Netto Theory Substitutions

LIBRARY
UNIVERSITY OF CALIFORNIA.
Deceived
Accessions No. *T<
8?
Ctetf Aô.
THEORY OF SUBSTITUTIONS
AND
ITS
APPLICATIONS TO ALGEBRA,
BY DR.
EUGEN NETTO,
PROFESSOR OF MATHEMATICS IN THE UNIVEHSITY OF GIESSEX.
Revised by the Author and Translated with his Permission

BY
F.
N.
COLE.
PH. D.,
Assistant Frofessor of Mathematios in the University of Michicjan.
ANN ARBOR,
1892.
MICH,:
THE REGISTER PUBLISHING COMPANY.

tEbe flnlanb press.
6 o
PREFACE.
The presentation of the Theory of Substitutions here given differs in several essential features from that which has heretofore been custom
It will accordingly be proper in this place to state in brief the ary. guiding principles adopted in the present work. It is unquestionable that the sphere of application of an Algorithm
extended by eliminating from its fundamental principles and its general structure all matters and suppositions not absolutely essential to its nature, and that through the general character of the objects with
is
employment in the most varied That the theory of the construction of groups admits of such a treatment is a guarantee for its far-reaching impor tance and for its future. If, on the other hand, it is a question of the application of an aux iliary method to a definitely prescribed and limited problem, the elab oration of the method will also have to take into account only this one purpose. The exclusion of all superfluous elements and the increased usefulness of the method is a sufficient compensation for the
it
which
deals, the possibility of its

is
directions
secured.
lacking, but not defective, generality. a smaller sphere of action.
A greater efficiency is attained in
The following treatment is calculated solely to introduce in an elementary manner an important auxiliary method for algebraic inves
tigations.
is
By
the employment of integral functions from the outset,
it
not only possible to give to the Theory of Substitutions, this operat ing with operations, a concrete and readily comprehended foundation, but also in many cases to simplify the demonstrations, to give the various conceptions which arise a precise form, to define sharply the principal question, and what does not appear to be least important to
limit the extent of the work.
The two comprehensive treatises on the Theory of Substitutions which have thus far appeared are those of J. A. Serret and of C. Jordan.
The fourth
to this subject.
section of the "Algebre Superieure of Serret is devoted The radical difference of the methods involved here
"
work
and there hardly permitted an employment of this highly deserving for our purposes. Otherwise with the more extensive work of
Jordan, the "Trait6 des substitutions et des Equations algebriques." Not only the new fundamental ideas were taken from this book, but it
is
proper to mention expressly here that
many
of
its
proofs and pro^
PREFACE.
in the cesses of thought also permitted of being satisfactorily employed difference of the general treat essential the of in work spite present " " ment. The investigations of Jordan not contained in the Traite which have been consulted are cited in the appropriate places. " But while many single particulars are traceable to this "Traits his to and to these investigations, nevertheless, the author is indebted honored teacher, L. Kronecker, for the ideas which lie at the foundation
of his entire work.

benefit
He has striven to employ to best advantage the which he has derived from the lectures and from the study of the works of this scholarly man, and from the inspiring personal inter course with him; and he hopes that traces of this influence may appear in many places in his work. One thing he regrets: that the recent im
portant publication of Kronecker, "Grundzuge einer arithmetischen Theorie der algebraischen Grossen," appeared too late for him to derive from it the benefit which he would have wished for himself and his
readers.
The plan of the present book is as follows: In the first part the leading principles of the theory of substitutions are deduced with constant regard to the theory of the integral func tions; the analytical treatment retires almost wholly to the background, being employed only at a late stage in reference to the groups of solvable
.
equations. In the second part, after the establishment of a few fundamental principles, the equations of the second, third and fourth degrees, the Abelian and the Galois equations are discussed as examples. After this
follows a chapter devoted to an arithmetical discussion the necessity of which is there explained. Finally the more general, but still elementary questions with regard to solvable equations are examined.
STRASSBURG,
1880.
the preceding I have now to add that the present translation from the German edition in many important particulars. Many new investigations have been added. Others, formerly included, which have shown themselves to be of inferior importance, have been omitted. Entire chapters have been rearranged and demonstrations simplified. In short, the whole material which has accumulated in the course of time since the first appearance of the book is now turned to account. In conclusion the author desires to express his warmest thanks to Mr. F. N. Cole who has disinterestedly assumed the task of translation and performed it with care and skill.
To
differs
EUGEN KETTO.
GIESSEN, 1892.
TRANSLATOR
NOTE.
The translator has confined himself almost exclusively to the function of rendering the German into respectable English. My thanks are especially due to The Register Publishing Company for their gener ous assumption of the expense of publication and to Mr. C. N. Jones, of Milwaukee, for valuable assistance while the book was passing through the press.
F. N.
COLE.
Ann
Arbor, February
27, 1892.
TABLE OF CONTENTS.
PART
I.
THEORY or SUBSTITUTIONS AND OF INTEGRAL FUNCTIONS.
CHAPTER
SYMMETRIC OR SINGLE-VALUED FUNCTIONS
t
I.
ALTERNATING AND TWO-VAL-
UED FUNCTIONS.
.
1-3
4.
5-10.
11.
Symmetric and single-valued functions. Elementary symmetric functions. Treatment of the symmetric functions.
Discriminants.
12
Euler
formula.
13. 14. 15.
Two-valued functions; substitutions. Decomposition of substitutions into transpositions.

Alternating functions.
.
16-20
Treatment and group of the two-valued functions.
CHAPTER
22 24
25
.
II.
MULTIPLE-VALUED FUNCTIONS AND GROUPS OF SUBSTITUTIONS.

Notation for substitutions. Their number. Their applications to functions. Products of substitutions.
26-27
28
.
29-32
34. 35.
Groups of substitutions. Correlation of function and group. Symmetric group.

Alternating group. Construction of simple groups. Group of order pf.
36-38
39-40.
CHAPTER
III.
THE DIFFERENT VALUES OF A MULTIPLE-VALUED FUNCTION AND THEIR

ALGEBRAIC RELATION TO ONE ANOTHER.
41-44
.
Relation of the order of a group to the of the corresponding function.
number
of values
viii
CONTENTS.
45
46-47.
48-50.
51
.
Groups belonging
Transformation.
to the different values of a function.
The Cauchy-Sylow Theorem.

Distribution of the elements in the cycles of a group. Substitutions which belong to all values of a function.
p- valued function. Discriminants of the functions of a group.
52.
53.
55.
Equation for a
56-59.
Multiple-valued valued.
functions, powers of
which are
single-
CHAPTEK
TRANSITIVITY AND PRIMITIVITY.
IV.
SIMPLE AND COMPOUND GROUPS.
ISOMORPHISM.
60-61
.
Simple transitivity.
Multiple transitivity.
Primitivity and non-primitivity.
62-63.
64
65-67
68.
Xon-primitive groups.
Transitive properties of groups.
69-71.
72-73.
Commutative substitutions;
Isomorphism.
Substitutions which affect
self -con jugate
subgroups.
74-70
all
the elements.
77-80.
Limits of transitivity.
Transitivity of primitive groups.
81-85.
86.
87.
88-89.
91
.
Quotient groups. Series of composition. Constant character of the factors of composition. Construction of compound groups.
92.
The alternating group

Groups of order p a
.
is
simple.
93
94
95.
Principal series of composition. The factors of composition equal prime numbers.
96.
.iT
Isomorphism.
.v
.
The degree and order
equal.
99-101
Construction of isomorphic groups.
CHAPTER
ALGEBRAIC RELATIONS BETWEEN
GROUP.
103-105.
V.
TO
FUNCTIONS BELONGING
THE SAME
FAMILIES OF MULTIPLE-VALUED FUNCTIONS.
106.
107.
Functions belonging to the same group can be rationally expressed one in terms of another. Families; conjugate families. Subordinate families.
CONTENTS.
108-109.
IX
Expression of the
subordinate.
principal
functions in terms of the
110.
111.
The resulting equation binomial. Functions of the family with non-vanishing discriminant.
CHAPTER
112. 113.
VI.
THE NUMBER OF THE VALUES OF INTEGRAL FUNCTIONS.

Special cases. in the
Change
.
form of the question.

is less
114-115
116.
Functions whose number of values

Intransitive
than their degree.
and non-primitive groups.
117-121.
122-127.
Groups with substitutions of four elements. General theorem of C.Jordan.
CHAPTER
128
.
VII.
CERTAIN SPECIAL CLASSES OF GROUPS.

Preliminary theorem. with r = n --p. Groups Groups 12 with r = n=p.
.
129. 130. 131.
Cyclical groups.
q.
132-135.
136.
Groups 12 with r = n=p 2 Groups which leave, at the most, one element unchanged. Metacyclic and semi-metacyclic groups. Linear fractional substitutions. Group of the modular
equations.
137-139.
Groups of commutative substitutions.
CHAPTER
140
141
.
VIII.
ANALYTICAL REPRESENTATION OF SUBSTITUTIONS.
THE LINEAR GROUP.
The
analytical representation.
143
Condition for the defining function. Arithmetic substitutions.
144. 145.
Geometric substitutions.
Condition
among
the constants of a -geometric substitution.
146-147.
Order of the linear group.
PAKT
II.
APPLICATION OF THE THEORY OF SUBSTITUTIONS TO THE ALGEBRAIC

EQUATIONS.
CHAPTER
IX.
THE EQUATIONS OF THE SECOND, THIRD AND FOURTH DEGREES. OF AN EQUATION. RESOLVENTS.
148.
GROUP
The equations
of the second degree,
p$
CONTENTS.
1 J
..
150.
1
">-
The equations of the third degree. The equations of the fourth degree. The general problem formulated. Galois
Affect equations.
resolvents.
153-154.
156.
157.
Group of an equation.
Fundamental theorems on the group of an equation. Group of the Galois resolvent equation.
General resolvents.
158-159.
CHAPTER
f
X.
THE CYCLOTOMIC EQUATIONS.

101
.
162.
Definition and irreducibility. Solution of cyclic equations.
163.
164-165.
10<>.
Investigation of the operations involved. Special resolvents. Construction of regular polygons by ruler
and compass.
167.
168.
The regular pentagon. The regular heptadecagon.

Decomposition of the cyclic polynomial.
169-170.
CHAPTEE
171-172.
1
XI.
THE ABELIAN EQUATIONS.
73
One root of an equation a rational function of another. Construction of a resolvent.
174-175.
176. 177.
Employment of special resolvents Second method of solution.

