some other element

than the point

just as

we have

interpretations of the

non-Euclidean geometries, which cannot, however, take

the place of their ordinary interpretation.* As long ago as 1846 it was pointed out by Pliicker that four variables
the Congress at St. Louis, he says, "Une seule objection pouvait fctre faite Tabsence de toute base reele, de tout substratum," etc. (Bulletin des sciences mathematiques, ser. 2, vol. 28, p. 261, Congress of Arts and Sciences edited by H. J. Rogers,
. .

Houghton, Mifflin and Co., Boston, vol. I, 1005, p. 557). But Darboux himself has made important contributions to the geometry of n dimensions see, for example, his Leqons sur les systemes orthogonaux, 2d ed., Paris, 1910; in particular, Bk. I, chap. 6,
:

and Bk.

II,

chap.

i.

Poincare", in speaking of the representation of

two complex variables

in space of

four dimensions, says,

sens."

"On

est expose* a rebuter la plupart des lecteurs et

de plus

on ne possede que 1'avantage d'un langage commode, mais incapable de parler aux
Acia Mathematica, vol. o, 1886-1887, p. 324. the other hand, we have the following from Sylvester : "There are many who regard the alleged notion of a generalized space as only a disguised form of algebraic

On

formulization

or of impossible lines, or lines

but the same might be said with equal truth of our notion of infinity, making a zero angle in geometry, the utility of dealing
will

with which no one

myself in

be found to dispute.

Dr. Salmon in his extension of Chasles*

theory of characteristics to surfaces, Mr. Clifford in a question of probability, and

theory of partitions, and also in my paper on barycentric projection, of the practical utility of handling space of four dimensions as if it were conceivable space" ("A Plea for the Mathematician,"

my

have

all felt

and given evidence

;

Nature, vol. i, 1869, p. 237 Mathematical Papers, vol. II, p. 716). A statement of Cayley's has been given in a previous footnote (p. 5). For other expressions of his views we may refer to the first paragraph of the "Memoir " on Abstract Geometry mentioned above, and to a statement quoted by Forsyth
in his "Biographical Notice," Cayley's Mathematical Papers, vol. VIII, 1895, p. xxxv.

As to the existence of a higher space, Gauss also is said to have considered it a possibility (W. Sartorius von Waltershausen, "Gauss zum Gedachtniss," Gauss
Werke, Gottingen, vol. VIII, 1000, p. 267). Segre, referring to the first of the two remarks that we have quoted from Darboux, says, "Maintenant nous faisons usage de 1'espace a quatre dimensions sans nous
pre*occuper de la question de son existence, que nous regardons

comme une

question

tout-a-fait secondaire, et personne ne pense qu'on vienne ainsi a perdre

de la rigeur."

Mathematische Annalen, vol. 24, 1884, p. 318. *See Emory McClintock, "On the Non-Euclidean Geometry," Bulletin of the New York Mathematical Society, vol. 2, 1892, pp. 21-33.

EXTENT AND VARIETY OF APPLICATIONS

II

can be regarded as the coordinates of a line in space.* Another four-dimensional geometry that has been suggested
is

that of spheres. f

But

this higher

geometry

is

now

recognized as an indis-

pensable part of mathematics, intimately related to

many

other branches, and with direct applications in mathematical physics. The most important application for
the application as analytic geometry to algebra and analysis : it furnishes concise terms and expressions, and by its concrete conceptions enables him
is

the mathematician

to grasp the
tricate

meanings of complicated formulae and inThis is true of all the geometries as well relations.

as the geometry of four dimensions. The latter is of special use in connection with two complex variables, both in the

study of one as a function of the other, and when

it is

desired to study functions of both considered as indeAnother very important applicapendent variables. J tion of geometry of four dimensions is that mentioned by

d'Alembert, making time the fourth dimension: within a few years this idea has been developed very fully, and

has been found to furnish the simplest statement of the

new

physical principle of relativity.

* System der Geometric des Raumes, Diisseldorf, p. 322. " t See article by Professor Keyser, A Sensuous Representation of Paths that Lead from the Inside to the Outside of a Sphere in Space of Four Dimensions/' Bulletin of the American Mathematical Society, vol. 18, ign, pp. I&-22. I See reference given on the preceding page to Poincare"'s memoir in the Acta
Mathematica; also Kwietnewski,
Ueber Flachen des vierdimensionalen Raumes,
deren sdmtliche Tangentialebenen untereinander gleichwinklig sind, und ihre Beziehung zu den ebenen Kurven, Zurich, 1902.

theory has been developed somewhat as follows : If time is represented by t measured on an axis perpendicular to the hyperplane of the spaceaxes, the /-axis itself or any parallel line will represent a stationary point, and uniform motion will be represented by lines oblique to the /-axis, forming an angle with

The

a coordinate

the /-axis which depends on the rate of the motion.

of light) is taken as the greatest possible velocity

A certain velocity
and the same

(the velocity

for all systems of

measurement.

The

lines

this velocity are the elements of a conical hypersurface.

through the origin, or through any point, representing All lines not parallel to

12

INTRODUCTION

With these various applications have been developed many methods of studying the higher geometries, besides the ordinary synthetic and analytic methods. We now have the synthetic and analytic projective geometries, we have including the projective theories of measurement
;

the theories of transformations and transformation groups the geometry of algebraic curves and algebraic functions; the geometry associated with the representation of two complex variables; differential geometry and the trans-

formation of differential expressions; analysis situs, enumerative geometry, kinematics, and descriptive geometry the extensive magnitudes of Grassmann and different kinds
;

of vector geometry; the application of quaternions to four dimensions; and the very recent application of fourdimensional vector analysis to the principles of relativity.*
these elements are divided into two classes
the lines of one class, less inclined to

represent possible motions, while the lines of the other class can represent only imaginary motions. The system may be regarded as a non- Euclidean geometry in which the conical hypersurface plays the part of absolute for angles, while dis-

the

/-axis,

tances along lines of the two classes are independent and cannot be compared. Now a point moving uniformly may be regarded as stationary, and the points which are

moving uniformly in the opposite space-direction. This change represented by a transformation of coordinates, the new /-axis being the In this theory the angles of planes line representing the given uniform motion.
really stationary as
is

of view

play an important part, and line and plane vectors are freely used. This application of four-dimensional geometry was developed by Minkowski.
further elaboration sec article

For

by E. B Wilson and G. N. Lewis, "The Space-time Manifold of Relativity. The Non-Euclidean Geometry of Mechanics and Electromagnetics," Proceedings of the American Academy of Arts and Sciences, vol. 48,
No. n, Nov., 1912.

*On the projective theory of measurement see d'Ovidio, Le funzione metriche fondamentali negli spazii di quantesivogliano dimension! e di curvatura costante,'
k<

Atti detta

Accademia de Lincei,

ser. 3, vol. i,

1876, pp. 133-103;

abstract in the

pp. 403-418. important series of memoirs by Poincarg Journal de I'Ecole Polytechnique, vol. 100, 1894; Rendiconti del Circolo Matematico di Palermo, vol. 13, 1899; Proceedings of the London Mathematical Society, vol. 32, 1000; Bulletin de la Societi Mathematique de France, vol. 30, 1002 Journal de mathematiques pure*
12, 1877,

Mathematische Annalen, vol.

On analysis situs there is an

et appliquees, ser. 5, vol. 8,

1902;

Rendiconti di Palermo, vol. 18, 1904; Comptes

Rendus,

vol. 133, 1901.

The enumerative geometry has been developed

chiefly

by Schubert.

has

ESSENTIAL PART OF GEOMETRY

All these interpretations

13

and methods that have been

applied to the study of the higher geometries, and all these uses to which they have been put, are interesting and valuable to a greater or less degree but the greatest advantage
;

to be derived

from the study of geometry

of

more than three

dimensions
geometry.

is

Our plane and

a real understanding of the great science of solid geometries are but the

beginnings of this science. The four-dimensional geometry is far more extensive than the three-dimensional, and all
the higher geometries are more extensive than the lower. The number and variety of figures increases more and more rapidly as we mount to higher and higher spaces, each space

extending in a direction not existing in the lower spaces, each space only one of an infinite number of such spaces
in the next higher.

study of the four-dimensional geometry, with its hyperplanes like our three-dimensional space, enables us
to prove theorems in geometry of three dimensions, just as a consideration of the latter enables us to prove theorems Such theorems may come from much in plane geometry.

of

simpler theorems relating to the four-dimensional figures which the given figures are sections or projections.*
articles in the

M athematische Annalen, vols. 26, 38, and 45,

'

1886, 1891

and 1894;

in

the Acta Mathematica, vol.

of a Rigid

In kinematics we may No Forces T System in an A -fold Homaloid," Proceedings of the London Mathematical Society, vol. 7, 1876, Mathematical Papers, No 26, pp. 236-240; Beltrami, "Formules fondamentales de cine"matique dans les espaces de courbure constante," Bulletin des
science mathematiques, vol.

1886; and elsewhere. mention: Clifford, "On the Free Motion under
8,

see also articles

vols. 20

and

22,

n, 1876, pp. 233-240, Opere, vol. Ill, 1911, pp. 23-29; by Craig and Hatzidakis in the American Journal of Mathematics 1898, and IQOO

Quaternions have been applied to geometry of four dimensions by Hathaway, Bulletin of the American Mathematical Society, vol. 4, 1897, pp. 54-57; Transactions of the American Mathematical Society, vol. 3, 1902, pp. 46-59; and by
Stringham, Transactions,
437-439vol. 2,

1901, pp.

183-214;

Bulletin, vol.

n,

1905, pp.

Other methods are illustrated in memoirs already referred to. * See Cayley's article in Crelle's Journal, vol. 31, and the articles of Veronese and athematische Annalen, to which we have already referred. Segrc in the

14

INTRODUCTION

Indeed, many theorems and processes are seen only partially or not at all in the lower geometries, their true nature and
four dimensions

extent appearing hi the higher spaces. Thus in space of is found the first illustration of figures

which have two independent angles, and of different kinds of parallelism and different kinds of perpendicularity. Another example is the general theorem of which a particular case is given in Art 31, namely, that a section of a simplex of n dimensions is one of the two parts into which i dimensions (that is, its interior) may be a simplex of n divided by a section.* There are also many properties in which spaces of an even number of dimensions differ from spaces of an odd number of dimensions, and these differences would hardly be recognized if we had only the ordinary geometries. Thus in spaces of an even number of dimensions rotation takes place around a point, a plane, or some other axis-space of an even number of dimensions,
while in spaces of an odd
of a rotation
is

always of

number of dimensions the axis an odd number of dimensions

(see chap. IV).

The study of these geometries gives us a truer view of the nature of geometrical reasoning, and enables us to break away from intuition. This is especially true if we
adopt the synthetic method. The analytic geometry may seem to be free from difficulty, and many feel a higher
degree of certainty in the results of their algebraic processes. But we are apt to attach the terms of geometry to our
algebraic forms without
their significance.

There

any attempt at a realization of is, indeed, an abstract geometry

which the terms are regarded as meaningless symbols; but the interest and usefulness of geometry depend on the clearness of our perception of the figures to which it may be
in

applied,

and so we prefer

to study

some concrete geometry,

vol. II,
i,

* See Schoute, Mehrdimensionale Geometric,

Nr.

6.

SYNTHETIC METHOD
some interpretation
could have obtained
to its terms.
of the abstract

15

geometry which we

And

by giving a particular interpretation then the abstract geometry and other

can all be obtained from the concrete is really the same absolute certainty There geometry.* to synthetic geometry if it is developed logically from the axioms, and in the synthetic study of four-dimensional geometry we are forced to give up intuition and rely entirely
interpretations

on our

logic, f

it is doubtful whether we can ever picture to ourselves the figures of hyperspace in the sense that we can picture to ourselves the figures of ordinary space, yet we can reason about them, and, knowing that the validity of

Although

our geometry depends only on the logical accuracy of our reasoning, we can proceed to build it up without waiting and then we may in time acquire for a realization of it
;

facility in handling the geometrical proofs of the theorems and in stating precisely the forms and properties of the figures that it is almost as if we could see them. For
* Some portions of our study are treated by themselves as new interpretations As soon as the fundamental propositions which of geometries already studied. correspond to the axioms of some such geometry have been established, so as to
justify this

such

mode of procedure, we have only to translate its theorems in accordance with these propositions to have in our possession a complete development of the Examples are, the three-dimensional Point Geomparticular subject considered.
etry (Art. 64), the two-dimensional Edge Geometry (Art. 78), systems of isocline planes at a point (Art. 112).

and the theory

of

We do

not seek to

in laying the foundations

know which of several geometries is the true geometry, and we do not seek for the true system of axioms, or even the
and
relations.

All geometries are equally true, and somebe built up equally well in several different ways. A complete treatise on geometry should consider not only the different geometries, but different methods of building up each geometry. An example of such a treat-

true system of elements

times a particular geometry

may

Elcmentargeometric, edited by Enriques by H. Thieme of Question* riguardanti la geometria See also the chapters on this subject by Enriques and clementare, Bologna, IQOO). others in the French and German Encyclopedias (Encyklopadie der math. Wiss., vol.
is

ment

the

first

volume

of

Fragen der

(Leipzig, IQI

i,

German

translation

IIIi, Leipzig,

different systems of

1907; Encyclopedic des sci. math., vol. IIIi, Leipzig, IQII). A list of fundamental elements and relations is given in a footnote at

the beginning of Coolidge's Non-Euclidean Geometry, Oxford, 1009.

16

INTRODUCTION

studying the geometry synthetically our attention is fixed upon the figures themselves, and this takes us directly
in

to the heart of the difficulty

and keeps

it

before us until

we have mastered
tion*

it.

Thus

in its results this

greatly increases our power of intuition

geometry and our imagina-

following pages have been written with the object of meeting as far as possible the difficulties of the subject.

The

No

knowledge of higher mathematics

is

necessary;

yet

we do not

believe that the simplest way is to avoid a mathematical treatment. The confidence gained from a study
of the proofs,
if

do more

for the student

they can be made clear and precise, will than a mere description of four-

dimensional space. will indicate how this purpose has influenced us in our choice of subject-matter and the form
of presentation.
.

We

have adopted the synthetic method and made no use of analytic proofs, feeling, as we have already explained, that this study of the figures themselves will serve best to
help us understand them. We have confined ourselves to the fourth dimension, although it would have been easy to cover a much wider
field,

We

We

hope that in

this

way

the four-dimensional

space will be made to appear as a concrete matter to be studied by itself, and not as one of an indefinite series of
spaces, each understood only in a vague general way. have wished to give to these pages a familiar appear-

We

and so have endeavored to follow the popular textbooks and build up a structure that will rest on the foundations laid in the schools. Our geometry might have been to the axioms of some modern investigation, or adapted
ance,
See C. J. Keyser, "Mathematical Emancipations/' Monist, PP' 65-83, particularly pp. 81-82. t See, for example, the Mehrdimensionalc Geometric of Schoute.

vol.

16,

1006,

FOUNDATIONS

17

have attempted to establish a system of axioms, but either course would have raised questions quite different from those of four-dimensional geometry. The methods employed in this book are methods which the student has used freely in the past, even though he may be ignorant of their
true significance

and

justification

there

in their application here,

and

their

nothing new employment without

is

question leaves him free to fix his attention upon the difficulties inherent in the subject.

There is, however, one part of the foundations which has been presented with considerable care, namely, that which relates to the definitions and the intersections of certain
elementary
figures.

It is here that the four-dimensional

geometry
for

begins to contradict our experience, declaring,

example, that two complete planes may have only a point in common, and that a line can pass through a point of the interior of a solid without passing through any other
of its points. It is true that these facts and many others not easy to realize are easily proved, and require only a few of the

theorems given in this connection.

On

the other

hand, the theorems for which most of these details are needed are so "evident'' that they are usually ignored alNow a statement of these theorems, with a together.
of what is to be proved, of out the logical working proofs themselves, will give the student more confidence in all the results of his Similar considerations have led us in the fourth study.
realization of
is

what

assumed and

and a

of

chapter to take up symmetry, order, and motion in space two dimensions and in space of three dimensions.

Great assistance comes from the analogies that exist in geometry, and so we have gone back in some cases and given proofs which are not well known, and to which more difficult proofs that follow are analogous;* and we have
* See, for example, the theorems of Arts. 61 and 62.

18

INTRODUCTION
comparison of chapters and sections one another by adopting in them the same

tried to facilitate the

analogous to

arrangement of paragraphs and the same phraseology. Not much use can be made of diagrams, and so far as
they are given they must be regarded as indicating the relations of different parts of a figure rather than as showing
in

any way its appearance. A figure can be accurately determined by its projections, and the descriptive geometry of four dimensions will be helpful to those who are familar

with the methods of descriptive geometry.*

also be learned

Much

can

by studying the sections of a figure. section of a four-dimensional figure is that part which

A
lies

in a three-dimensional space or hyperplane, and is, therecan suppose that fore, like the figures of our space.

We

are able to place ourselves in any hyperplane, and so to examine any hyperplane section: in connection with the diagrams we shall sometimes call attention to those parts which lie in any one section, speaking of them as "what we can see in a hyperplane." One way of studying a
figure is to let it pass across our space, giving us a con-

we

tinuously varying section, as

if

time were the fourth diit

turn, or our section of it, so that the direction of our view changes. It is along these lines, if at all, that we are to acquire a perception of

mension.

Another way

is

to let

hyperspace and

its figures.

Some explanation should be made in regard to the arrangefar as they are considered,

ment, the particular form chosen for the foundations so and the fundamental conceptions

as

we have

presented them.

We have given only the Euclidean Geometry, except that

the geometry of the hypersphere, and of the hyperplane at infinity, and the geometry in a few less important cases,
* See Schoute, Mehrdimensionale Geometric,
vol. I,
5.

COLLINEAR RELATION
are themselves non-Euclidean.

19

It has been found, however, that several chapters can be completed before we make

any hypothesis

much
and

in regard to parallels, and that, too, without variation from the usual treatment. Perpendiculars

kinds of angles, symmetry and order, and those hypersurfaces (the hyperpyramid, the hypercone, and the in fact, hypersphere) which do not involve parallels
all

all

of "restricted"

geometry

can be taken up before the

In the chapter on the hyper-

introduction of parallels.*

sphere, its geometry, being elliptic, is stated as such, and a group of theorems is given from the non-Euclidean geom-

etry;

and

in the last

erties of the

chapter the non-Euclidean prophypersphere are used quite freely. Although

these portions of the

will find it

book may be omitted, the student

an advantage to make himself familiar with the Hyperbolic and Elliptic geometries, f We have started with points only as elements, regarding
figures as classes of points,

all

and so defining a

figure

simply by stating what points constitute the class. To do this we assume first a relation by which with any two points certain points are said to be collinear. Then for line we take two points and the class of points collinear with them,
points collinear with any two that we have, and thus continue, at each step adding to our class of points all that are collinear with any two already
to the

add

group

all

now

in the group, so that the line includes every point which it is possible to get in this way. Thus any two points determine a class consisting of the points which are collinear

with them, and any two points determine a class of the kind which we call a line.J By the axioms of Art. 3 the
* See the author's Non-Euclidean Geometry, Ginn and Co
in particular, p. 6.
,

Boston, 1001, chap. I

that

t The Hyperbolic and Elliptic geometries are the only non-Euclidean geometries we have referred to at all. t That two points determine a line does not mean, as in some of our text-books,
"\

20

INTRODUCTION
classes are identical
:

two

the line consists only of the points

two given points, and there are no additional points to be obtained by taking any two of these In fact, any two points of a line determine the points. same class of points as collinear with them, and the same
collinear with the
line.

But

until

we have adopted

these axioms

we must
:

suppose that the line might be a much more extended class that, if we have the points collinear with two given points,
be quite different;
every point of the

the class of points collinear with any two of these might and that, while a line must contain
line

determined by any two of

its points,

Thus we make
and

the latter might not contain every point of the former. a distinction at the beginning between the
line,

notion of collinear points and the notion of points of a

this distinction

analogous to plane and of than three dimensions. to more and spaces hyperplane, But after we have adopted our first two axioms we are
line

makes

employ the word collinear in its commonly accepted thus to avoid the introduction of a new term for and sense, two relations. of these one A careful distinction has been made between the points of a closed figure and the points of its interior. Thus a triangle is made to consist of three vertices and the points of its sides, a tetrahedron of its vertices and the points of This is only carrying to its edges and faces, and so on.
able to

the limit the tendency to regard a circle as a curve rather than as the portion of the plane enclosed by the curve, and

a sphere as a surface.

The
is

figure

of

one-dimensional

geometry corresponding to the triangle and tetrahedron,

the one-dimensional simplex,
the segment.
of

Therefore,
points,

we
let

have defined segment as consisting

two

and

that the line contains the two points, or that no other line contains them. A figure may be determined in various ways. Thus a line in the ordinary plane geometry may be determined by two points as the locus of points equidistant from them.

FIGURE AND ITS INTERIOR

21

the points between them constitute the interior of the segment.* On the other hand, a side of a polygon or an edge
of a

polyhedron is the interior of a segment, consisting of the points between two vertices and not including the vertices

a face consists of the interior of a triangle or polygon; a half -line is defined so as not to include its and extremity, a half-plane so as not to include its edge
themselves
;

so on.

We speak, indeed, of the sides and vertices of a face,

and
of the area of a triangle,

of the length of a segment,

we have used the terms of ordinary without definition, and employed freely all the geometry words and phrases of its everyday language. But the
just as, in general,

between the points of a closed figure and the points of its interior is of great importance, and has been
distinction

carefully observed.
* Hilbert defines segment (Strecke) as a "system of two points," but he speaks between A and B as "points of the segment AB," although he also

of the points

speaks of them as points "situated within the segment" (Grundlagen der Geometric, Leipzig, 1899, p 6, 4th ed 10,13, p 5)
,

In the Encyclopedic des sciences mathematiques, vol. IIIi, p. 23, Enriques defines segment upon a line as "having its extreme points at two given points A and B of
the line and containing the intermediate points."
di Keometria of Enriques

More

definitely, in the Elementi

and Amaldi (Bologna, 1911), half-line is defined so as to include its extremity, and then the segment AB is the part common to the halflines A B and BA (p 3) E. 11. Moore defines the segment AB as consisting of points "distinct from A and

B"

etc.

that

is,

A and B

are not included

among

the points of his segment

("On

the Project ive Axiom-* of Geometry," Transactions of the American Mathematical

See also Veblen, "A System of Axioms 3, 1002, p ijy, Axiom 2], Geometry," Transactions, vol 5, 1004, p. 354, Definition i, and "The Foundations of Geometry," Monographs on Modern Mathematics, edited by J. W. A. Young, New York, 1911, p. 5. Most writers who use the word segment in this connection regard a segment as an entity, a piece of a line, without considering whether the end-points are included
Society, vol.

for

or not.

Many writers speak of the segment as the "measure of the distance" between the two points (see Schotten, Inkalt und Methode des Planimetrischen
Unterrichts, Leipzig, vol. II, 1893, chap,
i,

2).

" Veblen, in the Foundations of Geometry" just referred to, defines triangle and tetrahedron in the same way that we have defined them (pp. 29 and 45).

22
*

INTRODUCTION

remarkable memoir on geometry of n dimensions

is

Theorie der

vieljachen Kontinuitdt, by L. Schlafli, edited by J. H. Graph, Bern, This was written in the years 1850-1852, but the author did 1911.

not succeed

in getting* it published, apparently on account of its length, and it remained among his papers for fifty years, until after his death (see Vorbemerkung).

other things he works out the theory of perpendicularity kinds of angles, giving, in particular, a generalization of the theorems which we have given in Arts. 66 and 67 (15). He proves the polyhedroid formula and the corresponding formula for any number

Among
all

and

of dimensions, and he constructs the six regular convex polyhedroids and the three regular figures which exist in each of the higher spaces,

proving that these are the only regular figures of this kind ( 17). He makes an extensive study of the hypervolume of a spherical simplex,

of

showing the difference between the cases of an even number and an odd number of dimensions, and giving the formula for a pentahedroid to which we have referred at the end of Art. 165 ( 22). In the third part of the memoir he takes up quadratic hypersurfaces,
the classification of these hypersurfaces, confocal hypersurfaces, etc. The methods are analytical but the language and conceptions are
,

purely geometrical.
* This note was written after the rest of the Introduction was in type.

CHAPTER

THE FOUNDATIONS OF FOUR-DIMENSIONAL GEOMETRY

I.

POINTS

AND LINES

Points. Figures regarded as classes of points. of elements The geometry are points. We do not define It is impossible to build up a system of the term point.
1.

geometry without undefined terms, and if we can give different meanings to this word we shall be able hereafter to give to our geometry different interpretations (see, for example, Arts. 64 and 78 and Introduction, p. 15). The objects which we study are to be regarded as conas classes of points selected according to various laws from the class which includes all points. Any selected class is said to constitute a figure, although
is,

sisting of points, that

the word figure is also used to denote a drawing or picture. points of a figure may also be said to lie in the figure or belong to it. One figure is said to lie in another when

The

all

of its points are points of the second.

It will often

be

convenient to speak of a figure as consisting of certain other figures, but this expression should always be understood
as meaning that
figures.
it

consists of the points of these other

Two

figures intersect

when they have a point

or points in

common, and

their intersection consists of such

common
... R,

point or points.
2.
. .
.

The

collinear relation.

Geometries of
23

x, a, 3,

dimensions.

Points have an undefined relation de-

24

POINTS AND LINKS

[i. I.

Given any two points, there noted by the term collinear. of the is a class consisting points collinear with them.
Geometries of
i,-2, 3,
. . .

n,

dimensions are geom2,

etries of figures lying in spaces 0/ i,

3,

j,

dimensions; that is, spaces of more than three dimensions.

in a line, plane, hyperplane,

and in

consists of the points that we get if we take two distinct points, all points collinear with them, and all points
line

collinear

with any two obtained by this process.

plane consists of the points that we get if we take three points not points of one line, all points collinear with any two of them, and all points collinear with any two obtained

by

this process.

A hyperplane consists of the points that we get if we take four points not points of one plane, all points collinear with
any two
of them,

and

all

points collinear with

any two

obtained by this process. A space of four dimensions consists of the points that we get if we take five points not points of one hyperplane,
all

points collinear with any two of them, and

all

points

with any two obtained by this process. Continuing in this way, we can define a space of n + 1 dimensions after we have defined a space of n dimensions.
collinear

All spaces

have this property: that the points collinear with any two points of a space belong to the space. Two distinct points are said to be independent. In
if it is

general, a point
class

is independent of the points of a given included not among the points that we can

get by taking these points, points collinear with any two of them, and points collinear with any two obtained by
this process.

The

points of a given set are absolutely in-

and

* In the ordinary interpretation of these terms, line alone is used for straight line, the complete line is meant, not that part of a line which we shall speak of as the interior of a segment (Art. 5).

2,3l

AXIOMS OF COLLTNEARTTY

25

dependent if it is impossible to obtain them all in this way from a smaller number of points.* The different kinds of

space are distinguished by the number of absolutely independent points that they can contain.
3.

Segments.

Two axioms
consists

concerning
of

the

collinear
points.

relation.

segment f

two

distinct

Any two
AB.

points are the points of a segment. The segment consisting of the points A and B will be called the segment

point is said to be collinear with a segment collinear with the two points of the segment.

when

it is

Concerning the collinear relation we shall

now make
segment

two axioms

AXIOM

The

class of points collinear with the

AB includes

the two points of this segment.

AXIOM 2.J If a point O, not the point B, is collinear with the segment AB, then any point P collinear with the segwill be collinear with the segment OB. ment

AB

and

a segment OB), then all points collinear with the segment AB will be collinIn particular, A itself will be ear with the segment OB.
(so that there is

In other words, if is not the point

is

collinear with the

segment

AB

collinear with the

all

points collinear with the

segment OB, and therefore, conversely, segment OB will be collinear

of the rest

* If the points of a set are absolutely independent, each point will be independent and in this geometry we shall find that the points of a set are absolutely
;

independent
etc.).

if

each point

is

independent of the rest (Art.

(Elements of Geometry,

4,

Th.

Art. 10, Th. 2

We may,

therefore, speak of such points simply as independent points.

t Halsted
also uses the

calls this

sect

word

straight for line, taking these

New York, 1885, p. Q). He terms from the German Strecke

and Gerade

(see Rational Geometry, New York, 1907, pp. i and 6). the Elliptic Non-Euclidean Geometry this is assumed with certain restrictions. Thus on the sphere, whose geometry is the elliptic two-dimensional

tin

geometry, collinear meaning "on a great circle with," the statement given here as Axiom 2 is not true when A and B are opposite points

26

POINTS AND LINES

[i. i.

with the segment AB. The class of points collinear with one segment is the same as the class of points collinear
with the other segment. ably in connection with

We can use A or O interchangeB as one of the two points with

is

respect to which the class

tinction

selected.

In this second axiom we do not intend to make any disbetween the points A and B. Except when O is

one of these points, ments OA and OB

collinear with the
4.

P will be
;

collinear with

and when

is

the point

both of the seg-B, it will be

segment

OA

Lines.

Only one

line

Given two
obtained

distinct points

and 5, the

contains two given points. line is the line

AB

when we

start with these points

and carry out

the process described in the definition of line (Art. 2). If A' and B' are two distinct points of the line AB, then
points of the line A'B' will be points of the line AB; for the process of obtaining the second line is but a continuation of the process of obtaining the first. shall now
all

We

are points of the line prove that all points of the line in A'B' other words, that the two are the same line, and
;

AB

that two points can both be points of only one line.

THEOREM

i.

with the segment

Any AB.

point

of the line

AB

is collinear

PROOF. We shall prove this theorem We know that it is true of the points A and B

by induction.
(Art. 3, Ax. i).

Let

By

be any other point of the line. hypothesis, O, being a point of the

line
line.

two points and N of the that the theorem is true of and N. from one, at least, of the points A and
ear with

AB,

is collin-

M must be distinct
J5,

We

assume

distinct. Let us suppose that for A and say that substitute

since they are

is

not B.

Then we can

the

segment

AB

points collinear with are collinear with the segment MB.

all

3-sl

ORDER ON A LINE

27

one of these points and is not the point N for B, and say that all of these is points are collinear with the segment MN. Now collinear with the segment MN. Therefore, reversing
Again, since
is

M, we can

substitute

this process of substitution,

we have O

collinear with the

segment

MB,

and

finally

with the segment

f

AB.

line

THEOREM 2. If A' and B' are two distinct points of the AB, then the line A'B is the same as the line AB. PROOF. We may suppose that A' is not B. Then,
A'
is

since

collinear with the

ear with the segment same as the line AB.

line

segment AB,A will be collinand the line A'B will be the A'B, In the same way we prove that the

A'B'

is

the

same

line.

from these two theorems that the class of points collinear with the segment AB constitute the line A B, and that collinear with means lying on a line with. It follows also that the two points A and B do not play any We can particular part among the points of the line.
It follows

speak of a set of collinear points, or of points collinear with one another, without specifying any two particular points
can also speak of as special points of the class. or more points as collinear with one given point.
5.

We

two

Order

of

points

on a

line.

Half-lines.

Another

relation in geometry, a relation of the points of a line, is that of order* This may be explained somewhat in detail

as follows
If

and B are two distinct points, then A comes before B lies beyond A in one direction along the line AB, while B comes before A and A lies beyond B in the opposite If A comes before B and B before C in a given direction. direction along a line, then A comes before C in the same

and

* Veblen uses the word order to mean order on a

or "Foundations of Geometry," p. 5

"
line,

System of Axioms,"

p.

344 ;

(full references are given above, p. 21).

28

POINTS

AND LINES
between

[l

T.

direction,

and

is

said to

lie

three points of a line, one of

and C. Given any them lies between the other

two.

This relation of order belongs to other classes of points beFor example, it belongs to the class of sides those of a line.
points constituting

what

is

commonly

called

a broken

line.

points of a line have also relations of density and continuity, but it will not be necessary to explain these

The

terms here, nor to give the axioms and theorems by which

these relations and the relations of order are established.*

The

interior of

between the two points

a segment consists of the points which of the segment. a line which

it.

lie

half-line or ray is that portion of

line

lies

in

one direction along the

is

from a given point of

The

called the extremity of the half-line, but this given point not is itself a point of the half-line. The half-line point

which has the extremity A and contains B, the half -line which we can describe as drawn from A through 5, is called and that portion of this half-line which the half-line AB
;

lies

beyond
y

is itself

a half-line, called the continuation

of

AB

or

AB

a given point of a

produced. The two half-lines into which line divides the rest of the line are called

opposite half-lines.
6.

Cyclical order.

There

is

another kind of order called

the points of a class are in cyclical of these points A and B (unless they are conorder, two secutive points) divide the rest of them into two sub-classes,
cyclical order.

When

those of one sub-class lying from A towards B in one direction and those of the other sub-class lying from A towards

in the opposite direction.

A and B

are said to

lie

* See Introduction, p. 16. For a treatment of these subjects we will refer to " " Veblen's "System of Axioms or Foundations of Geometry," and to R. L. Moore, "Sets of Metrical Hypotheses," Transactions of the American Mathematical Society,
vol. g, 1008, pp.

487-512

S-7l

LINE AND TRIANGLE

29

between the points of the

two

sub-classes

and

them.

If

is

a point of one sub-class and

to separate a point of

the other sub-class, that is, if we have these points in the DA, then we can say that the segments cyclical order

ACS

AB

and

CD

separate each other.

If

we

think of a class

of points in cyclical order as cut at some point, they will then have the relations of order described in the preshall have many illustrations of cycliceding article.

We

cal order (see Arts. 7

and
II.

14) .*t

TRIANGLES
of Pasch.

7.

Triangles.

Axiom

Intersection of a line

and a

triangle. triangle consists of three non-collinear of the three segments whose points the interiors points and are these points taken two at a time.

three points are the vertices, segments are the sides.

The

and the

interiors of the

Any
angle.

three non-collinear points are the vertices of a

tri-

In particular, two vertices of a given triangle and a point in a side which does not lie between them are themso also are a point in each selves the vertices of a triangle of two sides and the vertex where these two sides meet.
;

The

points of a triangle are in cyclical order in the

6).

tri-

angle (Art.

For a complete treatment of the intersections of and triangles the following axiom is required f
:

lines

* The points of a

circle are in cyclical order.

Indeed, the points of a line are in

cyclical order in Projective

We do not wish to

Geometry and in the Elliptic Non-Euclidean Geometry. exclude the Elliptic Geometry by assuming that the points of the
In

entire line are not in cyclical order.

many

cases of cyclical order there will be

an "opposite" to every element, and on a line of this kind we can determine the two directions at any point by regarding the line as cut at the opposite point (see, for example, the definition of "between" and "side produced" in Art. 122) or we
;

can suppose that we are considering only a "restricted" portion of the line; that one of the sub-classe* determined by two points suitably chosen in any given case. t See Veblen, "System of Axioms," p. 351.

is

30

TRIANGLES

[i.

n.

AXIOM.

line intersecting one side of a triangle

intersects the third side.

and

another side produced This will be referred to as the

it is

Axiom of Pasch.

As stated,

little

broader than

is

necessary.
intersect

THEOREM.
and two
of a triangle.

No

line

can

one side of a triangle

sides produced , nor can a line intersect all three sides

PROOF.

The

first

Axiom
angle

of Pasch.
line

To

statement follows directly from the prove the second statement, let us
tri-

suppose that a

could intersect the three sides of the

5'coming between A' and C' on the line. In the triangle A BC' we should then have the line AC intersecting one side, A'C'> But this is conat 5', and the other two sides produced.
',
f

ABC, BC

at A',

AC

at

and

AB

at C",

trary to the
line

first

statement of the theorem.

all

Therefore a

cannot intersect

three sides of a triangle.

COROLLARY.

No
it

line

of a triangle unless
8.

can contain more than two points contains one of the sides of the triangle.

Interior of a triangle.

The

interior of a triangle con-

segments whose points are points of the triangle, except of those segments which are collinear with two vertices of the triangle, that is, whose interiors The interior of a triangle does not also lie in the triangle.
sists of

the interiors of

all

include the triangle itself hence, whatever is in the triangle cannot be in the interior of the triangle.
;

THEOREM i. If two segments lying in a triangle separate each other in the cyclical order of the points of the triangle (Art. 6), then their interiors intersect, unless the interior of
one of these segments lies in the triangle. PROOF. When each of the two given segments has a vertex for one of its points, each segment with the third
vertex determines a triangle to which

we can apply

the

7,8]

INTERIOR OF A TRIANGLE
of

Axiom

Pasch and thus prove that the interior of

intersected

this

segment

is

by

the line containing the other

segment. From this it follows that the interiors of the

two segments have a point in

common.
of the two given has a vertex for one segments

When

one

of its points

and the other

a triangle

does not,

we form

by taking one point of the latter and the two vertices of

the given triangle which are
collinear with its other point.

This
the

triangle

contains

also

segment, or a segment whose interior is a part of the interior of the first segment.* We can prove the theorem, then, by proving it
first

for this

When
tex

segment and the second given segment. neither of the two given segments contains a verin the

we proceed
2.

same way, reducing

this case first to

the preceding.

A half-line drawn from any point P of a a triangle through point O of the interior intersects the triangle in a point of PO produced.
THEOREM
In proving this theorem we follow the methods of the preceding proof, taking first the case where O is between one vertex and a point of the opposite side and P is at
another vertex.

COROLLARY.
a point

// one of two opposite half-lines

drawn from

of the interior of

a triangle intersects the triangle,

the other does also.

* According to theorems of order referred to in Art. see Veblen, "System of 5 Axioms," p. 357, Lemma 6, or "Foundations of Geometry," p. n, Cor. 4.
;

TRIANGLES
9.

[i.

n.

The

relation, collinear with a triangle.

point

is

said to be collinear with a triangle

when

it is

collinear with

any two points

This
of all
is true,

of the triangle.

points of the sides produced,

in particular, of all points of the triangle, and of all points of the

interior.

THEOREM

i.

If a point
is

is collinear

with a triangle

A BC,
least

and

if P

any

point of this triangle not

a vertex and

not the point O, then the line

PO

will intersect the triangle at

in a second point Q.

PROOF.
Otherwise,

The line PO may contain one side of

if

the triangle.

a point of the triangle, it will itself be the second point Q, and if O is a point of a side produced, the theorem is the same as the Axiom of Pasch. If O is a point
is

of the interior, the

theorem

is

the same as Th.

2 of

the

preceding

article.

There remains, therefore, only the case where O is not a point of any one of the lines AB, BC, or AC, nor a point of and the interior of the triangle. Let then be the two a point of the side A B, points with which O is collinear, a point of the side AC, or the vertex or the vertex JB, and and K, not both vertices, however. C,

H H

POINTS AND TRIANGLE

33

We may suppose
Then

also that

O is

a point of

H K produced.

OAB or on which AC, passes through K, will intersect the interior of the segment BO in a point G. Now O is a point of BG produced. We have the triangles A BG and CBG, and in one or both of these triangles the line PO, intersecting BG produced and one other side,
will lie in the interior of the triangle

the side

OB, and the

half-line

will intersect the

third side. Therefore, in

all

cases this

line will intersect the triangle

A BC

in a

second point Q.

THEOREM
is collinear

2. If a point O, not a point of the line with the triangle ABC, then any point P collinear

with

the

triangle

ABC
O

will

be collinear

with the triangle

First let

OBC.
PROOF.
point
of

be a
If

the line

AB.

O
is

is

at A, the triangle the same as the triangle

OBC

line

ABC. If P is a point of the AB, it is collinear with O

and B.
that
line,

is

Let us suppose then not a point of this

is a point of a AB, point of BA produced, or a point of AB

and that

the side

produced.

a point of the AB, then a line through P and a point of

If
is

side

OB
will
or,

will intersect

BC
AC,

or

pass

through

C,
will

intersecting

intersect

OC

(Axiom

of

Pasch).

34
If

TRIANGLES

[i.

H.

is

a point of
of

BA

produced, then a line through

and a point
If

AB

will intersect

BC or will pass
OC.

through C,

or, intersecting

AC,

will intersect

is

a point of
of

and a point

AB produced, then a line through P AB will intersect AC or will pass through C,

BC,
will intersect

or, intersecting

OC.

will

be collinear

OAC, and therefore, by the first case, with the triangle OBC. In the same way we prove the theorem when O is a point
with the triangle
of the line
If

AC.

AC,

not a point of any one of the lines then a line through O and a point
is

AB, BC,

or

of the side

BC

will

will intersect

pass through A or one of the sides

AB or AC.

Let the point of

intersection be E,

and

let

us

suppose that it is the point A or a point of AC. Then

triangle

ABC,

P, being collinear with the will be collinear with the triangle A DC

y

then with the triangle EDC, with the triangle finally with the triangle OBC.

ODC, and

and

In other words, if is collinear with the triangle ABC is not a point of the line BC (so that there is a triangle
all

OBC}, then
self will

points collinear with the triangle be collinear with the triangle OBC. In particular,

ABC
A

will
it-

be collinear with the triangle OBC, and therefore, conversely, all points collinear with the triangle OBC will
collinear with

be collinear with the triangle ABC. The class of points one triangle is the same as the class of points

collinear with the other triangle.

We

can use

terchangeably in connection with the segment

the three points with respect to which the class

A or O inBC as one of
is

selected.

9, io]

DETERMINATION OF A PLANE

35

In this theorem we do not intend to make any distinction between the point A when O is collinear with

be collinear with

all

and the points B and C. Except two of these three points, P will three of the triangles OBC, OAC, and
III.

OAB.
PLANES
10.

Planes.

Only one plane contains three given non-

As already stated, a plane consists of the points that we get if we take three points not points of one line, all points collinear with any two of them, and all points collinear with any two obtained by this process.
collinear points.

ABC

Given three non-collinear points, A, B, and C, the plane is the plane obtained when we take these points and

carry out the process described in the definition. Now we can take for two of the three given points any r two points of their line. That is, if B' and C are points
of the line

EC, the plane

2).

AE C
f

ABC (Art. 4, Th.

we take
If

We

the same as the plane can 'think of a line as one of the

is

things with which we start in the process of obtaining a plane. Given the points of a line and a point not a point of the line,
all

points collinear with

points collinear with any two of these and any two obtained by this process.

all

A', B', and plane ABC, then

all

are three non-collinear points of the points of the plane A'B'C' will be
;

points of the plane

ABC

for the process of obtaining the

second plane
ing the
first.

is

but a continuation of the process of obtain-

shall now prove that all points of the are plane points of the plane A'B'C'-, in other that the are the same plane, and that three two words,

We

ABC

non-collinear points can all be points of only one plane.

THEOREM

i.

Any

point of the plane

ABC

is collinear

with the triangle

ABC*
same as
co planar

* Collinear with a triangle is therefore the in the usual sense of the word coplanar.

with the

triangle,

36
PROOF.

PLANES

[i.

m.

We

prove

this

theorem
that
it is

Th.
let

of Art. 4.

We, know

just as we proved true of all points of

the triangle and of all points of the sides produced. We be any other point, and prove by induction that the

theorem is true for O. This point is, by hypothesis, collinear with two points and N of the plane ABC. We assume that the theorem is not a and N. We can suppose also that is true of point of the line BC and that N is not a point of the line MB. It follows (Art. 9, Th. 2) that we can substitute for A, and say that all points collinear with the triangle

A BC are collinear with

is

the triangle

MBC.

Again, since

one of these points and is not a point of the line B, we for C, and say that all of these points are can substitute Now O is collinear with collinear with the triangle MNB.

the segment
fore,

MN and
this

so with the triangle

MNB.

There-

reversing

process of

substitution,

we have

collinear with

the triangle

MBC,
and

and

finally

with the

triangle

ABC.
2.

THEOREM
as the plane

// A',

',

are three non-collinear

points of the plane

ABC,

then the plane

A'B'C*

is the

same

ABC.

PROOF.

We

will

Th.

2 of Art. 4.

We

prove this theorem just as we proved ' can suppose that A is not a point

BC, and that B' is not a point of the line Then, since A' is collinear with the triangle ABC, A will be collinear with the triangle A'BC (Art. 9, Th. 2), and the plane A'BC will be the same as the plane ABC.
of the line

A'C.

In the same

way we prove
is

finally the plane A'B'C',

11.

that the plane A'B'C, and the same plane.

triangle in

Intersection of a line

and a

a plane.

THEOREM.

In

the

plane

ABC

any

line intersecting a

side of the triangle

ABC

will intersect this triangle at least

THE TWO S1DKS OF A LINE

in a second point,

37

and any half-line drawn from a point

of the interior of the triangle will intersect the triangle.

PROOF.
Th.
i,

The first statement follows directly from Art. 9, since the line contains points collinear with the

the second statement, let the half-line AO intersect the side BC in a point /), a point of produced (Art. 8, Th. 2). Since the given plane can also be
triangle.

To prove

AO

or as the plane ACD, it regarded either as the plane follows that any line of this plane through O other than the that line A will intersect both of these triangles again

ABD

or some is, that any line through O, whether the line other line of the plane, will intersect the given triangle; and, therefore, that any half-line drawn in the plane from 0, as well as
its
8,

AD

opposite half-line, will intersect the

tri-

angle (see Art.

12.
line.

Th.

2,

Cor.).

The two
i.

parts of a plane lying on opposite sides of a

THEOREM
*

Any

line of a plane divides the rest of the

into two parts, so that the interior of a segment lying plane one point in each part intersects the line, and the interior of a segment lying both points in the same part does not intersect the line.

PROOF, f Let a be the given line, and A a point of the plane which is not a point of a. We divide the points of
the plane which are not points of a into two classes, putting A into the first class, putting a point B into the first class
if

the interior of the segment A B does not contain a point of a, and putting a point B' into the second class if the
interior of the
*

segment

AB

does contain a point of

it

a.

Or at least any restricted portion of the plane through which same statement applies to the first theorem of Art. 23, and to Art.
t This proof
for

passes.

The

28.

is given in Halsted's Rational Geometry, p. 8. The point A used purposes of proof does not play any particular part in the actual separation of the points into two classes, any more than do any two points play a particular part in the class of points which we call a line.

PLANES

[i.

in.

Any

line of the

plane through

will

contain only points

of the first class, or will contain a point of a

by which

the rest of

its

points are separated into two classes, so

that the interior of a segment with one point in each class itself contains a point of a, and the interior of a

segment with both points in the same

tain a point of a.* Now in a triangle
f

class

does not con-

a point of the first class, and B a point of the second class, the line a, intersecting the side AB' and not AB, must intersect BB' (Art. n).

ABB', B being

In a triangle ABC, B and C being points of the first class, the line a does not intersect either side coming to A
,

and therefore

it

cannot intersect
,

BC

(same

reference).

points of the second class, the line a intersects both of the sides r f which meet at A, and therefore it cannot intersect B C
(Art. 7).

Finally, in a triangle

AB'C' B' and C' being

The two
and the

parts into which any line of a plane divides the

rest of the plane are said to lie

line is said to lie between

* This is

on opposite sides of them.

5.

the line,

one of the theorems of order referred to in Art.

I2,i3l

HALF-PLANES AND TRIANGLE

2.

39

THEOREM
of each,

// two lines

intersect, Hie opposite half-lines

drawn from

their point of intersection, lie in their

plane on opposite sides of the other.

13.

Half-planes.

The three

half-planes of a triangle.

which lies on one side of a line of it is called a half-plane, and the line is the edge of the halfThe half-plane which has the line AB for its edge plane. and contains the point C will be called the half-plane A B-C. The two half-planes into which any line of a plane divides
of a plane

That portion

the rest of the plane are called opposite half-planes.

THEOREM i. Every point of the plane belongs to one, at least, of the three half-planes BC-A, AC-B, or AB-C. PROOF. All points of the plane which do not belong to

ABC

BC-A belong to the line BC or to the halfplane opposite to BC-A All points of the line BC belong to the half -line CB or to
the half-plane
.

the half-line BC, or to both, and so to one, at least, of the half-planes AC-B or AB-C.

The

be a point of the half-plane opposite to BC-A. segment AP has a point Q in common with the line BC, a point of one or both of the half-planes

Let

interior of the

and AB-C. Now the interior of the segment PQ cannot have points in common with both of the lines AC and A B for PQ produced contains their common point A. Therefore P belongs, with Q, to one or both of these two
;

AC-B

half-planes.

That

is,

every point of the plane which

the half-plane
half-planes

BC-A belongs AC-B and AB-C.

2.

is not a point of to one, at least, of the two

THEOREM

Any

point

of the interior of the triangle

ABC

belongs to all three of the half-planes

BC-A, AC-B,

and AB-C.

40

CONVEX POLYGONS
3.

[i.

w.

THEOREM
of the

Conversely, if a point

belongs to all three

it

half-planes BC-A, AC-B, and AB-C, then

is

point of the interior of the triangle ABC. PROOF. Let O be any point other than
of the triangle.

P of the interior

Then O

like

planes
of Th.

BC-A, AC-B, and

belongs to each of the halfAB-C, and the interior of the

segment
i

PO

cannot intersect the triangle (see statement

of Art. 12).

from O, the half-line do intersect the triangle (Art. n), intersecting it in the two points of a segment whose interior contains O, and therefore P and lies entirely in the interior of the triangle.
7

But the opposite half-lines drawn OP and the half-line PO produced,

Therefore

P lies

in the interior of the triangle.

IV.

CONVEX POLYGONS
half-planes of a convex polygon. number of points, three or
this

14.

Polygons.

The

polygon consists of a finite

more, taken in a definite cyclical order, and the interiors

of the segments

whose points are consecutive points of

order.

The

riors of the

points are the vertices of the polygon, and the inteare segments are its sides. If A, B, C,
.

the vertices in order, the sides are the interiors of the segments AB, BCj ., and the polygon may be described
.
.

as the polygon The entire class of points beto the are in longing polygon cyclical order (Art. 6). The triangle is a particular case of the polygon.
. .

ABC

A diagonal of a polygon is the interior of a segment whose

points are two non-consecutive vertices of the polygon. A polygon must have at least four vertices to have a diagonal.

A plane polygon is a polygon which lies entirely in a plane.

If

two points

line of the plane,

of such a polygon are on opposite sides of a each of the two portions into which these

i3-i5l

LINK AND POLYGON

41

points divide the rest of the polygon will intersect the line, for in each portion there will be a vertex on the line or a
side that has points

on both

sides of the line.

A polygon is a simple polygon when no point occurs twice

as a point of the polygon.

By polygon we shall always mean

a simple plane polygon. A convex polygon is a simple plane polygon no point of which is a point of a side produced.
i // each pair of consecutive vertices of a polyin the edge of a half-plane which contains all of the gon the other vertices, polygon is a convex polygon.
.

THEOREM
lie

THEOREM

2.

Conversely, in a convex polygon each pair

of consecutive vertices lies in the edge of a half-plane which contains all of the other vertices.

PROOF.
let

Let

and

be two consecutive vertices, and

be any other vertices. If AB-D and AB-E were opposite half-planes, each of the two portions into

and

which

and

divide the rest of the polygon would con-

But the side tain at least one point of the line AB. lies entirely in one of these two portions, and no point of the other portion can be a point of the line if the

AB

AB

polygon

is

a convex polygon.

COROLLARY. // A and a convex polygon, and if P

point of the line
those of the line
15.

B
is

any point of

are two consecutive vertices of the polygon not a

AB,

then all points of the polygon except

points of the half-plane

AB will be

AB-P.

Intersection of a line

and a convex polygon.

THEOREM i. No line can contain more than two points of a convex polygon unless it contains one of the sides. PROOF. Let us suppose that a line a, not containing a
side of the polygon, could contain the three points H, and of the polygon, and coming between

K, on

42
the
line.

CONVEX POLYGONS
That
side of the

[i.

iv.

polygon of which
if

or one of the two sides which meet at

is

M M M
K

is

a point,

is

a vertex,

a part of a line i, the common edge of two opposite halfand respectively. planes which contain the points

But

this is

theorem.

contrary to the corollary of the preceding Therefore, a line cannot contain three points

of the convex polygon unless it contains one of the sides.

are two points of a convex 2. and If at in each of the two portions with vertex least one polygon,
into

THEOREM

which they divide

the interior of the

the rest of the polygon, then

and

and

segment

HK

will

form with each of

these two portions a convex polygon.

Let A, 5, ... be the vertices in one of these These points with taken in order from H. portions, and will then be the vertices of a polygon KH. This is a simple polygon, for no point of the interior

PROOF.

H
.

HAB

of the segment

HK H
;

(Th. i). continuation of

It is also a

(same reference)
point of

a point of the original polygon convex polygon: no point of either K can be a point of the original polygon can be a moreover, no point of
is

HK

some other side produced, for the line containing such a side would be the common edge of two opposite half-planes, one containing the point // and the other
the point K, which is contrary to the corollary in the last In the same way we prove that the other polygon article.
is

a convex polygon.

i Si

LINE AND POLYGON

43

THEOREM
points,

3.

If a line intersects a convex polygon in two

of the sides of the polygon, the two portions into which these points divide the rest of the polygon lie on opposite sides of the line.

and does not contain one

the two points is a point of a side and not a vertex, the extremities of this side will lie on opposite sides of the given line and as the polygon can inter-

PROOF.

If either of

sect the line in only the

into which
lie

points, the two portions these points divide the rest of the polygon must

two given

one entirely on one side and the other entirely on the

opposite side. It only remains, therefore, to prove the theorem in the case of a diagonal.

Let
tive

A E be any diagonal, D, E, and F being three consecuvertices. No three of the points A, D, E, and F can
i).
Z), like all the vertices except

be collinear (Th.

Now

and F,
,

of the half-plane

EF-A

(Art. 14,

Th.

2).

a point is a Again,
is

and being consecutive vertices, belong to one of the two portions into which A and F divide the rest of the polygon, and lie on the same side of the line AF. Then if D and F were on the D would be a point of the halfsame side of the line A plane A E-F, and would therefore be in the interior of the But this would put A triangle AEF (Art. 13, Th. 3).
point of the half-plane

AF-E\

for

44

CONVEX POLYGONS

[i.

iv.

and F on
2),

which

is

(see Art. 8, Th. opposite sides of the line It of Art. 14. theorem to the second contrary

ED

D and P lie on opposite sides of the line AE\ and the same must be true of the entire portions into which A and E divide the rest of the polygon.
follows that

THEOREM

4.

Any

line in the plane of a convex polygon

intersecting a side will intersect the polygon at least in a

second point. If the polygon

with

it

not a triangle, a diagonal will form two convex polygons, each having a smaller number
is

of vertices than the given polygon. Thus the theorem, of a can be true being triangle (Art. n), proved by induc-

tion to be true of

any convex polygon.

The interior of a con16. Interior of a convex polygon. vex polygon consists of the interiors of all segments whose points are points of the polygon, except of those whose
interiors also lie in the polygon.

// two segments lying in a convex polygon separate each other in the cyclical order of the points of the polygon (Art. 6), then their interiors will intersect unless the
i.

THEOREM

interior of one of these segments lies in the polygon.

PROOF. Let EF and be two segments separating each other in the cyclical order of the points of the polygon.

HK

The

and points and the points E and

lie

on opposite lie on opposite

sides of the line sides of the line

EF,

HK,

unless one of these lines contains a side of the polygon The line of each segment therefore passes (Art. 15, Th. 3).

between the points of the other, and the segments intersect.

interiors of the

two

THEOREM
vex polygon

2.

Any

half-line

drawn in

the plane of

a con-

from a point

of the interior will intersect the

polygon.

iS-irl

PLANE AND TETRAHEDRON

45

PROOF. The point O is in the interior of a segment H K whose points are points of the polygon. It lies, then, in the side H K of two convex polygons, all of which, except this side and the points H and K, lie on opposite sides of the line H K (Art. 15, Ths. 2 and 3). The theorem can therefore be proved in the same way that the theorem of
Art. ii
is

proved.

THEOREM 3. Any half-plane whose edge contains two consecutive vertices of a convex polygon, and which itself contains the other vertices (Art. 14, Th. 2), contains also all
the points of the interior.

THEOREM

4.

Conversely, if a point

lies

in each of the

half-planes whose edges contain two consecutive vertices of a convex polygon and which themselves contain the other vertices,

then

is
is

in

ttie

interior of the polygon.

The proof

the same as that of Th. 3 of Art. 13.

V.

TETRAHEDRONS
Intersection of a plane and a tetraconsists of four non-coplanar
interiors of the four triangles

17.

Tetrahedrons.

hedron.
points

tetraJiedron

and the

sides

and

whose

vertices are these points taken three at a time. The four points are the vertices of the tetrahedron, the
sides of the triangles are the edges,

and the

interiors of the

triangles are the faces. Any four non-coplanar points are the vertices of a tetrahedron.

We shall sometimes speak of the vertices and sides of a face, but it should always be remembered that a face of a tetrahedron is the interior of a triangle and does not include
the triangle
itself.
i.

THEOREM

The plane of
itself the

three non-collinear points of

plane of one of the faces, intersects the tetrahedron in a triangle or a convex quadrilateral.

a tetrahedron, if not

46
PROOF.
plane of

TETRAHEDRONS

[i.

v.

Two

points of the given intersection in the

any one face determine a line which contains a point in each of the planes of at least two other faces. If then we have another point of the given intersection in
the plane of one of these faces, we have a second line; and we can sometimes continue in this way and trace the
intersection completely

around the tetrahedron.

always be the case when one of the given points is a vertex. Any plane through a vertex and two points of the tetrahedron not collinear with the vertex, if not itself the plane of one of the faces, will intersect the tetraThis
will

hedron in a

triangle.

The
will

following construction, with slight modifications, provide for all other cases:

Let

A BCD
ABC,
will

K be three points of
the face
of the edge

be the given tetrahedron.

AD

Let H, F, and the given intersection a point of F a point of the face BCD, and a point or of one of the faces or ACD. A
:

K
',

ABD

produced produced in a point K', H' a point of the side interior of the segment H' K' lying in the interior of
y

meet the

triangle BCD in a point H

and A K BC and the

1

this

triangle.

I7l

PLANE AND TETRAHEDRON

If

47

F is a point of

the line

K', the given plane will pass

Let us suppose that F lies on the opposite side of this line from C. We have a point G common to the interiors of the segments CF and H' K' (Art. 12, Th. i, then in connection with and Art. 8, Definition and Ths.)
through A.
;

the triangle

AH

'

K' we have a point

common

to the
;

interiors of the

and finally, we have in and the segment

FM

(Art. 8, Th. i) segments AG and with in the interior of the triangle ACF, produced a point L of the side AC. But F

HK

tion.

H K lie in the plane of the given intersecTherefore, the line FM lies in this plane, and L
a point of the given intersection.
:

itself is

Now assuming that the given plane does not pass through
a vertex, we have the following cases
First,
if

we can take any one

the three given points lie in three different faces, of these points for the point and

obtain a point of the intersection in one side of this face, and then a line in which the given plane intersects the plane of this face. These lines will intersect the edges of the tet-

rahedron in three points or in four points, and the interThe given section will be a triangle or a quadrilateral. cannot three of intersect the sides a plane triangle (Art.
Th.), and therefore it cannot intersect of the edges of the tetrahedron.
7,

more than four

different faces

Again,

if

two

of the given points lie in

two

and the

third lies in the edge which is not a side of either of these faces, there will be lines in the planes of these two
faces intersecting in their common side, and intersecting two other edges of the tetrahedron in two points which,

with the third given point, determine the rest of the intersection. The intersection will in this case be a quadrilateral.

In the third place,

if

and the other two

in edges

one of the given points lies in a face which are not sides of this face,

48
let

TETRAHEDRONS

[i.

v.

lies in the face ABC, F in the edge us suppose that the in and BD, edge AD. This change in the position

of

will

not affect
line

th'e

determination of the point

there

will

be a

EL intersecting the edge BC, and the intersec-

The half -line cannot meet tion will be a quadrilateral. and BD. the edge AB, for the plane already intersects if in three lie three not all the given points edges Finally,
meeting at one vertex nor lying in one plane, we may supis a point of BC, F a point of BD, and a pose that

LH AD

point of
cide,
;

AD.

but this
still

L we

In the construction above, and coinnot affect the determination of the point have a point of the intersection in each of four
will

edges, and the intersection will be a quadrilateral. All cases in which the intersection cannot be determined
directly
four.

from the given points can be reduced to one

the intersection
is

of these

When

a quadrilateral,

it is

a convex

any any other face, and no point of any edge lies in the plane of any face except of the two faces which have this edge for a common side. Therefore no point of
quadrilateral. lies in the plane of

No

point of

face of the tetrahedron

any

side of the quadrilateral can be a point of a line containing another side, and no vertex can be collinear with
vertices.
2.

two other

THEOREM
PROOF.

No

line
it

a tetrahedron unless

can contain more than two points of lies in the plane of one of the faces.
line

Let a given

a intersect the tetrahedron in

two points
point of
a.

H and F, and

let

be a vertex which

is

not a

The plane A HF will contain

faces.

the line a and will

intersect the tetrahedron in a triangle,

if it

does not contain

the tetrahedron

one of the
will

The intersection of a and

its

then be the same as

intersection with the triangle,

and

will consist only of the

two points

and

(Art. 7,

Cor.).

17-iQl

POINTS

AND TETRAHEDRON
The
all

49
a
tetra-

18.

Interior of a tetrahedron.

interior of

hedron consists of the interiors of

segments whose points

are points of the tetrahedron, except of those segments whose interiors also lie in the tetrahedron.

THEOREM.

half-line

drawn from any point

P of a tetra-

hedron through a point O of the interior intersects the tetrahedron in a point of PO produced. PROOF. The point O is given as in the interior of a segment whose points are points of the tetrahedron. If P is

one of these points, the other

half-line

PO
If

intersects

produced.

is

will be the point in which the the tetrahedron, a point of PO not one of these points, these points

plane intersecting the tetrahedron in a triangle or convex quadrilateral, and the theorem follows as a theorem concerning this intersection (see Art. 16,
will lie in a

and

Th.

2).

19.
is

The

relation, collinear with

a tetrahedron.

point

said to be collinear with a tetrahedron

when

it is

collinear

with any two points of the tetrahedron. This is true, in particular, of all points of the tetrahedron,
of all points in the planes of its faces,
its interior.

and

of all points of

THEOREM

If a point

is collinear

with a tetrahedron

A BCD, and
the line

PO

not the point O, then will intersect the tetrahedron at least in a second

if P is any point of a face,

point Q.

be the two points with which O is collinear. Through these two points and P we can pass a plane intersecting the tetrahedron in a triangle or convex

PROOF.

Let

and

or containing a face of the tetrahedron. will the intersection is a triangle or quadrilateral, be a point of a side, and thus in any case the line PO will
quadrilateral,

When

50

HYPERPLANES

[i.

vi.

contain at least a second point

9,

of the tetrahedron

(Art.

Th.

i,

and Art.

15,

Th.

4).

THEOREM
is collinear

2. If a point O, not a point of the plane BCD, with the tetrahedron A BCD, then any point P

collinear with the tetrahedron

the tetrahedron

A BCD

will be collinear with

OBCD.

PROOF.
with

Take a point and P. The line

H in

the face

BCD,

not collinear
tetrahedron

HO will intersect the

in a second point F,

The
in

and the line HP in a second point K. and K are not collinear, and do not lie F, H, points the plane of any one face of the tetrahedron A BCD.

The plane intersects this tetrahedron in a triangle or convex quadrilateral, one side of which lies in the face and contains the point H. That is, the plane intersects the triangle in a segment and therefore

HFK

BCD

the tetrahedron
of the side

BCD MN, OBCD in a triangle OMN, with H a point MN. The line HP must then intersect this
and

triangle in a second point,

must be

collinear with

it.

Therefore

P is collinear with

the tetrahedron

OBCD.

Remarks

made

at this point.

similar to those at the end of Art. 9 may be The class of points collinear with the

the same as the class of points collinear with the tetrahedron A BCD. We can use A and O

tetrahedron

OBCD

is

interchangeably in this connection.

VI.

HYPERPLANES
Only one hyperplane contains four

20.

Hyperplanes.

given non-coplanar points. Our space a hyperplane. A hyperplane consists of the points that we get if we take four points not points of one plane, all points collinear

with any two of them, and obtained by this process.

all

points collinear with

any two
Z>,

Given four non-coplanar points, A, B, C, and

the

ig, 20]

DETERMINATION OF A HYPERPLANE

51

hyperplane
definition.

A BCD is the hyperplane

these points

obtained when we take and carry out the process described in the

THEOREM

i.

// two points of a line

lie

in a given hyper-

plane, the line lies entirely in the hyperplane ; and if three non-collinear points of a plane lie in a given hyperplane,
the plane lies entirely in the hyperplane.

For the

line or

plane can be obtained from these points

by

the process used in obtaining the hyperplane.

a plane having two points in a given hyperplane, but not lying entirely in it, will intersect the hyperplane in the line which contains these two points.
It follows that

THEOREM
we

2.

From

the points of the figures given in each the points of

a hypertwo plane if any of them and all points collinear with any two obtained by this process : (1) A plane and a point not in it, or a plane and a line
of the following cases
take all points collinear with
that intersects
(2)
it

we can obtain just

but does not lie in it;

lines not in one plane; Three lines through one point but not in one plane; (3)
(4)

Two

Two

planes intersecting in a line.

can, indeed, speak of a line or a plane as one of the things with which we start in the process of obtaining a

We

hyperplane (compare this with Art. 10). It follows from (i) that a line and a plane which do not
in a hyperplane do not intersect at all, and from (4) that two planes which do not lie in a hyperplane cannot have more than one point in common.
lie

THEOREM
points

3.

//

',

B', C", and

are four non-coplanar

of the

A'B'C'D'

is the

hyperplane A BCD, then the hyperplane same as the hyperplane A BCD.

52

HYPERPLANES

[i.

vi.

The proof

follows the lines of proof of the

two theorems

given in Art. 10.

In regard to Th. 2 we can now say that the hyperplane obtained in each case is the only hyperplane that contains the given figures.

Three non-collinear points can be points of two

different

hyperplanes. The intersection of the hyperplanes will then be the plane of the three points (see Art. 27, Th. 2).

Apparently we get

all

the points of ordinary space

all

by

taking four non-coplanar points,

any two of them, and all tained by this process. The space

points collinear with with any two obcollinear points

of our experience will

therefore be regarded as a hyperplane.

21.

Intersection of a line or plane and a tetrahedron in

a hyperplane.
i. In the hyperplane A BCD, any line intera secting face of the tetrahedron A BCD will intersect this tetrahedron at least in a second point, and any half-line drawn

THEOREM

from a point
PROOF.

of the interior of the tetrahedron will intersect

the tetrahedron.

Th.
line

i,

follows directly from Art. 19, since the line contains points collinear with the

The first statement

tetrahedron.

To prove
18,

the second statement, let the half-

AO intersect the face BCD in a point P

Th.).
If

produced (Art.
half-line

a point of AO we have given any other

y

in the hyperplane, any point 0' of this half-line will be collinear with the tetrahedron (see

drawn from

reference in connection with Art. 20, Th. 3), and the line PO' will intersect the tetrahedron at least in a second point Q not collinear with A and (Art. 19, Th. i). The plane

APQ will

intersect the tetrahedron in a triangle;

and the

given half-line,

being drawn

in this plane

from a point in

20-231

INTERSECTION OF PLANES

53

the interior of the triangle, will intersect the triangle, and therefore the tetrahedron.
2. In the hyperplane a tersecting face of the tetrahedron

THEOREM

ABCD,

any plane

in-

ABCD

(if not itself the

of the interior of the tetrahedron, will intersect the tetrahedron in a triangle or convex quadrilateral.

plane of this face), or any plane containing a point

For

in either case

by drawing

lines in the

plane

we can

obtain three non-collinear points of the intersection.

22.

THEOREM.
a point
O.

Intersection of two planes in a hyperplane. // two planes lying in a hyperplane have

in common, they have in

common a

line through

Let a and /3 be the two given planes. Let C be two points of a not collinear with O, and take P a point between C and D, B a point of PO produced, and A any point of the hyperplane not a point of a. The point
PROOF.

and

lies in

the face

BCD

of the tetrahedron

ABCD, and we

can consider the given hyperplane as the hyperplane of

this tetrahedron. The plane /3, containing O, will intersect the tetrahedron in a triangle or convex quadrilateral (Art. 21, Th. 2), and a, the plane of the face BCD, in the line

which contains one side

a line through O.

of the triangle or quadrilateral,

23. Opposite sides of a plane. Half-hyperplanes. The four half-hyperplanes of a tetrahedron. THEOREM i. Any plane of a hyperplane divides the rest

of the hyperplane into two parts, so that tlie interior of a segment lying one point in each part intersects the plane, and
the interior of

a segment lying both points in

i,

the

same part

does not intersect the plane. The proof of Art. 12, Th.

applies almost without

54

HYPERPLANES

[i.

VI.

change, the reference to Art.

Art. 22.

holding true by virtue of

into which any plane of a hyperplane the rest of the hyperplane are said to lie on opposite divides sides of the plane, and the plane is said to lie between them.

The two parts

That portion

of a hyperplane

which

lies

on one side of

a plane of it is called a half-hyperplane, and the plane is the face of the half-hyperplane. The half-hyperplane for its face and contains the which has the plane

ABC

point

will be called the half-hyperplane

ABC-D.

The

two half-hyperplanes into which any plane

hyperplanes.

of a hyperplane

divides the rest of the hyperplane are called opposite half-

// two planes intersect in a line, the opposite half-planes of each, which have this line for a common edge, lie in their hyperplane on opposite sides of the other.
2.

