Lie Groups And Lie Algebras For Physicists Ashok

Das Susumu Okubo download

https://ebookbell.com/product/lie-groups-and-lie-algebras-for-
physicists-ashok-das-susumu-okubo-4913512

Explore and download more ebooks at ebookbell.com

Here are some recommended products that we believe you will be
interested in. You can click the link to download.

Problems And Solutions For Groups Lie Groups Lie Algebras With
Applications Wh Steeb

https://ebookbell.com/product/problems-and-solutions-for-groups-lie-
groups-lie-algebras-with-applications-wh-steeb-4421316

Lie Groups And Lie Algebras Chapters 79 Bourbaki N

https://ebookbell.com/product/lie-groups-and-lie-algebras-
chapters-79-bourbaki-n-2599780

Lie Groups And Lie Algebras A Physicists Perspective 1st Ed Adam M

Bincer

https://ebookbell.com/product/lie-groups-and-lie-algebras-a-
physicists-perspective-1st-ed-adam-m-bincer-4379940

Lie Groups And Lie Algebras Part Ii Chapters 46 Nicolas Bourbaki

https://ebookbell.com/product/lie-groups-and-lie-algebras-part-ii-
chapters-46-nicolas-bourbaki-5432730
An Introduction To Lie Groups And Lie Algebras 1st Edition Alexander
Kirillov

https://ebookbell.com/product/an-introduction-to-lie-groups-and-lie-
algebras-1st-edition-alexander-kirillov-1492676

Theory Of Groups And Symmetries Finite Groups Lie Groups And Lie
Algebras 1st Edition Alexey P Isaev

https://ebookbell.com/product/theory-of-groups-and-symmetries-finite-
groups-lie-groups-and-lie-algebras-1st-edition-alexey-p-isaev-10437566

Transformation Groups And Lie Algebras Nail H Ibragimov

https://ebookbell.com/product/transformation-groups-and-lie-algebras-
nail-h-ibragimov-4990700

Lie Algebras And Lie Groups Jeanpierre Serre

https://ebookbell.com/product/lie-algebras-and-lie-groups-jeanpierre-
serre-42860150

Basic Theory Of Algebraic Groups And Lie Algebras Gerhard P Hochschild

https://ebookbell.com/product/basic-theory-of-algebraic-groups-and-
lie-algebras-gerhard-p-hochschild-38406462
 Lie Groups and
Lie Algebras
for Physicists

9169_9789814603270_tp.indd 1 23/7/14 8:42 am

July 25, 2013 17:28 WSPC - Proceedings Trim Size: 9.75in x 6.5in icmp12-master

This page intentionally left blank

 Lie Groups and
Lie Algebras
for Physicists

Ashok Das
Susumu Okubo
University of Rochester, USA

World Scientific
NEW JERSEY • LONDON • SINGAPORE • BEIJING • SHANGHAI • HONG KONG • TA I P E I • CHENNAI

9169_9789814603270_tp.indd 2 23/7/14 8:42 am

 Published by
World Scientific Publishing Co. Pte. Ltd.
5 Toh Tuck Link, Singapore 596224
USA office: 27 Warren Street, Suite 401-402, Hackensack, NJ 07601
UK office: 57 Shelton Street, Covent Garden, London WC2H 9HE

British Library Cataloguing-in-Publication Data

A catalogue record for this book is available from the British Library.

LIE â•›GROUPS â•›AND â•›LIE â•›ALGEBRAS â•›FOR â•›PHYSICISTS

Copyright © 2014 Hindustan Book Agency (HBA)
Authorized edition by World Scientific Publishing Co. Pte. Ltd. for exclusive distribution worldwide
except India.

The distribution rights for print copies of the book for India remain with Hindustan Book Agency (HBA).

All rights reserved. This book, or parts thereof, may not be reproduced in any form or by any means,
electronic or mechanical, including photocopying, recording or any information storage and retrieval
system now known or to be invented, without written permission from the Publisher.

ISBN 978-981-4603-27-0

Printed in India, bookbinding made in Singapore.

Lakshmi - Lie Groups and Lie Algebras.indd 1 8/9/2014 2:52:18 PM

 To
Mary Okubo
July 25, 2013 17:28 WSPC - Proceedings Trim Size: 9.75in x 6.5in icmp12-master

This page intentionally left blank

Preface

Group theory was discovered by Évariste Galois in the 19th century

for the case of finite dimensional symmetric group. It has been suc-
cessfully generalized subsequently for the infinite dimensional case
(i.e. Lie group and Lie algebra) by Sophus Lie. Before the advent
of quantum mechanics in the early 20th century, group theory was
thought by many physicists to be unimportant in the study of physics.
It is indeed of some interest to note the following anecdote narrated
by Freeman Dyson (as quoted in Mathematical Apocrypha, p 21, by
S. A. Kranz), “In 1910, the mathematician Oswald Veblen (1880-
1960) - a founding member of the Institute for Advanced Study - and
the physicist James Jean (1877-1946) were discussing the reform of
the mathematics curriculum at Princeton University. Jeans argued
that they ‘may as well cut out group theory,’ for it ‘would never be of
any use to physics’.” The real fundamental change in thinking truly
occurred with the development of quantum mechanics. It was soon
realized that a deep knowledge of group theory and Lie algebra in the
study of of angular momentum algebra is crucial for real understand-
ing of quantum mechanical atomic and nuclear spectral problems. At
present, group theory permeates problems in practically every branch
of modern physics. Especially in the study of Yang-Mills gauge theory
and string theory the use of group theory is essential. For example, we
note that the largest exceptional Lie algebra E8 appears in heterotic
string theory and also in some one-dimensional Ising model in statis-
tical mechanics, some of whose predictions have been experimentally
verified recently.
Our goal in this monograph is to acquaint (mostly) graduate
students of physics with various aspects of modern Lie group and
Lie algebra. With this in view, we have kept the presentation of
the material in this book at a pedagogical level avoiding unnecessary
mathematical rigor. Furthermore, the groups which we will discuss
in this book will be mostly Lie groups which are infinite dimensional.
We will discuss finite groups only to the extent that they will be nec-

vii
viii Preface

essary for the development of our discussions. The interested readers

on this topic are advised to consult many excellent text books on
the subject. We have not tried to be exhaustive in the references.
Rather, we have given only a few references to books (at the end of
each chapter) that will be easier to read for a student with a physics
background. We assume that the readers are familiar with the mate-
rial of at least a two semester graduate course on quantum mechanics
as well as with the basics of linear algebra theory.
We would like to thank Dr. Brenno Ferraz de Oliveira (Braz-
ingá) for drawing all the Young tableau diagrams in chapter 5.

A. Das and S. Okubo

Rochester
Contents

Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vii
1 Introduction to groups . . . . . . . . . . . . . . . . . . . . . 1
1.1 Definition of a group . . . . . . . . . . . . . . . . . . 1
1.2 Examples of commonly used groups in physics. . . . 3
1.2.1 Symmetric group SN . . . . . . . . . . . . . . . 3
1.2.2 One dimensional translation group T1 . . . . . 6
1.2.3 One dimensional unitary group U (1) . . . . . . 9
1.2.4 U (N ) and SL(N ) groups . . . . . . . . . . . . 10
1.2.5 O(N ) and SO(N ) groups . . . . . . . . . . . . 13
1.2.6 Symplectic group Sp(2N ) . . . . . . . . . . . . 20
1.3 Group manifold . . . . . . . . . . . . . . . . . . . . . 24
1.4 References . . . . . . . . . . . . . . . . . . . . . . . . 28
2 Representation of groups . . . . . . . . . . . . . . . . . . . 29
2.1 Matrix representation of a group . . . . . . . . . . . 29
2.2 Unitary and irreducible representations. . . . . . . . 41
2.3 Group integration . . . . . . . . . . . . . . . . . . . . 51
2.4 Peter-Weyl theorem . . . . . . . . . . . . . . . . . . . 55
2.4.1 Fully reducible representation . . . . . . . . . . 55
2.4.2 Unitary representation . . . . . . . . . . . . . . 57
2.5 Orthogonality relations . . . . . . . . . . . . . . . . . 58
2.6 Character of a representation . . . . . . . . . . . . . 61
2.7 References . . . . . . . . . . . . . . . . . . . . . . . . 62
3 Lie algebras . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
3.1 Definition of a Lie algebra . . . . . . . . . . . . . . . 63
3.2 Examples of commonly used Lie algebras in physics. 66
3.2.1 Lie algebra of gl(N ) and sl(N ) . . . . . . . . . 66
3.2.2 Lie algebra of so(N ) . . . . . . . . . . . . . . . 67
3.2.3 Lie algebras of u(N ) and su(N ) . . . . . . . . . 71
3.3 Structure constants and the Killing form . . . . . . . 78
3.4 Simple and semi-simple Lie algebras . . . . . . . . . 87
3.5 Universal enveloping Lie algebra. . . . . . . . . . . . 97

ix
x Contents

3.6 References . . . . . . . . . . . . . . . . . . . . . . . . 98
4 Relationship between Lie algebras and Lie groups . . . . . 99
4.1 Infinitesimal group and the Lie algebra . . . . . . . . 99
4.2 Lie groups from Lie algebras . . . . . . . . . . . . . . 104
4.3 Baker-Campbell-Hausdorff formula . . . . . . . . . . 111
4.4 Ray representation . . . . . . . . . . . . . . . . . . . 120
4.5 References . . . . . . . . . . . . . . . . . . . . . . . . 124
5 Irreducible tensor representations and Young tableau . . . 125
5.1 Irreducible tensor representations of U (N ) . . . . . . 125
5.2 Young tableau . . . . . . . . . . . . . . . . . . . . . . 132
5.3 Irreducible tensor representationns of SU (N ) . . . . 139
5.4 Product representation and branching rule . . . . . . 142
5.5 Representations of SO(N ) groups . . . . . . . . . . . 149
5.6 Double valued representation of SO(3) . . . . . . . . 154
5.7 References . . . . . . . . . . . . . . . . . . . . . . . . 156
6 Clifford algebra . . . . . . . . . . . . . . . . . . . . . . . . . 157
6.1 Clifford algebra . . . . . . . . . . . . . . . . . . . . . 157
6.1.1 Dimension of the representation . . . . . . . . 157
6.1.2 Reducible representation . . . . . . . . . . . . . 161
6.1.3 Irreducible representation and its uniqueness . 164
6.2 Charge conjugation . . . . . . . . . . . . . . . . . . . 166
6.3 Clifford algebra and the O(N ) group . . . . . . . . . 171
6.4 References . . . . . . . . . . . . . . . . . . . . . . . . 179
7 Lorentz group and the Dirac equation . . . . . . . . . . . . 181
7.1 Lorentz group . . . . . . . . . . . . . . . . . . . . . . 181
7.1.1 Proper orthochronous Lorentz group . . . . . . 184
7.1.2 Orthochronous Lorentz group . . . . . . . . . . 185
7.1.3 Improper Lorentz group . . . . . . . . . . . . . 185
7.2 Generalized Clifford algebra . . . . . . . . . . . . . . 185
7.3 Dirac equation . . . . . . . . . . . . . . . . . . . . . . 187
7.3.1 Charge conjugation . . . . . . . . . . . . . . . . 195
7.3.2 Weyl and Majorana particles . . . . . . . . . . 197
7.4 References . . . . . . . . . . . . . . . . . . . . . . . . 199
8 Yang-Mills gauge theory . . . . . . . . . . . . . . . . . . . . 201
8.1 Gauge field dynamics . . . . . . . . . . . . . . . . . . 201
8.2 Fermion dynamics . . . . . . . . . . . . . . . . . . . . 209
8.3 Quantum chromodynamics . . . . . . . . . . . . . . . 212
8.4 References . . . . . . . . . . . . . . . . . . . . . . . . 218
Contents xi

9 Quark model and SUF (3) symmetry . . . . . . . . . . . . . 219

9.1 SUF flavor symmetry . . . . . . . . . . . . . . . . . . 219
9.2 SUF (3) flavor symmetry breaking . . . . . . . . . . . 226
9.3 Some applications in nuclear physics . . . . . . . . . 238
9.4 References . . . . . . . . . . . . . . . . . . . . . . . . 243
10 Casimir invariants and adjoint operators . . . . . . . . . . 245
10.1 Computation of the Casimir invariant I(ρ) . . . . . . 248
10.2 Symmetric Casimir invariants . . . . . . . . . . . . . 256
10.3 Casimir invariants of so(N ) . . . . . . . . . . . . . . 262
10.4 Generalized Dynkin indices . . . . . . . . . . . . . . 273
10.5 References . . . . . . . . . . . . . . . . . . . . . . . . 280
11 Root system of Lie algebras . . . . . . . . . . . . . . . . . . 281
11.1 Cartan-Dynkin theory . . . . . . . . . . . . . . . . . 281
11.2 Lie algebra A` = su(` + 1):. . . . . . . . . . . . . . . 297
11.3 Lie algebra D` = so(2`): . . . . . . . . . . . . . . . . 306
11.3.1 D4 = so(8) and the triality relation . . . . . . . 313
11.4 Lie algebra B` = so(2` + 1): . . . . . . . . . . . . . . 319
11.5 Lie algebra C` = sp(2`): . . . . . . . . . . . . . . . . 323
11.6 Exceptional Lie algebras . . . . . . . . . . . . . . . . 327
11.6.1 G2 . . . . . . . . . . . . . . . . . . . . . . . . . 328
11.6.2 F4 . . . . . . . . . . . . . . . . . . . . . . . . . 331
11.6.3 E6 . . . . . . . . . . . . . . . . . . . . . . . . . 333
11.6.4 E7 . . . . . . . . . . . . . . . . . . . . . . . . . 335
11.6.5 E8 . . . . . . . . . . . . . . . . . . . . . . . . . 336
11.7 References . . . . . . . . . . . . . . . . . . . . . . . . 339
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 341
July 25, 2013 17:28 WSPC - Proceedings Trim Size: 9.75in x 6.5in icmp12-master

This page intentionally left blank

Chapter 1

Introduction to groups

In this chapter, we introduce the concept of a group and present

in some detail various examples of commonly used groups in physics.
This is helpful in establishing the terminology as well as the notations
commonly used in the study of groups which will also be useful in
further development of various ideas associated with groups.