Examples. Abelian equations.
Solution of the simplest Abelian equations. for the solution.
178-180.
181.
182. 183.
184-186
187
.
Their solvability. Their group. Solution of the Abelian equations; first method. Second method.
Analytical representation of the groups of primitive Abelian equations.
188-189.
Examples.
CHAPTER
190-193
194.
XII.
EQUATIONS WITH RATIONAL RELATIONS BETWEEN THREE ROOTS.
Groups analogous to the Abelian groups. Equations all the roots of which are rational functions of
196.
197.
two among them. Their group in the case n The binomial equations.
=p.
CONTENTS.
199.
XI
200-201
202.
Triad equations. Constructions of
compound
triad equations.
203-205
206.
Group Group
of the triad equation for of the triad equation for
n n
7.
Hessian equation of the ninth degree.
CHAPTEE
207-209.
XIII.
THE ALGEBRAIC SOLUTION OF EQUATIONS.

Rational domain. Algebraic functions. Preliminary theorem. Roots of solvable equations.
Impossibility of the solution of general equations of higher
degrees.
218.
219.
210-211
212-216
217.
Representation of the roots of a solvable equation.
220-221
The equation which is satisfied by any algebraic expression. Changes of the roots of unity which occur in the expres
sions for the roots.
222-224
Solvable equations of prime degree.
CHAPTER
226. 227.
228.
XIY.
THE GKOUP OF AN ALGEBEAIC EQUATION.

Definition of the group.
Its transitivity. Its primitivity.
229.
230.
Galois resolvents of general and special equations. Composition of the group.
231
Resolvents.
232-231.
235.
Reduction of the solution of a compound equation.
236-238.
Decomposition of the equation into rational factors. Adjunction of the roots of a second equation.
CHAPTER XV.
ALGEBEAICALLY SOLVABLE EQUATIONS.
239-241
242 243
. .
Criteria for solvability.
Applications.
Abel
244.
245
246.
248-249
s theorem on the decomposition of solvable equations. k Equations of degree p their group. Solvable equations of degree p. Solvable equations of degree p 2 Expression of all the roots in terms of a certain number of them.
; .
ERRATA.
p.
7,
footnote, for transformatione read transmutatione.
p. p. p. p.
15, line 10,
read t
16, line 5, read
= S A/H ^ ^ = 2S
.
\/ J.
28, line 9, for
<p
read
0.
v -.
29, line 12, for V 29, line 9,
read
p.
p.
from bottom, for ^ read
V.
(a? 2 cc 4 ),
31, line 8, read
G=
[1,
(x,x^ (x 3 a? 4 ), (a?^)
.
(a^)
(a? 2
aj 3
)].
p.
p. p. p. p.
41, line 7, for
read
pf
52, line 13, read

52, line 5,
from bottom, for
read
89, line 2, for not 93, line 9, for a

94, line 2, for
more read
read a primitive group H. k. q-\-2 read n q 6 3, read >,.
group
p.
p.
98, line 19, for 3 a

L>.,
p. 101, line 3, for
read
&V
p. 103, line 14, read [1, (z&)]. p. 125, THEOREM XI, read: Jn o?-der ^/ia^ there
may be ap/o-valued function y a prime power y* of which shall have p

values, etc.
p. 159, lines 10,
11 from bottom, read:
Since
,
belongs to r
at
least
one of the exponents // ?,..., belongs to i\ must be prime to r

,
the s of which
p. 106, line 4,
read r 2
Cj
/2 r 2
3
.
r2
p. 174, line 2, for

p.
read 2c
210, foot note, for No. XI,
etc.,
read 478-507, edition of Sylow

a).
and
p. 224, line 2
Lie.
p. 219, line 10, for
^cos a) read (^(cos
from bottom, read: which leaves two elements with

from bottom, read
successive indices unchanged.

p. 248, line 3
+ +
= Fa +
(Fa +
.).
PART
I.
THEORY OF SUBSTITUTIONS AND OF THE INTEGRAL

FUNCTIONS.
CHAPTER
I.
SYMMETRIC OR SINGLE- VALUED FUNCTIONS. ALTERNA TING AND TWO- VALUED FUNCTIONS.
1.
ments
a?!,
In the present investigations we have to deal with n ele oc.2 a? n which are to be regarded throughout as entirely
j
.
independent quantities, unless the contrary

is
is expressly stated. It easy to construct integral functions of these elements which are unchanged in form when the x\s are permuted or interchanged in
any way.
For example the following functions are
of this kind
etc.
Such functions are
called
symmetric functions.
We
,
,
confine
ourselves, unless otherwise noted, to the case of integral functions. If the a.Vs be put equal to an any arbitrary a, a 2
quantities,
. .
.
so that
a 15 x,
=a
2,
a?,,
it
is
clear that the symmetric
functions of the X A S will be unchanged not only in form, but also in value by any change in the order of assignment of the values a A to the XA S. Such a reassignment may be denoted by
Xi
=
.
a (l
x2
a,-
...
xn
a,- n
where the a^
a,-
denote the same quantities a ly a 2
... in
any
one of the possible n orders.

!
Conversely,
^(xj,
it
can
of
be
shown
that
every integral function,

1?
x2
a?,,)
n independent quantities x
x2
... #,
which
THEORY OF SUBSTITUTIONS.
is
unchanged
in value
.r/s, is
by
values of the
also
the possible permutations of arbitrary unchanged in form by these permutations.

all
Theorem
2.
I.
Every
single -valued
jmlrjirmlcnt elements x^
x>, ...
tl
is
integral function of n symmetric in these elements.

of
this
The reasoning on which the proof

it
theorem
is
based
will
be of frequent application in the following treatment.

here in
full detail.
It seems, therefore, desirable to present
(A).
(1)
If in the integral function

A
= 71
the coefficients aA are equal to zero, then / (x) vanishes identi Conversely, cally, i. e.. f(x) is equal to zero for every value of x.
all
if
/ (.
vanishes for every value of x, then
all
the coefficients a\ are
equal
to zero.
if
that for every real

r,,
,
not identically zero, then there is a value ^ such of which the absolute value \x\ is greater than the value of the function f(jf) is different from zero. we For
For
f(x)
is
!:
may
take the highest of the absolute values of the several roots of
the equation f(x)
0.
Without assuming the existence of roots

also obtain a value of ^ as follows:*
coefficients
of algebraic equations,
we may
Let a k be the numerically greatest of the n

in (1),
ac a
,
1?
and denote
an
0,
by
r.
We
have then
,..+i)
Hence,
for
any value of x not lying between

"
~.<"
r
t>
and
x
-J- r,
-f
+a
<
so that the sign of
/ (.r)
/.
is
the same as that of a n
x".
we may
take
Consequently,
*L. Kronecker. Crelle
101. p. 347.
SYMMETRIC AND TWO-VALUED FUNCTIONS.

(B).
(2)
If
no two of the integral functions
/.(*),/,,;../.,(*)
are identically equal to each other, then there is always a quantity C such that for every x the absolute value of which is greater than
C
,
the values of the functions (2) are different from each other.
For,
if
/JB
we denote by
)>
r a8 the value
determined for the function

for C the
fa (x)
as r
was determined in (A), we may take
greatest of the quantities r a p.

(C).
If in the integral function
/O l,
all
V-2,
U)
=2
Al A2.
A,,,
*/
ff,>
the coefficients a are equal to zero, then the function / vanishes identically, i. e., the value of / is equal to zero for every system of XH values of *\ .r 2
, ,
. .
To prove the converse proposition we put

(3)
&if=rfc
2
x*
v
Gf
,
*3
<T,
...x n
g,
=
the
ff-
x^
A
. .
.r./
.r,,,
then becomes a power of
exponent of
1
which
is
r\i A2
. .
A>,,
^-i
+V+^+
v X>H
^X""
From
ues of
v,
(B),
we can find
a value for are
such that for
all
greater val
the various r Al A2 ...
all different
from one another.
We
have then
f(x ly x2
...
xm)
=2a
AlA2
A)n
r g *
^
we can
proposition
But, from (A), if take g so large that /

is
all
is
the coefficients a do not disappear,

different
from
0.
The converse
then proved.
(D).
If a product of integral functions
a- 2>
(4)
is
/i (#1,
...
O/
and
0i, #2,
/< (#i,# 2
XH )
equal to zero for
all
systems of values of the x\s, then one of the
factors is identically zero.
For,
select
if
we employ again the

va
substitution
f actor
such values g a ... X H ) which 2 any does not vanish identically, that for every system of values which
for
,
we fa (x lt x
(3),
can,
from
(C),
arises
from
from
zero.
v
If then
same time
when g > g a and * > v a the value of fa is different we take g greater than g ly g. g m and at the v v m we obtain than n greater systems of values
(3
)
2
, .
. .
v. 2 ,
4
of the a\ s for
J
which
(4)
does not vanish, unless one of the factors
vanishes identically.
The proof
arises
of
Theorem
I follows
now
directly
from
(C).
.
For
.
.
if
xn ) and if c^ (x^x.2 then it follows from of the x\s, rearrangement by any the fact that $? has only one value, that the difference
,{*) is
a single- valued function,
from
<?
cr
OM
,,
XH )
9>,
On
<T 2 ,
XH )
vanishes identically.
If the elements x\ are not independent,
Theorem
is
no longer
necessarily true.
3
For
instance,
if
all
arbitrary function of the

a*,
x\s
is
is
single-valued.
single- valued
the x\s are equal, then any Again the function

if
H- 3,*Y x,
4^
in
x-2
-f-
3 ir/
x^
=
all
2,r 2
although
it is
unsymmetric.
3.
If,
any symmetric function, we combine
terms which
only one of these terms,
differ in their coefficients into a single term,
and consider any
Cxf x./ aV,

it
of the function requires that
then the symmetric character should contain every term which can
.
be produced from the one considered by any rearrangement of the If these terms do not exhaust all those .TA S. present in the func
tem
C xf x.f x/ ... in which the sys not the same as in the preceding case. This term then gives rise to a new series of terms, and so on. Every symmetric function is therefore reducible to a sum of simpler sym
tion there will
still
be some term,
of exponents
is
metric functions in each of which

single one
all
among them by rearrangement
the terms proceed from any of the x^s. The several
terms of any one of these simple functions are said to be of the same type or similar. Since these functions are deducible from a
single term,
it
will suffice to write this
one term preceded by an
S.
Thus S (x 2 ) denotes in the case of three x* -j- ay -(- x3 2

4.
case of two elements

,
x?
-f x,
2
,
in the
etc.
the elements x lt x.2 ... x n as the roots of an th equation of the n degree, this equation, apart from a constant f a<N tor, has the form
If
,
we regard
(5)
f(x)
(x
a?,)
x2 )
...
(x
xn)
the left
member
of which
expanded becomes