THEOREM

THEOREM

3.

Every
or

point

of

the

hyperplane

belongs to one, at least, of the four half-hyperplanes

A BCD BCD- A,

ACD-B, ABD-C,
The proof
is

ABC-D.
Th.
i.

like that of Art. 13,

THEOREM
hedron

4.

Any

point

of the interior of the tetra-

A BCD

belongs to all four of the half-hyperplanes

BCD- A, ACD-B, ABD-C,

versely, if
it is

and

ABC-D

and,

con-

a point belongs
2

to all

four of these half-hyperplanes,

a point of

the interior of the tetrahedron

A BCD.

See Art. 13, Ths.

and

3.

We
and

shall

sometimes speak of a tetrahedron as a surface

of its interior as a solid.

tetrahedron divides the

interior

rest of its

hyperplane into

two portions,

and

ex-

terior to the tetrahedron.

23, 2 4 ]

PLANE AND PYRAMID

55

VII.

CONVEX PYRAMIDS AND PENTAHEDROIDS

Intersection of a plane and a convex pyramid. pyramid consists of a simple plane polygon which has an interior, and a point not in the plane of the
24.

Pyramids.

polygon, together with the interior of the polygon and the interiors of the segments formed by taking the given point

with the points of the polygon. The point is called the vertex of the pyramid, and the The interiors of the interior of the polygon is the base.

segments formed by taking the vertex and the vertices of

the polygon are the lateral edges, and the interiors of the triangles determined by the vertex and the sides of the
to the vertex

base are the lateral faces. Often the term vertices is applied and the vertices of the polygon taken together,

the term edges to the lateral edges and the sides of the polygon, and the term faces to the lateral faces and the base. Sometimes, also, we shall speak of the vertices and
sides of the base or of a lateral face (see

remark at the begin-

ning of Art. 17).

The tetrahedron

When
pyramid.

a particular case of the pyramid. the polygon is convex, the pyramid is a convex
is

We

shall consider only

convex pyramids.

THEOREM. The plane of any three non-collinear points of a convex pyramid, if not itself the plane of one of the faces, intersects the pyramid in a convex polygon*
PROOF. If the plane passes through the vertex, the intersection will be a triangle, as in the case of a tetrahedron
(Art. 17).

Let

be the vertex, and

let

A, B, C,

be the vertices

of the base.

Suppose the plane, not passing through O, intersects the base in the interior of the segment H K, and K

It

may

be a

triangle,

which

is

a particular case of the convex polygon.

56

PYRAMIDS AND PENTAHEDROIDS

[i.

vn.

being points of the polygon

AB

dividing the rest

of this polygon into two portions with each of which they form a convex polygon and the interior of the segment
(Art. 15,

Th.

2).

The
into

HK triangle OHK

will divide the rest

two portions, with each of which it pyramid and its interior will form a convex pyramid. Now the base and the vertex of one of these pyramids will lie on
of the

opposite sides of the given plane, while all of the other and the points of its pyramid except the segment The given plane interior lie on the same side of this plane. will not intersect the second pyramid except along the

HK

line

H K,

and the given

intersection

is

the same as the

intersection of the plane with a pyramid and base lie on opposite sides of the plane.

whose vertex

O-ABC

Let us suppose, then, that we have a plane and a pyramid whose vertex and base lie on opposite sides
.

The plane may contain one vertex of the base, or a side and two vertices, but the base itself does not intersect the plane, and all the rest of the intersection In particular, lies in the lateral faces and lateral edges.
of the plane.

any
will

lateral

edge

OA

will intersect the

meet the plane at A.

plane in a point A or Then the face OAB will in'

tersect the plane in the interior of a

segment

A'B

',

or

AB',
;
.
. .

unless (in one particular case) A B itself lies in the plane and the entire intersection will be a polygon A'B'C'

a one-to-one correspondence of the points of this corpolygon and the points of the polygon with O. There collinear is, indeed, responding points being
is

There

ABC

a correspondence of the entire planes of these two polygons so far as there are points in both collinear with O. There
is

also

a correspondence in the order of the points of the

in the order of the points of

two polygons, and

any two

corresponding lines (Art. 8, Ths. i and 2). If, then, a point P' were to be found twice among the points of the

24, 25l

PLANE AND PENTAHEDROID

. .

57

., the corresponding point P would be polygon A'B'C' or if a point found twice among the points of ABC of the former polygon were also a point of some side produced, the same would be true of the corresponding point
.

in the plane of the base.

It follows that the intersection

must be a simple convex polygon, the polygon

being such a polygon (Art.
25.
terior.

ABC

14, Defs.).

Pentahedroids.

Intersection
relation.

with a plane.

In-

pentahedroid consists of five points not points of one hyperplane, and the edges, faces, and interiors of the five tetrahedrons whose vertices
are these points taken four at a time. The five points are the vertices, the edges and faces of the tetrahedrons are edges and faces of the pentahedroid,

The

collinear

and the

interiors of the tetrahedrons are its cells.

Any

five

points, not points of one hyperplane, are the vertices of a pentahedroid.

We shall sometimes speak of the vertices, edges, and faces

should always be remembered that a cell of the pentahedroid is the interior of a tetrahedron and does
of a cell,

but

it

not include the tetrahedron

itself.

THEOREM.
a pentaltedroid,

The plane of

three

if it does not itself lie

non-collinear points of in the hyperplane of one

of the cells, intersects the pentahedroid in

a convex polygon.

PROOF.
first

As

in the case of a tetrahedron (Art. 17,

Th.

i,

part of the proof),

we can sometimes determine

the

intersection directly from the three given points, two points in the hyperplane of one cell determining a line which

contains a point in each of the hyperplanes of at least two other cells. In particular, this will be true when one of

the given points

is

a vertex.

be the pentahedroid, and let Let not in the plane of the given points. If

ABCDE

be a vertex

is

any point

58
of the

PYRAMIDS AND PENTAHEDROIDS

pentahedroid other than A, the half- line A of the opposite cell in a point

[i.

vn.

will
//'.

meet the hyperplane

This point
the

we

shall call the projection of //

from A.

//'is

same
all

as //

when
2,

// itself
is

In

other cases //'

a point of the cell BCDE. a point of the tetrahedron

is

BCDE

(see Art. 8,

Th.

and Art.

18).

Let the three given points be //, F, and A', and let //', r be their projections from A. These last three F', and

points are not collinear;

for, if

they were, the plane de-

and A would be a plane the three given points and the vertex A containing They are not all in the plane of any one face of the tetrahedron BCDE for, if they were, //, F, and K would be in the hyperplane determined by this face and A, and the given plane would lie entirely in this hyperplane. Therefore
termined by their

common

line

the plane H'F' K' intersects the tetrahedron BCDE in a triangle or convex quadrilateral (Art. 17, Th. i). This triangle or quadrilateral is the base of a pyramid with

vertex at A which lies entirely in the hyperplane A H'F' K' and which is, indeed, the intersection of the hyperplane and pentahedroid (see Art. 31, Th. i). The points //, F, and K are points of this pyramid, and
,

the plane

HFK
it is

lies

entirely in the hyperplane of this

pyramid.
droid
is

The intersection of the plane and the pentahethe same as the intersection of the plane and this
a convex polygon (Art.
24, Th.).

pyramid:

As there are only five cells, the intersection can only be a triangle, a quadrilateral, or a pentagon.

We

define interior of a pcnlahedroid

and

collinear with

a pentahedroid exactly as we define the corresponding expressions for the tetrahedron, and for the pentahedroid we have theorems analogous to those of Arts. 18 and 19.

25, 26]

RESTRICTION TO FOUR DIMENSIONS

59

four.

In a pentahedroid each tetrahedron is adjacent to each of the other If we move along a line through the interior of one of the tet-

rahedrons until

we come

to a face, a continuation of our

movement

take us entirely away from the pentahedroid. along But we can change the direction of our path so as to pass into the interior of an adjacent tetrahedron, thus following a broken line that
this line will

we move along a
to

This is like what we do belongs entirely to the pentahedroid. line in one face of a polyhedron until we

when come

an edge, and then change the direction of our path so as to pass an adjacent face. It may not be very difficult to think of two adjacent tetrahedrons, even though they lie in different hyperplanes, but it is quite impossible for us to form any picture of the pentahedroid All that we should try to do now is to remember in a as a whole. mechanical way the numbers and relations of the different parts.
across

VIII.

SPACE OF FOUR DIMENSIONS

The hyperspace
of this

26.

Space of four dimensions.

book.
that

we

space of four dimensions consists of the points get if we take five points not points of one hyper-

plane,

all

points collinear with

points collinear with any two of them, and any two obtained by this process.

all

a series of theorems analogous to others already given for the plane and hyperplane, we can prove that any five points of a given space of four dimensions, not points of

By

one hyperplane, determine the same space of four dimensions.

We

shall limit ourselves to a discussion of space of four

dimensions.

We

shall
lie

and that

all

points

assume that there in one such space.*

is

such a space,

The preceding
:

theorem can, therefore, be stated as follows

any five points not points of one hyperplane, all points collinear with any
i.

THEOREM
* This

We

get all points if

we

take

is

of the phrase "in

a mere matter of convenience, enabling us to avoid frequent repetition one space of four dimensions," and to state many theorems in a

simpler form.

60

SPACE OF FOUR DIMENSIONS

all

[i.

vm.

two of them, and

points collinear with

any two obtained

by

this process.

The word

hypersfiace

is

used for any space of more than

three dimensions, but as the only hyperspace that we shall consider is the space of four dimensions we shall use these

terms interchangeably.

THEOREM 2. Any line intersecting a cell of a pentahedroid will intersect the pentahedroid at least in a second point, and any half-line drawn from a point O of the interior of
a pentahedroid will
intersect the pentahedroid.

THEOREM
droid, if
it

Any plane intersecting a cell of a pentahe3. does not itself lie in the hyperplane of this cell,

or

any plane containing a point

of the interior of the penta-

hedroid, will intersect the pentahedroid in a convex polygon.

These two theorems correspond to the two theorems of For theorems on the intersections of pentaheArt. 21. droids and hyperplanes see Art. 31.
Intersection of a plane and a hyperplane hyperplanes. The linear elements of planes.
27.

and

of

two

THEOREM

O in

If a plane and a hyperplane have a point common, they have in common a line through O. The proof is like that of the theorem of Art. 22. We
i.

construct a pentahedroid with a cell lying in the given hyperplane and containing O. The given plane intersects

the pentahedroid in a convex polygon and the given hyperplane in the line which contains one side of this polygon.

THEOREM
PROOF.

2.

mon, they have in

// two hyperplanes have a point common a plane through 0.

/8

in com-

Let a and

be the two hyperplanes.

Any

in a line through 0, plane of a through O will intersect the theorem a and second by preceding plane of a through
;

26,27)

PLANES IN HYPERSPACE

61
first

O, drawn so as not to contain the line in which the plane intersects /3, will intersect /? in a second line.

We

have two

lines

through

O common

to the

two hyperplanes.

The hyperplanes

therefore intersect in a plane (see Art. 20).

common
and

Three hyperplanes having a point in common have in at least one line, a line lying in one hyperplane

in the plane of intersection of the other two. hyperplanes may also have a plane in common.

Three

planes which do not lie in one hyperplane contain a set of lines, one and only one through each point of either plane which is not a point of the other plane, and any two of these lines coplanar.
3.

THEOREM

Two

PROOF.
point of
of a.

Let a and
Let

/3

one hyperplane.
|8,

be two planes which do not lie in be any point of a which is not a

be any point of /? which is not a point The hyperplane determined by and A intersects
let

and

in a line a,
/3

intersects

in a line b (Th. i).

lie

and the hyperplane determined by a and B The lines a and ft, each
in the plane of intersection

lying in both hyperplanes, of the hyperplanes.

the other hand, no two lines lying in one of the given planes and coplanar with lines in the other can intersect
in a point
for, if

On

which is not common to the two given planes they did, both of them and the entire plane in which
;

they

lie

would

lie

in the

hyperplane determined by their

point of intersection

and the other given plane. and might them the linear When the two planes have a
these lines,
shall call

The planes a and j8 are covered with

be said to consist of them.
elements of the

We

two

planes.

point in
point.

common, the If any plane

linear elements all pass

through this
lines,

intersects the

two planes in

these lines are linear elements.

62

SPACE OF FOUR DIMENSIONS

The student may prove
If

[i.

vm.

the following theorem

in

two planes have only one point

common, then through any

point not a point of either plane passes one and only one plane intersecting both the two planes in lines.

28.

Opposite sides of a hyperplane. Half-hyperspaces. hyperplane divides the rest of hyperspace just as a plane

in a hyperplane divides the rest of the hyperplane (Art. can speak of the opposite sides of a hyperplane, 23).

We

have, for example, the lying on one side of the hyperhalf-hyperspace A and the BCD, plane opposite half-hyperspace lying on the
of a half-hyperspace.

and

We may

ABCD-E

opposite side of this hyperplane. The hyperplane is the If two hyperplanes intersect, cell of the half-hyperspace.
the opposite half-hyperplanes of each which have the plane
of intersection for a

common

face

lie

on opposite

sides of

the other.

Given a pentahedroid, each of the five tetrahedrons determines the cell of a half-hyperspace which contains the opposite vertex and all points of the interior; and,
conversely,
it will lie
if

a point

lies in all five of

these half-hyperspaces

in the interior of the pentahedroid.

Every point

of hyperspace is a point of at least

one of these half-hyper-

spaces.

sometimes speak of a pentahedroid as a hypersurface (see Art. 33), and of its interior as a hyper solid. A pentahedroid divides the rest of hyperspace into two
shall

We

portions, interior

and

exterior to the pentahedroid.

As a

line divides a plane into

two

parts, but not ordinary space (a

hyperplane), so a plane divides a hyperplane, but not hyperspace.

some one point

point in
this in

In hyperspace we can pass completely around a plane, keeping near of it, without touching the plane at all. We can do this, for example, by keeping in another plane which has only one

common with the given plane. The student should bear mind when he comes to consider absolutely perpendicular planes (Art. 42) and rotation around a plane (Art. 81).

27-30]

HYPER PYRAMID AS POLYHEDROID

IX.

63

HYPERPYRAMIDS AND HYPERCONES

Knowledge
will this

29.

Introduction.

posed.

In

section

polyhedrons presupbe introduced the hyper-

of

pyramids and hypercones of four-dimensional geometry, with a few theorems analogous to those already given for convex polygons and convex pyramids. We shall assume
that the term polyhedron has been defined, at least so far

concerned, together with vertices, edges, and faces, and the expressions convex polyhedron and interior of a polyhedron. We shall also speak
as the hyper plane polyhedron
is

and surface, and, in particular, of the circle, the sphere, and the circular conical surface; and we shall use such theorems as we need concerning these surfaces, and
of curve

concerning polyhedrons (see Introduction to chap.

II).

We may

find

some

difficulty in

sidered, but that need not deter us

understanding the forms here confrom going on, as the next three

of

chapters do not depend on this part of our text. Indeed, the study hyperpyramids and hypercones could be omitted entirely until

we come again

to their treatment in chap. V.

We
the

30. Hyperpyramids. Interior of a hyperpyramid. Figures in hyperspace which correspond to the polyhedrons of geometry of three dimensions are called polyhedroids. shall not attempt to define this term, except to say that
vertices,

a polyhedroid consists of
cells

edges, faces,

and

cells,

being the interiors of certain hyperplane polyhedrons joined to one another by their faces so as to enclose

a portion of hyperspace, the interior of the polyhedroid. We shall apply the term polyhedroid only to certain simple
figures

which we

shall define individually.

The pentahe-

droid

is

the simplest polyhedroid.

hyperpyramid consists of a hyperplane polyhedron enclosing a portion of its hyperplane, and a point not a

64

HYPERPYRAMIDS AND HYPERCONES

[i.

ix.

point of this hyperplane, together with the interior of the polyhedron and the interiors of the segments formed by taking the given point with the points of the polyhedron. The point is the vertex, and the interior of the polyhedron
is

the base.

The meaning

of other

terms used in connection

with the hyperpyramid

definitions of Arts. 24

may
and

25.

be readily inferred from the The pentahedroid is the

simplest hyperpyramid. The interior of a hyperpyramid can be defined as con-

segments formed by taking the vertex with the points of the base, but in the case of a convex hyperpyramid the interior of any segment whose
sisting of the interiors of the

points are points of the hyperpyramid will lie entirely in the interior of the hyperpyramid unless it lies entirely in the hyperpyramid
itself.

No

line

can intersect a convex

hyperpyramid in more than two points unless it lies in the hyperplane of one of the cells, and any half-line drawn from a point O of the interior will intersect the hyperpyramid in one and only one point.
In the
in
half-line

attempting to understand

we have Since a geometry can be drawn from O through any point of the hyperpyramid,
last

statement

is

involved one of the

difficulties that

of four demensions.

each point of the hyperpyramid plays a part in separating the interior from that portion of hyperspace which is exterior to the hyperpyramid. When we think of a solid as forming a part of the boundary of a figure,

we

terior

are apt to think of one face of the solid as coming next to the inand the other face as on the outside, and we think of the thick-

in a very different
terior

ness of the solid as representing the thickness of the boundary. It is way that the cells of a polyhedroid separate the in-

from the

outside.

Each

point,

no matter how

far within the

its re-

exposed both to the interior and to the outside, and moval would allow free passage to a half-line from 0.
solid, is

Hyperplane sections of a hyperpyramid. A hyperpyramid or any polyhedroid can be cut by a hyperplane

31.

30, 3i]

SECTIONS

65

in a section*

which divides the rest of the polyhedroid into two parts lying on opposite sides of the hyperplane (Art.
28).

The

sections of a polyhedroid are polyhedrons

whose

faces are the sections of the cells of the polyhedroid made by the planes in which the hyperplane intersects the

hyperplanes of the
are proved

cells.

Thus the following theorems

hyperpyramid orpentahedroid
sections of their cells
:

in regard to the sections of a

by considering the plane

i.

THEOREM

section of a convex

hyperpyramid made

by a hyperplane containing the vertex, is a convex pyramid whose base is the corresponding plane section of the base of In the case of a pentahedroid this applies the hyperpyramid.
to

any

vertex.

When

one vertex of a pentahedroid

lies

in

a hyperplane and two vertices of the opposite cell lie on each side of the hyperplane, the section will be a quadrilateral pyramid. In all other cases the section of a pentaliedroid
by a hyperplane containing one vertex and not containing a cell will be a tetrahedron.

THEOREM

2.

hyperplane passing between one vertex

the opposite tetrahedron will intersect

of a pentahedroid

and

the pentahedroid in

a tetrahedron.
lie

THEOREM
will be

3.

// two vertices of a pentahedroid

on one

side of a hyperplane

and

three

on

the opposite side, the section

a polyhedron in which there are two triangles separated

by three quadrilaterals.
PROO**. Let ABCDE be the pentahedroid, A and B on one side of a hyperplane a, and C, D, and E on the opposite side. The hyperplanes of the tetrahedrons ACDE and BCDE are cut by a in planes which pass between the
* The word section is somewhat synonymous with intersection, but in general not be used except for an intersection which divides the rest of a figure into two parts and completely separates these parts. Thus a triangle or convex quadrilateral would be a section of a tetrahedron.
it

will

66

HYPERPYRAMIDS AND HYPERCONLS

[i.

ix.

and the points A and B (Art. 28), and which triangle The hypertherefore cut these tetrahedrons in triangles. planes of the other three tetrahedrons are cut by a. in planes
which pass between their common edge AB and the opposite edges CD, DE, and CE, and which therefore cut
these tetrahedrons in quadrilaterals.
triangles to be
will
If

CDE

we suppose

the

C'D'E' and C"D"E", the quadrilaterals be C'D'C'D", D'E'D"E", and C'E'C"E", and the

section of the pentahedroid will be C'D'E'-C"D"E". If a. intersects the line in a point P, the three lines

AB

C'C", D'D", and E'E" will pass through P, and the section will be a polyhedron which may be called a truncated
tetrahedron.
this

In any case the section

p.

will

be a figure of
exercise at

type (see Introduction,

14,

and the

the end of Art. 128).

section of a figure in hyperspace

is all

that

we can

see in

any

hyperplane (see Introduction, p. 18). We can, however, see any section, and we can learn much about a figure by studying its various
sections.

32.

Double pyramids.

hyperpyramid whose base

is

the interior of a pyramid may be regarded in two ways as a hyperpyramid of this kind, the vertex of the base in

one case being the vertex of the hyperpyramid in the other

case.

Thus

mon

there are two pyramids having themselves a combase, and we can say that the hyperpyramid is deter-

mined by a polygon and two points neither of which is in the hyperplane containing the polygon and the other Looked at in this way the hyperpyramid is called point.
a double pyramid. A double pyramid consists, then, of the following classes
of points
:

(i) the points of a convex polygon, or of any plane

31,32]

WITH PYRAMID BASE

67
its in-

polygon which has an interior, and the points of

terior
(2)
;

two points not in *a hyperplane with the polygon, segment formed of these two points, ind the interiors of the segments formed by taking each 3f these points with the points of the polygon (3) the interiors of the triangles formed by taking each point of the polygon with the two given points (4) the interiors of two pyramids each formed by taking the polygon with one of the two given points. The interior of the segment of the two given points is called the vertex -edge of the double pyramid, and the inthe interior of the
; ;

terior

of the

polygon

is

the base.

The

interiors of the

triangles (3) are called elements, and, in particular, those elements whose planes contain the vertices of the polygon

are lateral face elements or lateral faces of the double pyramid. The two pyramids (4) are called the end-pyramids.

The vertex-edge and the sides of the base are opposite edges of a set of tetrahedrons. These tetrahedrons are in a definite cyclical order corresponding to the sides of the polygon, and are joined, each to the next, by the faces
which are adjacent to the vertex-edge. They arc joined to the end-pyramids by the faces which are adjacent to
the sides of the base.
are the lateral
cells,

The

interiors of these tetrahedrons interiors of the end-

and these and the

is

pyramids are the

cells of

The pentahedroid

the double pyramid. the simplest double pyramid.

Certain cases of intersection of double pyramids are given by the following theorems
:

THEOREM

i.

edge and intersecting or a plane containing

plane containing a point of the vertexthe base in the interior of a segment y the vertex-edge and a point of the base y

will intersect the double

pyramid in a

triangle.

68

HYPERPYRAMIDS AND HYPERCONES

[x.

DC.

In the latter case two sides of the triangle are in the

interiors of the end-pyramids.

THEOREM
a pyramid.

2.

hyperplane containing the base and a

point of the vertex-edge will intersect the double

pyramid in

THEOREM
and
a tetrahedron.

3.

hyperplane

containing

the

vertex-edge

intersecting the base will intersect the double

pyramid in

In this case two faces of the tetrahedron are sections of

the interiors of the end-pyramids.

We

have already referred to the importance of studying the

sec-

Another way of studying a tions of a figure of hyperspace (Art. 31). polyhedroid is to examine its cells. These cells can be cut apart
sufficiently to spread

them out

in

a hyperplane where we can see

are the interiors of polyhedrons; and we can think of a polyhedroid as formed by taking a suitable set of polyhedrons, plac-

them.

They

ing them upon one another, and folding

them away from our hyperplane until they come together completely. In a pentahedroid, for example, there are five tetrahedrons whose twenty faces fit together in pairs, each tetrahedron having a face in
with each of the other four.

common

We

can take any one of these

tetrahedrons and place the other four upon it, all in one hyperplane, and then we can turn the four outside tetrahedrons away from this

hyperplane without separating them from the

in

fifth

or distorting

them
their

any way

(Art. 81), until

faces.

we have brought
five

together every pair of together

corresponding
interiors

The

tetrahedrons

with

analogous to the process of forming a tetrahedron by placing three of its faces in a plane around the fourth, turning them about the sides which lie upon
is

now

enclose a portion of hyperspace.

This

the sides of the fourth, and bringing them together completely.* Another way of constructing a pentahedroid is by means of

its

ten edges. Any tetrahedron will have six of these edges and will be connected with the opposite vertex by the remaining four. We can take these four and cause them to stand out from the vertices of the
tetrahedron,
all in

one hyperplane.

In hyperspace these four edges

* Other examples of this method of studying polyhedroids are given in notes in

Arts. 135, 136,

and

141.

32,331

HYPERSURFACE WITH LINE ELEMENTS

69

can be turned so that their ends come together at a point, and with the faces of the given tetrahedron determine the other four tetrahedrons and the complete
figure.

Hyperconical hypersurfaces. Hypercones. Hypersurface is the term applied to a figure in hyperspace which
corresponds to the surfaces of geometry of three dimensions. We shall not attempt to define hypersurface, and

33.

we

shall

use the word only in connection with certain

simple figures which we shall define individually. hyperplane is the simplest hypersurface.

The

mined by the points

hyperconical hypersurface* consists of the lines deterof a hyperplane surface and a point

not in the hyperplane of this surface.

The point
surface,

is

the vertex,

the surface

is

the directing-

and the

lines are the elements.

The hyperconical

hypersurface has two nappes.

The only hyperconical hypersurfaces which we have to consider at present are those in which the directing-surface is a plane, a sphere, a circular conical surface, or a part or
combination of parts of such surfaces. surface is a plane, the hypersurface
portion of a hyperplane.

When
is

the directing-

a hyperplane or a

hypercone consists of a hyperplane surface, or portions

of hyperplane surfaces, forming a closed hyperplane figure, and a point not a point of the hyperplane of this figure,

together with the interior of the latter and the interiors of the segments formed by taking the given point with the points of the hyperplane figure.
the vertex, the interiors of the segments are elements, and the interior of the hyperplane figure is the
is

The point

base.
* See foot-note, p. 220.
cal

We shall

sometimes use the word conical

for hyperconi-

when

there

is

no ambiguity.

70

HYPERPYRAMIDS AND HYPERCONES

[i.

ix.

A hypercone, or the hypersolid which we call the interior of a hypercone (see below), can be somewhat inaccurately described as cut from one nappe of a hyperconical hypersurface by the hyperplane of the direct ing-surface.
The hyperpyramid may be regarded as a particular case The only other cases which we have of the hypercone.
to consider at present are those in which the base interior of a sphere or of a circular cone.
is

the

plane containing the vertex of a hypercone and intersecting the base in the interior of a segment, will intersect the hypercone in a triangle; and a hyperplane containing the vertex and intersecting the base, will intersect the hypercone in a cone.

The

interior of a hypercone consists of the interiors of the

segments formed by taking the vertex with the points of the base, but in the case of a convex hypercone the interior
of

any segment whose points are points

the hypercone
itself.

of the

hypercone

will lie entirely in the interior of the

lies entirely in

No

hypercone unless it line can intersect

a convex hypercone in more than two points if it passes through a point of the interior, and any half-line drawn

from a point O of the interior in one and only one point.

34.

will intersect the

hypercone

Double cones.

hypercone whose base

is

the

interior of a cone

be regarded in two ways as a hypercone of this kind, the vertex of the base in one case being

may

the vertex of the hypercone in the other case. Thus there are two cones having themselves a

common

base, and we can say that the hypercone is determined by a closed plane curve and two points neither of which is in the hyperplane containing the curve and the other point. Looked at in this way the hypercone is called a double
cone.

33-351

HYPERSURFACE WITH PLANE ELEMENTS

71

double cone consists, then, of the following classes of

:

points (1) the points of a closed plane curve and the points of
interior
(2)
;

its

two points not in a hyperplane with the curve, the interior of the segment formed of these two points, and the interiors of the segments formed by taking each of these
points with the points of the curve (3) the interiors of the triangles formed
;

by taking each
;

point of the curve with the two given points (4) the interiors of two cones each formed

by taking the

curve with one of the two given points.

The interior of the segment of the two given points is the vertex-edge of the double cone, and the interior of the curve is the base. The interiors of the triangles (3) are the
elements
j

and the two cones

(4) are

the end-cones.

35. Piano-conical hypersurfaces. piano-conical hypersurface consists of the planes determined by the points of a plane curve and a line not in a hyperplane with this curve.

The
curve,

line is the vertex-edge,

the curve

is

the directing-

and the planes are the

elements.

Each element meets

the plane of the directing-curve in only one point, the point where it meets the directing-curve itself. There are two

nappes to the hypersurface. The only case which we have to consider at present the case where the directing-curve is a circle.

is

THEOREM.

hyperplane which contains the directing-

curve of a piano-conical hypersurface and a point of the vertexintersects the hypersurface in a conical surface.
y

The line containing the vertex-edge of a double cone, and the curve whose interior is the base, are the vertex-edge
and directing-curve
of a piano-conical hypersurface.

72

HYPERPYRAMIDS AND HYPERCONES

[MX.

double cone, or the hypersolid which we

call

the interior of a

double cone, can be somewhat inaccurately described as cut from one nappe of a piano-conical hype rsurface by two hyperplanes each of which contains the directing-curve and a point of the vertex-edge.

There are theorems on certain cases of "intersection of double cones corresponding to the theorems of Art. 32.

CHAPTER

II

PERPENDICULARITY AND SIMPLE ANGLES

36.

Introductory.

From

this

point

the

theorems

of

ordinary geometry presupposed. We shall now take up perpendicularity and various kinds of angles in very much
the

same way that these subjects are taken up

in the

text-books.
line

We shall find the

relation of the perpendicular

and hyperplane analogous to the relation of the perpendicular line and plane of three-dimensional geometry, but a new kind of perpendicularity in the case of two planes
absolutely perpendicular in hyperspace, every line of one

plane through their point of intersection being perpendicThe ular to every line of the other through this point.
relation of perpendicular planes in a hyperplane now takes a very simple form, as that of a plane intersecting in a line each of two absolutely perpendicular planes. This chapter
will

conclude with a treatment of perpendicular planes

of

and hyperplanes, and

hyperplane angles, which are

analogous to dihedral angles. Chap. Ill will take up what

sional angles,

may

be called two-dimen-

and angles which correspond

After finishing

angles;

and chap. IV symmetry, and motion.

to polyhedral will consider questions of order,

these

chapters

we

shall return to the polyhedroids

and other hypersur-

faces already introduced.

at the beginning of the preceding section (Art. 29), the three chapters which follow do not depend on that section. Indeed, only a few simple facts are needed

As indicated

from the entire

first

chapter besides what are already

73

74

LINES PERPENDICULAR TO A HYPERPLANE

[n.

x.

familiar to the student.

These

facts are easily stated

and

understood, and a study

of their application to the

theorems

which follow

will help us to realize their truth.

Except for the properties of points on a line referred to in Art. 5, the development of most of chap. I has been complete in itself, no theorem being based on theorems not given there. The subjects now to be taken up belong

more particularly to metrical geometry; and we shall assume the axioms of metrical geometry,* and employ
its

assume

terms without special definition. In fact we shall all of the theorems of the ordinary geometry, except,

for the present, those

parallels.!
I.

which depend on the axiom of

LINES PERPENDICULAR TO A HYPERPLANE

Existence of perpendicular lines and hyperplanes. i. The lines perpendicular to a line at a given
lie

37.

THEOREM
PROOF.

point do not all

in one plane.

Every point in hyperspace lies in a plane with the given line, and in every plane which contains the line
if

is a perpendicular to the line at the given point. Now these perpendiculars were all in one plane, that plane and the given line would determine a hyperplane con-

there

taining

all of

the planes which contain the line (Art. 20,

Th.
of

and so all points of hyperspace. But the points hyperspace do not all lie in one hyperplane.
i),
2.

THEOREM

line perpendicular at

a point

to

each of

three non-coplanar lines, is perpendicular to every line through

this point in the

hyperplane which the three lines determine.

* For a set of axioms of metrical geometry see R. L Moore, "Sets of Metrical " Hypotheses," referred to on p. 28, or Veblen, Foundations of Geometry," pp. 27, ff The theorems of are true in any plane and of any plane figures, f plane geometry and the theorems of geometry of three dimensions are true in any hyperplane and
.

of

any hyperplane

figures, for all the

axioms are assumed to be true

in every portion

of hyperspace.

36-38]

LINES PERPENDICULAR TO A LINE

75

PROOF. Let m be a line perpendicular at a point O to each of three non-coplanar lines, a, &, and c, and let d be any other line through O in the hyperplane determined by
these three lines.

The plane
6, is

of ab in a line h (Art. 22, Th.).

of cd will intersect the plane The line w, being perc

pendicular to a and

perpendicular to h lying in the

plane of aft;
it is

and then, being perpendicular to perpendicular to d lying in the plane of ch*

3.

and

A,

THEOREM

a given point lie PROOF. Let m be the given line and O the given point. Three non-coplanar lines perpendicular to m at O determine a hyperplane a such that m is perpendicular to every line of a through O. Now let a be any line perpendicular to m at O. The plane of am intersects a in a line b (Art. 27, Th. i), also perpendicular to m at O. In this plane, then, we have a line m and the two lines a and b perpendicular to m at O. Therefore a must coincide with b and lie in the
hyperplane a.
is perhyperplane at a point pendicular to the hyperplane when it is perpendicular to the hyperall lines of the hyperplane which pass through O The point is be to the line. also said to plane perpendicular

All lines perpendicular in a single hyperplane.

to

a given

line at

line intersecting a

is

called the foot of the perpendicular.

38.

One hyperplane through any

line.

point perpendicular to

a given

THEOREM

Planes in a perpendicular hyperplane. i. At any point of a line there is one and only
to the line.

one hyperplane perpendicular

This follows immediately from the theorems of the preceding

*
O,
is

article.

A line through a point O, perpendicular to each of two lines intersecting at perpendicular to every line through O in the plane which the two lines determine. This is always true, for the plane and the line lie in one hyperplane.

76

LINES PERPENDICULAR TO A HYPERPLANE

[n.

I.

THEOREM THEOREM

2.

Through any point outside of a

line passes

one and only one hyperplane perpendicular

3.
to

to the line.

line perpendicular to
the

a hyperplane

is

hyperplane passing every plane of through the foot of the line; and every plane perpendicular to a line at a point lies in the hyperplane perpendicular to
perpendicular
the line at this point.

THEOREM
to

4.

// a line and plane

intersect,

a line per-

both at their point of intersection is perpendicular pendicular to the hyperplane determined by them; or if two planes intersect

in a line, a line perpendicular

to

both at

any point

of

their intersection is perpendicular to the

hyperplane

deter-

mined by them.
39.

THEOREM

Lines perpendicular to a given hyperplane. i. At a given point in a hyperplane there

is

one and only one line perpendicular to the hyperplane. PROOF. Consider three non-coplanar lines lying in the hyperplane and passing through the given point. The
three hyperplanes perpendicular to these lines at the point

have in common at least a line (Art. 27, Th. 2 and remark), and any such line must be perpendicular to the given
hyperplane because perpendicular to the three lines. If there were two lines perpendicular to the hyperplane
at the given point, they would both be perpendicular to the line in which their plane intersects the hyperplane.

We

should have in a plane two lines perpendicular to a third at the same point, which is impossible.

THEOREM
lie

2.

Two

lines perpendicular to

a hyperplane
20,

in a plane.

PROOF.
2

Any two

lines lie in

a hyperplane (Art.

Th.
lines

(2)),

and a hyperplane containing the two given

intersects the given hyperplane in a plane to

which the

38, 39)

ONLY ONE THROUGH A POINT

77

both perpendicular (Art. 38, Th. 3). Therefore, as proved in geometry of three dimensions, the two lines
lines are
lie

in a plane.*

THEOREM

3.

Through any point outside of a hyperplane

to the

passes one and only one line perpendicular

,P

hyperplane.

-\

be the point, and a some line perpendicular to the hyperplane. If a does not pass through P,

PROOF.

Let

a and

P determine
c,

in a line

and
c,

in this plane there

a plane which intersects the hyperplane is a line b through P perf

Let b be the intersecting c at a point O. line perpendicular to the hyperplane at O. a and 6' lie
pendicular to
the plane containing a and the point 0. But the plane determined by a and P passes through O. Therefore b' lies in the plane determined by
in a plane (Th. 2),

which

is

a and P, and in this plane is perpendicular to c at 0. In other words, b' coincides with 6, and b must be perpendicular to the hyperplane. If there were two lines through

perpendicular to the

hyperplane,

we should have two

pendicular to the line this is impossible.

* Euclid proves

lines through P perwhich passes through their feet, and

this in connection

ular to a plane are parallel (Elements, Bk. XI, Prop. 6). ever, depend on the axiom of parallels.

with the theorem that two lines perpendicThe proof does not, how-

78
40.

LINES PERPENDICULAR TO A HYPERPLANE

[XL

i.

Projection of a point upon a hyperplane. Distance from a hyperplane. The projection* of a point upon a

hyperplane is the foot of the perpendicular from the point to the hyperplane. The perpendicular may be called the
projecting line.

The distance from any point outside of a upon the hyperplane is less than hyperplane the distance from the point to any other point of the hyperplane. The distance between a hyperplane and a point outside of the hyperplane is the distance between the point and its
i.f
to its projection

THEOREM

projection

upon the hyperplane.

COROLLARY.J // the distance between two points is less than the distance of one of them from a hyperplane, they lie on the same side of the hyperplane in hyperspace.
2. Given any point P outside of a hyperplane, O its projection upon the hyperplane, then any two points of the hyperplane equally distant from P will be equally distant from O, and any two points equally distant from O will

THEOREM

and

be equally distant from P; and if two points of the hyperplane or O, that point which is are unequally distant from either

nearer

to

one of them will be nearer

to the other.

PROOF.

The perpendicular PO, and

the lines through

and any two points of the hyperplane, lie in a second hyperplane which intersects the given hyperplane in a plane a. The perpendicular is perpendicular to a, and the three lines intersect a in the points where they intersect the given hyperplane. The theorem is therefore a theorem in the
three-dimensional geometry of the second hyperplane.
* That
lines
is,

orthogonal projection.
(see, for

We

sometimes have cases of projection by

through a point

example, the proof in Art. 25), but

when we use

the

word projection alone we shall mean by it orthogonal projection. t The two theorems of this article are true at least when the distances referred to are "restricted" (see Introduction, p. 19, and reference given there in the footnote).
I

Used

in the proof of the first

theorem of Art. 96*

40,

PROJECTION OF A LINE
From any
point of our three-dimensional space

79

we can move

off

into hyperspace without passing through or approaching any other point at the centre of a sphere, for example, point of our space.

moving any point

off

on a

line perpendicular to

our space, would not approach

It would, indeed, be moving farther and of the sphere. farther from all points of the sphere, the distances from all of these points being the same, and all increasing at the same rate. An object

completely enclosed within a surface in our three-dimensional space would be entirely free if it could pass off in any direction out of this space.

41.

Projection of a line upon a hyperplane.

Angle of a

and hyperplane. The projection of any figure a upon hyperplane consists of the projections of its points.
half-line

a line and hyperplane are not perpendicular, the projection of the line upon the hyperplane is a line or a part of a line.
i
.

THEOREM

When

and a the perpendicular projecting some point of m upon the hyperplane. Any other perpendicular b projecting a point of m upon the hyperplane, lies in a plane with a (Art. 39, Th. 2), and this plane, containing two points of m, is the plane determined by m and a. Therefore, all the perpendiculars projecting
PROOF.

Let

be the

line,

points of

by
is

m upon the hyperplane lie in the plane determined and a, and the projection of m upon the hyperplane the same as its projection upon the line in which this
m

plane intersects the hyperplane.*

* In plane geometry, when two lines are not perpendicular, the projection of one upon the other is the latter line itself or a single connected part of it. That is, if two given points are the projections on a line a of points of another line m, any point between them will be the projection of a point of m. For the perpendiculars at the three points of a do not intersect one

another, and therefore it follows that the perpendicular which lies between the other
intersect m (Art. 12, Th. i) any point of a between the projections of two points of m is itself the projection of a point of m.

two must
is,

that

80

ABSOLUTELY PERPENDICULAR PLANES

COROLLARY.

[u. n.

When a half-line drawn from a point O of a does not lie in the hyperplane and is not perpenhyperplane dicular to it, its projection upon the hyperplane is a halfline

drawn in

the

ment which has

O for

hyperplane from O, or the interior of a segone of its points.

half-line

THEOREM
pendicular

2.

When a

a hyperplane does not

to it,

drawn from a point O of in the hyperplane and is not perthe angle which it makes with the half-line
lie

which

drawn from O containing its projection is less than the angle it makes with any other half-line drawn in the hyper-

plane from O.

When
does not

lie

a half-line drawn from a point O of a hyperplane in the hyperplane and is not perpendicular to it,
it

the angle which

angle of the half-line

makes with its projection is called the and hyperplane. A half-line drawn
right angle

from a point
plane
is

of a hyperplane perpendicular to the hyper-

said to

make a

with the hyperplane.

II.

ABSOLUTELY PERPENDICULAR PLANES

42.

Existence of absolutely perpendicular planes. THEOREM i. A plane has more than one line perpendicu-

it at a given point. For the plane is the intersection of different hyperplanes, and the lines in these hyperplanes perpendicular to the

lar to

plane at the given point must be different lines (see Art. 20).

THEOREM

2.

Two

lines perpendicular to

a plane

at a

given point determine a second plane,

and

the two planes are

so related that every line of one through the point is perpendicular to every line of the other through the

point*

prove without the axiom of parallels that every point of a is the projection of a point of m, for the perpendiculars at some points of a may not intersect m at all. The projection of m may be the entire line a, but it may be only a halfline of a, or the interior of

We cannot

a definite segment of this

line,

* See foot-note, p 75.

41-43)

LINES PERPENDICULAR TO A PLANE

3.

8l

THEOREM
PROOF.

a given point

All the lines perpendicular to a plane at in a single plane. Let a be the given plane, and the given point.
lie

Two

lines perpendicular to

at

determine a plane

in

is a line perpendicular to a. which every line through Now let a be any line perpendicular to a at O. The hyperplane determined by a and a intersects /3 in a line b (Art. 27,

Th.

i), also

perpendicular to

containing Therefore a coincides with b and

a.

only one

line

a at O. But in a hyperplane can be perpendicular to a at O.

lies in
(3.

Two
is

planes having a point in

common

are absolutely

perpendicular

when every

line of

one through that point

perpendicular to every line of the other through that

point.

These planes have only the point in common, and do not intersect We can never see both planes in a single hyperplane like the space in which we live. The most that we could see would be one plane and a single line of the other.
in a line.

One plane through any point absolutely perpendicua given plane. THEOREM i. At any point of a plane there is one and only one plane absolutely perpendicular to the given plane.
43.
lar to

This follows immediately from the theorems of the preceding

article.

plane and a point outside of the plane lie in one and only one hyperplane. Through any such point, therefore,
line perpendicular to the

by three-dimensional geometry, passes one and only one and the projection of the plane
;

point upon geometry of three dimensions, is the foot of this perpendicular. A point that lies in a given plane is its own projection upon the plane.

the plane, as in

82

ABSOLUTELY PERPENDICULAR PLANES

[n.

u.

THEOREM
plane.

2.

Through any point outside of a plane passes

to the

one and only one plane absolutely perpendicular

given

PROOF.
plane.

P be the point, and O its projection upon the The absolutely perpendicular plane at O will then
Let

contain the perpendicular line

over,

OP, and

therefore P.

More-

any absolutely perpendicular plane containing must contain a perpendicular line through P, and there
only one such
44.
line.

P
is

Planes absolutely perpendicular to planes which

i

intersect in a line.

THEOREM
in

// two planes intersect in a line and so

lie

hyperplane, their absolutely perpendicular planes at any point of their intersection intersect in a line and lie in

a. and /3 be two planes intersecting in a line let and and a /?' be the planes absolutely perpendicular a, to a. and j3 respectively at a point O of a. a and 0' are, then, both perpendicular to the line a at 0, and lie in a
f

a hyperplane. PROOF. Let

hyperplane perpendicular to a at
fore

(Art. 38,

Th.

3).

There-

a and

]8'

intersect in a line (Art. 22, Th.).

THEOREM
line lie in

2.

// three planes have a line in

common

their absolutely perpendicular planes at

any point of

this

a hyperplane; and if three planes lie in a hyperand a point in common, their absolutely perpendicuhave plane lar planes at this point have a line in common. This is proved by Art. 38, Th. 3, the line being perpendicular to the hyperplane.

COROLLARY.
lie

// three planes have a line in

common and

any point of this

plane.

in a hyperplane, their absolutely perpendicular planes at line have a line in common and lie in a hyper-

43~4S]

TWO PLANES

IN A HYPERPLANE

83
differ-

45.

Planes absolutely perpendicular to a plane at

ent points.

THEOREM.

Two

planes

absolutely

perpendicular

to

third lie in a hyperplane.

PROOF.
line

Let a and

]8

lar to a given

OO', and

and
lie

plane 7 let a and b be the lines perpendicular at O a and b to the hyperplane determined by a and c.

at

be the planes absolutely perpendicutwo points and 0'. Let c be the

in a plane (Art. 39, Th. 2), and this plane is 7, for a, being perpendicular to a, must lie in 7, and only one plane

can contain a and the point O'. b is then perpendicular to so that lies a in the with |8 hyperplane to which b is per;

pendicular (Art. 38, Th. 3), the hyperplane determined by

a and

c.

All the planes absolutely perpendicular a plane at the points of any line of it, lie in a hyperplane.

COROLLARY.

to

It should be noted that the figure given here is merely to serve as a diagram, and does not represent the actual appearance of two planes absolutely perpendicular to a third. The most that we could ever see in any one hyperplane would be the plane 7 and a line in each of

the other two planes, or the two planes a and

of 7.

and the

single line c

84
46.

ABSOLUTELY PERPENDICULAR PLANES

[n.

n.

Projection of a line upon a plane. As in other cases of projection, the projection of any figure upon a plane consists of the projections of its points (Art. 41).

THEOREM

i.

The perpendiculars projecting

the

points

of a line upon a plane do not lie in a single plane, unless the line itself lies in a hyperplane with the plane upon which it
is projected.

PROOF. If any two of the perpendiculars were in a plane, that plane, having two points in common with the given plane, would intersect the latter in a line and lie with it
in
line
is

a hyperplane (Art. 20, Th. 2 (4)). Therefore, the given would lie in a hyperplane with the plane upon which it

projected.

When two
0,
all

planes are absolutely perpendicular at a point the points of one project upon the other in the single

point O.

We may

regard projection upon a plane as

it,

made

by

planes absolutely perpendicular to

just as in geometry

of three dimensions

regard projection upon a line as made by planes perpendicular to the line. We shall sometimes speak of projecting planes and think of a point
as projected

we may

upon a plane
2.

in this

way.*
line

THEOREM
line or

The projection of a

upon a plane

is

a part of a

line, unless the given line lies

in a plane

absolutely perpendicular to the given plane.

PROOF.

Let

be the given
of

line

and y the given plane.

Let a be the plane absolutely perpendicular to 7 which

projects

some point
Th.).

jecting a point of
(Art.

45,

upon 7. Any other plane prom upon 7 lies in a hyperplane with a But this hyperplane containing two

* In a hyperplane, on the other hand, where the perpendiculars projecting points of a line upon a plane always lie in a single plane, the latter is sometimes called the These two uses of the expression projecting plane, projecting the line as a whole. "projecting plane" should be carefully distinguished.

46,47]

PERPENDICULARITY AS INTERSECTION

85

and a; points of m, is the hyperplane determined by same as its projection and the projection of upon 7 is the upon the line in which this hyperplane intersects 7.*

The
lines.

given line and

projecting lines form a curved surface which contains the its projection and lies in the hyperplane of these two

The student may prove the following theorem Given any two lines, we can pass through either a plane upon
:

which

it

will

be the projection

of the other.

III.

SIMPLY PERPENDICULAR PLANES

47.

Planes intersecting in lines two absolutely per-

pendicular planes.
perpendicular,
line in the

Two planes are perpendicular, or simply

lie

when they

in one hyperplane

and

in this

hyperplane form right dihedral angles.

Each contains a
one of two absothe point

other and lines perpendicular to the other, f

i.

THEOREM

A
is

plane perpendicular

to

lutely perpendicular planes,

and passing through

to the other.

where they meet, PROOF. Let

perpendicular a' be two absolutely perpendicular planes meeting in a point O, and let /3 be a plane passing and perpendicular to a. Two planes which are through

a and

perpendicular

geometry

in a hyperplane, and, of three dimensions, a line in

lie

by a theorem

of

to their intersection

is

perpendicular to the other.

one perpendicular There

* See foot-note, p. 79. t In a sense two planes are perpendicular whenever one of them contains a line perpendicular to the other. Such planes have also been called half-perpendicular,
in distinction
pletely

from absolutely perpendicular planes, which are then said to be com-

Two

ehrdimensionale Geometric, vol. I, p. 49). (see Schoute, planes in a hyperplane forming a right dihedral angle might be described as We shall find that such planes are of great imporPerpendicular in a hyperplane. tance in our geometry, and we shall, at least for the present, apply to them the word
perpendicular

perpendicular.

nary geometry

Perpendicular, then, has with us the same meaning as in the ordiand the only new concept with which we have to become famil-

iar at this point is that of absolute perpendicularity.

86
is,

SIMPLY PERPENDICULAR PLANES

then, a line in
/?

[n. ra.

perpendicular to

at O, that

is,

a line

common to ft and the absolutely perpendicular plane a', so that ft intersects & in a line and lies with a' in a hyperNow the line in which ft intersects a, like all the plane.
lines of

a through

perplane of ft Therefore to a'.

0, is perpendicular to a! ; so in the hyand a! we have a line lying in ft perpendicular

ft

is

perpendicular to a'.

THEOREM
PROOF.

2.

plane intersecting in a line each of two

absolutely perpendicular planes, is perpendicular to both.

Let a and a

planes meeting both of these planes in

ft

in a point O,
lines.

be two absolutely perpendicular and let ft be a plane intersecting

These two

lines

and the plane

pass through 0;
ft

otherwise the hyperplane determined

by

and

sible.

which

O would contain both a and a:', which is imposNow the line in which intersects a! is that line is perpendicular to a at O in the hyperplane deterft

line lying in

mined by a and ft. In this hyperplane, then, we have a Therefore is perft and perpendicular to a.
i,

pendicular to a. In the same way, or by Th.

we prove

that

ft

is

perpendicular

to a'.

THEOREM

3.

// two planes are perpendicular, their ab-

solutely perpendicular planes at

any point of

their intersec-

tion are perpendicular.

PROOF.
let

Let a and
ft'

ft

be two perpendicular planes, and

a, being perpendicular
j3',

a and

be

their absolutely perpendicular planes at

a point
to
/?,

of

their intersection,
;

is

perpendicular to 0'

and then
.

being perpendicu-

lar to a, is perpendicular to a'

The student may prove

ular planes, both.

If a plane is perpendicular to

one of two absolutely perpendicit

and contains a point

of the other,

is

perpendicular to

47-491

PLANES INTERSECTING IN A LINE

87

48.

The

different possible positions of

solutely perpendicular planes at a point.

of absolutely perpendicular planes Then in common. point
(1)

two pairs of abLet two pairs *

their intersection

have

in common they may have only the point (2) each plane of one pair may intersect in a line one plane of the other pair or (3) each plane of one pair may be perpendicular to both
;

planes of the other pair. In the last case the four lines of intersection are mutually perpendicular, and, taken two at a time, determine also a

The planes third pair of absolutely perpendicular planes. of each of the three pairs are then perpendicular to all the
planes of the other two pairs. We have, in fact, four mutually perpendicular lines, any two of them determining a plane, and any three a hyperplane.
of

Each

line is perpendicular to the

hyperplane de-

termined by the other three. In each hyperplane are three the six planes, three mutually perpendicular planes

forming a trirectangular trihedral angle. We shall call this figure a rectangular system (see Art. 71).
49.

Common

perpendicular planes of two planes inter-

secting in a line.

THEOREM.
any point
PROOF.
a! and
/?'

planes which intersect in a line have at of that line one and only one pair of common

Two

perpendicular planes.

Let a and

|8

be the two given planes, and

let

plane which

be their absolutely perpendicular planes at O. is a plane is perpendicular to a and /S at

which
we

intersects these four planes in lines.

The

planes

* In general, we shall use the word "pair" in speaking of two planes only when have in mind two planes absolutely perpendicular to each other. Thus we shall

speak in this

way of a

pair of planes perpendicular to a given plane,

and of a pair of

common

perpendicular planes

when we have two given

planes.

88
a.

SIMPLY PKRPKNDICULAR PLANKS

ft

IH.

m.

and

intersect in a line

and

lie

in a hyperplane,

and the
ft' is

planes a'
(Art. 44,

and
Th.

&
i).

intersect in a line

and

lie

in a hyperplane

The

line of intersection of a!

and

perpendicular to the hyperplane of a and /3, and the line of intersection of a and ft is perpendicular to the hyperplane
of a' and/3' (Art. 38, Th. 4).

In a hyperplane two intersecting planes have a

common

any point perpendicular plane perpendicular to the intersection. Thus

perpendicular to

at

of

their intersection,

we have one plane

lying in their hyperplane. This plane cannot contain the intersection of a' and #',
ft

a and

at

for the intersection of

perplane of
a' and
ft'

a and

in

a and ft' is perpendicular to the hyThe plane must therefore intersect separate lines, and lie also in their hyperplane.
ft.

common perpendicular plane with which are familiar in the case of two intersecting planes, perpendicular to their intersection at a point 0, is the plane
In other words, the

we

of intersection of their hyperplane with the hyperplane of their absolutely perpendicular planes at O.

perpendicular plane is the plane absolutely perpendicular to the plane already found (Art. 47, Th. i). This plane contains the lines of intersection of a

second

common

and

ft

and

of a!

and

ft'

and

may

be regarded as determined

by them.
It remains to

prove that these two planes are the only

0.
lines

planes perpendicular to

a and ft at Any plane intersecting a and ft in

intersecting

must pass through

in lines

their line of intersection or lie entirely in their hyperplane,

and any plane

through their

a and
But, as

ft'

must pass

line of intersection or lie entirely in their

i).

hyperplane (Art. 20, Th.

we have

seen, a plane

two of these planes can not pass through the line of intersection of the other two.
lying entirely in the hyperplane of

Any common

perpendicular plane, therefore, must

lie in

4Q]

PLANES INTERSECTING IN A LINE

89

tion

both hyperplanes or pass through both lines of intersecthat is, it must be one of the two planes already found.
:

These are therefore the only planes which can be perpendicular to a and /3 at O.

When

planes have

the two given planes are perpendicular, the four all together four lines of intersection, and there

are four different hyperplanes which contain two of the four planes. The planes can be associated in two ways so

that two planes shall intersect in a line and the other two planes be their absolutely perpendicular planes. But the

plane determined by the two lines of intersection in one case is the plane of intersection of the two hyperplanes in the other case. Therefore we get only two planes perpendicular to the

two given planes.

That

is,

if

a and

a and /3 and the intersection of a' and /3' is the plane of intersection of the hyperplane determined by a and ft' and the hyperplane^determined by a' and j8 and the plane determined by the intersection of a and /3' and the intersection of a! and & is the plane of intersection of the hyperplane determined by a and ft and the hyperplane determined by a' and /3'. Thus we have only two planes intersecting in lines the four planes, a, j8, OL /3', and so perpendicular to the two given planes a
;

perpendicular to /8' by the intersection of

are perpendicular, so that a is also and /3 to a', then the plane determined
/3

',

and/3.
four planes and the two common perpendicular planes in this case are the six planes of a rectangular sys-

The

tem

(Art. 48).

// each of two planes having a common point a line the plane absolutely perpendicular to the other at O, then these planes have one and only one pair

COROLLARY.

intersects in

of

common

perpendicular planes.

pO

PERPENDICULAR PLANES AND HYPERPLANES

The student may prove the
:

[n. iv.

following theorem Given two planes with a point O in common, through any line containing O in one of these planes can be passed a plane perpendicular In other words, to this plane and intersecting the other in a line. is the projection upon this plane of a any line of one plane through
line of the other.
If the

of intersection,

two planes intersect in a an infinite number

line

of planes
if

the conditions of the theorem.

Or,

and the given line is this line can be drawn satisfying one of the two planes lies in a

hyperplane with the plane absolutely perpendicular to the other at O, and the given line is the intersection of the other given plane with
this hyperplane, there will be

Except

in these

an infinite number of these planes. two cases there is only one plane satisfying the

conditions of the theorem.

IV.

PERPENDICULAR PLANES AND HYPERPLANES

Perpendicular planes and hyperplanes.

50.

The planes

perpendicular or absolutely perpendicular to planes lying A plane intersecting a hyperplane is in the hyperplanes.
perpendicular to the hyperplane at a point of their intersection, if the plane absolutely perpendicular to the given plane The hyperplane is at this point lies in the hyperplane.
also said to be perpendicular to the plane.

THEOREM
at
this line.

i.

If a plane

is

perpendicular
it is

to

a hyperplane
all

one point of their intersection,

perpendicular

along

For the

line of intersection

and the plane absolutely per-

pendicular to the given plane at the given point determine the hyperplane (Art. 20, Th. 2 (i) ), which, therefore, contains the planes absolutely perpendicular to the given

plane at

all

points of the line of intersection (Art. 45, Th.

and

Cor.).
2.

THEOREM

If a plane
c,

is

perpendicular
the

to

a hyper-

plane along a line

any plane in

hyperplane perpendic-

ular to c is absolutely perpendicular to a,

and any plane ab-

49,5o]

PLANES IN THE HYPERPLANES

91

solutely perpendicular to
lies entirely

PROOF.

a, through a point of the hyperplane in the hyperplane. The planes absolutely perpendicular to ot at

the points of c lie in the hyperplane, and in the hyperplane they are planes perpendicular to the line c. Now in the
line at
fore,

hyperplane only one plane can be perpendicular to a given a given point. Any plane in the hyperplane, thereperpendicular to the line
c,

must be one

of the planes

absolutely perpendicular to a. at the points of c. On the other hand, if a plane a! is given as passing through a point of the hyperplane and absolutely perpendicular to
a,

we can draw a plane /3 in the hyperplane through the same point perpendicular to c. )3 is then absolutely perpendicular to a, and a must coincide with /3 and lie entirely in the
we cannot have two planes through a point absolutely perpendicular to a given plane (Art. 43,
hyperplane, since
Ths.).

THEOREM
along a line
is

3.
c,

If a plane a is perpendicular to a hyperplane any plane in the hyperplane passing through c

to a,

perpendicular
c,

ing through

or through
itself

and any plane perpendicular to a passany line which lies in the hyperperpendicular
to c, lies entirely

plane and is not

hyperplane.

in the

PROOF.
through
c,

In

intersects
it

hyperplane,

given plane, passing and, lying in the given intersects in a line the plane absolutely per-

the

first

case,

the
;

in this line

pendicular to ular to a.

at

any point
if

of

c.

It

is

therefore perpendic-

On
to

the other hand,

a and passing through

the plane is given as perpendicular c, or through any line b which lies

must
to

in the hyperplane and is not itself perpendicular to c, it intersect in a line the plane absolutely perpendicular

a through any

point of the given

line,

a plane which

lies

92

PERPENDICULAR PLANES AND 1IYPERPLANES

It

[11.

iv.

entirely in the hyperplane (Th. 2). contain two lines of the hyperplane
(Art. 20,

must, therefore,
lie

and

entirely in

it

Th.

i).

THEOREM 4. // two hyper planes are perpendicular to a plane at a point O, they intersect in the absolutely perpendicular plane at O.

For the absolutely perpendicular plane

perplanes,
51.

lies in

both hy-

by hypothesis.

Lines lying in the plane and perpendicular to the

hyperplane, or in the hyperplane and perpendicular to the

plane.

THEOREM
any
to

i.

If a plane

is

perpendicular

to

a hyperplane,

line in the plane perpendicular to their intersection is

to the

perpendicular
the plane.

hyperplane, and any line perpendicular the hyperplane through a point of the plane lies entirely in

lying in the given plane, is perpendicular not only to the intersection, but also to the absolutely perpendicular plane at the point where
first case,

PROOF.

In the

the

line,

it

meets the intersection.

The

line is therefore

perpen-

dicular to the hyperplane (Art. 38, Th. 4). On the other hand, if the line is given as passing through a point of the plane and perpendicular to the hyperplane,

we can draw a

line in the

plane through the same point

perpendicular to the intersection, and the two lines must coincide, since they are both perpendicular to the hyper-

plane (Art. 39, Ths.

and

3).

COROLLARY. // two planes are perpendicular to a hyperplane at a point 0, they intersect in a line which is perpendicular to the hyperplane at O.

For,

by the theorem, the

line perpendicular to the hy-

perplane at

lies in

both planes.

50-52]

ONLY ONE THROUGH A LINE

2.

93

THEOREM
any
is

If a plane is perpendicular

to

a hyperplane,

line in the hyperplane perpendicular to their intersection

to the plane, and any line perpendicular a point of the hyper plane lies entirely through plane

perpendicular

to the

in the hyperplane.

PROOF.

In the

first case,

the line, lying in the hyper-

plane and perpendicular to the intersection, lies also in the plane which in the hyperplane is perpendicular to the But this plane is absolutely intersection at the same point.

perpendicular to the given plane (Art. 50, Th. 2). Therefore the given line is perpendicular to the given
plane.

On the other hand, if the line is given as passing through a point of the hyperplane and perpendicular to the given plane, it lies in the absolutely perpendicular plane which passes through the same point, and therefore in the hyperplane (same reference).
52.

Through any

line

one plane perpendicular

to

a given

hyperplane, or one hyperplane perpendicular to a given

plane.

THEOREM
any

i.

If a line

is

perpendicular

to

a hyperplane,
the

plane passing through the line is perpendicular to

hyperplane. For the plane absolutely perpendicular to the given plane at the point where the line meets the hyperplane is per-

pendicular to the
(Art. 38, Th.
3).
2.

line,

and therefore

lies in

the hyperplane

Through a line not perpendicular to a hyperplane passes one and only one plane perpendicular to the
hyperplane.

THEOREM

PROOF. The plane perpendicular to the hyperplane along the projection of the given line upon the hyperplane, and we cannot contains the given line (Art. 51, Th. i)
;

94

PERPENDICULAR PLAJNEb AMD

HYFfcKFLAIMfc&

in. IV.

have two perpendicular planes containing the given line, for that would make the line itself perpendicular to the hyperplane, which
is

contrary to hypothesis (same reference,

and

Cor.).
3.

THEOREM

Through a
to

lutely perpendicular one hyperplane perpendicular

line not lying in a plane absoa given plane, passes one and only
to the

given plane.

PROOF.

The hyperplane perpendicular

;

to the given plane

along the projection of the given line upon the plane, conand we cannot have tains the given line (Art. 51, Th. 2)

two perpendicular hyperplanes containing the given line, for that would put the line itself into a plane absolutely perpendicular to the given plane, which is contrary to hypothesis (Art. 50, Th. 4).
53.

Planes with linear elements

all

perpendicular to a

hyperplane.

Given two planes not in one hyperplane, if any two of their linear elements (Art. 27) have a common perpendicular line, they all have a common perpendicular
hyperplane,
dicular.
to

THEOREM.

which

the two given planes are also perpen-

PROOF.

Let

a.

and

ft

a be one of the linear elements in

elements in

be the two given planes, and let a. and b one of the linear

If a and b have a common perpendicular /?. the line, hyperplane perpendicular to either of these lines at the point where it meets the perpendicular line, is per-

pendicular to the plane of the two (Art. 52, Th. i), and therefore perpendicular to the other (Art. 51, Th. i).

This hyperplane
]8,

is

also perpendicular to the planes

a and

as well as to every plane containing a or b (Art. 52, Th. i). Now any other linear element in j8 is the intersection of
|8

with a plane through

is

element in a

the intersection of

a, and any other linear a with a plane through b.

52-54]

HYPERPLANE ANGLES

95

The hyperplane

is therefore perpendicular to all of these elements (Art. 51, Th. i, Cor.). If two elements in one of the two given planes are given as having a common perpendicular line, then the perpen-

dicular hyperplane is perpendicular to the plane in which they lie and to every plane containing either one of them.

But any element of the other plane is the intersection of two planes through the two given elements, and is thereThus fore, as before, perpendicular to the hyperplane.
the hyperplane
is

and

so perpendicular to all the elements

perpendicular to elements in both planes, and to the two

given planes.*
V.

HYPERPLANE ANGLES
Interior.

54.

plane angle consists of

mon

A hyperPlane angles. two half-hyperplanes having a comface but not themselves parts of the same hyperDefinition.

plane, together with the common face. The common face is the face of the hyperplane angle, the two half-hyperplanes are the cells.

and

The
of all

interior of a hyperplane angle consists of the interiors

segments whose points are points of the hyperplane

angle, except those segments

;

whose

interiors also lie in

the hyperplane angle that is, the interior of a hyperplane consists of the of those segments which have interiors angle

a point in each
rest of
*

cell.

The hyperplane
two

angle divides the

hyperspace into

regions, interior

and

exterior

point

The linear elements of two planes not in a hyperplane will all pass through a common to the two planes, or will all be perpendicular to a hyperplane per-

where there are

pendicular to the two planes, except in the Hyperbolic Non-Euclidean Geometry, lines in a plane which do not intersect and do not have a common

It follows from the perpendicular, namely, the parallel lines of this geometry theorem that in this geometry the linear elements of two planes are all parallel if two of them arc parallel. They will then be axes of a set of boundary-hypersurfaces,

and

their

geometry

will

correspond to the Euclidean three-dimensional geometry

of boundary-hypersurfaces (see the author's

Non-Euclidean Geometry,

p. 52).

gO

HYPERPLANE ANGLES

In. v.

lies

to the hyperplane angle. Each cell of a hyperplane angle on one side of the hyperplane of the other cell and the
;

portion of hyperspace which lies on that side of the hyperplane of each cell oh which lies the other cell, lies between
the two cells and constitutes the interior of the hyper-

plane angle.

drawn from a point O in the face of a hyperplane angle, one in each cell, and each perpendicu-

Two

half-lines

lar to the face, are the sides of

an angle which

is

called the

plane angle at

of the hyperplane angle.

THEOREM
point

i.

The plane absolutely perpendicular

at

to the

face of a hyperplane angle, intersects the hyper-

plane angle in the plane angle at O.

THEOREM

2.

hyperplane perpendicular

to

the face of

a hyperplane angle intersects the hyperplane angle in a dihedral angle which at any point of its edge has the same plane
angle as the hyperplane angle. PROOF. The intersection consists
of

two half^planes

with a

common

edge lying in the face of the hyperplane

The plane absolutely perpendicular to the face angle. of this edge, lies in the perpendicular hyperat any point in and this hyperplane is perpendicular to the edge plane,
of the dihedral angle
50,

formed by the two half-planes

it

(see Art.

Th.

2).

This plane intersects the dihedral angle in the

in which
is,

same angle

intersects the hyperplane angle

an angle which

therefore, the plane angle at

of both.

THEOREM

3.

Two
is

plane angle of one

hyperplane angles are congruent if a equal to a plane angle of the other.

PROOF. We have given two hyperplane angles with the plane angle at a point of one equal to the plane angle at a point O' of the other. If we make these angles coincide,
the faces of the two hyperplane angles will coincide, each face being absolutely perpendicular to the plane of the

54]

THEIR PLANE ANGLES

97

common plane angle at the vertex of this angle. Now one side of the common plane angle lies in one cell of each
of the hyperplane angles, and the other side lies in the other cell of each. The two cells of one hyperplane angle, therefore, coincide

with the two

cells of

hyperplane angles coincide throughout

(Art. 20, Th. 2 (i)

the other, and the two and are congruent

and Th.

3).

If we make a cell of one of these hyperplane angles coinf cide with a cell of the other, with the points O and O coinciding and with the other cells on the same side of the

hyperplane of the common cell, the hyperplane angles will coincide throughout. For the planes of the given will be absolutely perpendicular to the plane plane angles

common face at the same point and will coincide, and the plane angles themselves will coincide, having one side in common and the other sides in the same half-plane
of the

with respect to the line of the

common

side.

hyperplane angles coincide, every point in the face of one coincides with a point in the face of the other,

When two

and the plane angles at any such common point coincide and are equal.

THEOREM
the

4.

The plane angle of a hyperplane angle

is

same

at all points of the face.

PROOF. The absolutely perpendicular planes at any two points of the face are in a perpendicular hyperplane (Art. 45, Th.), in which the corresponding angles are two plane
angles of the same dihedral angle. plane angles are equal.*
*

Therefore the two

The theorem

of the edge is usually

that the plane angle of a dihedral angle is the same at all points made to depend on the axiom of parallels. It can be proved,

however, without the use of this axiom : Let A and B be two points of the edge, and let be the point them. We can turn the dihedral angle so that the half-line OA
original position of the half -line at upon its own original position.

midway between
will fall

OB

and the
This

half -line

upon the which bisects the plane angle

will

make

the plane angle at

coincide

98
Hereafter

HYPERPLANE ANGLES

[n. v.

we shall speak of the plane angle of a hyperwithout plane angle thinking of it as located at any particular point of the face.

COROLLARY. Two hyper plane angles which are congruent in any position will always coincide as soon as they have a

common

cell

and
this

the other cells lie

on

the

same

side of the

hyperplane of
55.

common

cell.

The hyperplane angle

as

a magnitude.