1.1 Definition of a group

Let us start with the formal definition of a group G as follows:

(G1): For any two elements a and b in a group, a product is defined

in G satisfying

ab = c ∈ G, ∀a, b ∈ G. (1.1)

(G2): The group product is associative so that

(ab)c = a(bc)(≡ abc), ∀a, b, c ∈ G. (1.2)

(G3): The group has a unique identity (unit) element e ∈ G such

that

ea = ae = a, ∀a ∈ G. (1.3)

This implies that

ee = e ∈ G. (1.4)

(G4): Any element a ∈ G has a unique inverse element a−1 ∈ G so

that

aa−1 = a−1 a = e. (1.5)

1
2 1 Introduction to groups

Any set of elements G satisfying all the axioms (G1)-(G4) is defined

to be a group. On the other hand, a set of elements which satisfies
only the first three axioms (G1)-(G3), but not (G4), is known as a
semi-group. (More rigorously, a semi-group is defined as the set of
elements which satisfy only (G1)-(G2). However, one can always add
the identity element to the group since its presence, when an inverse
is not defined, is inconsequential (see (1.3)-(1.4)) and we will adopt
this definition commonly used in physics.)
Some comments are in order here. We note that the definition
of a group does not require that the product rule satisfy the commu-
tativity law ab = ba. However, if for any two arbitrary elements of
the group, a, b ∈ G, the product satisfies ab = ba, then the group G
is called a commutative group or an Abelian group (named after the
Norwegian mathematician Niels Henrik Abel). On the other hand, if
the product rule for a group G does not satisfy commutativity law
in general, namely, if ab 6= ba for some of the elements a, b ∈ G, then
the group G is known as a non-commutative group or a non-Abelian
group. Furthermore, it is easy to see from the definition (G4) of the
inverse of an element that the inverse of a product of two elements
satisfies

(ab)−1 = b−1 a−1 6= a−1 b−1 , (1.6)

in general, unless, of course, the group G is Abelian. Equation (1.6)

is easily checked in the following way

(ab)(b−1 a−1 ) = a(b(b−1 a−1 )), by (1.2) with c = b−1 a−1 ,

= a((bb−1 )a−1 ), by (1.2),
= a(ea−1 ), by (1.5),
= aa−1 , by (1.3),
= e, by (1.5). (1.7)

Similarly, it is straightforward to verify that (b−1 a−1 )(ab) = e.

In order to illustrate the proper definitions in a simple manner,
let us consider the following practical example from our day to day
life. Let “a” and “b” denote respectively the operations of putting
on a coat and a shirt. In this case, the (combined) operation “ab”
would correspond to putting on a shirt first (b) and then putting
on a coat (a) whereas the (combined) operation “ba” would denote
putting on a coat first and then a shirt. Clearly, the operations are
not commutative, namely, ab 6= ba. If we now introduce a third
1.2 Examples of commonly used groups in physics 3

operation “c” as corresponding to putting on an overcoat, then the

law of associativity of the operations (1.2) follows, namely, (ca)b =
c(ab) = cab and corresponds to putting on a shirt, a coat and an
overcoat in that order. It now follows that the operation “bb” denotes
putting on two shirts while “b(bb) = (bb)b = bbb” stands for putting
on three shirts etc. The set of these operations would define a semi-
group if we introduce the identity (unit) element e (see (1.3)) to
correspond to the operation of doing nothing. However, this does
not make the set of operations a group for the following reason. We
note that we can naturally define the inverses “a−1 ” and “b−1 ” to
correspond respectively to the operations of taking off a coat and a
shirt. It follows then that a−1 a = e = b−1 b as required. However, we
also note that the operation “aa−1 ” is not well defined, in general,
unless there is already a coat on the body and, therefore, aa−1 6= e,
in general (but a−1 a = e always). As a result, the set of operations
so defined do not form a group. Nonetheless, (1.6) continues to be
valid, namely, (ab)−1 = b−1 a−1 6= a−1 b−1 , whenever these operations
are well defined. In fact, since ab corresponds to putting on a shirt
and then a coat, if we are already dressed in this manner, the inverse
(ab)−1 would correspond to removing these clothes (come back to
the state prior to the operation ab) which can only be done if we first
take off the coat and then the shirt. This leads to (ab)−1 = b−1 a−1 .
Furthermore, let us note that if “d” denotes the operation of putting
on a trouser, then the operation of putting on a shirt and a trouser
are commutative, namely, db = bd and so on.

1.2 Examples of commonly used groups in physics

In this section, we would discuss some of the most commonly used

groups in physics. This would also help to set up the conventional
notations associated with such groups.

1.2.1 Symmetric group SN . Although most of the commonly used

groups in physics are infinite dimensional (Lie) groups, let us begin
with a finite dimensional group for simplicity. Let us consider a set
of N distinct objects labelled by {xi }, i = 1, 2, · · · , N and consider
all possible distinct arrangements or permutations of these elements.
As we know, there will be N ! possible distinct arrangements (per-
mutations) associated with such a system and all such permutations
(operations) form a group known as the symmetric group or the per-
mutation group of N objects and is denoted by SN . Since the number
of elements in the group is finite, such a group is known as a finite
4 1 Introduction to groups

group and the number of elements of the group (n!) is known as the
order of the group SN . Permutation groups are relevant in some
branches of physics as well as in such diverse topics as the study of
the Rubik’s cube. Note that every twist of the Rubik’s cube is a
rearrangement of the faces (squares) on the cube and this is how the
permutation group enters into the study of this system. There are
various possible notations to denote the permutations of the objects
in a set, but let us choose the most commonly used notation known as
the cycle notation. Here we denote by (ij) the operation of permuting
the objects xi and xj in the set of the form

(ij) : xi ↔ xj , i, j = 1, 2, · · · , N, (1.8)

with all other objects left unchanged. In fact, the proper way to read
the relation (1.8) is as

(ij) : xi → xj → xi . (1.9)

This brings out the cyclic structure of the cycle notation (parenthesis)
and is quite useful when the permutation of more objects are involved.
It is clear from this cyclic structure that under a cyclic permutation

(ij) = (ji), (1.10)

and when permutations of three objects are involved we have

(ijk) : xi → xj → xk → xi , (1.11)

and

(ijk) = (kij) = (jki) (6= (jik) = (kji) = (ikj)) . (1.12)

Let us clarify the group structure of SN in the case of N = 3,

namely, let us consider the set of three objects (elements) {xi } =
(x1 , x2 , x3 ) where i = 1, 2, 3. With the cycle notation we note that
the 3! distinct permutations of these elements can be denoted by the
set of operations {P } with the elements in the set given by

P1 = e = (1), P2 = (12), P3 = (23), P4 = (13),

P5 = (123), P6 = (132), (1.13)

where we have introduced the identity element P1 = e = (1) to

correspond to doing no permutation (every object is left unchanged
1.2 Examples of commonly used groups in physics 5

under this operation). The product rule for the permutations can
now be understood in the sense of operations. For example,

P2 P3 = (12)(23), (1.14)

implies carrying out the permutation (23) first followed by the per-
mutation (12). We note that under this combined operation
(23) (12)
x1 −−→ x1 −−→ x2 ,
(23) (12)
x2 −−→ x3 −−→ x3 ,
(23) (12)
x3 −−→ x2 −−→ x1 , (1.15)

so that the combined operation is equivalent to

P2 P3 = (12)(23) = (123) = P5 ∈ {P }, (1.16)

and the group is closed under multiplication (see (1.1)). Physically

this can be understood from the fact that since {P } represents all
possible distinct arrangements of the objects in the set, any two ar-
rangements would lead to an element of the set. Following (1.15), we
can also show that

P3 P2 = (23)(12) = (132) = P6 ∈ {P }, (1.17)

which shows that the product rule for the permutations is not com-
mutative

P2 P3 6= P3 P2 . (1.18)

The associativity of the product of permutations (1.2) can also

be easily verified. For example, we note that

(P2 P3 )P4 = P5 P4 = (123)(13) = (23) = P3 ,

P2 (P3 P4 ) = (12)((23)(13)) = (12)(123) = (23) = P3 , (1.19)

so that associativity holds and we have

(P2 P3 )P4 = P2 (P3 P4 ) = P2 P3 P4 . (1.20)

The identity element P1 = e of the group corresponding to no per-

mutation is clearly unique and satisfies (see (1.3))

Pα P1 = P1 Pα = Pα , α = 1, 2, · · · , 6. (1.21)
6 1 Introduction to groups

Finally, the inverse of the permutations is determined easily as

follows. For any pair permutation (ij), we see that doing the permu-
tation twice brings it back to the original arrangement of the objects.
Therefore, we have

P2 P2 = P3 P3 = P4 P4 = e = P1 , (1.22)

and we can identify (identity is its own inverse, (P1 )−1 = P1 )

(P2 )−1 = P2 , (P3 )−1 = P3 , (P4 )−1 = P4 . (1.23)

On the other hand, for permutations involving all three objects we

see that (we are using (1.16) and (1.17) as well as (1.22))

P5 P6 = P2 P3 P3 P2 = P2 P2 = P1 , (1.24)

which determines that P6 is the inverse of P5 and vice versa. This

shows that S3 indeed satisfies all the postulates of a group and a
similar proof can be extended to the case of SN as well.
After the discussion of this simple finite dimensional group, let
us next move onto infinite dimensional (Lie) groups (named after the
Norwegian mathematician Sophus Lie) which are more commonly
used in the study of modern physics.

1.2.2 One dimensional translation group T1 . Let x denote a real vari-

able (namely, x ∈ R and it can be either a spatial coordinate or the
time coordinate, for example). For any real constant parameter a,
we define an operation T (a) which acting on the coordinate x simply
translates it by an amount a such that

T (a)x = x + a, ∀x ∈ R. (1.25)

This is to be interpreted in one of the following two ways. First, we

can think of T (a) as the active transformation
T (a)
T (a) : x −−→ x0 = x + a. (1.26)

Alternatively, we can interpret T (a) as a linear operator satisfying

(1.25) with the action of T (a) leaving any constant parameter b un-
changed (a linear operator does not act on constants, only on vectors),
namely,

T (a)b = b. (1.27)
1.2 Examples of commonly used groups in physics 7

In either case, the combined effect of two translations can now

be determined. Viewed as an active transformation, for example, we
can write
T (b)
T (a)T (b) : x −−→ x0 = x + b, followed by
T (a)
x0 −−→ x00 = x0 + a = x + b + a, (1.28)

which, in the notation of (1.25), corresponds to

(T (a)T (b)) x ≡ T (a) (T (b)x) = (T (b)x) + a = x + b + a

= x + (b + a) = T (b + a)x = T (a + b)x. (1.29)

One can represent the transformations in (1.28) diagrammatically as

shown in Fig. 1.1. On the other hand, considered in terms of linear
operators, it is easy to check with the use of (1.27) that

T (b) T (a)

x b x! a x!!

Figure 1.1: The combined operation of two translations T (a)T (b)

where a translation T (b) is followed by another translation T (a).

(T (a)T (b)) x ≡ T (a) (T (b)x) = T (a) (x + b) = T (a)x + T (a)b

= x + a + b = T (a + b)x = T (b + a)x, (1.30)

which coincides with (1.29).

We note that we may identify T (a) with the differential operator

d
T (a) = 1 + a , (1.31)
dx
acting on x so that
 
d
T (a)x = 1 + a x = x + a,
dx
 
d
T (a)b = 1 + a b = b. (1.32)
dx
8 1 Introduction to groups

Here 1 denotes the identity operator. It follows now that

  
d d
T (a)T (b)x = 1+a 1+b x
dx dx
d2
 
d
= 1 + (a + b) + ab 2 x
dx dx
 
d
= 1 + (a + b) x = T (a + b)x, (1.33)
dx

consistent with (1.30) (or (1.29)).