(6)
af
(o^-f
(x, x,
*V+
(
...
3
+x
9
)x".
.
+
The
+ ^ # + x, x +
lYx.
powers of x in
.....
+ av
... -f
coefficients of the
this equation are there
fore simple integral symmetric functions of the x^s: (7)

c1
= S(aj ),
1
02
= 8(0^),
S (x x 2
l
. .
.
CA
=
.
S to
<* 2
av),
.....
cn
xn}
=x
xn
These combinations
CA
are
called
the elementary symmetric
functions. They are of special importance for the reason that every symmetric function of the x\s can be expressed as a rational inte
gral function of the C A

5.
S.
Among
the
many
1
proofs of this proposition
we
2
select that
of Gauss. *
We call a
the
first
term x
x/" 2
x3 " 3
higher than
/^ m.2
,
(
x^
2
,
,r/
s
x3 M
(
when
which
of
the differences
is
This amounts then to assigning an arbi trary standard order of precedence to the elements a\. In accordance with this convention, c n c 2 c 3 have CA,
does not vanish
positive.
,
,
.
for their highest terms respectively
Xj_
x.2
x x2
1
.
x.A
.
#!
its
x z xs
a? A
..."
and the function

"
c x a c.f c^ . + ? + y + x^
has for
highest term
xf
.
. .
+ v
xj
+
; -.
In order, therefore, that the highest terms of the two expressions,

tt
<a
cf
c 3?
and
a
Cj
cf
cj
may
be equal, we must have
+ P + r +...=
+ +r +
...
+
that
.is,
a=
=$
,]>
= /,
...
.
two different systems of exponents in c^ c/ cJ. give two different highest terms in the x\8. Again it is clear that
It follows that
.
xf x./ x,y
is
*
(a^p^r^d...)
c^"^ c/~ ^c^~
5,
the highest term of the expression
s
. . .
and that
Cf.
Demonstratio nova altera etc. Gesammelte Werke III, ecker, Monatsberichte der Berliner Akademie, 1889, p. 943 seq.
pp. 37-38.
Kron-
6
all
the terms in the expansion of this expression in terms of the X^B
are of the
6.
same degree.
If
now
is
a symmetric function
S be
given of which the
highest term
A
the difference
Xi
a
x.J*
x,y
xf...
cjf-* ...
^
1
,3
>= Y
=d
S
will
left,
d--* c/-v
;
=S
in the subtrahend on the again be a symmetric function and if, the values of the C A S given in (7) be substituted, the highest
will
term of been
be removed, and accordingly a reduction will have
effected.
If the highest
term of
A, cf
Si is
now
-y
x?
-*
x.f
. . .
xj xf
then
S
is
-V
c.f
c,y
=&
The again a symmetric function with a still lower highest term. S.2 and Si are clearly not greater than that of S, and since of degrees
there
is
only a
finite
number
of expressions
degree which are lower than

repetition of the
xf x.f xj
at the
x^ x.f- x v ... of a given we shall finally arrive by

3
;
same process
symmetric function
(k)
.
. .
that
is
Sk
s
Ak
*&-!&>
c/-y
0;
and accordingly we have
=A
cf-ficf-y
... -f A.2 Ci*
-Pc.J"-y
...
+ ...
4-^.c^-^c/^-^...
S
7.
It is also readily
shown that the expression

S
function of the x^s as a rational function of the CA

in only one way.
of a symmetric can be effected
an integral symmetric function of x^ ar 2 ... x n could be reduced to two essentially different functions of c c 2 ... c n ... c,,) and </ (C M c_, ,. C H ), then we should have, for all
For,
if
, ,
} ,
values of the
.r A s,
cr
the equation
(c^c,,,
<p
.
.
c,,)
0(ci,c 2
c,,)
The
difference
f,
which, as function of
the CA
tp
S,
</
is
not
identically zero, since otherwise the
two functions
and
would
coincide, must, as function of the x-\s, be identically zero.
& those terms in c c 2 ... c which Suppose, now, that in <p cancel each other are removed, and let any remaining term be
:
lt
cf
C;P c-^
This term on being expressed in terms of the x^s
will give as highest
term
+ ^ + ? +
B
Now
a
l
xf + y+~- xj+ r.
.-
the different remaining terms cf cf c^ .give different Consequently among these highest highest terms in the x\a (5). terms there must be one higher than the others. But the coefficient
. .
of this
<P
term is not zero; and consequently (C)) the function have therefore zero. We cannot be $ identically
II.
(2
Theorem
An integral symmetric function of x^

,
x,2
...
xn
can always be expressed in one and only one tuay as an integral

function of the elementary symmetric functions c n c 2
8.
... c n
If
we
write SA
S (xf)
.
for the
sum
of the / powers of
a? we might by the above method. It is however simpler to obtain this result by the aid of two recursion formulas first given * by Newton and known under his name. These formulas are
the n elements
x lt
x.2
rt ,
attempt to calculate the s\a
as functions of the cfs
A)
8,.
C1
Cj
,._
+C
8 ._
I
B)
s,.
$,._!+
c_,s,._
+(l)"Cn 8r _ H = Q ... +(-_l)r r<?r =0

.
r=> n
(r^7i)
These two formulas can be proved in a variety of ways. The formula A) is obtained by multiplying the right member of (6) r ~n = 1, 2, .n by x replacing x\yy x^, and taking the sum over /
,
.
The formula
B may
)
be verified with equal ease as follows.

x.2
,
If
.
we
by
represent the elementary symmetric functions of

c/, c,
, .
.
xn
<?
_!,
we have
C2
= X + C/, G!
l
=X
n,
1
C/
+ C/,
1
C3
= X, C/ -f C /,
;
and accordingly,
x,
if
^
c,
we have
2
. . .
=
-f
r xf- -f c x^l) ( xf 0*i + c/) xf- + (x.2Cl + c/) ^ly (x^r^+c^) = ( i)V (
c,.
2
. . .
and hence, replacing

pondingly, c/ by
s
a? x
successively
. .
c,.", c,. ",
a? M and, corres by x ^ 3 of the n the sum c,W, and taking

2
, ,
. . .
resulting equations
.
c 1 s ._
l
=
*
+c )r (C/ + C/ + C/" +
1
s,._
l)
. . .
r
<?r
n
C,.(" ).
Newton
Arith. Univ.,
De Transformatione Aequationum.
The right member

.
is
symmetric in x
x.2
the terms of c and no others.

t
Moreover, the term

r times.
consequently every term, occurs n

sr
X H and contains all x x and x,. Accordingly we have

.
c sr
l
.
_
.
+ CoS,._
ci s
>
l)
c,.n
(
= (!)
l)
r
r)c r
s,.
._
-f-
CoS,._2
...
c,.r = 0,
and formula B) is proved. * fied in the same way.

9.
The formula A) can obviously be

A) and B)
veri
The
solution of the equations

S
for the successive
values of the S A
of the CA S.
gives the expressions for these quantities in terms The solution is readily accomplished by the aid of
determinants.
We
add here a few of the
results. f
C)
S
*1
s2 s3 s s5
= = = = =
n
<*!
c,
2c 2
c?
3 Cl e 2 -f 3c 3
c^-5
4cfc 2 -f
8
4^3
2
-f 2c 2
4c 4
2
C!
5cj c 2 -f- 5cj c 3 -(-
Sc^
Sc^^
5e_,c 3 -j-
5c 5
It is to
be observed here that
all
the c\s of which the indices are
This is obvious if we add greater than n are to be taken equal to 0. to the n elements x x.2 ... x u any number of others with the
l ,
value 0; for the
CA S
.
. .
up
will
to c
lt
will not
be affected by this addition,

a
while
cH +
c,,
+2
be
0.
10.
The observation of
+ P + v
5 that
term x^
x/
+ v +
xj
+
.
c/ c 3 ? gives for its highest *. can be employed to facili

Cj
. . .
tate the calculation of a
symmetric function in terms of the C A S. We may suppose that the several terms of the given function are of the same type, that is that they arise from a single term
.
among them by interchanges of the aVs

homogeneous; suppose
its literal
it
The function
v.
is
then
to
be of degree
We
can then obtain
part at once.
the function contains one element, and consequently
all
For,
if
and no higher power, then every term of the at the high corresponding expression in the C A S will be of degree est. and For, in the first place, two different terms Cj c/ c.J
elements, in the
mth
tt
Another, purely arithmetical, proof is given by Euler; Opuscula Varii Argument!. Demonstrat. genuina theor. Newtonian!, II, p. 108. tC/. Faa di Bruno: Formes Binaires.

a
Cj
give different highest terms in the x^s, so that two such terms cannot cancel each other; and, in the second place,
c/
c3>
c \ a c/ csY
gives a
a power x\
+ & + ? +
---,
so that
wi.
a? 3 ?
+P+r+
Again the degree
of
xf
+ & + v +
;
+ y + x.2 P
+
-.
.
is
+
and
since
;
3-f3
+...=v
x
.
-f-2
is homogeneous, the sum a ?-|-3^-f- ... must be equal to v for every term c c/ c 3 ? These two limitations imposed on the exponents of the C A S that
the
given
expresssion
+?+r+
2
i m,
,-3
3 r
exclude a large number of possible terms. The coefficients of those that remain are then calculated from numerical examples. The
quantity
ci
a c
o.
-f-
/?
-f-
3 y
-\-
is
called the weight of
the term
all
f
2
V
x
l
an(i a function of the C A S

is
whose several terms are
of
the same weight
called isobaric.
7
For example
c5 c 2
3
S S
2
2
(x,
[(
x, x,
2
= g c + qi c
2
ct
a? 2
(x 2
2
ar s )
(a? 8
2 5
+g
ci c :
+& +3
c3
c 3 c 2 Cj
+g = ^ ] g +g +g +g
-f g 2
2
c, c 3
(m
c,
3
=
2 8
2,
7)
c6
c5
7
c3 cx
c2
+g + q c/ cf
c4 c2
(m
4,
6)
where the q s are as yet undetermined numerical coefficients. In the second example we will calculate the g s for the case w = 3,
for
which therefore
c4
= c =c =
5 6
0.
It is obvious that for different

different.
Cj
7
values of
I.
n the
coefficient
2
s will
be
Taking
#i
l,
=
3
1,
xs
7
;
= 0,
we have
= 0,
c,
c2
=
c2
1,
c3
= =
2
II.
III.
IV.
.-. S-4 = -g x, = x = 1, x = 0, .-. S = 0= -4 + 4g = = = x 1, # x, 1, :. S -0 = 4g +4-27; ^ = ^ = 2, x = 1,