Supple-

mentary hyper plane angles are those which can be placed so as to have one cell in common while their other cells
are opposite half-hyperplanes. Each of them is then the of the A other. supplement right hyperplane angle is one

congruent to its supplement. The hyperplanes of the cells of a right hyperplane angle are said to be per-

which

is

pendicular.
If
cells

two hyperplane angles have a common face and the of one lie in the interior of the other, or if they have
cell in

one

common

while one

cell of

one

lies in

the interior

one hyperplane angle is a part of the interior of the other. We shall find it convenient to speak of one hyperplane angle as a part of the other and as less in magnitude.
of the other, then the interior of

Let two hyperplane angles be placed so as to have one

common while their other cells and their interiors on opposite sides of the hyperplane of this common cell. Then, if either hyperplane angle is less than the supplement of the other, these two cells and the common face will form a hyperplane angle which we can call the sum of
cell in
lie

the given hyperplane angles.

as a whole with its original position and the dihedral angle as a whole will occupy its original position. Then the plane angle at A will take the former position of

the plane angle at B; and the two plane angles must be equal. This theorem should have been proved in the author's A' on- Euclidean Geometry at some point before Art. 1 1, p. 26, as has been pointed out by Professor J L. Coolidge.

54-56]

THEIR MEASUREMENT
i.

99
angles have

THEOREM

Two supplementary hyper plane

supplementary plane angles; a right hyperplane angle has a right plane angle; the smaller of two unequal hyperplane angles has the smaller plane angle; and the plane angle of
the

sum

angles.

of two hyperplane angles is the sum of their plane The converse theorems are also true.
2.

// we divide the plane angle of a hyperplane number any of equal parts, the lines of division will determine additional cells by means of which the given hyperplane angle is divided into the same number of equal parts;
angle into
in particular, given a hyperplane angle a, plane angle may be divided into a sufficient

THEOREM

any other hypernumber of equal

parts so that one of these parts shall be less than a. Thus we can build up a complete theory of the measure-

ment

of

hyperplane angles, taking any particular one as

the unit of measure.

The student will notice the analogy of the hyperplane angle to the dihedral angle of the ordinary geometry. The theory of measurement is identical in the two cases, and so does not involve any new
conceptions.

We may

consider the class of half-hyperplanes with a

common
which

face as having a geometry, a one-dimensional geometry, in

these half-hyperplanes are the elements (compare Point Geometry," Art. 64, and "Edge Geometry," Art. 78, and see Introduction, p.
15).

"

Disregarding the

common

face,

we may say

in this

geometry

that the hyperplane angle consists of two half-hyperplanes. The hyperplane angle corresponds to the segment, and the magnitude of

the hyperplane angle to the length of the segment.

Hyperplane angles measured by their plane angles. THEOREM. Two hyperplane angles are in the same ratio as their plane angles, and the hyperplane angle may be meas56.

ured by

plane angle. If we divide the plane angle of a hyperplane into some number of equal parts, the half -lines of angle

its

PROOF.

100

HYPERPLANE ANGLES

[n. v.

division, taken with the face of the hyperplane angle, will

determine half-hyperplanes which

plane angle
parts.
(its interior)

will

divide the hyperof equal

into the

same number

We

can therefore prove in the usual

way

that

hyperplane angles are proportional to their plane angles,

first

when the plane angles are commensurable and then method of limits or some equivalent method) when the (by the plane angles are incommensurable.

a right hyperplane angle has a right plane angle Therefore, the measure of the hyper(Art. 55, Th. i). in of a right hyperplane angle is always terms plane angle
the

Now

same as the measure

of its plane angle in terms of

right angle.

57.

The

bisecting half -hyperplane.

i.

THEOREM

The half-hyper plane

angle is the locus of points

bisecting a hyperplane in the interior of the hyperplane

angle equidistant from the hyper planes of the cells. PROOF. Through any point lying in the interior of a

hyperplane angle we can pass a plane absolutely perpendicular to the face of the hyperplane angle (Art. 43, Th. 2).
This plane will intersect the cells of the hyperplane angle in the sides of a plane angle, and the bisecting half -hyperplane in the half-line which bisects the plane angle. Now
this plane is perpendicular to the

hyperplanes of the

cells

of the hyperplane angle (Art. 50, Def.), and the distances of the given point from these hyperplanes are the dis-

tances of the point from the sides of the plane angle (Art. 51, If, then, the point is in the half -hyperplane bii). the secting hyperplane angle, it is in the half-line bisecting

Th.

the plane angle, and these distances are equal or, if these distances are equal, the point is in the half-line bisecting the plane angle, and therefore in the half -hyperplane
;

bisecting

the hyperplane angle.

That

is,

the bisecting

5&-581

PERPENDICULAR HYPKRPLANES
is

IOI

the locus of points in the interior of the hyperplane angle equidistant from the hyperplanes of its

half-hyperplane
cells.

THEOREM 2*

The distance of a point in one

cell

of a

hyperplane angle from the bisecting half-hyperplane, is greater than one-half the distance of the point from the hyperplane of the other
cell.

proved, like the preceding theorem, by passing a plane through the given point absolutely perpendicular to the face of the hyperplane angle.
is

This

Perpendicular hyperplanes. perpendicular to the other.

58.

Lines lying in one and

THEOREM
to the other,

i.

// two hyperplanes are perpendicular, any

line through a point of one perpendicu-

line in one perpendicular to their intersection is perpendicular

and any

lar to the other lies entirely in the first.

PROOF. In the first case, we have a line along which lies one side of a plane angle of each of the four right hyperplane angles whose cells lie in the two hyperplanes, and,
since these plane angles are right angles, the line is perpendicular to that line of the second hyperplane along

That is, the given line is perin to a line the second pendicular hyperplane as well as to the plane in which the two hyperplanes intersect, and
which
lie

their other sides.

therefore

it

is

perpendicular to the second hyperplane

(Art. 38, Th.

4).

The proof

of the second part of the

theorem

is

like the

corresponding proofs given in Art. 51.

THEOREM

2.

If a line

is

perpendicular

to

a hyperplane,

any hyperplane which contains

the hyperplane.
* Used in the proof of the

the line is perpendicular to

first

theorem of Art.

96.

102

HYPERPLANE ANGLES
3.

[n. v.

THEOREM

a line common

to all three

// three intersecting hyper planes with only are perpendicular to a given hyper-

plane, the line of intersection is perpendicular to the given

hyperplane.
59.

Planes lying in one and perpendicular to the other

of

two perpendicular hyperplanes.

i. // two hyperplanes are perpendicular, any in one, perpendicular to their intersection, is perpenplane dicular to the other, and any plane through a line of one, perpendicidar to the other, lies entirely in the first unless the

THEOREM

line itself is perpendicular to the second.

PROOF.

In the

first

case, a line

can be drawn in the

given plane perpendicular to the plane of intersection, therefore perpendicular to the second hyperplane.

and

The

given plane must then be perpendicular to the second hyperplane (Art. 52, Th. i).
the other hand, if the plane is given as perpendicular to the second hyperplane, and contains a line b which lies
in the first

On

hyperplane and

is

not perpendicular to the second,

b in the first

we can pass a plane through

hyperplane per-

pendicular to the intersection, and therefore perpendicular to the second hyperplane, by the first part of this theorem. But through a line not perpendicular to a hyperplane there
passes only one plane perpendicular to the hyperplane (Art. Therefore the given plane must coincide with 52, Th. 2).
the plane just drawn,

and lies entirely in the

first

hyperplane.

If a plane is perpendicular to a hyperplane, which contains the plane is perpendicular to any hyperplane
2.

THEOREM

the hyperplane.

For

in the plane are lines perpendicular to the given

hyperplane, and any hyperplane which contains the plane will contain these lines, and must itself be perpendicular
to the hyperplane (Art. 58, Th. 2).

58-60]

PROJECTION OF A PLANE
3.

103

THEOREM
pendicular
to

// two intersecting hyperplanes are pera given hyperplane, their intersection is also
given hyperplane.

perpendicular

to the

60. Projection of a plane upon a hyperplane. Angle of a half -plane and a hyperplane. The projection of a plane upon a hyperplane THEOREM i
.

is

a plane or a part of a plane;

it

does not

lie entirely

in one

line unless the given plane is perpendicular to the hyperplane.

PROOF. The given plane and any point of its projection determine a hyperplane which is perpendicular to the given hyperplane (Art. 58, Th. 2). The lines which project

same as the

the points of the plane upon the given hyperplane are the lines which project the points of the plane

upon the plane of intersection of the two hyperplanes (Art. The projection upon the 58, Th. i, and Art. 38, Th. 3). given hyperplane is therefore the same as the projection

upon the plane

of intersection.

Now

the given plane

is

not perpendicular to the plane of intersection if it is not perpendicular to the given hyperplane (Art. 59, Th. i). Therefore the projection is not a line, but is the plane of
intersection itself, or a part of this plane.*

When a half-plane with its edge in a given does not lie in the hyperplane and is not perpenhyperplane dicular to it, its projection upon the hyperplane is a halfCOROLLARY.
plane, or a portion of a half-plane, having the

same

edge.

edge in a given hyperplane does not lie in the hyperplane and is not perpendicular to it, the dihedral angle which it makes with the half2.

THEOREM

When

a half-plane with

its

less

plane having the same edge and containing its projection is than the dihedral angle which it makes with any other halfplane of the hyperplane having the same edge.

* It is a single convex part, any point lying between the projections of two points being also the projection of a point of the given plane. See foot-note, p. 79.

104

HYPERPLANK ANGLES

[n. v.

PROOF. Let a be the given half-plane, /3 the half-plane with the same edge containing the projection of a upon the hyperplane, and 7 q.ny other half-plane lying in the hyperplane and having
its

edge in

common

with

a and

/8.

hyperplane perpendicular to this common edge will intersect the three half-planes in three half-lines, a, 6, and c,

and these

half-lines,

taken two at a time, are the sides of

plane angles of the dihedral angles

whose faces are any

two of the three half-planes.

Now
of

a and

the last-named hyperplane and the hyperplane contain the half-lines a and 6, and are both per-

pendicular to the given hyperplane (the former by Art. 58, Th. 2). Hence they intersect in a plane which contains a and b and is perpendicular to the given hyperplane (Art.
59,

Th.

3),

and therefore contains the projection

of a

upon

In other words, the given hyperplane (Art. 51, Th. i). the half-line b is the half-line which contains the projecand a makes an angle tion of a upon the given hyperplane
;

with b which
(Art. 41 ,

is less

Th.

2).

than the angle which it makes with c Therefore the dihedral angle which a makes
it

with

)8

is less

than the dihedral angle which

its

makes with 7.

When
does not
it,

a half-plane with
lie

edge in a given hyperplane

in the hyperplane and is not perpendicular to the plane angle of the dihedral angle which it makes

its

with the half-plane having the same edge and containing projection is called the angle of the half-plane and hyperWhen a half-plane with its edge in a given hyperplane.
it is

plane is perpendicular to the hyperplane, a right angle with the hyperplane.

said to

make

The student may prove the

following theorem

When

a plane intersects a hyperplane but does not

lie in it

and

is

not perpendicular to it, that plane of the hyperplane which passes through the intersection and is perpendicular to the projection, is perpendicular to the given plane.

CHAPTER

III

ANGLES OF TWO PLANES AND ANGLES OF HIGHER ORDER

I.

THE COMMON PERPENDICULAR OF TWO LINES

Existence of

61.

minimum

distance between two lines.

to do with the common perpendicular two lines not in the same plane are usually made to depend on the axiom of parallels. We shall give a form of statement and proof for these theorems that will show

The theorems having

of

independence of this axiom, so far as they are independent of it. We shall give the proofs in full, because
their

they will be a guide to us in proving a similar set of theo-

rems

for planes in hyperspace.*

THEOREM.
there is

a point of a whose distance from b

distance from b of

Given two lines a and b not in the same plane, is less than or

equal

to the

PROOF.

Let

from

M to the

M be some point of
with

any

other point of a.
a.

line b lies

sects a at the point

The perpendicular a plane which interso that the half-lines of a from

b itself in

forming an angle with this plane, move away from it indefThat is, we can take a definite portion of a, a initely.

segment and
side of this
*

and say that any point of a outportion willbe distant from the plane, and thereits interior,

Elliptic

In this method of treatment one or two modifications are necessary for the Geometry. Thus in the proof in this article the statement that the line a
off indefinitely is

moves

Euclidean Geometry, p. 22).

not true in the "third hypothesis" (see the author's NonBut in this case we can take the entire line for AB,

putting A and B together at some point and regarding them as the two ends of the line Another slight modification will be pointed out In connection with Th. 2
of the next article.

105

106

COMMON PERPENDICULAR OF TWO

from
ft,

LINES
of

[in. i.

fore distant

by more than the distance

M from

b*
Let
all

AB

point of

AB.

denote such a portion of a, and let P be any The distance of P from the line b has for

a lower limit positions of that for every positive number

o.

By

this

we mean

there are positions of

where the distance, though never less than /, will be less than l + e. This is a theorem of irrational numbers, proved as follows If we separate all numbers into two classes, putting into
the first class all negative

numbers and every positive number which is never greater than the distance of P from b and into the second class every positive number which is greater than the distance of some point P from 6, the numbers of the first class will all be less than those of the second class, and we determine in this way a number which This occupies the point of separation of the two classes. / is, then, never greater than the diss the number /. but / + e is a number of the second tance of P from b :lass, and there is some position of P for which its distance
y

from b

is less

than

On any
ower limit

portion of
for

+ AB
e.

we prove
from
b

the distance of

a point

same way that has a lower limit, and the

in the

any

portion of

AB

taken

is

greater than or

*qual to the lower limit for any portion within xmtained, and so greater than or equal to /.

which

it is

Now we separate the points of AB into two classes, >utting A into the first class, and putting any other point
Z into the
>f

first class if for

is

the points of
I,

AC

the lower limit

the distance from b

greater than

but into the second

:lass if this
:lass all
,his

equal to /. The points of the first come before those of the second class; and by lower limit
is

separation

we determine a point

C', the point of sepa-

* See reference in previous foot-note.

DISTANCE BETWEEN LINES

ration,

107

which may be and B.

itself,

or

or

some point between

The distance of C" from b is exactly /. For suppose this distance to be / + 5, where 6 is some Take C\ on a towards A and Cz towards positive number. B, each at a distance of ^ 5 from C'. If C' coincides with A or is at a distance from A less than | 5, we take C\ at -4 or if C' coincides with 5, or is at a distance from B less
,

than ^5, we take C% at -B. In ACi, when Ci does not coincide with A, the lower limit of the distance from b
is

a number greater than

/.

In

AC

this lower limit is

/.

Therefore in CiC 2 the lower limit is /, and hence there is whose distance from b is less than CiC 2 a point " 6 being a / Now 2 5, positive number/' a value of e.
in

Therefore the distance of C' from the foot of the peris less than I d (one pendicular drawn to b through

side of a triangle less than the

sum

of the other two).

6,

That

is,

the distance of C' from one point of

is

and

there-

* In an equation or inequality "C'Z>"

of the segment

C'D

(see Introduction, p. 21), that

used, as in our text-books, for the length is, for the distance between the

points C' and D.

108

COMMON PERPENDICULAR OF TWO

from
b, is less

LINES

[m.

I.

fore certainly

than

assumed

this distance to

be equal to

+ 5, +
I
f

whereas we have
S.

This proves that the distance of C from b is exactly /, and is less than or equal to the distance from b of any other
point of
a.
/ is

The
62.

distance

not zero, since a and b do not intersect.

Existence of a

common

perpendicular.

An

infinite

number if more than one. THEOREM i. Given two

plane, the line along

lines a

and
the

b not in the

same

which we measure

minimum

distance

from

a point of a is perpendicular to both. PROOF. Let be a point of a whose distance from b

b of

is less

than or equal to the distance from b of any other of a, and let N be the projection of upon b, so point b If is that the line is perpendicular to at N.

MN

MN

not perpendicular to a, then the projection of upon a will be a point whose distance from one point of b, and
therefore certainly from
6, is less

than the distance of

from
line

b.

MN
b.

But this is contrary to hypothesis; hence the must be a common perpendicular to the lines a
// two lines not in the same plane have

and

THEOREM

2.

more than one common perpendicular, they have an infinite number of common perpendiculars; along all of these perpendiculars the distance between them is the same, and any two of the perpendiculars cut of the same distance on them*
* In the Elliptic Geometry we should say, If two lines not in the same plane have more than one common perpendicular within a restricted region," and in the proof we must assume that the points M, N, M', and N' are all within
such a region. This theorem
will

"

prove that

it is

cipal geometries,

see

Introduction, p. 19, foot-note) that

only in the Elliptic Geometry (of the three printwo lines not in the

this

same plane can have more than one common perpendicular. For it is only in geometry that two lines not in the same plane can be everywhere equidistant (see first paragraph of proof in Art. 61 and the reference in the foot-note on p. 105).

61,62]

WHEN THEY HAVE MORE THAN ONE

IOQ

PROOF. Let a and b be two lines not in the

same plane, and let and M'N' be two

MN

common
lars,

perpendicuperpendicular to

a at

M and M
distance

',

and
N'

perpendicular to b at

N and N'.
The

M'N'

is

then equal to the distance

f

MN.

For suppose we had

M'N < MN.

which have

Then

in the right triangles

NN'

in

common, we

should have

M'N < MN',

and
in the right triangles

which have
1

MM'

in

common,

MN
Our hypothesis

< M'N.
We
must have

leads to a contradiction.

M'N' = MN,

M'N = MN',
and
in

the right triangles which have

MN

in

common

MM' =
Now on

NN'.

a and b lay off equal distances from and N to 1 and both towards and N' or points Q respectively, both away from M' and N', and draw PQ, MQ, etc. The right triangles and MNQ are equal, having the two legs of one equal respectively to the two legs of

M M

MNP
MQ

the other.

Therefore

= PN.

In the same way we prove

M'Q =

PN

9
.

110

COMMON PERPENDICULAR OF TWO

LINES

[m.

i.

Q^

N'

On

we

is

from

take Q' in the same direction from N' that so that

N'Q'

= NQ,

and therefore so that QQ'

= NN'.

Then we have equal

right triangles

MNQ
= PN,

and M'N'Q',

M'Q' =
the triangles

MQ

PNN'

and M'Q'Q mutually

equilateral,

and

and M'Q' N' have two Finally, now, the triangles sides and the included angle of one equal respectively to

PNQ

two

sides

and the included angle

of the other,

and

JLM'N'Q' = Z.PQN.

But the angle M'N'Q'

is

Therefore PQN is a right angle. a right angle, and PQ is perpendicular to 6. In the same way we prove that PQ is perpendicular

to a.

PQ

is

also equal to

MN and to M 'N'
all of

the lines a and b

are equidistant along

these perpendiculars.
for

The last theorem is true and the proof holds good two lines in the same plane.

62,63]

OF A LINE AND PLANE

III

63. The common perpendicular line of a line and plane, and the common perpendicular plane of two planes which have a common perpendicular hyperplane. THEOREM i Given a line a and a plane /3 not in the same
.

hyperplane, there is
less

a point of a whose distance from

distance from
j8

j8

is

than or equal

to the

of

any

other point of a,

and
is

the line along

which we measure

this

minimum
lies

distance

perpendicular

to both.

PROOF.
of a

Let
j8.

be the

line

along which

the projection

perpendicular to

and b will be and line /3, any perpendicular to a and j3 will be perpendicular to b, and will be the projecting line of a point of a. Now a and b lie in a hyperplane but not in a plane (Art. 46, Th. i), and the distance of any point of a from its projection upon j8 is less than its distance from any other point of /3. Therefore, any point of a whose distance from b is less than or equal to the distance from 6 of any other point of a, will be a point whose distance from/3 is less than or equal to the distance from /3 of any other point of a and the line along which we measure this minimum distance will be perpendicular to a and |8, being perpendicular to a and b.

upon

Any

line perpendicular to a

THEOREM

2.

If a line

and plane not in one hyperplane

have more than one

perpendicular line they have an infinite number of these perpendiculars, one through every
point of the line ; along all of these perpendiculars the distance between them is the same, and any two of the perpendiculars intersect the line and plane at the same distance from

common

each other.

THEOREM
a

common

If two planes not in one hyperplane have perpendicular hyperplane (Art. 53), they have
3.

a common perpendicular line and a common perpendicular

plane.

112

POINT GEOMETRY

[ni. n.

PROOF.

The common perpendicular hyperplane

inter-

sects the two planes in lines which do not lie in one plane (Art. 27, Th. 3), and which have one common perpendicular This line is perpendicular to the two given planes line. (Art. 51, Th. 2), and the plane containing this line and the linear elements which it intersects is perpendicular to the two given planes.*

THEOREM

a common common perpendicular plane, they have an infinite number of common perpendicular planes, one through each linear
element;

// two planes not in one hyperplane have perpendicular hyperplane and more than one
4.

and any two of

these planes are equidistant along the

intersections of the given planes with the given perpendicular

hyperplane or with all the perpendicular there can be more than one.\

hyperplanes

if

II.

POINT GEOMETRY

geometry whose elements are the half-lines drawn from a given point. We shall make a particular study of the angles formed at a point O by the lines, planes, and hyperplanes which pass through O and our presentation of the subject will be simpler if we omit all mention of
64.
;

* In the Hyperbolic Non-Euclidean Geometry we have planes which do not and do not have a common perpendicular hyperplane, namely, planes with parallel elements. These planes, however, do have a common perpendicular
intersect

plane.

We can

hypersurface.

prove this by considering the geometry of an orthogonal boundarySee foot-note, p. 95.

t It is only in the Elliptic Geometry that a line and plane not in one hyperplane can have more than one common perpendicular line, and it is only in this geometry that two planes not in one hyperplane having a common perpendicular hyperplane
can have more than one common perpendicular plane (see foot-note, p. 108). Two planes in the Elliptic Geometry always have a point in common, a point common to
all their linear

slightly "restricted," applied to

;

Th. 3, remark). The above theorem should be some restricted region containing a part of the perpendicular hyperplane because two planes having a point in common always have two common perpendicular planes at such a point But this case will be fully treated
elements (Art.
27,

in Arts. 66-69.

63,64]

THE HALF-LINES AT A POINT

be understood that the

113

this point, leaving it to

and hyperplanes which pass through

considered.

lines, planes, are the only ones

In this way we can avoid frequent repetition " of the phrase pass through O." Thus we can speak of two planes as intersecting only when they intersect in a
line,

since all lines, planes,

always understanding that they intersect at O; and and hyperplanes have O in common,

we can always say

sect,

that two planes in a hyperplane interand that any plane and hyperplane or any two hyper:

planes intersect
that
it exists.*

not be necessary in each particular case to assume such intersection or to mention the fact
it

will

geometry of the various kinds of angles have at a point Point Geometry.^ may As O completely separates the two opposite half-lines drawn from it along any line, it is better to consider the half-line as one of the elements of Point Geometry, rather than the entire line. To every half-line there is, then, one opposite and a plane or hyperplane containing one of two opposite half-lines always contains the other. Two in a or and any plane planes hyperplane, hyperplane, Each hyperplane intersect in two opposite half-lines. has two perpendicular half-lines, one opposite the other. Point Geometry in space of four dimensions is a threeshall call the

We

which we

dimensional geometry. of the Point Geometry

all

That
if

we

get all the elements take four elements not in one

is,

we

elements coplanar with any two of them, hyperplane, and all elements coplanar with any two obtained by this
process.

We

can interpret the two-dimensional and three-

dimensional geometries given in chap. I as point geometries, if we give a proper meaning to their undefined
*
26,

Compare the remark on the

foot-note, p. 50.

limitation of hyperspace to four dimensions, Art.

and

t Klein, Nicht-cuklidische Geometric, Gottingen, pression "Metrik im Punktc."

I

893, vol.

I, p. 8,

uses the ex-

114

THE ANGLES OF TWO PLANES

[m.

m.

terms and confine ourselves when necessary to a restricted

(angular) region.
III.

THE ANGLES OF TWO PLANES*

in regard to perpendicular planes stated

65.

Theorems

in the language of Point

Geometry.

In this section we use

all planes and hyperplanes assumed to being pass through a given point O, and all half-lines to be drawn from O. For perpendicular planes we have proved certain theorems which can here be stated as follows (Arts. 43, 44, 47, and

the language of Point Geometry,

49):

THEOREM THEOREM
THEOREM

i.

Each plane has one and only one

two
inter sect y
their

absolutely

perpendicular plane.
2.

//

planes

absolutely

perpendicular planes
3.

intersect.

A
is

plane perpendicular

to

one of two ab-

solutely perpendicular planes is perpendicular to the other.

This theorem

included in the following.

Given two pairs of absolutely perpendicular planes, if either plane of one pair is perpendicular to one plane of the other pair, or if either plane of one pair intersects both
4.

THEOREM

planes of the other pair, then each plane of either pair intersects both

planes of the other pair and is perpendicular

to

both planes of the other pair.

THEOREM

5.

// two planes have a

to the

common

perpendicular

plane, the plane absolutely perpendicular to the latter is also

a common perpendicular plane

two given planes.

* This subject is treated by C. J. Keyser, "Concerning the Angles and the Angular Determination of Planes in 4-Space," Bulletin of the American Mathematical Society, vol. 8, 1902, pp. 324-329, and by I. Stringham, "On the Geometry of Planes in a Parabolic Space of Four Dimensions," Transactions of the American

Mathematical Society, vol.

tirely

2,

IQOI, pp. 183-214.

Stringham gets his results en-

by quaternians.

5 4-66]

THE MINIMUM ANGLE

6.

115

THEOREM

inly one pair of

Two planes which intersect have one and common perpendicular planes.
last

theorem, the two given planes intersect in a pair of opposite half-lines and lie in a hyper-

In the case of the

and their absolutely perpendicular planes intersect n a pair of opposite half-lines and lie in a hyperplane. 3ne common perpendicular plane is the plane containing these four half -lines, and the other is the plane of intersection of the two hyperplanes (see proof of theorem of The two given planes contain the faces of four \rt. 49). iihedral angles, and their absolutely perpendicular planes
Diane,
:

The plane angles :ontain the faces of four dihedral angles. )f all of these dihedral angles lie in the plane of intersection
two hyperplanes, since this plane lies in both hyperand is perpendicular to the edges of the dihedral The two hyperplanes contain the cells of four ingles. lyperplane angles whose common face is their plane of and each of the plane angles of these hyperntersection
)f

the

planes

)lane angles has for one side a half-line of the intersection

)f
>f

the two given planes, and for the other side a half-line the intersection of their absolutely perpendicular

)lanes.

>f

When the given planes are perpendicular, so that each the common perpendicular planes can be determined
we have
two hyperplanes, which are at the same time angles

)y half-lines or as the intersection of

,hen in each

he plane angles of a set of dihedral angles and the plane ingles of a set of hyperplane angles.
66. Existence of a minimum angle between two planes. The theorems and proofs of this article and the next two ollow the analogy of the theorems and proofs of Arts. 61

md

62.

The two

theories are, indeed, the same, as

we

hall see later (Art. 123).

Il6

THE ANGLES OF TWO PLANES

[m. in.

THEOREM. Given two planes a. and /3, there is a half-line of a whose angle with /3 is less tlian or equal to the angle made
with
j8

by any other

ftalf-line of

a*

PROOF.
f in

We

will write angle

of 360 a and b together in some position in a and regarding them

a (around

ab to denote a certain angle the point O), putting two half -lines

two sides of the angle. Let m be any half-line of a, either a or b or a half-line in some position between a and b. The angle which m makes
as the

with

/3

has for
for

all

positions of

m
e

a lower limit

<f>

o.

every positive angle where this angle, though never

is,

That

there are positions of

than

<t>

This

is

less than </>, will be less theorem of irrational numbers

proved as on p. 106. Within any angl^e of a, that

ab,

we prove

with j8

is, any portion of the angle same way that the angle which m makes has a lower limit and the lower limit for any portion

in the

of the angle ab is greater than or equal to the lower limit for any portion within which it is contained, and so greater

than or equal to

<f>.

separate the half -lines of the angle ab into two classes, putting a into the first class, and putting any other half-line c into the first class if for all positions of in the

Now we

is

angle ac the lower limit of the angle which it makes with greater than $, but into the second class if this lower
limit
is

equal to </>. The half-lines of the first class come before those of the second class; and by this separation we determine a half -line c' occupying the place of separa-

two classes, c being a itself, or b, or some halfline between a and b. The angle which c' makes with /3 is exactly 0. For suppose this angle to be + 8, where 5 is some
tion of the
<t>

* a and ^ pass through O, and a half-line is always a half-line drawn from 0. t Using the term angle as it ib used in trigonometry.

66]

THE MINIMUM ANGLE

In

117

positive angle.

take half-lines

c\

towards a and

c2
f

If c 6, each making an angle of \ 5 with c'. coincides with a, or makes with a an angle less than 6, we take c\ along a or if c' coincides with 6, or makes with

towards

an angle

less
c\

aci,

when

along ft. In the angle does not coincide with a, the lower limit of

than

5,

we take

c2

makes with |8 is greater the angle which the half-line than 0. In the angle ac 2 this lower limit is 0. Therefore
in the angle
c\c<t

the lower limit

is

0,

and hence there

is

a half-line
/?

in the angle c\c^

which makes an angle with

"
positive angle," a value

less
.

than

5,

\ S

being a

of

Now
^c'fl*'

5,

half-line c', forming a trihedral angle with w' and the projection of m' upon /?, makes with the projection an angle less than the sum of the other two face angles of the

and the

trihedral angle, that

is,

with one

half-line of #,

angle less than

less than + d. Thus c' makes and therefore certainly with )8, an 5, whereas we have assumed this
<f>

angle to be equal to

+ 5.

This proves that the angle which c' makes with ]8 is exactly 0, and is less than or equal to the angle made with
j8

by any other

half-line of a.

Il8

THE ANGLES OF TWO PLANES

ft

[m. in.

When a and
some
ft

intersect,

<t>

in all other cases

is

definite positive angle,

an acute angle unless a and

are absolutely perpendicular.

67.

common perpendicular planes. THEOREM. When two planes a and /3 do not intersect, tfie plane of the minimum angle which a half-line of a makes
Existence of
with
ft

is

perpendicular

to

a and ft.

PROOF.* Let m be a half-line of a which makes with ft an angle 0, less than or equal to the angle made with ft by any other half-line of a, and let n be the projection of m upon ft, so that the plane mn is perpendicular to ft along n. If this plane is not perpendicular to a, the projection of n upon a will be a half-line of a. which makes with one halfline of ft, and therefore certainly with ft, an angle less than But this is contrary to hypothesis and the plane mn 0,
;

must therefore be a common perpendicular plane and ft.

to

The plane mn,

posite half-lines,

the

common

perpendicular plane of the

a. and ft in two pairs of opin this plane we have two pairs of vertical angles, one pair of acute angles each equal to 0,

preceding theorem, intersects

and

and one pair

of obtuse angles (unless

a and

ft

are absolutely

perpendicular).

absolutely perpendicular to the plane mn is also perpendicular to a and to ft (Art. 65, Th. 5). Let 0' be one of the acute (or right) angles lying in the intersec-

The plane

m' in n and n 1 (when 0' is not a right angle) are the projections of m and m' upon ft, and that portion of a which lies within the right angle mm' will be projected upon that portion of
* Given by C. J. Keyser, see reference on
p. 114.

tion of this plane with a and ft. be the half-lines forming the angle 0',

We may

m' and n 1 a and n' in ft.

let

66-68]

COMMON PERPENDICULAR PLANES

lies

IIQ

|8

which

within the right angle nn

(see foot-note, p.

79)-*
68.

Planes with an

infinite

number

of

common

per-

pendicular planes. In the following proof we shall use a figure drawn to represent points at a given distance from O. Half-lines (drawn from 0) are represented in this
figure

by

points,

of the figure
lines

and planes by lines. Any three points not appearing on one line represent half-

which are the edges of a trihedral angle; or we can think of the points as vertices of spherical triangles, the different triangles of the figure lying on equal spheres but
not on the same sphere.
This figure
points taken at
;

is

not

all visible in

any one hyperplane.

Only four

random

at the given distance from

O could be seen at

any one time or along with a complete view of one spherical triangle would appear only those points of the figure which lie on the sphere
entire figure does, however, lie in a hypersphere not difficult to imagine a figure lying in a portion of a hypersphere as a slightly curved space.
of this triangle.

The

(chap. V),

and

it is

THEOREM. // two planes a and /3 cut out equal angles on a pair of common perpendicular planes, they have an infinite number of common perpendicular planes, the plane projecting any half-line of either upon the other being perpendicular
to both. On all of these common perpendicular planes they cut out equal angles; and if a and /3 are not absolutely per-

pendicular,
to the

any two of

these planes cut out

on & angles equal

angles which they cut out on a.

Conversely, if a. and 0, not being absolutely perpendicular, have more then two common perpendicular planes, the acute

angles which they cut out on pendicular planes are equal.

any pair of

these

common

per-

* Combining this theorem with the theorem of Art. 49, with Th. 3 of Art. 63, and with the first foot-note on p. 112, we can say that any two planes have a

common

perpendicular plane.

I2O

THE ANGLES OF TWO PLANES

[ill.

m.

a and /3 have a pair of common on which they cut out the acute perpendicular planes and </>'. Let the half-lines m and n be the sides angles of one of the angles 0, and let m' and n' be the sides of one In the of the angles </>', m and m' in a, and n and n' in /?.
PROOF.
planes
<t>

Two

right angles

mm' and

ww' lay

off

equal acute angles

mp

and nq. The right trihedral angles O-mnp and O-mnq have the face angles including the right dihedral angle of
one equal respectively to the face angles including the right dihedral angle of the other. Therefore

Zmq =
In the same way we prove

/.pn.

m'q

Z.pn'.

Now supposing <' to be equal to 0, we take the half-line r 1 q in the same direction (around O) from n that q is from n, so that
/.n'q'

and therefore so that

f

Z.qq

Then we have

right trihedral angles

O-mnq and

0-m'n'q',

in which the face angles including the right dihedral angle of one are equal respectively to the face angles including

the right dihedral angle of the other.

Therefore
pn.

m'q'

= Z mq = Z

68]

COMMON PERPENDICULAR PLANES

121

the trihedral angles O~pnn and O-m'q'q have the three face angles of one equal respectively to the three
face angles of the other, and so in these trihedral angles the r dihedral angles along the half-lines g and n are equal. f r Finally, the trihedral angles 0-pnq and 0-m q'n have

Now

these
of

same dihedral

angles,

and the adjacent face angles

one equal respectively to the adjacent face angles of the other, Z pn = Z.m'q', and Z qn = Z n'q'. There-

fore the dihedral angles along q

and n

in these trihedral

angles are equal, and as the latter is a right dihedral angle the former is also a right dihedral angle, and the plane

pq

is perpendicular to /8. In the same way we prove that the plane pq

is

perpendic-

ular to a.

The angle pq
mn, that

is

equal to the angle m'n' and to the angle

being the acute angles <t> and <//, let p and q be half-lines within the angles mm' and nn* q the projection of p upon 6 and p the projection of q upon a, so that the plane pq is perpendicular
',

equal to <. the Conversely, angles mn and

is, it is

m'n

both to

a and

to
is

/3.

Then

the angle pq

is

angle mn, that

Suppose, for

and the angles 4> and example, that we had

to <,

equal to the <j>' are equal.

Z pq

< Z

fif.

Then

in the right trihedral angles angle nq in common we should have

which have the face

Z pn < Z w#,
but in the right trihedral angles which have the face angle

mp in common,
Z mq
Our hypothesis

< Z

/w.

leads to a contradiction.

We

must have

Z pq = Z

urn.

122

THE ANGLES OF TWO PLANES

[m.

m.

In the same

way we prove Z pq = Z

m'n',

and therefore

Z m'n' = Z #w,
or

<'

<.

Now in the right trihedral angles angle mn in common, since also

Z pn = Z
w<?,

which have the face

we have Z
w/>

= Z

#.

Isocline planes. 69. The angles between two planes. which two planes cut The acute (or right) angles and
</>

</>'

out on their
is

common

perpendicular planes are called the

angles between the two planes. When one of these angles zero the two planes intersect and lie in a hypcrplane. The other angle is then the measure of the acute (or right)

dihedral angles whose faces lie in the two planes. When one angle is zero and one a right angle the planes are simply perpendicular. When one angle is a right angle the planes

are sometimes said to be perpendicular even if the other angle is not zero, but we have used the word perpendicular

alone as applied to planes only when the other angle is zero (see foot-note, p. 85). When both angles are right
angles the planes are absolutely perpendicular.

The angles which a plane makes with one i. two of absolutely perpendicular planes are Hie complements the of angles which it makes with the other; and any two
planes
planes.

THEOREM

make

the

same angles as

their absolutely

perpendicular
their

PROOF.

Let

a and

be two planes and a' and

/3'

absolutely perpendicular planes. In a plane perpendicular to these four the complete intersection consists of two pairs

68, 6g]

ISOCLINE PLANES
that
is,
f

123

of perpendicular lines;

of eight half -lines

j8

can be taken in the order

a0a

/3'a

a'/?'

which around the point

O, and which are alternately perpendicular.

When
an

the angles of two planes are equal the planes have

infinite

number

of

common

perpendicular planes, but

they cut out the same angle on

them

all.

The two

planes

Absolutely perpendicular planes are always isocline, and a plane isocline to one of two absolutely perpendicular planes is isocline to the other.

are then said to be isocline*

Any two

lines

pendicular

taken one in each of two absolutely perplanes determine a common perpendicular

plane (Art. 65, Th. 4), but in the case of two isocline planes which are not absolutely perpendicular only one of the

common
either.

perpendicular planes passes through any line of

Any two of these common perpendicular planes cut out equal angles on the two given planes and are themTHEOREM
* See
t

selves isocline.
2.f // two half-lines in a plane a make equal with a angles plane |8, the half-line bisecting the angle between
p. 125, and foot-note. Used by Veronese in finding the perpendicular

planes, Grundztige,

150.

124

THE ANGLES OF TWO PLANES

[in.

m.

them and the half-line bisecting

jections

the angle between their pro-

upon

/3

will lie in one of the planes perpendicular to

1

a and

/3.

PROOF.
are equal.

Let p and p be the half -lines

their projections

upon
if

/?,

and q and q so that the angles pq and p'q*

in a,

Then

is

the half-line bisecting the angle

*
pp' and n the half-line bisecting the angle qq the plane mn will be one of the common perpendicular planes of a
,

and

j8.

To prove

this

we

consider in succession the following

:

pairs of trihedral angles

O-pqq' and O-p'q'q,

0-p'pq

0-pp'q',

O-pqn

0-p'q'n,

0-pmn O-mpq O-mnq

0-p'mn,
O-mp'q',
1

O-mnq

In each case two face angles and the included dihedral angle of one trihedral angle are equal respectively to two face angles and the included dihedral angle of the other,
or the three face angles of one trihedral angle are equal respectively to the three face angles of the other.

COROLLARY.

lines in one of two planes

// more than two pairs of opposite halfmake any given angle with the other

plane, the two planes are isocline.

69)

THE "ISOCLINAL ANGLE"

The student may work

125

out the details in proof of the following theorems

:

Given that the angles ' mn and m'n' are the angles and 0' made by two planes a and 0, let p be
a half -line within the right

p
n

'

angle mm', and q its proThen if the angle pq is greater than jection upon ft. than 0'. That is, we have in succession the inequalities

it will

be

less

/.mn,

/.pn

/.mq,

/.mp

and then,

further,

/.pm'
/.pn'

/.pq

< < <

/.qn',

Zw'<7,
/.m'n'.
;

This requires that 0' shall be greater than the angle pq is always between and 0'.

and the magnitude

of

Now instead of taking

q as the projection of p upon 0, let us take p and
q as in Art. 68, so that

/.mp = /.nq. Then we shall have equal dihedral angles at p and q,

namely,

dihedral angle

mpq

dihedral angle nqp

and the plane

of the half-lines bisecting the angles

mn and

m'n' will

intersect the plane

pq at

right angles

and

will contain the half-line

bisecting the angle pq.

been called an

angle pq, because of the equal dihedral angles at p and g, has isoclinal angle of the planes a and 0, and the angles and 0' are its minimum and maximum values.*

The

* See p. 1 09 of article by Stringham referred to at the beginning of this section. Stringham says that "Two planes may be said to be mutually isoclinal when their isoclinal angle is constant/' and he speaks of two such planes as "isoclines" (p. 210). We have used the term isocline as an adjective. Stringham' s isoclinal angle is not
of special importance in this

development of the subject.

126

POLYHEDROIDAL ANGLES
IV.

[in. iv.

POLYHEDROIDAL ANGLES
Interior of a polyhedroidal

70.

Polyhedroidal angles.
Vertical

angle.

polyhedroidal

angles.

polyhedroidal

angle consists of the half-lines drawn through the points of a polyhedron (Art. 29) from a given point not in the

hyperplane of the polyhedron, together with

point.
is

this given

The

half-lines are called elements, the

is

polyhedron

the directing- polyhedron, and the pont

the vertex.

The

elements which pass through the vertices of the polyhedron are called edges, the elements which pass through the points
of

any edge of the polyhedron lie in the interior of a face angle, and the elements which pass through the points of any face of the polyhedron lie within a polyhedral angle

of the polyhedroidal angle.

The

interiors of the polyhedral

angles are the cells. Adjacent polyhedral angles are joined by their face angles, these lying in the planes of intersection of their hyperplanes. polyhedroidal angle is convex when

each of these hyperplanes contains no element of the polyhedroidal angle except those which belong to the
polyhedral angle of this hyperplane and to its interior. The polyhedroidal angle is convex when the directing-

polyhedron

is

convex.

We

shall

consider

only

convex

polyhedroidal angles.
lines

The interior of a polyhedroidal angle consists of the halfdrawn from the vertex through the points of the interior of the directing-polyhedron. The interior of a

convex polyhedroidal angle lies within any one of its hyperplane angles; and a point lying within all of the
hyperplane angles
angle.
lies

The polyhedroidal

in the interior of the polyhedroidal angle separates the rest of

hyperspace into two portions, interior and exterior to it. The polyhedroidal angle whose elements are half-lines
opposite to the elements of a given polyhedroidal angle
is

70, ?i]

TETRAHEDROIDAL ANGLES

127

vertical to

the latter. In two vertical polyhedroidal angles the face angles, dihedral angles, and polyhedral angles of one are all vertical to the corresponding parts of the other, and the face angles and dihedral angles of one are equal to the

corresponding face angles and dihedral angles of the other. A polyhedroidal angle taken together with its vertical
polyhedroidal angle may be regarded as a particular case of a hyperconical hypersurface (Art. 33).

A hyperpyramid, or the hypersolid which we call the interior of a hyperpyramid, may be described as cut from the interior of a polyhedroidal angle by a hyperplane that cuts all of its elements and does
not pass through
its vertex.

angles. The rectangular system. polyhedroidal angle is the tetrahedroidal angle, simplest for a tetrahedron directing-polyhedron. Any four having
71.

Tetrahedroidal

The

all

hyperplanes with a point but not a line common to them are the hyperplanes of a tetrahedroidal angle, and any

four half-lines

drawn from a point and not

in one hyper-

plane are the edges of a tetrahedroidal angle. The four hyperplanes, or the lines of the four half-lines, determine

a set of sixteen tetrahedroidal angles filling completely the hyperspace about the point, and associated in eight
In a tetrahedroidal pairs of vertical tetrahedroidal angles. are three of six there planes, pairs* angle opposites, the two
planes of a pair meeting only at the vertex. Otherwise the six planes all intersect, three in each edge. We may also speak of each half-line as opposite to the trihedral angle formed

by the other

three half-lines.

if it is within point any three of its dihedral angles whose edges lie in one plane, or if it is within two dihedral angles whose edges contain a pair of opposite edges of the tetrahedron. In the same

is

in the interior of a tetrahedron

We

do not mean here that they are absolutely perpendicular.

128

POLYHEDROIDAL ANGLES

[ra. iv.

is in the interior of a tetrahedroidal angle if within any three of its hyperplane angles whose faces lie in one hyperplane, or if it is within two hyperplane angles whose faces are the planes of two opposite face angles.
it is

way, a point

At each vertex

of

a pentahedroid
is

A
lines

particular case

a tetrahedroidal angle. the rectangular system: four halfis

mutually perpendicular,

six

face angles

which are

right angles lying in three pairs of absolutely perpendicular planes, the trihedral angles rectangular trihedral angles,

the dihedral angles right dihedral angles, and the four hyperplanes mutually perpendicular (see Art. 48).
72. Tetrahedroidal angles with corresponding face angles equal. We can make two tetrahedrons correspond, the four vertices of one to the four vertices of the other,

in

any order, corresponding

edges, faces, etc., being those

determined by corresponding vertices. Then we have the following theorem in regard to tetrahedrons with corresponding edges equal
:

THEOREM
equal
the
to the

to the

// each of the six edges of a tetrahedron is corresponding edge of a second tetrahedron, when
i.

four

vertices of

one are made

to

correspond in some order

four vertices of the other, then the faces and dihedral angles of one will be equal to the corresponding faces and
dihedral angles of the other.

For any two corresponding faces are triangles which are mutually equilateral and therefore equal, the angles of one
equal to the corresponding angles of the other; then any two corresponding trihedral angles have the face angles of one equal to the corresponding face angles of the other,

and therefore corresponding dihedral angles are equal.*

The axiom of parallels is usually employed in the proof of this theorem, but the proof can be modified so as to make it independent of this axiom (compare proof
of Th. 2). Indeed, any theorem in regard to angles at a point must be independent of the axiom of parallels (see the author's Non-Euclidean Geometry, p. 26).

7i,72)

TETRAHEDROIDAL ANGLES

129

In the same way we can make two tetrahedroidal angles correspond, the four edges of one to the four edges of the
other, in

any order, corresponding parts being those determined by corresponding edges. For these we have the
following theorem analogous to the theorem just given
:

THEOREM

2.

angle is equal

to

// each of the face angles of a tetrahedroidal the corresponding face angle of a second

tetrahedroidal angle,

when

the

four edges of one are made

to

correspond in some order to the four edges of the other, then corresponding dihedral angles and corresponding hyperplane
angles will all be equal.

PROOF.* Lay off convenient distances on the edges of the two tetrahedroidal angles, the distances in one equal to the corresponding distances in the other, forming two
pentahedroids.

Let

them

be

called

OABCD
all of

and

O'A'B'C'D'.

The

five tetrahedrons of

one correspond to
these tetraface
angles,

the five tetrahedrons of the other, and in

hedrons corresponding edges, corresponding

and corresponding dihedral angles are equal.

tetrahedroidal angles are all equal. we take a convenient point In the face at point on the line perpendicular to the edge

Thus, in

particular, the corresponding dihedral angles of the

two

OAB

H, say a
its

AB

middle

point, and through

to the plane
lines

H pass a plane absolutely perpendicular

This absolutely perpendicular plane

OAB.

OAB-C and OAB-D in halfof the hyperwhich are the sides of the plane angle at Th. In the halfC-OAB-D i). (Art. 54, plane 'angle from drawn half -line OAB-C the perpenhyperplane dicular to the plane OAB will meet the tetrahedron OABC at a point P which is not a point of the plane OAB (Art. 21, Th. i). If P is not a point of the half-plane
will cut the half-hyperplanes

* It will help the student to note the analogy of this proof to the proof of the corresponding theorem about trihedral angles as given in many of our text-books.

130

POLYHEDROIDAL ANGLES
it will

[m.

iv.

AB-C
edge

Ue in the face

OAC

or in the face

OBC

or in

the edge OC, and the half-plane

AB-P

will intersect the

OC

(Art. 8, Th. 2).

We

can then modify our construc-

tion, taking the point C at this intersection, so that the halfplane AB-P will be the same as the half-plane AB-C, and
will in any case meet the halfthe perpendicular from P. Now let A' be the middle the at AB-C point plane

point of AB. Then the angle at A' in the triangle the plane angle of the dihedral angle O-AB-C.

HKP

is

In the same way we can modify the position of Z>, if necin the half-hyessary, so that the half-line drawn from

perplane OAB-D perpendicular to the plane OAB will meet the half-plane AB-D at a point Q\ and in the triangle HKQ the angle at K will be the plane angle of the dihedral
angle

O-AB-D.

It is the angle
lars,

PHQ,

formed by these two perpendicu-

which

is

the plane angle at

H of

the hyperplane angle

C-OAB-D.
repeat the construction in the second pentahedroid, using the same letters with accents, making the distance f of from A'B' equal to the distance of from AB, and

We

changing the positions of C and D if necessary, so that their distances from O' are equal, respectively, to the final distances of C and D from O. We have, then, to prove that the angle P' H'Q' is equal to the angle PHQ.
f
f

',

In the tetrahedrons

OABC

and O'A'B'C' the dihedral

Therefore

angle equal to the dihedral angle A'B'. their plane angles are equal, that is,
is

AB

Z
and, as are equal.

HKP
KP HP

= Z H'K'P';

HK = H

A', the right triangles

H KP and H'K'P'

Thus we have

and

= K'P' = H'P'.

72, 73l

TETRAHEDROIDAL ANGLES

131

Similarly, triangles

H KQ and
KQ = HQ =

H' K'Q'
1
,

are equal,

K'Q

and

H'Q'

Now

in the tetrahedron

A BCD

the angle

PKQ is the

plane angle of the dihedral angle

AB.

Thus

Z
the triangles

PKQ

= Z P'#'(X,

PKQ and

P' K'Q' are equal, and

Then,

finally,

the triangles

PHQ

and P'H'Q' are mu-

tually equilateral,

and

Z P//(? = Z P'H'Q'.
In this way we prove that any hyperplane angle of one tetrahedroidal angle is equal to the corresponding hyperplane angle of the other.

We

shall see in the next section that the

of this article arc identical,

two theorems and that the proof of Th. i,

interpreted as a proof in Point Geometry, holds for Th. 2

(Art. 76).
73.

The

bisectors of the hyperplane angles of a tetra-

hedroidal angle.

The half-hyperplanes bisecting the six hyperplane angles of a tetrahedroidal angle contain a half-line common to them a//, tlie locus of points within the tetrahedroidal
angle equidistant from
its

THEOREM.

four hyperplanes.
a, 6, c,

PROOF.
edges
of

Let

be the vertex and

and d the four

a tetrahedroidal angle.

The half-hyperplane
is

bisecting the hyperplane angle c-ab-d

the locus of points within this hyperplane angle equidistant from the hyperplanes of its cells, namely, the hyperplanes abc and abd

132
(Art. 57,

POLYHEDROIDAL ANGLES
Th.
i).

[m.

iv.

The half-hyperplane
is

bisecting the hyper-

the locus of points within this hyperplane angle equidistant from the hyperplanes abc and acd.

plane angle b-ac-d

a whose edge

These bisecting half-hyperplanes intersect in a half-plane a lies entirely is the line which contains a.

within both of the hyperplane angles, and is the locus of points within both hyperplane angles equidistant from the
three hyperplanes abc, abd, and acd. Now points within the hyperplane angle c-ab-d are on the same side of the hyperplane abd as the half-line c\

and points within the hyperplane angle b-ac-d are on the same side of the hyperplane acd as the half-line 6. The
half-plane a, lying within both of these hyperplane angles, is therefore within the hyperplane angle b-ad-c (Art. 54),

and so

entirely in the half-hyperplane bisecting this hyper-

plane angle. There are four such half-planes, a, 0, 7, and 5, having for edges the lines containing a, 6, c, and </, respectively,

each lying within the three hyperplane angles whose faces contain its edge, and each common to the half-hyperplanes
bisecting the three hyperplane angles.

The half-hyperplane bisecting the hyperplane angle c-ab-d contains both a and 0. Therefore these half -planes,
whose edges pass through O, intersect in a half-line g drawn from (Art. 22, Th.), and the points of g are equiThe points of g distant from all four of the hyperplanes.
within five of the hyperplane angles of the tetrahedroidal In particular, they lie within each of the two oppoangle.
lie

site
lies

hyperplane angles b-ac-d and a-bd-c. Therefore g within all six of the hyperplane angles (Art. 71), is
to the four planes a,
/?,

and contains all the points within the tetrahedroidal angle which are equidistant from the four hyperplanes.
7,

common

and

5,

73, 74]

ANGLES FORMED WITH HALF-PLANES

133

V.

PLANO-POLYHEDRAL ANGLES
angles,

74. Piano-polyhedral

analogous to polyhedral

and to polygons. Piano-trihedral angles. We come now to another class of angles which are analogous to polyhedral angles, more so, indeed, than the polyhedroidal
angles
angles of the preceding section. Polyhedral angles are themselves analogous to polygons. Thus we may define a polyhedral angle as consisting of a
finite

number

of half-lines,

drawn from a point O, and taken

and in a definite cyclical order, together with the point are of whose the angles consecutive the interiors edges half-lines of this order (see Art. 14).
In the same
define a piano-polyhedral angle as consisting of these half-planes taken in a definite cyclical order, together with their common edge and the interiors of the dihedral angles whose faces are consecuThe half-planes are the faces, tive half-planes of this order.
their common edge is the vertex-edge, and the interiors of the dihedral angles are the cells. If a, /?, 7, are the faces in order, the cells are the interiors of the dihedral
.

planes having a

way we may have a common edge, and

finite

number

of half-

and the piano-polyhedral angle angles a/3, 187, be described as the piano-polyhedral angle af3y ....
. . . ,

may

In the analogy of polyhedral angles and piano-polyhedral angles, as in the analogy of dihedral angles and hyperplane angles, it is a face in the hyperspace figure which corresponds to an edge in the
figure of the ordinary geometry,
face.

and a

cell

which corresponds to a

In the ordinary geometry the polyhedral angle lies, of course, and so here the piano-polyhedral angle is entirely in a hyperplane
;

assumed to

lie

in one space of four dimensions.

lie within the cells and have the vertexand those which are the faces of the piano-

Half-planes which
edge
for edge,

134

PLANO-POLYHEDRAL ANGLES
all called

[in. v.

polyhedral angle, are order (Art. 6).

elements

and are

in cyclical

piano-polyhedral angle occurs twice as an element

;

is

simple
shall

when no
is

half-plane
it

we

always assume that

convex

is

simple piano-polyhedral angle simple. the hyperplane of each cell contains no element except those Each face of the cell itself and the two faces of the cell.

when

of a

the

common

convex piano-polyhedral angle is a half-plane lying in face of two half-hyperplanes which contain two
cells.

adjacent

These two half-hyperplanes are the

cells

of a hyperplane angle, one of the hyperplane angles of the

piano-polyhedral angle.
lies in the vertex-edge of a whose cells contain the adjacent angle piano-polyhedral cells of the polyhedroid, and the edges of a polyhedroidal angle lie in the vertex-edges of piano-polyhedral angles

In a polyhedroid each edge

which belong to the polyhedroidal angle. The piano-polyhedral angle whose elements are

half-

planes opposite to the elements of a given piano-polyhedral angle is vertical to the latter. In two vertical piano-polyheare
dral angles the dihedral angles and hyperplane angles of one all vertical to the corresponding parts of the other.

A piano-polyhedral angle with three faces is called a pianoAny three half-planes having a common not but edge lying in one hyperplane are the faces of a piano-trihedral angle. Any three hyperplanes which intersect but have only a line common to all three are the hypertrihedral angle.

planes of a piano-trihedral angle. The planes of three such half-planes, or three such hyperplanes, determine eight
piano-trihedral angles, completely filling the hyperspace about their line of intersection, and associated in four pairs
of vertical piano-trihedral angles.

The piano-polyhedral
piano-trihedral angles.

angles of a pentahedroid are

all

74, 7Sl

DIRECTING POLYHEDRAL ANGLE

135

75.

Right sections.

Polyhedral sections of a piano-polyhedral angle. A hyperplane intersecting the edge of a

piano-polyhedral angle, but not containing the edge, intersects the faces in half-lines which are the edges of a polyhedral angle and the piano-polyhedral angle may be con;

sidered as determined

by a polyhedral angle and a

line

through its vertex not in its hyperplane. When either the piano-polyhedral angle or the polyhedral angle is convex, the other is convex.
piano-polyhedral angle might be regarded as a polyhedroidal angle with a directing polyhedral angle instead of a directing-polyhedron that is, the half-lines drawn through
:

the points of a polyhedral angle from a given point not in

its

hyperplane form a certain portion of a piano-polyhedral

angle.

THEOREM.

the vertex-edge of

hyperplane perpendicular at a point a piano- polyhedral angle intersects the

to

latter

in a polyhedral angle whose face angles are the plane angles at O of the dihedral angles of the piano- polyhedral angle, and whose dihedral angles have at O the same plane angles as the

hyperplane angles of the piano- polyhedral angle.

PROOF.

The

hyperplane,

being perpendicular to the

vertex-edge at O, intersects the hyperplanes of the dihedral angles in planes perpendicular to their common edge at

Therefore the face angles of the polyhedral are the angle plane angles of the dihedral angles of the pianothis point.

polyhedral angle. The given hyperplane is also perpendicular to the planes of the faces of the piano-polyhedral angle These planes are the faces of the hyper(Art. 52, Th. i).

plane angles of the piano-polyhedral angle;

therefore the

dihedral angles of the polyhedral angle have at O the same plane angles as the hyperplane angles of the piano-polyhedral angle (Art. 54, Th. 2).

136

PLANO-POLYHKDRAL ANGLES

|m.

v.

is

The polyhedral angle in which the piano-polyhedral angle cut by a hyperplane perpendicular to the edge is a right
Theorems proved by means
a right section.

section of the piano-polyhedral angle.

76.

of

The

Geometry. rems in regard to piano-polyhedral angles follow, as analogous to theorems true of polyhedral angles
:

piano-polyhedral angle in Point

Certain theo-

THEOREM

i.

The sum of two dihedral angles of a pianothan the third.

trihedral angle is greater

THEOREM 2.* The sum of the dihedral angles of a convex piano- polyhedral angle is less than the sum of four right dihedral angles.
// two piano-trihedral angles have the three dihedral angles of one equal respectively to the three
3.f

THEOREM

dihedral angles of the other,

their

homologous hyperplane

angles are equal; and if two piano-trihedral angles have two dihedral angles and tlte included hyperplane angle of one equal respectively to two dihedral angles and the included hyper-

plane angle of the other, the remaining parts of one are equal
to the

corresponding parts of the other.

piano-polyhedral angle plays the part of polyhedral angle in the three-dimensional Point Geometry at any point of the edge, and the polyhedroidal angle plays the part of

The

polyhedron in the Point Geometry at its vertex. Thus Th. i of Art. 72 and its proof themselves become a state-

ment and proof

of the

second theorem of the same

article.

The student may prove the following theorems independently of the axiom of parallels The area of a triangle is greater than the area of its projection on
:

any plane containing one

itself (see

of its sides, not the plane of the triangle the author's Non-Euclidean Geometry, p. 7, Cor.).
parallels,

* The proof of this theorem can be made independent of the axiom of

f See foot-note,

p. 128.

75-77]

D1RECTING-POLYGON
of

137
than the sum of the

The area

any

face of a tetrahedron

is less

areas of the other three faces.

Any

trihedral angle in a tetrahedroidal angle is less than the

sum

of the other three.

77. The direc ting-polygon, and half-plane elements. THEOREM. A convex polyliedral angle can be cut by a

plane in a convex polygon.

PROOF.* In a convex polyhedral angle each pair of consecutive edges lies in the face of a half-space which contains Let the verall the other edges (compare Art. 14, Th. 2).
tex be

and the edges the

half -lines a, 6,
.

c, d,
. . .

.,

and on

All of these these edges take points A, B, C, D, points except A and B lie on one side of the plane ABO\

and the half-planes AB-C, AB-D,

half-plane

AB-O,

together with the are in a definite order around the line

. .

but the half-plane AB-O on the same side of the Therefore one of these half-planes, say A B-C, ABO. plane comes next after AB-O around the line AB, and lies be-

ABj

all

tween

AB-O
is,

and each

of the others, so that

it

cuts the in-

teriors of the

segments OZ),

....
all of

the edges of the polyhedral angle in points, and the polyhedral angle itself in a polygon, a convex polygon since the polyhedral angle is
cuts

That

the plane

ABC

convex.

convex piano-polyhedral angle may be considered as determined by a convex polygon and a line not in a hyperplane with the plane of the polygon. The
It follows that a

elements of the piano-polyhedral angle (Art. 74) are then

half-planes having the given line for common edge and containing each a point of the polygon. The polygon is
* This is a theorem of geometry of three dimensions, for a convex polyhedral angle is denned as lying in a hyperplane, and in the proof we speak of the hyperplane of the polyhedral angle as ''space," and of any half -hyperplane lying within this hyperplane as a "half-space."

138

PLANO-POLYHEDRAL ANGLES
and each

[m. v.

called the directing-polygon,

side of the polygon

lies in the interior of a cell of the piano-polyhedral angle.

Two vertical piano-polyhedral may be regarded as a particular

hypersurface (see Art. 35).

angles taken together case of a piano-conical

A double pyramid, or the hypersolid which we call the interior of a double pyramid, may be described as cut from the interior of a piano-polyhedral angle by two hyperplanes which contain the directing-polygon and each a point of the vertex-edge.

78.

Edge Geometry.

The elements

half-planes with a

edge. We can build up a geometry by taking for elements the half-planes which have a given line for edge.

common

geometry an Edge Geometry. The halfplanes which lie in the same plane on opposite sides of this common edge will be called opposite elements, and to every
will call this

We

element there

is

one which

is

opposite.

Any two

non-

opposite elements determine a hyperplane which contains the given line. The geometry will be the geometry of these
half-planes and hyperplanes. plane as consisting of half-planes
of this

We may
;

think of a hyperand in the development

mention of the given line, and think of two figures as intersecting only when they have one or more half-plane elements in common (compare
shall

geometry we

omit

all

Point Geometry, Art. 64).

By

confining ourselves

when necessary

to a properly

restricted region (angular region

about the vertex-edge), to

a region, for example, consisting of all the elements which make an angle of less than 90 with any given element, and by giving a suitable interpretation to the terms used,

we have the geometry developed in Arts. 1-16. Thus for point we say half-plane, for line hyperplane, and for segment dihedral angle. The half-planes in any hyperplane have
the relations of order referred to in Art.
5, all

of the half-

77,78]

EDGE GEOMETRY

139

planes in a hyperplane being, however, in cyclical order We get all of the half -planes of this geometry if (Art. 6).
three not in a hyperplane, all that lie in a hyperwith any two of them, and all that lie in a hyperplane with any two obtained by this process. This geomplane etry is, therefore, two-dimensional.

we take

We

Art. 4, Art. 3

have the following theorem equivalent to Th. 2 of and so essentialy equivalent to the two axioms of If a' and /?' are two distinct non-opposite half:

planes of the hyperplane a/3, then the hyperplane <*'' the same hyperplane (Art. 20).

is

We

can also prove a theorem corresponding to the Axiom

of Pasch.

Suppose that

a,

j8,

and 7 are three half-plane

elements not in one hyperplane, and that a hyperplane X, not containing any one of these three, contains a half-plane
in the interior of the dihedral angle
187,

and does not contain

a half-plane in the interior of the dihedral angle a/3. The half-planes /3 and 7 are then on opposite sides of the hyperare on the same side. plane X, and the half -planes a and Therefore the half-planes a and 7 are on opposite sides of
this hyperplane,

and the hyperplane must contain a

half-

plane lying in the interior of the dihedral angle

ay

(Art. 28).

convex piano-polyhedral angle necessarily

lies

within

a restricted region, and therefore the geometry of convex polygons holds true in this way of convex piano-polyhedral Such an angle divides all the elements which do angles.
not belong to those outside.
it

into

two

classes, those in the interior

and

The

interior consists of the interiors of all

the dihedral angles whose faces are elements of the pianopolyhedral angle except those whose interiors also lie in it
;

and any half-hyperplane whose

face contains a half-plane

lying in the interior of a convex piano-polyhedral angle, will itself contain one and only one element of the
latter.

140

PLANO-POLYHKDRAL ANGLES

[HI. v.

Other theorems of triangles and convex polygons can be

interpreted
79.
in'

the

same way.
of a

convex piano-polyhedral angle. Dropping the language of our Edge Geometry, we may say that a convex piano-polyhedral angle divides the rest of
interior

The

hyperspace into two parts, interior and

terior contains the interiors of all

exterior.

The

in-

segments whose points

are points of the piano-polyhedral angle except those whose The interior belongs to the interior interiors also lie in it.
of each of its hyperplane angles
;

and any point which

is

in the interior of each of the hyperplane angles of a

convex

In piano-polyhedral angle is in the interior of the latter. the case of a piano-trihedral angle, if a point is in the interior of

two

of the hyperplane angles

it is

in the interior

of the piano-trihedral angle.

half-hyperplanes which bisect the a hyperplane angles of piano-trihedral angle intersect in a

Tfie

THEOREM.

three

half-plane, the locus of points in the interior of the pianotrihedral angle equidistant

from

the

hyperplanes of

its cells.

CHAPTER

IV

SYMMETRY, ORDER, AND MOTION

I.

ROTATION AND TRANSLATION

Rotation in a plane and rotation in a hyperplane. In a plane a figure can rotate around a point, or we may think
80.
of the entire plane as rotating on itself around one of its points, one direction of rotation being to the right and the

other to the

left.

In a hyperplane, or in the space of our experience, a plane perpendicular to a given line at a point O, rotating on itself around O, always remains perpendicular to the line,

say that the plane rotates around the line as axis, 01 On the other hand, a plane can rotate around one of its own lines as axis-line through a certain dihedral
axis-line.

We

and if a plane perpendicular to the axis rotates with a plane which contains the axis, the former rotates through an angle which is the plane angle of the dihedral angle through which the latter rotates. Thus we can compare the rotations of two planes perpendicular to an axis, since the plane angles of a dihedral angle are the same at any twc
angle
;

points of its edge (see foot-note, p. 97).

THEOREM.
in

When

all the

planes perpendicular

to

a line

a hyperplane rotate around the line in the same direction and through the same angle or at the same rate, figures in the hyperplane remain invariable, any two points being always at the same distance from each other. PROOF. Let A and B be two points, O and O' their projections upon the axis, and A' and B' their positions after
141

142

ROTATION AND TRANSLATION

[iv. i.

a rotation through an angle 0.

is

We are

to prove that

A'B'

equal to

AB.

half-plane OO'-B rotates in the same direction and through the same dihedral angle <j> as the half-plane OO'-A.

The

That

is,

dihedral angle and therefore

A-OO'-A' =

dihedral angle B-OO'-B',

f

dihedral angle A-OO'-B = dihedral angle A'-OO'-B Let a and be the planes in which A and B rotate, planes f perpendicular to the axis OO', and let C and C be the pro.

jections of

A and A upon
'

/3.

Then
last

the angles

CO'B and
triangles

C'O'B' are plane angles of the

angles,

mentioned dihedral
right tri-

and are equal. Thus we prove that the CO'B and C'O'B' are equal, and then that the angles ACS and A 'C'B' are equal. Therefore

AB

= A'B'.

That is, any figure in a hyperplane can rotate around a line, and we can think of the entire hyperplane as rotating on itself around one of its lines.
81.

rotation.

Rotation in hyperspace.* The axis-plane. Double In space of four dimensions one of two absolutely

* This subject is treated by F. N. Cole, "On Rotations in Space of Four Dimensions," American Journal 0} Mathematics, vol. 12, 1890, pp. 191-210.

8o, 8i]

THE AXIS-PLANE

143

perpendicular planes rotating on itself around the point where the two planes meet always remains absolutely perpendicular to the other. We can say, then, that the rotating plane rotates around the other plane, and we can call the other plane an axis-plane. Two planes absolutely perpendicular to a given plane at points O and O' lie in a hy-

perplane in which they are perpendicular to the line OO Thus we can compare the rotations of two (Art. 45, Th.).
r

planes absolutely perpendicular to the axis-plane by considering them as rotations in a hyperplane around an axisline.

THEOREM

i.

When

all the

lar to a given plane rotate

around

planes absolutely perpendicuthe given plane as axis-

plane in the same direction and through the same angle or at the same rate, all figures remain invariable, any two points
being always at the same distance from each other. PROOF. Any two points, with the absolutely perpendicular planes

through them,

lie

in a hyperplane in which the

rotation takes place around a line perpendicular to these Therefore, by the theorem of the preceding article, planes.

the distance between the two points remains unchanged.

Thus any

figure in hyperspace
2.

can rotate around a plane.

THEOREM

Rotations

around two absolutely perpenafter two

dicular planes are commutative;

such rotations

all

points of hyperspace take the

tion comes first.

same

positions, whichever rota-

PROOF.
tion

Let a and

|8

be the two planes.

it,

of these planes remains in

and

is

point in one subject only to the rota-

around the other, whichever rotation comes first. Let P be any point outside of a and /3. Let us suppose that the rotation around /3 turns P to P', and that the rotation around ot turns P to P" and P' to (); then the rotation
around
/3

will turn

P"

to Q, so that the final position of

144
will

ROTATION AND TRANSLATION

[iv. i.

be at Q, whether the rotation around a comes first, or P and P" are in a plane j3i absothe rotation around j8. lutely perpendicular to a at a point B, and P' and Q are

in a plane 0'i absolutely perpendicular to a at the point f to which B moves in the rotation around /3.