Equations (1.29) and (1.30) show that the product T (a)T (b),
the combined operation of two translations, is well defined and cor-
responds also to a translation, namely,

T (a)T (b) = T (b + a) = T (a + b) = T (b)T (a), ∀a, b ∈ R. (1.34)

This defines the product rule for the operation (see (1.1)) which, in
this case, is commutative. The associativity of the product (1.2) is
easily verified from

(T (a)T (b))T (c) = T (a + b)T (c) = T ((a + b) + c)

= T (a + b + c),
T (a)(T (b)T (c)) = T (a)T (b + c) = T (a + (b + c))
= T (a + b + c), (1.35)

where (1.34) has been used and (1.35) leads to

(T (a)T (b)) T (c) = T (a) (T (b)T (c)) . (1.36)

The unit element e (see (1.3)) in this case can be identified with

e = T (0), (1.37)

so that it satisfies

eT (a) = T (0)T (a) = T (a) = T (a)T (0) = T (a)e. (1.38)

Finally, the inverse of a translation T (a) by an amount a can be

defined as a translation by an amount (−a)

(T (a))−1 = T (−a), (1.39)

1.2 Examples of commonly used groups in physics 9

so that (see (1.5))

T (a) (T (a))−1 = T (a)T (−a) = T (0) = e

= T (−a)T (a) = (T (a))−1 T (a). (1.40)

Thus, we see that the set of translations T1 = {T (a)} with the

parameter of translation a ∈ R (namely, −∞ < a < ∞) satisfy all
the properties (G1)-(G4) defining a group. This is a continuous group
since the parameter of translation can take any continuous value in R
and is known as the one dimensional translation group. Furthermore,
since the product rule for this group is commutative (see (1.34)), this
is an example of an Abelian group. To conclude this subsection, we
note that if the parameters of translation were restricted to a finite
interval L1 ≤ a ≤ L2 , then the set of translations will not define a
group since two successive translations may not remain in the interval.

1.2.3 One dimensional unitary group U (1). Let z be a complex variable

and let us consider the phase transformation

z → z 0 = eiθ z, (1.41)

where θ is a real constant angular parameter 0 ≤ θ < 2π. We can

write the phase transformation (1.41) in terms of operators also as
(see also (1.25))

U (θ)z = eiθ z. (1.42)

Following the discussion in section 1.2.2 we can show from (1.41) (or
(1.42)) that the product

U (θ1 )U (θ2 ) = U (θ1 + θ2 ) = U (θ2 )U (θ1 ), (1.43)

is well defined if we identify U (θ + 2π) = U (θ) (since e2πi = 1)

whenever θ1 + θ2 = θ + 2π ≥ 2π. We note that in this case we can
identify († denotes the Hermitian conjugate, in this one dimensional
case, only the complex conjugate)

U † (θ) = U −1 (θ) = U (−θ), (1.44)

namely,

U † (θ)z = U −1 (θ)z = U (−θ)z = e−iθ z, (1.45)

so that the operator U (θ) is unitary

U (θ)U † (θ) = e = U † (θ)U (θ). (1.46)

10 1 Introduction to groups

The set of phase transformations U (1) = {U (θ), 0 ≤ θ < 2π}

forms a group called the one dimensional unitary group U (1) (also
written as U1 ). We see from (1.43) that it is a commtative group.
Let us also note that we can restrict the range of the parameter of
transformation to 0 ≤ θ < 2π since e2πi = 1 and this has to be
contrasted with the parameters of translations T1 discussed earlier.
Correspondingly, U (1) is known as a compact group (parameters have
a compact range) as opposed to T1 which is called a non-compact
group. We will discuss these concepts in more detail in section 1.3.

1.2.4 U (N ) and SL(N ) groups. We can generalize the group of one

dimensional phase transformations U (1) discussed in section 1.2.3 to
N dimensions as follows. Let A denote a N × N unitary matrix,
namely,
 
α11 α12 · · · α1N
 α21 α22 · · · α2N 
 
A= . , (1.47)
 
 .. .. .. .. 
 . . . 

αN 1 αN 2 · · · αN N
where the matrix elements αij , i, j = 1, 2, · · · , N are complex param-
eters such that

AA† = EN = A† A, (1.48)

Here EN (= 1N ) denotes the N × N unit (identity) matrix so that

the unitarity condition (1.48) leads to the condition
N
X N
X
∗ ∗
αik αjk = δij = αki αkj , (1.49)
k=1 k=1

where δij denotes the familiar Kronecker delta. For i = j, equation

(1.49) yields
N
X N
X
|αik |2 = 1 = |αki |2 , (1.50)
k=1 k=1

which, in turn, implies that

|αik | ≤ 1, i, k = 1, 2, · · · , N. (1.51)

Let us further note that

∗
det A† = det AT = (det A)∗ , det EN = 1, (1.52)
1.2 Examples of commonly used groups in physics 11

where († , T , ∗ ) denote respectively Hermitian conjugate, transpose

and complex conjugate of the matrix. With these, we can take the
determinant of both sides in (1.48) to obtain
det A (det A)∗ = | det A|2 = 1,
or, | det A| = 1, (1.53)
so that A−1 always exists and can be identified, as in (1.44), with
(see (1.48))
A−1 = A† , (1.54)
and it follows that A−1 is also unitary. For example,

†  †
A−1 A−1 = A† A† = A† A = EN , (1.55)

which follows from (1.48). Similarly, if A and B are both unitary

matrices, then their matrix product AB also is. Finally, the N × N
unit (identity) matrix EN is unitary. Thus, we conclude that the set
of all N × N unitary matrices U (N ) = {A} form a group known as
the unitary group U (N ) (or UN ).
In a similar manner, if we consider all N × N matrices A satis-
fying only the condition
det A = 1, (1.56)
we can show that they also form a group. From the matrix prod-
uct, it follows that, for any two matrices satisfying (1.56), we have
det(AB) = det A det B = 1 since each of the matrices has unit de-
terminant, namely, det A = 1 = det B. Furthermore, we note that
det A−1 = (det A)−1 = 1 and det EN = 1. (Note that since the ma-
trices have unit determinant, the inverses exist.) This group is known
as the special linear group in N dimensions and is denoted by SL(N )
(or SLN ). If the set of N × N matrices {A}, in addition to satis-
fying (1.56), also satisfy the unitary condition (1.48), the resulting
group is known as the special unitary group in N dimensions and is
denoted by SU (N ) (or SUN ) which is a subgroup of both SLN and
UN . (The notion of a subgroup will be discussed later.) For N ≥ 2,
these groups are not Abelian (non-Abelian).
The U (N ) group is of interest in the operator algebra of har-
monic oscillators in quantum mechanics. Let ai and a†j with i, j =
1, 2, · · · , N denote respectively the annihilation and the creation op-
erators for a N dimensional oscillator system satisfying
[ai , aj ] = 0 = [a†i , a†j ], [ai , a†j ] = δij . (1.57)
12 1 Introduction to groups

Here [a, b] = ab − ba denotes the commutator of two linear operators

a and b. Let us next consider the linear canonical transformation of
the annihilation and the creation operators in (1.57) of the form

N
X
ai → bi = αij aj ,
j=1

N
a†i → b†i = a†j αij
X
∗
, (1.58)
j=1

where αij denote constant parameters. It follows now that if the

parameters αij satisfy the unitarity condition (1.49), the transformed
operators satisfy

[bi , bj ] = 0 = [b†i , b†j ], [bi , b†j ] = δij , (1.59)

for i, j = 1, 2, · · · , N . The last relation in (1.59) (which is the only

one that needs checking), for example, can be seen as

N N N
" #
[bi , b†j ] = a†` αj` [ak , a†` ]
X X X
∗ ∗
αik ak , = αik αj`
k=1 `=1 k,`=1

N
X N
X
∗ ∗
= αik αj` δk` = αik αjk = δij , (1.60)
k,`=1 k=1

where we have used (1.57) in the intermediate step as well as (1.49)

in the last step.
Since the N × N matrix A defined by (1.48) (or equivalently
by (1.49)) is unitary, we conclude that the canonical commutation
relations (1.57) for the creation and annihilation operators of the
harmonic oscillator are invariant under the unitary group of transfor-
mations U (N ). More importantly, the Hamiltonian operator (normal
ordered although the presence of a zero point energy does not modify
the result) for the harmonic oscillator system defined by

N
a†i ai ,
X
H= (1.61)
i=1

is also invariant under U (N ). This is easily checked from the fact

1.2 Examples of commonly used groups in physics 13

that (see (1.58))

 
N N N N
!
b†i bi a†j αij
X X X X
∗
=  αik ak
i=1 i=1 j=1 k=1

N N
a†j ak
X X
∗
= αik αij
j,k=1 i=1

N N
a†j ak δjk a†j aj = H,
X X
= = (1.62)
j,k=1 j=1

where we have used (1.49) in the intermediate step.

The results in (1.59) and (1.62) can be seen more easily as fol-
lows. Let us arrange the two sets of N annihilation operators into
column matrices as
   
a1 b1
 a2   b2 
   
â =  .  , b̂ =  .  . (1.63)
   
 ..   .. 
   
aN bN

Correspondingly the two sets of creation operators ↠and b̂† would
be denoted by row matrices. With this notation, the transformation
(1.58) can be written in the matrix form (see also (1.47))

b̂ = Aâ, b̂† = ↠A† , (1.64)

and the relations (1.59) and (1.62) follow if the transformation matrix
A is unitary, namely, if it satisfies (1.48). For example, we note
that (the summation over all modes is automatically contained in
the matrix multiplication)

b̂† b̂ = ↠A† Aâ = ↠1â = ↠â = H, (1.65)

The invariance of the Hamiltonian of a dynamical system under a

transformation defines a symmetry of the system as we will discuss
later.

1.2.5 O(N ) and SO(N ) groups. In the earlier example of the unitary
group studied in section 1.2.4 we assumed the N × N matrix A to be
14 1 Introduction to groups

a complex unitary matrix. On the other hand, if we restrict further

the matrix A to be a real matrix, namely,

A∗ = A, (1.66)

then (1.48) leads to

AAT = EN = AT A, (1.67)

which follows from the fact that A† = (A∗ )T = AT for real matrices.
Here AT denotes the transpose of the matrix A, namely,

(AT )ij = Aji , (1.68)

so that explicitly we can write (see (1.47))

 
α11 α21 · · · αN 1
 α12 α22 · · · αN 2 
 
AT =  . . (1.69)
 
 .. .. .. .. 
 . . . 

α1N α2N · · · αN N
We note that the restriction on the matrix elements following from
(1.67) has the form
N
X N
X
aik ajk = δij = aki akj , (1.70)
k=1 k=1
∗ = α ). In
which can be compared with (1.49) (for real elements, αij ij
particular, for i = j, relation (1.70) leads to
N
X N
X
2
(aik ) = 1 = (aki )2 , i = 1, 2, · · · , N, (1.71)
k=1 k=1

so that the elements of the matrix A are restricted to have values

−1 ≤ aij ≤ 1 for any i, j = 1, 2, · · · , N .
We can readily verify as in the last section that the set of all N ×
N real matrices satisfying (1.67) forms a group which is called the real
orthogonal group in N dimensions and is denoted by O(N ) (or ON )
(the matrices A satisfying (1.67) are known as orthogonal matrices).
Furthermore, taking the determinant of both sides of (1.67), we note
that since det AT = det A, we have

(det A)2 = 1,
or, det A = ±1. (1.72)
1.2 Examples of commonly used groups in physics 15

If we require the orthogonal matrices A to satisfy the additional con-

dition

det A = 1, (1.73)

then it can be shown that all such matrices define the real special
orthogonal group in N dimensions which is denoted by SO(N ) (or
SON ). Sometimes it is also known as the group of rotations in N
dimensions identified as RN . This nomenclature can be understood as
follows. Let us consider the N dimensional Euclidean space labelled
by the coordinates
 
x1
 x2 
x =  . . (1.74)
 
 .. 
xN

Rotations define the maximal symmetry (which leaves the origin in-
variant) of such a space and the length of a vector in this space defined
as

xT x = x21 + x22 + · · · + x2N , (1.75)

is invariant under rotations. Let us note that if we define rotations

through N × N matrices R as

x → x0 = Rx, (1.76)

it follows from (1.75) that the length remains invariant, namely,

x0 T x0 = xT RT Rx = xT x,
or, (x01 )2 + (x02 )2 + · · · + (x0N )2 = x21 + x22 + · · · + x2N , (1.77)

only if

RT R = EN . (1.78)

Comparing (1.78) with (1.67) we recognize that the rotation matrices

R in N dimensional Euclidean space correspond to orthogonal ma-
trices and correspondingly the O(N ) group is also identified with the
rotation group RN . The difference between the O(N ) and the SO(N )
groups lies in the fact that the latter group does not include “mirror
reflections” (which render the matrices A to have det A = −1). (Note
that the length of a vector defined in (1.75) is clearly invariant under
16 1 Introduction to groups

“mirror reflections” xi → −xi for some of the coordinates.) Corre-

spondingly SO(N ) is also known as the group of proper rotations. We
note here that the rotation group SO3 (without “mirror reflections”)
plays an important role in the study of the Schrödinger equation with
central potentials in general and the group SO4 clarifies the origin of
the “accidental degeneracy” in the case of the Hydrogen atom (which
we discuss in chapter 3).
A group that is very important in the study of relativistic sys-
tems is the Lorentz group. It is the maximal symmetry group (which
leaves the origin fixed) of the four dimensional space-time Minkowski
manifold (space) just as the three dimensional rotations (which leave
the origin invariant) define the maximal symmetry group of the three
dimensional Euclidean manifold (space) (and are important in non-
relativistic phenomena). In the case of the four dimensional space-
time manifold, if we define a coordinate four vector as (t denotes the
time coordinate, x the three dimensional spatial coordinate vector
and c represents the speed of light)

xµ = (x0 , x)
= (x0 , x1 , x2 , x3 )
= (ct, x1 , x2 , x3 ), µ = 0, 1, 2, 3, (1.79)

then the length invariant under a Lorentz transformation (which in-

cludes three dimensional rotations as well as Lorentz boosts), is given
by

c2 t2 − x2 = c2 t2 − (x1 )2 − (x2 )2 − (x3 )2 , (1.80)

and is known as the invariant length of the Minkowski space. Denot-

ing the coordinate four vector by a four component column matrix
(see (1.74))
 
 0 ct
x  x1 
x= = , (1.81)
x  x2 
x3

and introducing a 4 × 4 constant diagonal matrix (metric)

 
1 0 0 0
0 −1 0 0
η= 0 0 −1
, (1.82)
0
0 0 0 −1
1.2 Examples of commonly used groups in physics 17

we note that we can write the invariant length (1.80) as

xT ηx = c2 t2 − (x1 )2 − (x2 )2 − (x3 )2 . (1.83)

Under a Lorentz transformation the coordinates transform as

 0  0 0  0
x x x
x1  x0 1 1
 → x0 =   = A x  = Ax,
 
x=x 2 0
x  2 x2  (1.84)
x3 x0 3 x3

where A is a 4 × 4 matrix implementing the Lorentz transformation.