2
g =
-4.
= 2,
= 1,
3,
c3
c3
Cl
g8_=l. .= 0, c 2 = q. = -27.
Cl
;
.S==0= -27-16
3
4 -27g c
=3, c = g = -4.
2
fi
0,
c3
= -4
c8
=
V.
= x, = x =
.-.
.
.
1,
c,
3,
c2
3,
S=0=
(a?!
-27
c3 c2 c :
+ 9g
4c 3
135; g 5
=18.
27c 3
2
2
x 2 ) 2 (# 2
x,) (x3
3
2
l
=
cx
+ 18
Cj
4c 2
-j- c 2
2
.
10
L>
a- ) (x 3 x.2 )~ ( x.2 This expression, (a^ J, is called x^ The characteristic the discrhuhnutt of the quantities ,r n x.2t a? 3 property of this discriminant is that it is symmetric and that its
:{
vanishing
of the
is
the sufficient and necessary condition that at least two

equal.
uVs are
11.
ties
,r 15
jc. 2
,
In general, we give the name "discriminant of n quanti a ,," to the symmetric function of the x^s the van
1 . .
ishing of which
is
the sufficient and necessary condition for the
equality of at least two of the the JL\ S is to vanish for x

r
x\s
x.2
,
it
If a symmetric function S of must be divisible by x x>,

l
and consequently by every difference x a

flf= (x,
x,) S,.
~) t ,
x$.
Suppose
x.2 S is unchanged if x^ and S, and consequently (a^ x., be But this changes the sign of x x.2 and there interchanged. fore of S Consequently Si vanishes if x = x.2 and accordingly
1
1 .
Xow
contains
x x.2 as a factor. The symmetric function S

t
is
and consequently by every (x a

-J
therefore divisible by (x x.y 2 that is it is divisible by a^,)

l
;
==
| |
AU
(x,
(*i
xj*
(/
<
/,;
1, 2,
n
.
1;
/,
2, 3,
n)
=
(8)
^)
(x,
(^2
x 3 y2 *3 2
(x,
x,y
(x,
xn y
2
. .
-!-This quantity J already satisfies the condition as to the equality of the aVs and, being the simplest function with this property, is
,
itself
the discriminant.
.
It contains
J n (n
1) factors of the
form
0<V any
rM )
2
;
its
degree
is
is
n (n
l)
th
1),
.
and the highest power
to
which
;r A
occurs
but, as
we
It is the square of an integral, shall presently show, unsymrnetric function, with which
the (n
we
shall hereafter frequently

12.
have
to deal.
consider another symmetric function in which the discriminant occurs as a factor.
Finally
we
will
Let the equation of which the roots are x lt

before, f(x)
= 0.
x>, ...
xn
be,
as
Then
if
we
write
SYMMETEIC AND TWO- VALUED FUNCTIONS.
11
we
/"
have, for
all
values
=
2*2)
1, 2,
.
n, the equation
a-A-i) (o?x
x+
1)
.
(.TA)
= (ir\ î) OA We attempt now to
OA
,/
,
.
OA
a?,,}.
express the integral symmetric function

(*,)
,
S [*,-./
in terms of the
(*,,)...
.
(*)]
coefficients c 1
c.2
c n of
/ ( a;),
Every one of the
n terms
of
is
divisible
by
^
,
x.2l
since either
(ojj)
or/
(x 2 )
occurs in every term.
Consequently, by the same reasoning as in 2 and therefore being a symmetric is S divisible 11, a\>) by (x 2 function, by every (x a Xp) that is by
l
,
J =TT(*A
AM,
- av)
(I
<
/,;
1, 2, ...
1;
P.
= 2,8,
e.,
....
n).
is
therefore divisible by the discriminant off(x),
i.
by the
dis
criminant of the n roots of
f (x).
?i
Now/ (as*)
other
Xfj,.;
is
of degree of
1 in
.T A
and of degree
1 in every
and therefore
î"
.r,
a
.
f /
(x-i)
(î)
f /
n
*&)
(-^s)
Consequently,
J
is
if
<
f (x> } f (x) n 1, S
t
is is
of degree a of degree of degree 2
+n
2n
n
1 in 3 in
^
x
l
.
is
3 in
l ,
while
it
of degree 2
2 in
But
since J
is
a divisor of $,
fol
lows that
(9)
is
in this case identically 0.

. . .
S [x^.f(x,) / (* 8 )
a
/ (*)]
0,
<
1.)
n 1, then S and J can only differ by a constant Again, if factor. To determine this factor we note that the first term of S is
of degree 2
2 in
l ,
while
all
the other terms are of

ll
lower
degree in
x lt
1
The
coefficient of Xi
8)
.
~2
is
therefore
. .
.
(- l)"-
^-*
(x,
-x
~
l
l
u)
(x 3
x2)
(x,
- *)
In J the
coefficient of
2
~2
is
(x,
x,)
(x.2
xy.
is
(djrt _,
x
(
2
)e
n(n
l)
The
have
desired numerical factor
therefore
1)
and we
12
l
(10)
"~
B[x *-\f (x ).f(xt )...f (xJ]=(-l

l
Formulas
(9)
and (10) evidently

<f
still
hold
if
we
replace
xf
or
by any integral function Moreover since

l
(x) of degree a
^ n respectively.
n(n-l)
(
1)-T-
=/
(*i)
(**)
.f (x u )
we have
according as the degree of

13.
<f
is less
than or equal to n
x^
,
1.
If
an integral function of the elements
x.2
...
t,
is
not symmetric, it will be changed in form, and consequently, if the oVs are entirely independent, also in value, by some of the possi
ble interchanges of the oVs.
The process The
of effecting such
an inter
change we
the
,r A
shall call a substitution.

call
Any
order of arrangement of
substitution whatever
;
we
a permutation.
substitutions are therefore
operations; the permutations the result.

leaves a symmetric function
Any
functions the form of
unchanged in form but there are other which can be changed by substitutions. For
example, the functions

I
x?
ay 4-
Xz
if
x?,
x.2
xs
-f
x x5
+ ê
+ x* + x
take
l
new
values
certain substitutions be applied to them; thus
if
x and
x, be interchanged, these functions

**!
become
5
~T
^*2
l~
^3
^*4
^1
^2 ^3
~T~
^l-^ s ~T" ^*6

}
^2
*^l
^*3
The
first
changed, the second also

Functions
are
two functions are unchanged if x and x, be inter if x and ar 5 be interchanged, etc.

designated
as
one-,
two-,
three-,
w-valued
according to the number of different values they take under the The existence of oneoperation of all the n\ possible substitutions. valued functions was apparent at the outset. enquire now as to the possibility of the existence of two-valued functions.
We
we have met with the symmetric function J, the dis criminant of the n quantities ar 1? x ... x n The square root of J
In
11
2
.
is also
a rational integral function of these
n quantities
*The formula
is
due to Euler; Calc.
Int. II
1169.
13
= (^
X 2 ) (X
X,) (X,
X.)... (X,
x,).. .(x 2
(x 2
~- x 3 ) (x 2
(x 3
xt ).. .(x
Xn ) XH ) xn )
Every
difference of two elements
xa
x n ). (#*_! occurs once and only x$
once on the right side of this equation. Accordingly if we inter change the aVs in any way, every such difference still occurs once and
only once, and the only possible change
is
that in one or
more cases
either
l
an x a
is
Xp may become
-f-
x-p
xa
But
The
>
result of
any substitution
is
therefore either
V ^ or
first
VJ
if,
e.,
the function */ J
one- valued or two- valued.
in particular,
we interchange x
its sign,
and
ir.j
the
first
factor of the first
row above changes
while
the other factors of the
ing factors of the other rows, since these do not contain either
then, for this substitution, \/ J
row are converted into the correspond the second row, and vice versa. No change occurs in
or
x.,
Since
becomes
\/ J
it
appears that
we
have in \/ J a two-valued function.

This function
is
specially characterized
values only differ in algebraic sign.

shall call alternating functions.
by the fact that its two Such two- valued functions we
Theorem
quantities
ties.
III.
2
,
The square root of the discriminant of the n
x lt x
...
xn
is
an alternating function of
all
these quanti
14
Before we can determine
the alternating functions, a
short digression will be necessary.
An
The
(x a xp).
interchange of two elements

,
transposition of
We
shall
we shall call a transposition. x a and x$ we will denote by the symbol now prove the following
Every substitution can be replaced by a
Theorem
IV.
series of transpositions.
Thus, if we have to transform the order x ly x.2 X BI ... x into the order x^, x,-2 a?/ 3 #/, we apply first the transposition
,
lt
(a?!
x-i
,
x^).
X( l + i, ...
.
. .
x,
The order of the xjs then becomes x^ x z x3 x^ i x n and we have now only to convert the order x in ... x n into the order x^ xi x^ - i, x x^ + By
, ,
.
. .
l ,
3 ,
14
THEOEY OF SUBSTITUTIONS.
repeating the same process as before, this can be gradually effected, and the theorem is proved. Since a symmetric function is unaltered by any substitution, we
obtain as a direct result
Theorem
transposition
15.
is
V.
function which
is
unchanged by every
which
symmetric.
is
There
therefore
at
least
one
transposition
changes the value of any alternating function into the opposite

value.
We
x,,
. .
will
denote this transposition by (x a x$)

v
,
and the
alter
nating function by
V- (a-,,
.
and accordingly we have

.
.
a?,.
if
. .
Ofc,.
,
#)
=-
(#!,
x2
...
0.
^,
...
xa
Accordingly,
xa
<r
x$ we must have
...Z, ...Xp,
=.
Consequently the
equation
(3C lt Xt,
....
X^
regarded as an equation in z has a root z

</
=
. .
x$ and the polynomial
is
therefore divisible by z
^ (a-j,
x@
,
.
The function
. .
x2
x.
a-0-
)
2
therefore contains
tains
(jc a
xa
x^ as a factor, and, consequently,

either remains
con
xpf
as a factor.
c
But
since, for all substitutions,

its
2
unchanged or
;
only changes
tain
sign, <
accordingly, since
all
factors of
contains the factor (x a 2 the form (a* A x^) i. e., ^

,
-
must be a symmetric function and, 2 it must con cr/g)