In the rotation around

and

a.

rotates

on

itself

around

the hyperplane containing P its intersection with ft, and

in this hyperplane,

the half-lines

BP

and B'P'> lying

rotates on

and

perpendicular to a, are coplanar.

containing

P" and a

itself

Also the hyperplane and the half -line

BP" is
turned.

coplanar with the half-line at

into which

it is

In the rotation around

into the latter.

the former hyperplane

is

is

turned

another hyperplane, the hyperplane containing fa and /3'i, which rotates on itself f around the line BB and in this hyperplane the halfthere
,

But

plane
f

BB'-P

is

turned to the position of the half-plane

half-line

BB -P", BP"

and the

B'P', which

lies in

the former,

to the position of B'Q, which must, therefore, lie in the latter. Thus the half-line B'Q, and the half-line at B' into which
is

turned by the rotation around

plane -BjB'-P", and perpendicular

coincide.

]8, lying in the halfto the edge 55', must

8i,82]

TRANSLATION ALONG A LINE

can break up the rotation around
into
a.

145

We

around

any number

of equal parts

and the rotation and take these

smaller rotations alternately, and by a limiting process we can derive a resultant

motion

of

which the two rotations are

call

the components.

We

shall

a combination of rota-

tions around two absolutely perpendicular planes a double rotation.

82.

Translation along a line.

Another

simple form of motion is translation along a line. In this motion all the points of the line move the same distance or at the

same

rate,

and
line.

in

the

same

direction

In the translation of a plane or of a plane figure along a line the plane as a whole remains constantly coinciding with itself, and each point of the

along the

plane or of the plane figure which is not a point of the line remains constantly on the same side of the line and at the same distance from it, moving in such a way that
its

projection

upon the

line is
itself.

as a point of the line

of translation.

translated along the line The line is called the line

The

instant,

and any

position of a point is fixed at each two points remain at a constant distance

figure along a line is a

from each other.

The

translation of

any

motion such

that each plane containing the line of translation, or the intersection of the figure by any such plane, is itself translated along the line through the given distance or at the

given rate.
tions

upon

Any two points, together with their projecthe line of translation, if not all in one plane, are

the vertices of a tetrahedron .which remains congruent to itself in the translation. Therefore, any two points remain

146

SYMMETRY
and any

[iv.

n.

at a constant distance from each other,

figure

remains invariable in the translation.*

In a hyperplane a translation along a line combined with a rotation around the same line gives a screw motion. After

same

a screw motion the points of the hyperplane all take the positions, whether the translation and rotation are

simultaneous or taken in succession.

A screw motion in a hyperplane can be regarded as a screw motion in hyperspace. In general, a screw motion in hyperspace consists of a translation along a line combined with a rotation around a plane containing the
figures in
line.

the hyperplane which is perpenHyperplane dicular to the plane along the line (Art. 50) move in this hyperplane in a screw motion along the same line.
II.

SYMMETRY t
Symmetry in a plane.
to

83.

Symmetrical positions.
lies

Two
is

points are symmetrically situated with respect

the point

which

midway between them.

The point midway

called their centre of symmetry.

cally situated with respect to

Two
a
line,

points are symmetriplane, or hyperplane

two points at their centre of symmetry. Such a line, plane, or hyperplane is called the line, plane, or hyperplane of symmetry.
which
is

perpendicular to the line of the

* For the points not on the on the theory of parallels

:

line of translation the character of

the motion depends

In the Elliptic Geometry translation along a line

is

a rotation around the pole,

polar line, or polar plane of this line, so that translation is not different from rotation. In the Euclidean Geometry the translation takes place along a system of parallel
lines

In the Hyperbolic Geometry points which are not on the line of translation move along a system of equidistant-curves. In the Hyperbolic Geometry we can also have translation along a system of boundary-curves, the curves cutting orthogonally

a system of parallel lines, planes, or hyperplanes. In this case there is no line of translation nor centre or axis of rotation (see the author's Non -Euclidean Geometry
chap. II, II). t See Veronese on order and symmetry, Grundziige,
146, 147,

and

references.

82-84]

IN A PLANE OR HYPERPLANE

147

Two

figures are symmetrically situated with respect to a

the points of the two figures can be made to correspond in such a way that all pairs of corresponding points are so situated.
point, line, plane, or hyperplane,

when

THEOREM

i.

Two

figures in a plane symmetrically situ-

ated with respect to a point can be

made

to

coincide point for

point by a rotation of one of them through 180

centre of symmetry.

around

the

THEOREM

2.

Two
to

figures in a plane symmetrically situlose this relation of

ated with respect

a line will not

try, if they are rotated in the plane around

symmeany point of the

line through the

same angle in opposite

directions.

We can in this way bring any point of one figure into coincidence with the corresponding point of the other, in the line of symmetry, and then by a second rotation around
this

new

point

we can bring a second

pair of corresponding

points into coincidence.

THEOREM

3.

When

a plane figure

is rotated

in a hyper-

plane which contains the plane through 180 around some line of the plane, it comes again into the same plane, to a position

which
line;

symmetrical to its first position with respect to the and so two figures in a plane symmetrically situated
is
to

with respect

a line of the plane can be made

to

coincide by
the

a rotation of one of them, in a hyperplane which contains plane, through 180 around the line of symmetry.
84.

Symmetry

in a hyperplane.

THEOREM i. In a hyperplane (or in ordinary space) any two figures symmetrically situated with respect to a line can be made to coincide point for point by a rotation of one of
them through 180 around
the line of

symmetry.

THEOREM

2.

Two

figures in a hyperplane symmetrically

situated with respect to

a point

can be put into positions of

148

SYMMETRY
respect to

[iv.

u.

symmetry with

of one of them through 180 this plane at O.

any plane containing O by a rotation around the line perpendicular to

Let a be the plane containing O, and c the line perpendicular to a at this point. Let A and A be two
PROOF.
1

corresponding points. If we rotate the figure to which A belongs through 180 around c, we shall bring A again into a plane with its first position and with A ' and c, to a position
will

where we will call it A ; and then this point and A ' be symmetrically situated with respect to the plane a.
i

figures in a hyper plane symmetrically situated with respect to a plane will not lose this relation of
3.

THEOREM

Two

symmetry,

if they are rotated

line of the plane through the

in the hyperplane around any same angle in opposite directions.

is rotated

THEOREM

4.

When
1

a hyperplane figure

in hy-

around some plane of its hyperplane perspace through (Art. 81), it comes again into tlie same hyperplane, to a position which is symmetrical to its first position with respect to
the plane;

80

and so two figures in a hyperplane symmetrically situated with respect to a plane of the hyperplane can be made
coincide by a rotation of one of them through 180

to

around

the plane of

symmetry.

84, 8 S ]

WITH RESPECT TO A PLANE OR POINT

figures

149

symmetrical could be made to coincide in space of four dimensions. This fact is mentioned and illustrated in nearly all popular See, for example, Fourth descriptions of the fourth dimension. Dimension Simply Explained (referred to on p. o), pp. 28, 48, 158,

Thus

which we

call

if

the space in which

we live were a hyperplane

214, etc.

85.

Symmetry

in hyperspace with respect to

with respect to a point. symmetry with respect to a point, a

a plane and In space of four dimensions we have

line,

a plane, and a

hyperplane (Art. 83).

THEOREM
respect to

i.

Any

two figures symmetrically situated with

a plane can be made to coincide point for point a rotation by of one of them through 180 around this plane as
axis-plane.

THEOREM

2.

Two

figures

symmetrically situated with

respect to a point can be made to coincide by a rotation of one of them through 180 around each of two absolutely perpen-

dicular planes through the point.

PROOF.

Let

be the centre of symmetry, and

let

a and

be two absolutely perpendicular planes through O. Let A and A ' be two corresponding points. If we rotate the
figure to

which

belongs through 180 around the plane a

SYMMETRY

[iv.

n.

we shall bring A to a position where we will call it A\. The three points A A ', and A are in a plane perpendicular to a and /3, all three at the same distance from a and at the same distance from /3. In this plane A and A\ are sym,

metrically situated with respect to /?, and a second rotation, around /8, will make the two figures coincide entirely.

We

can take for a and

/3

any two planes which are abso-

Either figure can be rotated lutely perpendicular at O. around one of these two planes, and then the same figure
or the other figure can be rotated around the other plane.

86.

Symmetry in hyperspace with respect to a line. THEOREM. Two figures symmetrically situated with
to

re-

spect

a line c can be put into positions of symmetry with

respect to

any hyperplane containing

around
AN
c.

by a rotation of one of

them through 180

hyperplane along

the plane perpendicular to this

A,

PROOF. Let be the hyperplane containing c, and a the plane perpendicular to 6 along this line. We are to prove that a rotation of one figure through 180 around a
brings any point A of it into a position where this point f and the corresponding point A of the other figure are symmetrically situated with respect to 6.

The two

points

with respect to the point

and A' are symmetrically situated O where the line A A' meets c.

8s, 86]

WITH RESPECT TO A LINE OR HYPERPLANE

151

If

we

around a we shall bring it Aij and we shall have

lar to
If

rotate the figure to which A belongs through 180 A to a position where we will call

A and A\
1

symmetrically situated

with respect to the plane

at

(see
/3,

A A

lies in

which is absolutely perpendicuTh. 2 of Art. 85). of proof A\ will coincide with A', and these two
/?

points will
If

be coinciding points in 0. ' does not lie in jS, the line A A\

is

is

bisected

by

/?

at a
6,

point O' and

perpendicular to
the line

a along

perpendicular to |8. The hyperplane c, contains the plane ]8 which

is

absolutely perpendicular to

at

(Art. 50),

and meets
0.

We
c
is

A'Ai&t its middle point O' where this line meets shall prove that the line A'Ai is perpendicular also
O'.

to

another plane of 6 at O', namely, the plane determined by

and the point

Thus we

shall
/?

perpendicular not only to

prove that the line A A\ but also to the hyperplane

1

9 (Art. 38, Th. 4). i lies in the line

drawn through

A
A,

perpendicular to a,

and therefore

in a hyperplane with

r
,

and

a.

More-

over, the line c, lying in the plane a, is perpendicular to 00' r lying in the absolutely perpendicular plane /?, and to

AA

by hypothesis. Assuming that A does not lie in /3, we have f the line c perpendicular to the plane A\, and then the f plane determined by c and O perpendicular to the same

AA

plane

dicular

AA'A\\ for two planes in a hyperplane are perpenwhen one contains a line perpendicular to the other.
f 7

Now A A\ is perpendicular to OO', the intersection of these two planes, since OO lies also in Therefore A* A\ is
ft.

perpendicular to the plane determined

line in

and O' for a one of two perpendicular planes perpendicular to

by

their intersection

is

perpendicular to the other.

1

This proves that A and A\ are symmetrically situated with respect to 0; that is, that a rotation through 180 around a puts any point A of one of the two figures into

152

SYMMETRY

[iv.it.

a position where this point and the corresponding point of the other figure are symmetrically situated with respect
to
8,

the hyperplane perpendicular to

a along

c.

In a single hyperplane we can see A, A', AI, the plane o, the line but we cannot see the rest of symmetry c, and the line OO' in we can see only the plane determined by ft and of the hyperplane c and O'.
of
;

87.

Symmetry
i.

in

hyperspace with respect to a hyfigures

perplane.

THEOREM

Two

symmetrically

situated

with

respect to a hyperplane will not lose this relation of symmetry,

around any plane of the hyperplane through same angle in opposite directions. PROOF. Let A and A be two corresponding points, and let a be some plane in the hyperplane of symmetry. The
if they are rotated

the

'

line

A A'

is

perpendicular to the hyperplane of symmetry,

and the plane

is

through
lies in

absolutely perpendicular to

also perpendicular to this hyperplane

(Art 50, Def.).

Therefore

A A'

the plane

(Art. 51,

Th.

i).

in-

tersects the hyperplane of

symmetry

in a line c

which
the

passes through

O
'

where

meets a, and in the plane

points the line

and

are symmetrically situated with respect to

c.

Now when the two figures rotate around a, the points A and A rotate in around the point O, and if the rota'

tion takes place through the same angle in opposite directions these points remain symmetrically situated with In the new position respect to the line c (Art. 83, Th. 2).
of the figures the line

A A'\

and

is

bisected

by

c.

This

lying in 0, is perpendicular to c, line is therefore perpendicular

(Art. 51,

to the hyperplane of points A and A' are

symmetry
still

Th.

i),

and the

symmetrically situated with re-

spect to this hyperplane.

86-89]

ORDKR ON A LINE
2.

153

with respect

// two figures are symmetrically situated a hyperplane we can bring any Jour non-coplanar points of one into coincidence with the corresponding points of the other, in the hyperplane of symmetry, without
to

THEOREM

disturbing this relation of symmetry. PROOF. Taking any plane in the hyperplane of

symmetry
;

bring a point A into the hyperplane of symmetry, and so into coincidence with its corresponding point A ' then taking a plane through A in the hyperplane of sym-

we can

metry we can bring a second point B into this hyperplane a plane through A and B enables us to bring a third point
;

into this hyperplane; and, finally, the plane enables us to bring the fourth point into this hyperplane. THEOREM 3. // two figures are symmetrically situated

ABC

with respect to a point, line, plane, or hyperplane, any segment of one is of the same length as the corresponding segment
of the other,

and any two corresponding

angles, dihedral angles,

or hyperplane angles are equal.

III.

ORDER

88.

The two

directions along a line.

Two

points which

are distinguished in
to distinguish

some way from each other enable us the two directions along the line determined

by them

(Art. 5).

When we

speak of the line

AB we
B
is

shall

often have

in

mind the

direction

on

this line that

from

We can speak of the line BA

as a different line.

Some-

when we wish to call attention to a direction along the line, or when there may be some question as to our meaning, we shall say order AB instead of line AB.
times
89.

Right and

left in

principles of order.

a plane. The two fundamental In a plane * the two half-planes which

and
so with all of this section.

are on opposite sides of a line (Art. 12) can be distinguished *

*

Or

at least in a restricted portion of a plane,

54

ORDER

[iv.

m.

as right and left with respect to a particular direction along the line. This notion we shall associate with the notion of

a right and

left direction of

rotation in the plane,

by con-

Two opposite half-lines sidering the rotation of half-lines. drawn in a plane from a given point would have to turn in
opposite directions around this point in order to pass to the same side of the line on which they lie and two half;

lines

drawn from a point along

different lines

would have

to

turn in opposite directions in order to turn towards each other through the interior of the angle of which they are
the sides.* We have then two principles on which we can base the theory of order in a plane
:

I.

is

A and B being any two points of a plane, a point which on one side of the line AB is on the opposite side of the
BA.
,

line

II. O, A and being any three non-collinear points of a is on one side of the line and A is on the opposite plane, side of the line OB.

OA

We

shall

speak of II as holding true of the half-lines

OA

and OB, and we can express it by saying that two lines drawn from a point along different lines lie in plane on opposite sides of each other.

half-

their

be the right side of a line and which the left is. an arbitrary matter that cannot be determined; but, having assumed the two sides of one line, we shall be able by means of I and II to determine the two sides of every
shall
line in the plane.

Which

We prove
and
line

this in

what

follows.

90.

The

right
i.

left sides of lines

through a point.
half-line

THEOREM
*

Given a line

BC

and a

OA

drawn

along some other

from a point

of the interior of the

6).

The half-lines drawn from a

point in a plane are in cyclical order (Art.

8g, QoJ

IN A PLANE

155

segment BC, if II holds true of the half-lines

of the half-lines

OA

and

OB and

OA

and OC, then I

will hold true of the line

BC.

B and C are on opposite sides of the line OA (Art. 12, Th. 2), and by II it follows that A must be on opposite sides of the lines OB and OC, or, what is the same thing, of the lines CB and BC.
For,

OA

COROLLARY. // / holds true of BC and II holds and OB, then II will hold true of OA and OC.
2.

true of

THEOREM
if

Given three half-lines,

OA, OB, and OC,

drawn in a plane from a point

I holds true of
the line

along three different lines;

OA

lines

OA

and

OB

and of

the half-lines

will hold true of the half-lines

and II holds true of the halfOA and OC, then II OB and OC.
,

OA C lie on the same side of the line OB, we can take A' in place of A that is, we can assume that A and C lie on opposite
;

A and be the half -line opposite to OA are on opposite sides of any other line through O (Art. 12, Th. 2), and if I holds true of the line OA, we can substi' tute for OA in any application of II. Hence, if A and
PROOF.

Let

OA

'

'

sides of the line

OB,

so that there

is

a point B\ of this line

lying in the interior of the segment AC, a point of the halfline or of the opposite half-line. Then

OB

I5^>

ORUKR
Bi and

|iv.

m.

C A

lie

on the same
"
k<

side of the line

OA

and

"
l

"

OC;
of each other,

or, since

OA

and

OC

lie

on opposite sides

Bi

lies

on opposite

sides of the lines

OA

and OC.

Now BI and B lie on the same side of every line through O except the line B\B itself, or they lie on opposite sides of every line through O except this line, and in any statement like the last we can substitute B for B\. That is,

lies

on opposite sides of the

of the

lines

OA

and OC.

But by the conditions

theorem

OA
C

and

OB
lie

lie

on opposite sides of each other,

and by hypothesis
and

on opposite sides
finally,

of

OB.

Therefore we must have,

OB

and

OC

lie

on opposite

sides of each other.

Given a half-line OA, we have only to assume that I holds true of the line in which this half-line lies and that
II holds true of all the half-lines
lines of the plane
will

drawn from

in other

taken with OA.

By

the

first

theorem

the lines through O, and by the second theorem II will hold true of any two non-opposite halflines drawn from O.

hold true of

all

91.

Right and
notation,

left

with respect to the sides of a triangle.

The
to
of

points of a plane,

ABC. Given three non-collinear A, B, and C, we shall write order ABC denote that side of a line, right or left, on which A is the line BC.
order
J5C, order

Then if I holds true of the line

ACB is the oppo-

go, 91]

WITH RESPECT TO A TRIANGLE

order

157

site of

A BC,

and

if

II holds true of the half -lines

BC

the opposite of order ABC. In is a line indicated according other words, the opposite side of to I by interchanging the last two letters, and according

and BA, order

CBA

is

to II

by writing the

set of three letters

backwards.

THEOREM.
C, if

Given three non-collinear points, A, B, and

the lines

I holds true of

CB

holds true of the half-lines and CA, then II will

BC, CA, and AB, and if II BC and BA and of the half-lines hold true of the half-lines AB and

AC.

have given that order CBA is the opposite of order ABC, and that order BCA is the opposite of order ACB\ also that ABC and ACB, BCA and BAG, and
PROOF.

We

CAB

and

CBA
We
order

are pairs of opposite orders.

We

are to

prove that order CAB is the opposite of order BAC. This will appear at once if we arrange these expressions
in

a table.
left, is

find, in fact, that

one side of a

line, right

or

denoted by any one of the three expressions

ABC,

order

BCA,

order

CAB,

and the other

side

by any one
order

of the three expressions

order

ACB,

BAC,

order

CBA.
and
II of all

When

I holds true of all three of these lines

three pairs of half-lines,

we can say

that a cyclical per-

158

ORDER
letters

[iv. in.

mutation of the three

does not change the meaning

of the expressions order

ABC and order ACS.

a line BC,
II

Given the right a&d

left sides of
line.

we can assume

that I holds true of this

Then by

we can determine

the right and left sides of all lines intersecting BC. I will hold true of all these lines, and II will hold true of any
half-lines

drawn from a common point along two

a line

of

these lines.

does not intersect BC, we can draw a transversal, say the line AB, and determine the right and left sides of this new line by applying II to the half-line A B
if

Now

AD

and the two opposite half-lines drawn from A along this If we draw any other transversal DC, II, being true line. of AD and AC and of CA and CD (Art. 90, Th. 2), will hold true of DA and DC, as also of the opposite half-line, AD produced, and DC. Thus the right and left sides of the line AD are determined by means of the line BC and

any

transversal.
shall say, therefore, that I

We

and II constitute a

defi-

nition of right and left for all lines of the plane, given the right and left sides of one particular line.

Order

A BC may

be used to denote a direction of rotation

If

in the plane

angle

ABC,

around the point B. passing along AB from

we

traverse the

tri-

shall turn at each vertex (through

A to B, and so on, we an exterior angle) in the

by order ABC.
in the

same
92.

direction, the direction indicated

Order

in a plane

unchanged by any motion

plane.

THEOREM
the

Figures symmetrical in a plane. i. Given two triangles

is the

ABC

and

A 1B'C'

in
if

same plane, order A'B'C'

where h

each of the three

same as order segments A A', BB', and CC' is

ABC
less

than

h,

is the shortest altitude of either

of these triangles.

i,

92]

UNCHANGED BY ANY MOTION

159
'

PROOF.
is less

and
is

'

are on the

than the distance of

same side of BC, since A A from the line BC. Therefore

order A' EC

the same as order

ABC.

and n be the perpendicular lines which bisect the four angles formed by the lines AC and BC. Then A is
Let
at a distance greater than \ h from each of these bisecting lines, the perpendicular

from
line

to

a bisecting

being one-half of an

oblique line from A to the line BC. It follows

that

and

lie

to-

gether in

one of the four

right angles formed

by

the two bisecting lines. Likewise and B' lie together in one of these four right angles, in one of the two, indeed, which are adjacent to the right angle containing A and

1
.

Now

the line

A'C

consists of the point

C and

two

opposite neither of which

lying right angles, the angle containing B and B'. Therefore, no point of the interior of the segment BB' can be a point of the line A'C, and order A'B'C is the same
is

half-lines

in

two

vertical

as order

A'BC.
,

Starting now with order A'B'C' we prove that this is the same as order A'B'C, and therefore that order A'B'C'
is

the same as order

ABC.

160

ORDER
the two triangles are equal

[iv.

m.

When
plane.

we can

consider

them

as representing two positions

of a triangle

moving

in the

We

shall say, then, that the order of

a triangle

cannot be changed by any motion of the triangle in its plane,* regarding this statement, however, as in part a In particular, definition of the phrase motion in a plane.
order
is

not changed by a rotation or translation of the

triangle in its plane.

THEOREM

2.

Two

figures

in

plane

symmetrically
coincide point

situated with respect to a line cannot be

made

to

for point by any motion in the plane unless they are groups
of collinear points.

PROOF.

We

can make any two points of one coincide

respectively with the corresponding points of the other in the line of symmetry, without disturbing the relation of

symmetry

(Art. 83,

Th.

2),

but in this position any cor-

responding points not collinear with these two will be on opposite sides of the line. Therefore any two corresponding
triangles are in opposite orders,f

and the order of one can-

not be changed to the order of the other by any motion in

the plane.
If

two

figures in

a plane symmetrically situated with

respect to a line have each a line of symmetry with respect to which they are self -symmetrical (for example, two
isosceles
triangles),

one can be made to coincide with

the other as a whole

point for point.

by a motion

in the plane, but not

point of one not a point of its line of symmetry will coincide with that point of the other which

Any

corresponds to
*

its

symmetrical point.

Or

at least in

restricted portion of the plane.

orders
of

t When we say that two sets of points are in the same order or in opposite we assume that each point of one set is associated with one and only one point the other set, and then, taking the points of one set in any order we please, we
this order with the corresponding order of the points of the other set.

compare

92,93]

IN A

HYPERPLANE

l6l

93.

Order in a hyperplane.

principles.

The

notation, order

The two fundamental ABCD. In a hyperplane

the two half-hyperplanes which are on opposite sides of a plane (Art. 23) can be distinguished with respect to the

order of a triangle in the plane and by associating the two sides of a plane with the two directions of rotation in the
;

hyperplane of a half -plane around

derive, as before, two principles which together will serve to define the two sides of any plane when we have given the two sides

its

edge,

we can

of a particular plane

I. A, B, and C being any three non-collinear points of a hyperplane, a point which is on one side of the plane of order is on the opposite side of the plane of order ACB.

ABC
II.

B, C, and

of a hyperplane,

is

being any four non-coplanar points on one side of the plane of order ACD

and

A is on the opposite side of the plane of order BCD. We shall speak of II as holding true of the half-planes CD- A and CD-B, and we can express it by saying that two
lie

half-planes with a common edge,* lying in different planes, in their hyperplanes on opposite sides of each other.

Theorems exactly analogous to those stated in Art. 90 of the half-lines drawn in a plane from a point O hold true
of the half -planes with a

common edge lying in a hyperplane. to Art. 91, suppose that three planes intersect by twos in three different lines, a, b, and c y and let be a point
As

which is not a point of b or c, B a point of b which is not a point of a or c and C a point of c which is not a point of a or b and not collinear with A and B. It will make no
of a
y

difference

lines, and so the three planes, have a point common to them all or not. We can treat the half -planes aB, bC> bA, just as in Art. 91 we have treated the half-lines AB, J8C, BA,

whether the three

*
II

The edge taken

in

the same

"

order"

in both.

162

ORDER

[iv.

m.

Given the two sides of a particular plane in a hyperplane, and assuming that I holds true of this plane, we can by II determine the two .sides of any plane in the hyperplane intersecting this plane, and then of any other plane in the hyperplane by using a plane which intersects this plane and the given plane. That is, we determine the two sides
of every plane in the hyperplane methods of Art. 91.

by

following exactly the

write order

Given four non-coplanar points, A, B, C, and D, we shall A BCD to denote that side of a plane on which
is

of the plane of order

BCD.

for order in a plane,

In accordance with the definitions and theorems given we can say that a cyclical permutation

of the last three letters does not change the

meaning

of the

that a non-cyclical permutaexpression order tion of these letters changes the expression to one denoting the opposite order, if I, as stated for hyperplanes, holds true.

A BCD, and

On

the other hand,

if

II holds true, for example, of the

half-planes CD- A opposite of order

and CD-B, then order BACD is the That is, the order is changed to to the opposite, according II, by an interchange of the first two letters.

A BCD.

By combining these operations we find that there are twelve different arrangements * of the four letters for which the expressions of the form order A BCD denote one
side of a plane,

and twelve
side.

for

which these expressions

denote the opposite

We
a
line

associate the line with

cannot speak of the right and left sides of a line except as we some plane in which it lies but direction along is a property of the line itself and is independent of any plane
;

or space that contains

it.

Likewise, we cannot speak of one and the other side of a plane except as we associate the plane with some hyperplane in which it
*

Corresponding to the group of twelve even permutations of four things.

93, 94]

UNCHANGED BY ANY MOTION

"
order in the plane," or direction of rotation,
is

163
independent

lies

but

of

any hyperplane.
94.

in the hyperplane.

Order in a hyperplane unchanged by any motion Figures symmetrical in a hyperplane.

THEOREM i. Given two tetrahedrons, A BCD and A'B'C'D', in the same hyperplane, order A'B'C'D' is the same as order A BCD if each of tftefour segments A A', BB', CC', and DD' is less than % h, where h is the shortest altitude
of either of these tetrahedrons.

A and A' are on BCD, since A A' is less than

PROOF.
plane.

the same side of the plane the distance of A from this

is

Therefore order

A' BCD

the

same as order

A BCD.
Let a and ,5 be the perpendicular planes which bisect the four dihedral angles formed by the planes A and BCD. Then we prove, exactly as in Art. 92, that A and A' lie

CD

together in one of the four right dihedral angles formed by f these two bisecting planes, and that B and B lie together in an adjacent right dihedral angle. It follows that no

point of the interior of the segment BB' can be a point of the plane A 'CD, and that order A'B'CD is the same as order A' BCD.
Starting
this is the

now with
same

order A'B'C'D',

we prove

first

that

as order

A'B'C'D, and then the same as

order

A'B'CD. Thus we have,

shall

finally,

order

A'B'C'D' the same as order

A BCD.
say that the order of four non-coplanar points of a hyperplane figure cannot be changed by any motion of the figure in its hyperplane, thus defining in part the
phrase motion in a hyperplane.

We

THEOREM 2. Two figures in a hyperplane symmetrically situated with respect to a plane cannot be made to coincide

164

ORDER

[iv.

m.

by any motion in the hyperplanc unless they are plane

figures.
Figures which we tall symmetrical in the ordinary geometry, "with the parts of one equal to the corresponding parts of the other but arranged in the opposite order/ can be placed in posifigures
1

tions of symmetry with respect to a plane. They will be so placed as soon as we have placed three non-collinear points of one upon the
rical figures in

corresponding points of the other. But it is better to define symmeta hyperplane as figures which are not plane figures,

by a motion
95.

but can be placed in positions of symmetry with respect to a plane of one or both in their hyperplane.

Order in hyperspace. The two fundamental principles.

notation,

The

order

ABCDE.

In hyperspace we shall

distinguish the two sides of a hyperplane (Art. 28) with respect to the order of four non-coplanar points of it. We

proceed as before and write

principles which two sides of any hyperplane when we have given the two sides of a particular

down two

together will serve to define the

hyperplane
I.

being any four non-coplanar points, a which on is one side point of the hyperplane of order A BCD is on the opposite side of the hyperplane of order ABDC.
II.

A, B, Cj and

A,
is

hyper plane ,

C, Z), and E being any five points not in one B is on one side of the hyperplane of order ACDE,
9

and A

on

the opposite side of the hyperplane of order

BCDE.

can express II by saying that two half-hyperplanes with a common face,* lying in different hyperplanes, lie

We

on opposite sides of each other. Theorems analogous to those

of Art.. 90 follow at once

for half-hyperplanes having a common face. As to Art. 91, suppose that three hyperplanes intersect

by twos

in three different ..planes, a, 0,

*

and

7,

and

let

A, be

The

face -taken in the

same order

in both.

94-96]

IN HYPERSPACE

165

a point of a which is not a point of /8 or 7, B a point of ft which is not a point of a or 7, and C a point of 7 which is not a point of a or /? and not collinear with A and JB. We

can then carry through the methods of Art. 91 just as we do in the case of a hyperplane (Art. 93). not points of one Given five points, A B, C, /?, and
, ,

hyperplane, we
order
order

shall write order

ABODE
is

to denote that

side of a hyperplane

on which

of the hyperplane of

BCDE.

Order

ABDEC denotes the same side of a hyperplane; ABC ED and order BACDE denote the opposite
Thus we can obtain
sixty different

side of a hyperplane.

permutations of the above form denote one side of a hyperplane, and sixty for which these expressions denote the opposite side.
96.

of the five letters for

which the expressions

Order in hyperspace unchanged by any motion. Symmetrical figures. THEOREM i. Given two pentahedroids, A BCDE and A'B'C'D'E', order A'B'C'D'E' is the same as order

A BCDE
and

if each of the five

is less

EE

than J

h,

segments A A', BE', CC', DD', where h is the shortest altitude of

that is, the shortest distance

either of these pentahedroids;

between a vertex and the hyperplane of the opposite cell. PROOF. A and A' are on the same side of the hyper1 plane BCDE, since A A is less than the distance of A from
this

hyperplane (Art. 40, Th.

the same as order

i,

Cor.).

Therefore order

considering the perpendicular hyperplanes which bisect the four hyperplane angles formed by the

A 'BCDE is Now by

A BCDE.

and BCDE, we prove, exactly as in hyperplanes Art. 92, that order A'B'CDE is the same as order 'BCDE (using in this case Th. 2 of Art. 57).

ACDE

Corresponding to the group of sixty even permutations of

five things.

66

ORDER

[iv. in.

Moreover, we find in this proof that B is at a distance In the same greater than | h from the hyperplane A 'CDE. way we prove that C. is at a distance greater than f h from
the hyperplane A'BDE. Then, since the segments BB' f and CC are each less than \ h, we can bisect the four

hyperplane angles formed by these two hyperplanes and prove that no point of the segment CC" is a point of the hyperplane A'B'DE.
Therefore order

A'B'C'DE

is

the

same

as order

Starting that this is the

it is

A'B'CDE. now with order A'B'C'D'E', we prove

first

same as order A'B'C'D'E, and then that

as

the same as order A'B'C'DE. Thus we have, finally, order A'B'C'D'E' the same

order

ABCDE.

We shall say that in any figure the order of five points not in a hyperplane cannot be changed by any motion of the figure in hyperspace, thus defining in part the phrase
motion in hyperspace.

THEOREM

2.

Two

figures

symmetrically

situated

with

respect to a hyperplane cannot be

made to coincide point for point by any motion in hyperspace unless they are hyperplane

figures.

In studying geometry of four dimensions we shall use the word symmetrical only with reference to figures which are not themselves hyperplane figures and which can be
placed in positions of symmetry with respect to a hyperplane by some motion in hyperspace. All the segments,
angles, dihedral angles, and hyperplane angles of one figure are equal to the corresponding parts of the other (Art.
87,

Th.

3),

but the points of the two figures are arranged

vertical

in opposite orders.

We

have in two

piano-polyhedral angles an

example

of symmetrical figures, the

two piano-polyhedral

g6, 97l

MOTION IN A PLANE

167

angles being symmetrically situated with respect to a line

(Art. 86, Th.).

Two vertical polyhedroidal angles, on the other hand, are equal and can be made to coincide, being symmetrically situated with respect to a point (Art. 85, Th. 2).
IV.

MOTION IN GENERAL

a plane on itself. In the ordinary geommotion of a plane figure in its plane, etry and of figures in space. With a plane figure we can associEven in the case of motion in a plane ate the entire plane.
97.

Motion

of

we speak

of the

we can think

of the entire plane as

moved on

itself.

In the

motion of a plane on itself the order of any three noncollinear points cannot be changed (Art. 92, Th. i, remark).

THEOREM
first to their
its first to

i.

Given any two

positions

of a
it

plane,

motion which takes three non-collinear points of

from
it

their

second positions will take every point of its second position.

from

PROOF.

In the

first

place, each point of the line con-

taining two of the given points remains collinear with them and at the same distances from them. Moreover, since the plane as a whole after such a motion coincides with its

second given position, any point not a point of this line remains at the same distance from it and with the same
it must come to its second position or projection upon it to a position symmetrical to its second position with respect
:

to this line.

But the

latter is impossible, as its distance

from the third given point would then be changed. COROLLARY.

// after a motion of a plane on
it

itself

one

occupied before, then every point will occupy the position that it occupied before, or every point will occupy a position that could have been

point of the plane occupies the position that

l68

.MOTION IN (iKXKRAL
itself

|iv. iv.

reached by a rotation of the plane on

angle around
the given point.
2.

through a certain

THEOREM

In.

a motion of a plane on
let

itself,

if two

positions of any point differ, second position of such a point, and

tion of the point

be the first
let

and

the

be the second posi-

whose

first position

was B.

If the line of

symmetry of A, then the second position could have been reached by a rotation of the plane on itself through 180 around the centre of symmetry of the two points. If the lines of symmetry do not coincide but do
is,

symmetry of A and B, that

and

coincides with the line of

if

coincides with

intersect in

a point P, the second position could have been reached by a rotation of the plane around P. If the lines of not coincide do have a do but common symmetry perpendicular line, the second position could have been reached by a translation along this perpendicular (see the author's Non-Euclidean Geometry, chap. II, I, 8, proof, p. 42).*

Motion of a hyperplane on itself. We can now speak of the motion of a figure in hyperspace. With a
98.

hyperplane figure we can associate the entire hyperplane, whether the motion of the figure be anywhere in hyperspace or within the hyperplane
itself.

Indeed, in the ordi-

nary geometry we can associate with a figure any additional points, and so speak of all space as moved about on itself, even though we know of no hyperspace within which it lies. In a motion of a hyperplane on itself the order of any four non-coplanar points cannot be changed (Art. 94, Th. i,
remark).

THEOREM i. Given two positions of a hyperplane, a motion which takes four non-coplanar points of it from their
* In the Hyperbolic Geometry the lines of symmetry may be parallel. The second position is then one which could have been reached by a translation along a system of boundary-curves (see the author's Non-Euclidean Geometry, chap. II, II).

97,98]

IN A

HYPERPLANE
it

169

first to their

its first to its

second positions will take every point of second position.

first

from

PROOF.

In the

place, each point of the plane con-

taining three of the given points will come to its second Moreover, since the hyperplane position (Art. 97, Th. i).
as a whole coincides with its second given position, any point not a point of this plane remains at the same distance

from

it

and with the same projection upon

it:

it

must

to its second position or to a position symmetrical to its second position with respect to this plane. But the latter is impossible, as its distance from the fourth given

come

point would then be changed.

COROLLARY
the other

i.

Given bwo positions, one obtained from

by a motion of a hyperplane on itself, then any motion of the hyperplane on itself which takes three noncollinear points of
it

will take every point

from their from its

first to their

second positions second

position.

first

to

its

In

particular, a motion of a plane

on

itself (Art. 97,

Th.

2)

will determine a

motion on

itself of

any hyperplane contain-

ing the plane.

COROLLARY 2. // after a motion of a hyperplane on itself each of two points of the hyperplane occupies the position that
it

occupied before, then every point will occupy the position

it

that

occupied before, or every point will occupy a position

that could have been reached

on

itself

by a rotation of the hyperplane through a certain angle around the line containing

the two given points.

THEOREM
one point of

2.
it

// after a motion of a hyperplane on itself occupies the position that it occupied before,

then every point will occupy the position that it occupied before, or every point will occupy a position that could have been

reached by a rotation of the hyperplane on itself through a

certain angle

around a

line containing the given point.

170

MOTION IN GENERAL

[iv. iv.

The

By and

proof follows that of Th. 2 of Art. 97. The points C lie on a sphere, on a circle whose axis is the
.

axis of the rotation.

99.

of

Every motion in a hyperplane equivalent a plane on itself or to a screw motion.

to

a motion

THEOREM.
there is
before,

// after a motion of a hyperplane on itself no point which occupies ttie position that it occupied then every point will occupy a position that could
itself

have been reached by the motion on hyperplane or by a screw motion.

of some plane of the

PROOF.
position of

Let

be the

first

some point, let C whose first position was B and let D be the second point of the point whose first position was C. We will position assume that A, B, C, and D are not collinear, and if they are coplanar that the triangles ABC and BCD are in
y

the second position and be the second position of the

opposite orders in their plane.

Then

AB
and

= BC = CD,

A ABC

BCD.

Let b and c be the half-lines bisecting the angles ABC and BCD. The lines containing these half-lines have a common perpendicular m, even if coplanar, ABC and BCD

N on- Euclidean Geometry, chap. I, I, 10). The figure DCBA taken with c and b is congruent to the figure A BCD taken with b and c
being in opposite orders (see the author's

and

if

we

place the former

upon

the latter, each of these

98,99]

SCREW MOTION

171

half-lines will take the position

before.

If the lines containing these half-lines

which the other occupied have only

one

common
If

also fall

upon

equal.

m, this perpendicular will and its distances from B and C are these lines have more than one common perperpendicular
itself,

one of these perpendiculars will be a line through B intersecting the half-line c, and another will be a line through C intersecting the half-line b. Then the
pendicular,

perpendicular mid- way between them, being equidistant from B and C, can be taken for the line (see Art. 62, Th. 2). When A, B, C, and are coplanar, a screw motion along

with a rotation of 180

and a translation that puts

its

upon

will

put the entire hyperplane into

second

position.

Supposing that A, B C, and D are not coplanar, let f C' and D' be their projections upon m. Then, A', B
y j

in the first place, the tetrahedrons

CBB'C
C

and BCC'B'

are congruent.
triangle

Furthermore,

and

of the isosceles

ABC

are symmetrically situated with respect to

6,

the line containing

respect to this

and

same

line.

is symmetrical to itself with Therefore A' and C' are sym-

hedrons

metrically situated with respect to this line, and the tetraABB' A' and are congruent (Art. 84,

CBB'C

Th.

i).

It follows that the tetrahedrons

ABB' A

and

BCC'B' are congruent. In the same way we prove that these are congruent to the tetrahedron CDD'C'. Then
a screw motion along

m putting A

upon

B will put ABB' A

\

upon BCC'B' and BCC'B' upon CDD'C' that is, it will put B and C into their second positions and move the entire
hyperplane from
its first to its

second position.
:

The student may prove

Two
an

the following theorems successive rotations of a hyperplane on itself around

different axis-lines are together equivalent to a single rotation

axis-line, if the

two given

lines intersect.

When

two around the two axis-

172

MOTION IN GENERAL

[iv. iv.

lines are perpendicular to a plane the two rotations together will be equivalent to some kind of motion which the plane can have on itself

(Art. 07, Th. 2).

Conversely, if two successive rotations of a hyperplane on itself are together equivalent to a single rotation, the axes of the three rotations will be the three lines of intersection of three planes inter-

they will meet in a point if two of them secting two by two in a point ; or they will be perpendicular to a plane if two of are perpendicular to a plane.*
:

meet

them

100. Motion in hyperspace. In the motion of a figure in hyperspace we can associate with it any additional points, and so speak of all hyperspace as moved about on In this article we shall speak of a figure rather itself.

than of
that

all

we need

hyperspace, but we shall assume that all points In any to consider are included in the figure.

motion in hyperspace the order of five points not points of one hyperplane cannot be changed (Art. 96, Th. i, remark).

THEOREM

i.

Given two positions of a figure, one of which

can be obtained from the other by a motion in hyperspace, then any motion in hyperspace which takes Jour non-coplanar
points from their first
to their

second positions will take every

point of the figure from its first to its second position.

For each point of the hyperplane of the four given points comes to its second position in this motion (Art. 98, Th. i) and any point which is not a point of this hyperplane remains at the same distance from it, on the same side of it,
and* with the

come
on

same projection to its second position.

i.

upon

it,

so that

it

must

also

COROLLARY
itself will

In

particular,

a motion of a hyperplane

determine a motion in hyperspace of any figure the containing hyperplane or containing four non-coplanar
points of
it.

"The only other axes may be parallel.

possibility is in the Hyperbolic

Geometry, where the three

99, ioo]

IN HYPERSPACE

173

COROLLARY 2. // after a motion of a figure in hyperspace each of three non-collinear points occupies the position that it occupied before, then every point of the figure will occupy
occupy a position that could have been reached by a rotation of the figure through a certain angle around the plane of the three points.
the position that it occupied before, or every point will

THEOREM 2. // after a motion of a figure in hyperspace each of two points occupies the position that it occupied before, then every point of the figure will occupy the position that it
occupied before, or every point will occupy a position that could have been reached by a rotation of the figure around

a certain plane containing the two given points. PROOF. Each point of the line containing the two given
points occupies the position that

occupied before, and any hyperplane perpendicular to this line taken as a whole occupies the position that it occupied before (Art. 38, Th. i).
it

Now

such a hyperplane could be put into this second position by a motion on itself, since each point of the given line

except its intersection with the hyperplane always remains on the same side of the hyperplane with respect to any

given order of points in the hyperplane (Art. 96, Th. i). But in the hyperplane one point occupies the position that
it

occupied before, namely, its intersection with the given line. Therefore, the second position of the hyperplane could have been reached by a rotation around an axis-line

through this point (Art. 98, Th. 2), and the second position of the figure could have been reached by a rotation around
the plane determined by this axis-line and the line containing the two given points.

of Art. 97 as a
*
line is

This theorem might also be proved by interpreting Th. theorem of Edge Geometry (Art. 78).*
Edge Geometry
;

is Elliptic and in the Elliptic Geometry a translation along a a rotation around the pole of the line. In other words, the lines of symmetry of AB and BC always meet in a point.

174
101.

MOTION IN GENERAL
Motion
itself.
i
.

[iv. iv.

in

which one plane remains fixed or moves

only on

THEOREM

// after a motion of a figure in hyperspace

it

one point occupies the position that

every point will

occupied before, then

it

occupy

the position that

occupied before,

or every point will occupy

a position that could have been

reached by a single or double rotation (Art. 81). This can be proved by interpreting the theorem of Art. 99 as a theorem of the three-dimensional Point Geometry at

A, B, C, and D are to be interpreted as and c as the bisecting half -planes of the dihedral angles A BC and BCD, and m as a common perpendicular plane of the plane containing b and the plane containing c. The screw motion on m is then to be interpreted as a double rotation around m and the plane absolutely perpenthe given point.
half -lines, b

dicular to

at this point.

THEOREM 2. In a motion of a figure in hyper space, if two positions of any points differ, let four of these points be are their first positaken in such a way that A, B, C, and tions, and B, C, D, and E their second positions (see Art. 99).

There

is

then through

and

E and for B

a plane y, a plane of symmetry for and D. Let 8 be the corresponding plane

through D, lying in the second position of the plane whose Then if y and 8 intersect, or have first position is that of y.

a common perpendicular hyperplane,

of the figure is one

the

second

position

which could have been reached from the a in which one plane remains fixed or moves motion first by only on itself*
* In the Hyperbolic Geometry we can have planes y and & which do not intersect and do not have a common perpendicular hyperplane. These planes do, however, have a common perpendicular plane (see first foot-note, p. 112). The theorem is therefore true without restriction, whatever our theory of parallels, and can be
*?

stated very simply as follows:

In a motion of a figure in hyperspace, if two positions of any points differ, the second position of the figure is one which could have been reached from the first by a motion in which one plane remains fixed or moves only on itself.

IQI]

ONE PLANE MOVES ON ITSELF

175

PROOF. The proof need be given here only for the case where the four points are non-coplanar and their positions
are in different hyperplanes or in opposite orders in the

same hyperplane.

Then

AB = BC = CD = DE, Z ABC = Z BCD = Z CDE, and tetrahedrons A BCD and BCDE are congruent. Moreover, the tetrahedron DCBA is congruent to the tetrahedron A BCD, and EDCB to BCDE (Art. 93). The
hyperplane angle

E-DCB-A
and the

When these figures do not lie in one have the same order (Art. 95), and when hyperplane they lie in one do they hyperplane, A BCD and BCDE being in opposite orders, any order of points of one figure in this
hyperplane
180
is

angle A-BCD-E, the figure A BCDE.

the same as the hyperplane A is congruent to figure

is

EDCB

changed to the order of the corresponding

points of the other

by a

around one
centre of
of

of its

rotation of the hyperplane through planes (Art. 84, Th. 4, and Art. 94).

symmetry of A and E and the centre of and D do not change their positions when symmetry we reverse the figure, placing EDCB A upon A BCDE. Moreover, these two centres of symmetry cannot be

The

they were, the entire figure A BCDE one hyperplane, and, taken in two ways as above, would have symmetrical positions with respect to a line, so that corresponding points would be in the same
collinear with C.
If

would

lie

in

order in the hyperplane (Art. 84, Th. i). There is then a plane 7 determined by these two points and C, a plane of symmetry of A and E and of B and Z>, and the figure is

by a rotation in hyperspace through 180 around 7. At A, B, D, and E are also planes corresponding to 7, say a, j8, 8, and e; 5, for example, being in the second position of the plane whose first position is that of 7. The
reversed

176

MOTION IN GENERAL

[iv. iv.

figure consisting of consisting of J5, C,

B, C, and
8.

7
It

is
is

congruent to the figure

also congruent to the

D, and

Therefore the last figure consisting of /, Z>, C, and 7. two figures are congruent to each other.

and planes lie in a hyperplane, then the middle points of AB, BC, CD, and DE lie in a plane 77 perpendicular to all the planes a, /3, 7, 8, and e. For the line determined by the middle points of AB and DE and the line determined by the middle points of BC and CD are both perpendicular to 7, and therefore lie in a plane
If all these points

perpendicular to

(see foot-note, p. 77)

and, since this

plane
is

determined by any three of the middle points, it perpendicular to 8 and to /3, and then also to a and to e.
is

same distance and B and D from rj, A C, and E lying on one side of on the other side and the projections of AB, BC, CD, and be the DE upon 77 are equal. Let A', B' C", >', and If we rotate projections upon t\ of A, B, C, D, and E. and at the the hyperplane through 180 around the plane same time move this plane on itself so as to move A to B and B' to C', then we shall have moved A to B and B to C. Four non-coplanar points of the hyperplane will be in
points A, B, C, D, and
y

The

are at the

17

'

',

rj

'

their second positions:

that

is,

four non-coplanar points

of the given figure will be in their second positions, and the given figure itself will be entirely in its second position
(Art. 100,

Th.

i).
8,

The planes 7 and

not lying in one hyperplane,

may

intersect in a point 0, a point common to 8 fore also to all five of the planes, a, j3, 7,

responds to itself in the symmetry of with respect to 7, and in the two congruent figures BCDS and EDCy. O is therefore at the same distance from B,

and ft, and thereO cor8, and the planes ft and 8

.

C, and D, and so also from A and E. The projections of and upon the hyperplanes of the tetrahedrons A

BCD

xoi]

ONE PLANE MOVES ON ITSELF

will

177

EDCB
angle

be corresponding points in the congruence of

these two tetrahedrons.

will lie

within the hyperplane

A-BCD-E

plane angle. hyperplanes of orders

or within the vertically opposite hyperThat is, O will lie on opposite sides of the

A BCD

and

same
(Art.

side of the hyperplanes of orders

93).

EBCD, and so on the A BCD and BCDE

itself

will

then coincide with

in the
is

positions of the figure,

and the second position

two one which

could have been reached by a single or double rotation

around
If

(Th.

i).

7 and 6, not lying in one hyperplane, do not intersect but do have a common perpendicular hyperplane, they have a common perpendicular plane rj (Art. 63, Th. 3) and if there is only one common perpendicular plane, this
;

plane will coincide with itself when we place EDCy upon BCDd. This plane cuts y and 5 in linear elements, and
If y and 8 have at the same distance from C and D. more than one common perpendicular plane,* one of these planes passes through C and another through Z>, and there is a plane mid-way between these two (Art. 63, Th. 4)
is

which can be taken for ?/, coinciding with itself in these two positions of the figure, and therefore at the same distance from C and D. The plane 17, being perpendicular to y and passing through a line of 5, will intersect ft in a corresponding line and be perpendicular to ft also, since ft and 5 are symmetrically situated with respect to y. In other
words, the plane
t\

is
/3,

a
7,

common
5,

five of the planes a,

and

and

perpendicular plane to all is at the same distance

.

from all five of the points A,B,C,D, and Let A', B', C", Then Z?', and E' be the projections of these points upon 17.

A'B' = B'C' =
*

CD

= />'';

Two planes not in a hyperplane can have more than one common perpendicular plane only in the Elliptic Geometry, and in the Elliptic Geometry two planes always Therefore it is not intersect at least in a point (see second foot-note, p. 112).
really necessary to consider this case.

178

MOTION IN GENERAL
f

[iv. iv.

for example, are symmetrically situated with respect to the plane 7, or, we may say, to the interThe hyperplane angles B-y-C and section of 7 and ?j.

B'C and C'ZX,

are equal, and so therefore are all four of the hyperplane angles A-rj-B, etc. These four hyperplane angles are in the same order around the plane 77, and, if we rotate

C-y-D

the figure around rj through the hyperplane angle A-ij-B the half-hyperplanes fj-A, rj-B, etc., will be turned to the
If at the positions of the half-hyperplanes rj-B, ij-C, etc. same time we move TJ on itself so that A' shall move to B'

and B' to

C",

will

move

to

and

to

four non-

coplanar points of the figure will move to their second positions, and the entire figure will move to its second
position (Art. 100, Th. i).
102.

Composition of rotations.
i.

THEOREM

Two

successive

rotations

around two

dif-

ferent axis-planes are together equivalent to a single rotation

around an axis-plane, line (Art. 100, Th. 2).

if the

two axis-planes intersect in a

equivalent to

2. // two successive rotations are together a single rotation around an axis-plane, the axisplanes of the two rotations are in a hyperplane, and when they have a point in common they intersect in a line.

THEOREM

PROOF.
tions.

The
by

tion are left

points of the axis-plane of the third rotathe second rotation in their original posi-

Let

the

first

rotation

be the position of one of these points. If moves this point to another position JB,

the second rotation will

move

it

the hyperplane of

symmetry

of

back from B to A Then A and B will contain the

.

axis-planes of both of these rotations.*

* The three axis-planes will all intersect in one line, or will all be perpendicular to one plane, or (in the Hyperbolic Geometry) they will be parallel. The first theorem of Art. 101 and the two theorems of this article, so far as ro-

tations around a fixed point are concerned, were proved

by N. F. Cole,

see reference

on

p. 142.

IOI-I03]

CONGRUENT ARRANGEMENTS
V.

179

RECTANGULAR SYSTEMS

103.
to

Ways

itself.

point

0*

in which a rectangular system is congruent Four mutually perpendicular lines through a forma rectangular system (Art. 48), a system of

rectangular axes in hyperspace.

Distinguishing positive

and negative directions along lines through O, we let a, 6, c, and d be the positive half-lines of such a system. Any
three of these half-lines form a rectangular trihedral angle in a hyperplane perpendicular to the fourth and, without
;

disturbing the fourth, we can permute the three cyclically by a rotation in the hyperplane around the half-line which

makes equal angles with them.

of this type

By

we

find twelve different

combining relations ways in which the

four half-lines can be thus permuted. The system is congruent to itself in all of the twelve arrangements of the same order, f

the other hand, we can obtain any arrangement of three of these half-lines by a rotation in hyperspace. If,

On

we take an axis-plane through the first and the angle formed by the second and third, a bisecting rotation through 180 around this plane will permute the second and third half-lines without disturbing the first,
for example,

but the direction of the fourth half-line

will

be reversed

by
*

this process,

and the system

its

of four half-lines will not

completely occupy

original position:

in fact, it will

can regard this section as a section in Point Geometry ; and in the next shall use exclusively the language of Point Geometry. t Arrangements of the four half-lines in the three-dimensional Point Geometry at O. This geometry is "restricted," but is sufficiently extended to iiiclude four
section

We

we

A B, C, and on the four given half -lines, the various arrangements of the half -lines will correspond to the various arrangements ot the five points in which O comes first
,

mutually perpendicular half-lines. If we take with four points

Thus
then

if we say we have

order abed
"

abdc

= -

order "

OABCD,.

OABDC,

etc.

l8o

ISOCLINE PLANES

[iv. vi.

occupy a symmetrical position with respect to the hyperplane of the

first

three half-lines.
.

rectangular system determines three pairs of abWe can write them as solutely perpendicular planes.
the planes be and ad, ca and bd, and ab and cd. If we let be denote that order of rotation which, turning b through an angle of 90, would make it coincide with c, and so
for the other planes,

The

then
of

we can

express the different

possible

arrangements

four

mutually

perpendicular

arrangements of order in these pairs of abIn the case of any one pair solutely perpendicular planes.
half-lines as

there are four congruent arrangements, congruent arrangements symmetrical to

and four other

the
first

four.

Thus, in the pair of planes determined, one

the other

by

and

and

by a and

d,

we have

the four arrangements

congruent to one another, be and ad, ad and be, cb and da, and da and eb

and symmetrical
ments,
be

to these the four other congruent arrange-

da and be, cb and ad, and ad and cb. The three pairs be and ad, ca and bd, and ab and cd, are congruent to one another, each pair taken in any one of

and

da,

four different ways.

VI.

ISOCLINE PLANES*

104.

Rectangular systems used in studying the angles

of two planes. In this section, as in the third section of Ill (Arts. 65-69), we shall use the language of Point chap.

Geometry, all lines, planes, and hyperplanes being assumed to pass through a given point O.
*

study of the hypersphere and

of the

its relation

to the Point

Geometry

at

its

centre

(Art. 123) will help the student to understand this section.

Many

theorems of this section are proved by Stringham in the paper

re-

ferred to on p. 114.

103, 104]

THK ANGLES OF TWO PLANES

181

and d be four mutually perpendicular halflines of a rectangular system. Suppose in the planes ba and 0'. The and cd we lay off from b and c the angles half-lines which terminate these angles determine a plane and 0', a which makes with the plane of be the angles
Let
a, b y c,
<t>

ba and cd being the pair of

common

perpendicular planes

perpendicular to

a and

to be.

Given any two planes a. and /3, with their common perpendicular planes 7 and 7', we can take for be the plane /3 and for ba and cd the planes y and 7'. ad will be the plane
#' absolutely perpendicular to

and the angles and 0' will be laid off as above in the planes ba and cd. When we say that the plane a makes with be the angles and 0', we imply a sense of rotation in a corresponding to the order be. If p and q are the terminal half-lines of
/3,

these angles, then

is

the plane pq, with a sense of rotation

which turns p through 90 to the position of q. and 0' can then be any angles whatever, positive or negative. With a particular plane a (of order pq} each of the angles and 0' can be changed by any multiples of 2 TT, or both
at the

are angles

same time by odd multiples of TT. made with be by the plane qp.

and 0'

fl-

1 82

ISOCLINE PLANES
angles which
of
<f>,

[iv. vi.

The

a makes with the plane ad

0'.

are the

complements

and

When
by

0'

a. is isocline to the planes be

<

and ad; and

giving different values to

we have an

infinite

number

of planes isocline to be and ad and isocline to one another. These planes are all perpendicular to ba and to ed, and

constitute
105.
series.

what

is

called a series of isocline planes.

of

The common perpendicular planes

an

isocline

to be,

Conjugate series. When the plane a is isocline these two planes have an infinite number of common

0,

perpendicular planes on which they cut out the same angle and any two of the common perpendicular planes cut out the same angle on a as on be (Art. 69).

If,

from

for example, in the planes be and a we lay off an angle ^ b and p, the terminal half-lines of these angles will

themselves form the angle and will determine a plane to the be and to a. perpendicular plane

This plane y, being perpendicular to be, is perpendicular to its absolutely perpendicular plane ad, and therefore is one of the common perpendicular planes of a and ad.

The terminal

half-lines of the angles

angle 0, determine in

\l/, forming in y an a of sense rotation y corresponding

I04-io6]

SERIES OF ISOCLINE PLANES

183
of 90 be a half-

to ba

and the

half -line of 7,

making an angle
lies in be, will
\l/

with that terminal half-line which

line of ad,

making with a the angle in the same way that was formed in the planes be and a. The plane 7 may, then, be regarded as determined by laid off on the planes be and ad. the angles equal to But this construction is independent of the angle and the
this angle
\f/

position of a. The plane 7 is therefore perpendicular to all the planes of the isocline series obtained by giving different values to <f> and laying off these angles in ba and

cd from b and

c.
\l/

By

giving different values to

of planes

we have an

infinite

number

7 perpendicular

to all the planes

of the

isocline series.

Starting with ba,

we

construct these planes

in the same way that the a-series was constructed, and so they themselves form an isocline series with the planes of the cK-series for their common perpendicular planes, each plane a perpendicular to all of them.

Thus we have,
two
perpendicular to
shall call

series of isocline planes,

all

associated with a rectangular system, each plane of either series

the planes of the other series.

We

The
to be.

conjugate series of isocline planes. * a are not the only planes which are isocline planes can rotate the rectangular system around be

them

We

as

an

any

axis-plane, the half -lines a and d rotating in ad through angle to new positions, and in this new rectangular

system we can construct a new series of planes isocline to be and ad, with a new series of common perpendicular planes, perpendicular to all of these but not perpendicular
to

any plane of the first series except

106.

to be

and ad themselves.

The two senses

*

in

Conjugate series isocline

which planes can be isocline. in opposite senses. There are

Not the only planes through O.

184

ISOCLINE PLANES
in

[iv. vi.

which a plane can be isocline to a given plane corresponding to the two possible arrangements of a rectangular system. With a given rectangular system, using the construction of Art. 104, we can say that the plane a is isocline to be in one sense when we make <' = 0, and in

two senses

the opposite sense

when we make

0'

0.

Starting with a plane

pendicular planes

and to 7 and 7' respectively. two in the same direction from b in the angles lay same from c in y', or if we lay and in the direction plane 7 off two angles in opposite directions from b in 7 and in opposite directions from c in 7', we shall have two planes in the same sense. But if we take the same isocline to direction in one of the two perpendicular planes and opposite directions in the other, we shall get two planes
be half-lines
If

7 and common to

of absolutely per7' perpendicular to /3, let b and c

and a pair

we

off

isocline to

/3

When two
senses

we

in opposite senses. planes are isocline to a given plane in opposite can speak of one as positively isocline and the

other as negatively isocline. If a is the plane pq of Art. 104 and is isocline to be, we can determine the sense in which it is isocline by considering
the order of the four half-lines
ft,

c,

p,

and

q.

Now

in this

take in each plane, instead of the two given half-lines, any two non-opposite half-lines, determining their order by a positive rotation of less than 180.

determination

we can

That

half-lines in the plane pq such r that a positive rotation of less than 180 turns p to the position of g', we shall have order bcp'q' = order bcpq;
is, if

f p and q are two

to the positions of p' and q without becoming opposite, and so without changing this order (Art. 94, Th. i). In the same way we can take for
for

p and q can be turned

and c any two non-opposite half-lines in the plane be such that a positive rotation of less than 180 will turn the
first

to the position of the second.

io6J

THE TWO SENSES

we can determine the order of four noncoplanar half-lines drawn from O with reference to the
Conversely,
order of two isocline planes or of any two planes which have only the point O in common.

THEOREM i. If a is isocline in the same sense.

THEOREM
2.

to

ft

/3

will be isocline to

Two

conjugate series of isocline planes are

isocline in opposite senses.

For these correspond to two arrangements of the forms bead and bacd, which are of opposite orders.

In

fact,

if

we

rotate

the rectangular system around the plane which passes through b and
bisects the angle ac, we shall interchange a and
c

and the planes ba and be. d will have its direction reversed so that the plane ad will
coincide with the original position of (fc,not with the original One of the angles ^ (of Art. 105) is now laid position of cd.

the direction of one of the angles in the original of the while the position figure, present position of the
off in
<

86

ISOCLINE PLANES

[iv. vi.

other angle ^ and the original position of the other angle <t> are opposite.
absolutely perpendicular planes are isocline in both but in only one sense when we distinguish in each senses, a particular direction of rotation. Thus in the rectangular

Two

system ad and da are isocline to be in opposite senses.

Planes through any line isocline to a given plane. Planes to which given intersecting planes are isocline.
107.

THEOREM i. Through any half-line not in a given plane nor perpendicular to it* two planes can be passed isocline, one positively and the other negatively, to the given plane.

PROOF.

Let p be a

half-line

perpendicular to /8. If we pendicular to ]8, we can determine a rectangular system with four mutually perpendicular half-lines, 0, 6, c, and d,
so taken that
is

not in a given plane /J nor pass a plane through p per-

the plane be

terior of the angle ba.

and p a half-line in the inThen we can take equal to the

<t>

angle bp in this plane,

*

and lay

off

<t>

and

<t>

from

c in the

It is always to be remembered that in this section, as in the third section of chap. Ill, all half-lines are supposed to be drawn from the point O and all planes to pass through O.

io6, 107]

TWO

ISOCLINE TO A THIRD

I8 7

plane cd. The terminal lines of these angles determine with p two planes isocline to /3 in the two senses.

THEOREM

2.

Two

intersecting planes determine

of planes (absolutely perpendicular to each other) they are isocline in one way in the two senses respectively, and another pair to which they are isocline in the other way

a pair to which

in the two senses respectively.

""-~^^

PROOF.

the given planes intersect, and

Let p be one of the opposite half-lines in which let 7 and 7' be their common

perpendicular planes, 7 passing through the half-line p, and 7' the plane of the plane angles of the dihedral angles

which they form (Art. 49). Let q and q' be the half-lines which form one of these plane angles, and let c be the halfline bisecting the angle qq '.
l

In 7 and 7' we establish directions of positive rotation. Then in 7' the half -lines q and q' form with c angles which

may be

called

<f>

and

<t>.

If

now

in

bp plane be to which the two given planes are isocline in the

b so that the angle

shall

be equal to

7 we take a half-line 0, we shall have the

two

senses, as also to its absolutely perpendicular plane.

If,

on the other hand, we take

b so that the angle

bp shall

88

ISOCLINE PLANES

[iv. vr.

we shall have another plane be to which </>, the two given planes are isocline in the two senses, as also
be equal to
to its absolutely perpendicular plane.

The common perpendicular planes when two are isocline to a third. planes THEOREM i. // two planes are isocline in the same sense to a plane a, the common perpendicular planes which they
108.

through any half-line of a form a constant dihedral angle, the same for all positions of the line in a.
have with

and 7 be two planes isocline to a in the same sense, and let n be any half-line of a. The common perpendicular planes of a and ]8 form a series of the opposite sense, and the common perpendicular planes of a and 7 form a series of the opposite sense. Through the line containing n passes one plane from each series, two planes forming two pairs of vertical dihedral angles along this line and intersecting the absolutely perpendicular plane a' in two pairs of opposite half-lines which are the sides of the plane angles of these dihedral angles. Let mm' be f one of these angles, m and m so taken that /3 shall intersect the interior of the right angle mn and 7 the interior
PROOF.

Let

)8

of the right angle m'n.

Now

to a second position of n,

107, io8]

THE PERPENDICULAR PLANES

189

forming a certain angle with its first position, will correspond f second positions of m and forming the same angle with

their first positions, the angles being laid off in the same direction (around O) in the plane a'. Therefore the angle mm' will be the same in its second position as in its first,

and the corresponding dihedral angles along n formed in the two positions of the perpendicular planes will be equal.
// two planes isocline in the same sense to a plane a have with a a single pair of common perpendicular planes, perpendicular to all three, then all the common per-

COROLLARY.

pendicular planes of either and a are perpendicular to all three, and the two planes with a. belong to a single series of
isocline planes.

THEOREM
to

2. // two planes are isocline in opposite senses a plane a, the two planes with a. have one and only one pair

of

common

perpendicular planes, perpendicular

to all three.

V,

/\
PROOF.

and 7 be two planes isocline to a. in oppon be any half-line of a. The common perpendicular planes of a and /3 and the common perpendicular planes of ot and 7 form two series of opposite senses. One plane of each series contains the half-line n, and these
Let
site senses,

and

let

1 90

ISOCLINE PLANES
half-lines,

[iv. vi.

respectively, in the absoso taken that j8 shall interlutely perpendicular plane a', sect the interior of the right angle mn and 7 the interior

two contain

and

of the right angle m'n.

Then
plane

the half-line which bisects

the angle

mm'

will lie in a

common
two

to the

two

series

for this half-line forms equal angles in opposite directions

with

and m', and the planes

of the

series

which con-

tain this half-line contain the half-lines of

a which form

the same angle in the same direction from n, and which We have then a plane, and therefore coincide (Art. 106).
so a pair of planes, perpendicular to is, a pair of planes common to the

and y and to a that two series of perpendic;

ular planes. These are the only planes that can be perpendicular to j8 and 7 and to a. Any such plane will be perpendicular

to

all

the planes of the a/3-series and to

7-series.

Through any

half-line of it

the planes of the not a half-line of a

all

or a! pass two distinct planes, one belonging to each of these series; that is, there are planes of one series that intersect planes of the other series, and two intersecting

planes can have only one pair of planes (Art. 49).

common

perpendicular

planes which are isocline to a plane a in opposite senses and make the same angle with a. always

COROLLARY.

Two

intersect.

For they intersect each of the two perpendicular planes which they have in common with a, and the common angle which they form with a is laid off in the same direction from a in one of these perpendicular planes, though in opposite
directions in the other.

109.

Two

planes isocline to a third in the

same sense

isocline to

each other.

io8, 109]

TWO
Two

ISOCLINE TO A THIRD
planes isocline
to

IQI
the

THEOREM.

a third in

same

sense are isocline to each other in this sense also.

q'

P'

and 7 be two planes isocline in the same sense to a plane a. )8 and 7 have at least one pair of common perpendicular planes mn and m'n' where m, n, m', and n' are half -lines taken in the intersections of /3 and 7
PROOF.

Let

/3

with these perpendicular planes in such a

angles

way

that the

mm' and nn

two planes, n ular to 7 and a, p and p being taken in a in such a way that the angles pn and p'n' shall be two equal angles of the isocline planes 7 and a. Then p and p lie in planes pq and p'q' perpendicular to a and /?, yand q' being taken in 13 in such a way that the angles pq and '#' shall be two equal angles of the isocline planes a and /3. ' are right angles, pp' and 99 'are also Since mm' and
r
f

are two positive right angles in these and n' lie in planes np and n'p' perpendic-

right angles, qq' a positive right angle like

104).

mm'

(see Art.

Then we have

Z mq = Z

wV
f

Again, the dihedral angle along p formed by the halfplanes containing n and g, is equal to the dihedral angle

along p' formed by the half-planes containing n and q', since j8 and 7 are isocline to a. in the same sense (Art. 108, Th. i). Therefore, the trihedral angles npq and n'p'q'

IQ2

ISOCLINE PLANES

[iv. vi.

have two face angles and the included dihedral angle of

one equal respectively to two face angles and the included
dihedral angle of the other, so that the third face angles are equal ; namely,
'

Z nq = Z

n'q

Then in the right trihedral angles mnq and m'n'q' there arc two face angles, one adjacent and one opposite to the right dihedral angle, which have the same values in one
trihedral angle as in the other. The third face angles in these two trihedral angles are therefore equal namely,
;

Z mn = Z
This proves that
ft

m'n'.

and 7 are isocline. Now ft and 7 must be isocline to each other in the same sense as to a. Otherwise a and 7, being isocline to ft in opposite senses, would have with ft one pair of common
perpendicular planes perpendicular to
all

three (Art. 108,

Th.

2).

But

as

ft

and 7 are

isocline to

in the

same

sense,

there can be no planes perpendicular to all three unless the three belong to the same isocline series.

Poles and polar series. i. Given two conjugate series of isocline planes there is a pair of absolutely perpendicular planes (Art. 105), to which the planes of the two series are isocline in opposite
110.

THEOREM

senses, all at

and 7 a These planes intersect at right angles and are therefore both isocline, in opposite senses, to a pair of planes a and a', each forming an angle of 45 with a and the same angle with a' (Art. 107, Th. 2). All the planes, then, of the first series are isocline to a and a' in the same sense as )8, and all the planes of the conjugate Now every series in the same sense as 7 (Art. 109, Th.). of one of the series has plane conjugate pair opposite halfPROOF.
of the first series

an angle of 45. Let ft be any plane

series.

plane of the conjugate

109,110]

POLKS AND POLAR SKRFKS

193

lines in common with /3, and every plane of the first series has one pair of opposite half-lines in common with 7, and these half-lines all make an angle of 45 with a and the same

angle with a'. Therefore, all the planes of both series are isocline to a and a! at an angle of 45.