Since the coordinates are real, it follows that the matrix A represent-
ing the Lorentz transformation is also real, namely,

A∗ = A. (1.85)

Furthermore, the invariance of the length (1.80) (or (1.83)) now leads
to (the constant matrix η does not transform under a Lorentz trans-
formation)

x0 T ηx0 = xT AT ηAx = xT ηx, (1.86)

which is possible only if the matrix A denoting a Lorentz transfor-

mation satisfies the relation

AT ηA = η. (1.87)

Relation (1.87) is quite similar to (1.67) (see also (1.78)) if we replace

the 4 × 4 constant diagonal matrix η by E4 (= 14 ). Therefore, we sus-
pect that the Lorentz transformations are related to the orthogonal
matrices in some manner.
To understand this better, we note that the diagonal elements
of the relation (1.87) lead explicitly to
3
X
(a0µ )2 − (aiµ )2 = 1, µ = 0, 1, 2, 3. (1.88)
i=1

Comparing with (1.71) we see that the matrix elements of a Lorentz

transformation satisfy the same relation as those for the orthogonal
matrices except for the relative negative signs in the last three terms
(which basically arise from the structure of the metric η). Corre-
spondingly, the Lorentz transformations are said to belong to the
orthogonal group O(3, 1) which is different from the group O(4) in
that the metric in the case of the former has three negative and one
18 1 Introduction to groups

positive signature (as opposed to the case of O(4) where the metric
is the identity matrix).
We point out here that it is possible to identify

x4 = −ix0 = −ict, (1.89)

and introduce a coordinate four vector as

x̃µ = (x, x4 )
= (x1 , x2 , x3 , x4 )
= (x1 , x2 , x3 , −ict), µ = 1, 2, 3, 4, (1.90)

and denote it as a four component column matrix as

 1 
x
 x2 
 
x
x̃ = =  x3  .
 (1.91)
x4
−ict

In this case the invariant length would have the standard form with-
out (the metric) η (see (1.75))

−x̃T x̃ = −(x2 + (x4 )2 ) = −((x1 )2 + (x2 )2 + (x3 )2 − c2 t2 )

= c2 t2 − (x1 )2 − (x2 )2 − (x3 )2 . (1.92)

If we now define a Lorentz transformation through a 4 × 4 matrix Ã

as
 1  0 1  1
x̃ x̃ x̃
x̃2  0 2 2
0 x̃  x̃ 
  
x̃3  → x̃ = x̃0 3  = Ã x̃3  = Ãx̃,
x̃ =  (1.93)


x̃4 x̃0 4 x̃4

then the invariance of the length (1.92) under the Lorentz transfor-
mation (1.93) would require

−x̃0 T x̃0 = −x̃ÃT Ãx̃ = −x̃T x̃, (1.94)

which is possible only if

ÃT Ã = E4 . (1.95)

This is precisely the orthogonality condition in (1.67) and this would

suggest that the Lorentz transformations belong to the orthogonal
1.2 Examples of commonly used groups in physics 19

group O(4). However, as we have already seen, this is not true and
the difference arises from the fact that because the coordinates in
(1.90) or (1.91) are now complex (and the Lorentz transformation
preserves the nature of the coordinates), the transformation matrix
à is no longer real (see (1.66)),

Ã∗ 6= Ã. (1.96)

In fact, from the nature of the coordinates in (1.91), it follows that

the matrix elements of the transformation matrix would have the
structure

ã∗ij = ãij , ã∗44 = ã44 , ã∗4i = −ã4i , ã∗i4 = −ãi4 , (1.97)

where i, j = 1, 2, 3. As a result, the diagonal elements of the orthog-

onality relation in (1.87) in this case lead to
4
X
(ãνµ )2 = 1, (1.98)
ν=1

for any fixed value of µ = 1, 2, 3, 4. For example, for µ = 4, relation

(1.98), together with (1.97), leads to
3
X
(ãi4 )2 + (ã44 )2 = |ã44 |2 − |ãi4 |2 = 1, (1.99)
i=1

which brings out the O(3, 1) nature of the Lorentz transformations

discussed in (1.88). Note that (1.99) implies that in this case we can
have |ã44 |, |ãi4 | ≥ 1 in contrast to the U (N ) case (see (1.51)).
Let us conclude this section by simply noting that when we
combine space-time translations with the Lorentz transformations,
we obtain the Poincaré transformations and they form an interest-
ing group. This is an important symmetry group in the study of
relativistic systems. In four dimensions, for example, the Poincaré
transformations act on the space-time coordinates as
 0  0 0  0  0
x x x a
x1  0 1 1 1
x x
 → x0 =   = A   + a  ,
  
x= x2  x0 2  x2  a2  (1.100)
x3 x0 3 x3 a3
where the 4 × 4 matrix A defines a Lorentz transformation (discussed
in (1.84)) and aµ = (a0 , a1 , a2 , a3 ) = (aµ )∗ denotes the four vector
representing space-time translations. We will go into more details of
these two important groups in physics in a later chapter.
20 1 Introduction to groups

1.2.6 Symplectic group Sp(2N ). Let us introduce the 2N ×2N matrix

 of the form
 
0 1 0 0 0 0 0 0

···

 −1 0 0 0 0 0 0 0 
 0 0 0 1 0 0 0 0 
 ··· 
 0 0 −1 0 0 0 0 0 
2N = ,
 0 0 0 0 .. 0 0 
 . ··· 

 0 0 0 0 0 0 

 0 0 0 0 0 0 0 1 
···
0 0 0 0 0 0 −1 0
(1.101)

which is a block
diagonal matrix with the anti-symmetric 2×2 matri-

ces iσ2 = −1 0 1 along the diagonal (σ represents the second Pauli
0 2
matrix). By construction, this matrix is anti-symmetric,

T2N = −2N , (2N )2 = −E2N , (1.102)

where E2N = 12N denotes the 2N × 2N identity (unit) matrix. Such

a matrix is known as a symplectic structure (also known as a complex
structure) and is by definition even dimensional. Let us next consider
a 2N × 2N matrix A which preserves the symplectic structure 2N ,
namely,

AT 2N A = 2N = A2N AT . (1.103)

(The relation (1.103) can be compared with (1.67) with 2N → E2N =
12N .) Such a matrix is known as a symplectic matrix in 2N dimen-
sions and the set of all 2N × 2N matrices satisfying (1.103) can be
shown to define a group known as the symplectic group in 2N dimen-
sions denoted by Sp(2N ) (or Sp2N ). Any symplectic matrix satisfying
(1.103) is known to have unit determinant (det A = 1) so that it is
automatically special. We do not go into the general proof of this
here since it is rather involved.
The symplectic groups are of some relevance in the study of
quantum mechanics. For example, it is straightforward to show that
the real symplectic group Sp(2N, R) (which consists of symplectic
matrices which are real) is the invariance group of the N -dimensional
Heisenberg algebra. Let us recall that if qi and pi with i = 1, 2, · · · , N
denote the canonical coordinate and momentum (Hermitian) opera-
tors, they satisfy the canonical (Heisenberg) commutation relations

[qi , qj ] = 0 = [pi , pj ], [qi , pj ] = i~δij , (1.104)

1.2 Examples of commonly used groups in physics 21

for i, j = 1, 2, · · · , N . This is commonly known as the Heisenberg

algebra. Let us next introduce a 2N -dimensional vector operator
ξµ , µ = 1, 2, · · · , 2N with the components

ξµ = (q1 , p1 , q2 , p2 , · · · , qN , pN ). (1.105)

In terms of ξµ , the Heisenberg algebra (1.104) can be written com-

pactly as

[ξµ , ξν ] = i~µν , µ, ν = 1, 2, · · · , 2N, (1.106)

where µν corresponds to the (µ, ν) element of the matrix  defined

in (1.101).
Let us now make a transformation of the vector ξµ as
2N
X
ξµ → ξµ0 = βµν ξν , (1.107)
ν=1

where βµν are constant parameters. It follows now that

" 2N 2N
# 2N
X X X
[ξµ0 , ξν0 ] = βµλ ξλ , βνσ ξσ = βµλ βνσ [ξλ , ξσ ]
λ=1 σ=1 λ,σ=1

2N
X
= i~ βµλ βνσ λσ . (1.108)
λ,σ=1

This implies that the transformed variables ξµ0 would also satisfy the
Heisenberg algebra

[ξµ0 , ξν0 ] = i~µν , (1.109)

if the transformation parameters βµν satisfy

2N
X 2N
X
βµλ βνσ λσ = βµλ λσ βνσ = µν . (1.110)
λ,σ=1 λ,σ=1

Introducing the 2N × 2N matrix

 
β11 β12 ··· β1(2N )
 β21 β22 ··· β2(2N )
 

A= . , (1.111)
 
 .. .. .. ..
 . . . 

β(2N )1 β(2N )2 · · · β(2N )(2N )
22 1 Introduction to groups

we note that we can write (1.110) in the matrix form as (compare

with (1.103))

AAT = . (1.112)

Namely, we see that the real symplectic group of transformations

Sp(2n, R) defines the invariance group of the N -dimensional Heisen-
berg algebra (the reality condition arises from the fact that the op-
erators are Hermitian). Unfortunately, however, the Hamiltonian,
say for the harmonic oscillator system, is not invariant under the
Sp(2n, R) group of transformations defined by (1.107) and (1.112).
More explicitly,
N N
1X 2
1X
pi + qi2 6= (p0i )2 + (qi0 )2 ,


H= (1.113)
2 2
i=1 i=1

where

ξµ0 = (q10 , p01 , q20 , p02 , · · · , qN

0
, p0N ), (1.114)

so that it is not a symmetry of the harmonic oscillator system.

The special case of the symplectic group for N = 1 is of some
interest because it arises in various branches of physics. In this case,
we can identify Sp(2) ≡ SL(2) as follows. Let
 
α β
A= , (1.115)
γ δ

denote a 2 × 2 symplectic matrix. Therefore, by definition (see

(1.103)) it satisfies

AT 2 A = 2 ,
     
α γ 0 1 α β 0 1
or, = ,
β δ −1 0 γ δ −1 0
   
0 (αδ − βγ) 0 1
or, = ,
−(αδ − βγ) 0 −1 0
   
0 1 0 1
or, det A = ,
−1 0 −1 0
or, det A 2 = 2 , (1.116)

where we have used the definition of the determinant,

det A = αδ − βγ. (1.117)

1.2 Examples of commonly used groups in physics 23

Equation (1.116) determines that

det A = αδ − βγ = 1, (1.118)

for a mtrix to belong to Sp(2). However, this is also the condition for
a matrix to belong to SL(2) and this proves the equivalence (isomor-
phism) between the two dimensional groups Sp(2) and SL(2). This
fact can be used to construct 2-component spinors in physics as we
will show in a later chapter.
These two dimensional groups are also quite relevant in the
study of Bogoliubov transformations (among other things). For ex-
ample, let a and a† denote the annihilation and creation operators
for a harmonic oscillator system satisfying (see, for example, (1.57))

[a, a] = 0 = [a† , a† ], [a, a† ] = 1. (1.119)

Let us consider a transformation of the operators of the form

a → b = αa + βa† ,

a† → b† = β ∗ a + α∗ a† , (1.120)

where we assume that the constant parameters of the transformation

α, β satisfy

αα∗ − ββ ∗ = 1. (1.121)

In this case, it is straightforward to check that the transformed op-

erators satisfy the algebra

[b, b] = 0 = [b† , b† ], [b, b† ] = 1. (1.122)

For example, we note that

[b, b† ] = [αa + βa† , β ∗ a + α∗ a† ]

= αα∗ [a, a† ] + ββ ∗ [a† , a] = αα∗ − ββ ∗ = 1. (1.123)

We note that if we define the matrices

     
a b α β
â = , b̂ = † , A = , (1.124)
a† b β ∗ α∗

then we can write the transformation (1.120) in the matrix form as

â → b̂ = Aâ, A2 AT = AT 2 A = det A 2 . (1.125)

24 1 Introduction to groups

Here we have used the identification in (1.116). The commutation

relations in (1.119) can now be written compactly as

[âi , âj ] = (2 )ij , i, j = 1, 2. (1.126)

Calculating the commutation relations for the transformed operators

in (1.125) (see (1.108)), it follows now that

[b̂i , b̂j ] = A2 AT ij = det A (2 )ij ,

(1.127)

for i, j = 1, 2. The invariance condition for the commutation rela-

tions, (1.121), is now seen to correspond to

det A = 1. (1.128)

A transformation mixing the annihilation and creation operators is

an example of a Bogoliubov transformation and we see that it is
implemented through a matrix A ∈ SL(2). (This is useful in the case
of harmonic oscillator, for example, in the construction of coherent
states.) We note here that the matrix A is not unitary, namely,
A† A 6= E2 .
To conclude this discussion, we note that the groups SL(N ), U (N ),
O(N ), SO(N ), Sp(2N ) are termed classical groups and discussed in
detail by Hermann Weyl in his famous book Classical Groups. The
book also includes a discussion of the symmetric group SN as well as
a more general linear group GL(N ). Without going into details, we
simply note here that the general linear group GL(N ) consists of all
N × N matrices A with only the restriction det A 6= 0 (nonsingular)
so that the inverse A−1 always exists.