,
contains J as a
factor, and consequently c contains \/ J as a factor. is determined ing factor of by aid of the following
(.
The remain
Theorem
form S. \/ J
,
VI.
Every alternating integral function

is the
where \/ J
is of the square root of the discriminant and
is
an
integral symmetric function.
if c
That S. \/ J is an alternating function is obvious. Conversely, is an alternating function, it is, as we have just seen, divisible
Let
by V-J-
(V- )
1
"
be the highest power of
V J which occurs
as a factor in
</*.
Then the quotient
is
either a one- or a two-valued function, since every substitution

either leaves both
15
numerator and denominator unchanged or changes But this quotient cannot be twothe sign of one or both of them. J which is con valued, for then it would be again divisible by
trary to hypothesis.
It
must therefore be symmetric, and we have
accordingly
V
.
s,
(V J)
this
Now
tion,
an even number, the right member of and consequently the left, would be symmetric.
if
m were
We
S,
equa must
therefore have
m =
2;?,
-j-
1.
And
if
we
write
.^
we have
X =
Corollary.
S. A/J
form of an alternating function it remains unchanged or is changed in follows that such a function
the
From
sign simultaneously with

16.
V J for all substitutions.

to
Having now shown how
form
all
the alternating func
tions,
we proceed
<r 2
<?
to the examination of the two-valued functions in
general.
Let f(a? n two values of
a? M
be any two-valued function, and
let
the
be denoted by
PiO*?!,
x zj
...
<r M )
and
foOû
se2 ,
#).
value,
These two functions must

since the
differ in
form as well as in
if
and
x\s are any arbitrary
quantities,
we apply
to
and ? 2
any substitution whatever, the resulting values
Fif
will also
^
l
.
and
the result
But whatever substitutions are applied to & always <f or .2 In other (),
be different.
is
.
:
words
Those substitutions which leave the one value of a tiuo-valued function 
into the other, also
convert the second value into the

17.
is
first.
it
From
the preceding section
follows at once that
^+c
a symmetric function
-
03)
+* = 2ft.
16
Again the
difference
is
cr
2
<f l
cr
is
a two-valued function of which

(<f l
the second value
ft
2).
This difference
is
therefore an alternating function, and accordingly, from Theorem
VI,
(r)
we may
write
s,
= 2s ^J.
2
From
ft
(/5)
and
(f)
we
obtain
8t
+ s, v j,
it
?2
Si
v j,
$.
$ y^.
is is
Conversely,
two- valued.
is
plain that every function of the last form
the
Theorem VII. = S S. form

c1 l
Every two-valued integral function

\/ J, where Si and
the square root of
2
of
are integral symmetric
functions and \f J
is
the discriminant.
Con
versely every function of this
form
is
two-valued.
is
Corollary.
18.
Every two-valued integral function
unchanged
or changed simultaneously with
V ^ by every substitution.
we recog
which leave
From
the corollaries of the last two theorems
nize the importance of determining those substitutions
the value of \/ J unchanged.
of
13) that the transposition (aô*;,) changes the sign ( In the arrangement of the factors of \f~J in the same Section, we might equally well have placed all the factors containing x a or j-p in the first and second rows, x a x$ taking the place of
We
know
TJ
x.2
etc.
The
sign of \f J
ment, but whatever this sign

transposition (XgXp),
transposition.
may be changed by this rearrange may be, it will be changed by the

its
Consequently \/ J changes
if
sign for every
This result
is
easily extended.
For,
we apply
changed
is
/*
successively
times, that is
if
/*
any
,a
transpositions to \/ J, its sign will be

(
V J becomes
odd,
1)^ \/ J
If
/->
is
even,
VJ
unchanged;
is
V-J becomes
*/
J.
We have therefore
VIII. All substitutions which are formed from an odd number of transpositions change the value \f J into J;
Theorem
all substitutions
sitions leave
which are formed from an even number of transpo J unchanged. Similar results hold for all two-val
ued functions.

19.
17
Every substitution can be reduced

is
to a series of transpo
sitions in a great variety of ways, as
readily seen, and as will be
shown
in detail in the following Chapter.

it
theorem
follows that the

is
number
of transpositions into
But from the preceding which a
substitution
resolvable
is
always even, or always odd, according as

-J
the substitution leaves
unchanged or changes
its sign.
Tlieorem IX.
an even (odd) number of an even (odd) number of

20.
If a given substitution reduces in one transpositions, it reduces in every

transpositions.
way way
to
to
Theorem X.
of an equation of
the second degree of
Every two-valued function is the root which the coefficients are
rational symmetric functions of the elements
x ly x 2
xn
From
the equations of
ft
17,
=*+
s.2
VJ
ft
Si
J,
we have
for the elementary
symmetric functions of
c^
and
ft.
We
tion
recognize at once that
c^
and
<^
are the roots of the equa
p2__
It is
2^? +
($"-- ASi)
0.
however to be observed here that
it is
not conversely true
that every quadratic equation with symmetric functions of the x\s as coefficients has two-valued functions, in the present sense, as
roots.
It is further necessary that the roots

is
should be rational in
the elements x\, and this
not in general the case.
CHAPTER
II.
MULTIPLE-VALUED FUNCTIONS AND GROUPS OF

SUBSTITUTIONS.
The preliminary explanations enable us to indicate now the coarse

21.
tions, at least in their
general outline. one-valued and two-valued functions and have determined those
substitutions
Chapter our further investiga Exactly as we have treated

of
of the preceding
which leave the
latter class of functions
unchanged, so
we
shall have further, either to establish the existence of functions
having any prescribed number of values, or to demonstrate their

impossibility; to study the algebraic
form of these functions; to
determine the complex of substitutions which leave a given multiplevalued function unchanged; and to ascertain the relations of the
various values of these functions to one another.
Further,
we
shall
attempt to classify the multiple- valued functions; to exhibit
them
possibly, like the two-valued functions, as roots of equations with
symmetric functions of the elements as coefficients; to discover the relations between functions which are unchanged by the same sub stitutions; and so on.
22.
At the outset
it
is
necessary to devise a concise notation
for the expression of substitutions.
tities
Consider a rational integral function of the n independent quan x x.2 ... x,, which we will denote by y (x x 2 ... X H ). If
l
l ,
in this expression
such a way that
we interchange the position of the elements x\ in for x x ... x n we put x o?, x in respectively,
l ,
!} ,
where the system

function ^
of
numbers i M
1, 2,
i.
in
mutation of the numbers

(
...
?i,
we
denotes any arbitrary per obtain from the original

(l ,
.
... x n ) the new expression y> (x X,- H ). x^ now the manner of representing by symbols such a transition from x ly x ... x n to x xh x in to this transition we have already given the name of substitution.
: ,
x.2
We
consider
fl ,
CORRELATION OF FUNCTIONS AND GROUPS.
19
A.
symbol
In the
first
place
we may represent
,
this substitution
by the
#"35
... 3
which shall indicate that every element of the upper line is to be replaced by the element of the lower line immediately below it. In
this
mode
all
is
of writing a substitution
we may
.
obviously, without loss,
omit
that
those elements which are not affected by the substitution, In the latter case the entire x all those for which x k
i]e
number
of elements
is
not
known from the symbol, but must be
otherwise given, as is also true in the case of <p itself, since, for example, it is not in any way apparent from the form of
=
whether other elements
ation, as well as those
a? 5
,
x,x 2
,
+ x^x,
may
in
rb
...
not also be under consider
which appear
<p.
B.
Secondly,
we may make use

we denote such
of the result of the preceding
Chapter, that every substitution can be resolved into a series of

transpositions.
If
a transposition,
i.
e.,
the inter
change
of two elements,
by enclosing both in a parenthesis, every

(xp x q ).
substitution
may be
written as a series
(x a oc b ) (x c x
(l
This reduction can be accomplished in an endless variety of l^L of the ways. For, as is shown in preceding Chapter, we can
first
bring any arbitrary element to its proper place, and then pro ceed with the remaining n 1 elements in the same way. Indeed we may introduce any arbitrary transposition into the series and can
cel its effect
by one or more
it
later transpositions,
which need not
immediately follow
or each other.
C.
Thirdly,
. .
(x ai x a ,x ((3
we may also write every substitution in the form x ak ) (x x b ,x h x b j (x x x x ) .....

.
bl
Cl
c2
.
C3
c>
Here each parenthesis indicates that every element contained

it except the last is to be replaced by the next succeeding, the last element being replaced by the first. The parentheses are called cycles, the elements contained in each of them being regarded as
in
forming a closed system, as if they were, for example, arranged order of succession on the circumference of a circle.
in
20
If
we wish
to pass
from the notation
A to the present
. .
one, the
resulting cycles would read
(X X;
1
X,-.
(<
.,.</
Xft
.....
Here too
it is
clear that those elements
itself,
the substitution, every one of may be omitted in the symbol.
which are unaffected by which therefore forms a cycle by

In the notation
A these ele
ments are the same as those immediately below them. A fourth system of notation which is indispensable in many
important special cases will be discussed later. 23. It is obvious that each of the three notations, A, B, and C, In the order of arrangement contains some arbitrary features.
of the elements in the first line
is
entirely arbitrary; in
the
reduction to transpositions is possible in a great number of ways; in C the order of succession of the cycles, and again the first element of each cycle, may be taken arbitrarily.
The
that the
first
of the three notations, in spite of its apparent sim

;
plicity lacks in clearness of presentation
the second
is
defective, in
same element may occur any number of times, so that the " important question, by which element is a given element replaced,""
cannot be decided at
stitutions is not
first
glance,
immediately clear
and the equality of the two sub from their symbols. We shall,
therefore, in the following investigations,
employ almost exclusively the representation of substitutions by cycles. The following example, for the case n 7, will serve as an illus
tration of the different notations.