THEOREM
PROOF.

2.

an angle of 45

lie

All the planes isocline to a given plane at in two conjugate series.

Two

planes which are isocline to a plane

in

opposite senses, and make with a an angle of 45, intersect and are perpendicular to each other (Art. 108, Th. 2, Cor. and Art. 107). Thus all the planes /3, isocline in one sense,
are perpendicular to any one of the planes 7, isocline in the opposite sense and through each half-line of 7 passes one
;

and only one of these planes (Art. 107, Th. i). But any two that are not absolutely perpendicular determine a
series of planes all perpendicular to 7,

one through each

pair of opposite half -lines of 7,

45

with a.

planes /3 all that the planes 7, isocline in the opposite sense,

to one series.

making an angle of Therefore these are the planes ]8, and the belong to one series. In the same way we prove
all

and

all

belong

The two
series,

isocline planes forming

an angle of 45

with

all

the planes of a given series may be called the poles of the and the series may be called one of their polar series.

Two

absolutely perpendicular planes always have two polar

series,

conjugate to each other.

COROLLARY.
series determined

When

two series of the same sense have a

pair of planes in common, these planes are tfie poles of the by tJie poles of the two given series.

In a complete system of all planes * isocline to a given plane in a given sense, any two series have a pair of planes in common.

THEOREM

3.

*AtO.

194

ISOCLINE PLANES
of the

[iv. vi.

For the poles

conjugate
are planes
111.
series

two given

whose

determine a pair of poles, being 45 from each of them,

series

common

to the

two given

series.

Planes intersecting two isocline planes. THEOREM i. If a plane intersects two isocline planes,
each of the four dihedral angles formed about one line of intersection is equal to the corresponding dihedral angle of the

four formed about the other line of intersection (see Art. 104).

PROOF, Let a and /3 be two isocline planes, and let a and b be half-lines lying in the intersections of these planes with a third plane. Through each of the half-lines a and Let n and m b passes a plane perpendicular to a and (8. be half-lines in the intersections of these planes with ft and

respectively, so taken that the angles

na and bm

shall

Then

be two of the equal angles formed by the isocline planes. the angles am and nb will also be equal, and the two trihedral angles abn and bam will have the three face angles
of one equal respectively to the three face angles of the other. It follows that the dihedral angle along a of the

trihedral angle bam is equal to the dihedral angle along b of the trihedral angle abn. But the latter dihedral angle

which corresponds to the dihedral former. These corresponding angles are therefore equal, and each of the four dihedral angles along a is equal to the corresponding dihedral angle along b.
is

vertical to the dihedral angle

no, in]

LIKE PARALLEL LINES

2.

195

Given two isocline planes a. and |3, not absolutely perpendicular, with non-opposite half-lines a and a' in a and b and V in $ so taken that

THEOREM

then the planes ab

sense.

and

a'b' will be isocline in the opposite

PROOF. If these planes are isocline, they are isocline in the sense opposite to that of a and /3, for the orders aba'b' and aa'W are opposite (Art. 106).

Now

the half-lines a' and &'

make equal
upon

angles with the

form with the four given half-lines two equal trihedral angles. Moreover, there is a plane X to which a and j8 are isocline in the
plane ab, for their projections
this plane

given sense and ab

is isocline

107, Th. 2). These form the same angle with

in the opposite sense (Art. planes and the four given half -lines all
X.

Then the plane

/?,

plane through a' isocline in the opposite sense to ab

X,

and the and to

form 108, Th.

this
2,

same angle with X and must intersect (Art. Cor.) in two opposite half-lines which form with

the plane ab the angle that a' forms with ab. In other words, the plane a'b' is this isocline plane through a', isocline to ab in the sense opposite to that of a and ]3.

196

ISOCLINE PLANES
four half -lines a,
b, a',

[iv, vi.

The

and

b' are

the edges of a

polyhedral angle in hyperspace, not in a hyperplane, having properties somewhat analogous to those of the parallel-

ogram
112.

of the Euclidean Plane

Geometry.

Isocline rotation.
f

In a simple rotation around a,

or in a double rotation around

pendicular plane a planes isocline to a

itself,

a and its absolutely perthe (Art. 81), complete system of all in a given sense is transformed into

the planes generally into one another. When the two rotations around a and a' are equal, all the planes isocline in the sense corresponding to the rota-

on themselves, the series conjugate to any series of these planes moving as a series on itself. Every halfline (drawn from O) rotates in a plane isocline to a, and any one of these planes and its absolutely perpendicular plane can be regarded as the axis-planes of the rotation, no partiction rotate

ular pair of planes playing in this way a special part. shall call this rotation isocline rotation, and the

We

com-

mon

angle of the

two rotations the angle of

isocline

the isocline

rotation.

THEOREM
opposite

i.

In an

rotation

every

does not rotate on itself remains isocline

to the rotation.

to itself

plane that in the sense

Let 7 be any plane rotated to a position Two half-lines a and b of 7 rotate to positions a f and rotating in two isocline planes, and we have
PROOF.

7'.
6',

Z aa'= Z
Therefore

W.

7 and

7' ^re isocline in the sense opposite to the

2).

rotation (Art.

in, Th.

THEOREM 2. A simple plane a moves any plane of

it

rotation

of angle 9
to

around a

a polar series
its

a position where

makes an angle of % 6 with

original position.

m,ii2]

THEIR TWO-DIMENSIONAL GEOMETRY

In an isocline rotation of angle
0,

197

PROOF.

conjugate to a series of the system with respect to

the rotation takes place
series

any plane which

is rotated on the planes of this 6. Now the isocline rotation can an angle through be decomposed into two equal simple rotations around the two poles, say a and a', of the series. Either of these

simple rotations, therefore, moves the given plane to a 6 with its original position where it makes an angle of
position.

But the effect of a simple rotation around a is same on both of its polar series. After such a rotation any plane of either series in its final position makes an angle
the
of 5

with

its original position.

a given plane, in a given sense, constitute a conical hypersurface of angle, double revolution. The hypersurface contains also all the
isocline to

The planes

and at a given

planes isocline to the given plane in the opposite sense and at the same angle, but the planes of either set contain all the half-lines which make this angle with the given plane,

and therefore completely fill the hypersurface (see Art. 1 1 8). In particular, two conjugate series lie together in such a hypersurface, which therefore may be said to consist of

the planes of either of these series (see Art. 124).

isocline in

will find it useful to think of the planes a given sense to a given plane as the elements of a two-dimensional geometry which is exactly like the

The student

geometry

of the sphere, absolutely perpendicular planes

of planes to great circles.

corresponding to opposite points on the sphere, and series For distance between two ele-

ments we should take double the angle between the two planes, and to measure the angle between two series we can measure the dihedral angle between conjugate planes
intersecting in one of the
108, Th.
i).

common

planes of the series (Art.

These conjugate planes are not themselves

198

ISOCLINE PLANES

liv. vi,

a part of the two-dimensional geometry, but the measure of the angle so determined will be the same as the distance
intercepted on the, polar series of its vertex-plane. An isocline rotation with respect to the planes of this system is
to be regarded as

no motion at

all

in this geometry, but a

simple rotation or a double rotation which is not an isocline rotation corresponds to a rotation of the sphere, in the latter case through an angle equal to the difference of the

two component

rotations.*

* Stringham calls a series of planes an ordinal system, and the set of planes isocline to a given plane at a given angle forms with him a cardinal system. He uses these terms, however, with reference to a particular pair of absolutely perpendicular planes They correspond to meridian and parallel circle taken on the sphere with reference

to a particular axis.

See p. 212 of the paper referred to on p.

14.

CHAPTER V
HYPERPYRAMIDS, HYPERCONES, AND THE HYPERSPHERE
I.

PENTAHEDROIDS AND HYPERPYRAMIDS*

:

113. Pentahedroids

the

point

equidistant

five vertices, the point equidistant

from the

five cells,

from the and

the centre of gravity. THEOREM i. In a pentahedroid, if two of the tetrahedrons can be inscribed in spheres, the lines drawn through the centres of these spheres perpendicular to their hyperplanes lie in a
plane;

when

they meet in a point this point is equidistant

from
can

the five vertices of the pentahedroid, the five tetrahedrons

all be inscribed

in spheres, and the

five lines

drawn through

the centres of these spheres perpendicular to their hyperplanes

all

pass through the same point .f

PROOF.

The common

face of the

two tetrahedrons

is

the interior of a triangle inscribed in a circle common to the two spheres. The absolutely perpendicular plane at the centre of this circle contains the centres of the spheres

and

is perpendicular to the hyperplanes of the .two tetrahedrons. It therefore contains the lines perpendicular to

the hyperplanes at these points (Art. 51, Th. i). Now the line drawn through the centre of a sphere perpendicular to its hyperplane is the locus of points equidisThis section and the next are continuations of the last section of chap. I, and the latter should be read again at this point. t There are other possibilities in the Hyperbolic Geometry, which the student may investigate if he is familiar with this geometry.
*

199

200

PENTAHEDROIDS AND HYPERPYRAMIDS

[v. z.

tant from the four vertices of any tetrahedron inscribed

in the sphere (Art. 40, Th. 2). If, then, the two perpendiculars meet, the point where they meet must be at the same distance from the five vertices of the pentahedroid ;

and a

through this point perpendicular to the hyperplane of any one of the five tetrahedrons contains a point of this hyperplane equidistant from the vertices of the
line

tetrahedron.

THEOREM

2.

The

half-hyperplanes

bisecting

the

ten

hyperplane angles of a pentahedroid all pass through a point within the pentahedroid, a point equidistant from the hyperplanes of
its five cells.

PROOF. At a vertex A we have a tetrahedroidal angle with six hyperplane angles, and the bisecting half-hyperplanes of these hyperplane angles have in common a halfline a, the locus of points within the tetrahedroidal angle

equidistant

from

its

four

hyperplanes

(Art.

73,

Th.).

Drawn from
by

another vertex

B we

have another

half-line

the locus of points within the tetrahedroidal angle at this vertex equidistant from the four hyperplanes of this

Now the piano-polyhedral angle angle. contains three hyperplane angles which belong to the tetrahedroidal angle at A and also to the tetrahedroidal
tetrahedroidal

AB

angle at 5, and the bisecting half-hyperplanes of these three hyperplane angles intersect in a half -plane a which must contain both a and b (Art. 79, Th.). The plane of a
intersects the pentahedroid in a triangle, and, as the half-

a and b pass within the angles at A and B of this triangle, they must intersect in a point O within the pentahedroid (see Art. 8, Th. i). We have, then, a point O
lines

equidistant from the five hyperplanes of the pentahedroid, lying within the pentahedroid, and lying in each of the ten

half-hyperplanes bisecting the ten hyperplane angles of the

pentahedroid.

113]

THE CENTRE OF GRAVITY

3.

2OI
the vertices of

THEOREM
cells

The half-lines drawn from

pentahedroid through the centres of gravity of the opposite

PROOF.

meet in a point. Given the

pentahedroid

ABCDE, we

will

write half-plane is the line

AB
and

AB

to denote the half -plane whose edge which itself contains the centre of

gravity of the triangle

to denote the triangle

CDE, and we will write triangle whose vertices are A and B and

AB
this

same centre of half-plane and

face.

gravity.

In this way we can speak of the determined triangle by any two vertices

of the pentahedroid

of gravity of the opposite are in All points in the interior of the triangle

and the centre

AB

the interior of the pentahedroid. The four half -planes AB, AC,

AD, and

AE

contain

respectively the half-lines drawn from the vertices J5, C, D, and through the centres of gravity of the opposite faces

in the tetrahedron

contain, therefore, the centre of gravity of this tetrahedron, as well as the vertex A. In the same way the four half-planes AB, BC, BD,

BCDE.

They

and the centre of gravity of the tetrahedron ACDE. Thus we have in the halftwo half-lines drawn from the vertices A and B AB plane and
contain the vertex

BE

through points in the opposite sides of the triangle AB. These two half-lines, therefore, intersect in a point P in the

and so in the interior of the But A and B are any two i). pentahedroid (Art. 8, vertices. Hence the half-lines drawn from the vertices
interior of the triangle
y

AB

Th.

of the pentahedroid

opposite of the pentahedroid.

half-lines

cells

must

all

through the centres of gravity of the intersect one another in the interior

These intersections all coincide. If, for example, the from A, B, and C intersected in three different points, these half-lines, and so the vertices A, B, and C, would lie in the plane of these three points. But this is

202

PENTAHEDROIDS AND HYPERPYRAMIDS

[v.

i.

impossible, the intersections being in the interior of the

pentahedroid.*

The

point where these half -lines intersect

is

called the

centre of gravity of the pentahedroid.

corresponding edges equal. can make two pentahedroids Regular pentahedroids. correspond, the five vertices of one to the five vertices of

114.

Pentahedroids

with

We

the other, in

any

order, just as
72.

we have made two

tetra-

hedrons correspond in Art.

// each of the edges of a pentahedroid is equal to the corresponding edge of a second pentahedroid, when the five vertices of one are made to correspond in some order to the
the five vertices of the other,

THEOREM.

pentahedroids will be congruent

all

or symmetrical. PROOF. All the faces

and

the face angles, dihedral

angles, and hyperplane angles of one are equal to the corresponding parts of the other, as proved in the two theorems

and any two corresponding tetrahedrons can be made to coincide, even though they happen to be in a
of Art. 72
;

hyperplane in opposite orders (Art. 84, Th. 4). Then by putting two such tetrahedrons together we can prove that
the pentahedroids will coincide entirely if their orders are the same, and that they will be symmetrically situated
* The corresponding theorem for tetrahedrons is proved in the same way. The theorem which determines the centre of gravity of a triangle is usually made to depend on the axiom of parallels. The theorem is true, however, in the

Hyperbolic and Elliptic Geometries, and therefore is independent of the axiom of It can be proved very simply in these two geometries and in the Euclidean Geometry by means of trigonometrical formulae (see Chauvenet's TrigonomPhiladelphia, 1881, Part II, Art. 188, p. 253). Or, having proved the etry, gth ed
parallels.
,

theorem

for

Euclidean Geometry by means of parallels, we can prove

it

for

any

projection from the centre of the sphere upon the plane of the plane jtriangle which has the same vertices, and then in the Non-Euclidean Geometries for any plane triangle which can be inscribed in a circle, and so in a sphere,
spherical triangle

by

by reversing

this projection.

centre of gravity is used here without any reference to the physical properties of the point. Another name is centroid.

The term

II3-H5]

RIGHT AND REGULAR

203

with respect to the common hyperplane of these tetrahedrons if their orders are opposite.
the pentahedroids are symmetrical, corresponding tetrahedroidal angles and corresponding piano-polyhedral angles are symmetrical.

When

its

we take a regular tetrahedron and draw a line through centre perpendicular to its hyperplane, every point of this line will be equidistant from the four vertices of the
If

tetrahedron, and we can take a point at a distance from the four vertices equal to one of the edges of the tetrahedron.

We
all

have then a pentahedroid in which the ten edges are All the parts of any one kind, face angles, diequal.
etc.,

hedral angles, faces,

is

are equal;

for the

pentahedroid

congruent to itself in sixty different can be made to coincide with itself,

hedroid

with any other part of is called a regular pentahedroid (see Art. 166).

ways (Art. 95), and any part coinciding the same kind. Such a penta-

115. The terms right and regular as used of hyperpyramids and double pyramids. When the base of a hyperpyramid is the interior of a regular polyhedron, the interior of the segment consisting of the vertex and the centre of the base is called the axis of the hyperpyramid ; and when

the line containing the axis is perpendicular to the hyperplane of the base the hyperpyramid is regular.

hyperpyramid the lateral pyramids are equal regular pyramids. The axis of any one of these lateral pyramids is the hypothenuse of a right
i.

THEOREM

In

regular

triangle

whose

legs are the axis of the

hyperpyramid and a

radius of the sphere inscribed in the base.

The
of

slant height of a regular

hyperpyramid

is

the altitude

any one of the

lateral pyramids.

204

HYPERCONES AND DOUBLE CONES

the base of a double pyramid (see Art. 32)

[v.

n.

When

is

the

interior of a regular polygon, the interior of the triangle

determined by the vertex-edge and the centre of the base

is

called the axis-element of the double

pyramid; and when

the plane of this triangle is absolutely perpendicular to the plane of the base we have a right double pyramid. A
right double

pyramid

is isosceles

when

the extremities of

the vertex-edge are at the same distance from the plane of the base. Such a double pyramid is also called regular.

THEOREM

2.

In a

right double

pyramid

(the base being

regular) the lateral faces are congruent, the lateral cells are

In congruent, and the two end-pyramids are regular. double the are pyramid regular end-pyramids congruent.
II.

HYPERCONES AND DOUBLE CONES

A Spherical hypercones and right hypercones. cone is one is the whose base interior of a spherical hyper sphere.
116.

The

axis of a spherical hypercone is the interior of a consisting of the vertex and the centre of the base.

segment

A right

spherical hypercone, or simply a right hypercone, is one whose axis lies in a line perpendicular to the hyperplane of the

base.

section of a spherical hypercone by a hyperplane containing the vertex and any point of the base is a circular

cone.

THEOREM

i.

When a right

triangle takes all possible posi-

tions with one leg fixed, the vertices

and

the points of the other

two sides of the triangle make

up a right spherical hypercone.

an element, and

The fixed

side is the axis, the hypothenuse is

the other leg is

a radius of the base.

// in the hyperplane of a cone of revolution
its

THEOREM

2.

we pass a plane through

axis

and

rotate

around

this

plane

IIS-"?]

RIGHT DOUBLE CONE

205

which lies on one side of it, we shall have all of a right spherical hyper cone except that portion which makes up the section of the cone by the plane.
that portion of the cone

The slant height of a right spherical hypercone is the distance from the vertex to any point of the sphere whose interior is the base; it is the length of the hypothenuse of the right triangle of Th. i.
117.

Circular double cones

circular double cone is one

and right double cones. A whose base is the interior of a

circle. The axis-element of a circular double cone is the interior of the triangle determined by the vertex-edge and

the centre of the base.

right circular double cone, or

is

simply a right double cone,

base,

one whose axis-element

lies

in a plane absolutely perpendicular to the plane of the

and the double cone

is

also isosceles

when

the extremi-

the vertex-edge are at the same distance from the of the base. plane
ties of

THEOREM

i.

In a

right

double cone the elements are

congruent, and the two end-cones are cones of revolution. an isosceles right double cone the end-cones are congruent.

In

THEOREM

2.

right double cone

may

be generated by

a tetrahedron which has an edge and face in a perpendicular line and plane, the rotation taking place around
the rotation of

the latter.

perpendicular to the plane ABC. In the rotation around this will generate the set of elements of plane the face the double cone, the face will be the axis-element,

Let

A BCD

be the tetrahedron with the

line

CD

ABD
and

ABC

the faces

ACD

BCD

will generate the interiors of the

end-cones, and the edge except the centre.

CD

will generate all of the

base

206
118.

HYPERCONES AND DOUBLE CONES

[v.

n.

Hypersurfaces consisting of planes through a with point only this point common to any two of them. The conical hypersurface of double revolution. When
sists of lines (as, for

the directing-surface of a hyperconical hypersurface conexample, in the case of a conical sur-

face), the hypersurface consists of planes or portions of

planes, every point collinear with the vertex and a point of any one of these lines being a point of the hypersurface.

planes are not included in this way, we can consider the hypersurface which does consist of the entire planes, defining it as consisting of these planes and calling
If the entire

the planes elements.

can also form a hypersurface of planes through a O and the points of a plane curve, the planes being determined in some way so that they shall have only
fixed point

We

the point O common to any two of them and shall intersect the plane of the directing-curve only in the points of this For example, the planes can be isocline to a given curve.

plane through O. These various hypersurfaces should

all

be regarded as

hyperconical hypersurfaces (see foot-note, p. 220). We shall consider only the hypersurface generated by the rotation of one of two isocline planes around the other, the conical hypersurface of double revolution of Art. 112.

being the vertex (where the two planes meet), the plane through any other point of the hypersurface, absolutely perpendicular to the axis-plane, will intersect the hypersurface in a circle which can be taken for the directing-circle.

As we have seen, there are two sets of planes in the hypersurface, and two generating planes, and the hypersurface rotates on itself in any double rotation around the given
axis-plane and its absolutely perpendicular plane at O, or in any simple rotation around either one of these planes. Thus the hypersurface has a pair of axis-planes.

n8,

119]

THE HYPERSPHERE, SECTIONS

also

207

The hypersurface can

all

the half-lines

drawn from

be described as consisting of which make a given angle

from

with a given plane through O. All other half-lines drawn O are divided into two classes, those which make a

smaller angle with the given plane, and those which make a smaller angle with its absolutely perpendicular plane, than do the half-lines of the hypersurface itself. The hypersurface, therefore, divides all the remaining points of hyperspace into two classes, orue containing all the points except

of

one of the axis-planes, and the other containing

all

the points except

of the other axis-plane.

can pass around either axis-plane without passing through a point of the hypersurface, just as in a hyperplane we can pass around the axis-line of a conical surface of revolution without passing through
of the

We

a point of the surface. In the case of the hypersurface, however, each two regions into which it divides the rest of hyperspace is completely connected, so that in either one of these regions we can pass from any point to any other point without passing through a point

of the hypersurface ; while the axis of a conical surface of revolution in a hyperplane lies partly in one and partly in the other of two vertical regions that are completely separated.

III.

THE HYPERSPHERE
circles in

119.

Spheres and

a hypersphere.

Tangent

hyperplanes. A hypersphere consists of the points at a given distance from a given point. The terms centre,
radius, chord,

and diameter are used as with

circles

and

spheres.

THEOREM
is

i. Any hyperplane section of a hypersphere a sphere having for centre the projection of the centre of the

hypersphere upon the hyperplane (Art. 40, Th.

2).

the hyperplane passes through the centre of the hypersphere the section is a great sphere. Other spheres
of the hypersphere are small spheres.

When

208

THE HYPERSPHERE

[v.

m.

2. Four non-coplanar points of a hypersphere a determine sphere of the hypersphere, and three points not coplanar with the centre of the hypersphere determine a great

THEOREM

sphere.

THEOREM 3. Any plane having more than one point in a hypersphere intersects the hypersphere in a circle having
for centre the projection of the centre of the hypersphere
the plane.

upon

This is proved by considering the hyperplane which contains the plane and the centre of the hypersphere.
a hypersphere is a great circle when passes through the centre of the hypersphere.
circle of

its

plane

THEOREM
circle,

4.

and two points not

Three points of a hypersphere determine a collinear with the centre of the

circle.

hypersphere determine a great

THEOREM
intersect,

5.

Two

great circles

on

the

same

great sphere

and two

great circles

which

intersect lie

on one great

and Art. 20, Th. 2 (4) ). and a great sphere always intersect, intersecting in the extremities of a diameter, and two great spheres intersect in a great circle (Art. 27, Ths. i and 2).
sphere (Art. 22, Th.,

great circle

Distance in a hypersphere between two points not the extremities of a diameter is always measured on the arc
less than 180 of the great circle containing them. The distance between the extremities of any diameter is 180.

THEOREM

6.

All the circles of a hypersphere which pass

through a given point are perpendicular, that is, their tangents are perpendicular, to the radius of the hypersphere at this

These tangent lines, therefore, all lie in the hyperplane which is perpendicular to the radius at this point.
point.

A hyperplane perpendicular
at its extremity
is

to a radius of a hypersphere

tangent

to the

hypersphere.

i ig,

120]

THEOREMS OF VOLUMES

209

Spherical dihedral angles and spherical tetraheA great circle of a sphere, dividing the rest of the into two hemispheres, may be called the edge of sphere Two hemispheres of great either of these hemispheres.
120.

drons.

common edge form a a double convex lens, and enclose figure a portion of the hypersphere, a definite volume. Along the edge we have a spherical * dihedral angle, which we can
spheres in a hypersphere having a

somewhat

like

think of as consisting of a restricted portion of the edge and The tangent halfrestricted portions of the hemispheres. a which have common planes edge tangent to the edge of
the spherical dihedral angle form an ordinary dihedral angle whose measure can be taken as the measure of the
former.

THEOREM
measure

i.

spherical dihedral angle has the

same

at all points of its edge.

THEOREM 2. The volume enclosed by the hemispheres of a spherical dihedral angle is to the volume of the hypersphere as the dihedral angle is to four right dihedral angles.
The edge of a spherical dihedral angle has on each face a pole, and the arcs of great circles drawn through these poles from any point of the edge determine a spherical
angle by which the spherical dihedral angle can be measured, just as the dihedral angle formed by two half-planes is

measured by its plane angle. The spherical angle is itself measured by the distance between the two poles of the edge, so that this distance can be considered a measure of the spherical dihedral angle, and also, if we take corresponding units, as a measure of the volume enclosed by its
hemispheres.
It is hardly necessary to define spherical trihedral angle
*

We might have
is

said hyperspherical, but

we

shall use the shorter

word where

there

no ambiguity.

210

THE HYPERSPHERE
spherical
tetrahedron.

[v. ill.

and
than

We

shall

suppose

that

the

sides of a triangle
1

and the edges

of a tetrahedron are less

80.

The

four great spheres which contain the faces

of a spherical tetrahedron determine a set of sixteen tetrahedrons, eight pairs, the two tetrahedrons of a pair being

symmetrically situated with respect to the centre of the hypersphere, and therefore congruent (Art. 85, Th. 2). That
half-hypersphere which lies on one side of any one of the four great spheres (on one side of the hyperplane of the great sphere) contains the interiors of eight tetrahedrons,

one from each

pair.

spherical tetrahedron has six edges, each lying in the edge of a spherical dihedral angle whose interior contains the interior of the tetrahedron. The interior of one of these
spherical dihedral angles contains also the interiors of three of the fifteen tetrahedrons associated with the given tetra-

hedron as explained above, and its volume is equal to the sum of the volumes of the four tetrahedrons whose interiors
are within
it.

Writing

'

for the opposite point to

the other extrem-

ity of the diameter to A, and so for other points, we let denote the volume of the tetrahedron A BCD, Ti the volume

of
is

A 'B'CD, and so on. A BCD' congruent to A'B'CD, and we have r 34 = r ]2 etc. The interior of the dihedral angle C-AB-D contains the A 'BCD,
Ti 2 the volume of
,

whose volumes are T, TI, of the dihedral angle the measure T%, in terms of a right dihedral angle, and if we take for unit of volume one-sixteenth of the volume of the hyperinteriors of the four tetrahedrons

and TU.

If 612 is

AB

sphere,

we

shall

have the relation

T!

T+
There are

+T +T
2

12

six of these equations,

and

in addition

one equa-

tion expressing the fact that the

sum

of the eight different

120,121]

POLES AND POLAR CIRCLES

is

21

volumes
namely,

equal to the volume of a half-hypersphere,

r+2

Ti

+2

Ti 2

8.

These seven equations reduce to the following equivalent

system
:

T + T\ =2(612 + 0i3 + 0i4 2), etc., four equations, T Ti2 = 2(013 + 0H + 023 + 024 ~ 4), etc., three equations.
Given the volume of one tetrahedron, we can find the volumes of the others; but we have no simple formula for the volume of a single tetrahedron as we have for the
area of a spherical triangle.*

Poles and polar circles. Duality in the hypersphere. The diameter of a hypersphere perpendicular to the hyperplane of any sphere of the hypersphere is called
121.

the axis of the sphere, and the extremities of the axis are the
poles of the sphere.

THEOREM

i.

Each pole of a sphere of a hypersphere

all the

is

equidistant from

points of the sphere.

The plane through

is

perpendicular to the plane of

the centre of a hypersphere absolutely any circle of the hypersphere

called the axis-plane of the circle,

this plane intersects the

and the great

is

circle

iii

which

hypersphere

the polar

circle of the given circle.

THEOREM
hypersphere
circle.

2.

is equidistant

Each point of the polar circle of a circle of a from all the points of the given

THEOREM
polar between
is

3.

great circle

of a hypersphere is itself

to the great circle

which

is its polar,

and

the distance

any two a quadrant.

points, one in each of two polar great circles,

* See Coolidge, Non-Euclidean Geometry,

p. 181.

212

THE HYPERSPHERE

[v.

m.

4. great sphere contains all the points at a distance from either of its poles, and each oj two quadrant's polar great circles contains all points at a quadrant's distance

THEOREM

from

the other.
5.

THEOREM

Great circles which pass through the poles

of a sphere are perpendicular to the sphere, and any great circle perpendicular to a sphere passes through the poles of
the sphere.
If

a point moves a given distance along an arc of a great

polar great sphere will rotate around the polar and generate a spherical dihedral angle whose

circle, its

great circle

measure is this same distance. If a great circle rotates on a great sphere through a given spherical angle around one
of its points (and the opposite point), its polar great circle, lying on the polar great sphere of the given point and pass-

ing through the pole of the given great sphere, will rotate around the latter through the same angle. In fact, the
rotation takes place in the hypersphere around the great circle (see Art. 124) determined by the two fixed points, and
in hyperspace

around the plane of

this great circle.

We have a principle of duality in

the hypersphere, points

and great spheres corresponding to each other, and great circles to great circles, one great circle considered as made up of points and the other as common to a set of great
Corresponding figures are called reciprocal figures. This correspondence can be realized by taking the relations
spheres.
of pole

and polar as

its basis.

The student may

tetrahedrons.

investigate the properties of polar spherical

122. Geometry of the hypersphere as an independent three-dimensional geometry. Starting with its points and great circles, and with certain fundamental theorems, we

i2i, 122]

AN INDEPENDENT GEOMETRY

213

can build up the geometry of the hypersphere without

further reference to the hyperspace in which it lies. These fundamental theorems play the part of axioms and we shall
statue

them here

as axioms.*

The
axiom
:

of a pair being called opposites.

points of this geometry are paired, the two points express this as an

We

namely,
i
.

AXIOM
site point.

To each

point there is one and only one oppo-

With any two non-opposite points

determined by these two points. the following axioms
:

and

B is associated a
a great
circle

class of points as points of the great circle

AB,

For great

circles

we have

AXIOM

2.

The
Th.

great circle determined by

any two non-

opposite points of a given great circle is the given great circle

itself (Art. 10,
2).

AXIOM

3.

All great circles which contain a given point

its

contain also

opposite point.

In points of a great circle are in cyclical order. particular, any two pairs of opposite points separate each
other (Art.
6).

The

one great

spherical triangle consists of three points not points of circle and all points between any two of them on

that portion of a great circle which does not contain their

opposites.
circle containing one side of a spherical trithat angle, portion which lies between one vertex and the opposite point to the other will be called the side produced,

On

a great

produced in one direction or the other as the case

may

be.

AXIOM

4.

great circle intersecting

one side of a spheri-

cal triangle

and another

side produced intersects the third side.

* Compare the chapter on "Pure Spherics" in Halsted's Rational Geometry (chap. XVI of the second edition).

214

THE HYPERSPHERE

[v.

m.

of Art. 3 are included in Ax. 2, modified " the term non-opposite." Ax. 4 is the same as only by of Pasch the Axiom (Art. 7), with the restriction placed " side produced." It follows that in above on the phrase restricted the theorems thus of the hypersphere any portion

The two axioms

of the first four sections of chap. I all hold true, if in place " " collinear we use the phrase " on a great of the word " " " line circle with/' and for say great circle."

Certain other forms of expression, also, are changed as a matter of convenience. Thus we shall speak of a great
circle as

a transversal

to

two points
"

of the triangle,
to

a triangle when it contains at least and then we can speak of a point

a triangle where in chap. I we have said collinear with a triangle." Again, in place of the " we must now say great sphere. This we word " plane define as consisting of the points that we get if we take
as on a transversal

three points not points of one great circle,

great circle with any two of them, and

circle

all

points on a points on a great

all

with any two non-opposite points obtained by this

process.

We
is

prove then that any point of the great sphere

to the triangle

ABC

that any three not of one of a points great circle, degreat sphere, points termine the same great sphere.

on a transversal

ABC, and

Now

in order to confine

our geometry to a single hyper:

points of a hypersphere if we take four points not points of one great sphere, all points of
the great circles determined by

sphere we make this axiom AXIOM 5. We gel all the

of the great circles determined

any two of them, and all points by any two non-opposite points

obtained by this process.

Thus we see that the hypersphere is a space of three dimensions, and its geometry is a geometry of three dimensions.

122]

ITS

GEOMETRY DOUBLE ELLIPTIC

215

Finally,

we

and determine

are able to consider the entire hypersphere, its relation to the theory of parallels, by
:

introducing the following axiom

AXIOM
intersect.

6.

Any

two great circles of the same great sphere

and a point
Th.),

passing a great sphere through a given great circle of a given great sphere we prove from these axioms that the circle and sphere always intersect (Art. 22,

By

and then that any two great spheres

circle.

intersect in a

great

These " axioms/' and such axioms as are necessary to establish the properties of points on a great circle (see Arts. 5 and 6) determine the nature of the hyperspherical geomeThe result we express in the following theorem try.
, :

THEOREM. The geometry of the hypersphere is the same as the Double Elliptic Non-Eitclidean Geometry of Three
Dimensions*
the great circles

sphere being taken for lines

and great spheres of and planes.

III.

the hyper-

* See the author's Non-Euclidean Geometry, chap.

This book has been

;

very properly criticised for giving only the Single Elliptic Geometry but in any "restricted region" the two geometries are the same. In the Double Elliptic Geometry two lines in the same plane intersect in two points, and a line meets any plane in which it does not lie in two points, the distance
of the line.

each of these cases being one-half of the entire length is most conveniently taken as 2n. If we start at an intersection of two lines and follow one of them until we come again to the other, we shall come, not to the same intersection point, but to an "opposite" point. We have traversed only one-half of the line, and we arrive at
in

between the two points

The

length of the line

the starting point only when we have gone the same distance further. The geometry of the sphere is the same as the Double Elliptic Geometry of Two Dimensions.

In the Double Elliptic Geometry a line divides a plane in which

it lies,

and a plane

divides space of three dimensions, into two entirely separate parts, which is not the case with the Single Elliptic Geometry. Symmetrical figures in the Single Elliptic Geometry can be made to coincide by moving one of them along the entire length of

Double Elliptic Geometry a movement from intersection point two lines only puts a figure into the opposite region of space, and two symmetrical figures can never be made to coincide. Two polar lines are still everywhere at a quadrant's distance from each other, and the locus of points at a given distance from a given line is also the locus of points
a
line,

but

in the

to intersection point of

2l6
123.

THE HYPERSPHERE

[v.

m.

Point Geometry the same as the geometry of the

hypersphere.

THEOREM.
sphere is the

PROOF.

Th$ Point Geometry at the centre of a hypersame as the geometry of the hypersphere. In the hypersphere points, great circles, and

great spheres are its intersections with half-lines drawn from the centre and with planes and hyperplanes through

the centre, and the distances and angles in the hypersphere are the same as the corresponding angles at the centre.

Therefore the two geometries are the same.

In particular, to a great sphere of the hypersphere and

its

poles correspond at the centre a hyperplane and its perpendicular half -lines to two polar great circles correspond
;

two absolutely perpendicular planes; and two simply perpendicular planes correspond to two great circles intersecting at right angles.

The theorems
shall

of Point

Geometry

can, then, be stated

as theorems of the geometry of the hypersphere. mention only some of the more important results
(Arts. 67, 68, 106, 107,

We

Any two great

right angles.

circles

ular great circles,

and 109) have a pair of common perpendictwo polar circles which intersect them at
:

When two great circles cut out equal arcs on a polar pair of common perpendicular great circles, they have an infinite number of common perpendicular great circles, on all of
common perpendicular great a quadrant which they arcs not than the circles, greater of them and on its polar circle are equal. cut out on any one
at the complementary distance from its polar line, namely, a surface of double revo-

which they cut out the same circles have more than two

arc.

Conversely,

if

two great

two polar lines for axes. But a line intersecting the axes meets the surface in four points instead of two ; and a plane through one axis, and so perpendicular to the other, cuts the surface in two circles instead of one.
lution with the

123]

AND POINT GEOMETRY

217

There are two distances between two

pair of

great circles, the dis-

tances not greater than a quadrant measured along a polar

common perpendicular great circles. When the distances are equal the given circles are parallel
by Clifford.*
Parallel great circles, there-

in the sense used

fore, correspond to isocline planes of the Point Geometry. There are two senses in which great circles can be parallel,

perpendicular to both of two parallel great circles (which are not polar) are themselves parallel in the opposite sense. Through any point not a point of a
circles

and two great

given great

circle

nor a point of

its

polar great circle pass

two great circles parallel in the two senses to the given Two great circles parallel to a given circle and to its polar. same circle in the sense are parallel to each other in great and the set of all the great circles parallel this sense also to a given great circle in a given sense completely fills the hypersphere, one and only one such circle passing through
;

each point.

We

can prove in another

:

way

that the geometry of the

hypersphere and the Double Dimensions are the same f

Elliptic

Geometry

of

Three

is the same as the Point Point But Geometry Geometry is the same of our and therefore the geometry whatever theory parallels, of /the hypersphere is the same whatever our theory of par-

The geometry of

the hypersphere

at its centre.

allels,
.

Now in- the Double Elliptic Geometry of Four Dimensions

all the lines perpendicular to a hyperplane meet in a pair of opposite points, the poles of the hyperplane. This is true because any two of these lines lie in a plane (Art. 39, Th. 2)

in

which they are perpendicular to a line, and therefore they meet in two points each at a given distance on both
* See the author's Non-Euclidean Geometry, p. 68. t See the author's Non-Euclidean Geometry, pp. 63 and
26.

2l8
of them.

THE HYPERSPHERE
Thus the hyperplane
is

[v.

ra.

a particular case of a

hypersphere having either pole as centre. Any point is one of the poles of a hyperplane, and the Point Geometry
is

the same as the hyperplane geometry

Elliptic

it is the Double Non-Euclidean Geometry of Three Dimensions.

:

124.

double revolution.
is

The surface of Rotation of the hypersphere. Rotation of the hypersphere on itself

centre.

the same as the rotation of the Point Geometry at its In any simple rotation a certain great circle, the

axis of rotation, remains fixed in all of its points; while its polar great circle, the circle of rotation, rotates or slides

on

itself (Art. 81).*

Rotation around a great circle in the hypersphere is not a distorted rotation, such as we might have in the case of a flexible object rotated around a curved axis, but an actual hyperspace rotation

around the plane of the great

circle.

A double rotation is a combination of two simple rotations around two polar great circles. A double rotation can also
be regarded as a double rotation along the two polar great circles, or as a screw motion along either one of them. When the two rotations of a double rotation are equal it
is

a parallel motion, corresponding to an isocline rotation at the centre of the hypersphere. In a parallel motion all
great circles parallel to the circles of rotation in the sense of the rotation rotate on themselves, and the motion can be regarded as a parallel motion along any polar pair taken
hypersphere can be moved freely on itself. Therefore, without attempting we can say that a hypersphere is a space of constant curvature. If our space were a hypersphere in Euclidean space of four dimensions we should realize the Elliptic Geometry. The Elliptic Geometry is therefore
to define curvature in hyperspace,
*

sometimes supposed to assume that our space is a space of constant curvature like hypersphere, not a space of no curvature like a hyperplane. Elliptic Geometry of Three Dimensions, however, does not depend on any assumption of a Euclidean space of four dimensions. We might suppose our space to be an elliptic space lying in elliptic space of four dimensions and with no curvature whatever.

123, 124]

SURFACE OF DOUBLE REVOLUTION

219

from

this set of circles.

We

can also think of the motion

as a parallel motion with respect to the set of circles, without thinking of any particular pair as the circles of rotation

or as the axes of rotation (Art. 112).

THEOREM

i.

Any

position of a hypersphere can be ob-

tained from any other position with the same centre by a simple
or a double rotation (Art. 101, Th. i).

simple rotations are equivalent to a their axes intersingle simple rotation sect (Art. 102, Ths. i
2.

THEOREM

Two

when and only when and 2).

surface of double revolution consists of the points of a hypersphere at a given distance from a given great circle,

and so at the complementary distance from the polar great circle.* These two great circles are the axes and the two The surface is distances are the two radii of the surface. of covered with two sets parallel great circles, those of one set parallel in one sense and those of the other set parallel
in the other sense.

Through each point

of the surface

passes one and only one circle of each set, and the surface can be regarded as consisting of the circles of either one of
great sphere containing one of the axes intersects the surface in a meridian circle ; and any circle of
these sets.

Any
two

either of the

sets of parallel great circles

forms with the

meridian

an angle equal to the corany of the surface (see proof of Th. 2 of Art. responding radius The 107, or the author's Non- Euclidean Geometry p. 67).
circle at

point of it

meridian
circles

with respect to one axis are the parallel of the surface regarded as a surface of revolution
circles
;

around the other axis

*

and two great

circles,

one from each

See the author's Non-Euclidean Geometry, p. 68. The surface of double revolusomewhat like the anchor ring of the ordinary geometry. This surface is of importance in the theory of functions of two complex variables. See Poincare*,
tion
is

"

Sur

les residus

des integrates doubles,"

A eta Mathematica, vol. 9,

1886-1887, p. 359.

220
of the

THE HYPERSPHERE

[v.

m.

two sets of parallel great circles, form at either of their two points of intersection two pairs of vertical angles, the angles of one pair the double of one radius of the surface, and the angles of the other pair the double of the other
radius.

The

surface of double revolution

is

the intersection of the

hypersphere and a conical hypersurface of double revolution having its vertex at the centre of the hypersphere When the two radii of the surface are equal, (Art. 118).

each being equal to -> the

circles

of each set of parallel

great circles on the surface are perpendicular to the circles of the other set. They correspond to two conjugate series
of isocline planes at the centre (Art. 105).*
* have defined hyperconical hypersurface as consisting of lines through a point with a hyperplane directing-surface (Art. 33), and this is sufficient for any hypercone with a hyperplane base but unless we make provision for certain special cases, like that of the conical hypersurface of double revolution, we should get a
;

We

definition by making the hypersurface a "surface" of the three-dimenGeometry, or by defining it as consisting of the half-lines drawn from the centre of a hypersphere through the points of any hyperspherical surface, together

more general

sional Point

with the centre

itself.

CHAPTER
EUCLIDEAN GEOMETRY.

VI

FIGURES WITH PARALLEL

ELEMENTS
125.

The axiom

of parallels first introduced at this point.

The development of the preceding chapters has been made independent of the axiom of parallels. They may be called chapters in Pangeometry* We shall now make a study of
with parallel elements, confining In other ourselves, however, to the geometry of Euclid. words, we shall assume an axiom of parallels which it will be convenient to put in the following form
parallels
of figures
:

and

AXIOM. Through any point not a point of a given line passes one and only one line that lies in a plane with the given line and does not intersect it.
I.

PARALLELS

Parallel lines and parallel planes. Lines and are planes parallel to one another as in the ordinary geometwo when they lie in one plane and do not interlines try
126.
:

a line and a plane or two planes hyperplane and do not intersect.

sect,

when they
the

lie

in one

THEOREM THEOREM
THEOREM
to

i.

Two

lines perpendicular to

same hyone of two

per plane are parallel (see Art. 39, Th. 2).

2.

hyperplane perpendicular

to

parallel lines is perpendicular to the other.

// two planes through a point are parallel a given line they intersect in a parallel line.
3.
*

A title used by

Lobachevsky

in 1855.

222

PARALLELS
4.

[vi. i.

THEOREM
allel

// a hyperplane intersects one of two parthe

the

planes and does not contain it, other plane also and the two
y

hyperplane intersects

lines of intersection are

parallel.

the hyperplane of the parallel planes in a plane which intersects the parallel
intersects

For the hyperplane

lines.

planes in parallel

If a plane meets one of two parallel planes in a single pointy it will meet the other in a single point.
5.

THEOREM

THEOREM
THEOREM

6.

Two

planes absolutely perpendicular to a

third are parallel (see Art. 45, Th.).

7.

plane absolutely perpendicular

is absolutely

to

one of

two parallel planes

perpendicular
to

to the other.

THEOREM
to

8.

Two

planes parallel

a third are parallel

each other.

For a plane absolutely perpendicular to the third is absolutely perpendicular to the first two, and they are parallel by Th. 6.

THEOREM

9.

line, they all lie

// three parallel planes all intersect a given in one hyperplane.

THEOREM

10.

Two

planes

absolutely

perpendicular

to

two parallel planes are parallel, and two planes parallel respectively to two absolutely perpendicular planes are absolutely
perpendicular.

// two planes intersect in a line, planes through any point parallel to them intersect in a parallel line and form dihedral angles equal to the dihedral angles formed by
the two given planes.

THEOREM n.

PROOF.

The

parallel planes are parallel to the line of

intersection of the

two given planes, and therefore intersect in a parallel line, by Th. 3. Now a hyperplane perpendicular to these parallel lines (see Th. 2) cuts the planes in lines

126]

LINES

AND PLANES

223

which contain the sides of the plane angles of the various Cordihedral angles formed about the two parallel lines. and dihetherefore corresponding responding plane angles,
dral angles,* are equal.

COROLLARY.

// two

planes

are

perpendicular,

planes

through any point

parallel to

them are also perpendicular.

THEOREM
(Art. 69).

12.

// two planes have a point in common,

parallel planes through

any

other point

make

the

same angles

PROOF. Let a and ft be the two given planes having a point O in common, and let OL and ft' be planes through a second point Q' parallel respectively to a and ft. The planes through O' parallel to the common perpendicular
a. and ft are themselves common perpendicular a and ft' (Th. n, Cor.). On each of these common perpendicular planes the same angles are cut out as on the corresponding planes at O, since the intersection of any two

planes of
planes of

planes intersecting in a line at O' is parallel to the intersection of the parallel planes at O, and two intersecting lines
at O'
lie

in

a hyperplane with the parallel lines at 0, forming

angles equal to the angles formed

by the

latter.

COROLLARY.

plane isocline

to

is isocline to the other

and makes

the

one of two parallel planes same angle with both.

one

THEOREM common

13.

Two

lines not in the

same plane have only

perpendicular line (see Art. 62).

Since the two lines lie in a hyperplane this is always a theorem of geometry of three dimensions, and is proved
as in the text-books.

THEOREM

14.

// a line and plane do not

lie

in one hy-

perplane, they have only one common perpendicular line. See proof of Th. i of Art. 63.
faces are parallel half-planes lying in their hyperplane

Corresponding dihedral angles have corresponding faces. Two corresponding on the same side of the plane determined by the two parallel lines.

224
127.
lie

PARALLELS
Half-parallel planes.

[vi.

i.

Two

planes which do not

in one hyperplane

and do not

intersect are said to be

half-parallel or semi-parallel.

THEOREM THEOREM
the only lines

i.

The linear elements of two

to

half-parallel

planes are all parallel

2.

one another (see Art. 27).

lie

The linear elements which

in one of two

half-parallel planes are parallel to the other plane,

and

these are

which lie in one plane and are parallel

3.

to the other.

THEOREM

Through any point passes one and only

to

one hyperplane perpendicular

planes (see Art. 53).

each of two half-parallel

THEOREM
one

4.

Two

half-parallel planes have one

and only

common

perpendicular plane.
is one such plane, by Th. 3 of Art. 63. we have given a plane perpendicular to each

PROOF.
of

There

Suppose, then,

two half-parallel planes. It will intersect these planes in linear elements, the edges of various right dihedral angles, each with one face in the perpendicular plane and one in
one of the half-parallel planes.

perpendicular hyper-

plane intersects the planes in lines which contain the sides of the plane angles of these dihedral angles, that is, it intersects the perpendicular plane in the

common

perpenhalf-

dicular line of the lines in which

parallel planes.

it

intersects the

two

There

is

only one such

common

perpen-

dicular line, and the given plane is the plane determined as in Art. 63 by this common perpendicular line and the linear

elements which

it intersects.

THEOREM 5. The only common perpendicular lines of two half-parallel planes are those which lie in the common
perpendicular plane.

The

tween two half-parallel planes

perpendicular distance or simply the distance, beis the distance between

1271

HALF-PARALLEL PLANES

22$

the points where they are cut by a common perpendicular line. It is the same for all of these lines, since the common perpendicular plane cuts the given plane in
parallel lines (Th. i).

The perpendicular distance between two half-parallel planes is less than the distance measured along any line which intersects both and is not perpendicular to both. PROOF. The perpendicular distance between two elements lying one in each of the two given planes is the distance measured along some line lying in a perpendicular hyperplane, the distance measured along some line between the intersections of the given planes and this hyperplane. It is less than the distance between the two elements along any line which does not lie in a perpendicular hyperplane. But the intersections of the given plane and the perpendicular hyperplane have for common perpendicular only the line in this hyperplane which is perpendicular to the two
6.

THEOREM

given planes.

Therefore the perpendicular distance be-

tween the two given planes is the perpendicular distance between these two intersections, and is less than the distance between the two planes measured on any line that is not perpendicular to both.

THEOREM
tively to

Two planes through a point parallel respec7. two half-parallel planes intersect in a line which u

parallel to their linear elements.

For the line through the point parallel to the linear elements is parallel to the two given planes, and therefore lies in both of the two planes which are parallel to them
through the point.

THEOREM
in a line,
it

8.

planes intersects
Q

// a plane distinct from each of two parallel one in a line and does not intersect the other

will be half-parallel to the second.

226

PARALLELS
If the given

[vi.

i.

plane were in a hyperplane with the second parallel plane, this hyperplane, containing the line in which the given plane intersects the first parallel plane,

PROOF.

must be the hyperplane

of the parallel planes;

or

if

the

given plane intersected the second parallel plane in a point, it would lie entirely in the hyperplane of the parallel planes.

Thus, in either case, we should have a plane lying in the hyperplane of the two parallel planes, intersecting one in a

As the given plane line, and therefore the other in a line. does not intersect the second parallel plane in a line, it cannot lie in a hyperplane with it nor intersect it at all.

They must,

therefore, be half -parallel.

9.

// a plane perpendicular to one of two absolutely perpendicular planes does not contain their point
of intersection,
it

THEOREM

is half-parallel to the other.

Let a and a' be two absolutely perpendicular planes intersecting in a point O, and let /3 be a plane perpenThen cannot lie in dicular to a but not containing O.
PROOF.
a hyperplane with a', for such a hyperplane would intersect a only in a line through 0. Nor can /3 intersect a!

even in a point, for then it would contain the line through such a point perpendicular to a, and so contain the point 0.
j8

is

therefore half-parallel to a'.

Lines and planes parallel to a hyperplane. Parallel hyperplanes. A line and a hyperplane, a plane and a hy128.

perplane, or two hyperplanes, are parallel

when they do

not intersect.

THEOREM
is parallel to

i.

If a

line,

not a line of a given hyperplane,

line of the hyperplane, it is parallel to the

hyperplane; and if a plane, not a plane of a given hyperplane,

is parallel to

a plane of

the hyperplane, it is parallel to the

hyperplane.

127,

i2]
2.

HYPERPLANES
// a line
is parallel to

227

THEOREM
through
it

parallel to the intersection

of the

a hyperplane, it is hyperplane with any plane

or with

is parallel to

any hyperplane through it; and if a plane a hyperplane, it is parallel to the intersection of any hyperplane through
to
it.

the hyperplane with

THEOREM

3.

If a line is parallel

a hyperplane, a

line

parallel through any point of line lies wholly in the hyperplane; and if a plane is parallel to a hyperplane, a plane or line through any point of the hy-

the hyperplane

to the given

perplane parallel
plane.

to the

given plane lies wholly in the hyper-

THEOREM THEOREM
pendicular
to

4.

Two

hyperplanes perpendicular

to the

same

line are parallel.

5.

// one of two parallel hyperplanes is perline, the other is also

a
6.

perpendicular

to the line.

THEOREM

Through a point, not a point of a given hy-

perplane, can be passed one and only one parallel hyperplane.

In general, we can pass through a point a hyperplane parallel to a given hyperplane, to a given line and plane,
or to three given lines through a line, a hyperplane parallel to a given plane or to two given lines ; through a plane, a hyperplane parallel to a given line. In some cases, how;

ever, the construction will give us a hyperplane containing

some or

all

of the given figures,

and

in

some

cases

more

than one hyperplane can be obtained.

THEOREM
allel

7.

All the lines and planes in one of two parto the other,

hyperplanes are parallel

and

all the lines

lie

and planes through a

a parallel hyperplane.

point, parallel to

a hyperplane,

in

8. // a plane intersects two parallel hyperor a planes, if hyperplane intersects two parallel planes, the

THEOREM

228

PARALLELS

[vi. I.

lines of intersection are parallel;

sects

and if a hyperplane intertwo parallel hyperplanes, the planes of intersection are

parallel.

THEOREM
are

g.

// three non-coplanar lines through a point

parallel

to three other non-coplanar lines two sets of lines determine the same hyperthrough a point, the

respectively

plane or parallel hyperplanes; or if an intersecting line and plane are respectively parallel to another intersecting line and
plane, they determine the

same hyperplane

or parallel hy-

perplanes.

THEOREM
parallel each

10.
to

Two trihedral angles having their sides each and extending in the same direction *
equal, f

from their vertices are congruent. For the corresponding face angles are

THEOREM n.

If a line

is parallel to

a hyperplane,
the

all

points of the line are at the

same distance from

hyperplane;

or if a plane is parallel to a hyperplane, all points of the plane are at the same distance from the hyperplane.

THEOREM
equidistant.

12.

Two

parallel hyperplanes are everywhere

The student may prove the

Let

following theorem

ABCDE be a pentahedroid cut by a hyperplane a so that the edge AB lies on one side of a and the face CDE on the other side. Then if a is parallel to the line AB and to the plane CDE, the section
will

be a prism

if

is parallel

section will be a truncated prism

to the line but not to the plane, the if a is parallel to the plane but not
;

to the line, the section will be a frustum of a

pyramid

or

if

is

not

parallel to the line nor to the plane, the section will be pyramid (see Art. 31, Th. 3).
*
tie

a truncated

Two
The

parallel half-lines

on the same side of the

t

line

extend in the same direction when in their plane they determined by their extremities.

proof given in our text-books that the two dihedral angles are equal does
the' trihedral angles shall lie in

not require that

one hyperplane.

.
,

128, i2 9 ]

ISOCLINE PROJECTION

2 29

Projection from a plane upon an isocline plane produces similar figures.

129.

THEOREM
an
PROOF.

i.

Any

plane polygon and

its

projection

upon

isocline plane are similar.

be the point of intersection of the two Let A and B be any two points of the original polygon, and A' and B' their projections. 0-4-4' and OBB' are two right triangles with equal acute angles

Let

isocline planes.

at O (Art. 69). They are similar, and the sides OA and OB are proportional to the sides OA and OB'. The angle AOB is also equal to the angle A'OB (same refer1
f

Therefore the triangles and OA'B' are themence). if the triangles formed selves similar triangles. by in its joining the vertices of a plane polygon to a point plane are respectively similar to the triangles formed in the

OAB

Now

same way from another polygon, the two polygons are similar. Therefore, the given polygon and its projection
are similar.*

COROLLARY.

The projection of a
a
circle.

circle

upon a plane

iso-

cline to its plane is

Conversely, if a plane polygon is similar projection upon another plane the two planes are isocline or parallel.
2.

THEOREM

to its

PROOF.

The

planes (Art. 46) projecting a figure upon

one of two parallel planes project the same figure upon the other (Art. 126, Th. 7), and the two projections are equal.

We can then suppose that the plane of the projection passes

through a vertex of the given polygon, and complete our
proof by reversing the steps of the preceding proof and using the corollary to the second theorem of Art. 69.
It can be proved in any case of projection of a plane polygon upon another plane that the area of the projection is equal to the area of the original polygon multiand <' are the angles between the planes of the two plied by cos 4> cos ', where
<

figures.

230
II.

THE "HYPERPLANE AT INFINITY"

[vi.

n.

THE "HYPERPLANE AT INFINITY"

in

which expressions in regard to inwe sometimes speak of points, at and a lines, plane infinity, so now we can speak of planes at infinity in different hyperplanes, and of a hyperplane This we shall regard only as a matter of lanat infinity. guage. We introduce these expressions without introduc130.
finity

The sense

are used.

Just as

ing

any new elements

"

into our geometry or a

number "

in-

into our number-system. express certain facts finity of parallelism as if they were matters of intersection, from

We

which, indeed, they are derived by limiting processes. Thus, we say that two lines intersect at infinity only as another way of saying that they are parallel. These forms
of expression appear to simplify the conception of parallelism, and they enable us to generalize certain theorems of
intersection.

We

must be

careful to

remember, however,

are not really introducing nor a region to be called inany points, lines, and planes, finity, and that we are not really making any change in our

that from this point of view

we

conception of parallelism.* Points at infinity are sometimes called ideal points,

lines at infinity ideal lines,

and so on.
infinity,

We

shall give

a brief account of the geometry at

relation to infinity of certain figures studied in the following pages; but the rest of this book

and point out the

be entirely independent of the present section, and all reference to infinity can be omitted without disturbing the
will

continuity of our chain of theorems. f

*

It

would be perfectly legitimate to assume

points, lines, etc

at infinity,

and a

number infinity, just as we assume other points, lines, and numbers. Such a course may seem no more unreasonable than the assumptions which distinguish the Elliptic and Hyperbolic geometries from that of Euclid, or the assumption of a fourth dimension. But the point of view here presented is better for this geometry. See note by Professor Bdcher, Bulletin of the American Mathematical Society, vol. 5,
1898-1899, p, 182. t The elements at infinity are try of Veronese.

made the basis of many of

the proofs in the Geome-

130,131]

A SINGLE HYPERPLANE

231

all comprised in a single at infinity a has single point at infinity, its inhyperplane. A plane has a line at tersection with any parallel line. intersection its with any parallel plane. The line infinity,

131.

The elements

line

any plane is made up of the points at infinity of its lines, and is determined by any two of them. In space of three dimensions, and so in any hyperplane, we have a plane at infinity, the intersection of the hyperat infinity of

plane with any parallel hyperplane. The plane at infinity of any hyperplane consists of the points at infinity of its
lines,

and contains the

lines at infinity of its planes.

Or,

we can say that a plane at infinity consists of the points that we get if we take any three non-collinear points at
infinity, all points collinear

with any two of them, and points collinear with any two obtained by this process. Lines which intersect a given hyperplane, but do not
it,

all

lie

in

have points at

infinity

which do not

at infinity of this hyperplane. That is, dimensions the points at infinity are not all points of a Thus we have a hyperplane at single plane at infinity.
infinity.

in the plane in space of four

lie

if

hyperplane at infinity consists of the points that we get four non-coplanar points at infinity, all points collinear with any two of them, and all points collinear

we take

with any two obtained by this process.

THEOREM.
sions
lie

All points at infinity in space of four dimenin a single hyperplane.

Through a given point O can be drawn a line Therefore, all points at infinity parallel to any given line. are the points at infinity of the lines through O. Through any point of hyperspace passes a line that goes through O and has a point at infinity.
PROOF.

Now we

can get

all

the lines through

O by

taking four

232

THE "HYPERPLANE AT INFINITY"

[vi.

n.

which are not in one hyperplane, all coplanar with any two of them, and all coplanar with any two obBut tained by this process (see last paragraph of Art. 64).
of these lines

any four non-coplanar points

infinity of four lines through

at infinity are the points at not in a hyperplane, and if

its

a line through
infinity
is

O
all

is

coplanar with two others,

collinear with their points at infinity.

point at There-

fore

we can

get

the points at infinity

all

by taking

four non-

coplanar points at infinity,

of them,

points collinear with any two

and

all

points collinear with

all

any two obtained by

lie

this process.

Hence

the points at infinity

in

a single

hyperplane.
132.

Distance and angle at

infinity.

We
O

can define

distance between two points at infinity as proportional to which is not the angle which they subtend at any point

There are two supplementary angles infinity. at 0, and so two distances between two points at infinity. We shall usually mean the smaller of these two distances when they are not equal. Taking the entire length of a
a point at
line at infinity as equal to

at infinity the

TT, we shall give to all distances same measures as to the corresponding angles

atO.
a plane at infinity are the lines at infinity in a hyperplane, and always inin a tersect, intersecting single point. They divide the rest of their plane into two separate regions, each a lunelines in

Two

of

two non-parallel planes

shaped region with the extremities coinciding and the two

angles actually vertical angles. the same as an entire line, but

A
"

half-line at infinity

is

it will

be more convenient

here to regard
the half-line.
half-lines

it

as a

"

restricted
its

in one direction from one of

portion of a line, lying points, the extremity of

An

angle will then be formed

by two such

with a

common

extremity lying along different

131-133]

ANGLES AT INFINITY

233

lines through the points of such an angle and not at infinity will determine a definite pair of vertical dihedral angles, and the measure of the angle at
lines.

The

a point

infinity will

be the same as the measure of either of these

dihedral angles, the

sum

of the angles
2
TT.

around a point in

a plane being always equal to

planes at infinity intersect in a line, and divide the rest of the hyperplane at infinity into two regions each some-

Two

what

like

a double convex

lens.*

Along the edge

of such

region
finity.

we have what may be

called a dihedral angle at in-

We
"

stricted

" can think of a half-plane at infinity as a reportion of a plane lying on one side of a line, the

edge of the half-plane, and think of a dihedral angle at infinity as formed by two such half-planes with a common

edge but not lying in the same plane (compare Art. 120).

To the dihedral angle will then correspond a hyperplane angle

whose measure can be taken as the measure of the former. Since two lines in a plane at infinity intersect in a single point, the geometry at infinity is the Single Elliptic GeomeThis geometry is, indeed, the try (see foot-note, p. 215). same as the Point Geometry of the half-lines from a point
O, except that two opposite half-lines correspond to a single point at infinity. Thus all the theorems of Point Geometry

can be stated as theorems of the geometry at infinity, just as they can be stated as theorems of the geometry of the
hypersphere (Art.
133.

Some
and

generalizations

now made

possible.

In-

tersection
*

perpendicularity.
is

We
two

will

now show how

The volume

of one of these regions

to the entire volume of the hyperplane at

right dihedral angles (not to
is

infinity as the corresponding dihedral angle is to

four right dihedral angles, because the geometry at inanity

the Single Elliptic

Geometry).
t

We could

for each of

have assumed that a line has two points at infinity, a separate point two opposite half -lines. The geometry at infinity would then be the
Geometry.
See Veronese, Grundzugc, Ft.
I,

Double

Elliptic

Bk.

I.

234

THE "HYPERPLANE AT INFINITY"

[vi.

n.

the forms of expression introduced in this section enable us to generalize certain theorems of geometry
:

Two points determine a line, three non-collinear points determine a plane, and four non-coplanar points determine a hyperplane, even when some or all of the points are at
infinity.

Two

lines in a

which intersect

lie

plane always intersect, and any two lines A line and a plane in a hyin a plane.

perplane intersect, and any line and plane which intersect Thus a line and plane which do not lie in a hyperplane.

do not intersect even at infinity. Two planes which lie in a hyperplane intersect in a line. Two planes which do not lie in a hyperplane intersect in a point when the point is at infinity the planes are halflie

in a hyperplane

parallel.

in a point

and a hyperplane always intersect, intersecting a plane and a hyperplane always intersect, inand two hyperplanes always intersect, tersecting in a line
line
; ;

intersecting in a plane.
is the intersection of a hyperthe hyperplane at infinity. A polyhedral plane angle by angle at infinity is the intersection of a piano-polyhedral angle by the hyperplane at infinity. The parts of the

dihedral angle at infinity

angle have the same relation to the pianopolyhedral angle as do those of any right section (Art. 75).

polyhedral

Two

perpendicular

lines,

not lines at

infinity, are lines

whose points at
each other.

infinity are at a quadrant's distance from line and plane are perpendicular when their

point and line at infinity are pole and polar in the plane at Two planes are absolutely infinity of their hyperplane.

perpendicular

when

their lines at infinity are polar lines.

When
it is

line at infinity intersects

both of two polar

lines at infinity

lines

perpendicular to both.

Any two

have

133, i34l

PRISMOIDAL HYPERSURFACES

235

two common perpendicular lines, and the distances between two lines are measured on these common perpendicular
lines.
III.

HYPERPRISMS
Secof

134.
tions.

Their interiors. Prismoidal hypersurfaces. Axes. A prismoidal hypersurface consists

system polyhedron but not lying in the hyperplane of the polyhedron.

of parallel lines passing through the points of a given

The polyhedron

is

called the directing-polyhedron,

the parallel lines are the elements, and the elements which shall pass through the vertices are lateral edges.

We

assume that the directing-polyhedron

is

a simple convex

polyhedron. The elements which pass through the points of a face of the directing-polyhedron constitute the interior of a pris-

matic surface and a

cell

of the hypersurface.

The elements

which pass through the points of an edge of the directingpolyhedron constitute what may be called a strip, which is that portion of a plane that lies between two parallel
lines.

This strip

is

a face of the hypersurface, the

surfaces.

common

face of

two adjacent prismatic

The
lines

interior of a prismoidal hypersurface consists of the

directing-polyhedron

which pass through the points of the interior of the and are parallel to the elements.
the interior of

The hypersurface being convex,

whose points are points

any segment
lie

of the hypersurface will

it lies

in the interior of the hypersurface unless

entirely entirely in

the hypersurface itself, and a half-line drawn from a point of the interior and not parallel to the elements will intersect the hypersurface in

one and only one point.

hyperplane passing through a point of a prismoidal hypersurface and parallel to the elements intersects the hypersurface in a prismatic surface.
i.

THEOREM

the interior of

236

HYPERPRISMS

[vi.

m.

For the hyperplane intersects the directing-polyhedron in a convex polygon, and intersects the hypersurface in the elements which pass through the points of this polygon.

THEOREM

2.

hyperplane which

is not parallel to the

elements of a prismoidal hypersurface intersects the hypersurface in a polyhedron, and any such polyhedron can be

taken as directing-polyhedron.

Each element
of hyperspace

of the hypersurface, and, indeed, each line

which is parallel to the elements, meets We have a correspondence the hyperplane in a point. of the between the points hyperplane and the lines which
are parallel to the elements of the hypersurface, as also between the points of the intersection and the points of

Each cell of the hyperthe given directing-polyhedron. surface intersects the given hyperplane in a face, each face
of the hypersurface in

an edge, and each

lateral

edge in a

vertex, of the given intersection. right section is a directing-polyhedron

whose hyperplane

is

perpendicular to the elements.

THEOREM

3.

Directing- polyhedrons lying in parallel hy-

perplanes are congruent, and any two homologous points of two such polyhedrons lie in a line parallel to the elements.

// a prismoidal hypersurface has a parallelopiped for directing- poly hedr on, it will have three pairs of equal opposite lateral cells lying in parallel hyperplanes, and
4.
all

THEOREM

directing-polyhedrons will be parallelopipeds. PROOF. Any two opposite faces of the given parallelepiped
of
its

are equal parallelograms lying in parallel planes. They of surfaces are, then, directing-polygons equal prismatic

lying in parallel hyperplanes (Art. 128, Th. 9). Therefore, any directing-polyhedron will have three pairs of parallel

opposite faces (Art. 128, Th. 8), and will be a parallelepiped.

134, i3Sl

THE PARTS OF A HYPERPRISM

237

5. // any directing-polydedron of a prismoidal a centre of symmetry, the line through this has hypersurface point parallel to the elements is an axis of symmetry, meeting the hyperplane of every directing-polyhedron in a point which

THEOREM

a centre of symmetry of this polyhedron. Each point of the line is, in fact, a centre of symmetry for the entire hypersurface, and the line as a whole is a line of symmetry.
is

For

this line lies

mid-way between the two

it

any plane containing

line intersecting it

which intersects the hypersurface, and any

lines in
it

determines with
Interior
of

such a plane.

A a hyperprism. hyperprism consists of that portion of a prismoidal hyper135.

Hyperprisms.
lies

surface which

between two parallel directing-polyhe-

drons, together with the directing-polyhedrons themselves

and

their interiors.

The
whose

interiors of the directing-polyhedrons are the bases.

In each hyperplane of the hypersurface we have a prism interior is one of the lateral cells of the hyperprism.
lateral faces

The

and edges

of these prisms are the lateral

faces and lateral edges of the hyperprism. The lateral edges are all equal; the bases are congruent (Art. 134, Th. 3). The interior of a hyperprism consists of that portion of

the interior of the prismoidal hypersurface which

lies

be-

tween the bases.

interior of

The hypersurface being convex, the

of the hy-

any segment whose points are points

lie

perprism will
unless
line
it lies

entirely in the interior of the hyperprism entirely in the hyperprism itself, and a halfwill intersect the

drawn from a point of the interior hyperprism in one and only one point.

hyperprism

is

a right hyperprism

when

the lateral

edges are perpendicular to the hyperplanes of the bases. When also the bases are the interiors of regular polyhedrons the hyperprism is regular.

HYPERPRISMS

[vi. ra.