1.3 Group manifold

In this chapter, we have discussed a finite group (with a finite number

of elements) as well as several (Lie) groups which are quite impor-
tant in physics. The (Lie) groups, as we have seen depend on some
continuous parameters and correspondingly they are known as con-
tinuous groups which have an (nondenumerably) infinite number of
elements. Let us introduce the concept of a group manifold for such
groups which would help us in classifying infinite dimensional contin-
uous groups (examples of which we have studied in the last section)
into compact and non-compact groups. Basically, the parameters
on which elements of a continuous group depend define a topologi-
cal space known as the parameter space. Since the group elements
1.3 Group manifold 25

depend on these parameters, every element can be put in one-to-

one correspondence with points on the parameters. In this way, a
continuous group inherits the topology of the parameter space and
correspondingly is also known as a topological group.
We will illustrate these ideas through examples. Let us consider
the one dimensional translation group T1 discussed in section 1.2.2.
In this case an element of the group T (a) represents a translation
of the coordinate (which can be a space or a time coordinate) by a
real continuous parameter “a” with −∞ < a < ∞. Therefore, the
parameter space, in this case, corresponds to the real line and it is
clear that we can represent every group element of T1 diagrammati-
cally as a point on the real line. For example, a specific element T (a)
would have the coordinate “a” on this real line from the origin which
can be identified with the identity element e = T (0) (see (1.37)) as
shown in Fig. 1.2. Topologically the real line can be identified with
a Cartesian axes, say for example, the x-axis. Therefore, this shows
that there is a one-to-one correspondence between the elements of
the group T1 and points on the x-axis and we can identify the group
manifold of T1 with the whole (entire) real line.

e = T (0) T (a)

0 a

Figure 1.2: Diagrammatic representation of one dimensional transla-

tions on a real line.
Let us next consider the example of the one dimensional unitary
group U1 discussed in section 1.2.3. A typical group element U (θ) =
eiθ , 0 ≤ θ < 2π corresponds to a simple (one dimensional) phase
rotation by an angle θ (see, for example, (1.41) and (1.42)). As a
result, the parameter space, in such a case, can be identified with the
unit circle (labelled only by an angle) and every element U (θ) of the
group U1 can be assigned a unique angle θ on the unit circle (angle
which a vector from the origin to the unit circle makes with the x-
axis) as in Fig. 1.3. Namely, there is a one-to-one correspondence
between the elements of the group U1 and points on the unit circle
and we can identify the group manifold of U1 with the unit circle. It is
worth noting here that since U (θ + 2πn) = U (θ) for any integer N , if
the parameter θ were allowed to take any arbitrary value (without the
restriction 0 ≤ θ < 2π), there can no longer be a simple one-to-one
correspondence between the group elements and the unit circle. In
26 1 Introduction to groups

this case, the parameter space coincides with that of the real line and
the topological space known as the covering group of U (1) becomes
equivalent to the group manifold of T1 .

θ
x

Figure 1.3: Diagrammatic representation of the one dimensional uni-

tary group on a unit circle.

We now turn to the more complicated case of the group UN

discussed in section 1.2.4. Here the group elements are realized as
N × N unitary matrices A whose elements αij , i, j = 1, 2, · · · , N are
complex parameters. In general, a matrix with N 2 complex param-
eters (elements) can be equivalently thought of as described by 2N 2
independent real parameters. However, the unitarity condition (1.48)

AA† = EN = A† A, (1.129)

imposes N 2 real conditions on the elements of a N ×N unitary matrix

(note that the identity (unit) matrix is real which is the reason for
the N 2 real conditions). As a result, the matrix A can be thought
of as being parameterized by N 2 independent real parameters which
we can take to be, for example, Re αij , i, j = 1, 2, · · · , N . Therefore,
it would appear that the parameter space can be identified with a
Euclidean manifold with N 2 independent axes. However, we note
from (1.51) that the magnitudes of the elements αij are bounded by
unity so that all the points in this space must satisfy

|Re αij | ≤ 1, (1.130)

and the space corresponding to the unit ball in N 2 dimensions is

bounded just like the unit circle (in the case of U (1)). Since every
1.3 Group manifold 27

element of U (N ) can be mapped to a point in this unit ball, in this

case the group manifold can be identified with the unit ball in N 2
dimensions. The group manifolds for more complicated continuous
groups can similarly be constructed and are much more complicated
in general. Nonetheless we can always assign a correspondence be-
tween the elements of a given group and points on a topological space
(defined by the parameters) which is known as the group manifold
for the particular group.
A topological space with a given topology (see, for example,
the book by J. L. Kelley) is called compact if it has a bounded and
closed topology. The space is known as non-compact otherwise. The
continuous groups are correspondingly classified as compact or non-
compact depending on the nature of the group manifold (topological
space) associated with them. It is clear from this that the group of
translations T1 is non-compact since the real line associated with this
group is not bounded. On the other hand, the unitary group U (N )
is compact since the group manifold is bounded. In a similar man-
ner, it can be shown that the groups SU (N ), SO(N ) and O(N ) are
also compact. However, it is worth noting that orthogonal groups
of the form O(m, n) with nontrivial m, n are non-compact. For ex-
ample, in the case of the group O(3, 1) we note from (1.88) that the
parameters of the transformation (matrix elements aµν ) do not have
to be bounded in order to satisfy (1.88) (because it involves a differ-
ence). As a consequence, Lorentz transformations are non-compact
(the boosts are not bounded). Similarly, the Poincaré group is also
non-compact since the space-time translations are not bounded (in
addition to the fact that the Lorentz transformations are, in general,
not bounded). We note here that the group SLN is also not compact.
This can be easily seen by restricting to n = 2 where we can label
the matrix as (see (1.115))
 
α β
A= , (1.131)
γ δ

satisfying (see also (1.117) and (1.118))

det A = αδ − βγ = 1. (1.132)

However, without the unitarity condition (1.48) (for example, in the

case of SU (2)) or the orthogonality condition (1.67), the matrix ele-
ments α, β, γ, δ satisfying only (1.132) are no longer bounded (namely,
they can be arbitrarily large and still satisfy (1.132)). As a result,
the group manifold (and, therefore, the group) is non-compact. We
28 1 Introduction to groups

note that since the groups SL(2) and Sp(2) are equivalent, Sp(2) is
non-compact as well (so is Sp(2N )). Finally, we note that any finite
group can be thought of as compact since it is bounded and closed
(in the sense of a finite set).

1.4 References

H. Weyl, The Classical Groups, Princeton University Press, Prince-

ton, NJ (1939).

H. Weyl, The Theory of Groups and Quantum Mechanics, (trans-

lated from German by H. P. Robertson), Dover Publication,
NY (1950).

J. L. Kelley, General Topology, Van Nostrand, NY (1957).

Chapter 2

Representation of groups

While the structure of the abstract group is quite interesting, in

physics the representation of the group itself is more important. The
most common representations of (Lie) groups are in terms of finite di-
mensional n×n matrices, although sometimes the infinite dimensional
(unitary) representations may be more meaningful for non-compact
groups. In this chapter we will discuss the matrix representations
as well as representations of groups by linear operators in a vector
space.

2.1 Matrix representation of a group

The n×n matrix representation of a group G is defined as follows. For

every element of the group a ∈ G, if we can construct an invertible
n × n matrix U (a) such that

(i) we can identify

U (e) = En , (2.1)

where En = 1n denotes the n × n identity (unit) matrix,

(ii) and for any two elements a, b ∈ G, the matrices U (a) and U (b)
satisfy the relation

U (a)U (b) = U (ab), (2.2)

then, the set of all such n × n matrices {U (a)} forms a group and
provides a (n × n matrix) representation of the group G. We note
that in (2.2) the product on the left hand side is the natural associa-
tive matrix product which defines the product rule for such a matrix
group.
Indeed it is straightforward to check that such a set of matrices
forms a group in the following manner.

29
30 2 Representation of groups

(i) We note from the defining relation (2.2) that

U (a)U (b) = U (ab) = U (c), (2.3)

is well defined with the matrix product rule and is in the set of
matrices {U (a)}. Here we have used the group property (1.1),
namely, for any two elements a, b ∈ G of the group, the group
multiplication leads to ab = c ∈ G.

(ii) Since the matrix product is known to be associative, it follows

that for any three matrices in the set we have

(U (a)U (b))U (c) = U (a)(U (b)U (c)), (2.4)

which should be compared with (1.2). In fact, given the group

associative property (1.2) we note that using (2.2) we can write

(U (a)U (b))U (c) = U (ab)U (c) = U ((ab)c),

U (a)(U (b)U (c)) = U (a)U (bc) = U (a(bc)), (2.5)

so that we have

(U (a)U (b))U (c) = U (a)(U (b)U (c)) = U (abc), (2.6)

since (ab)c = a(bc) = abc (see (1.2)).

(iii) By definition (2.1), the set has a unique identity (unit) element
U (e) = En which, using the matrix product, satisfies

U (e)U (a) = En U (a) = U (a) = U (a)En = U (a)U (e), (2.7)

which should be compared with (1.3). The relation (2.7) can

also be viewed from the point of view of group multiplication
as

U (e)U (a) = U (ea) = U (a) = U (ae) = U (a)U (e), (2.8)

corresponding to ea = a = ae.

(iv) Since by definition

U (a)U (a−1 ) = U (aa−1 ) = U (e) = En

= U (a−1 )U (a), (2.9)
2.1 Matrix representation of a group 31

it follows that

det U (a) det U (a−1 ) = det U (a)U (a−1 )

= det U (e) = det En = 1. (2.10)

This shows that

det U (a) 6= 0, (2.11)

so that (U (a))−1 exists and satisfies

U (a)(U (a))−1 = En = (U (a))−1 U (a). (2.12)

Comparing (2.9) and (2.12) we conclude that

(U (a))−1 = U (a−1 ). (2.13)

Conversely, we can multiply (2.12) with U (a−1 ) either from left

or from right and use the group composition to show that the
identification in (2.13) follows.

This shows that the set of n × n matrices constructed satisfying (2.1)

and (2.2) forms a group and provides a matrix representation for the
group G.
We note here that if the set of matrices {U (a)} defines a n × n
matrix representation of a group G, then the set {S −1 U (a)S}, where
S denotes any arbitrary n × n invertible matrix, also provides a n × n
matrix representation of the group G. This is easily seen by noting
that if we identify

W (a) = S −1 U (a)S, (2.14)

then, using the matrix product

(i) we have

W (e) = S −1 U (e)S = S −1 En S = S −1 S = En , (2.15)

(ii) and, furthermore,

W (a)W (b) = S −1 U (a)SS −1 U (b)S = S −1 U (a)En U (b)S

= S −1 U (a)U (b)S = S −1 U (ab)S = W (ab), (2.16)
32 2 Representation of groups

which can be compared with (2.1) and (2.2) respectively. Follow-

ing the same arguments as before, we conclude that the set {W (a) =
S −1 U (a)S} also defines a group and provides a n×n matrix represen-
tation of the group G. This suggests that the matrix representation
of a group G may not be unique since we can construct infinitely
many sets of matrices similarly related to the set {U (a)} for different
invertible matrices S. However, we regard all such sets of matrices,
related by a similarity transformation, to be equivalent and do not
consider {W (a)} to provide a distinct representation of G. This can
be understood in the following manner.
In the last chapter, we have pointed out that abstract groups can
be realized as groups of transformations acting on appropriate linear
vector spaces. For example, we have seen that the SO(N ) group
can be associated with the group of rotations in a N -dimensional
coordinate space while the U (N ) group rotates the creation and an-
nihilation operators of N harmonic oscillators. Transformations in
a linear vector space may be implemented through linear operators
acting on vectors. (Only for a very unique discrete transformation
is the corresponding operator anti-linear.) Therefore, given a group
G, we may identify every element of the group (a ∈ G) with a linear
operator A in a finite dimensional vector space Vn ,

a → A. (2.17)

Let us consider an orthonormal basis {xi }, i = 1, 2, · · · , n in Vn .