It is required that the order x,
,
x,,
: ,
replaced by the order x s
x^
xs x,, x^ X Q x& xz
, ,
.
,X~ shall be
The
first
method gives us
x2 x3 X X.
7
x, x, x, a?A __ fx, X, X l X, X.J \^X 3
x x X X.
2
1
x.
xA
X, X.J.
By
the second
we have
(x.2 x
7)
variously
( Xl
(avr 8 ) (x.x,)
x 3 ) (X&)
(x,x 2 ) (x.x,)
(x^)
(X2 X6 )
Since the given substitution resolves into

tions,
3, 5, 9,
transposi
always an odd number, we have here an example of the prin-

ciple of
21
Chapter
I,
19,
and
it
appears that this substitution
changes the sign of \/ J. The third method gives us

(ajjOJgffs)
(x 2 x ? )
(a? 4 )
(r6 )
= (x^x
(X 7 X 2 )
all
24.
We
determine now the number of
the possible substi
tutions
by finding the entire number of possible permutations. Two elements x lt x.2 can form two different permutations, x^x 2 and x.2 x If a third element xa be added to these two, it can be
,
x-,x { x,>
placed, 1) at the beginning of the permutations already present x-i x 2 x or 2) in the middle x^x.^ x.i x^x or 3) at the
, l
,
: ,
l ,
end:
x^x^x-^
x^x^x^.
There are therefore
2-3
!
3!
it
of three elements.
If a fourth element be added,
permutations can occur in
the
first,
second, third or fourth place of the 3
permutations already
obtained, so that from every one of these proceed 4

tions.
new permuta
permu
There are therefore in
this case 2
4! permutations,
and again,
tations.
for 5 elements, 5!, in general, for
elements, n\
If now, in the notation
A, we take
xs
,
for the
,
upper
line the natural
order of the elements,

cessively all the n\
x n and for the lower line suc x_>, possible permutations, we obtain all the possible
a? 1?
...
distinct substitutions of the

It is to
station
n elements.
be noticed that among these there is contained that sub for which the upper and lower lines are identical. This
substitution does not affect
any element;
it
is
denoted by
1,
and
regarded as unity or as the identical substitution.
Theorem
tions.
I.
For n elements
there are n\ possible substitu
To obtain the same

case of the notation aid of induction.
result
from the notation

for
more elaborate
investigations would be necessary
which
this is not the place; in
it is
easy to establish the number n\ by the
We
arrive in the latter case at a series of interesting relations,
of which at least one
may be noted
here.
If a substitution in the
expression for
which
all
the elements occur contains

ft
a cycles of
a elements, b cycles of
elements,
22
a a
this
where
b ?
...
n,
N)
we can obtain from

and by
cyclical
by
rearrangement of the order of the cycles
permutation of the elements of each single cycle

a! a"
b\?
expressions for the same substitution.

a! a" b\
distinct substitutions
Consequently there are

...
which contain a cycles of

on.
a elements, b cycles
of
,5
elements, and so
The summation
Hence
of these
respect to all possible
modes, N), of distributing the
numbers with number n gives
us
all
the possible n\ substitutions.
\a"b\
25.
p ...
=1
If
.
(.
,,
,r,,
now we apply all the 71! substitutions to the function x )j i. e., if we perform these substitutions, which
lt
may
be denoted by
1
- "loo Ay 2, Ojj
xn
o
.
a>
o o n jj
among
the
xn
x.2
...
in the expression c,
we
obtain n\ expres
l
sions, including that produced
by the substitution s
1.
These
expressions
we may denote by
?*!
?1, ?*,,
is
<P*s
<P, a
?*,,
or simply, where no confusion

r
likely to occur,
by
fl, f.M 9 3,
->a, ... P
all different
These values are not necessarily
from one another.

l
Some of them may coincide with
We
this
direct our attention at first to the
the original value ^>(x ^x.^ ... x n ). complex of those substitu
tions
which do
not change the
value of
<f>.
If

symmetric
is
complex
will
comprise
all
the n\ substitutions;
if
a two-
valued function, the complex will contain all substitutions which are composed of an even number of transpositions, and only these.
Again, for example, consider the case of four elements x lt x^
a- 3 ,
#4
and suppose
Cauchy: Exercices d analyse,
III, 173.
23
This function
is
unchanged by 8
of the possible
24 substitutions,
namely by
By
the remaining 4 8 16 substitutions ^ fact, is converted into either

!
is
changed, and, in
a?! a? 3 -f- a? 2
or
#1 #4
+ ^2
#3.
note then, in passing, that we have found here a three-valued function of four elements which is unchanged by 8 of the 24 possi
ble substitutions of the latter.
26.
We
Those substitutions which leave a function
<p
(x
l ,
x.2
oj u
unchanged, the
shall indicate
number
of
which we
shall
always denote by
r,
we
by
Si
G)
s1
1,
83,
*,-;
=1
is
of course contained
among
them.
Following the notation
of the preceding Section

* 2
=^ =
3
=

By
Si,s 2 ,s 3
supposition
,
there
is
no
substitution
different
i.
from
e.,
... s
which leaves the value of
unchanged;
we
have always
^+f>;
If
if S
+S
(A
1,2, 3, ...r).
,
now we apply two

c?,
substitutions s a
s^ of
our series
s l5 s 2
sr
successively to
and denote the
result, as
<?, a
above in the case of a
single substitution
by
be

<p
the result of the
two operations
will
V*a*P
(â)^
=
ft
and from
G).
that s a sp also occurs in the above series Every substitution therefore which is produced by the succes
this
we conclude
sive application of
two substitutions of
G) occurs
is
itself in
G).
What
is
true of two substitutions of the series
further clearly
true for any
number whatever.
which
results
from the successive application of two or more substitutions we call their product, and we write the
substitution
The
24
substitution
<r
which produces the same

,
effect
on the order of the
elements
JL\
o? 2
...
XH
as the successive application of s a
and
s^,
as the product
<?
s a Sp.
the series G) s n
The product of any number of the operations s occurs again in T/ie succession of the operations in s 3 ... s r s., = is to be reckoned from left to right. <? s a Sp s y a product
,
.
27.
The expression
of such a product in the cycle notation

:
which we have adopted is obtained as follows If the two factors of a product are
Sa
= (X = (x Sp
a QC (I1
Xa
.
(X h (X k
X hl X ...)...,
]I2
X bl X b
Xkl X k ...)...,
,
.
then in
s a Sp that
Suppose, for instance, that this
element will follow x a by which Sp replaces x ai element is x Again in s a Sp that

.
tll
element will follow x

example, x k,
,
jtl
by which
obtain
Sp replaces
jls
Let
this
be for
etc.
We
s a sp
=
xg
(x n
x,, 1 a? A 2
.
.....
replaces every index g of
If the substitution s a
be such that
,
it
i
the elements
1 ,
x.2
...
,
x n by
,
ff ,
and
if
Sp
be such that
it
replaces every index g by kg

sa
l
il
or, in formulae,
a oc ia
.
if
= (x x x ...) (x ..)..., = sp (x x x^ .. .) (x ,x kb .. .)...,

-.
1
kl
then the product will be of the form
The following may

sa
serve as an example:
xx
We have
tion
now arises how

a
The ques introduced here the expression "product." far the fundamental rules of algebraic multipli
b
cation
a,
(b
c)
b)
remain valid in
this case.
An
is
examination of this matter will show

fails,
that the former, the commutative law, in general
while the
second, the associative law,
retained.
In fact the multiplication of

sa
25
=
,
(x x x
1
ilL
iti
...)...,
sp
(x l
x kl x kki ...)...,
for every a
as performed above shows that it is only in the special case where, that the order of the two factors s a and s$ is i k,- a ka
,
indifferent.
This occurs, for example, as
is
a priori clear,
elements.
if
the
expressions for s a
and
s/g
contain no
common
interchange the individual cycles of a substi In tution in any way, since these contain no common elements. of page 19, on the other hand, this is not allowable. the notation
We may therefore
Passing
Sa
8
is
to the associative law, however,

S/3
if
=(x X ...)...,
(X 8
X ka
Sy
= (X^
ki>!
.,.)...,
we
have the following series of products,

sp s y
S a (SfiS y
)
= (x = (x
x
llfs
...)..., ...)...,
sa
sp=
x,
...)...
X lk
= (s a Sp) S y
(x a Xik *8 ...),..,
from which follows
Theorem
II.
In the multiplication of substitutions a
col
lection of the factors into sub-products without
of the factors, is permissible.

the contrary, generally
An
change in the order interchange of the factors, on

is
however permissible if
28.
Such an interchange the factors contain no common elements.

alters the result.
it
From
the preceding developments
appears that those
substitutions
G)
sl
I,s 2 ,s 3
sa
,
s,.
unchanged, form a closed group in this respect, that the multiplicative combination of its substitutions with one another leads only to operations already
which leave a given function ? (x
l ,
x.2
...
a?,,)
contained in the group.
The name "group" * we

ducing
itself
shall always use to denote a
system of
substitutions which possesses this characteristic property of repro
number
of elements operated
by multiplication of its individual members. The on is called the degree of the group.
It is not however necessary that all the elements should actually occur in the cycles of the substitutions. Thus
*Cauchy, who gave the first systematic presentation of the Theory of Substitutions " cl Analyse et de Physique Math6matique, employs the name system of conjugate substitutions." Serret retains this name in his Algebra, The shorter name, " group," was introduced by Galois.
in the Exercices
26
s,
=
if
1, s,
(j-jx,)
s.2
(xs x t )
s2
,
form a group, for we
have
s,s, =
sl
s.,s.2
=
,
^
,
1.
This
group
is
of degree 4,
consideration.
But we might
,r,
,
.
.
example, 6 elements
write for
&>
only the elements x x, x 3 a? 4 be under also regard the group as affecting, for in which case we may, if desired, .r, a\,
: ,
.
S,
(X,X,) (X.X,)
(Xr,)
(X6 ).
The degree of the group is then 6. The "number of substitutions contained in a group is called its order- as already stated, this number will always be denoted by r. The entire system of substitutions which leave the value of a
function
?>(;r,
,
x.2
x,,)
unchanged
is
called the group of substi
tutions belonging to the function y, or,

c.
.
more
the
briefly, the
group of
of
,,
The degree of the group expresses asunder consideration; its order .._,, ...
<p
number
of elements
gives the
number
substitutions which leaves the function
unchanged.
a? 3
,
Thus, given the four elements x

<p
l ,
x.2
3
and the function
x.2 -{-
x x i9
is
the degree of the group belonging to y

25, is 8.
4; its order, as
shown
in
For the
one.
five
elements x
l ,
x.2
x3 x
,
a? 5
the same function ^ has
a group of degree 5 and of order 8, identical with the preceding
For the
degree
tions
6.
elements x^ x 2 o? x6 the same y has a group of To the group above we must now add all those substitu
six
,
,
which arise from combining the former with the interchange of

(
x 5 and x
The group now contains beside the eight 25 the following eight new ones:
..
substitutions of
a? 5 a; 6 ),
The order
It is
of the group of
if
 4 ele ments, the order of the corresponding group becomes 8 (n 4) the group being obtained by multiplying the 8 substitutions of 25
easy to see that
we regard
as
by
all
the substitutions of the elements
a? 5
x&
...
xn