We

can cut apart a hyperprism, cutting

it

along

its faces, suffi-

can do this, for ciently to spread it out into a single hyperplane. example, so that the lateral prisms shall remain attached to one base, while the other base rests upon one of them. The figure below represents a tetrahedroidal hyperprism cut apart and spread out in this way. The four prisms rest upon the four faces of a tetrahedron, and a second tetrahedron equal to the first (symmetrical in this hyper-

We

plane) rests on the other end of one of these prisms. Now in hyperspace we can turn these prisms around the faces of the tetrahedron

upon which they

rest

away from the hyperplane

of this tetrahedron,

and the other tetrahedron around the face by which it is attached to one of the prisms: we can do this without separating any of the figures or distorting them in any way, until we bring them all together, each prism with a lateral face resting upon a lateral face of each of
ing

the others, and each of the four faces of the second tetrahedron restupon one of the prisms. The figure will then enclose completely

a portion of hyperspace (see note in Art. 32).

The student may investigate the conditions necessary in order that three given prisms and two tetrahedrons may be the lateral prisms and tetrahedrons of a tetrahedroidal hyperprism.

i3S i3fl

HYPERPARALLELOPIPEDS
Special forms of hyperprisms.

239
Hyperparallelo-

136.
pipeds.

hyperprism whose bases are the interiors of prisms can be regarded in two ways as a hyperprism of this kind for the lateral prisms correspond;

The hypercube.

ing to the ends of the bases are parallel (Art. 128, Th. 9) and congruent, and the remaining lateral prisms are parallelepipeds, which can be regarded as having their bases on this second pair of prisms and their lateral edges those edges
is

which belong also to the first pair of prisms. This figure a particular case of a double prism, and will be studied

in the next section (see Art. 144).

A hyperparallelepiped is a hyperprism whose bases are the interiors of parallelepipeds. In a hyperparallelopiped

there are four pairs of opposite equal parallel parallelepipeds whose interiors are the cells, and the interiors of any
pair can be taken as bases.

There are four

sets of eight

parallel edges, each set joining the vertices of two opposite cells, becoming the lateral edges when these cells are taken

section of a hyperparallelopiped made by a hyperplane intersecting all eight of the edges of a set will be

as bases.

The

a parallelepiped (Art. 134, Th.

4).

THEOREM
sect

The diagonals of a hyperparallelopiped bione another, passing through a point which is a centre
i.

all

of

symmetry for

the hyperparallelopiped.

right hyperparallelopiped
is

whose base

is

a rectangular

parallelepiped

a rectangular hyperparallelopiped.

The

edges which meet at any vertex lie in the lines of a rectanguThe lengths of these four edges are lar system (Art. 48).
the dimensions of the hyperparallelopiped.

THEOREM
squares of the

2.

The square of

the length of
to

a diagonal of a
the

rectangular hyperparallelopiped 4s equal

sum

oj the

four dimensions.

240

HYPERPRISMS

[vi.

in.

A
base

is

hypercube is a rectangular hyperparallelopiped whose the interior of a cube and whose altitude is equal to
;

the edge of the cube

equal.
1 66)
:

that
is

is,

its

four dimensions are

all

The hypercube
it

a regular polyhedron (see Art.

has eight equal cubical cells, twenty-four equal faces each a common face of two cubes, thirty- two equal
edges,
lines

and sixteen vertices. There are four axes, lying in which also form a rectangular system.
3.

THEOREM

The diagonal of a hypercube

is

twice as

long as the edge.

Six squares which can be folded so as to form a cube.

Eight cubes which can be folded so as to form a hypercube.

If we place six equal cubes upon the six faces of a cube, and one more outside of one of these, just as we put together four prisms and two tetrahedrons in the note in Art. 135, we can turn these cubes around the faces upon which they rest and bring them together so

as to form a hypercube.

This

is

analogous to the process of forming

a cube by folding

six squares together.

136-138]

PLANO-PRISMATIC HYPERSURFACES

241

There are two ways of projecting a hypercube that will both assist us in forming some conception of it. One is by an oblique projection, and the other is by projection from a point at a little distance
from the hypercube in the
the
first

line of

one of

its axes.

We

can think of

as representing the appearance of the hypercube when we stand a little to one side, and the second as we look down into it. In

each case we can pick out the projections of the eight cubes whose interiors form the cells of the hypercube. The hypercube has become one of the most familiar of the figures
of hyperspace.

The reader

will find it

mentioned in nearly

all

popu-

lar descriptions of the fourth dimension.

See, for example, Fourth

Dimension Simply Explained, pp. 46,

137.

72, 88, 92,

and

113.

Relation of the prismoidal hypersurface to infinity. Taking the point of view explained in Art. 130, we can say that a prismoidal hypersurface is a polyhedroidal angle

whose vertex

is

at infinity.

eight hyperplanes of a hyperparallelopiped intersect the hyperplane at infinity in four planes which contain

The

the faces of a tetrahedron.

is

Each vertex

of this tetrahedron

the vertex of one of the prismoidal hypersurfaces connected with the hyperparallelopiped, and the opposite face
is

of the tetrahedron

the plane of intersection of the hyper* (see Art. 145). planes of the corresponding pair of bases

IV.

DOUBLE PRISMS

Sections. Piano-prismatic hypersurfaces. shall use the word layer to denote that portion of a hyperplane which lies between two parallel planes, and call the
138.
parallel planes the faces of the layer. piano-prismatic hypersurface consists of a finite

We

of parallel planes taken in a definite cyclical order, layers which lie between consecutive planes of this order.

number and the

* There are, indeed, four tetrahedrons at infinity, each of which has itS vertices at these four points and its faces in these four planes, but we do not need to consider

them

here.

242

DOUBLE PRISMS
parallel planes are faces
If a,
/3,

[vi. IV.

The
cells

and the layers are

a/3,

cells of

the

hypersurface.

7,

...

are the faces in order, the

187,
.
.

can be described as the layers

and

the hypersurface as the piano-prismatic hypersurface a(3y .... The faces and all parallel planes within the layers
are the elements of the hypersurface, and are in cyclical

order (Art.

6).

The hypersurface is a simple piano-prismatic when no plane occurs twice as an element.

when,
also, the

hypersurface It is convex

hyperplane of each

cell

contains no element

except those of this cell and the two which are its faces. We shall consider only hypersurfaces which are simple and convex.

THEOREM

i.

hyperplane containing an element of a

piano-prismatic hypersurface or any parallel plane will intersect the hypersurface in elements if at all (Art. 128, Th. i).

hyperplane intersecting but not containing an element of a piano-prismatic hypersurface will intersect the hypersurface in a prismatic surface.
2.

THEOREM

For the hyperplane intersects all the elements in parallel lines (Art. 126, Th. 4), and so the cells in strips (see Art.
134),

which are the faces

of a prismatic surface.

Thus the hypersurface can be regarded

of a directing-polyhedron.

as a prismoidal

hypersurface, with a directing prismatic surface instead

THEOREM

3.

Two

parallel prismatic sections of a piano-

prismatic hypersurface are congruent. .PROOF. The strips in which the hyperplanes of the two parallel sections intersect any one of the cells of the

hypersurface
fore of the

lie

in parallel planes (Art. 128, Th. 8)

cell),

tween parallel planes (the faces of the

same

width.

They

lie

also

and beand are thereon the same side of

i3,

i39l

DIRECTING- POLYGONS

243

do in the same layer. Therefore corresponding dihedral angles of the two prismatic sections are equal (Art. 126, Th. n; see also foot-note, p. 223), and the sections themselves are
either pair of corresponding edges, lying as they

congruent.
139.
face.

Directing-polygons of a piano-prismatic hypersurInterior of the hypersurface.

i.

THEOREM

plane which contains a line parallel

to

the elements of a piano-prismatic hypersurface, but is not

itself parallel to

them, will intersect in a line every element

which

it

intersects at all.
2.

A plane which does not contain a line elements of a piano-prismatic hypersurface parallel will intersect every element in a point, and will intersect the
THEOREM
to

the

hypersurface, the latter being convex, in a convex polygon. PROOF. The plane will intersect each element in

In particular, it will interpoint, by Th. 5 of Art. 126. sect the faces in points which are the vertices of a polygon,

and the layers

is

in the sides of this polygon.

The polygon

sect the

a simple convex polygon, since no two elements can intersame plane in the same point.
in

The polygon

which a plane containing no

line parallel

to the elements intersects the hypersurface can

be called

a directing-polygon, and the hypersurface can be described as consisting of a system of parallel planes passing through
the points of a given polygon and intersecting the plane of the polygon only in these points. We can think of the polygon as going around the hypersurface.

THEOREM

3.

the elements of

Through any line which is not parallel to a piano- prismatic hypersurface can be passed

planes intersecting the hypersurface in directing-polygons. PROOF. Through a point A of the line pass a plane

244
parallel to

DOUBLE PRISMS
the elements.

[vi. iv.

The

line

intersects the plane

only in this point,

is

hyperplane. not a point of the hyperplane will intersect the hyperplane only in the given line, and the plane which we have drawn parallel to the elements only in the point A it will
:

and plane determine a Arty plane through the line and a point which
line

and the

be the plane of a directing-polygon.

THEOREM
points

4.

Two

parallel directing- polygons of a piano-

prismatic hypersurface are congruent, and any two fwmologous lie in one of the elements.

In

fact, their interiors are

lateral surface is

the bases of a prism whose cut out from the prismatic surface in

which their hyperplane intersects the hypersurface.

If a piano-prismatic hypersurface has a parallelogram for directing-polygon, it will ham two pairs of equal opposite cells (layers of the same width) lying in
5.

THEOREM

parallel hyper planes,

and

all of its directing- polygons will be

9).

parallelograms (see Art. 128, Ths. 8 and

We

can study the properties of these hypersurfaces as a

part of the two-dimensional geometry' whose elements are a set of parallel planes. The theorems in chap. I which
relate to triangles

and convex polygons can

all

be inter-

preted as theorems concerning these hypersurfaces (compare this with Arts. 78 and 112). But we can also prove
these theorems

by means

of a directing-polygon,

making

planes correspond to the points where they intersect the plane of the polygon. Thus we have the interior of the hypersurface as consisting
all parallel

the elements and

of those planes

which correspond to the interior of the

directing-polygon. The interior of any segment whose points are points of the hypersurface will lie entirely in

the interior of the hypersurface unless it lies in the hypersurface itself, and a half-line drawn from a point of the

139, 140]

DIRECTING-POLYGONS
and not

245

interior

parallel to the elements will intersect the

one and only one point. Also, that portion of a plane between two parallel lines of the hypersurface (a strip, Art. 134), or a layer between any two elements,
hypersurface in
lies

entirely in the interior unless

it

lies

entirely in the

hypersurface.

THEOREM

6.

When

the layer between two elements of

lies entirely

piano-prismatic hypersurface

in the interior of

the hypersurface, it separates the rest of the interior into two

portions lying on opposite sides of the hyperplane, and with its faces and each of the two parts into which they separate
the rest of the hypersurface
it

forms a convex piano-prismatic

hypersurface (see Art. 15, Ths. 2

and

3).

COROLLARY.
in

By

taking the diagonal layers which have

common one
set of

of the lateral faces of the hypersurface

we can
the

form a

triangular piano-prismatic hypersurfaces, their

layers,

interiors,

together with the diagonal

making up

interior of the given hypersurface.

140. Right direc ting-polygons. Axis-planes. A directing-polygon whose plane is absolutely perpendicular to the planes of the elements of a piano-prismatic hypersurface is called a right directing-polygon.*

THEOREM
upon

i.

The projection of any directing-polygon

a right directing-polygon
is the right direct-

the plane of

ing-polygon itself (see Art. 46).

2. plane isocline to the plane of a right direc tnot but ing-polygon, parallel to the elements, intersects the

THEOREM

hypersurface in a polygon similar

to the right

directing-poly-

gon;
*

and, conversely, any

will

directing-polygon similar to a

We

not

call it

two parts

of the hypersurface, nor

a right section, since it does not completely separate any even of a restricted portion of the hypersurface

(see foot-note, p, 65).

246

DOUBLE PRISMS
which
i

[vi. iv.

right directing-polygon lies in a plane

is isocline to the

plane of the latter (see Art. 129,

directing-polygon is the

Ths.

and

2).

The

right

minimum of all these similar polygons.

THEOREM 3. // any directing-polygon of a piano- prismatic hypersurface has a centre of symmetry, the plane
through this point parallel
to the

elements is

an axis-plane of
a

a centre of symmetry of this polygon. Each point of the plane is, in fact, a centre of symmetry for the entire hypersurface, every line of the plane is a line of symis

symmetry, point which

meeting the plane of every directing-polygon in

metry,

and

the plane as

a whole

is

For the plane lies mid- way which any hyperplane containing it intersects the hypersurface, and any line intersecting it but not lying in it
determines with
141.
it

a plane of symmetry. between the two planes in

such a hyperplane.

The

Intersection of two piano-prismatic hypersurfaces. two sets of prisms. When the elements of a piano-

prismatic hypersurface intersect the elements of a second piano-prismatic hypersurface only in points, the intersection of the

two hypersurfaces

consists of the lateral sur-

faces of a set of prisms joined together in succession

their bases,

by

together with the polygons whose interiors In another way, also, the same intersecare these bases.
tion consists of the lateral surfaces of a set of prisms joined

together

by

their bases, together with the polygons

whose

interiors are these bases.

hypersurface are parallel planes intersecting the second hypersurface in a set of equal parallel directing-polygons of the latter, and the cells of
fact,
first

In

the faces of the

the

first are layers, each layer intersecting the second in the lateral surface of a prism whose bases are interiors of two of these directing-polygons (see Art. 139, Th. 4).

In the same

way

the faces of the second hypersurface

i 4 o,

i4i]

TWO HVPERSURFACES
first

247

are parallel planes intersecting the

set of parallel directing-polygons,

and each

hypersurface in a cell of the second

intersects the first in the lateral surface of a prism whose bases are the interiors of two of these directing-polygons.

Thus the

entire intersection consists in

two ways of the

lateral surfaces of a set of prisms joined in succession

by

their bases, together

with the polygons whose interiors are

these bases.

The

lateral faces of

interiors of parallelograms,

any prism of the first set are the and are the intersections of one
hypersurface with the different set of corresponding faces of these
are, then, the intersections

particular cell of the cells of the second.

first

prisms, one from each prism,

of the different cells of the first hypersurface

with one

particular cell of the second. Thus the faces of any particular prism of either set form a set of corresponding faces of the different prisms of the other set, and every lateral
face of a prism of one set
is

a lateral face of some prism of prisms of one set

the other

set.

The

lateral edges of the

are the sides of the bases of the prisms of the other set, and the prisms of one set can be said to be joined crosswise by
their lateral faces to the prisms of the other set.
teriors

The

in-

and the bases of the prisms of the first set lie in the first hypersurface and in the interior of the second, and the interiors and bases of the prisms of the second set lie in the second hypersurface and in the interior of the first. We can think of the first set of prisms as going around the first
of

hypersurface, while any base or interior of a cross section any of these prisms is a piece cut out of an element of the first hypersurface. In the same way the second set

of prisms goes

around the second hypersurface.

double prism consists of the intersection of two piano-

prismatic hypersurfaces whose elements intersect only in

24$

DOUBLE PRISMS

[vi. iv.

points, together with all that portion of each hypersurface which lies in the interior of the other ; that is, it consists
of all of the prisms of
interiors.

both sets described above and their

the elements of one hypersurface are absolutely perpendicular to the elements of the other the double prism
is

When

a right double prism. When also the prisms of the two sets are regular the double prism is regular.
C

In a right double prism the prisms of each set can be put together in one hyperplane so as to form a single right prism with lateral edge equal to the perimeter of the base of the prisms of the other set. In
forming the double prism these two single prisms are folded towards each other in such a way that the upper base of each occupies the same position
as
its

lower base and the lateral surface of one

coincides as a whole with the lateral surface of

the other.
in this

Thus one
until the

way

single prism alone, folded ends come together, gives

us the other prism and the entire figure. When we have an oblique double prism and
spread out the prisms of one set in a single hyperplane, we have, not a prism, but a set of
prisms with equal bases resting upon one another somewhat like a broken column. The upper

base of the highest prism, however, and the

i 4 i, 142]

THE INTERIOR

249

lower base of the lowest prism will lie in parallel planes and will be so placed that each is the projection of the other.

142.

Interior of a double prism.

The

directing-polygons.

The
are

interior of

common

a double prism consists of the points which A to the interiors of its two hypersurfaces.

plane lying in the interior of one of the hypersurfaces parallel to its elements intersects the other in a directing-polygon

whose

interior belongs to the interiors of both,

and so to

the interior of

the double prism.

The

vertices of this

polygon are a set of corresponding points of the bases, and the sides lie in the interiors, of the prisms of the set which goes around the former hypersurface.

THEOREM

i.

Any

of a convex double prism

plane through a point of the interior intersects the double prism in a

convex polygon.

PROOF. The plane will intersect each hypersurface in a convex polygon, in two parallel lines, or not at all (Art. When the plane intersects each hyper139, Ths. i and 2). surface in a pair of parallel lines the intersection with the double prism will be a parallelogram. In all other cases there will be at least one convex polygon, and by applying

one or more times the second theorem of Art. 15 we can prove in all cases that the intersection of the double prism will be a convex polygon.

COROLLARY. The interior of any segment whose points are points of a double prism will lie entirely in the interior of the double prism unless it lies entirely in the double prism,

and a

half-line

drawn from a point of

the interior will inter-

sect the double

prism in one and only one point.

In a double prism the directing-polygons of each hypersurface

whose planes are elements of the other are

prism (just as

called

the directing-polygons of the double

we say

2$0

DOUBLE PRISMS

[vi. iv.

the vertex of a pyramid or hyperpyramid).

Any two

polygons intersecting in a single point and lying in planes which have only this point in common can be taken each
as a directing-polygon with the plane of the other as element of a piano-prismatic hypersurface, and so the two

together as the directing-polygons of a double prism.

can say that the surface of intersection of the two hypersurcommon lateral surfaces of the two sets of prisms as described in the preceding article, is generated by moving one of these
faces, the

We

Each point of polygons kept parallel to itself around the other. the polygon moves along the prisms of one set and around one of the prisms of the other set. The interior of the polygon generates the
interiors of the prisms along

which

it

moves.

The

interiors of the

other prisms will be generated by the interior of the other polygon moving in the same way around the first. The surface of intersection
it

is

covered with the polygons of each

set,

the two sets forming on

a net.

143.

Cutting a double prism so as to form two double

is cut

prisms.

Doubly triangular prisms. THEOREM. When a double prism

by a hyperplane

passing through points of the interior and containing elements of one hypersurface, the intersection is a prism, and the rest of it is separated into two portions, which, each combined
with the prism and its interior, form two double prisms whose interiors, with that of the prism, make up the whole interior.

PROOF.

The
lies

layer between the

two

elements of the

hypersurface
139, Th.

entirely in the interior of this hypersurface,

two hypersurfaces of the same kind (Art. This 6). layer and its faces intersect the other hypersurface in the lateral surface of a prism and two directit

and forms with

ing-polygons whose interiors are the bases of the prism.

The polygons belong

of the second,

to both hypersurfaces, their interiors,

the bases, belong to the first hypersurface and to the interior and the lateral surface belongs to the second

142, i43l

SUBDIVISION

251

hypersurface and to the interior of the first. The entire prism, therefore, belongs to the double prism, and is its intersection with the hyperplane of the layer, separating the rest of it into two portions which lie on opposite sides

The prism and its interior form with two portions two double prisms, whose hypersurfaces are the two hypersurfaces formed from the first given hyperof this hyperplane.

these

The

surface each taken with the second given hypersurface. interior of the prism lies entirely in the interior of the

given double prism, and separates the rest of this interior into two portions which are the interiors of the two new

double prisms.

The prisms

of

one

set,

surface, are separated,

some

the set which goes around the first hyperof them going to one of the two new

double prisms and the rest to the other. One or two of them may be divided, cut into two shorter prisms, one of these shorter prisms going to each of the new double prisms. The prisms of the other set
are all cut lengthwise, one part of each prism going to one of the new double prisms and the other part to the other.

COROLLARY.

By

cutting a double prism diagonally

the

we can

form double prisms in which

cutting these in another

prisms of one set are triangular,

so that those of the other set are three in

way

diagonally,

number; and then, we can form double

prisms in which the prisms of both

interiors of all of these double

sets are triangular, the

prisms together with the interiors the intersection prisms of making up the whole interior. of

double prism in which the prisms of both sets are triangular is a doubly triangular double prism, or simply a doubly triangular prism. Such a double prism is formed
three hyperplanes intersecting by twos in three parallel planes are cut by three other hyperplanes which intersect by twos in three parallel planes, any plane of one
set intersecting

when any

any plane

of the other set only in a single

point.

252

DOUBLE PRISMS

[vi. IV.

In Art. 163 it will be shown that a doubly triangular prism can be cut by hyperplanes so as to form six pentahedroids.

Hyperprisms with prism bases as double prisms. Centre of symmetry. A hyperHyperparallelopipeds. interiors of prisms is a double the bases are whose prism of the bases and the two lateral the two prism, prisms prisms corresponding to their ends forming one of the two
144.
sets of prisms of the double prism, while the prisms of the other set are parallelepipeds (see Art. 136).

set are parallelepipeds (and

Conversely, a double prism in which the prisms of one therefore the prisms of the

other set are four in number, two pairs of opposites) can be regarded in two ways as a hyperprism, the bases in each
case being the interiors of a pair of opposite prisms of the second set.

prisms of both sets in a double prism are parallelopipeds, or, what is the same thing, when
i.

THEOREM

When

the

both sets of directing- polygons are parallelograms, the figure

a hyperparallelo piped. Indeed, the hyperparallelopiped can be regarded in three ways as a double prism, the parallelis

opipeds of two pairs of opposite

of prisms and
the other

cells

forming one of

the sets

four
the

parallelepipeds the other set.

THEOREM

2.

When

two hypersurfaces of a double

prism have axis-planes of symmetry, the point of intersection of these planes is a centre of symmetry of the double prism;

and any hyper plane through

prism
146.

this point intersects the double

prism in a polyhedron which divides

the rest of the double

into two congruent parts (Art. 85,

Th.

2).

Relation

of

double

the point of view of Art. 130,

prisms to infinity. Taking we can say that a piano-

143-146]

HYPERCYLINDRICAL HYPERSURFACES
is

253

prismatic hypersurf ace tex-edge at infinity.

Two

a piano-polyhedral angle with versuch hypersurfaces give us a

double prism when the vertex-edge of one does not intersect the vertex-edge of the other, both being lines at infinity. When the prisms of one set in a double prism are parallel-

epipeds, so that the double prism can also be regarded in two ways as a hyperprism, there will be two points at
infinity

which are the vertices of two prismoidal hypersurfaces, and the line containing these two points will be the vertex-edge of a piano-prismatic hypersurface whose

directing-polygons are parallelograms. As stated in Art. 137, the eight hyperplanes of a hyperparallelopiped intersect the hyperplane at infinity in the

planes of the faces of a tetrahedron whose vertices are the vertices of the four prismoidal hypersurfaces in which the hyperparallelopiped lies. Each of the six edges of the

tetrahedron

in the vertex-edge of a piano-prismatic the three hypersurface, pairs of opposite edges corresponding to the three ways in which the hyperparallelopiped can be
lies

taken as a double prism.

at infinity corresponding to a rectangular hyperparallelopiped is a rectangular tetrahedron. Each

The tetrahedron
is

the pole of the opposite face, opposite edges are polar lines, and at each vertex there is a trirectangular
trihedral angle.
V.

vertex

HYPERCYLINDERS

146.

Sections.
of a

Hypercylindrical hypersurfaces. Their interiors. Axes. A hypercylindrical hypersurface consists

system of parallel lines passing through the points of a hyperplane surface, but not lying in the hyperplane of the
surface.

The

surface

is

called the directing-surface,

and the

parallel lines are the elements.

We shall consider only those cases in which

the directing-

254
surface
is

HYPERCYLINDERS

[vi. v.

a surface of elementary geometry, a plane or a sphere, or a conical or cylindrical surface with directingcircle, or a part or combination of parts of such surfaces.

A prismoidal

hypersurf ace can also be regarded as a partic-

ular case of a hypercylindrical hypersurface. Many of the properties of the hypersurface correspond In particular, to the properties of the directing-surface.

the hypersurface has an interior when the directing-surface has an interior, the interior of the hypersurface consisting

which pass through the points of the interior of the directing-surface and are parallel to the elements.
of the lines

Sections of the hypersurface are like those of the prismoidal hypersurface: a hyperplane passing through a point of the interior and parallel to the elements intersects

which

the hypersurface in a cylindrical surface, and a hyperplane is not parallel to the elements intersects the hypersurface in a surface, or at least in a system of points which
,

can serve as directing-surface.

surface whose hyperplane
is

A right section is a directinghypercylindrical hypersurface not parallel to the elements are

perpendicular to the elements.

THEOREM. Sections of a made by parallel hyperplanes

congruent.

PROOF.
section
section.
is

The distance between any two points of one equal to the corresponding distance in the other Then any tetrahedron whose vertices are points

of one section will be congruent to the corresponding tetrahedron in the other section (see Art. 72, Th. i), and if we

make

these tetrahedrons coincide, every point of one section (forming a tetrahedron with three of these points) will

coincide with the corresponding point of the other section (see proof of Th. i of Art. 98).
147.

Hypercylinders.

Special forms.

Relation of the

hypercylindrical hypersurface to infinity.

hypercylinder

146, i47l

SPECIAL FORMS

255

consists of that portion of a closed hypercylindrical hyper-

surface which

lies

between two

parallel directing-surfaces,

together with the directing-surfaces themselves and their

interiors.

The
lies

interiors of the directing-surfaces are the bases,

that portion of between the directing-surfaces is the lateral hyperThe interior of the hypersurface of the hypercylinder. cylinder consists of that portion of the interior of the hypercylindrical hypersurface

and the hypercylindrical hypersurface which

which

lies

between the bases.

a right hypercylinder when the elehypercylinder ments are perpendicular to the hyperplanes of the bases.
is

spherical hypercylinder is one whose bases are the The axis of a spherical hypercylinder is the interior of the segment whose points are the centres
interiors of spheres.
of the bases.
i. When a spherical hypercylinder is cut by a which passes through a point of the interior and hyperplane is parallel to the elements, the intersection is a circular cylinder.

THEOREM

THEOREM
three sides

2.

When a

rectangle takes all positions possible

with one side fixed, the vertices and the points of the other

make up a

right spherical hypercylinder.

fixed side is the axis, the opposite side is

other two sides are radii of the bases.

an

element,

The and the

THEOREM

3.

// we pass a plane through the axis of a

around this plane that poron one side of it, we shall form all of a right spherical hypercylinder except that portion which is the intersection of the cylinder and plane.
cylinder of revolution
tion of the cylinder
rotate

and

which

lies

A hypercylinder whose bases are the interiors of cylinders

can be regarded in two ways as a hypercylinder of this for there are two lateral cylinders corresponding kind
;

256

PRISM CYLINDERS AND DOUBLE CYLINDERS

[vi. vi.

to the ends of the bases, and these can be taken as the bases and the given bases as parts of the lateral hypersurface. In fact, the two lateral cylinders are congruent and lie in parallel hyperplanes, with the elements of one parallel to the elements of the other and the planes of the bases of

one parallel to the corresponding planes of the other. those elements of the hypercylinder whose lines intersect any element of one of its bases lie in the
Moreover,
interior of a parallelogram

which bears the same relation

This figure is a particular case be studied in the next section

to both pairs of cylinders.

of a prism cylinder, (see Art. 150).

and
of

will

Taking the point

view of Art. 130, we can say that

is

hypercylindrical hypersurface face with vertex at infinity (see Art. 153).

a hyperconical hypersur-

VI.

PRISM CYLINDERS AND DOUBLE CYLINDERS

Piano-cylindrical hyper surfaces.
Sections.

148.

Right
con-

directing-curves.
sists of

piano-cylindrical

hypersurface

a system of parallel planes passing through the points of a plane curve ajid intersecting the plane of the curve only in these points.* The curve is the directingcurve,

piano-prismatic hypersurface can be regarded as a particular case of a

piano-cylindrical hypersurface. Except for this, we shall consider only the case in which there is a directing-circle.

and the planes are the

elements.

As the piano-cylindrical hypersurface is analogous to the piano-prismatic hypersurface, many of the theorems
correspond.

We

shall
:

only state some of them briefly

(see Arts. 138-140)

* If we had undertaken to give a definition and some of the properties of curves, we might have followed the analogy of the piano-prismatic hypersurface and defined
this hypersurface independently of

any

directing-curve, as a part of the two-dimen-

sional

geometry of a system of

parallel planes (see Art. 138).

147,

US]

PLANO-CYLINDRICAL HYPERSURFACES

257

have the interior of the hypersurface consisting of those planes parallel to the elements which pass through the points of the interior of the directing-curve, with the
usual theorems in regard to the interior.

We

The

section

containing an element will be one or two elements, but the section made by a hyperplane which does not contain an element will be a cylindrical
surface, so that the hypersurface can also be regarded as a

made by a hyperplane

hypercylindrical hypersurface with a directing cylindrical surface. Likewise, as in the case of the piano-prismatic hypersurface, a plane which do.es not contain a line parallel
to the elements will intersect every element in a point,

and

the hypersurface in a curve which can be taken as directing-

curve;

and directing-curves which

lie

in parallel planes

are congruent.

directing-curve whose plane is absolutely perpendicular to the planes of the elements is called a right directingcurve ;

and any plane

isocline to the plane of a right direct-

ing-curve but not parallel to the elements, or, what is the same thing, any plane isocline to the elements, intersects
the hypersurface in a curve which
directing-curves.
is

similar

* to the right

the hypersurface has a directing-circle, the plane centre parallel to the elements is an axis-plane through the of hypersurface, and every point of it is a centre of
its

When

When the right directing-curve is a circle, the hypersurface can be generated by the rotation of one of the elements around the axis-plane, that is, by the rotasymmetry.

two parallel planes around the other. It is then a piano-cylindrical hypersurface of revolution. For such a hypersurface we can say that any plane isocline to the elements, or to the axis-plane, intersects the hypertion of one of
*

Two

curves are similar

all

proportionally

half -lines

when they can be placed in positions where they cut which intersect them from a given point.

258

PRISM CYLINDERS AND DOUBLE CYLINDERS

;

[vi. vi.

and through any line which is not pernor parallel to the elements pass two such pendicular
surface in a circle

planes (Art. 107, Th.

i).

THEOREM.

Any

directing-curve

of

piano-cylindrical

hypersurface of revolution is
cylindrical surface*

directing-curve of a circular

line of the plane of the directa to the planes of the elements, isocline ing-curve pass plane and so intersecting the hypersurface in a circle. This
lie in a hyperplane, and the directwhich they intersect the hypersurface are directing-curves of the cylindrical surface in which the Since one of hyperplane intersects the hypersurface.

PROOF.

Through some

plane and the given plane

ing-curves in

a circular cylindrical surface and the given directing-curve is a directing-curve of a

these
is

circle,

the surface

is

circular cylindrical surface.

149.

a piano-cylindrical hypersurface.

Intersection of a piano-prismatic hypersurface and When the elements of a

piano-prismatic hypersurface intersect the elements of a piano-cylindrical hypersurface only in points, the intersection of the

two hypersurfaces consists

of the lateral sur-

face of a set of cylinders lying in the cells of the prismatic

hypersurface and joined together by their bases, together with the curves whose interiors are these bases.
hypersurface are parallel planes intersecting the cylindrical hypersurface in a set of equal parallel directing-curves of the latter, and the
the prismatic hypersurface are layers, each layer intersecting the cylindrical hypersurface in the lateral
cells of

In

fact, the faces of the prismatic

of these directing-curves.

surface of a cylinder whose bases are the interiors of two The interiors of the bases of
lie

the cylinders

in the prismatic hypersurface

*

and in the

And

is

therefore a circle or ellipse.

i 4 8,

149]

PRISM CYLINDERS

259

A set of correspondone from of each cylinder, are the elements cylinders, ing the sides of a polygon which is a directing-polygon of the
interior of the cylindrical hypersurface.

prismatic hypersurface and the interiors of these polygons lie in the cylindrical hypersurface and in the interior of the prismatic hypersurface. We can think of the set of cylin;

ders as going around the prismatic hypersurface, while any base or interior of a cross section of any of these cylinders is a piece cut out of an element of the prismatic hypersurface.

prism cylinder consists of the intersection of a piano-

prismatic hypersurface and a piano-cylindrical hypersurface whose elements intersect only in points, together with all that portion of each hypersurface which lies in the interior
of the other.

The

interior of a

prism cylinder consists of the points

to the interiors of its two hypersurfaces. in the interior of the cylindrical hypersurface lying its to elements intersects the prismatic hypersurparallel face in a directing-polygon whose interior belongs to the

which are

common

A plane

interiors of

both hypersurfaces, and so to the interior of the prism cylinder. The vertices of this, polygon are a set of corresponding points of the bases, and the sides lie

in the interiors, of the set of cylinders described above.

the elements of one hypersurface are absolutely perpendicular to the elements of the other the prism cylin-

When
is

der

a right prism cylinder.

When

also the cylinders are

and when any set of corresponding elements, one from each cylinder, form a regular polygon,
cylinders of revolution,

the prism cylinder

is

regular.

In a right prism cylinder the cylinders of the surface of intersection of the two hypersurfaces can be put together in one hyperplane so as to form a single right cylinder. On the other hand, if
this surface is cut along a set of corresponding elements of the cylinder,

and so along a directing-polygon of the prismatic hypersurface,

it

can

260

PRISM CYLINDERS AND DOUBLE CYLINDERS

[vi. vi.

be spread out so as to form a single right prism. The cylinder and will each have their elements equal to the perimeters of the bases of the other.. In forming the prism cylinder this single right cylinder is folded on the planes of certain cross sections until the ends
prism
are brought together, and the prism is bent around the cylinder. The two are linked together, the lateral surface of one coinciding as a whole with the lateral surface of the other (see note to Art. 141).

150.
lar

Directing-polygons and directing- curves.

prism cylinders. prism cylinder the directing-polygons of the prismatic hypersurface whose planes are elements of the cylindri-

Prism cylinders

of revolution.

TrianguIn a

and the directing-curves hypersurface whose planes are elements

cal hypersurface,

of the cylindrical of the prismatic

hypersurface, are called the directing-polygons and directingcurves of the prism cylinder.

We
faces

is

can say that the surface of intersection of the two hypersurgenerated by moving the polygon around the curve, or by

moving the curve around the polygon. In the first case each point moves around one of the cylinders in the second case each point of the curve moves along them all. The interior of the
of the polygon
;

curve generates the interiors of the cylinders; the interior of the polygon generates that portion of the prism cylinder which belongs
to the cylindrical hypersurface

and

to the interior of the prismatic

hypersurface.

The

surface of intersection
sets

gons and with the curves, the two

covered with the polyforming on it a net.

is

THEOREM

i.

When a prism

cylinder is cut by a hyper-

plane passing through points of the interior and containing

elements of the prismatic hypersurface, the intersection is a cylinder, and the rest of it is separated into two portions,

which, each combined with the cylinder and its interior, form two prism cylinders whose interiors, with that of the
cylinder,

Art. 143, with only such changes in the terms used as are necessary because

make up the whole interior. The proof is the same as that of

is

the second hypersurface

cylindrical

and not prismatic.

149-151]

SUBDIVISION

261

The cylinders of the prism cylinder are separated, some of them going to one of the two new prism cylinders and some to the other. The other portion of the prism cylinder, that portion which goes around the cylindrical hypersurface, is cut lengthwise, one part going
to each of the

new prism

cylinders.

COROLLARY. By cutting a prism cylinder diagonally we can form prism cylinders in which the directing-polygons are triangles and the cylinders, three in number, triangular
prism cylinders.

A hypercylinder whose bases are the interiors of cylinders

is

and a prism cylinder in which the are parallelograms (and therefore the directing-polygons cylinders are four in number) can be regarded in two ways
a prism cylinder;
as a hypercylinder. When the two hypersurfaces of a prism cylinder have axes-planes of symmetry, the point of intersection of these

planes

is

a centre of symmetry (as in Art. 144, Th.

2.

2).

THEOREM

of one base, the

rest of the

// we rotate a right prism around the plane prism will generate a right prism

The lateral cylinder having circles for its directing-curves. the bases the ttie cylinders of prism cylinder, edges generate of
each lateral face generates the interior of one of these cylinders, and the moving base generates that portion of the prism cylinder which
lies

interior of the prismatic hypersurface.

the interior of the

in the cylindrical hypersurface and in the The fixed base and

prism belong

to the interior

of the prism

cylinder.

When

Intersection of two piano-cylindrical hypersurfaces. the elements of a piano-cylindrical hypersurface intersect the elements of a second piano-cylindrical hyper151.

surface only in points, each element of one hypersurface intersects the other hypersurface in a directing-curve, and

262

PRISM CYLINDERS AND DOUBLE CYLINDERS

[vi. vi.

* the surface of intersection consists of the curves of either

one of these
lie

sets.

The

interiors of the curves of each set

in

one of the hypersurfaces and in the interior of the

other.

A double cylinder consists of the intersection of two pianocylindrical hypersurfaces

whose elements intersect only in

lies

points, together with that portion of each which

in

the interior of the other.

The

directing-curves of each

hypersurface whose planes are elements of the other are called the directing-curves of the double cylinder.

The

interior of a double cylinder consists of the points

which are

common

to the interiors of its

two hypersurfaces.

plane lying in the interior of one hypersurface parallel to its elements intersects the other hypersurface in a direct-

ing-curve whose interior belongs to the interior of the double

cylinder.

the elements of one hypersurface of a double cylinder are absolutely perpendicular to the elements of the other the double cylinder
is

When

right double cylinder.

The
by

surface of intersection of the two hypersurfaces is generated moving a directing-curve of one system around a directing-curve

If this surface is cut along a directing-curve, it can be spread out in a hyperplane, and in the case of a right double cylinder it will then form the lateral surface of a right cylinder. This can

of the other.

be done in two ways, and in the double cylinder we have two cylinders bent around each other, the lateral surface of one coinciding as a whole with the lateral surface of the other, and this common surface, together with the interiors and bases of the two cylinders, making

up the double cylinder. When the two bases of a single cylinder come together, any two corresponding points of these bases and the points which were on a line between them become the points of a directing-curve whose interior belongs to the interior of the double
cylinder.
*

Without defining surface

in general,

we assume

that the intersection of

two

hypersurfaces is a surtace.

iSi, IS*]

DOUBLE REVOLUTION

263

Cylinders of double revolution. i. If we rotate a cylinder of revolution around the plane of one base, the rest of the cylinder will generate a
152.

THEOREM

right double cylinder with directing-circles ;

and

the double

cylinder can be generated in two ways by the rotation of a cylinder of revolution around one of its bases.

The

right double cylinder with directing-circies

is

there-

fore called a double cylinder of double revolution, or simply a cylinder of double revolution.

THEOREM

2.

In a cylinder
it is

of double revolution the in-

tersection of the two hypersurfaces lies in

a hypersphere, and

in this hypersphere
124).

a surface of double revolution (Art.

PROOF.

Any point of the intersection, its projection upon

one of the axis-planes, and the centre of the double cylinder, are the vertices of a right triangle whose legs are radii of

two systems. The hypothenuse, therefore, and the angle which it makes with the axisplane are the same for all points of the intersection, so that the intersection lies entirely in a hypersphere, and in this hypersphere is a surface of double revolution whose axiscircles are the two great circles in which the axis-planes of
directing-circles of the

the double cylinder intersect the hypersphere.

A plane passing
just

through the hypothenuse of the triangle

isocline

the axis-planes of the double cylinder, intersects the hypersurfaces in the same circle. The system of all these planes forms the conical
to

considered, and

hypersurface of double revolution whose intersection with the hypersphere is this same surface of double revolution.

THEOREM
tion in

3.

Conversely,

any surface of double

revolu-

a hypersphere

is the surface of intersection

of the two

hypersurfaces of a cylinder of double revolution.

264

PRISM CYLINDERS AND DOUBLE CYLINDERS

[vi. vi.

For the points of this surface, being all at the same distance from each of the axis-circles of the surface, are in hyperspace all at the same distance from each of the planes of these circles, and lie, therefore, in two piano-cylindrical hypersurfaces of revolution which have these planes for
axis-planes.

cylinder of double revolution can be regarded as inscribed in the hypersphere.

153.
infinity.

The

Relation of prism cylinders and double cylinders to Taking the point of view explained in Art. 130,
that a piano-cylindrical hypersurface is a pianohypersurface with vertex-edge at infinity.

we can say
conical

We

have a prism cylinder or double cylinder when the vertexedge of one hypersurface does not intersect the vertex-edge
of the other.

We

have a right prism cylinder or double

is polar to the other. the bases of a hypercylinder are cylinders there are two points at infinity, the vertices of two hypercylindri-

cylinder

when one vertex-edge

When

which we get by taking the figure in two as a ways hypercylinder. The line determined by these two points is the vertex-edge of the piano-cylindrical hypercal hypersurfaces

surface which belongs to the figure regarded as a prism The other vertex-edge in this case is the line cylinder.
of intersection of the planes of the bases of the four cylinders.

CHAPTER
IN

VII

MEASUREMENT OF VOLUME AND HYPERVOLUME

HYPERSPACE
I.

VOLUME

Lateral volumes of hyperprisms and hyperpyramids. Volume of the double prism. The cells of the polyhedroids
164.

that

we have

studied are polyhedrons of three-dimensional

geometry, and it is only necessary to state the theorems which concern their volumes.

THEOREM
to the

i. The lateral volume of a hyperprism is equal area of a right section multiplied by the lateral edge.

The

lateral

volume of a right hyperprism

*

is

equal

to

the

area of the surface of the base

multiplied by

the altitude.

THEOREM
is

2.

equal

to the

The lateral volume of a regular hyperpyramid area of the surface of the base multiplied by one-

third of the slant height, the

common

altitude of the lateral

bvramids.

THEOREM

3.

The

lateral

volume of a frustum of a regular

sum of the surface areas of the hyperpyramid bases plus a mean proportional between them, multiplied by
is equal to the

one-third of the slant height.

THEOREM
set of

4.

In a double prism

prisms

is equal to the

common

the total volume of one area of their bases multi-

plied by the perimeter of a right directing-polygon of the hypersurface around which the set of prisms extends.
*

Area

of the

commonly used

in mensuration will

polyhedron whose interior is the base. Many forms of expression be employed freelv in -this chapter.

265

266

VOLUME

[vn.

i.

PROOF.

Any
a

two

faces of

cell of

prism of the given set has its bases in the the hypersurface around which this

prisms extends, and its altitude is the distance between these two faces. Now a right directing-polygon of the hypersurface is a polygon whose plane is absolutely
set of

perpendicular to the elements, and the side which lies in this cell is perpendicular to the faces and measures the dis-

Therefore the volume of this prism is equal to the area of its base multiplied by this side of the right directing-polygon of the hypersurface, and the
tance between them.
total

volume

common

of the given set of prisms is equal to the area of their bases multiplied by the perimeter

of the right directing-polygon.

COROLLARY.
is equal to the

The

total

volume of a right double prism

area of one directing-polygon multiplied by the

perimeter of the other, plus the area of the second multiplied

by the perimeter of the first.

165.

surfaces.

Lateral volumes of cylindrical and conical hyperIn the case of curved hypersurfaces we have to

of limits or some other equivalent in to extend our definition of volume. and, fact, theory, Without going into details, we state the following theorems

employ the theory

THEOREM
the altitude.

i.

The

lateral
to the

volume

of

a right spherical

hypercylinder is equal

area of the base multiplied by

It is given by the

formula

R being the radius and H the altitude.

THEOREM
cone is equal
2.

The
Its

to the

lateral volume of a right spherical hyperarea of the base multiplied by one-third of

the slant height.

formula

is

H'

being the slant height.

154-156]

THEOREMS OF VOLUME
3.

267

The lateral volume of a frustum of a right is given by the formula cone spherical hyper

THEOREM

r being the radius of the

upper

base.

THEOREM

4.

In a
volume

right
is

prism cylinder with directingto the

circles the total

equal

area of the directing-

circumference of the directingpolygon multiplied by circle, plus the area of the directing-circle multiplied by the
perimeter of the directing-polygon.

the

The
to

volume of a cylinder of double revolution is equal the area of one directing-circle multiplied by the circumtotal

ference of the other, plus the area of the second multiplied by

the circumference of the first.
2
2

It is given

by the formula

T RR'(R

R'),

and R' being

156.

the radii of the two circles.

Volume of the hypersphere. THEOREM. The volume of a hypersphere

formula
2
TT

is given

by the

R*,

R being the radius.

PROOF.

Let

AB

be a quadrant of a great

circle.

Two

cut the hyperhyperplanes perpendicular to the radius in interiors are two spheres, whose the bases of sphere

OB

Let

a frustum of a hypercone inscribed in the hypersphere. f be the slant height of this frustum, and r\ and r2

the radii of the bases.

great by the formula
circle.

'

is

The

lateral

volume

the length of a chord of the of the frustum is given

Let from
H'.

H be the altitude of the frustum, and K the distance

to the middle point of the chord

which represents

From

similar triangles

we have

268

VOLUME

[vii. i.

H
or
ri

+ r,

= 2_KH_

H
2

Also,

from right

triangles,

H'2 =
and
R*
Therefore
r,r,

H +

(r,

r2 ) 2 ,
2
-

= =

^+
(r,

(J ff O

+r

2
2)

For an inscribed hypercone with vertex at radius of the base, we have only to make r^
these expressions.
If

and

r\

the

=o

in all of

AB

arcs symmetrically situated on the arc with respect to P, the middle point of this arc, we shall
triangles with

we take two

have two right

hypothenuse equal to H'

symmetrically situated with respect to the radius OP, and therefore equal. But the legs denoted by are non-

homologous

sides in the

two

triangles,

and so the sum

of

269

their squares

is

'2
.

Therefore,

if

we

write

down
of

the for-

mula given above

expressions will be

for each of

two divisions

AB

metrically situated with respect to P, the

sum

of the

symtwo

and the sum of the volumes of the frustums corresponding to the two arcs will be

n equal parts, we i pairs of shall have a hypercone and frustum, and n in a all inscribed and when frustums, half-hypersphere n is increased indefinitely the sum of their lateral volumes will have for limit the volume of the half-hypersphere.*
if

Now

we

divide the arc

AB

into 2

This sum

is

But
lim

nH' =

arc

AP = ^,

and
lim
n
*

00

Un,^,o. n' =
vt*

00

Here again we omit

details involving essentially a definition of the

volume

of the

hypersphere.

270

HYPERVOLUME
of the half-hypersphere is

[vn. n.

Thus the volume

and the volume

of the hypersphere is
2
7T
2
3
.

COROLLARY. The volume of a cylinder of double revolution circumscribed about a hypersphere is twice the volume of the
hypersphere.
II.

HYPERVOLUME

167. The terms hypersolid and hypervolume. We shall use the term hypersolid for that portion of hyperspace which constitutes the interior of a polyhedroid or of a simple

closed hypersurface such as a hypercone, hypersphere, or double cylinder. hypersolid is supposed to have hyper-

volume, which can be computed from the measurements of certain segments and angles, and which can be expressed in terms of the hypervolume of a given hypercube taken as a unit. The theory of hypervolume is exactly the same as the theory of volume in the ordinary geometry. We shall omit all discussion of this matter, and, as in the pre-

ceding section,

we

shall use freely the

forms of expression

commonly employed

in

mensuration.

The

distinction

we

between hypersurface and hypersolid is important, but shall often use these terms interchangeably, speaking,
for example, of the

hypervolume

of a given hypersurface,

and, on the other hand, of the vertices, edges, faces, or

a hypersolid. the ratio of two hypersolids we mean the ratio of their By hypervolumes. Thus the ratio of any hypersolid to the unit hypercube is the same as the hypervolume of the
cells of

hypersolid. Two hypersolids which have the same hypervolume are equivalent; and if a hypersolid is divided into two or more parts, the hypervolume of the whole is equal

156-159]

HYPERPARALLELOP1PED

271

to the
solids

sum of the hypervolumes of the parts. which are congruent are equivalent.
Congruent and equivalent hyperprisms.
i.

Two

hyper-

168.

THEOREM
PROOF.

Two

right

hyperprisms are congruent when

altitudes.

they have congruent bases

and equal

given base of one can be made to coincide of the other, and in one of these two base with either
positions the hyperprisms will lie on the same side of the hyperplane of the coinciding bases and will coincide throughout.

THEOREM
right
its altitude

2.

An

oblique hyperprism is equivalent to

its

hyperprism having for a lateral edge of the oblique hyperprism.

base a right section

a and for

two hyperprisms cut from the same prismoidal hypersurface with equal lateral edges are equivalent.

COROLLARY.

Any

159.

THEOREM
PROOF.

Hypervolume of a hyperparallelopiped. i. The hypervolume of a rectangular hyperits

parallelopiped is equal to the product of

four dimensions.

place, rectangular hyperparallelopipeds having congruent bases are to each other This we can prove when the altitudes as their altitudes.

In

the

first

two

are commensurable

and then when the

altitudes are in-

commensurable.

Then we prove that when they have two dimensions

in

common

they are to each other as the products of the other

in

two dimensions; when they have one dimension

com-

mon

they are to each other as the products of the other three dimensions; and, finally, in any case, they are to

each other as the products of their four dimensions. From the last statement, by taking for the second hyperparallelopiped the unit hypercube, we have the theorem
as stated.

272

HYPERVOLUME
2.

[vii.

n.

THEOREM
is equal ing altitude.
to the

The hypervolume of any hyperparallelopiped volume of any base multiplied by the correspond-

We shall prove the theorem by proving that an equivalent rectangular hyperparallelcan construct we opiped with base equivalent to the base and altitude equal
PROOF.
to the altitude of the given hyperparallelopiped. shall speak of a pair of opposite cells and the eight edges which join the vertices of one to the vertices of the

We

other as a corresponding pair of bases and

136)-

set

of edges (Art.

produce a set of edges and cut off an equivalent hyperparallelopiped by two hyperplanes perpendicular to these edges (Art. 158, Th. 2). The set of edges which are produced are taken on the same lines in the two hyper-

We

The other parallelopipeds, and are of equal lengths. three sets of edges of the given figure are replaced by edges perpendicular to the set produced. If the edges of any set were already perpendicular to the edges produced, they
are replaced by a set of edges parallel and equal to them, and if the edges of any set were perpendicular to the cor-

responding bases, and so to

will

all of

the other edges, the same

be true of the edges by which they are replaced.

The
is re-

pair of bases corresponding to the edges produced placed by a pair of bases perpendicular to these lines.

The

bases of the other three pairs are parallelepipeds replaced by equivalent right parallelepipeds lying in the same
hyperplanes.
Starting with the second hyperparallelopiped, we produce second set of edges, forming a third hyperparallelopiped

in the

same way;

and, finally, from the third hyper-

parallelopiped, producing a third set of edges,

we form a

fourth hyperparallelopiped, the four being all equivalent. The edges of the first set in all four hyperparallelopipeds

159, 160]

HYPERPRISM WITH PRISM BASES

273

are equal and parallel, and in the second, third, and fourth they are perpendicular to the corresponding bases, and therefore to all the other edges. The edges of the second
set in the second, third,

and fourth hyperparallelopipeds are equal and parallel, and in the third and fourth they are perpendicular to the corresponding bases, and therefore to all the other edges.

The edges

of the third set in

the third and fourth hyperparallelopipeds are equal and taken along the same lines, and in the fourth they are

perpendicular to the corresponding bases. Therefore, in the fourth hyperparallelopiped the edges of each set are perpendicular to the edges of all the other
sets,

and the hyperparallelopiped

is

rectangular.

the hyperplanes of the fourth pair of bases remain the same throughout this process, and with these bases
all have the same altitude Th. Moreover, these bases do not differ (Art. 128, 12). in the last two they are congruent; they and in volume, are, in fact, parallelepipeds formed by this same process

Now

the four hyperparallelopipeds

of producing successively different sets of edges.

have constructed an equivalent rectangular hyperparallelopiped with base equivalent to the base and alti-

We

and

tude equal to the altitude of the given hyperparallelopiped, this is sufficient to prove the theorem.
160.

THEOREM

Hypervolume of any hyperprism. i. The hypervolume of a hyperprism whose

to the

base is the interior of a prism is equal base multiplied by its altitude.

volume of

its

PROOF.
prisms,* the base
*

As any prism can be divided

it is is

only necessary to

into triangular the theorem when prove

we mean that a make up the volume
;

a triangular prism.
prisms can be formed whose volumes together

We

are using the term prism here for the solid (see Art. 157)

set of triangular

of the given prism.

274

HYPERVOLUME
a triangular prism

[vn. n.

On

we can

build a parallelepiped

by

joining an equal triangle to its base so as to form a parallelogram, and drawing a fourth lateral edge. We join
to the given prism a second triangular prism, and the two are symmetrically situated with respect to the centre of

the parallelepiped, and therefore equal (Art. 84, Ths. 2 On the hyperprism we can, then, build a hyper4). parallelopiped having this parallelepiped as base.

and

We

join to the given hyperprism a second hyperprism, and the two hyperprisms are indeed congruent, since the centre
of

symmetry

of the hyperparallelopiped lies in the diagonal

hyperplane along which the two prisms are joined (Art. The hypervolume of the 136, Th. i, and Art. 85, Th. 2).
given hyperprism is therefore equal to one-half of the hypervolume of the hyperparallelopiped, and so to the volume of its own base multiplied by its altitude.

THEOREM 2. Two right hyperprisms are equivalent if they ham equal altitudes and if their bases are the interiors
of tetrahedrons which can be so placed that as triangular pyramids they shall have equivalent bases and equal altitudes. PROOF. The lateral cells are the interiors of triangular

prisms.

hyperprism upon one of these be called a cells, may prismatic base. The opposite lateral edge will then be a vertex-edge, and the altitude will be the distance of the vertex-edge from the prismatic base.
will rest the

We

which

The

original base is

now

at one end, and

is itself

the in-

a triangular pyramid with one of the ends of the prismatic base for its base, and for its altitude the altitude
terior of

of the

hyperprism in

its

present position.

Looking at the hyperprism in this way, we have a series of theorems analogous to the theorems by which we determine the volume of a pyramid in geometry of three
dimensions
:

i6o]

TETRAHEDRAL HYPERPRISMS

275

(1)

is itself

and

its

hyperplane section parallel to the prismatic base a prism of the same length as the prismatic base, end is a section of the end-pyramid parallel and

similar to the base of the pyramid. The volume of this parallel hyperplane section is, then, proportional to the

square of
(2) If

its

two

distance from the vertex-edge. right hyperprisms with tetrahedral ends,

when

placed as explained above, have equivalent prismatic bases and equal altitudes, hyperplane sections parallel to the
prismatic bases and at the same distance from the vertexedges are equivalent. (3) Given the two hyperprisms just considered, we divide the common altitude into equal parts and construct a series

and circumscribed prisms to the pyramids at the ends, and so a series of hyperprisms having these prisms for ends, inscribed and circumscribed to the original
of inscribed

hyperprisms.

hyperprisms

is

The hypervolume of either of the given the limit of the sum of the hypervolumes of

the hyperprisms inscribed or of the hyperprisms circumscribed to it, when the number of subdivisions of the alti-

tude

is

increased

indefinitely.

Any
is

inscribed

or

cir-

equivalent to the corresponding hyperprism of the other figure, and therefore the two given hyperprisms must be equivalent.

cumscribed hyperprism of one figure

two right hyperprisms have equal altitudes, and for bases the interiors of triangular pyramids with equivalent bases and equal altitudes, the hyperprisms satisfy the conditions of (3) and are equivalent.

Now when

THEOREM
equal
to the

3.

The hypervolume of any hyperprism

its

is

volume of

base multiplied by

its altitude.

PROOF.
hedrons, the base
is

As any polyhedron can be divided

it is

into tetra-

only necessary to prove the theorem

when

a tetrahedron.

276

HYPERVOLUME

[vii.

n.

On

a tetrahedron we can build a triangular prism by

it

joining to

two other equivalent tetrahedrons, and so

on the given hyperprism we can build a hyperprism with a triangular prism for base, composed of three hyperprisms with tetrahedral bases. If we produce the lateral edges, we can form a right hyperprism with a triangular prism for base, composed of three right hyperprisms with tetrahedral bases, the four right hyperprisms equivalent respectively to the hyperprisms from which they were produced (Art. 158, Th. 2). The three tetrahedral bases of

the right hyperprisms are equivalent, and any two of them can be so placed that as triangular pyramids they shall

have equivalent bases and equal

altitudes.

Hence the

three right hyperprisms with tetrahedral bases are equivalent, and the first three hyperprisms with tetrahedral

bases

are

equivalent.
is,

The hypervolume

of

the

given

hyperprism

volume of which
of its

therefore, equal to one-third of the hyperof the hyperprism with triangular prism for base
it

own

forms a part, and so it is equal to the volume base multiplied by its altitude.

COROLLARY.
to the

The hypervolume of a hyperprism is equal volume of a right section multiplied by the lateral edge.

Hypervolume of a hyperpyramid. For the hypervolume of a hyperpyramid we have, as in the proof of Th. 2
161.
of the preceding article, a series of theorems analogous to the theorems by which we determine the volume of a

pyramid

in

geometry
i.

of three dimensions

THEOREM
proportional

parallel to the

A hyperplane section of a hyperpyramid base is similar to the base, and its volume is
cube of
its

to the
2.

distance from the vertex.

THEOREM

// two hyperpyramids have equivalent bases

bases

and equal altitudes, hyperplane sections parallel to the and at the same distance from the vertices are equivalents

i6o, 161]

HYPERPYRAMID
3.

277

THEOREM
alent bases

Two

pentahedroids are equivalent if they

altitudes.

can be so placed that as hyperpyramids they shall have equiv-

and equal

Dividing the altitude into some number of equal parts, we can construct a series of inscribed and circumscribed hyperprisms and prove that the hypervolume

PROOF.

of either pentahedroid is the limit of the sum of the hypervolumes of the set of hyperprisms inscribed or circumscribed

to

it

when the number

of subdivisions of the altitude is in-

creased indefinitely. Thus we prove our theorem in the same manner as we prove the corresponding theorem in geometry of three dimensions.

THEOREM
equal
to the

4.

The hypervolume of any hyperpyramid

its

is
its

volume of

base multiplied by one-fourth of

altitude.

PROOF.
hedrons, the base
it for
is

As any polyhedron can be divided

it is

into tetra-

only necessary to prove the theorem when the interior of a tetrahedron, that is, to prove

pentahedroids. a pentahedroid taken as a hyperpyramid we can build a hyperprism having the same base and one lateral

On

edge the same. In doing this we join to the given pentahedroid a hyperpyramid with a triangular prism for base,
a hyperpyramid which can be divided into three pentahedroids, the triangular prism being divided into three tetrahedrons. One of these pentahedroids can then be

proved equivalent to the given pentahedroid, so that the four pentahedroids are all equivalent. Let A-A'B'C'D' be the given pentahedroid. Drawing lines through J3', C", and D' parallel to A' A, and a hyperplane through

parallel to the hyperplane of the tetrahe-

dron A'B'C'D', we have a hyperprism ABCD-A'B'C'D' composed of the given pentahedroid and the hyperpyramid

278

HYPERVOLUME
This hyperpyramid

[vn. n.

A-BCDB'C'D'.

we

divide into three

equivalent pentahedroids by dividing the prism BCDB'C'D' into three equivalent tetrahedrons, the common altitude of the three pentahedroids being the distance of
the vertex

from the hyperplane

of the triangular prism.

Now

of these three pentahedroids

one can be regarded as

having A BCD as its base, and as its vertex one of the points
B', C", or

of the lower base of the hyperprism.

Re-

seen to be equivalent to the original since the bases of the two are the bases of pentahedroid, their common altitude is the altitude the hyperprism, and

garded in this

way

it is

of the hyperprism.
is, therefore, one of four equivwhich alent^ pentahedroids go to make up the hyperprism and its hypervolume is one-fourth of the hypervolume of the hyperprism, and so equal to the volume of its own base
;

The given pentahedroid

multiplied by one-fourth of

its altitude.

The figure on page 238 represents the cells of the hyperprism spread out into a single hyperplane. Dividing the three prisms on the left by the planes AB'C, AB'D', and AC'D', we have all the cells except
the

common

cell

AB'C'D'

of the given pentahedroid

and of the

hyperpyramid A-BCDB'C'D'.

Hypervolume of a frustum of a hyperpyramid. THEOREM. The hypervolume of a frustum of a hyper162.

pyramid

is

given by the formula

i
where

H(B + #
the

#V +

ft),

and b are
Let

volumes of the bases and

is the

altitude.

PROOF.
r

be the volume of the lower base, and

.

let

be the ratio of an edge of the upper base to the correspond3 The ing edge of the lower base, so that we have b = Br formula of the theorem can, then, be written |

HB(i

+r+r +r
2

).

161,162]

FRUSTUM OF A HYPERPYRAMID

279

By dividing the bases into tetrahedrons we can divide the frustum into frustums of pentahedroids,* in all of and r have the same values. Therefore, it is which

only necessary to prove the theorem when the bases are the interiors of tetrahedrons, that is, for a frustum of a
pentahedroid. Now the frustum of a pentahedroid is of the same general form as a hyperprism with tetrahedral bases, and can be
divided into four pentahedroids in the same way that the hyperprism of the preceding proof is divided into four
pentahedroids. f Then it can be proved that the hypervolumes of these four pentahedroids form a geometrical progression in which the ratio is r and the first term \ HB,

the hypervolume of the pentahedroid whose base lower base of the frustum.
',

is

the

Let the given frustum be ABCD-A'B'C'D* ABCD being the upper base and A'B'C'D' the lower base. We
divide this into four pentahedroids

by the three hyperplanes

AB'C'D', ABC'D', and ABCD';

namely, into

AA'B'C'D', ABB'C'D', ABCC'D', and ABCDD'.

hyperplane, cutting off the pentahedroid r AA'B'C'D', leaves the hyperpyramid A-BCDB'C'D That is, the last three of the four pentahedroids together
first
.

The

form a hyperpyramid with vertex at

and base the frustum

the hyperpyramid is divided into three pentahedroids this frustum is divided into three tetrahedrons whose volumes are proportional
Therefore, the hypervolumes of the three 2 pentahedroids are proportional to i, r, and r
i, r,
. .

BCD-B'C'D'

of a tetrahedron.

When

to

and

r2

See Art. 157, and foot-note, p. 273. A truncated pentahedroid is divided in this

"On

the Centre of Gravity of a Truncated Triangular

5.

way by Sylvester in the memoir, Pyramid" (p. 172), referred

to on p.

280

HYPERVOLUME

[vn. n.

Similarly, the first three pentahedroids form a hyperpyramid with vertex at D' and base the frustum

ABC-

A'B'C'

of

a tetrahedron, and their hypervolumes are

2
.

proportional to i, r, and r In other words, the hypervolumes of the four pentahe2 droids are proportional to i, r, r and r3
,

But the hypervolume

of the first pentahedroid is \ of the given

r*

HB.

Therefore the hypervolume

i

frustum

is

HB(i

+r+

+ r).

The formula can

tracting from

also be derived algebraically by subthe hypervolume of a hyperpyramid the

hypervolume

of a smaller

hyperpyramid cut

off so as to

leave a frustum.

163.

Hypervolume
i.

THEOREM

of a double prism. The hypervolume of a doubly triangular

prism is equal to six times the hypervolume of the pentahedroid whose vertices are the points obtained by taking the
vertices of

the

a base in one of the two sets of prisms vertices of a base in the other set (Art. 141).

together with

PROOF. We shall prove this theorem by dividing the double prism into two hyperpyramids with triangular prisms for bases, and then into six equivalent pentahedroids.
their faces

Certain of these pentahedroids will be found to have among a base from each of the two sets of prisms of

the double prism. Let the nine vertices of the double prism be

A'

ABC C
B'
as written

A" B"
where the three
parallel triangles,
lines

C",
represent three equal

and the three columns represent three

162, 163]

DOUBLE PRISM

281

equal parallel triangles. Any two triangles of either set lie in a hyperplane, and their interiors are the bases of a
prism.

The point A and the plane of the parallelogram B C'B"C"

r

determine a hyperplane which divides the double prism into two hypersolids. We can indicate this by writing
their vertices

C
The hyperplane cannot contain any
or
of the points J5, C, -4',

A"

for it contains at least

one vertex of each of the

six triangles,
it

and

if it

contained any one of these triangles,

triangles parallel to it

would contain the two

and so

all

the nine vertices of the double prism. Now the hyperplane intersects the hyperprism ABC-A'B'C' in the plane

which separates B and C from A and it intersects lf in the plane the hyperplane of the prism ABC-A"B"C AB"C" which separates B and C from A". Therefore

AB'C

',

and

are separated in hyperspace from

28).

A and A 11 by
r

this

hyperplane (Art.

The two
its

hypersolids are hyperpyramids, each having vertex at A and a triangular prism as base. Each of

these hyperpyramids can be divided into three equivalent pentahedroids, the base being divided into three equivalent tetrahedrons. This can be done in such a way that one
of the pentahedroids of the first set shall

be the pentawrite

hedroid

ABCC'C", and one

of the pentahedroids of the

second set the pentahedroid

these

AA'A"B"C". We

ABC

and
1

A
A'

C
C'

A" B"

C".

282

HYPERVOLUME
will take the first
its

[vn. n.

We

with vertex at C" and

pentahedroid as a hyperpyramid base as a pyramid with vertex at

C", and the second pentahedroid as a hyperpyramid with vertex at A' and its base as a pyramid with vertex at A. The two pyramids will then have equal bases, ABC and A"B"C", and equal altitudes, the distance between the planes of these bases. In fact, they are two of the three equivalent tetrahedrons into which the prism ABCA"B"C" can be divided. The altitudes of the two
hyperpyramids will be equal to the distance of the plane A'B'C' from the hyperplane of this prism, to which it is Therefore, these two hyperpyramids are equivparallel.
alent pentahedroids that is, the six pentahedroids into which the double prism has been divided are all equivalent,
;

and the hypervolume of the double prism is equal to six times the hypervolume of any one of these pentahedroids, for example, of the pentahedroid ABCC'C" which has among its faces the triangles ABC and CC'C".
,

figure on page 248 represents the cells of the double prism in a single hyperplane. Dividing the prisms on the right out spread and left by the planes AB'C' and AB"C", and the upper two prisms in front by the planes AB'B" and ACC"\ we have all the cells except

The

the

common
first

cell

AB'CB"C"

of the

two hyperpyramids

into which

we

divide the double prism.

2. The hypervolume of a double prism is equal a base area to the of of any of the prisms of either set, multiplied by the area of a right directing-polygon of the hyper-

THEOREM

surface around which this set of prisms extends. can divide the bases of the prisms of the PROOF.

We

given set into triangles, the given prisms into triangular prisms, and the double prism into double prisms having each a set of triangular prisms for the given set of prisms.

The hypersurface around which

these prisms extend is the same for the given double prism and for all of the double

iC 3 ]

DOUBLE PRISM

283

prisms into which it is divided (Art. 143, Th. and Cor.). Therefore it is only necessary to prove the theorem for double prisms in which the given set of prisms is a set of
triangular prisms.

Again, taking a double prism with a given set of triangular prisms for one of its sets of prisms, we can divide the right
directing-polygons of the hypersurface around which these prisms extend into triangles. The diagonals which divide

one of these directing-polygons into

triangles,

together

with the faces of the hypersurface at their extremities, determine layers which form, with the parts into which they divide the hypersurface, triangular hypersurfaces,

and so divide the double prism

prisms. it is true of any
If the

theorem

into doubly triangular true of doubly triangular prisms, double prism in which the given set of
is

a set of triangular prisms. Therefore, it is only necessary to prove the theorem for doubly triangular prisms. Proceeding as in the proof of Th. i, we have a pentaprisms
is

hedroid

ABCC'C", one

of

six

equivalent pentahedroids

The volume into which the double prism can be divided. of the tetrahedron is equal to the area of the triangle multiplied by one-third of the distance of

ABCC"

ABC

C"

and the hypervolume of the pentahedroid is equal to the volume of the tetrahedron A BCC" multiplied by one-fourth of the distance of C' from the hyperplane of this tetrahedron. That is, the
from the plane
of this triangle
;

hypervolume

of the pentahedroid

the triangle multiplied of these two distances.

ABC

by

is equal to the area of one- twelfth of the product

Now the plane absolutely perpendicular to the plane ABC

intersects the hypersurface

which has

this

plane for one

side of this

of its faces in a right directing-triangle.

One

triangle measures the distance

and

A"B"C",

between the planes ABC and the corresponding altitude of the

284

HYPERVOLUME

[vn. n.

triangle measures the distance of the plane

the hyperplane of the other two planes. is, of the right directing-triangle of the hyperfor one of its faces, is surface which has the plane
triangle, that

A'B'C' from The area of the

ABC

then equal to one-half of the

product of these two distances

and the hypervolume of the pentahedroid

ABCC'C" is equal

to one-sixth of the product of the areas of the two triangles. But the hypervolume of the pentahedroid is also onesixth of the

hypervolume
is,

of the double prism.

Therefore,

the latter

is

exactly equal to the product of the areas of

to the area of a base in one set

the two triangles, that

of prisms multiplied by the triangle of the hypersurface

area of a right directing-

around which

this

set

of

prisms extends.

COROLLARY.
equal
to the

The hyperwlume of a

right double

prism

is

product of the areas of its two directing-polygons.