A linear operator in Vn takes one vector in the vector space into
another which can be expanded in the basis |xi i (the Dirac notation
for a state vector). In particular, we note that
n
X
Axi = x0i = αji xj , i = 1, 2, · · · , n. (2.18)
j=1

The n2 coefficients of expansion αij describing the action of the linear

operator A on the n basis vectors xi , i = 1, 2, · · · , n can be thought
of as the components of a n×n matrix giving a matrix representation
U for the linear operator A. In this way, we can construct a n × n
matrix representation for the group G through linear operators acting
on Vn , namely,

a → A → U (a), (2.19)

with

(U (a))ij = αij , (2.20)

2.1 Matrix representation of a group 33

which can also be written explicitly as

 
α11 α12 · · · α1n
 α21 α22 · · · α2n 
 
U (a) =  . . (2.21)
 
 .. .. .. .. 
 . . . 

αn1 αn2 · · · αnn

As a result, the relation (2.18) can also be written as

n
X
Axi = (U (a))ji xj . (2.22)
j=1

We can verify that these matrices indeed satisfy the group prop-
erties in the following manner.

(i) We note that if we identify (where 1 denotes the identity oper-

ator)

e → 1, (2.23)

then it follows that since the identity operator leaves every vec-
tor invariant

1xi = xi , i = 1, 2, · · · , n, (2.24)

and we can identify

(U (e))ij = δij ,
 
1 0 ··· 0
0 1 ··· 0
 
or, U (e) =  . = En . (2.25)
 
 .. . .. 
 0 .. .
0 0 ··· 1

(ii) Furthermore, for any two other elements of the group

b → B, c → C, (2.26)

where B and C are linear operators acting on Vn , if we have

ab = c, (2.27)
34 2 Representation of groups

then we can write

AB = C. (2.28)
On the other hand, following (2.18)-(2.22), we note that we can
write
X n Xn
Bxi = βji xj = (U (b))ji xj ,
j=1 j=1
n
X n
X
Cxi = γji xj = (U (c))ji xj , (2.29)
j=1 j=1

where we have identified

(U (b))ij = βij ,
 
β11 β12 ··· β1n
 β21 β22 · · · β2n 
 
or, U (b) =  . ,
 
 .. .. .. .. 
 . . . 

βn1 βn2 · · · βnn

and

(U (c))ij = γij ,
 
γ11 γ12 ··· γ1n
 γ21 γ22 · · · γ2n 
 
or, U (c) =  . . (2.30)
 
 .. .. .. .. 
 . . . 

γn1 γn2 · · · γnn

It follows now that (by definition, a linear operator acts only

on vectors and not constants)
n
X n
X
ABxi = A (U (b))ji xj = (U (b))ji Axj
j=1 j=1
n
X
= (U (b))ji (U (a))kj xk
j,k=1
n
X
= (U (a)U (b))ki xk . (2.31)
k=1
2.1 Matrix representation of a group 35

Comparing (2.31) with (2.28) and (2.29), we conclude that

(U (a)U (b))ij = (U (ab))ij = (U (c))ij . (2.32)

It follows now from (2.1) and (2.2) that the set of matrices
{U (a)} do satisfy the group properties and provide a matrix
representation of the group.

However, we note that the matrix elements of a linear operator

are basis dependent. For example, as we have seen in (2.20) and
(2.22), in the basis xi , i = 1, 2, · · · , n we can write
n
X
Axi = (U (a))ji xj . (2.33)
j=1

On the other hand, if we have another basis |yi i, i = 1, 2, · · · , n in

Vn , related to |xi i through a similarity transformation S (det S 6= 0
by definition of a similarity transformation, namely, it is invertible),
we can write
n
X
yi = Sji xj . (2.34)
j=1

It follows now that

n
X n
X
Ayi = A Sji xj = Sji Axj
j=1 j=1
n
X n
X
= Sji (U (a))kj xk = Sji (U (a))kj (S −1 )`k y`
j,k=1 j,k,`=1
n
X
= (S −1 U (a)S)`i y` , (2.35)
`=1

where we have used (2.22) as well as (2.34). The coefficients of ex-

pansions in the new basis |yi i are given by the elements of the matrix
S −1 U (a)S which would also provide a matrix representation of the
group, namely, W (a) = S −1 U (a)S (see (2.14)). Thus, we see that
the same linear operator can lead to different matrix representations
of the group depending on the choice of basis. Since the choice of a
basis does not alter the physical contents of operators, all such ma-
trix representations are considered to be equivalent to U (a) and not
distinct from U (a) unless stated explicitly.
36 2 Representation of groups

As an aside we note that in (2.34) we have represented the sim-

ilarity transformation as
n
X
yi = Sji xj , (2.36)
j=1

instead of the conventional notation

X n
yi = Sij xj . (2.37)
j=1

The two relations (2.36) and (2.37) are, of course, physically equiv-
alent and can be understood as associated with the active or the
passive description of transformations. For example, let us consider
a given vector V in the vector space which can be expanded in the
basis xi as
n
X
V = vi xi . (2.38)
i=1

Under a transformation
V → V 0 = SV, (2.39)
the change in the vector can be viewed in one of the two following
equivalent ways. We may expand the transformed vector in the same
basis as
Xn
V0 = vi0 xi , (2.40)
i=1

which can be thought of as implying (see (2.39)) that the basis (of
expansion) is unchanged under the transformation, rather the coef-
ficients of expansion change. In other words, the vector transforms,
but not the basis (of expansion). This is known as the active picture
of transformation. In this case, using the linear nature of operators
and (2.18) we determine from (2.38)-(2.39) and (2.40) that
n
X n
X
0
V = SV = vi Sxi = vi Sji xj
i=1 i,j=1
n
X n
X
= (Sv)j xj = vi0 xi ,
j=1 i=1
n
X
or, vi0 = (Sv)i = Sij vj , (2.41)
j=1
2.1 Matrix representation of a group 37

so that we can write

n
X n
X
V0 = vi0 xi = Sij vj xi . (2.42)
i=1 i,j=1

In particular, in this picture, if we consider the vector V = xi , for

which the expansion (2.38) takes the form (δij represents the Kro-
necker delta)
n
X
V = xi = δij xj , (2.43)
j=1

then it follows from (2.42) that under the transformation

n
X
xi → x0i = yi = Sij xj , (2.44)
j=1

which coincides with (2.37) and shows that this form is associated
with the active view of the transformation.
Alternatively, we can view the transformation in (2.39) as chang-
ing the basis of expansion leaving the coefficients of expansion un-
changed. This would then correspond to leaving the vector unchanged
while changing the basis and is known as the passive view of the co-
ordinate transformation. In this case, we can write
n
X
0
V = vi x0i . (2.45)
i=1

Using (2.42), we conclude that in this case we have

n
X
x0i = yi = Sji xj , (2.46)
j=1

which coincides with (2.36) and shows that this form is associated
with the passive view of the transformation. In our discussion we have
been using a passive transformation and (2.36) is correspondingly the
appropriate relation.
I Example (Translations in one dimension). Let us end the discussion of this sec-
tion with an example of the representation of the translation group in one dimen-
sion T1 through matrices as well as through linear operators. Let us recall that
the group of translations in one dimension is defined by (see (1.26))
T (a)
x −−−→ x + a, (2.47)
 Another Random Document on
Scribd Without Any Related Topics
 If the plate has been over-exposed, which will be indicated by the
image appearing quickly, wash off the developer with water, add
more bromide to the solution and immerse the plate again.
If, on the contrary, the plate has not had enough time, it will be
indicated by developing very slowly and with a very thin and weak
image; when such is the case add more of the iron solution, however,
not more than to make the proportion of the iron nearly p60 one-
fourth of the whole solution. More than this proportion will cause a
yellow precipitate, which will settle on the surface of the film and
destroy it. If by this course you succeed in securing sufficient detail in
the shadows, the negative may afterwards be strengthened.
It is as well, however, when a plate is found to be undertimed
(and it is possible to make another from the same subject) to throw it
away; it will scarcely repay further attention.
It will be seen that a considerable margin in time of exposure can
be allowed and provided for. Yet none the less it is of the utmost
importance to learn to give the correct and proper exposure to
secure the finest results.
This, it is true, is not always possible, and as an under-exposed
plate is lost, any error in exposure would better be in too much
rather than in too little time.
It is said by some that by the use of citrate of soda almost any
degree of over-exposure may be remedied, but there remains the
fact that for a restraining agent to be of any value its application to a
negative in development must be almost instantaneous upon the
earliest discovery of the over-exposure.
When a large number of plates have been exposed and the first
developed indicates that all have been overtimed, then a 50 per cent.
solution of citrate of soda, used with judgment, might result in saving
the lot.
When over-timing makes itself visible unexpectedly and it is feared
that the development will have proceeded too far before the retarder
can act, a good plan is to take the plate out of the dish and flow over
it a solution of citric acid, containing a little chrome alum, p61 which
will immediately stop the action of the developer and give time to
make other arrangements.

AMMONIA PYRO DEVELOPER.

No. 1.
Pyrogallic acid 1 oz.
Alcohol 4 oz.
No. 2.
Water 60 oz.
Bromide of ammonium 60 grs.
Strong liquor ammonia 2 drs.
To use, take of
No. 1 or pyro solution 1 dr.
No. 2 or ammonia solution 10 oz.
This will be enough to develop three or four 8×10 plates, if they
are properly timed.
The above is one of the best forms of ammonia pyro developer; it
gives negatives of very fine quality.
Many dry plates, on development with ammonia and pyro, show a
green fog by reflected light, and a pink or red fog by transmitted
light; this fog does not materially injure the printing quality of the
plate, but spoils its appearance to the eye, and for that reason many
establishments have given up the use of ammonia and in its place
use soda or potash.

INTENSIFICATION OF DRY PLATE NEGATIVES.

 From the difficulty of judging when the development has reached
the right point, in consequence of the opacity of the gelatine film,
many negatives are finished before they have acquired sufficient
density in development. As a result of this, much attention has been
given to the discovery and preparation of redeveloping or intensifying
solutions, many formulæ p62 for which have been published; among
the best are the following:

INTENSIFYING SOLUTION. A.
Bichloride of Mercury 120 grains.
Dissolved in hot water 16 oz.
Add to this a strong solution of Iodide of Potassium drop by drop,
until the red precipitate begins to redissolve, then add a few grains of
hyposulphite of Soda, or enough to clear the liquid.
When required for use pour part of this solution into a tray of a
suitable size for the plate, add an equal amount of water, then
immerse the plate and keep the solution in motion. You will
immediately notice a change in the color of the film; take the plate
out, and if it has gained enough strength, wash it and set it up to
dry; if not, give it more.
It is proper to state that this process is more effective if the plate
has been dried after development.
When the plate has been thoroughly fixed and fully washed, it will
intensify by this method, without change of color by transmitted light,
and it will have secured a lovely printing quality rarely seen even in
the most perfect negatives.
This solution may be used until it is exhausted.

ROCHE'S INTENSIFIER. B.
Water 10 oz.
Sulphate of Copper 100 grains.
 Bromide of Potassium 100
When dissolved, this solution is ready for application and can be
used repeatedly.
The negative, after fixing and washing, is immersed in the solution
until it turns white. p63
Now remove it and wash slightly, then immerse it in old ferrous
oxalate developer and allow it to remain until it becomes black
entirely through the film. This method gives fine results and good
printing color.

PRACTICALITIES.
Form the habit of noticing carefully the intensity of the illumination
of the subject on the ground glass of the camera, so as to judge
correctly the time of exposure, as much of the successful working of
dry plates depends on the exposure; for although slightly under-
exposed or much over-exposed plates may by suitable modification of
the developer be saved and print good pictures, yet they will lack the
glow and balance that are characteristic of the most perfect work.
Use a flat camel's-hair brush to remove any dust from the plate,
before putting it in the shield for exposure.
Always keep your developing solution in motion on the plate.
Fix your negative very thoroughly; leave it in the hypo 5 or 6
minutes longer than it is apparently necessary.
Never use hypo after it becomes discolored or turbid and deposits
a sediment.
Develop longer than you think necessary; a slow printing negative
is preferable to a weak one.
Keep your dark room and its contents very clean and free from
dust, and well ventilated.
 RETOUCHING, OR STIPPLING THE NEGATIVE.
No photographic establishment in these days is considered to be
well equipped that does not employ a skillful retoucher and provide
all the apparatus and p64 conveniences for the proper performance of
this very important branch of the art.
It is within the memory of many photographers when this work of
retouching was done on the positive, and some establishments were
compelled to employ a large force of skilled hands to work up and
finish the crude productions of the camera and the printing frame.
After the introduction of the carte de visite portrait, and later the
Imperial card, and the consequent reduction in price, the expense of
this work became such a burthen to photographers that they were
compelled to perfect their mechanical operations to the utmost
extent, and by every means to endeavor to avoid it.
It occurred then to some one of the retouching artists to do this
work upon the negative once for all, and from this beginning the art
of retouching the negative has reached its present high position.
Retouching, like stippling or hatching in miniature, is a work of art,
skill in which is gained only by assiduous practice controlled by good
taste, and it is the special work of the artist and not of the
photographer.
While then it is not considered wise for the photographer to be his
own retoucher any more than it is for every man to be his own
lawyer or doctor, yet every photographer should know enough about
the art and practice of retouching to appreciate good work and to
detect the bad. And for that reason he should make a study of it both
from the artistic and the mechanical points of view.
By such a course of study he will learn to produce that quality of
negative that will require the least amount of improvement at the
hands of the retoucher. p65
 The work of retouching
requires certain appliances to
facilitate it, and its own special
room or atelier. A small well-
ventilated room having a window
with a northern exposure is the
best adapted for this work. This
room should be supplied with the
necessary conveniences, such as
the retouching stand or frame,
which in its earlier form is as
represented in Fig 1, but which,
after a time, became much
improved upon; one form of
Fig. 1. improvement is shown in Fig. 2,
but neither of these forms
became universally used, as many artists fashioned their stands or
frames to suit their individual p66 taste or the requirements of their
room or light. The march of improvement, however, is never delayed
by such a condition of affairs, and was not in this case.
 Fig. 2.