29.
27
The following theorem

III.
is
obviously true:
Theorem
tion there is
For every
single- or multiple -valued
func
a group of substitutions which, applied
to the function,
leave
it
unchanged.
perfect correlation of the theory of multiple -valued
To show the
functions and that of groups of substitutions the converse theorem:
we
will
demonstrate
For every group of substitutions functions which are unchanged by all the substitutions of and by no others.
Theorem
IV.
there are
the
group
We begin by constructing a function <p of the n independent elements a?,, a? 2 x n which shall take the greatest possible num
,
.
ber of values, viz
<p
is
therefore to be changed in value by the
application of every substitution different from unity.
Taking n -f- 1 arbitrary and different constants form the linear expression
...
we
If
now two substitutions s a =^x^x

to ^,
/-v
il
...)... and 8p
result,
/i
= (XgXk ..)>
s
i
on being applied
gave the same

s
*.
we should have
\
But the
ax s
satisfied if
being arbitrary quantities, this equation can only be each parenthesis vanishes separately ( 2, C), that is we
^.
must have
x^=xk
But,
if
this
be the
case, since the
x\s are
also
independent quantities, the two substitutions s a and Sp both replace x k ^, so that s a and Sp are identi every x\ by the same element x ^
;
cally the same.
It is only in this case,
where
sa
and
Sp are identical,
c?
that they can produce from y the

distinct values.
30.
same value.
Accordingly
has n
If
now
a group
be given, composed of the substitu

...
Sa ,
tions
Sj
1,
S2 , S 3
...
S,.,
which we
will indicate symbolically
by the equation
to the n!- valued function with
we apply
n
-\-
all
the substitutions of
parameters
28
and denote the
results of these operations as before

cr
by correspond
ing subscripts attached to the

9%,
Pi 5 <Pv
^,
iV
Then
the product of these functions c
will
be one of the functions to which the group of substitutions

this
ff
belongs.
To prove
substitution
it
must be shown
and
1) that
is
is
unchanged by every
in value
of G,
2) that $
changed
by every
substitution r which does not occur in G.
In regard
to the first condition

<r
we have
9O..O-;
?*i<r
?W ?V
s^
and, from the definition of a group,
=
<p
<r,
s.2
ff,
s 3 ^,
.ff
t
are again
contained in G.
another; for
Moreover, these products are
all
different
from one
if s a v
and
S$<T
applied to
have the same
effect, this
must
also be the case with s a
and
Sp alone,
and therefore
l
sa
and
.
Sp
t
are identical.
Accordingly the substitutions s

l
ff=ff,
1,
&>,
s.2
ff,
s.3 <r,
.ff
are identical, apart from their order, with s
sa
...
s,.,
and
hence the functions
are identical with

?*!,
F*,.
and accordingly
as
was
asserted.
to the
As
are
all
second condition, the substitutions s^,
s.>r,
s 3 r,
.r
different
9" X1T
,
from
vT?
.
s ly s,, s 3
.
.
...
s,.,
tions
V*,,
cr y
.
^T
cr
,
are
all
and consequently the func different from the factors
0. s-\, 9%, F*,. of Moreover, this difference is such that no can be equal to the product of a c by a constant c/ in which
?
s/
case
would become equal

o
to
a
if
r,
c 2 c3
c,.
=
Oj
1.
For,
4a,, .>\,
if
-f
4-
aV. 4-
c,.(
ar,-,

then, since the
A s
29
at
are arbitrary quantities,
it
would follow
once
that
1,
..
and consequently we should have the impossible equation c? T = <p f 31. In many cases the calculation of # is impracticable, since
the multiplication soon becomes unmanageable even for moderately There is however, another process of construc large values of r. tion in which the product is replaced by a sum, and every difficulty
of calculation is removed.
We
begin by taking as the basis for further construction, instead
of the linear function c, the following function

</
\(x
,x.2 ,x.i
XH )
x^x.^x,^
x u a >i.
The
ties,
are to be regarded here, as
before, as arbitrary quanti

it
and, as the
x\s are
also arbitrary,
follows at once that ^
is
an n .-valued function.
For,
if
<r
<T
then we must have identically
and, from
2,
,-,
C, this
that
is
responding / We denote the functions which proceed from $ under the opera
tion of the substitutions of
is, if
only possible if every ,3 is equal to the cor the substitutions a and r are identical.
f
by
^=A,^V--&,
and form now the sum
^=^+^ + ^3+
The proof
of the correlation of
...
$?>
+tfV
proceeds then exactly as
$.
and
in the preceding Section in the case of
G and
REMARK.
s
By making
certain assumptions with respect to the
Thus we may assign to the </ s some new properties. select the a s in such a way that an equation between any two arbi
we can
trary systems of the a

,i
s,
,-
= %+
*-
>
necessarily involves the equality of the separate terms on the right
and
left.
This condition
is satisfied if,
for example,
we take
30
in particular
if
0,1=1,
E.
/8
02
= 2,
08=4,
/
8,
16, ....
?\
g.
If
,-,
-[-
a,-
-+-... 4-
== 13,
we must have
1,
?2
= 3,
= 4,
3.
EXAMPLE.
familiar group
We
will
apply the two methods given above to the
G = [l,
(ir,ar 2 ),
(ayr,),
(a^) Ow),
(aw,) (x>x 4 ),
( ?l
(*v
(a-iX 3 x>;r 4 ),
(aôvrg)],
4, r
= 8)
taking as fundamental functions
a-j -j-
2> 2
a- 3
ia? 4
and
x^x^x^x^
where, as usual,
1-
We
have then the following results
-f
ta- 4
a- 2
ea-,) (a- 4 -f-
i> 3
Neither of the two methods furnishes simple But from $ we may pass at once to the function
results directly.
and from
to the
l
two functions
a? 2
) (a- 3
(x 4-
4-
a- 4
and
x^x.2 4-
x^
It is clear also that by alter the latter being already known to us. Jr we can obtain a series of func occur in which the exponents ing
l
tions all of
which belong
a
a
) (
to
G.
Among
tt
these are included
all
functions of the form

(Xi
4~ #2
XZ a
4" ^4;
#l
^2 4~ ^3 ^4^
31
In general we perceive that to every group of substitutions there belong an infinite number of functions.
It
may be
to
observed however that

a given
we cannot
obtain all functions
belonging
function
group by the present methods.
Thus the
x&z
belongs to the group
-f-
x3 x
(xiX s
+ x x)
2
G=
but
the functions
;
[1,
(aw), (o%), Oi^) (0^4)],

these
.
cannot be obtained by
"
",
<!
,
methods.
More
to
generally,
if
v
.
<r
H",
",
and
if
the groups belong respectively the substitutions common to these groups

. .
(Cf.
44,
Theorem VII) form the given group
(7,
then the function
where the
32.
s are arbitrary,
We
. . .
now
belongs to the group G. proceed to consider the case where the elements
x x
l ,
xn
are no longer independent quantities.
Even where any system of relations exists x ... a?,,, excluding only the case of the among 11 x.2 two or more equality of elements, we can still construct nl-valued XH * functions of x x.2 Using the notation of the preceding Section, we start from the same linear function
the elements
,
Tlieoreiil V.
l ,
^2
..._
+
?>
#,
s
and form the product
of the differences of the
ff(*r
\
fr)
n (n
this
_
^
product being taken over the
Y) -
possible
combinations
of the ^ s in pairs.
Expanding we have
In no one of these factors can

otherwise either the substitutions
the the
<r
all
v
and
the parentheses vanish, since r must be identical, or else

.
s s
are not
,
all different. The product, regarded as a function of therefore cannot vanish identically ( 2, D). Consequently,
*Cf. G. Cantor: Math.
Annalen V,
133;
Acta Math.
I,
372-3.
32
(
2
).
there are an infinite

all
number
<p
of systems of values of the
for
which
the u\ values of
are different
from one another.

9
Corollary.
The only relations among the x
which can
determine the equality of linear expressions of the

<;
form
-f a^\
-j-
a.2 X.2
+.. .+
in<l<
LK>ndentlu of the values of the

.r s.
s,
are the relations of equality
of two or more
33.
With the Theorems III and IV the foundation is laid for a classification of the integral functions of n variables. Every func
group of substitutions; to every group of substi This relation tutions correspond an infinite number of functions. which are functions between ship is not the only connection
tion belongs to a
unchanged by the same substitutions; we

ponding
algebraic
relation,
shall find also a corres
belongs to a
group
every function which namely, can be rationally expressed in terms of every

that
other function belonging to the same group. It becomes then a fundamental problem of algebra to determine The general all the possible groups of substitutions of n elements.
solution of this problem, however, presents difficulties as yet insup
erable.
The
existence
is
of
functions which possess
prescribed
number
It discussed in one of the following Chapters. will appear that there are narrow limits to the number of possible groups. For example, in the case of 7 elements, there is no function
of values
deduce the general proposition that a function of n elements which has more than two values, will have at least n values, if n > 4. A series of other anal
which possesses
3, 4, 5,
or 6 values; and
we
shall
ogous results
will also
be obtained.
For the present we shall concern ourselves only with the con struction and the properties of some of the simplest, and for our
* purpose, most important groups. First of all we have the group of order n 34. composed of This group belongs to the symmetric func all the substitutions. tions, and is called the symmetric group.
!
In Chapter I we have seen that every substitution

to a series of transpositions.
*
is
reducible
all
Accordingly,
II, g
if
a group contains
Cauchy:
loc. cit.
Cf. Serret:
Cours d alg6brc sup6rieure.
410-429.

the transpositions,
identical with the
it
33
is
contains
all
the possible substitutions and
symmetric group.
To secure
all
l
,
this result
it is
how
ever sufficient that the group should contain
which
tions
affect
any one element, for example x

fax.:),
those transpositions that is the transposi
fax
3 ),
faxj,
fax,,).
of
For every other transposition can be expressed 1 in fact every (x a xp) is equivalent to these n
;
as a combination
a series of three
of the system above,
(x a xp)
fax a ) faxp)
fax*),
is
(where
it is
indifferent).
again to be noted that the order of the factors We have then
not
Theorem
contains the
VI.
1
group of n elements x^
(.Va-l), (x a X a+l ),
x.2
...
x n which
transpositions
(aV*l), (XaX
is
Ua^n)
identical with the symmetric group.
Corollary.
group which contains the transpositions

(X a Xp), (X a Xy)
t
.
(XaXfi)
contains all the substitutions of the symmetric group of the elements

*X~ai
^/3j
^y
3Ci}-
tions
all
a group composed of all those substitu which are equivalent to an even number of transpositions. For these substitutions, and only these, leave every two-valued func
35.
We
know further
tion unchanged,
this
and they therefore form a group. will call r the Its is as order unknown, alternating group. group yet
to determine
it.
We
and we proceed
I)
Let
,
S,=
all
1, S 2
S3
.. S,
be
II)
the substitutions of the alternating group, and let

S/,
S,
,
;/,
...
S/
I),
be
all
the substitutions which are not contained in
and which are
therefore
composed
of an
<r,
odd number
for
of transpositions.
ff
We
select
now any
series
transposition
example
S 3 T,
fax.2 ), and form the two
Sjtf,
)
S.,<7,
...
ff,
.
S r <T,
.
.
II
s/<r,
s2
<r,
S3
st
ff.
34
Then every
transpositions,
substitution of I
is
composed
)
of an
odd number of
and every
sa
ff
and every substitution of II