Hypervolumes of cylindrical and conical hyperFor the hypercylinder, hypercone, prism double cylinder we have the following and cylinder, theorems, derived from the corresponding theorems of the
164.

surfaces.

preceding articles

THEOREM
altitude.

i.

The

hyperwlume
by the formula

of

spherical

hyper-

cylinder is equal to the volume of the base multiplied by the

It is given

being the radius of the base

and

the altitude, as in Art.

155-

THEOREM
the altitude.

2.

The hyperwlume of a spherical hypercone

volume of
the base multiplied

is equal to the

by one-fourth of

It is given by the formula

163-165]

CURVED HYPERSURFACES
3.

28$

THEOREM
hypercone
is

The hypervolume of a frustum of a spherical given by the formula

*H(R*
or

+ Rr
2

THEOREM

4.

The hypervolume of a
to the

right

prism cylinder

with directing-circle is equal

product of the areas of the

directing-polygon and directing-circle. The hyperwlume of a cylinder of double revolution is equal It is given to the product of the areas of its directing-circles.

by the formula

166. Hypervolume of the hypersphere. THEOREM. The hypervolume of a hypersphere

to its

is

equal

volume multiplied by one-fourth of

its

radius.

PROOF.

We

use the construction and notation of Art.

We inscribe a frustum of a hypercone entirely on 156. one side of the centre 0, and on each base we place a hypercone with vertex at 0. The figure formed by adding to the frustum the hypercone with larger base and then
taking
solid

away the hypercone with smaller base is a hyperwhich we can use to determine the hypervolume of

the hypersphere. If we divide the arc into some number of equal parts and form these figures for all the chords

AB

#', we shall have a set of hypersolids fitting together within the half-hypersphere, and the limit of the sum of
their

hypervolumes we assume to be the hypervolume of

the half-hypersphere. Let x denote the distance from

of the nearer base of

the inscribed frustum, and let this be the base whose radius is r\. The hypervolumes of the two hypercones
are then

286

HYPERVOLUME

[vn. n.

and f

TT(X

H)r2
is

3
,

and the hypervolume

of the

frustum

Adding to this last the hypervolume of the first hypercone and substracting the hypervolume of the second, we have
i T
!

H(n +

r 2 )(r! 2

r2

+ xrf -(x
r2

Now
Hence

(*+ H}H and

r!

That

is,

r2 )

But
and
(r\

- H*

(Art. 156).

i6s]

HYPERSPHERE
finally,

287

Hence,

our hypervolume becomes

r2

- #
Now

(n

nr,

+r

2 2 )

the lateral volume of the frustum

is

so that the
lateral

hypervolume of our hypersolid is equal to the volume of the frustum multiplied by \ K.

all

Consider
the

now number of
of the

the subdivisions of the arc

is

AB.

When
ap-

subdivisions

increased indefinitely the

as

sum

hypervolumes

will,

we have assumed,

proach as limit the hypervolume of the half-hypersphere, the sum of the lateral volumes will approach what we have
the half-hypersphere, and will to the and limit Therefore, passing expressing the result for the entire hypersphere, we have the hypervolume of the hypersphere equal to its volume
called the
of

volume

approach R.

multiplied

by one-fourth
i.

of the radius.
the

The hypervolume of given by the formula

COROLLARY

hypersphere

is

COROLLARY
to

2.

The hypervolume of a hypersphere

hypervolume
of
the

is

one-half

the

circumscribed

equal double

cylinder,

and twice the hypervolume of the inscribed double It is equal to the hypervolume of with cylinder equal radii. double inscribed cylinder plus the hypervolumes of two any hyperspheres whose radii are the radii of the double cylinder.
The student may
investigate the

hypervolume of a pentahedroid

in elliptic hyperspace, following the analogy of Art. 120.

He

will

288

HYPERVOLUME

[vn. n.

find that there are eleven equations connecting the sixteen different

hypervolumes of a set of associated pentahedroids with the measures If he introduces also what we may of the five hyperplane angles. call the hypervolumes of the five tetrahedroidal angles, he will have five more equations,, so that he can express the hypervolume of the
angles.*

pentahedroid in terms of the hyperplane angles and the tetrahedroidal But a tetrahedroidal angle is like a spherical tetrahedron we cannot measure it directly, and we have no simple formula for its
:

hypervolume.
* M. Dehn, " Die Eulersche Formel
ticular,

in

Zusammenhang mit dem Inhalt

in

der Nicht-

Euklidische Geometric," Mathematische Annalen, vol, 61, 1905, pp. 561-586, in parpp. 583-584.

CHAPTER

VIII

THE REGULAR POLYHEDROIDS *

I.

THE FOUR SIMPLER REGULAR POLYHEDROIDS

Definition
of

166.

regular

polyhedroid.

The

regular
of

regular polyhedroid f equal regular polyhedrons together with their interiors, the polyhedrons being joined by their faces so as to enclose a portion of hyperspace, and the hyperplane angles formed at the faces by the half-hyperplanes of adjacent polyhedrons being all equal to one another.

pentahedroid.

consists

lar

regular polyhedroids, the regupentahedroid (Art. 114) and the hypercube (Art. 136). The interior of the segment which measures the altitude

We have already had two

of a regular pentahedroid is
is

one leg of a right triangle whose

the edge of one of its cells and whose other hypothenuse leg is the radius of the sphere circumscribed about the cell.

The

radius of the hypersphere circumscribed about the pentahedroid is equal to four-fifths of the altitude, and the
radius of the inscribed hypersphere is equal to one-fifth of the altitude. These theorems are proved in the same way

that the corresponding theorems are proved for the regular tetrahedron and the triangle.

Radii perpendicular to the

cells of

a regular pentahedroid
in five points

meet the circumscribed hypersphere

*

which

are the vertices of a second regular pentahedroid symmetriThis subject is treated by I. Stringham, "Regular Figures in -dimensional Space/' American Journal of Mathematics, vol. 3, 1880, pp. 1-14. t That is, a regular convex polyhedroid. We shall consider only convex polyhedroids in this chapter.

289

2QO

THE SIMPLER REGULAR POLYHEDROIDS

[vm.

i.

cally situated to the first with respect to the centre,

and

therefore equal to

(Art. 85, Th. The pentahedroid has

it

2).

5 vertices, 10 edges, 10 faces,

and

5 cells.

167.

by

The hypercube. The hypercube can be generated the motion of a cube in a direction perpendicular to its

hyperplane through a distance equal to its edge. The centfe of the cube generates the interior of a segment whose middle point is the centre of the hypercube, equally
distant from
all of

the sixteen vertices.

The

distance of

the centre from any cell (the radius of the inscribed hypersphere) is equal to one-half of the edge, the radius of the circumscribed hypersphere is equal to the edge, and the

diagonal of the hypercube (the diameter of the circumscribed hypersphere) is twice the edge (Art. 136, Th. 3). The hypercube has
1

6 vertices, 32 edges, 24 faces, and 8

usually
is

cells.

As the regular polyhedroids are number of cells, the hypercube

octahedroid.

named from

the

also called

a regular

We

can

fill

all

hyperspace with a set of hypercubes and

hypercubes coming together at any the centres of these hypercubes are Moreover, of a second set of hypercubes of vertices the themselves
their interiors, sixteen

vertex.

the same kind, and the two sets are in this

cally related.

way

recipro-

Two

hypercubes

of

one set whose centres

are the extremities of an edge of the other set have in common a cube, and the centre of the cube is the projection

upon

cubes.

hyperplane of the centres of the two hyperThe edge and the cube lie, therefore, in a perpenits

dicular line
is

and hyperplane, intersecting at a point which the middle point of the edge and the centre of the cube.

i66-i68]

HYPERCUBE AND I6-HEDROID

of

291

Four hypercubes

of a face of the other set

one set whose centres are -the vertices have in common a square. The

hypercubes are in cyclical order around the plane of this

its

square, and the centre of the square is the projection upon plane of the centres of the four hypercubes. More-

over, the vertices of this square are in turn the centres of four hypercubes of the second set, arranged in cyclical

order around the plane of the given face of this set. The planes of the two faces are therefore absolutely perpendicular, intersecting at a point which is the centre of both
(see Arts. 46

and

179).

set of polyhedroids filling

is

hyperspace without over-

every point of hyperspace is either a point of two or more of the polyhedroids or a point of the interior of only one. The two sets of hypercubes just
lapping
described are reciprocal nets.

called a net :

off

The hexadekahedroid or 16-hedroid. If we lay a given distance in both directions on each of four mutually perpendicular lines intersecting at a point 0,
168.

the eight points so obtained are the vertices of a regular polyhedroid which has four diagonal^ along the four given lines. In fact, the rectangular system contains sixteen

which

rectangular tetrahedroidal angles, and the four vertices lie on the edges of one of these angles are the vertices of

a regular tetrahedron congruent to the tetrahedron whose vertices lie on the edges of any other one of these

Thus the sixteen tetrahedroidal angles (see Art. 72, Th. i). form a polyheinteriors with their tetrahedrons together
0.

droid and enclose a portion of hyperspace about the point Now any face of this polyhedroid can be taken as the

base of each of the two tetrahedrons which have this face

in

common and

lie

in the cells of a

hyperplane angle of the

polyhedroid.

Thus the opposite

vertices of these

two

tet-

2Q2

THE SIMPLER REGULAR POLYHEDROIDS

lie

[vm.

i.

rahedrons
face,

and

as they

in the plane angle at the centre of the given lie also in one of the four lines of the rec-

tangular system', they and the centre of the given face are the vertices of an isosceles triangle whose sides and angles
are the

same whatever

face be taken.

The hyperplane

angles of the polyhedroid are, therefore, equal,* and the polyhedroid is regular. As there are sixteen cells the poly-

hedroid
write
it,

is called

a hexadekahedroid, or as

we

shall usually

a i6-hedroid.

Each tetrahedron has four faces, and each face is common The number of faces of the 16to two tetrahedrons.
hedroid
is

therefore

jc6_X_4
2

_ ~

32

'

Each

common
ties of

vertex, being one extremity of a diagonal, is a extremity of six of the edges. The other extremi-

these edges are the extremities of the other three The number of edges is therefore diagonals.

8X6 2

2A 24.

Thus the

regular i6-hedroid has

cells.

8 vertices, 24 edges, 32 faces, and 16

The
169.

diagonals of the i6-hedroid are also called

its axes.

Reciprocal relation of the hypercube and the 16hedroid. When a hypercube is inscribed in a hypersphere, the radii perpendicular to the hyperplanes of its eight cells
inscribed in the

are the radii to the eight vertices of a regular i6-hedroid same hypersphere, the two polyhedroids
lines.

having their axes along the same

*

Each vertex

of

the i6-hedroid and the centre of the hypersphere are symThey

are, in fact, angles of

120.

168-170]

HYPERCUBE AND I6-HEDROID

293

metrically situated with respect to the hyperplane of a cube, and the vertex is at a distance r from each of the
eight vertices of the cube, r being the radius of the hypershall speak of the vertex as corresponding to sphere.

We

the cube.

This relation is reciprocal. At a vertex of the hypercube there are four cubes lying in hyperplanes perpendicular respectively to the four axes
;

and the four

vertices of

vertices of one of its tetrahedrons.

the i6-hedroid corresponding to these four cubes are the These four vertices

are at a distance r from the vertex of the hypercube, and the radius to the latter point is perpendicular to the hyperplane of the tetrahedron, passing through its centre.

Therefore, the radii perpendicular to the hyperplanes of the sixteen cells of a i6-hedroid are the radii to the sixteen
vertices of a hypercube. The hypercube and the 16hedroid are said to be reciprocal polyhedroids (see Art. 177).

170.

The diagonals

of the

hypercube and the 16-hedroid

forming three rectangular systems. The hypercube has eight diagonals joining the eight vertices of any one cube to the eight vertices of the opposite cube. Since any
vertex of a i6-hedroid
is

at a distance r from each of the

eight vertices of the corresponding cube, the radius to the former makes an angle of 60 with each of the eight radii
to the latter.

Now

ciated in

the eight radii to the vertices of a cube can be assotwo sets of four each for on a cube the common
;

vertex of three adjacent squares and the three opposite vertices of these squares are the vertices of a tetrahedron

whose edges are diagonals of the six faces of the cube, one in each face, while the other four vertices of the cube are the vertices of a second tetrahedron of the same kind.

The diagonal

of a square subtends a right angle at the

2Q4

THE SIMPLER REGULAR POLYHEDROIDS

[vm.

i.

centre of the hypersphere, and the four radii of a set are

mutually perpendicular.
set,

Two

radii taken,

one from each

ties of

are radii to the extremities of an edge or to the extremia diagonal of the cube. Now an edge subtends

an angle of 60

at the centre, and a diagonal, its length

being r Vj, subtends an angle of 120. Therefore, any radius of one set makes an angle of 60 with three of the
radii of the other set,

and with the fourth an angle

of

120.

fact, the radius opposite to this fourth radius goes to the other extremity of the fourth edge of the hypercube at the given point. Thus we can say that the radii to the

In

extremities of the four edges which go out from a vertex of a hypercube are a set of four mutually perpendicular radii.

Putting these results together, we find that the four diagonals of the i6-hedroid and the eight diagonals of the

hypercube

lie in three sets of four mutually perpendicular line of one set making with each line of the other each lines, two sets the two supplementary angles of 60 and 120.

The

lines of

any one

of these three sets

can be taken as

the axis-lines of a hypercube, and the eight lines of the other two sets will pass through its vertices, for there are only eight lines that can make angles of 60 and 120 with the
half-line making four lines of a rectangular system. an angle of 60 with a half-line of a rectangular system will

make an angle of 30 with

is

the hyperplane to which the latter the and point at a distance r on it will perpendicular,

be at a distance - from the hyperplane.

2

But there are

just

sixteen points at the distance - from each of the four hy2

perplanes of a rectangular system, the sixteen points of

intersection of four pairs of parallel hyperplanes, and so the sixteen vertices of a hypercube which has its axes along

the four lines of the rectangular system.

I7Q> 171]

24-HEDROID

295

Associated with the twelve lines described above, there are, therefore, three hypercubes and three regular 16hedroids.

24-hedroid associated with a hypercube and a 16-hedroid. Given a hypercube and a regular i6-hedroid
171.

The

inscribed in the

lying in the same rectangular system, let vertices of the latter corresponding to the

same hypersphere and with their axes P and P be two two cubes which
1

have in common the square

A BCD.

Let O' be the centre

is at the distance r from each of of the square. Since the four vertices of the square, it determines with the square one-half of a regular octahedron. The same is

and indeed it is also true of the centre true of the point The altitude O'P of the pyramid of the hypersphere.
',

P-ABCD is

equal to -=, and

O'P

is

of the

same

V2

length.

subtends at O an angle of 90, and is Hence the line PP passes through O', and the points P and P', together with the square, lie in one hyperplane, and are the vertices of a regular octahe-

But the segment

PP

of twice this length.

dron with centre at 0'. From a cube can be formed

six equal

common
of the
is

vertex at the centre of

pyramids having a the cube and the six faces

The interior of the square base of two such pyramids, one from each of the two cubes which have this square in common. These
cube as bases.
the

ABCD

common

of

two pyramids do not lie in one hyperplane, nor does either them belong to a regular octahedron, but the two pyramids

are the projections of these pyramids of the the octahedron, projected from the upon hyperplane the of centre hypersphere. At a vertex P there are six

of the octahedron

PP

quadrangular pyramids, belonging to

drons, and the projections
of the six

six regular

octahe-

pyramids which have

296

THE SIMPLER REGULAR POLYHEDROIDS

[vra.

i.

common

The

interiors of these six

vertex at the centre of the corresponding cube. are the lateral pyramids at

cells of

cube, that

a hyperpyramid whose base is, a cell of the hypercube.

is

the interior of the

the cubes of the hypercube are formed in all twenty-four pairs of pyramids, the two pyramids of a pair having a common base. The interiors of these pyramids

From

and

their bases

can be projected as above from the centre

O into

the interiors of the twenty-four regular octahedrons, the cells of a polyhedroid which can be built up by placing a hyperpyramid as described above upon each cell of the

hypercube.

This polyhedroid

is

called

an

ikosatetrahe-

Its vertices are the sixteen vertices droid or 24-hedroid. of the hypercube and the eight vertices of the i6-hedroid, twenty-four in all. This polyhedroid can be built up from

any one and

it

of the three

hypercubes associated with a set of

is

twelve lines such as

described in the preceding article, has the same number and arrangement of parts at
it

a vertex of the hypercube that

hedroid.

has at a vertex of the 16-

To

hedroid

get the plane angle of a hyperplane angle of the 24we draw half-lines through the centres of two ad-

jacent octahedrons from the centre of their common face. One common vertex of the two octahedrons can be taken
as the point P, and their centres are the centres of two adjacent squares of the cube to which corresponds. These two points and the centre of the common face of the octa-

hedrons are the vertices of an isosceles triangle whose

sides are equal respectively to the sides of the isosceles triangle formed in the same way from any other pair of

adjacent octahedrons.
regular.
*

of the polyhedroid are all equal,*

Therefore the hyperplane angles and the polyhedroid is

These are angles of 120,

like those of the i6-hedroid.

i/i, 172]

RECIPROCAL 24-HEDROIDS
eight faces. therefore

297
of faces

The octahedron has

of the 24-hedroid
is

The number

24

7likewise 96.

=96.
is

Eight edges meet at a vertex, and the number of edges

Thus the regular 24-hedroid has

24 vertices, 96 edges, 96 faces, and 24
172.
cells.

Reciprocal 24-hedroids. There are twenty-four points O', the centres of the twenty-four octahedrons, the centres of the twenty-four squares of the hypercube,

and the middle points of the twenty-four edges of the 16hedroid. These twenty-four points lie on twelve lines
through 0.
Since the points 0' are the middle points of the edges of a i6-hedroid, the half-lines OO' bisect the twenty-four
right angles formed

by the
lie

four axes of this i6-hedroid.

in

Now

these right angles

perpendicular planes, and

three pairs of absolutely the half -lines which bisect the

eight right angles in any one of these three pairs of planes are themselves the half-lines of a set of four mutually perpendicular lines. Our twelve lines, therefore, consist
of three sets of four
If

mutually perpendicular
lie

lines.

in perpendicular planes and have right angles a common side along the intersection of these planes, their bisectors form an angle of 60. Now any two of the

two

not absolutely perpendicular, are perpendicular, intersecting in one of the four Therefore each of the bisectors of the four right lines.
six planes of
if

a rectangular system,

angles in one of these planes makes with each of the bisectors of the four right angles in the other an angle of 60 or

an angle

of

120.

208

THE SIMPLER REGULAR POLYHEDROIDS

line in

[vm.

i.

This proves that a

any one

of our three sets of

four mutually perpendicular lines makes with each of the eight lines of the other two sets the supplementary angles of 60 and 120, and that the radii of the hypersphere drawn

through the points

meet the hypersphere

in twenty-

four points which are the vertices of a second regular 24hedroid. The two 24-hedroids are related in the same way

as a hypercube and a i6-hedroid whose axes lie along the same four lines; namely, the radii perpendicular to the

hyperplanes of the cells of one are the radii to the vertices The two 24-hedroids are said to of the other (Art. 169). be reciprocal (see Art. 177).
173. The reciprocal nets of 24-hedroids and 16-hedroids. The interior of a cube of a hypercube is the base of a hyper-

pyramid with vertex at the centre O, congruent to the hyperpyramid built upon the same base with vertex at the corresponding point P *. From the hypercube can be formed eight such hyperpyramids, and therefore eight
polyhedroidal angles like that of the 24-hedroid at P can be placed together with a common vertex so as to fill the

hyperspace about this vertex. In other words, eight 24hedroids can be placed together at a point with their interiors filling the hyperspace about this point, and we have
in hyperspace a net of 24-hedroids like the net of hypercubes described in Art. 167. The centres and vertices
of the 24-hedroids are, indeed, the centres and vertices of a net of hypercubes. The eight polyhedroidal angles at

have their axes lying in the axis-lines of the hypercube. Therefore, the eight 24-hedroids put together at any point

have axis lines through this point forming a rectangular system, and their centres are the vertices of a i6-hedroid whose centre is this point.
*

The two hyperpyramids

of the cube,

and are congruent.

are symmetrically situated with respect to the centre Thus the hypervolume of the 24-hedroid is twice

that of the hypercube.

i?2, i73l

RECIPROCAL NETS

299

Let us suppose that the given 24-hedroid with centre at is one of eight 24-hedroids put together at a vertex P,

so that

is

the centre of the i6-hedroid and one of its

The edges of the i6-hedroid pass through the centres of certain cells of the 24-hedroids. In particular, the edges which come to O pass through points
vertices is at O.

which we have called 0', through those six which lie on the cube corresponding to P. Now the points O' which lie on a single cube lie on radii to the vertices of a single
cell of

ing article.
is

the reciprocal 24-hedroid constructed in the precedThis cell is therefore the base of a hyper-

pyramid with vertex at

just

O whose polyhedroidal angle at O the polyhedroidal angle at this point of the i6-hedroid
mentioned.

Therefore, twenty-four such polyhedroidal angles, and so twenty-four i6-hedroids, can be put together at a point with their interiors filling the hy-

perspace about this point; and we have in hyperspace a net of i6-hedroids reciprocal to the net of 24-hedroids.
adjacent 24-hedroids have in common an octaheits centre 0' the projection upon its hyperplane of the centres of the two 24-hedroids. These

Two

dron PP', with

centres are the extremities of an edge in the net of i6-hedroids. Thus the edges of the i6-hedroids correspond to

the

cells of

the 24-hedroids, an edge and a

cell

lying in a

and hyperplane, and intersecting at a is which the middle point of the edge and the centre point of the cell. Three 24-hedroids whose centres are the verperpendicular line
tices of

each having a

a face of the net of i6-hedroids are in cyclical order, cell in common with the next, and the hyper-

planes of the three cells, being perpendicular to the sides of the given face at their middle points, pass through the centre and intersect in the plane absolutely perpendicular
to its plane at this point. In other words, the three 24hedroids have in common a triangle and the face which is

300

THE POLYHEDROID FORMULA

[vra. n.

the interior of this triangle. The vertices of this triangle are in turn the centres of three of the i6-hedroids having the given face in 'common, and arranged in cyclical

order around the plane of this face. Each of these three i6-hedroids has in common with the next a cell whose
projection upon its hyperplane of the two i6-hedroids, so that each edge of the net of 24-hedroids and the corresponding cell of the net of 16hedroids lie in a perpendicular line and hyperplane, and intersect at a point which is the middle point of the edge and the centre of the cell (see Art. 179).

centre

is

the

centres of the

II.

THE POLYHEDROID FORMULA

Extension of the polyhedron formula. In a simple polyhedron the number of vertices and faces taken together
174.
is 2

more than the number

of edges.*

This relation can

be expressed very conveniently in the form

i

- No +

#1

- Ni +

o,

where

is

the

number

of vertices,

Ni the number

of

edges, and Nz the number of faces. For the five regular polyhedrons this formula becomes
for the tetrahedron
i i

"

4+
8

6
12

hexahedron or cube
octahedron

"
"

i
1 1

dodekahedron
ikosahedron

20

"

+ + +

4 6
8
12

12

30

12+30

20

+ 1=0, + 1=0, + 1=0, + 1 = o, + 1 =o.

*This theorem was discovered by Euler about 1750 and usually goes by his name. It was known to Descartes more than a century earlier at least it follows " left by Desdirectly from formulae in a manuscript, "De Solidorum Elementis, This memoir was not published, however, until 1860, a copy having been cartes. found only a few years earlier among the papers of Leibnitz (CEuvres infdites de See comDescartes, par M. le Comte Foucher de Careil, Paris, 1860, vol. II, p. 214. munication by Prouhet in the Comptes Rendus, vol. 50, 1860, p. 779, and several by E. de Jonquieres in vol. no, i8go).

173, i74l

PROOF FOR POLYHEDROIDS

first

301

In the

three cases the

first

members of these equations

are the expansions of

(i

i)

-(2-1)', and(i -2)^

their

1.

The polyhedrons take

to certain

names from the numbers N%.

so as to apply
to polyhedroids

The above formula can be generalized more complicated figures and

in space of four dimensions.

For a simple polygon we can write a similar formula,

namely,
i

- No + Ni -

o.

In proving the formula for a polyhedron we think of the polyhedron as built up by putting together a set of polygons,
taking them in succession in such order that each is joined to those already taken by a side or by two or more sides
line. As long as we have not comof enclosed any portion space and formed a polypletely hedron we have the relation

forming a single broken

- # + Ni - Ni =

o.

In the same

way we can build up a polyhedroid by putting

joined to those already taken by a set of polygons like the incomplete polyhedron. At each stage of the process we add to the number of veris

together a set of polyhedrons. sion in such order that each

We

take them in succes-

and faces already obtained the number of vertices, edges, and faces of the new polyhedron, and to the number of cells we add i and we subtract the number of vertices, edges, and faces of the connecting figure, which otherwise would be counted twice. Assuming that the figure had an equation of the form
tices,

edges,

- #0 + # - #2 + # 3 =
1

O,

302

THE POLYHEDROID FORMULA

[vm. n.

where N* is the number of cells, we add all but the first term of the equation of the new polyhedron, and then subtract all but the first term of the equation of the connecting
figure.

Thus we prove by induction that

is

as long as the

polyhedroid sumed.

incomplete

its

equation

is of

the form as-

When we come

to the last polyhedron there are

vertices, edges, or faces. Only the by i, and the equation will be true
i

no new number N 9 is increased if we write it

tfo
is

+ N l

N*

#3

o.

This equation
call it

true for a simple polyhedroid,

and we

shall

the polyhedroid formula. The relation can be stated as a theorem in the following words
:

In a simple polyhedroid the number of cells plus the number of edges is equal to the number of faces plus the number of vertices* For the four regular polyhedroids already considered the polyhedroid formula becomes
for the pentahedroid
i i i 1

THEOREM.

"

'

" "

hypercube
i6-hedroid

16

24-hedroid

24

+ + + +

1010+
32
24 24

1=0,
i

32

96

96

+ + +

8
16

i
1

24

= = =

o, o,
o.

In the

first

three cases the

first

numbers

of these equa-

tions are the expansions of

(i

-i),

In
*

all

cases the

4 -( 2 -i)*, and(i-2) -i. name of the polyhedroid comes from

the

number N*.
An interesting discussion of this law and its extension to geometry of higher dimensions is given in Schoute's Mehrdimensionale Geometric, vol. II, 2. The corresponding law for simple polyhedroids of any number of dimensions is proved by Stringham in the article referred to on p. 289. The theorem will not be used in proving the existence of any regular polyhedroid, nor even in computing the number of any of its elements. Its proof may be omitted by the student, but we have in

the formula

itself

a convenient mode of expressing these numbers.

174, i75l

NETS ON THE SPHERE

303

III.

RECIPROCAL POLYHEDROIDS AND RECIPROCAL NETS OF POLYHEDROIDS

Reciprocal polyhedrons and nets of polyhedral In geometry of three dimensions a regular polyangles. hedron can always be inscribed in a sphere. This is proved
175.

proved for tetrahedrons, and the proof holds also in the Elliptic Non-Euclidean Geometry,
in the

same way that

in the

it is

of the

geometry of the hypersphere. The vertices polyhedron are the vertices of a net of equal regular spherical polygons, a net of spherical polygons being a set

and so

covering the sphere so that every point of the latter

is

either a point of two or more of the polygons or a point of the interior of only one. Conversely, the vertices of a

net of equal regular spherical polygons are always the vertices of a regular polyhedron inscribed in the sphere. Since the sphere is the same in the Elliptic Geometry, the
regular polyhedrons of this geometry are of the types as those of Euclidean Geometry.*

same

When we have a net of equal regular spherical polygons, the centres and vertices of these polygons are the vertices
and centres
is

of a second net of equal regular polygons

reciprocal to the

And so when a regular polyhedron first. inscribed in a sphere, radii perpendicular to the planes of the faces are radii to the vertices of a second regular

polyhedron. Two polyhedrons which can be so placed that half-lines from the centre perpendicular to the planes of the faces of one are half -lines from the centre through the
vertices of the other, are said to be reciprocal polyhedrons.
* It can also be proved in the Hyperbolic Geometry that a regular polyhedron can be inscribed in a sphere, and that we have just the same types of regular poly-

hedrons.

But

in the

Hyperbolic Geometry

we have

also boundary-surfaces

and

equidistant-surfaces on which we can form other nets of polygons ; and the vertices of such a net are the vertices of an infinite broken surface formed of equal regular polygons and their interiors, like a regular polyhedron.

304

RECIPROCAL POLYHEDROIDS AND NETS

of edges,

[vm. m.

They have the same number

vertices of one other.
is

the same as the

and the number of number of faces of the

is

The polyhedron formula of one formula of the other written backwards.

Half-lines

the polyhedron

regular polyhedron through the vertices are the edges of equal regular polyhedral angles forming a net of polyhedral angles in the two-

from the centre

of

dimensional Point Geometry at this point, every halfline element of the Point Geometry being either a half-line

two or more of the polyhedral angles, or a half-line of the interior of only one. Corresponding to two reciprocal polyhedrons and to two reciprocal nets on the sphere we
of

reciprocal nets of polyhedral angles, the edges of the polyhedral angles of one net being the axes of the polyhedral angles of the other net. Conversely, a net of

have two

equal regular polyhedral angles in a hyperplane intersects any sphere of the hyperplane with centre at the vertex of
the net in a net of equal regular spherical polygons, and the edges of the polyhedral angles pass through the vertices of

a regular inscribed polyhedron which corresponds

to the net.

Reciprocal nets of polyhedrons. In a net of equal regular spherical polygons the angles at a vertex P are equal, vertices adjacent to P are vertices of a regular
176.

spherical polygon,

and the corresponding inscribed poly-

hedron has a regular polyhedral angle at P. If, then, a set of equal regular polyhedrons with their interiors fills a three-dimensional space about a point so that their
polyhedral angles at this point form a net, the centres of the polyhedrons, lying at a given distance on the axes
of of

these polyhedral angles,

are themselves the vertices

a regular polyhedron, reciprocal to the polyhedron which

corresponds to the net.

With any given net

of regular

I7S.I76]
is

NETS OF POLYHEDRONS

305

polyhedrons
of the

associated a reciprocal net, the vertices polyhedrons of one net being the centres of the

polyhedrons of the other. Any edge of a polyhedron of one net, joinitfg the centres of two polyhedrons of the other net, and the common face of these two polyhedrons,
lie

in a perpendicular line and plane and intersect at a point which is the middle point of the edge and the centre of the The number of vertices in a polyhedron of one face.

net

is

equal to the

at a vertex.

number of polyhedrons of the other The polyhedrons of two reciprocal nets

net
are

not, in general, reciprocal polyhedrons. the polyhedral angles of the polyhedrons of a net

Now

must be such

as occur in a net of polyhedral angles at the

centre of a regular polyhedron.

at the centre of a tetrahedron

Thus we have
4
6 8
12
trihedral angles,

" " " "

a cube

tetrahedral
trihedral

"

an octahedron a dodekahedron an ikosahedron

"

"
pentahedral
trihedral

20

"

The polyhedral
as follows
:

angles of the five regular polyhedrons are

in the tetrahedron trihedral angles,

"

cube
octahedron tetrahedral

"
"

dodekahedron trihedral
ikosahedron pentahedral

" "
"

"

Therefore the only sets of regular polyhedrons that can be used to form nets are
O

4 tetrahedrons, cubes, or dodekahedrons at a point, it U

20
6 octahedrons
12 ikosahedrons

" "

306

RECIPROCAL POLYHEDROIDS AND NETS

[vm.

m.

In two reciprocal nets the number of vertices in a polyhedron of one net is equal to the number of polyhedrons of the other net at a- point. These nets of polyhedrons are
associated, therefore, as follows
:

4 cubes reciprocal to 8 tetrahedrons, " " 20 dodekahedrons 4 " " 20 cubes; 8

also

4 tetrahedrons reciprocal to a net of the same kind, 8 cubes " 20 dodekahedrons

6 octahedrons
12 ikosahedrons

"

"

"

These theorems are true of the Non-Euclidean Geometries

as well as of Euclidean Geometry, but in the Hyperbolic Geometry the angles of a regular polyhedron are smaller,

and
in

in the Elliptic

Geometry they are

case as

larger,

than they are

the
the

Euclidean Geometry.
first

angles in the second case as large as

is

Moreover, we can make small as we please, and in

we

figures sufficiently large.

up Thus any
please

to

180, by taking the

of these combinations

possible in one of the three geometries, at least for a restricted portion of space, and two reciprocal nets must occur in the same kind of geometry. Any combination

which more than fills the part of Euclidean space about a point belongs to Hyberbolic Geometry, and any combination which does not fill the part of Euclidean space about a point belongs to Elliptic Geometry. Now in Euclidean Geometry we have a net of cubes,
eight at a point. Then eight dodekahedrons, twenty cubes, or twenty dodekahedrons would more than fill Euclidean space, and nets of these types must belong to the Hyperwhile nets with four cubes, four tetrahebolic Geometry
;

NETS OF POLYHEDRONS

307

drons, or eight tetrahedrons will belong to the Elliptic can also prove that the polyhedral angles Geometry.

We

of a regular octahedron in Euclidean Geometry are smaller than those of a net at the centre of a cube, so that the net

of octahedrons, six at a point, belongs to Elliptic Geometry. For example, the faces of the octahedron are equilateral triangles, while the centre and two adjacent vertices of a

cube are the vertices of an isosceles triangle in which the legs, each being the half of a diagonal, are shorter than the
base.

There remains to be considered the net of ikosahedrons, twelve at a point, and the reciprocal nets of four dodekahedrons and twenty tetrahedrons. In the net of twelve pentahedral angles and in the net of four trihedral angles there are three of these angles around an edge, and the dihedral angles must all be angles of 120. Therefore in
the net of ikosahedrons and in the net of dodekahedrons the dihedral angles
to determine in

must be angles which geometry

of

120.

Thus we have

the dihedral angles of the ikosa-

hedron, and in which geometry the dihedral angles of the dodekahedron, are angles of 120.

Let

ABC

be a spherical

tri-

angle of the net corresponding

to the ikosahedron, its centre the (on the sphere), and

A
2

middle point of the arc

AB.

Then

OD is the supplement

of the dihedral angle of the ikosahedron, and 2 is the supplement of the dihedral angle of the dodekahedron,

AD

Now

the dodekahedron being reciprocal to the ikosahedron. two sides of the triangles on opposite sides of the sphere lie on a great circle which also crosses four of

ABC

these triangles.

Hence, we have four arcs equal to OD,

308

RECIPROCAL POLYHEDROIDS AND NETS

[vra.

m.

four equal to
Therefore,

AD, and

four equal to

AO

on such a

circle.

OQ + AD + AO =
But
in the triangle

90.
60, and 90,

AOD the angles are 36,

OD< AD<
and
Again, in the triangle

AO,

OD <

30.

ACD the angles are 72, 36, and 90,

and

CD<
That
and
is,

AC.
2

OD + AO <

AD,

therefore

AD >

30.
Geometry the dihedral
;

It follows that in the Euclidean

angles of the ikosahedron are greater than 120, and the net but the of ikosahedrons belongs to Hyperbolic Geometry
dihedral angles of the dodekahedron are less than 120, so that the net of dodekahedrons and the reciprocal net of tetrahedrons belong to Elliptic Geometry.*

Summing up we

find that
;

Euclidean space can be filled with cubes, eight at a point Hyberbolic space can be filled with dodekahedrons, eight

at a point or twenty at a point, with cubes, twenty at a and point, or with ikosahedrons, twelve at a point Elliptic space, or at least any restricted portion of elliptic
;

space, can be filled with tetrahedrons, four at a point, eight at a point, or twenty at a point, with cubes, four at a point, with dodekahedrons, four at a point, or with octa-

hedrons, six at a point.

In the net of twenty trihedral angles the dihedral angles are angles of 72, and can be proved that the dihedral angles of the tetrahedron in Euclidean Geometry are less than 72, so that the net of twenty tetrahedrons'at a point belongs, indeed,
it

to the Elliptic Geometry.

i?6, i77l

NETS IN THE HYPERSPHERE

Reciprocal polyhedroids.

309
polyhedroid

177.

regular

can always be inscribed in a hypersphere.

For the perpen-

diculars to the hyperplanes of the cells at the centres of the cells all pass through a point which is at the same distance

on each
all

them, and, therefore, at the same distance from the vertices of the polyhedroid. This is proved in the
of
it is

same way that

Th.

proved for pentahedroids

(see Art. 113,

Since the hyperplane angles are all equal, no two i). adjacent cells lie in the same hyperplane, and the perpendiculars to the hyperplanes of two adjacent cells cannot be

parallel

and must therefore

intersect.*

regular polyhedroid is inscribed in a hypersphere, are the vertices of a net of equal regular hyperFor the vertices of any polyhedron spherical polyhedrons. of the polyhedroid lie on a sphere which lies entirely in the
its vertices

When a

hypersphere, and so they are the vertices of a regular hyperThe polyhedrons spherical polyhedron of the same type.
are equal, for the spheres are equal, and with their interiors they completely fill the hypersphere. They can, indeed, be regarded as the projections of the polyhedrons of the

given polyhedroid, projected by radii from the centre of the hypersphere. Conversely, the vertices of a net of equal
regular hyperspherical polyhedrons are always the vertices of a regular polyhedroid inscribed in the hypersphere.

The

on a sphere and are the vertices of a regular hyperplane polyhedron in the hyperplane of the sphere. These hyperplane polyhevertices of

any one

of the polyhedrons

lie

drons are

all

equal, with their interiors they enclose a por-

tion of hyperspace, and the hyperplane angles formed by the hyperplanes of any two which are adjacent are all equal.

When a regular polyhedroid is inscribed in a hypersphere,

*

The theorem

is

true in Elliptic

Geometry

of four dimensions, since

It

any two

lines in

a plane of Elliptic Geometry intersect.

can also be proved in Hyperbolic

Geometry.

310

RECIPROCAL POLYHEDROIDS AND NETS

[vm. m.

so that its vertices are the vertices of a net of equal regular hyperspherical polyhedrons, radii perpendicular to the hy-

perplanes of

its colls will

be
;

radii to the centres of these

is, they will be radii to hyperspherical polyhedrons the vertices.of the reciprocal net, to points which are, thereThe fore, the vertices of a second regular polyhedroid.

that

relation of the
lar

two polyhedroids is reciprocal and two regupolyhedroids which can be so placed that half -lines from
;

one are

the centre perpendicular to the hyperplanes of the cells of half-lines from the centre through the vertices of

the other, are called reciprocal polyhedroids. The number of vertices of one is equal to the number of cells of the other,

and the number

of edges of

one

is

equal to the

number
of

of
is

faces of the other.

The polyhedroid formula

one

the polyhedroid formula of the other written backwards. Moreover, the number of cells at a vertex of one equals the

number

of vertices to a cell of the other, the

number

of

of the other,

edges at a vertex of one equals the number of faces to a cell and so on. Whenever we have constructed

a regular polyhedroid, or proved its existence, we have proved the existence of a reciprocal polyhedroid. The two may, however, be polyhedroids of the same type.

The hypercube and the regular i6-hedroid are reciprocal polyhedroids. The regular pentahedroid and the regular
to the

These cases correspond 24-hedroid are self-reciprocal. two reciprocal nets of four cubes at a point and eight

tetrahedrons at a point, and to the two self-reciprocal nets of four tetrahedrons at a point and six octahedrons at a

have found that in Elliptic space of three dimensions, or at least in a restricted portion of such space, and so in the hypersphere or in a restricted portion of the
point.

Now we

hypersphere, there exists another pair of reciprocal nets, nets with four dodekahedrons at a point and twenty tetrahedrons at a point and so in Euclidean space of four di;

1771

THE N- AND SN-HEDROTD

311

mensions there can be at most only two types of regular polyhedroids besides the four which we have already found.
In the next section we shall construct a regular polyhedroid with twenty tetrahedrons at a point and a regular polyhedroid with four dodekahedrons at a point, one of these constructions being necessary to complete the proof of the

existence of the two.

Assuming that these polyhedroids exist, we know that they are reciprocal and that the polyhedroid equation of
one
is

wards.
cells,
it

the polyhedroid equation of the other written backLet us suppose that the second polyhedroid has

that

is,

that

it

contains

dodekahedrons, and

call

an N-hedroid. Each dodekahedron has twenty and at each vertex there are four dodekahedrons.

vertices,

There-

fore the

number

of vertices is

The
5

first

polyhedroid

is,

then, a $ N-hedroid, containing

in the AT-hedroid

N tetrahedrons.
The dodekahedron has twelve faces, and
is

each face
the

common

to

two dodekahedrons.

Therefore

number

of faces in the A^-hedroid is

In the same
is

way

the

number

of faces in the

found to be
2

Thus the polyhedroid formulae

of the

two are
i i

iand

6N - loN + $N i-sN+ioN- 6AT + N -

N+

= -

o,
o.

312
178.

RECIPROCAL POLYHEDROIDS AND NETS

Regular
is

[vin. in.

polyhedroidal

angles.

regular

poly-

one subtended at the centre of a hypersphere by a regular hyperspherical polyhedron whose circumscribed sphere is not a great sphere. But the hyhedroidal angle

perspherical polyhedron is the projection from the centre of the hypersphere of a hyperplane polyhedron which has the

same

vertices,

and through the

vertices of a given hyper-

pass a hypersphere having its centre at any point in the line perpendicular at the centre of the polyhedron to its hyperplane. Therefore the polyhedroidal angle subtended by a regular polyhedron at any point in the line drawn through the centre of the polyhedron perpendicular to its hyperplane, except at the centre

plane polyhedron we may

a regular polyhedroidal angle. The polyhedroidal angle at the vertex of a regular hyperpyramid is a regular
itself, is

polyhedroidal angle.

The
hedron,
line

half-line

drawn from the vertex

of a regular poly-

hedroidal angle through the centre of the subtending polyor, in the case of a regular hyperpyramid, the half-

which contains the axis of the hyperpyramid,

is

the

axis of the polyhedroidal angle. Let be the vertex of a regular polyhedroidal angle

and

0' the centre of the subtending polyhedron. If we project the polyhedroidal angle by orthogonal projection upon the
hyperplane of the polyhedron, the vertex
at the centre 0',
will

and the polyhedroidal angle

on

be projected will be pro-

jected into the net of equal regular polyhedral angles at

produced, and with a hypersphere passing through P. Radii from O will project the polyhedroidal angle into a net of equal regular hyperspherical polyhedral angles at
this point.

Take a point

PO

as centre construct

congruent to the net at O'. The polyhedral angles at P in the hypersphere can be regarded as in the tangent

hyperplane at P.

Any one

of the half-lines

from

P in

this

178]

POLYHEDROIDAL ANGLES

313

hyperplane, and the corresponding half-line from O', lie * in the intersections of a projecting plane through O with

two parallel hyperplanes, and are parallel (Art. 128, Ths. 4 and 8). Conversely, suppose we have a net of equal
regular hyperspherical polyhedral angles at P.
ject these
If

we

pro-

by

radii

at

some point

gruent net at O', centre at O' will intersect the edges of the polyhedral angles in the vertices of a regular polyhedron which subtends at

upon a hyperplane perpendicular to OP between O and P, we shall have a conand any sphere in this hyperplane with

a regular polyhedroidal angle.

regular polyhedroid is projected by radii upon the circumscribed hypersphere in a net of equal regular hyperspherical polyhedrons, and the edges of the polyhedroid

which come to a vertex

lie

are projected into arcs which

in the edges of the hyperspherical polyhedral angles these edges are of the same length, of the net at P. their extremities are the vertices of a -regular polyhedron,

Now

of the given polyhedroid subtended is a regular polyhedroidal angle by this polyhedron. In other words, the polyhedroidal angles at the

and the polyhedroidal angle at

vertices of a regular polyhedroid are regular polyhedroidal

angles.

The

half-line

from a vertex
is

through the centre at the vertex.

of a regular polyhedroid the axis of the polyhedroidal angle

Half -lines drawn from the centre of a regular polyhedroid through the vertices are the edges of a set forming a net
of polyhedroidal angles in the Point

every half-line of the Point

Geometry at this point, Geometry being a half-line of

a half-line of

two or more

of the polyhedroidal angles or

the interior of only one. Corresponding to two reciprocal nets of polyhedrons in the hypersphere, we have two re*

See foot-note,

p. 84.

314

RECIPROCAL POLYHEDROIDS AND NETS

[vnnn.

ciprocal nets of polyhedroidal angles, the edges of the polyhedroidal angles of one net being the axes of the polyhedroidal angles of the. other net. Conversely, a net of equal

regular polyhedroidal angles intersects any hypersphere with centre at the vertex of the net in a net of equal regu-

and the edges of the polyhedroidal angles the vertices of a regular inscribed polyhepass through droid which corresponds to the net.
lar polyhedrons,
If a set of equal vertex at P, with a common regular polyhedroids having their interiors, fill the part of hyperspace about this point so that their polyhedroidal angles at P form a net, the centres of the polyhedroids, lying at a given distance from P

179.

Reciprocal nets of polyhedroids.

on the axes

of these polyhedroidal angles, are themselves the vertices of a regular polyhedroid with centre at P, reciprocal to the polyhedroid which corresponds to the
net. With any given net of regular polyhedroids is associated a reciprocal net, the vertices of the polyhedroids of one net being the centres of the polyhedroids of the other.

Any edge
of

of a polyhedroid of

one net, joining the centres

two polyhedroids of the other net, and the common cell of these two polyhedroids, lie in a perpendicular line and hyperplane, and intersect at a point which is the middle point of the edge and the centre of the cell. Those polyhedroids of one net whose centres are the vertices of a face of the other net, have in common a face in a plane absolutely perpendicular to the plane of the given face at a point which is the centre of both faces. The vertices of

the second face are in turn the centres of polyhedroids of the other net which have the given face in common (see
Arts. 167 and 173). The number of vertices in a polyhedroid of one net is equal to the number of polyhedroids of the other net at a vertex. The polyhedroids of two

I7,

i79l

NETS OF POLYHEDROIDS

315,

reciprocal nets are not, in general, reciprocal polyhedroids.

Whenever we have constructed a net

of equal regular poly-

hedroids, or proved the existence of such a net,

we have

proved the existence of a reciprocal net. Now the polyhedroidal angles of the polyhedroids of a net must be such as occur in a net of polyhedroidal angles
at the centre of a regular polyhedroid. at the centre of a pentahedroid "

Thus we have
"

5 tetrahedroidal angles,

"

a hypercube a i6-hedroid
a 24"

8 6-hedroidal
16 tetrahedroidal

"

24 8-hedroidal

as#an N-

"
"

sN

tetrahedroidal

i2-hedroidal

The

polyhedroidal angles of the six regular polyhedroids

:

are as follows

in the pentahedroid "

tetrahedroidal angles,

" "
"

hypercube
i6-hedroid
8-hedroidal
6-

"
24-

SNN-

"
"

20-

"

tetrahedroidal

"

Therefore the only sets of regular polyhedroids that can be used to form nets are
5

pentahedroids, hypercubes, or ^V-hedroids at a point,

16

8 24-hedroids

"

24 16-

"

The number

of vertices in a polyhedroid of one of two reciprocal nets is equal to the number of polyhedroids of the

other at a point. These nets of polyhedroids are associated, therefore, as follows

:

316
5

RECIPROCAL POLYHEDROIDS AND NETS

hypercubes reciprocal to 16 " s# S #-hedroids
16
pentahedroids,

[viu. in.

N-

"
"

"

8 24also

"

$N hypercubes,
24 i6-hedroids;

5 pentahedroids reciprocal to a net of the same kind, 16 hypercubes

These theorems are true

of the

Non-Euclidean Geometries

as well as of Euclidean Geometry.

As

in the case of poly-

fill

hedrons (Art. 176), those combinations which more than the part of Euclidean hyperspace about a point belong to Hyperbolic Geometry, and those which do not fill the

part of Euclidean hyperspace about a point belong to Elliptic

Now

Geometry. in Euclidean Geometry we have a net of hyper-

Then sixteen AT-hecubes, sixteen at a point (Art. 167). or of of either these droids, 5 polyhedroids, would more

the Euclidean hyperspace, and nets of these types must belong to Hyperbolic Geometry; while nets with

than

fill

five

hypercubes, five pentahedroids, or sixteen pentahe-

droids, will belong to Elliptic Geometry. have also in Euclidean hyperspace the reciprocal nets of eight 24-hedroids and twenty- four i6-hedroids

We

There remain, therefore, to be considered only the reciprocal nets of five N-hedroids and 5^ pentahe(Art. 173).

droids.

We

shall find
its

when we have constructed

the 5

N-

hedroid that

edge subtends an angle of 36 at the centre (see Art. 182), and therefore the hyperplane angles of the AMiedroid in Euclidean hyperspace are the supplements

^-hedroid being reciprocal to the But in a net of five ^-hedroids there are S^V-hedroid. four of these around any face, and the hyperplane angles
of

36,

or 144, the

PLATE

I.

FIG.

2.

FIG.

i.

FIG. 4.
FIG.
3.

FIG.

5.

FIG.

6.

179, i8o]

600-HEDROID
of

317

must be angles

90.

belong to the Hyperbolic

Therefore these reciprocal nets Geometry. Or, we may say that

the face angles of the pentahedroid in Euclidean hyperspace are angles of 60, so that 5JV pentahedroids would more

than

fill

Euclidean hyperspace about a point.

up,

Summing

we

find that

Euclidean hyperspace can be filled with hypercubes, sixteen at a point, with 24-hedroids, eight at a point, or

with i6-hedroids, twenty-four at a point; Hyperbolic hyperspace can be filled with AMiedroids, at a point, with five at a point, sixteen at a point, or 5

pentahedroids,
at a point
;

5#

at a point, or with

hypercubes,

$N

and

Elliptic hyperspace, or at least

any

restricted portion

of elliptic hyperspace, can be filled with pentahedroids,

five at a point, or sixteen at
five at

a point, or with hypercubes,

a point.*

IV.

CONSTRUCTION OF THE REGULAR 6oo-HEDROID AND THE REGULAR I20-HEDROID

First half of

proving its existhyper sphere a net of equal at a regular tetrahedrons, twenty point, with their interiors completely filling the hypersphere. We have already
:

180.

the 600-hedroid

ence.

We

shall construct in a

proved that we can in this way fill the part of the hyper" " restricted sphere about a point, and so any portion of the hypersphere, with tetrahedrons. What we have to

show now
will
*

is

that a certain

number

of these tetrahedrons

fill

the entire hypersphere without overlapping.

in space of five dimensions there are only three possible types of regular

Thus

(convex) figures: the simplex, corresponding to the tetrahedron and pentahedroid, the orthogonal, corresponding to the cube and hypercube, and

the figure reciprocal to the latter, constructed on a set of mutually perpendicular diagonals and corresponding to the octahedron and i6-hedroid.

318

THE 600-HEDROID AND I2O-HEDROID

[vm.

iv.

is a three-dimensional space (Art. 122), our attention at any one time upon a suffifixing ciently small portion of it we can carry on our processes as if it were the space of our experience. We shall use the

The hypersphere

and by

language of the ordinary three-dimensional geometry and say "line" and "plane" instead of "great circle" and
"
great sphere.
"

We

start with a regular ikosahedron

i).

made up

of

twenty

tetrahedrons (Plate I, fig. any one of the vertices.

well

A be the The radii A B, as

Let

centre and

as the edges BB, are edges of the About each edge there are tetrahedrons.

five tetrahedrons.

adjacent tetrahedrons form a double triangular pyramid,* the common face of

call

Two

the two tetrahedrons being the interior of what a cross section of the double pyramid.

we may

Five tetrahedrons about an edge form a double pentagonal pyramid * with this edge as axis. The cross section is a pentagon whose sides and vertices are edges and vertices
of the tetrahedrons.

the ikosahedron.

Let a denote any one of the component tetrahedrons of We have

20 tetrahedrons a,
i

vertex

A and
"

12 vertices

J5,

12 edges

30 faces
*

AB ABB"
is

30 edges BB, 20 faces BBS.

The term double pyramid

not used here in the technical sense defined in

However, this may be regarded as a limiting case, obtained by rotating the end-pyramids around the plane of their common base until they come into a hyperplane with their vertices on opposite sides of this plane.
Art. 32.

i8o]

600-HEDROID
face

319
/3

To each
a and
/3

BBB attach a
We
new

tetrahedron

(Fig. 2).

The

which have

this face in

common form

a double

triangular pyramid. edges BC, and three

have a new vertex C, three new faces BBC. We have added

j8,

20 tetrahedrons 20 vertices C, 60 edges J3C,

and

60 faces

BBC.

Along an edge BB we have now two tetrahedrons a and two tetrahedrons 0. Along such an edge, therefore, we can put one more tetrahedron 7 (fig. 3). This gives us a new edge CC and two new faces BCC. The vertices C and the edges CC are the vertices and sides of a set of pentagons like those on the original ikosahedron. The
pentagons of the ikosahedron, however, overlap, while these form a regular dodekahedron. We have added

and

30 tetrahedrons 7, 30 edges CC, 60 faces BCC.

Along an edge BC we have now one tetrahedron /3 and two tetrahedrons 7. At a vertex B we have five tetrahedrons a, five tetrahedrons /?, and five tetrahedrons 7.

We

can

fill

the space about

B by inserting a double pentag-

We

onal pyramid made up of five new tetrahedrons 5 (fig. 4). have a new vertex D. The axis forms an edge BD, a

and there are five new edges CD. BCD between the tetrahedrons of the double pyramid, and five faces CCD coming to the point D. We have added
continuation of
five
y

AB

There are

faces

60 tetrahedrons
12 vertices D,

5,

12 edges

60 faces

BD and 60 edges CD, BCD " 60 faces CCD.

320

THE 600-HEDROID AND I2O-HEDROID

[vra. IV.

have one tetrahedron 7 and two tetrahedrons 5. Along this edge there is room, therefore, for two more tetrahedrons * forming a double triangular pyramid This gives us a new vertex E, two new edges (fig. 5). There is one new face CCE CJE, and two new edges DE. between the two tetrahedrons, and there are four new faces CDE coming to the point E. We have added now

Along

CC we

60 tetrahedrons

e,

30 vertices E, 60 edges CE and 60 edges DE,

30 faces

CCE

"

120 faces

CDE.

Along
hedrons

CD
.

we have two
Along
,

tetrahedrons 8 and two tetrais

CD

there

room, therefore, for one

with one new edge EE and two new more tetrahedron The vertices E and the faces CEE and DEE (Fig. 6). edges EE are the vertices and sides of a third set of pentagons like the two sets already mentioned. These pentagons, however, touch only at their vertices, and are separated by triangles. We have added

60 tetrahedrons f 60 edges EE, 60 faces

CEE and

60 faces

DEE.
and two we have one

Along each edge

tetrahedrons
?,

CE we

have two tetrahedrons

and along each edge DE and two tetrahedrons f Now the pentagons and triangles just mentioned lie all in one plane.* If at C we insert a tetrahedron rj having the interior of the triangle EEE for face, and at D the half of a double pentagonal pyramid formed by taking a half
tetrahedron
e
.

of each of five tetrahedrons

*

0,

with the interior of the

We

say plane instead of great sphere, as explained at the beginning of this

article.

i8o, i8ij

600-HEDROID

321

pentagon for base, we shall have along the edge EE one tetrahedron ?, one tetrahedron 17, and a half of a tetrahedron 0, forming dihedral angles whose sum is just equal to two Therefore it is not necessary to right dihedral angles. continue our process. We have a plane completely filled with triangles and pentagons and their interiors, and the half of space on one side of this plane completely filled with
the tetrahedrons which
If

we continued our
on the other
fill

we have taken, and their interiors. process we should have the same
and the two together

figure

side of the plane,

would completely There are

the elliptic space.

20 tetrahedrons

rj

and 60 tetrahedrons

0.

The

total

number

of tetrahedrons in the entire figure will

/?,
.

be the number of each of the kinds a,

twice, and the

77

counted

number
be
30

of the tetrahedrons

counted once.

That

is, it

will

2(20

20

60

60

60

20) 4-

60

600.

This construction in the hypersphere of 600 equal regular tetrahedrons, which, with their interiors, fill the hypersphere, determines in space of four dimensions a regular
polyhedroid containing 600 equal regular tetrahedrons. Its name is hexakosioihedroid, or 6oo-hedroid.

We can count the number of vertices, edges, and faces, but from any one of these numbers the other three can be " computed directly. In fact, the N" of Art. 177 is 120,
and the polyhedroid formula
i

of this polyhedroid

is

120

720

1200

600

1=0.

Completion of the 600-hedroid. Although it is not necessary to do so, we shall complete the figure by the
181.

process employed for the

first half.

The

different parts,

322
as

THE 600-HEDROID AND I2O-HEDROID

[vili. iv.

we come

to them, will correspond to the parts already

formed

for the first half,

and we

shall

denote them by the

same letters with

accents.

we had a figure (Fig. 6) on which were triangular cavities at the points C and pentagonal cavities at the points D. In the former we insert
After adding the tetrahedrons f

double triangular pyramids each made up of two tetrahedrons 17 and 77', and at each point D we put a double pentagonal pyramid formed by taking five tetrahedrons
(Plate II,
fig.

7).

The double

triangular

pyramid gives

us a new vertex C", three new edges EC', a new face

separating the two tetrahedrons, and three faces

EEE
EEC'

coming

to C'.

The double pentagonal pyramid

line

gives us a

new vertex ZX, an axis edge DD' in a and five new edges ED' coming to
faces

with

AB

and

BD

f
.

It gives us five

DED', common
axis,

faces

of

successive
1

tetrahedrons

around the

and

five faces

RED

coming to D'.

Corre-

sponding to themselves in the two halves of the figure are the tetrahedrons 0, the vertices E, the edges EE and Z)ZX,

and the

faces

EEE and

DED'.

Along EE there is room for one tetrahedron ', with a new edge D'C and two new faces ED'C (fig. 8). At the point E we can still insert a double triangular

pyramid (fig. 9), corresponding to the one which first produced the vertex E (fig. 5). This is formed of tetrahedrons e', giving us the new edge C'C', a dividing face f The points C" are the EC'C, and two new faces D C C
f f
.

vertices of twelve pentagons forming a regular dodekahe-

dron, outside of which our figure lies. At the point f we insert a double pentagonal pyramid (fig. 10), corresponding to that which first produced the

This is composed of five tetrahedrons 8'. We have a new vertex B', a new edge, the axis D'B', and five new edges C'B'. We have five new faces D'C'B', common
vertex D.

i8i, 182]

THE SEVENTY-TWO DEKAGONS

and

323
five

faces of successive tetrahedrons around the axis,

C'C'B' coming to the point B ' and Along the edge C'C" we have two tetrahedrons two tetrahedrons 5'. We have room, here, then, for one more tetrahedron 7', giving us one new edge B'B' and two new faces C'B'B' (fig. n). At the point C' we can still insert one tetrahedron /3',

new

faces

with the face B'B'B'

(fig.

12).

now encloses a regular ikosahedron like that with which we started. It contains twenty tetrahedrons twelve new edges B'A', and a', with one new vertex A thirty new faces B B' A' (fig. 13).
Our
figure
1
,

the following page regular 6oo-hedroid.

On

is

a table

of all the parts of the

182. The seventy-two dekagons in a 600-hedroid. The angle subtended at the centre by an edge. We have noticed that the edges AB, BD, and DD' lie along the same line. From A to A one of these lines contains five the ten edges, and the entire line must contain ten edges edges and their extremities make up the entire line. In the
'
;

complete figure there are seventy-two such

lines,

each line

running along ten of the 720 edges. In the hypersphere there are seventy- two great circles; and in hyperspace there are seventy-two planes through
the centre of the 6oo-hedroid, each intersecting the polyhedroid in a regular dekagon whose sides are all edges of

the polyhedroid. An edge of the 6oo-hedroid, therefore, subtends an angle of 36 at the centre and a set of regular pentahedroids, 600 at a point, having as they do face
;

angles of

60, with

their interiors

would more than

fill

the

part of Euclidean hyperspace about the point, so that such a set must belong to Hyperbolic Geometry (see
Art. 179).

3 24

THE 600-HEDROID AND I2O-HEDROID

[vin. iv.

183. in

Construction of the 120-hedroid.

In the manner
6oo-hedroid

which we have constructed

the regular

we can

construct a regular polyhedroid with four dodekahedrons at a point. Starting with a dodekahedron a, we attach a dodeka-

hedron
vertex

/8

to each face (Plate III,

of

fig.

14).

About any

a we have
face
-4's,

three

new dodekahedrons, with a

outwards from

common
vertex B.

edge which

extends

to

The
two

common
two
5's,

its vertices

to two adjacent /3's has for and a vertex C, the highest

PLATE

II.

D'

c'

FIG. 7.

FIG.

8.

FIG.

9.

FIG.

10.

B'
FIG.
12.

FIG.

ii.

FIG.

13.

182, 183]

I20-HEDROID

325
/3

point of this face. a face DD.

.

Each dodekahedron

has at the top

have now a triangular cavity in which we can dodekahedron 7 (Fig. 15). Each 7 is attached to The face common to two adjacent 7*5 has three others. The for vertices one C, two Z?'s, and two new vertices E. two 7*s form a figure with a neck across which this face Each y has three new vertices F, and at the very cuts. top one new vertex G. The outer face of the dodekahedron j8 now becomes the base of a pentagonal cavity in which we can insert a dodekahedron 8 (Fig. 1 6). Each 8 has at the top a face with five new vertices H. Two S's resting above two adjacent faces of the original dodekahedron a are separated by the neck The upper edge of this neck, joining two adjacent 7*5. of the base a at is now EE, cavity in which we can insert a dodekahedron e edgewise. If instead of a dodekahedron c we take a half, cut off by the plane of two opposite edges, we shall have along the one dodekahedron 5 and a half of a dodekahedron edge e, forming dihedral angles whose sum is just equal to two

At

B we

insert a

HH

The section of the e will right Dihedral angles (Fig. 17). lie in a plane with the upper face of the 5, and the half of
space on one side of this plane will be filled.* The plane contains a set of regular pentagons, and a set of hexagons with two sides equal to the sides of a pentagon and four
sides equal to the altitudes of a pentagon, the

pentagons

and hexagons with

The
in

their interiors filling the plane. highest point of a dodekahedron 7 does not

appear

being at a distance equal to one-half of the edge beneath the point common to three adjacent
this plane,

hexagons.

We

have one dodekahedron

a, twelve

j8's,

twenty 7

s,

* See foot-note, p. 320.

326

THE 600-HEDROID AND I2O-HEDROID

6's,

[vm. w.

and twelve
hedrons
e,

making

forty-five besides the half-dodekathirty.

of

which there are

Thus

the

total

number

in the entire figure will be

2

45

I2

The name of the figure is

and
its

hekatonikosahedroid or i2o-hedroid
is

polyhedroid formula
i

600

1200

720

120

o.

Below

is

a table for the i2o-hedroid, corresponding to

the table on p. 324.

PLATE

III.

FIG.

14.

FIG. 15.

FIG. 17.

FIG. 16.

TECHNICAL TERMS
In this
text
list

are

some

of the terms of four-dimensional

geometry not used in the

and

of it-dimensional geometry, also terms equivalent to some that are used, In most cases a reference is added. the principal abbreviations. For terms

and

used or explained in the text, see Index. The following are the authors most frequently mentioned, many of the references Clifford, being given in full in the preceding pages: Cayley, Math. Papers (p. 5) Dehn (p. 288) Enriques, Encyclopedic, vol. Math. Papers (p. 5) ; Cole (p. 142) Loria (p. 9) Pascal, Repertorium der hbheren MatheIIIi (p- IS) Jouffret (p. 9) matik, Ger. trans, by Schepp, vol. II, Leipzig, 1902 Poincare", Proc. London Math. So., vol. 32 (p. 12); Riemann (p. 6); Schlafli (p. 22); Schoute (p. 9); Stringham Veronese, Grundzuge, etc. (p. 9); Wil(p. 289); Sylvester, 1851 and 1863 (p. 5); son and Lewis (p. 12).
;
;

The numbers
Achtzell,

refer

always to pages.
tions are placed together, Schoute,
I,

Z8

regulare,
II,

hypercube;
202,
242.

netz;

Schoute,

88, 124.
(Ital.),

See

Zell.

Bipiano
577-

#,

in

Rn

Pascal,

Allomorph, allomorphic, two polyhedrons having the same number

of vertices and edges, and the same number of faces of each

C5 C8 C16 C24 C120 C600

, ,

the six

kind that
of

they have;

regular polyhedroids, Jouffret, 103. similarly Case, cell; hypercase, corresponding

Schoute, II, polyhedroids; 22-23. See Isomorph. Ankugel, Ankugelraum, Anradius, see
Kugel.

term
103Cell,

in space of five

Poincare",

278.

dimensions; See Jouffret, 96,

Apothema

a hypercone), height, Schoute, II, 302.

(of (of

slant

Zell;

Grenzraum, Seitenraum, a hyperplane angle) Schenkelraum.

case,

(of

Arele, edge, Jouffret, 96.

-cell,

-hedroid,

Maschke, Am. Jour.

1 8,

Axe

a piano-polyhedral vertex-edge, see Kant.

angle),

Math., vol.

181.

Configuration, Cay ley- Veronese (p. 5), Carver, Trans. Am. Math. So., vol.
6, 534-

Basis, base, of a pyramid, hyperpyra-

Confine, polyhedroid of n dimensions; 35 face of a (n i)-boundary; -raum, Schoute, II, 242 ; used also of the base of a linear system of rectangular simplex; prime with edges at one vertex prime spaces (e.g., the line common to the linear system formed from the equal and perpendicular to one another; Clifford, 603. equations of three hyperplanes in Cylinderraum, (Cy), of k dimenRJ, Schoute, I, 141. sions, spharisch, Bildraum, the space of the figure of hypercylinder ; zweiter with spherical bases; descriptive geometry, the space in which all the different projecStufe, with cylinders for bases (Art,

mid,

etc.,

Schoute,

II,

327

3*8

TECHNICAL TERMS

Stufe 5-ter $)-ter (k infinity (the "vanishing point" of 147) ; Dimension, Cy[(Fo)i> RV(s)], what perspective), Schoute, I, 2 ; Fiuchtraum (of Rn ), Schoute, I, 124. a prism of this kind becomes when the bases, (P0)fc are no longer Fold: two-fold, w-fold, applied to
entirely
linear
II,

(see

Prisma)

figures, angles, boundaries, spheres,

etc.,

Kreiscylinder piano-cylindrical hypersurface with (Art. 148) right directing-circle

Schoute,

293.

to

indicate

the

number

of

their dimensions,

Stringham;

see

Veronese, 557.

Manifold. A k-iold relation in space of m dimensions gives an (m k) -dimensional locus, Cayley,

Decke, Demi-,

vertex-face,

analogous to verespace, half-hyper-

VI, 458; seeOwo/.

tex-edge, Schoute, II, 4.

half-;

Ftinfzell,

Z6

pentahedroid, Schoute,

11,4; seeZell.

plane; Jouffret, 60. Difcdre d'espaces, hyperplane angle, g^, line at infinity, Schoute, I, 21. Gegen-Punkte or entgegengesetzte Jouffret, 60.

Ditheme, surface, see Theme.

Differentielle, g&>m6trie m6trique, restricted geometry, see Restreint.

Punkte, opposite points (as in the

Double

Elliptic

Geometry,

see

Dreikant, trihedral angle, see Kant.

footnote, p. 215), Veronese, 237. Gegenuber, opposite (as in a tri-

Droite-sommet, vertex-edge, Jouffret,

92.

hedral angle each edge is opposite the face which contains the other
two), Veronese, Schoute, I, 268.

449;

see

also

Eben,

an (n i) -dimensional Gemischt, having both proper and E*-i, the same as Rn-i, 577; improper points; gemischtes SimDreieben, vierdimensionale Ebenplex, S g (d) Schoute, I, 29. tripel, piano-trihedral angle, Schoute, Gerade, right, Schoute, II, 108, 293; schief when not gerade. II, 8, 4. Eigentlich, proper, not at infinity, see Gleichwinklige Ebene, isocline planes,
flat,
;

flat; Pascal for

Ebene, plane;

used by

UneigenUich, Schoute,

I,

20.

Veronese, 539.
itatsgrad, see Parallel, Orthogonal.

Entendue, hyperspace
sions), Jouffret, i.

(of four

dimen- Grad von Parallelismus, Orthogonal-

Entgegengesetzte Punkte, points (of the Double

opposite
Elliptic

Grenzraum,

cell, Schoute, I, 10; Grenztetraeder, Schoute, II, 218.

Geometry), see Gegen-Punkte. Espace, hyperplane, Jouffret, 2.

Half-, semi-, demi-, halb-.

He'catonicosae'drolde,

C120

i2O-he-

Face:

a deux dimensions, face

(of

droid, Jouffret,

105, 169.

angle (of a tetrahedroidal angle), -hedroid, Stringham.

half-plane

a angle); dihedral angle;

sions,

trois

a plano-trihedr.al HexacosildroTde, C600 , dimensions, Jouffret, 105, 169. a quatre dimen- Hexad6cae*drolde, C16 ,
Jouffret,
Jouffret, 105, 128.