It became necessary on the

part of those interested in the
manufacture of such apparatus
to produce something that
should meet all requirements,
and this has indeed been
accomplished by the production
of the desk represented in Fig 3,
which is called the Novel
Retouching Desk and is suitable
for any size of negative under
25×30; it has a false top which
can be pulled down to darken
the space in front of the ground
glass cut out.
The bed on which the
negative is laid can be set at any
inclination to suit the
convenience of the artist.
Fig. 3. By means of the slide-rest the
negative may be moved up or
down to any desired position.
It is handsomely made in hard wood, and would be quite an
ornament to any retoucher's room. p67
Next in importance would be a varnishing table, a small table or
bench supplied with a gas or oil stove. See Figs. 4 and 5.
Fig. 4. Common Gas Stove.

Fig. 5. Oil Stove.

Also Anthony's retouching varnish and varnish pot. See Fig 6, page
68.
A tin pot as shown in cut, with a filter and funnel in top; rubber
stopper in spout.
The retoucher's room should also be supplied with
 Faber's Siberian leads, full set,
Eagle lead pencils,
Metallic leads,
Camel's hair pencil brushes,
Camel's hair blender,
Guenther Wagner's retouch colors,
India ink,
Water colors, blue, white and red,
A coarse sharpening stone,
Fine emery paper.
p68

The principal work of retouching, the stippling of the face, neck

and shoulders when bare, and the hands, is done with Graphite or
metallic leads, and directly upon the varnished or film surface.
 Fig. 6. Peerless Varnish Pot and Filterer.

The gelatine emulsion film is strong enough to take the pencil

without varnish, and many prefer to retouch and even to print them
before varnishing; but it is pretty generally conceded that gelatine
negatives will receive the pencil better when varnished. p69
Before varnishing a gelatine plate, heat it as much as it will bear,
so as to drive off all moisture from the film, then let it cool to the
proper temperature before applying the varnish.
If the varnish used is too smooth or hard to take the pencil, a
better tooth may be given it by the application delicately, to the parts
to be retouched only, of a solution of rosin in turpentine, which
should be allowed to stand and dry an hour or so before the
retouching is done.
The work of retouching is very trying to the eyes; great pains
should therefore be taken to ascertain the most favorable conditions
of light and distance under which to work.
A negative should be kept at one angle and stationary while being
retouched.
There should be as little lead used on a negative as is consistent
with good work; make every touch tell to some purpose.
Note carefully the way the subject is lighted, and let the effect of
the stippling be in that direction; you will thus be less likely to change
the character or likeness of the face.
The lights and shadows should be caused to blend imperceptibly;
this will give a harmonious effect.
Let the grain or stipple be seen in all parts suitable in fineness to
the size of the head.
The refinements of negative retouching cannot be taught in books,
although much has been written concerning the same.
Try to light your subject and develop your negative so as to get
the utmost roundness or relief; then be careful not to destroy this
modeling, but improve p70 it by retouching, only doing so much work
as to give a soft, clean-looking complexion.
Patient perseverance, close observation and trying again, will bring
a certain degree of success, and while only the few can become first-
class retouchers all may learn to improve their work by this great
help to photography.

SILVER PRINTING.
It was long ago said that silver printing was doomed, that in a few
more years it would be among the things that were.
 Other processes have been introduced, have had their little day
and have passed off, some of them of much merit, and, it was
supposed, possessed of every element of permanency, but they were
found wanting in some very important respect, and so not being able
to hold their ground they passed into the limbo of neglected things.
Silver printing, however, still maintains the field against all comers,
and does so by its intrinsic merits alone. No other process has
equalled it in ease of production and in beauty of finish, and it is
doubtful if any process has results of greater permanence, taking the
best products of the process as samples.
The one defect of the process is the possibility of its products
fading, but we must not judge of a process by its poorest examples,
but rather by its best; if such should be the decision, there are at this
day thousands of silver prints that are co-existent with the process
and remain as perfect as it is possible for anything on paper to be
after such a lapse of time.
A process that is so simple and easy that it can be p71 acquired in a
few hours, is predestined to careless working, slovenly manipulation,
and to endless scamping at the hands of careless workers. What
wonder then if prints fade which were only half made.
Suffice it that the process, intelligently and conscientiously
worked, would never suffer from such an imputation.

THE PRINTING ROOM.

All professional photographers practice the silver printing process
with more or less of success, as they bestow care and intelligence in
the working and are zealous for the reputation of their work.
The proper handling of a negative by the printer requires as much
study and skill on his part as does the posing and lighting of the
subject in the studio; and the printing room should be as replete with
every convenience for forwarding and improving the operations
therein conducted, as any other department of the establishment.
 This room is generally situated at the top of the house, and
wherever situated, should be so arranged that the east, south and
west sides may be opened to the sun, that his direct rays may be
utilized from the earliest to the latest working hours. Part of the room
should be partitioned off in such a manner, and illuminated in such a
moderate degree, that the sensitized paper may not suffer from
excess of light while the frames are being filled and the changes
made.
The printing room outfit consists of the various sizes of printing
frames, cardboard, tissue paper and cotton-wool for vignetting, and
knives, long-blade scissors, and glass forms for cutting the paper to
various sizes, p72 light tight drawers to hold the prepared paper,
shelves, and racks for the negatives.

Mucilage or paste and brushes.

Flat brushes for removing dust from negatives.
India ink and Opaque, with pencil brushes and lead pencils
for spotting negatives.

THE SILVERING ROOM.

The silvering room may be also used for other manipulations in
the process, such as fuming the paper and washing the prints, and
should be furnished with all the conveniences for such work, which
consist of

Nitrate of Silver.
Albumenized and plain paper.
A flat dish to hold the silver solution, of the size of the
largest paper to be floated.
A fuming box.
A bottle of concentrated Ammonia.
A gas or oil stove for drying the paper.
 A dozen or more of Lockwood's patent photo clips. Bi-
Carbonate of Soda.
An Argento meter.
A 16 oz. Graduate.
A clock.
One agate evaporating dish, and 2 or 3 wide-mouth glass
bottles, that will hold the full quantity of silver solution in
use.

THE SILVER SOLUTION FOR PAPER.

A plain silver solution, of a strength of 30 grains of nitrate of silver
to the ounce of water in hot weather, increasing to 60 grains in the
coldest, with enough bi-carbonate of soda added from time to time to
keep a small sediment of carbonate of silver in the bottom of p73 the
bottle or dish, is all that can be desired as a silvering solution for
albumen paper. It only needs to be kept up to the proper strength by
the addition of silver nitrate and water, when required; and when it
shows signs of discoloration a little more soda and an hour's sunlight
will keep it in good order the year round.

TO SILVER THE PAPER.

Albumen paper is sensitized by being floated for a time on the
surface of this silver solution, albumen side down. For this purpose it
is necessary that the solution should be held in a flat dish, like a tray,
with sides from 1½ to 2 inches high, and of a size suitable to the size
of the sheet of paper to be floated; if a full sheet, then the dish
should be 19×24, and may be of porcelain, agate ware, hard rubber,
or any other kind that will resist the corrosive action of the silver
solution.
Pour the silver solution into the dish; there should be enough to
give a depth of at least half an inch. If there is any scum or dust on
the surface after the air bubbles have subsided, take a long strip of
tissue paper and draw the edge over the surface of the solution,
which will remove the scum or any floating dust or air bells that may
remain.
Place your paper on a bench or table, albumen up, and with a
large tuft of cotton rub the surface, using a light friction over the
whole; then turn the sheet of paper over and take between the
thumb and first finger of each hand opposite corners diagonally
across the paper, holding the corners up and near together; let the
sag of the paper sink one end first, until it touches the solution,
which can be plainly seen; then let the hands fall until the other end
has reached the p74 surface of the silver; then lower the corners
gradually, until the paper lies flat on the solution, care being taken
that none of the silver solution runs over the edge of the paper to the
back.
The corners of the paper may now be gently raised to see if any
air bubbles have attached themselves to the surface of the paper; if
so, remove them and let the paper fall again to the solution. It will
now be seen that the edges of the sides of the paper will begin to
curl back, as if repelled from the solution. This may be permitted to a
certain extent, but not so far as that the wet surface shall curl over
against the back of the sheet. After a short time this action ceases
and the rolled edges unroll and fall again upon the solution, which,
when they have reached and lie flat, may be taken as an indication
that the paper is sufficiently silvered, when it should be removed.
With a small pointed stick raise the left corner farthest from you from
the solution, take it between the thumb and first finger of the left
hand and raise it very slowly, until the right hand far corner is off the
solution. Take that corner between the right hand thumb and finger
and continue to raise the paper, still very slowly, until it is clear of the
solution; then hang it up to dry in a moderately dark place, or better
still, place it face down, on a sheet of clean blotting paper, put
another sheet over it and on that the next sheet that is silvered, and
so on alternate sheets, until you have floated as many sheets as may
be needed.
 When this is done turn the papers over, bringing the sheet first
silvered to the top, which on removing the blotting paper will be
found surface dry at least, and may now be completely dried by
artificial heat, or by p75 hanging it up two sheets together, back to
back, on lines with spring clips, until they dry spontaneously, when
they are ready to be fumed.

FUMING THE PAPER.

The fuming box is usually a light-tight box with two
compartments; the upper part has a door and should be sufficiently
large to hold the amount of paper necessary for a full day's work,
without crowding the sheets together or preventing a free circulation
of the fumes between their surfaces.
The paper is taken from the lines, each two sheets back to back,
the corners fastened with clips and set up on end in this
compartment or hung on lines, as the case may be. When all the
paper is in, shut and fasten the door and into the bottom
compartment (which is usually a shallow drawer, and separated from
the upper only by a lattice work of wood) place a saucer containing
an ounce or two of strong liquor of ammonia, push the drawer in,
thus closing the compartment, and leave for fifteen or twenty
minutes, or until you are ready to use the paper.
When you remove the paper from the fuming box do not expose it
to strong light until you have it in the printing frames, nor after that,
until you have toned and fixed it. Take the spent ammonia from the
box and pour it into a bottle; it is useful for other purposes. Many
printers cut the paper to size before printing, others simply tear each
sheet into halves, quarters, eighths or twelfths, and print them thus,
trimming them afterwards, some before toning, others after they are
finished and before wetting them for mounting. The most economical
method is to trim before printing; p76 all the trimmings should be
carefully saved for the silver they contain.

TONING THE PRINTS.

 A toning solution is composed of chloride of gold in water with
enough carbonate of soda to make it slightly alkaline, and a pinch of
common salt. The action of toning is merely the deposition upon the
metallic silver in the print of a certain minute quantity of metallic gold
from the solution, which deposit causes the change of color that may
be seen while the toning is taking place.
The normal toning solution is varied endlessly by the addition of
acetates, chlorides and nitrates of various metals and salts and of
borax and chalk and such, the object being to produce unusual
tones, such as sepia, blue, blue-grey, brown and black, but the
richest and most agreeable tones may be produced as well by the
simplest means.

THE TONING SOLUTION.

Water 1 quart.

Into this put half as many grains of pure chloride of gold as you
have sheets of paper to tone; after adding the gold make the solution
alkaline by adding enough of a strong solution of a carbonate or
bicarbonate of soda to turn red litmus paper just blue; then add a
pinch of salt, and the toning solution is ready for use. When half your
prints are toned, or the toning proceeds too slowly, add more gold,
allowing not over 1 grain of gold for each sheet of paper. When all
the prints are toned, put the toning solution into a bottle to keep for
the next time, when it will probably work better than at first; p77 use
it until it becomes badly discolored, then put it into a large bottle to
save the waste gold.
The toning must be done in a flat dish that is as large as the
largest print that requires to be toned.
Your toning solution being ready, the prints are now to be
prepared for toning by being deprived of all excess of silver and free
chloride not acted on by the light; in order to do this, put them one
by one in a large dish full of water, in this dish keep them in motion
for some little time, then prepare another dish containing the same
amount of water slightly acidified with acetic acid, and remove the
prints into this one by one as before, keeping them in motion; they
will at once commence to turn red, and when they are so they may
be taken out, the waters of these two washings must be saved for
the silver contained therein.
The prints require two more waters after the acid, when they are
ready for toning.
Immerse half a dozen or more of the prints, one at a time, in the
toning solution and watch the change of color.
They may at first become a little redder from the action of the salt
in the solution, but they will soon change to a brown and then to a
bluish or grey brown. And when the faces have become as clear and
white as to suit as finished pictures, remove them to fresh water and
proceed with another batch until all are done.

FIXING THE PRINTS.

Hyposulphite of soda, about 1 lb. to a gallon of water, constitutes
the fixing solution; its function is to dissolve all the chloride of silver
that may remain in the prints after their previous washing and
toning. p78
The prints should be immersed one at a time, and kept in motion
so that the solution may have an equal action. When there are a
large number of prints the dish containing the fixing solution should
be correspondingly large, so that the prints may not be massed
together but kept separate and free for the more perfect action of
the fixing solution; they should be kept in motion until the fixing is
completed, which will be in about 15 or 20 minutes. Some printers
ascertain that the action is complete by holding the print up and
examining it by transmitted light; if it looks mottled and uneven it is
not fixed, but if you can see the fibre of the paper and all is clear,
then remove the prints from the hypo into a dish containing brine or
common salt and water, one by one as before, and keep them in
motion in this brine until all are well saturated with it; then let fresh
water run into the dish, which will gradually change the brine to clear
water.
The prints are put into the brine, and this gradually changed to
clear water to prevent blisters or a separation of the albumen from
the paper in the form of blisters. The water may be allowed to run
into the dish, the prints being kept in motion, and in the course of an
hour they will be sufficiently washed; they may then be taken out
and placed between the sheets of blotting paper to remain until next
morning, when they can be overlooked, the blemished prints thrown
out, and the perfect may be mounted.
The fixing solution should not be used a second time, but should
be thrown into a large barrel with the washings, that at some future
time the silver may be recovered; when the barrel becomes full, a
small amount of saturated solution of protosulphite of iron p79 may be
thrown in, the water stirred well, and when the sediment has settled
to the bottom the clear liquid may be drawn off and the barrel is
ready to receive the next washings of hypo.

MOUNTING THE PRINTS.

The selected prints are again placed in water and permitted to
remain until they become saturated; in the meantime some starch
paste should be prepared. The prints are then removed from the
water and placed on a sheet of glass, face down; when all are thus
placed, squeeze out all the water and they are ready to mount. Now
with a wide bristle brush spread the paste evenly on the back of the
print, carefully removing any lumps or dust or fibre of any kind, then
insert under the edge of the print the point of a knife-blade, raise it
until you can take it in the fingers, then place it pasted side down on
the face of your mount, adjust it evenly and lay it on the table; now
cover it with a piece of clean white paper, and with the palm of the
hand rub it down until all parts are in contact and all air is expelled
from between the print and mount. As the prints are mounted they
should be placed in rows on clean white or blotting paper, one layer
over the other; on the top place a clean sheet of paper and over that
a board of the proper size with a weight on it to press the mounted
picture flat; in the course of an hour they will be dry enough to spot.

SPOTTING THE PRINTS.

With a fine pencil brush and Indian ink, go over the prints and
carefully touch up all the white spots and other blemishes that may
be found on the surface, so that they harmonize in color with the
surrounding p80 parts. When this is done the pictures may be
lubricated for burnishing.

LUBRICATING THE PRINTS.

Make a pad of Canton flannel, get a piece of white castile soap,
rub the pad on the soap until it is well covered, then with this soaped
pad rub the surface of each print separately, carefully covering every
part of the surface; keep the pad well soaped by rubbing it on the
soap after every 2 or 3 prints. While this is being done the burnisher
may be heated; when the tool is hot enough to hiss when touched by
a wet finger, proceed to burnish. Do not use much pressure; too
great pressure will degrade the whites of the picture.
When they are all burnished go over them again with a moderately
soft lead pencil, touching up and blending any streaks or spots that
may be found; if any streaks or dirty lines are discovered in the
direction of the draw of the burnisher, they may be removed by light
friction with a piece of Canton flannel moistened with alcohol. This
completes the photograph.

THE PHOTOGRAPHIC STUDIO.

Mr. Ernest Lacan, a prominent artist of Paris, France, about ten
years ago, wrote for the Philadelphia Photographer an account of
some of the prominent studios of that great city, from which I take a
description of the studio and establishment of the celebrated
Reutlinger.
 This establishment comprises the fifth and sixth stories of a fine
house on the Boulevard Montmartre.
A handsome and wide stairway leads to the studio. The first thing
that strikes you on entering the ante-chamber, p81 which is
transformed into an office, is the lowness of the ceiling and the want
of light. On the right is a room, larger and better lighted, for the sale
of choice specimens of his work. On the left are the exhibition and
waiting rooms, which are of medium size and whose principal
ornaments are the framed pictures, which cover the walls. A small
door leads to the skylight, of which the diagram at the head of this
article is a correct view as taken from a photograph.
 THE PHOTOGRAPHIC STUDIO.

The view is taken from the door at which you enter. This gallery is
formed of two mansards, which have been united by removing the
partition; is 39 feet p82 long by 13 in breadth; its height to the top of
the upper sash is about 16 feet; the light comes from the north. It is
by means of an ingenious combination of white and blue shades,
that the artist succeeds in obtaining the charming effects so much
admired in his productions. At the end of the gallery is a small room
for ladies. The door which is seen on the left leads to the laboratory,
which is divided into three small apartments. The first is used for
cleaning plates, the second for their preparation, and the third for
developing negatives.

This is a diagram of the studio of the famous p83 Loescher &

Petsch, of Berlin, who became so well known, some years ago,
through the style of picture called "Berlin Heads," which were among
the first samples of fine photographs from retouched negatives
brought to this country, and which certainly created a sensation.
 The shades are arranged so
as to show how some of the
most charming effects of
illumination are produced. The
room is filled with diffused light,
with a ray of direct light falling
so as to produce a clear high
light on the prominences of the
head of the sitter.
The next diagram is one of
the Biglow studios. Mr. Biglow is
the author of a book on lighting
and posing, which had a large
sale, and is a valuable book for
FIG. 1. positionists.
These three views represent
forms of skylight and side-light with north exposure, by which all the
finest effects possible are obtained, but other forms of exposure are
capable of being utilized with fine success.
The studio of Sarony, of New York, is lighted with a top light
similar to the top light of the Biglow studio, and without any side
light at all.
A prominent artist of Brooklyn produces very fine p84 work under
an east light, or rather a light a little south of east, which to the
ordinary photographer would be considered a very difficult light to
work. Good effects can be produced, however, under any form of
light by the use of shades, screens and reflectors, so that no
photographer need regret a favorable location, because unable to
have a northern exposure for his light.
The studio should be furnished with every requisite for the
production of the finest work, such as

Lenses and Camera boxes,

 Camera stands,
Pneumatic shutters for the lenses,
Scenic grounds; interior and exterior accessories, such as
balustrades, rocks, grass-mats, flowers, tables, chairs,
draperies, rugs, etc.; head-rests, screens, reflectors, and all
such articles as can be advantageously introduced in a
picture to improve and embellish.

All these things are of importance, but should be used with taste
and judgment. The photograph should never be a picture of a piece
of furniture, with a figure thrown in, but rather the accessories
should be used only to improve the figure and make it more
prominent by increasing the perspective, when possible or allowable.

CAMERA LENSES.

THE OPTICS OF PHOTOGRAPHY.

A solar beam of light is a bundle of rays; a ray being the smallest
portion of light which can emanate from a luminous body.
Each of these rays possesses distinctive characters, both as
regards their chemical functions and colors.
Sir Isaac Newton proved that the white light emitted p85 by the
sun is not so simple as it appears, but is composed of vivid colors, as
shown by his beautiful experiment, the Analysis of Light, which is
exemplified by the use of a glass prism. (See fig. 1.)
The ray of light A, E, being admitted into a darkened room
through a hole A, in the shutter, would fall upon the wall at E. As
soon as the prism, B, C, is placed in the path of the sunbeam so as
to allow it to fall on one of its angles B, the ray will be refracted, or
bent out of its course so as to pass through the prism (as in the line
D) and not in the same line, A, E, that it would have done had the
prism not been interposed.
 Another effect also takes
place; the ray of white light is
decomposed into its component
colors, and if you stand at a
short distance from the prism,
you will see that these colors are
spread out in a triangular form,
the base of which is on the wall
and the apex at the angle C of
the prism. Remove the prism
and it is seen that the splendid
display of colors upon the wall
has disappeared, and a round
spot of white light, E, is seen
below the place occupied by the
FIG. 1. spectrum.
The colored image on the
wall is called the prismatic or solar spectrum, which, according to Sir
Isaac Newton, is composed of seven different colors. The color at
the lowest portion of the image is red and the one at the other end
is violet, the intermediate parts being occupied by five other colors,
and the whole arranged p86 according to the table below, the
proportion of each color having been measured by Fraunhofer with
the greatest care, with the results placed opposite to each,
corresponding with the 360 degrees of a circle, the red ray being the
least and the violet the most refracted of this chromatic image.
TOP
Violet 109
Indigo 47
Blue 48
Green 46
Yellow 27
Orange 27
 Red 56
360
The sunbeam, the ray of white light, contains powers within it of
which the earlier philosophers had but a faint idea, besides its
accompanying heat. There is a principle associated intimately with it,
which has the power of decomposing and of determining the
decomposition of chemical compounds.
This principle is "Actinism" and is as perfectly distinct in the
nature of its properties from light, as light is from the principle of
heat, with which it is also closely connected.
Actinism may then be considered as the fundamental principle, on
which photography is based, and its power is exerted in forming the
image on the sensitized plate in the Camera, as well as subsequently
in forming or causing the impression on the sensitized paper
exposed to the light beneath the negative.
In this connection we have only to consider the p87 application of
this Actinism through the medium of a combination of lenses to form
an image on a sensitized plate in the Camera.
It has been shown that when light passes through a prism of
glass its colors are separated; this separation is caused by the
unequal refrangibility of the different colored rays, the violet being
the most and the red the least refrangible of the seven rays.
A ray of light passing through a vacuum progresses in a perfectly
straight line, but all matter, however attenuated it may be, has the
property of refracting or bending the ray of light.
The refractive power of some substances is immense, while that
of others is very trifling. The mode of the refraction depends on the
comparative density or rarity of the respective media. If the medium
which the rays enter be denser, they move through it in a direction
nearer to the perpendicular drawn to its surface; on the contrary,
when light passes out of a denser into a rarer medium, it moves in a
direction farther from the perpendicular. This refraction is greater or
less, that is the rays are more or less bent or turned from their
straight course, as the second medium through which they pass is
more or less dense than the first.
We next study the utilization of this power of refraction in the
manufacture of lenses to overcome the unequal refrangibility of the
colored rays of light.
A lens for use in photography is made of glass as pure and as
colorless as can be procured, and is ground into such a form as to
collect or disperse the rays of light which pass through it. Lenses are
of different shapes, and thence receive different names. p88

FIG. 2 E.
1. Double convex. 2. Plano-convex. 3. Concavo-convex.
4. Double concave. 5. Plano-concave. 6. Meniscus.

The figures 1 to 6 represent sections of the variously shaped

lenses which are combined for use in photography.
The design in forming lenses is to procure a medium through
which the rays of light from any object may pass and converge to a
corresponding point beyond; the manner in which the rays proceed
through the lens, and then centre in a focal point, will depend on the
form of the lens, its capacity for refraction and the distance of the
object.
The double convex lens may be viewed as a portion cut out of the
side of a sphere. Here, as in all cases of convexity, the focus of the
parallel rays passing through the lens is at the centre of the sphere.
(See fig. 3.)
 A plano-convex has only half of the
refractive power of the double
convex; the parallel rays, falling on
the convex side of the lens, would
converge at the distance of the whole
diameter of the sphere. Thus the focal
point at which the rays of light
converge is always regulated by the
degree of curvature of the lens. Thus
the double convex lens has the
FIG. 3. greatest power of converging the rays
of light; the plano-convex has only
half the power of the former. Both these lenses have also the power
p89 of magnifying the image of an object seen through them in the
same proportion.
The double concave and the plano-concave have the power of
dispersing the rays of light and of diminishing the image of an object
seen through them in the same proportion.
The meniscus lens has but a very
slightly dispersive power, and the
concavo-convex merely separates the
parallel rays to the thickness of the
lens and sends them on parallel as
they entered.
All these lenses, having something
of the prism in their shape, have the
power to a greater or lesser extent of FIG. 4.
decomposing the light that passes
through them. This is called chromatic
aberration, because the colored rays do not all converge to the same
focus; thus the image seen through them is surrounded by a fringe
or border of color.
Single lenses lack the power of producing a straight image of a
straight object; the image will have the curve of the lens through
Welcome to our website – the perfect destination for book lovers and
knowledge seekers. We believe that every book holds a new world,
offering opportunities for learning, discovery, and personal growth.
That’s why we are dedicated to bringing you a diverse collection of
books, ranging from classic literature and specialized publications to
self-development guides and children's books.

More than just a book-buying platform, we strive to be a bridge

connecting you with timeless cultural and intellectual values. With an
elegant, user-friendly interface and a smart search system, you can
quickly find the books that best suit your interests. Additionally,
our special promotions and home delivery services help you save time
and fully enjoy the joy of reading.

Join us on a journey of knowledge exploration, passion nurturing, and

personal growth every day!

ebookbell.com