6oo-hedroid,

i6-hedroidt

hyperplane angle;

Homaloid, flat, represented by an equation or by equations of the Flat noun and adjective), linear, first degree, Sylvester, 1851 writhomaloid (or omaloid), eben, flach. ten also omaloid; see Theme. Fluchtpunkt (of a line), the point at uf, polyhedron in which two faces
62-63.
First, vertex-edge, Schoute, II, 4.
;

TECHNICAL TERMS
are polygons of the same number of sides with one side in common (Ferse) and the remaining
sides of

329
Educational Times, vol.

from
10
:

the

100.

one connected with the Icosat6tra6droide, C24 Jouffret, 105, 137. remaining sides of the other by

24-hedroid,

quadri- Ideal, improper, uneigentlich. polyhedroid Inkugel, see Kugel. formed in a similar way. Partic- Ineunt points, the points of a locus, ular cases are the Prismenkeil Cayley, VI, 469. In the same way he uses the expression, "tangent and the Pyramidenkeil. Schoute,
laterals;

two

triangles in

and

by

Rn

II, 26, 41, 43.

omals of an envelope."
i2O-he- Inhalt,

Zmo, Hundertzwanzigzell, droid, Schoute, II, 213.

8,

volume, hypervolume, Schoute, II, 94.

etc.,

Hyper-: hyperlocus, Sylvester, 1851, Iper12;

plane,

planar,

(Ital.), hyper-, Loria, 302. Isocline planes, plans d'angles igaux,

pyramid,
try,

pyramidal,
theory,
1863,

geomeItal.

plans

une

infinite

d' angles,

ontological,

gleichwinklige Ebene.

Isomorph, isomorphic, allomorphic polyhedrons or polyhedroids are isomorphic when faces or cells Hypercone, hypercone de premiere which come together in one always espece, Kegelranm. correspond to faces or cells which Hypercone de premiere espfcce, hyin come the other. together percone; de seconde espece, double cone; Jouffret, 92. 22-23. See AlloSchoute, II,
Sylvester,
iper-.

172-177;

Hypercube,
Achtzell,

tessaract,

octaedroide,

Masspolytop,

morph. Oktaschem Isomorphic geometries are different

interpretations of the same stract geometry (see p. 15).

(see -schem).

ab-

Hypercylinder, Cylinderraum. _!, Pascal, Hyperebene, -Rw _i (or

577).

KI, Kugelraum, see this word.

Hyperparallelopiped, paralUlepipede Kant, edge, Schoute, I, 9; Dreikant, d quatre dimensions, Parallelotop, trihedral angle, Schoute, I, 271 ; tetrahedroidal Paralleloschem (see -schem). Vierkant, angle, Schoute, I, 267 ; w-Kant, Vielkant, Hyperplane, lineoid, quasi-plane, esSchoute, I, 279, 286; Scheitelkant, pace, plan, Hyperebene, Raum. vertex-edge, Schoute, I, 268; DreiHyperplane angle, diedre d'espaces, Raumwinkel. kant zweiter Art, piano-trihedral

Hyperprism, Prisma. Hyperpyramid, Pyramide.

Hypersolid, Confine, Polytop.

angle, Axe, its vertex-edge, Veronese,

540, 544; regulare w-Kant ^-ter Art, Schoute, II, 140.

Hyperspace, 4-space, I'entendue, Hy- Kantenwinkel (of a Vierkant), face perraum (Pascal, 577). angle, Schoute, I, 268. Hypersphere, quasi-sphere, Kugel- Kegel : Kreiskegel erster Art, pianoconical hypersurface of revolution; raum, n-Sphdre, Polysphare. 1*"" 8 i dimenzweiter Art, conical hypersurface of surface, of p (Hyper) sions in space of p dimensions; double revolution (Art. 112), Veronese, 557e.g., in space of five dimensions, hyper-hyper-surface. H. R. Greer, Kegelraum, (Ke)i, of k dimensions, zweiter Stufe, double hypercone; "Question 2503," Math. Questions

330
cone, j-ter Stufe (k
sion,

TECHNICAL TERMS

+ $)-ter Dimen-

what a Ke[(Po)i, S(s)], pyramid of this kind becomes when the base, (Po)t, is no longer enthe tirely linear (see Pyramide)
;

Kegelraum
isch,
if
if

(of

any kind)
is

is

sphar-

the base

also

spherical, gerade the centre of the base is

Kugelraum is the hypersurface, not the interior. In space of n dimensions it is of n i dimensions, and he writes n ,\, but later he writes n the subscript denoting the number of dimensions of the space. See foot-note, II, 263.

the projection of the entire vertex- Leit-, directing-; strahl, kurve, raum; Schoute, I, 98, 117, 199, simplex upon the space of the base, etc. regulare if, further, the vertexsimplex is regular and its centre Linear, flat, spaces as denned in Art. 2, represented by equations of is the projection upon its space of
the centre of the base, schief
if

the

first

degree.

Schoute, II, 292-293.! Lineoid, hyperplane, Cole, 192; colKeil, dihedral angle, Veronese, 444;! lineoidal, Keyser, Bull. Am. Math. 86. So., vol. 9 hyperplane angle (Keil von vierj Veronese, 544 ;! Losung, point, used by Schlafli to Dimensionen), denote a set of values of the varipiano-polyhedral angle, Dehn, 571.}
:

not gerade;

see Pyramidenkeil, Prismenkeil, these words. Kiste, (Ki)ij of k dimensions, rectangular par allelo piped or hyperparallelopiped,

ables
like

satisfying
for

given

equations,

and then

any

set of values,

Cauchy's "analytical point"

(p. 6), Schlafli, 6.

Kontinuum,

Schoute, II, 94. Lot, perpendicular line, Schoute, I, aggregate of all 44; Letvielcck, simplex with all solutions of an equation or of a angles right angles (in Elliptic Geometry), Schoute, I, 47. system of equations, locus or
the

spread, Schlafli, 6. Kreiscylinder, Kreiskegel, see Cylin- Manifold,

der, Kegel.

space of any number of dimensions, variety, Mannigfaltig-

Kreuzen, used of lines not parallel keit, Grassmann, Riemann, and others. and not intersecting, also of other spaces often used with senkrecht, Mantel, lateral boundary, of a hyperSchoute, I, 43-44pyramid, etc., Schoute, II, 35. Kugel, hyperspherc (of any number of Masspolytop, hypercube, Schoute,
;

With dimensions), Veronese, 592. Veronese Kugel denotes the interior, Kugeloberflache the hypersurface; see Kugelraum; net of regular polygons
-netz,

n, 93Monotheme,
Theme.

line,

Sylvester;

sec

a Netz, Achtzellnetz, etc., Schoute, II, 242. sphere, Schoute, II, 154; Ankugel, sphere tangent to the edges of a Normal, perpendicular; (of planes) stereoregular polyhedron, Inkugel, inabsolutely perpendicular;

on

scribed sphere, scribed sphere;

245-

Umkugel, circumAnkugelraum, An-

metrisch

normale
a
;

Ebenen,

per-

radius, etc.; Schoute, II, 151, 199,

in pendicular Schoute, I, 70-72

hyperplane;

see Senkrecht.

Kugelraum,
Schoute,

Kn
I,

-.\.

or

Kn

127.

With Schoute

hypersphere, Oberflache, Inhalt of the boundary, Schoute, II, 95, 263.

TECHNICAL TERMS
Octa6drolde,
83, 118.

331
hyperparallelopiped,
of k dimensions,

C8

hypercube, Jouffrct, Paralleloschem, see -schem.

Omal

or

omaloid (both noun

and Parallelotop, (Pa)*,

396

adjective), line, plane, etc., linear, the same as homaloid. If the

hyperparallelopiped,

Schoute,

II,

C , penlahedroid, Joufor (m k) -dimensional fret, 105, 132. dimensions) omaloid. Perpendicular, absolutely, simply, (in space of Omaloid is used absolutely to deCole, 195, 198; simplement ou innote the onefold or (m i)-dimenperpendiculaires, complStement
relation
is is

linear or omal, the locus Pentae"drolde,

fc-fold

Opera tionsraum, which all the

are supposed

sional omaloid, Cayley, VI, 463. n , the space in

figures considered to lie, as if there Plagioschem, spherical simplex, see were no space of higher dimensions -schem. Plan (Fr.), plane; used by Cayley (see Art. 26), Schoute, I, 4.

Jouffret, 34; see onal, senkrecht.

absolument ou completement Normal, orthog-

Opposite, gegenuber, entgegengesetzte. Order, Richtung, Sinne. Orthogonal, teilweise, Orthogonalisee tatsgrad, Schoute, I, 49;
Senkrecht.

in 1846 for hyperplane, for ordinary plane (I,

demi-plan
321),
see

also plane;

plans d'angles 6gaux

infinite"

ou plans a une

d'angles,

Orthogonal figure, cube, etc., Stringham, 5.

I?ao>

isocline planes, Jouffret, 77. hypercube, Plane, used by Cayley in five-dimensional geometry for a space

point at infinity, Schoute, I, 21. Parallel: planes are parallel "von der ersten Art" if they have the same line at infinity, "von der

four dimensions: dimensional geometry

of

"In

five-

we have:

space, surface, subsurface, supercurve, curve, and point-system according as we have between the six

zweiten Art"
point in

and line, subplane, superline, point," IX, 79. paralleles" if their lines at infinity Piano-trihedral angle, triedre de seconde espece, Dreieben (see Eben). have a single point in common, Wilson and "paralleles suivant le deuxieme Planoid, hyperplane, mode ou completement paralleles " Lewis, 446. r point (Fr.), the space of if their lines at infinity coincide, Point: i dimensions determined by r r (we should notice Jouffret, 31 that these two writers use "first" independent points (Art. 2); biand "second" in opposite senses); point, line, d'Ovidio, Math. Anteilweise parallel or halb parallel nalen, vol. 12, 403-404. are terms used by Schoute; in a Polvierkant, the edges perpendicular to the cells of a given Vierkant space of more than four dimensions lower spaces may be "ein Viertel, (analogous to supplementary trizwei Viertel or drei Viertel parallel," hedral angle), Schoute, I, 268. etc., Schoute, I, 34. (Po)*, of k dimensions, Polytop; when the boundaries are not all linear Parallel6pipfcde a quatre dimensions,
Jouffret, 82.

they have only a this point is at infinity, Veronese, 516; planes are "paralleles suivant le premier mode ou incompletement
if

common and

coordinates, o, i, 2, 3, 4, or 5 equaand so when the equations tions are linear, we have: space, plane,
:

CP0)*

is

used, Schoute, II, 28, 292.

332

TECHNICAL TERMS

Polyschem, polyhedroid, see -schem Quadri&dre a quatre dimensions, tetra&iroide, tetrahedroidal angle; Polysphare, see Sphere of n dimenit has six "faces a deux dimensions. sions," four "faces a trois dimenPolytop, (P0)t, a limited portion of and four "tri&ires de sions," any space, the boundaries (Po)t seconde espce," Jouffret, 63. it is generally understood that the boundaries are linear, then it Quasi-plane, /?n _i in /?n, Sylvester,
is a polyhedroid; Simplexpolytop, one bounded by simplexes, such in R* is the Tetraederpolytop

1863, 173;

quasisphere, sphere of
Clif-

any number of dimensions,

ford, 604, 605.

or
28.

Vierflachzell

Schoute, II,

i,

Rn
dimensions,
etc.;

linear space of

n dimensions

in

Prisma, (Pr)i, of k hyperprism, prism, zweiter Stufe, with bases (Art. 136);
(k

Schoute

prisms for s-ter Stufe

usually the Operationsraum, R& any space in Rn Rt+i * R<t, but Rw is used for w "w" referring to the Punktis
, ,

Rn

Schoute, 1, 4, 13. ; RV(s)], the bases parallel (Po)*, RV(s), the lateral elements, R9 of a the lateral elements parallel Rs, prism or cylinder "j-ter Stufe," Schoute, II, 37-39; pyramidales Schoute, II, 39. Prisma, hyperprism with pyramids Raum, space; used alone for #3,
,

+ *)-ter

Dimension, Pr[(Po)i,

wert

for bases, Schoute, II, 41. Prismenkeil, the Huf formed when a hyperprism is cut into two parts

Schoute,

I,

Basis-,

Bild-,

Cylinder-, Kegel-, Kugel-, Seiten-, etc., see these words.

by a hyperplane (or by an Rn Raumwinkel, angle, hyperplane which intersects a base, Schoute, Schenkelraume, its cells, Schoute,
II, 43I,

268.

Prismoid, w-dimensional polyhedroid Region limite'e, restricted region. bounded by two (Po) n -\ in parallel Restreint (Fr.), restricted, region /Zn-i, an<* a Mantel of simplexes, mitrique diflimite'e, gtomitrie
S(n), Schoute, II, 44. Proper, see EigenUich.
ferentielle, Enrique, 48, 112.

Richtung,
it

Sinne,

order,

Veronese,

Punktwert of a space, the number

of independent points in
2)
;

381, 456, 498, 550.

i)-ten 209 Rotation von den (p Potenz, when each point describes
;

(Art. Rotation, simple, double, Cole, 201,

for

R#

it is

-f i

Schoute,

1,12.

a 9 (see Art. 116, Th. i, and Art. Pyramide, (Py)i, of k dimensions, pyramid, hyperpyramid, etc.; 147, Th. 2), Schoute, II, 297. zweiter Stufe, double pyramid; 5-ter Stufe (k + s)-ter Dimension, SQ, Si, ^2, etc., point, line, plane, etc. Py((Po) t S(s)], with a polyhe(Stern), Veronese, 509. droid, (Po)t, for base and a vertex- S(d), simplex with d vertices, Schoute, Su (d), improper simplex, 10 simplex, S(s); Schoute, II, 35I,
,

36.

S0(d), gemischtes Simplex, Schoute,

Pyramidenkeil, the Huf formed when I, 28, 29. a hyperpyramid is cut into two St(P, q)> Simplotop. parts by a hyperplane (or by an Scheitel, vertex, Scheitelkant, vertexRn -\) which intersects the base, edge, Schoute, I, 268. Schoute, II, 41. Schein, contour of a figure as seen

TECHNICAL TERMS
from an outside point, Schoute, Sense,
II, 17-

333

order, Sime.

Simplex, triangle, tetrahedron, and in -schem, -hedroid, used by Schlafli general a polyhedroid whose verwith qualifying prefixes to denote tices are all independent (Art. 2). various polyhedroids : Polyschem, S(d -f i) the simplex in Ra with
i vertices, Schoute, I, 9-10, d 19 ; paralleloschem, hyperparallelepiped, 12; in space of four dimenSimplicissimum, prime confine. sions the six regular polyhedroids Simplex, spharisch, spherical tetrahedron, etc. are, Pentaschem, Oktaschem, HekSchoute, II, 291. Eikositetraschem, Simplexpolytop, bounded by simkaidekaschem,

Hekatonkaieikosaschem, and Hexaplexes, Schoute, II, 28. kosioschem, 46-52; a spharisches Simplicissimum, simplex, Sylvester, "Question 8242,*' Math. Questions Polyschem is a polyhedroid belonging to the geometry of the nfrom Educational Times, 47 : 53. sphere; it is a Plagioschem when Simplotop, St(p, q), obtained by it has n boundaries, the same as i) forming two simplexes S(p and S(q -f- i) with one vertex simplex; the single parts of the in common from a set of p -f q i boundary of a spharisches Polyschem are Perrischeme, 58. independent points (and so lying in a space Rp+q), and then letting Schenkel (of an angle) side, Scheneach move parallel to itself over kelraum (of a hyperplane angle) the other. A particular case is cell, Schoute, I, 268. the doubly triangular prism (see Schief, oblique, Schoute, II, 293.

Art. 143), Schoute, II, 44. Sechshundertzell, Zm> 6oo-hedroid; Schoute, II, 213. Sinne, sense, order, see Richtung. Sechszehnzell, Zi$, t6-hedroid; Sechs- Situs (Lat.), used by Gauss for direction of a plane, like Stellung, zehnzellnetz, Schoute, II, 202, 242. " Seiten (of a poiyhedroidal angle), circa Disquisitioncs generates von drei ebenen superficies curvas," Werke, IV, face angles,
,

Dimensionen, trihedral angles, Ve219. ronese, 544; (of a pentahedroid) Space, 3-space, 4-space,
Veronese, 547. Seitenraum, the # n _i of a simplex S(n -f i), Schoute, I, 142.
faces
cells,

etc., for /?a,

is

and

/?4,

etc.

the

word space

used by

for the highest space conlike Schoute 's Operasidered,

Cayley

tionsraum, see Plane. kreu- Sphere of n dimensions, hypersphere, Senkrecht, perpendicular; schneiden ; zugeordnet, zen, n-sphere; called by Schlafli Polyin R\ polar, used of P^ ; halb sphare or n-sphare, Disphare, cirused of planes with one angle a cle, Trisphare, sphere, 58. in space Spharisch Cylinderraum, Kegelraum, right angle (see Art. 69) of more than four dimensions Simplex, see these terms. lower spaces may be "ein Viertel, Spitze, vertex (point, line, or higher zwei Viertel, or drei Viertel ," space), Veronese, 466, 557, 606. Veronese Spread, surface, hypersurface, repreetc., Schoute, I, 40-49. sented by an equation or by a uses senkrecht or "senkrecht von der ersten Art" for planes absosystem of equations, used with a and number to denote the number 4, lutely perpendicular in " senkrecht von der zweiren Art" of its dimensions as 2-spread,
Semi-parallel, half-parallel*
, ;

for planes half -perpendicular, 521.

3-spread,

etc.;

see

articles

in

334
Am.
31,

TECHNICAL TERMS
Jour. Math,

by Carver, vo
Eisenhart,

"faces a quatre dimensions," Jouffret, 62.

Sisam,

33,

34

Eisland, 35.
Stellung, direction of a t plane, space
etc.,

Stern, consists of all points collinea with the points of a space and a

V* - ^*-i (see Rn ), Schoute, I, 45. Umkugel, see Kugel. point outside of the space, used b> Unabhangig, independent (points),
Veronese to define the differen
spaces.
erster Art.,

Schoute,

I,

26, 75.

Ua-i, the infinitely distant (uneigent_ lich) part of Rd , Schoute, I, 22.

Veronese, 256. improper,

ideal,

#3

>

zweiter Uneigentlich,

at

infinity; Punkt, Fluchtpunkt; Veronese, 424, 507 also unendlich; Schoute, I, 2, 21. Schoute, I, 190-192. Schoute, II Strahl, line, Haupt Variety, variett, Enrique, 66 ; Variet&t, 161 ; Halb , Schoute, I, 85. Schoute, I, 209; manifold. vector, line-vector Straight, flat; Vertex-edge, droite-sommet, First, plane-vector, etc. ; straight 3-space Scheitelkant. "Four DimenEuclidean, Lewis, in /?4 Vierflachzell, polyhedroid sional Vector Analysis," Proc. Am, bounded by tetrahedrons, TetraeAcad. Arts and Sci., v. 46 : 166, 1 73, derpolytop, Schoute, II, 28. Stumpf (of an angle), obtuse; (of a vierdimensionale letraVierkant, prism), truncated, etc., Schoute, II hedroidal angle, Schoute, I, 267. 43, 127Vierundzwanzigzell, Z24 24-hedroid; Subplane, subsurface, see Plane. netz; Schoute, II, 203, 242. see Plane. Superline, supercurve, Volumeinheit, unit of volume, hyperSurface (Cayley), see Plane. volume, etc., Schoute, II, 95. Volum, Inhalt, Schoute, I, 156. Teilweise or parallel, orthogonal senkrecht, see these terms. line or 6.

Art.,

/<U, etc.,

curve, Schlafli,

Tessaract,

figure designates the number of its axes: pentact, a figure with five Z5 , Z8 , Z16 , Z24 , 120, Zeoo, the regular axes, penta-tessaract, a regular polyhedroids, Schoute, II, 203, 213. i6-hedroid, T. Proctor Hall, Am. Zell, cell, polyhedroid, Schoute, II, Jour. Math., vol. 15: 179. Schoute uses the term Zell 196.

hypercube, Hinton, The Winkel, angle, Kan ten-, Flachen-, Fourth Dimension, London, 1904, Raum-, see these words; korTessaract belongs to a ter159. von vier Kan ten, tetraperlich minology in which the name of a hedroidal angle, Veronese, 544.

Tetrahedroidal

angle,

quadriedre

dimensions, quatre Vierkant.

tetraedro'ide,

sions,

alone for polyhedroid of four dimenbut with prefixes (Achtzell,

Fiinfzell, etc.) it
line,

Theme,
points,

spread;

mono-,

or

seems to refer to the three-dimensional boundaries.

curve, di-, surface, keno-, system of

homaloid theme, flat; terms

used by Sylvester in 1851.

Totalitat, hyperspace, Schlafli, 6. Trifcdre de seconde espfcce, pianotrihedral angle, it has three "faces

a deux dimensions, three "faces a trois. dimensions," and three

1 '

See Jouffret, 96. used of a polygon in space regarded as having two faces, Schoute, II, 148, 182. Iweiraum, hyperplane angle, Schoute, II, 8; used also of a polyhedron regarded as having two sides in Rt, Schoute, II, 186.
rweiflach,

INDEX
Numbers
refer to pages.

Absolutely independent points, 24. Absolutely perpendicular planes, 81

see Perpendicularity.

Aristotle
;

on the three dimensions of

magnitudes, i. Axes of a hypercube and a i6-hedroid,

240, 292.

Abstract

geometry
;

and

different

see Geometries, examples of diferent kinds. Alembert, d', on time as the fourth

interpretations, 14

Axiom of Pasch, 30

of parallels, 221

see Parallels, axiom of. Axioms of collinearity, restrictions in

Elliptic

dimension, 4. Analogies, their assistance, 17; hyand dihedral angles perplane

angles, 99; tetrahedroidal angles and trihedral angles, 129; piano-

Geometry,

25.

Axis of a sphere in a hypersphere, 21 1 ; of a regular polyhedroidal

angle, 312.

Axis-element of a double pyramid, 204; of a double cone, 205. polyhedral angles and polyhedral isocline planes and Axis-plane of a rotation, 142, see angles, 133 lines, Rotation; of a circle in a hyper194; piano-prisparallel matic hypersurfaces and polygons, sphere, 211 ; of a piano-cylindrical 242. hypersurface of revolution, 257. Analytic development of the higher Axis-planes of a conical hypersurface of double revolution, 206. geometry, 6. locus, terms used Analytical point,
;

by Cauchy,

6.

Base,

Angle, hyperplane, 95; see Hyperplane angle. Angle of a half-line and a hyperplane, 80; of a half-plane and a hyperplane, 104 ; minimum between two
planes, 116. Angles at infinity, 232-234.

prismatic, of a hyperprism with tetrahedral ends, 274. Beginnings of geometry of more than
three

dimensions,

synthetic,

4;

analytic, 6.

Beltrami, Hyperbolic
certain
13-

Geometry on
kinematics,

surfaces,

7;

Angles, the two between two planes, the associated rectangular 122; system, the associated sense of rotation, 181.

Bibliographies, Loria, Sommerville, 9. B6cher, our use of the term "infinity," 230.

Boundaries of a hypersolid are threedimensional, 64.

Applications of the higher geometries : to a problem in probability (Clifford) 5 geometries with other ele, ;

Boundary-hypersurface of Hyperbolic Geometry, 95, 112.

ments, lines in space, spheres, etc., 10 in connection with complex

;

to mechanics, variables, in proofs of theorems in three-di-

n;

n;

Cauchy, applied the language geometry to analysis, 6.

Cayley, early papers, 5; on Abstract Geometry,"
8.

of

"Memoir

mensional geometry, 13.

335

336
Cell of

INDEX
a half-hyperspace,
62.

Cylinders, the set in a prism cylinder, polyhe258. droid, 63; double pyramid, 67; Cylindrical, sometimes used for hy' angle, 95 ; hyperplane polypercylindrical or for piano-cylinhedroidal angle, 126; piano-polydrical, 258, 266, 284. hedral angle, 133; prismoidal hypersurface, 235 piano-prismatic Darboux, hesitated to use geometry of four dimensions, 9. hypersurface, 242. Centroid, 202 ; see Gravity, centre of. Dekagons, the seventy-two in a 600Classes of points constitute figures, hedroid, 323. 19, 23. Density, of points on a line, 28. Clifford, problem in probability, 5; Descartes, use of "sursolid," 2 ; knew the polyhedron formula, 300. "On the Classification of Loci," 8; kinematics, 13, Descriptive geometry of four dimenClosed sphere, passing out of, 79. sions, 1 8.
Cells of pentahedroid,

57;

Collinear relation, 23; distinguished at first from "on a line," 19, 27; the two axioms, 19, 25. Collinear with a segment, 25; tritetrahedron, 49 angle, 32 pen; ;

"Determine," meaning in geometry,

19.

Diagrams only indicate

Dihedral angle,
its

relations, 18.

tahedroid, 58. Complex variables

represented in space of four dimensions, n, 219. Cone, double, 70 ; see Double cone. Configurations of points (Cayley, Veronese), 5, 8. Conical hypersurface of double revolution inscribed in

plane angle the same at all points, proof independent of the axiom of parallels, 97 ; in a hypersphere, its volume, 209. Dimensions, early references to the number, i differences in spaces of an even number and of an odd
;

number,

14;

of

rectangular

a hypersphere,

239 ; only hyperparallelopiped, three regular figures in a space of

five dimensions, 317.

intersecting it in the same surface as the inscribed cylinder of double

Diophantus, use of "square-square,"

etc., 2.

revolution, 263. Conical, sometimes used for hyperconical, 69. Conjugate series of isocline planes, 183 ; see Isocline planes.

Directing-curve

of a piano-conical hypersurface, 71 ; 'piano-cylindrical hypersurface, 256; similar direct-

ing-curves, 257.

Continuity of points on a line, 28. Coolidge, list of systems of geometry,

IS-

Corresponding

dihedral

angles

of

parallel planes, 223.

Craig, kinematics of four dimensions,

13-

Curvature, Riemann, 7; of the hypersphere, constant, 218. Cyclical order, 28. Cylinder, double, 262; see Double
cylinder.

Directing-polygon of a piano-polyhedral angle, 137 ; piano-prismatic hypersurface, 243; similar directing-polygons, 245. Directing polyhedral angle of a pianopolyhedral angle, 135. Directing-polyhedron of a polyhedroidal angle, 126; prismoidal
hypersurface, 235. Directing-surface of a hyperconical hypersurface, 69; hypercylindrical hypersurface, 253. Direction on a line, opposite directions, 27.

Cylinder,
cylinder.

prism,

259;

see

Prism

INDEX
Distance between a point and a hyperplane,
surface of,

337

78; the minimum between two lines, 105 ; in a hypersphere, 208 ; at infinity, 232.

197; with plane elements, 206; its interior all connected, 207; cylinder of, 263, see Double cylinder; surface of, in a

Distances between two great circles in a hypersphere, 217; two lines

at infinity, 235.

Dodekahedrons, the net, four at a

point, 324.

Double cone, 70; vertex-edge, base, cut from revolution, 263. elements, end-cones, 7 1 a piano-conical hypersurface, 72; Double rotation, 145
;

hypersphere, its importance in the theory of functions, 219; the intersection of the hypersphere with an inscribed cylinder of double revolution and with an inscribed conical hypersurface of double
;

in the hyper-

circular, axis-element,

right,

isos-

celes, generated by the rotation of a tetrahedron, 205.

sphere, 218. Doubly triangular prism, 251.

Duality in the hypersphere, reciprocal figures, 212.

Double

cylinder,
right,

directing-curves,

interior,

generated by the

directing-curves

and

their interiors,

Edge Geometry, the elements

half-

spread out in a hyperplane, 262; of double revolution, cylinder inscribed in a hypersphere, 263;
relation to infinity, 264;

volume, 267; hypervolume, 285; ratio to circumscribed and to inscribed

hypersphere, 287.
Elliptic

planes with a common edge, 138; applied to the theory of motion with two points fixed, 1 73. Edge of a polyhedron, how defined,
21
;

of

a half -plane, 39 ;
;

hemi-

pyramid, polyhe55 pentahedroid, 57; Geometry, 215; see Elliptic Non-Euclidean Geometry. droid, 63. Double prism, the two sets of prisms, Elements, linear, of two planes, 61. 246 right, regular, its cells spread Elements of geometry, points, 19, out in a hyperplane, 248 interior, 23; of Point Geometry, 113. double pyramid, directing-polygons, 249; generated Elements of a

sphere, 209. Edges of a tetrahedron, 45

;

Double

by the

interiors,

directing-polygons and their cut into two double

prisms, 250; doubly triangular, 251 ; hyperprisms with prisms for bases as double prisms, relation to infinity, 252; volume, 265;

doubly triangular double prisms

cut
into six

67 hyperconical hypersurface, hypercone, 69; piano-conical hypersurface, double cone, 71 ; polyhedroidal angle, 126; piano-polyhedral angle, 134 piano-prismatic hypersurface, 242; piano-cylindrical hypersurface, 256.
; ;

equivalent pentahevertex-edge,

Elliptic

droids, 280; hypervolume, 282.

Double pyramid, 66;

base, elements, end-pyramids, lateral faces, lateral cells, intersection

with a plane, 67; with a hyperplane, 68; cut from a piano-polyhedral angle, 138; axis-element, right, isosceles, regular, 204; in a
hyperplane, 318.

Non-Euclidean Geometry due to Riemann, 7 restrictions to the the axioms of collinearity, 25 points of a line are in cyclical order, 29; modification of proof of minimum distance between two lines, 105; lines with more than one
:

common
lines

perpendicular

line,

108;

and planes with more than

Double revolution,

conical

hyper-

one common perpendicular line, planes with more than one common

338
perpendicular
plane,

INDEX
Geometries of
i, 2, 3,

112; Edge 173 the most general motion in hyperspace, 174, 177 ; volume of a tetrahedron, 211; the geometry of the hypersphere is the Double Elliptic, 215, 217; difference between the Single Elliptic and the Double Elliptic, 215

...,...
dif-

Geometry

is elliptic,

dimensions, 24.

Geometries of different kinds,

ferent interpretations of

an abstract

geometry, 14, 15; the geometry of half-hyperplanes with a common Point Geometry, 113; face, 99;

Edge Geometry, 138;

isocline

hyperplane in four dimensions, 217 space of constant

poles
of
;

planes,

197

system of the .hyper-

sphere,

212;

system of parallel
7.

curvature, 218; the geometry at infinity is the Single Elliptic, 233 hypervolume of a pentahedroid, 287; the possible nets of poly;

planes, 244.

Grassmann, Au$dehmmgslehre,
Gravity, centre
vester, 5
;

hedrons in Elliptic Geometry, 306 a regular polyhedroid can be inthe scribed in a hypersphere, 309 possible nets of polyhedroids, 316; see also Non-Euclidean Geometry,
;
;

Sylof a pentahedroid, 201 tetrahedron, triangle, 202.

of,
;

memoir by

Green, problem in attraction,

* '

6.

Half-hyperplane,

Parallel axiom, Restrictions. Enriques, the foundations of geometry, 15; definition of segment, 21. Euler's name usually associated with the polyhedron formula, 300.

half-hyperplane halfopposite hyperplanes, 54; half-hyperplanes with a common face form a onedimensional geometry, 99. " Half -hyper space, half -hyperspace

ABC-D"

face,

Face angles of a polyhedroidal angle,

126.

cell, opposite hyperspaces, 62. Half-hypersphere, 210.

ABCD-E,"

half-

Face of a half-hyperplane, 54

hyper-

Half-line, "half-line AB," produced," opposite half-lines, 28.

"AB

plane angle, 95. Faces of a tetrahedron, 45 ; pyramid, 55; pentahedroid, 57; polyhedroid, 63 piano-polyhedral angle, 133 Figure is regarded as a class of points,
;
' ,

Half -parallel planes, 224; see Parallelism.

Half-perpendicular planes, 85
Perpendicularity.

see

Half-plane, "half-plane

AB-C" edge,

19, 23; belong to, lie in, 23. Five dimensions, only three regular
figures, 317.

opposite half-planes, 39. Halphen, geometry of n dimensions, 8. Halsted, use of the terms "sect" and "straight," 25; proof that a line
divides a plane, 37.

Foundations, different systems, 15;

definitions

Hathaway, application of quaternians to geometry of four dimensions, 13. elementary figures particularly conintersections
of
sidered, 17.

and

Four dimensions, space

restriction to, 16, 59.

of,

24;

our

Fourth dimension as time, 4, Fourth Dimension Simply

plained, 9.

The Ex- Hexadekahedroid,

hedrold.

Hatzidakis, kinematics of four dimensions, 13. Hekatonikosahedroid, 326; see 120hedroU.

291
;

see
see

16-

Hexakosioihedroid,

321;

600-

Gauss considered the higher space a

possibility, 10.

hedroid.

Hilbert, definition of segment, 21.

INDEX
Hyperbolic
etry
:

339
hypersurface,
di-

Non-Euclidean

Geom- Hypercylindrical

planes with parallel elements, their common per112; 95, pendicular plane, 112; boundaryhypersurfaces, 95, 112; translation

recting-surface, elements, 253; inrelation to terior, sections, 254;

infinity, 256.

along a
curves,

line, 146,

along boundary-

rotations in a hyperpiane around parallel axes, 172;. the most general motion in
146,

168;

Hyperparallelopiped, its diagonals all bisect one another; rectangular, its dimensions, the square of the length of its diagonal equals the sum of the squares of its four
as a double dimensions, 239; prism, 252; hypervolume when
rectangular,
272.

hyperspace, 174; rotations around parallel axis-planes, 1 78 pentahedroids which have no point equidistant from the five vertices, 199 the possible nets of polyhedrons, 306 a regular polyhedroid can be
;

271,

when
"

oblique,

Hyperpiane,

24

hyperpiane

A BCD"
mine
it,

inscribed in a hypersphere, 309; the pdssible nets of polyhedroids,

50; figures which deteronly one contains four-

316

see also Non-Euclidean Geindif-

ometry.

Hypercone,
terior,

intersections, 69; with a cone for base,

given non-coplanar points, 51; ordinary space a hyperpiane, 52; divided by a plane, 53; intersection with a plane, 60 ; intersection
of
site

regarding it, 70; axis, generated by the rotation of a half-cone, 204; lateral volume, 266; lateral volume of a frustum, 267 hypervolume, 284 ; hypervolume of a frustum, 285. Hyperconical hypersurface, directing-surface, elements, 69 of double revolution, 197 its plane elements, 206 its interior all connected, 207 intersection with a hypersphere,
;
;

ferent

ways

of

two hyperplanes, 52, 60; opposides of a hyperpiane, 62;

face, cells, interior,

at infinity, 231.

Hyperpiane angle,
divides

spherical directing-surface, 220.

Hypercube, the diagonal twice the edge, its cells spread out in a hyperplane, two forms of projection, as a regular polyhedroid, 240;
reciprocal nets, relation to the

of hyperspace, intersection with a hyper95; plane perpendicular to its face, 96 ; plane angle, 96, 98; two hyperplane angles are congruent when they have two equal plane angles, 96 the plane angle is the same at all points of the face, 97; as a magnitude, supplementary hyperplane angles, right hyperpiane anal- % angles, the sum of two, 98 ogous to a dihedral angle, in the
rest
; ;

the

290;

reciprocal

of geometry half-hyperplanes, measured by the plane angle, 99;

i6-hedroid,

292;
three

the bisecting half-hyperplane, 100.

diagonals

and

axes

form

rectangular systems, 293; associated 24-hedroid, 295. Hypercylinder, 254; lateral hypersurface,
interior,

Hyperprism, lateral cells, etc., 237; its cells spread out in a hyperpiane, the 238; with prisms for bases, dif-

spherical,

axis,

generated by a rectangle, by rotation of a half-cylinder, 255 with

;

cylinders for bases, different ways of regarding it, 255, 261 lateral
;

ways of regarding it, 239, 252; lateral volume, 265; congruent and equivalent hyperprisms, 271 ; hypervolume when the bases are prisms, 273 ; when the bases
ferent

volume, 266; hypervolume, 284.

are tetrahedrons, 274; hyperprism with tetrahedral ends, prismatic

340

INDEX

double revolution, 263; volume, base and vertex edge, 274; hypervolume of any hyperprism, 275. 267; hypervolume, 285. Hyperpyramid, 63; base, interior, Hypersurface, 69 of a pentahedroid, 62. sections, 64; with a pyramid for base, different ways of regarding it, Hypervolume, 270; of a rectangular 66 ; cut from a polyhedroidal angle, 271 ; hyperparallelopiped, any hyperparallelopiped, 272; hyper127; axis, regular, 203; lateral prism with prisms for bases, 273 ; volume, of a frustum, 265; hywith tetrahedral bases, 274; any cut frustum 276; pervolume, into pentahedroids, hypervolume, hyperprism, 275 ; hyperpyramid, double prism, 276 frustum, 278 278. 280; cylindrical and conical hyHypersolid, the interior of a pentahedroid as a hypersolid, 62; boun284 ; persurfaces, hypersphere, daries are three-dimensional, 64; 285; ratio of the hypersphere to inscribed and to circumscribed hypervolume, ratio of two, equivadouble cylinders, 287; pentahelent hypersolids, 270. droid in elliptic hyperspace, 287. Hyperspace, term used to denote the space of four dimensions, 60; divided by a hyperplane, not di- Ideal points, lines, etc., at infinity, vided by a plane, 62. 230. Hypersphere, great spheres and small Ikosatetrahedroid, 296, see 24-he; ; ;

Independent points, 24. hypersphere, tangent hyperplanes, Infinity, sense in which the term
spherical dihedral angle, its tetrahedron, the volume, 209; sixteen associated tetrahedrons, their volumes, 210; axis and poles

207; great circles and small circles, 208; distance in a

spheres,

droid.

is

208;

used, 230;
infinity,

points, lines, etc., at all at points infinity

;

in

a single hyperplane, 231

dis-

tance, angle, 232;

of

cle of

a sphere, axis-plane and polar cira circle, 211 their motion in

;
;

a rotation of the hypersphere, 212

duality,

reciprocal figures, 212; the geometry of the hypersphere as

233; the Single Elliptic, generalizations made possible by the use of these forms of expression, 233; distances between two lines, 235;
;

dihedral angle, the geometry at infinity is

an independent three-dimensional relation to infinity of the prismoidal geometry, 212; it is the Double hypersurface, 241 hyperparalleloElliptic Non-Euclidean Geometry, piped, 241, 253; piano-prismatic hypersurface, double prism, 252; 215; the Point Geometry at the the distances between centre, 216 hypercylindrical hypersurface, 256 ; two great circles, parallel great piano-cylindrical hypersurface, circles, 217; proof from Point prism cylinder, double cylinder, 264. Geometry that the geometry of the hypersphere is the Double Elliptic, Interior of a figure as distinguished from the figure itself, 20 see Seg217; rotation, double rotation,
;

screw motion, parallel motion, 218 ment, Triangle, Polygon, etc. curvature constant, 218; inter- Intersect, intersection, 23. section with a conical hypersurface Isoclinal angle (Stringham), 125. of double revolution, 220; inter- Isocline planes, 123, 180; have an section with an inscribed cylinder infinite number of common perof double revolution a surface of pendicular planes, 123, 182; series
;

INDEX
of isocline planes, 182; conjugate series, 183 ; the two senses in which

341

Jouffret,

Jordan, geometry of n dimensions, 8. Gtomitrie a quatre dimensions, 9.

planes can be isocline, 184; conjugate series isocline in opposite senses, 185 ; through any line pass two isocline to a given plane in opposite senses, 186; two intersecting planes are isocline to two
pairs of planes, 187; when two planes are isocline to a given plane in the same sense the common per-

Kant, reference to the number of dimensions of space, 3.

Keyser, the four-dimensional geometry of spheres, 1 1 ; our intuition of hyperspace, 16; the angles of planes, 114; proof that two planes

have

common
of

perpendicular

pendicular planes which they have with the given plane form a constant dihedral angle, 188; when two planes are isocline to a given plane in opposite senses, there is

plane, 118.

Kinematics
articles

four
Clifford,

dimensions,
Beltrami,

by

only one pair of common perpendicular planes, perpendicular to all sions, ii. three, 189; two planes isocline to a third in the same sense are Lagrange, time as the fourth dimenisocline to each other in this sense sion, 4. also, 190; poles and polar series, Lateral edges, faces, cells, hypersurall the planes of two conjugate face, etc., see Pyramid, Hyperseries are isocline at an angle of pyramid, Hypercone, etc. Layer, 241. 45 to a single pair of planes, 192 the converse also true, 193; in a Left, right and left in a plane, 154. system of planes isocline in a given Lewis, G. N., Wilson and Lewis on sense any two series have a pair relativity, 12. if a of planes in common, 193 Line, 24; "line AB," only one line contains two given points, 26 ; propplane intersects two isocline planes in lines the corresponding dihedral erties of its points, order, 27 ; den; ;

Craig, Hatzidakis, 13. Kwietnewski, complex variables represented in space of four dimen-

angles are equal, analogy to parallel

lines,

sity

and

continuity, 28;
;

opposite

sides in a plane, 38 at infinity, 231. 194-196; a system of planes isocline in a given sense forms a Linear elements of two planes, 61. two-dimensional geometry, 197 Lobachevsky, Pangeometry, 221. "ordinal" and "cardinal" system Loria, bibliography, 9. (Stringham), 198; a series cuts a
;

hypersphere (with centre at O) in McClintock, interpretations of Nona surface of double revolution with Euclidean Geometry, 10. equal radii, 220; projection from Methods of studying the higher one upon the other of two isocline geometries, 12.
planes produces similar figures, 229. Isocline rotation, every plane remains isocline to itself, 196.
Isosceles double pyramid, 204
;

Minkowski developed application

relativity, 12.

to

double

M6bius, symmetrical figures, 4. Moore, E. H., definition of segment,

21.

cone, 205.

Moore, R. L., properties of points on a line, 28; axioms of metrical Jacobi, generalizations of geometrical
formulae, 6.

geometry, 74.

342
More,
Henry,
3.

INDEX
spirits

are

four-di-

Opposite directions on
half-lines,

line,

27;

mensional,

Motion

in a plane does not change order in the plane, 160; in a hyperplane, does not change order in the hyperplane, 163; in hyperplanes, 54; sides of a hyperplane, half-hyperspaces, 62; elements of space, does not change order in Point Geometry, 113; elements of hyperspace, 166; in a plane with one point fixed, 167; the most Edge Geometry, 138; points in a general, 168; in a hyperplane, two hypersphere, 210, 213. equivalent if equivalent for three Order of points on a line, 27 ; "order non-collinear points, motion with 153 ; Veblen's use of the term " one point fixed, 169; every motion order," 27 ; cyclical, 28 order in a in a hyperplane equivalent to a two fundamental prinplane, 1 53 motion of a plane on itself or to a ciples, right and left sides of lines

cyclical order, 29; sides of a line in a plane, 38; half-planes, 39; sides of a plane in a hyperplane, half-hyper-

28;

in

AB"

screw motion, 170; in hyperspace, two equivalent if equivalent for four non-coplanar points, 172; motion with two points fixed, 173; with

through a point, 154; with respect to a triangle, "order ABC," 156; unchanged by any motion in the
158; independent of any hyperplane, 162 ; order in a hyperwith respect to a tetraplane, 161 hedron, "order 162; unchanged by any motion in the hyperplane, 163; order in hyperspace, 164; with respect to a penuntahedroid, "order changed by any motion, 165 order in Point Geometry, 179.
plane,
;

one point

fixed,

74

every motion

equivalent to a motion in which one plane remains fixed or moves only

ABCD"

on

itself,

174.
of, 24.
;

n dimensions, space
same
as

^-hedroid and 5 JV-hedroid, 311

i2o-hedroid

the
600-

and

ABCDE"
;

hedroid, 321.
;

Nets of hypercubes, 290 24-hedroids and i6-hedroids, 298; spherical Ovidio,

polygons, 303; polyhedral angles, polyhedrons, 304 ; polyhedroidal angles, 313; polyhedroids, 314; net of twenty tetrahedrons at a
point, 317
;

d',

projective geometry, 12.

Ozanam, higher products imaginary,

3-

Paciuolo, use of
2.

"primo

relato," etc.,

Pangeometry, term used by Lobaa fx>int, 324. chevsky, 221. Non-Euclidean Geometry used in the Parallel axiom, proofs which do not theory of relativity, 12; not pardepend on it, 77, 97, 105, 128, 136, ticularly considered in this text, 202; restrictions due to its omis18;
translations in, 146;

four dodekahedrons at

nets of

spherical polygons, 303;

see also

Elliptic Geometry, Hyperbolic Geometry, Parallelism, Restrictions.

sion, 37, 78, 79, 103, 108, 112, 138, *39 iS3 l6 J s66 Elliptic Geometry

and

Restrictions.

Non-Euclidean Geometry by the author, 19.

Parallelism taken

up after many other

Ndther, birational transformations,

8.

subjects, 19 ; parallel great circles in a hypersphere, 217; parallel

Object and plan of this book, 16, 73. Octahedroid, regular, 290; see Hypercubc.

motion in a hypersphere, 218; axiom of parallels, 221 parallel

;

lines

and

planes, 221

half-parallel

planes, their

common perpendicular

INDEX
lines

343
intersecting in a line,

and

mum
allel

planes, 224; distance, their

their mini-

two planes

linear

ele-

ments, 225; lines and planes parto a hyperplane, parallel hyperpianes, 226. Pasch, Axiom, 30. Pentahedroid, edges, faces, cells, 57 intersection with a plane, 57, 60;
;

87; perpendicular planes and hyperplancs, perpendicular along a line, 90; the planes perpendicular or absolutely perpendicular to planes lying in the hyperpianes, 91 ;
lines lying in either

and perpendic-

interior, coliinear with, 58;

pass-

59 intersection with a line, 60; the five half-hyperspaces and the interior, 62 sections, 65 ; its cells or its edges
cell to cell,
; ;

ing from

ular to the other, 92 ; planes with linear elements all perpendicular to a hyperplane, 94 ; perpendicular

spread out in a hyperplane, 68; the point equidistant from its vertices, 199; the point equidisits centre tant from its cells, 200
;

hyperpianes, 98; lines or planes lying in one and perpendicular to the other, 101 ; the common perpendicular line of two lines not in

one plane; lines with more than one common perpendicular line,
1 08;

of

pentahedroids with corresponding edges equal, 202 hypervolume in elliptic hyperspace, 287; regular, 203, 289; radii of circumscribed and ingravity,
; ;

201

line of

mon

the common perpendicular a line and plane ; the comperpendicular plane of two

planes

which

have

common

scribed hyperspheres, reciprocal pentahedroids, 289. Perpendicularity : lines perpendicular to a line at a point, 74; perpendicular line and hyperplane, 75; planes perpendicular to a line at a

in; perpendicular hyperplane, the common perpendicular planes of two planes which intersect only in a point, 118; planes
with an
infinite

number

of

com119,

mon
182.

perpendicular planes,

two lines perpendicular point; to a hyperplane lie in a plane, 76 lines perpendicular to a plane at a
;

Plan and object of this book, 16, 73. Plane, 24; "plane only one contains three given non-collinear divided by a line, 37 ; points, 35

ABC"

point, 80;

absolutely perpendicuif two planes interlar planes, 81

;

two planes with only odd point

;
;

in

sect in a line, their absolutely perpendicular planes at any point of

this line

intersection of common, 51, 81 two in a hyperplane, 53 opposite

sides of

a plane

in a hyperplane,

intersect in a line,

82

54;

intersection

with a hyper-

two planes absolutely perpendicular to a third lie in a hyperplane, 83

;

plane, 60;
planes, 61
;

linear elements of

two

perpendicular planes, simply perpendicular, half-perpendicular, or

perpendicular in a hyperplane; a plane perpendicular to one of two
absolutely perpendicular planes at their point of intersection is perpendicular to the other, 85; a
.plane intersecting

absolutely perpendicular planes, 81; perpendicular, 85; see Perpendicularity; if two not in a

hyperplane have a

common

per-

two absolutely perpendicular planes in lines is perpendicular to both, 86; the common perpendicular planes of

pendicular line, they have a common perpendicular 94 ; hyperplane, isocline planes, 123; see Isocline planes; planes at infinity, 231. Plane angle of a hyperplane angle, 96 ;
see Hyperplane angle.

Piano-conical
edge,

directing-curve,

hypersurface, vertexelements,

344

INDEX
;

intersection with a hyperplane, 71

Geometry
179;
in

see Double cone.

of a rectangular system, the theory of isocline

Piano-cylindrical
recting-curve,

hypersurface,

di-

256 ; interior, right directing-curves, similar directing-curves, hypersurface of 257; revolution, axis-plane,
elements,

planes, 180 ; the same as the geometry of the hypersphere, 216. Poles and polar series of isocline
planes, 193 ; poles of a sphere and polar circles in a hypersphere, 211 ; their motion in a rotation, 212 ; of

with piano-prismatic hypersurface, the set of cylinders, 258; intersection of two pianocylindrical hypersurfaces, 261 ; the
intersection

a hyperplane in Elliptic Geometry * of Four Dimensions, 217.

Polygon,
sides,

diagonals,

cyclical

surface of intersection, 262; relation to infinity, 264; see Prism

cylinder

order, 40; simple, convex, intersection with a line, 41 ; divided into

and Double

cylinder.

Piano-polyhedral angle, faces, vertexedge, cells, 133; elements, simple, convex, its hyperplane angles, 134 ; vertical piano-polyhedral angles, 134; directing polyhedral angles, polyhedral angles which are right sections, 135 ; theorems proved by means of them, 136; directingpolygons, 137 ; interior, 139, 140. Piano-prismatic hypersurface, 241 ; faces, cells, elements, simple, convex, sections, 242; directing-polygons,

two polygons, 42 interior, 44; the and interior, 45. Polyhedral angle, 133; convex, can
;

half -planes

be cut in a convex polygon, 137 ; nets of polyhedral angles, 304. Polyhedroid, edges, faces, cells, interior, 63 regular, definition, 289 ; can be inscribed in a hypersphere,
;

the associated net of hyperspherical polyhedrons, 309 ; reciprocal polyhedroids, 310; its polyhedroidal angles are regular, 313; nets of

243;

triangular,

similar

polyhedroids, 314; list of possible nets, 315; the nets in each of the

directing-polygons, 245; intersection of two piano-prismatic hypersurfaces, the two sets of prisms, 246; intersection with a piano-

non-Euclidean geometries, 316.

Polyhedroid formula, 302 proved by Schlafli, 22 by Stringham, 302. Polyhedroidal angle, elements, di; ;

cylindrical hypersurface, 258; see Double prism and Prism cylinder. Piano- trihedral angle, 134.

recting-polyhedron, polyhedral angles,

face
cells,

angles,
interior,

Plucker, the four coordinates of a line

in space, 10. Poincare' avoided

use of geometry

126; vertical polyhedroidal angles, 127; regular, axis, 312; the polyhedroidal angles of a regular polyhedroid are regular, net, reciprocal
nets, 313.

of four dimensions, 10 ;
situs, 12;

on

analysis

double integrals, 219. Polyhedron, 63; regular, can be inthe associated scribed in a sphere Point, 23; independent and absonet of spherical polygons reciprolutely independent points, 24; at cal polyhedrons, 303; nets, 304; infinity, 231. Point Geometry, 113; theorems in list of possible nets, 305 the nets in each of the non-Euclidean geomregard to perpendicular planes stated in the language of Point etries, 306.
; ; ;

formula, Descartes, Geometry, 114; applied to the Polyhedron Euler, 300. study of the angles of two planes, 114; piano-polyhedral angles and Popular interest in the fourth dimension, 9. polyhedroidal angles, 136; Point

INDEX
Powers of a number in early algebra, 2. Rectangular 239Prism, the two sets of prisms in a
double
prism,

345
hyperparallelopiped,

Double Rectangular system, 87, 89; as a tetrahedroidal angle, 128; ways in which it is congruent to itself, Prism cylinder, the set of cylinders, right, regular, spread out in a hy179; the different arrangements, the directing-polynotation, used in studying the perplane, 259 angles of two planes, 180; three gons and the directing-curves, belonging to the hypercube and generated by them and their i6-hedroid, 293. interiors; cut into two prism cylinders, 260; triangular, hyper- Regular hyperpyramid, 203; pentahedroid, 203, 289; double pyrcylinder with cylinders for bases as a prism cylinder, generated by the amid, 204; hyperprism, 237; the rotation of a prism, 261 volume, hypercube is regular, 240, 289; regular double prism, 248; prism 267; hypervolume, 285. Prismatic base of a hyperprism with cylinder, 259; polyhedroid, 289; octahedroid (hypercube), 290; 16tetrahedral ends, 274. Prismoidal hypersurface, directinghedroid, 291; 24-hedroid, 295; a regular polyhedroid can be inpolyhedron, edges, faces, cells, inscribed in a hypersphere, the assoterior, sections, 235 ; with parallelciated net of spherical polyhedrons, epiped for directing-poiyhedron,
246;
see

prism.

236.

Projecting line, 78
for area, 229.

plane, 84

factor

Projection of a line
;

upon a hyperplane, 78; a line or a part of a line, ii. 79, 84 upon a plane, 81 a line and its projection upon a plane not Restricted geometry, 19. coplanar, 84; of a plane upon a Restrictions to the second axiom of collinearity in Elliptic Geometry, hyperplane, 103; from a plane necessary in Edge Geometry, upon an isocline plane produces 25 similar figures, 229. 138; a convex piano-polyhedral
is
;

309; regular polyhedroidal angle, 312; in space of five dimensions only three regular figures, 317; 6oo-hedroid ,317; 1 2o-hedroid, 3 24. Relativity and the fourth dimension,

Protective geometry, points of a line

in cyclical order, 29. Ptolemy, the number of distances, i. Pyramid, base, edges, faces, inter-

angle

is restricted,

139;

in Point

Geometry a rectangular system is restrictions due restricted, 179; to omission of the axiom of parallels,

section with a plane, 55; double pyramid, 66 see Double pyramid.

;

see Parallel

axiom;

see also

Elliptic Geometry.

Revolution, surface of double revolu-

Quaternions applied to the study of geometry of four dimensions by

tion in a hypersphere of importance

in the theory of functions, see Double revolution.

219;

Hathaway and Stringham,

13.

Riemann

Ray or

half-line, 28.

Reciprocal figures in a hypersphere, 212; pentahedroids, 289; hypercube and i6-hedroid, 293; 24hedroids, 297; polyhedrons, 303; polyhedroids, 309 ; nets, see Nets.

on the foundations of geometry, Elliptic Geometry due to him, 6. Right and left in a plane, 154; see
Order.

Rotation in a plane, in a hyperplane, figures remain invariable, 141;

346

INDEX
plane, of a plane !n a hyperplane a property of the hyperplane, 162.

in hyperspace, the axis-plane, 142 ; figures remain invariable, rota-

tions around absolutely perpendicular planes commutative, 143 ; double rotation, 145 ; right and left in a plane, 154 ; when two rotations

Similar figures produced by projection from one upon the other of

two isocline planes, 229; directingpolygons of a piano-prismatic hypersurface, 245 ; directing-curves of a piano-cylindrical hypersurface,
257.

are equivalent to a single one in a hyperplane, 171, in hyperspace, 178; isocline, 196; of the hypersphere, the axis-circle and the circle of rotation, double rotation, screw

Simplex, sections, 14.

Simplicius, reference to Aristotle

and

motion, parallel motion, 218. Rudolph, use of terms representing powers of a number, 2.
Schlafli, multiple integrals, tiple continuity, 22.

Ptolemy,

i.

Solid, the interior of

a tetrahedron as
9.

solid, 54.

Sommerville, bibliography,
6;

mul- Space of
59;
52.
;

i, 2, 3,

...

w,

... dimen-

sions, 24;

Schotten, definitions of segment, 21.

of four dimensions, 24, ordinary space a hyperplane,

its

Schoute, etrie, g

Mehrdimensionale Geomsections of a simplex, 14

;

Sphere,

geometry

is

elliptic,

25;
in

descriptive geometry, 18 different kinds of perpendicularity, 85 ; the

closed, passing out of,

79;

polyhedroid formula, 302. Schubert, enumerative geometry, 12. Screw motion, the translation and rotation are commutative, 146; in the hypersphere, 218.
Sect, used

hypersphere, 207. Spherical sometimes used for hyperspherical, spherical dihedral angle,
Stifel

tetrahedron, 209. regards the higher powers as "against nature," 3.

by Halsted for segment,

25.

Straight, used

by Halsted

for line, 25.

Sections, study of a figure by them, 1 8 ; divide a figure into completely separated parts, 65, 245; of a

pentahedroid, hyperpyramid, etc., see these terms; of a piano-polyhedral angle, 135. Segment as defined by different writers, Hilbert, Enriques, E. H.

Stringham, application of quaternions to geometry of four dimensions, 13 on the angles of two planes, 114; use of the term "isoclinal angle," "ordinal" and "cardinal 125;
;

Moore,

Veblen,

Schotten,

21

systems," of isocline planes, 198; gave a proof of the polyhedroid formula, 302. Strip, the portion of a plane between

coldefinition, "segment AB," linear with, 25 ; interior, 28. Segre, the use of geometry of four

two parallel lines, 235. Surface, the tetrahedron as a surface, 54 ; of double revolution in a hypersphere, 219; see Double revolution. Sursolid in early algebra, sursolid
loci, 2.

dimensions, 10, 13.

Semi-parallel, the
lel,

same as

half-paral-

224.

Separate, of cyclical order, 29. Series of isocline planes, 182;

Isocline planes. Sides of a polygon,

see

Sylvester, early papers, 5 ; defence of the use of geometry of four dimensions, 10.

congruent, figures Symmetrical Mobius, 4, 149; defined as those triangle, 29; that can be placed in positions of line in a plane a property of the

defined, 21 ; polygon, 40; of a

how

INDEX
symmetry with respect
164.

347

to a plane,

two, 128; spherical, its volume, 210; net of tetrahedrons, twenty

Symmetry, 146 ;

in a plane, in a hy;

perplane, 147, in hyperspace, 149 rotations which leave the symmetrical relation undisturbed in a
plane, 147, in a hyperplane, 148, rotations in 152; hyperspace,
whicji bring into coincidence figures

Time
n.

at a point, 317. as the fourth dimension, La-

grange, d'Alembert, 4;

relativity,

Translations

symmetrical in a plane, 147, in a hyperplane, 148; symmetrical figures of ordinary geometry are really congruent, 148; symmetry in a hyperplane with respect to a point can be changed by rotation to symmetry with respect to a plane,
147 figures symmetrical in hyperspace with respect to a point or plane are congruent, 149; symmetry in hyperspace with respect to a line can be changed by rotation to symmetry with respect to a hy;

along a line, figures remain invariable, 145; different kinds of translation in Non-Euclidean Geometry, 146.

Triangle, sides, cyclical order, 29 ; intersection with a line, 30, 37 ; interior,

half-planes

30; collinear with, 32 ; the three and the interior, 39.

Use of studying the higher geometries,

13
;

see also Applications.

Veblen, definition of segment, 21 ; use of the term "order," 27 ; the properties of points on a line, 28 Axiom of Pasch, 29; axioms of metrical
;

perplane, 150;

in every kind of

geometry, 74.

symmetry corresponding segments Vector analysis of Grassmann, 8. and angles are equal, 153; figures Veronese, Fondamenti, 5, 8, 9; apsymmetrical in a plane cannot be made to coincide by any motion in the plane, 160 figures symmetrical in a hyperplane cannot be made to coincide by any motion in the hy;

geometry to theorems of ordinary geometry, 13 use of the elements at infinity, 230. Vertex-edge of a double pyramid, 67 double cone, piano-conical hy; ;

plication of the higher

perplane, 163

symmetrical in hyperspace cannot be made to coincide by any motion in hyper;

figures

persurface,
angle, 133
;

71

piano-polyhedral

hyperprism with tetra-

hedral ends, 274.

space,

1 66.

Synthetic development of the higher geometry, 4; advantages of the synthetic method over the analytic,
14.

polyhedroidal angles, 127; coincide, 167; piano-polyhedral angles, 134; cannot be made to coincide, 166. Vieta, use of terms representing

Vertical

can be made to

powers of a number,
Tetrahedroidal angle, 127; two with corresponding face angles equal, 1 29 the bisecting half-hyperplanes of its hyperplane angles have a common half-line ,131. Tetrahedron, edges, faces, intersection with a plane, 45 with a line, 48, 52 ; interior, collinear with, 49; the four half-hyperplanes and the
;
;

2.

Volume
209;
at

of a spherical dihedral angle,

interior,

54;

correspondence

of

210; 233; lateral volume of a hyperprism, hyperpyramid, frustum of a hyperpyramid, 265; hypercylinder, hypercone, 266, frustum of a hypercone, 267 ; volume of a double prism, 265, 266, prism 267; cylinder, double cylinder, hypersphere, 267.
spherical tetrahedron,
infinity,

348

INDEX
ciated

Wallis on the geometrical names of the higher powers, 3. Wilson and Lewis, relativity, 12.

with a hypercube and a

295
;

i6-hedroid,

reciprocal

24-

hedroids, 297; of i6-hedroids

298.

reciprocal

nets

and

24-hedroids,

i6-hedroid or hexadekahedroid, 291 ; axes, reciprocal relation to the hypercube, 292; diagonals of the hypercube and the i6-hedroid form three rectangular systems, 293 the
;

i2o-hedroid or hekatonikosahedroid,
construction,
parts, 326.

324;

table

of

its

reassociated 24-hedroid, 295; ciprocal nets of i6-hedroids and

6oo-hedroid, or hexakosioihedroid, construction of the first half, 317;

its seventy- two completed, 321 dekagons, 323; table of its parts,
;

24-hedroids, 298.

24-hedroid or ikosatetrahedroid, asso-

324.

>

HE

following pages contain advertisements of a

few of the Macmillan books on kindred subjects

GEOMETRY
BY

WALTER BURTON FORD

JUNIOR PROFESSOR OF MATHEMATICS, UNIVERSITY OF MICHIGAN

AND

CHARLES AMMERMAN

THE WILLIAM MCKINLEY HIGH SCHOOL, ST. Louis Edited by EARLE RAYMOND HEDRICK, Professor of Mathematics
in the University of Missouri
Plane and Solid Geometry, Plane Geometry,
cloth,
cloth,

i2mo 31 Q PP; $1.25

t

net

izmo, 213 pp., $oJSo net

Solid Geometry, cloth, i2mo, 106 pp. t $0.80 net

STRONG POINTS
authors and the editor are well qualified by training and experience to prepare a textbook on Geometry. II. As treated in this book, geometry functions in the thought of the
I.

The

pupil. III.

It

means something because

its

IV.

The logical as well as the practical side of the The arrangement of material is pedagogical.

practical applications are shown. subject is emphasized.

V. Basal theorems are printed in black-face type. VI. The book conforms to the recommendations of the National Committee on the Teaching of Geometry. VII. Typography and binding are excellent. The latter is the ree'nforced tape binding that is characteristic of Macmillan textbooks.
" Geometry is likely to remain primarily a cultural, rather than an informa" But the intimate connection tion subject," say the authors in the preface. of geometry with human activities is evident upon every hand, and constitutes
fully as much an integral part of the subject as does its older logical and scholastic aspect." This connection with human activities, this application of geometry to real human needs, is emphasized in a great variety of problems

and constructions, so
throughout the book.

that theory

and application are inseparably connected

These illustrations and the many others contained in the book will be seen to cover a wider range than is usual, even in hooks that emphasize practical
applications to a questionable extent. This results in a better appreciation of the significance of the subject on the part of the student, in that he gains a truer conception of the wi<le scope of its application. The logical as well as the practical side of the subject is emphasized. Definitions, arrangement, and method of treatment are logical. The definitions are particularly simple, clear, and accurate. The traditional manner of presentation in a logical system is preserved, with due regard for practical Proofs, both formal and informal, are strictly logical. applications.

THE MACMILLAN COMPANY

Publishers

64-66 Fifth Avenue

New York

THE CALCULUS
BY

ELLERY WILLIAMS DAVIS

PROFESSOR OF MATHEMATICS, THE UNIVERSITY OF NEBRASKA
Assisted

by WILLIAM CHARLES BRENKE, Associate Professoi Mathematics, the University of Nebraska

Edited by EARLE RAYMOND HEDRICK
Cloth, semi-flexible, xxi

of

+ 383 pp.

4- Tables (63), i2mo, $2.00 net

Edition

De

Luxe,

flexible leather binding,

India paper, $240

ttet.

as

This book presents as many and as varied applications of the Calculus it is possible to do without venturing into technical fields whose subject
is

matter

itself

unknown and incomprehensible

to the student,

and without

abandoning an orderly presentation of fundamental principles. The same general tendency has led to the treatment of topics with a view
toward bringing out their essential usefulness.
tion are not insisted upon, especially

Rigorous forms of demonstra-

where the precisely rigorous proofs

Rather the
is

would be beyond the present grasp of the student.

stress is laid

upon the

student's certain comprehension of that which

done, and his conuseful.

viction that the results obtained are both reasonable

and

At the

same

time,

an

effort

has been

made

to avoid those grosser errors and actual

misstatements of fact which have often offended the teacher in texts otherwise
attractive

and teachable.

Purely destructive criticism

are just as dangerous as ultra-conservatism.

and abandonment of coherent arrangement This book attempts to preserve

the essential features of the Calculus, to give the student a thorough training
in mathematical reasoning, to create in him a sure mathematical imagination,

and

to

meet

fairly

the reasonable

demand

for enlivening

and enriching the

subject through applications at the

tains

expense of purely formal

work

that con*

no

essential principle.

THE MACMILLAN COMPANY

Publithtr.

64-66 Fifth Avenue

Kew York

Analytic Geometry and Principles of Algebra

BY

ALEXANDER ZIWET
PROFESSOR OF MATHEMATICS, THE UNIVERSITY OF MICHIGAN

AND

LOUIS ALLEN HOPKINS

EARLE RAYMOND HEDRICK
,

INSTRUCTOR IN MATHEMATICS, THE UNIVERSITY OF MICHIGAN

Edited by

Cloth , viii -f 369 pp., appendix answers, index, I2mo, $1.60 net

This work combines with analytic geometry a number of topics traditionally treated in college algebra that depend upon or are closely associated with
geometric sensation. Through this combination it becomes possible to show the student more directly the meaning and the usefulness of these subjects.

The idea of coordinates is so simple that it might (and perhaps should) be Real explained at the very beginning of the study of algebra and geometry. analytic geometry, however, begins only when the equation in two variables
is

interpreted as defining a locus.

This idea must be introduced very gradu-

ally, as it is difficult for

the beginner to grasp.

The

familiar loci, straight

line

and

circle,

are therefore treated at great length.

these curves are given in the text

in connection with the circle.

In the chapters on the conic sections only the most essential properties of ; thus, poles and polars are discussed only
analytic geometry follows the

The treatment of solid

more usual
is

lines.

But,

given some the and of a function of variables two prominence; representation by contour lines as well as by a surface in space is explained and illustrated by practical
in view of the application to mechanics, the idea of the vector

examples.

The

exercises have

been selected with great care

in order not only to fur-

nish sufficient material for practice in algebraic

work but

also to stimulate

independent thinking and to point out the applications of the theory to concrete problems.

to

The number of exercises is sufficient to make a choice. To reduce the course presented in this book to about
on

allow the instructor

half

its

extent, the

parts of the text in small type, the chapters

solid analytic geometry,

and

the

more

difficult

problems throughout

may be

omitted.

THE MACMILLAN COMPANY

Publishers

84-66 Fifth Avenut

New York

4121-3

TRIGONOMETRY
BY

ALFRED MONROE KENYON

PROFESSOR OF MATHEMATICS, PURDUE UNIVERSITY

AND LOUIS INGOLD

ASSISTANT PROFESSOR OF MATHEMATICS, THE UNIVERSITY OF MISSOURI

Edited by EARLE RAYMOND HEDRICK

Trigonometry, flexible
cloih, pocket

Trigonometry

+ / 32 pp.) with Complete + 124 pp.), $1.35 net Tables (xviii + 12 pp.), $1.00 net (xi + 132 pp.) with Brief
size,

long

i2mo

(xi

Tables (xviii

Macmittan Logarithmic and Trigonometric Tables,

flexible cloth, pocket size, long

i2mo

(xviii

+ 124 pp.), $0.60 net

FROM THE PREFACE

The book contains a minimum of purely theoretical matter. Its entire organization is intended to give a clear view of the meaning and the immeThe proofs, however, are in a form that diate usefulness of Trigonometry.
will not require essential revision in

the courses that follow.

very large, and the traditional monotony is broken by illustrations from a variety of topics. Here, as well as in the text, the attempt is often made to lead the student to think for himself by giving of exercises
is

The number

suggestions rather than completed solutions or demonstrations. The text proper is short; what is there gained in space is used to
tables very complete
plete

make

the

and

usable.

Attention

is

called particularly to the

com-

and handily arranged table of squares, square roots, cubes, etc. ; by its use the Pythagorean theorem and the Cosine Law become practicable for actual computation. The use of the slide rule and of four-place tables is
encouraged for problems that do not demand extreme accuracy. Only a few fundamental definitions and relations in Trigonometry need be memorized; these are here emphasized. The great body of principles and
processes depends upon these fundamentals; these are presented in this book, as they should be retained, rather by emphasizing and dwelling upon that

dependence. Otherwise, the subject can have no real educational value, nor indeed any permanent practical value.

THE MACMILLAN COMPANY

Publishers

64-66 Fifth Avenue

New Tork