)
of an even number.
II),
Consequently every substitution of I
is
contained in
sa
ff
one of II
is contained in I)
Moreover,
s^ff,
and
+ V^
V
for otherwise
Sa Sa
we should have
-
= Sa = 8a
(ff (ff
ff)
ff)
= (S a = (S a
2)
<r)
ff=(Spff)ff=
ff
Sp(ff
ft
ff)
ff)
=
t
(Sp
ff
ff)
=S
t,
(ff
ff)
= S0, = S0
r
since
ff
ff
= (x^.,)
(xiX
1.
It follows
from
this that r
all
and r ^L
that
is
t
since I)
and
II) contain
the substitutions, r -^
= =n
t.
Again,
.
Hence
r==
y
^
.
We will note here that there is no other group F of order ^ For a function y belonging to such a group would be unchanged by It would -TT substitutions, and would be changed by all others.
l
therefore possess other values beside
Suppose
?<,
to
be one of
<f l
these values, and let
ff
be a substitution which converts
into
^2
= ?.
?!
If
now
(f
is
unchanged by the group

,
III)
*VP-*
must be converted into
ff,
*/,
...^.i],
all
then
<?.,
by
the substitutions
n! (r;
IV)
for SA leaves
si
ff
s.>, s
:i
V,
ff
... sS
y>,
unchanged and
<f> l
converts
^> l
into
^.j,
consequently
sa
ff
will also convert
into
<f>. 2
the series IV) are different

it
Again all the substitutions from one another. For, if s a ff

.
of
Spff,
would follow that

from the s/
sa
sj.
The
substitutions
sA
are
also
different
s, for the latter
have a different
effect
on
Consequently III) and IV) exhaust all the possi ble substitutions, and y> is therefore a two -valued function, for there is no substitution remaining which could convert î into a third
l
from the former.
value.
The group F
is
therefore the alternating group.
Theorem
order.
This
VII.
is
For n
elements there
is
only one group of

to the two-
the alternating group.
It belongs
valued functions.
can generalize this proposition. The proof, being exactly parallel to the preceding, may be omitted.
We
35
Theorem
Corollary.
VIII.
Either
all,
or exactly half of the substi
tutions of every group belong to the alternating group.
Those substitutions of any given group of order
r which belong to the alternating group,
form a group within

or
.
the
given group, the order of which
is either r
The simplest
alent to
substitutions belonging to the alternating
group
contain three elements in a single cycle, (x a xpxy ).
They
are equiv
two transpositions, (x a xpxy )
(x a x y ) (xpXy).
(aj fl
,
substitution containing only one cycle
!o ,
i>r
we
shall call a circular substitution of order
m.
2
Theorem
IX.
If a group of n elements contains the n
circular substitutions
(x.x^,
it is
(XiX^Xi),
(x^Xn),
either the alternating or the
symmetric group.
For since
(XaXftX y )
it
faxjCp) (x&jXp) (x^Xy) (XtXoXa) (x^X*)
(x^X^
follows that the given group contains every circular substitution
of the third order.
And
again, since
it
follows that
composed
ber of
substitutions occur in the given group which are and consequently o four, six, or any even num The theorem is therefore proved. transpositions.
all
of two,
We add the following theorem Theorem X. If a group contains

:
all circular substitutions
of order
sequently
m -f 2,
it
it
will contain also all those of order m,
and con
will
contain either the alternating or the symmetric
group, according as
is
odd or
even.
For we have
(35^2
.X m X n X b )
2
.
.
= (XiX
Finally,
(X^
1
X m X Xb )
tt
(x m X m _
X&tjXiXa)
X m _ X w ).
give the criterion for determining whether
we can now
a given substitution, expressed in cycles, belongs to the alternating
group, or not.
The proof
is
hardly necessary.
36
Theorem
cycles,
it
XI.
// a substitution contains
elements in k
does or does not belong to the alternating group according

is
as in
k
36.
even or odd.
single substitution at once gives rise to a group,
itself
/.
Any
it
if
multiply
by
e., if
we form
its
successive powers.
we The
meaning of the term "power" is already 27. We must have opments of

Sl
6
n>
1
.
fully
denned by the devel

!><
Q o
,
o S
Q o
ll
G &
>H
Q o
Q o
Qa o
&
n1
tt
= s"
The process
clear
third, fourth,
.
s?
of calculation of the powers of a substitution

If
is
also
from the preceding Sections.

th
.
.
we wish
of
to
form the second,

th
power of a
cycle, or
any substitution, we
following element of each
. . .
write after each element the second, third, fourth,
element of
the
corresponding cycle,
the
first
cycle being regarded as following the last.

(.iV
Thus, from the cycle

)
V V
.i
.-,
)
. . .
we
),
obtain for the second power UY*V-,

for the fourth (x^x-^x^
.
for the
third (xjXifa
),
etc.
It is
obvious
that in this process a cycle
may break up
a
into several.
This will
occur
when and only when

cycles
the
number of elements
of the cycle
and
the exponent of the power have
common
d.
divisor d
>
1.
The
number of resulting For example,
is
then equal to
,)
=
(XiX*JCsX^
(x^Xr^ (-Aô), (x3 x
(l
),
If the number of elements of a cycle be m, then the

th
(3///.)
,
.
m th
1.
(2w)
to 1.
th
,
powers of
g.
(.r 1
.r 2 .r
:
the cycle,
and no
others, ivill be
.
equal
.
E.
If
tl
.r v r,
:=
(x^x^ ^jc^- =
with
=
x ,
substitution contains several cycles
m,
elements respectively, the lowest power of the substitution which is equal to 1 is that of which the exponent r is the least common
multiple of
m,
m.,
Thus
This same exponent r
is
also the order of the
group formed by
the powers of the given substitution.
For
if
we
calculate
37
a further contimiation of the series gives merely a repetition of the
same terms
Q o
)
in the
same order:
<-
-Q --
Q> o, o
- 6,6
<}
o-
+3
-O
<p
y
Q o
2r
-o
2r
,6
-o _
r
_L
-|
Moreover the powers of

another, for
if A
from
A
to s
are different
from one
_ SA +
u.
SM
<.
(l _|_
;J
r)
then we should have contrary to hypothesis

8"
(PL
<
r).
The
extension of the definition of a power to include the case of

is
negative exponents
now
s-*
easily accomplished.
We
write
=s
o
ofc
fr
-*
.
so that
we have
o
fc
1
.
The
tion
substitution s therefore cancels the effect of the substitu

fc
,
s~
st
and
vice versa.
The negative powers

. .
of a substitution are
formed
(
in the
d
,
same way
d
,
as the positive powers, only that in
forming
powers, we pass backward in each cycle the last ... element being regarded as next pre 1, 2, 3, elements, ceding the first. It maybe noted that (st)~ t~ s~\ For (sfp (st)=l, and ~ the and then members of this by multiplying equation first into / ~~ into s *, we obtain the result stated.
l)
(
2)
3)
The simplest function belonging

<?
to the cycle (x l
x2
ni
ul )
is
x xr +
l
x.2
xf
+ xm sa
x,n
+x
if
x,
37.
Given two substitutions
and
sp
we wish
sa
to deter
,
mine the group of lowest order which contains not only to form all the powers s a A sp** and together, but we must form all the combinations
,
and
sp
we have
these
to multiply
A 10 &a
)
Q-fJj
e/T
>
O A<j U O^M-O A <j AQ Uo S a Sp^, bp^S a , S a 8pr
V
,
8fT8a 8p
0-/JtO
AQ V
,
Of the substitutions thus forrAed we retain those which are dif ferent from one another, and proceed with the construction until all
substitutions which arise from a product of
factors are contained
among
the preceding ones.
For then every product
of
m+1
fac-
38
THEOBY OF SUBSTITUTIONS. one of
tors is obviously reducible to
m factors,
and
is
also
contained
among
those already found.
The group
is
consequently is then
complete. In case sps a
= s as/,
the corresponding
group
exhausted by
in this case
2/j.
all
the substitutions of the
form
s a Sp\
For we have
>
Sp*S a
S/3
S a Sp 2lJ
= S aSp^, =
"
Sft
Sa
Sa
Sa
S aSp">* S a
=S
Sp"^
factors Consequently any product of three
is
reducible to a product
of two.
Thus
and the theorem
is
proved.
For example,
let
then
The group
the
of
lowest order
which contains
s,
and
s.2
contains
therefore at the most 5
4=
20 substitutions.
To determine whether
it is
number
is less
than
this,
we examine whether
possible that
If this
were the
case,
it
would follow that
But in the
er of
s,
,
series of
powers of s 2 there
is
only one which
and
/?
this is the zero power.
is also a pow must have we Consequently
= Y and = 5.
tutions.
The group therefore actually contains 20 substi These are the following, where for the sake of simplicity
we
write only the indices:
COEEELATION OF FUNCTIONS AND GROUPS.
39
= 1, = (1 2345), s* = (13524), = (14253), s* = (15432),

n
8l
Si
8l
Sl
a,
^
2
Sl
= (2354), s = (25) (34), = a s^ = (15) (24), (1325), s = (1534), V = (14) (23), = (1243), V = (13) (45), \ = (1452), ^ V = (12) (35),
s2
2
2 a
Sl Sl
1
8l
s,
8l
V= V=
= (2453), s = (1435), V = (1254),

s2
3 3
2
(1528)
(1342).
Analogous results may be obtained, for example, for the case
In case every s^s a

Sa sp
K
(/JL
=1,
2, 3,
.)
can be reduced
to the
form
x
,
the group of lowest order
which contains s a and

s a K Sp\
Sp is
exhausted
by the substitutions of the form
For by processes similar

stitution s^Sa" to
to those above
theforms/s/.
if
q
we can bring every sub The proof is then reduced to the

2
preceding.
Furthermore
sp
is
the lowest power in the series

sa ,
sp, s^
which occurs among the power of

as
then the group contains q times

.
many
,
substitutions as the order k of s a

A A in s a *sp is
For
1,
in the first place,
if
the exponent
greater than q
we can
at the
a s^sp" where v^Lq
1.
There are therefore
replace s^ by most q k differ
ent substitutions s a K s^.
Again
t
if
s.V = sa
then we mrist have,
if
we suppose
Sj8
and consequently
order r of
= Sa*-K =/ A =
*-"
v, p.
.
(V<g > v, <g (A

A
v
1),
1),
There are therefore actually q k

q
is
different substitutions.
s$.
It is readily seen that
a divisor of the
If three substitutions
s a , sp, s y , a? e stec/i s a sp^ s v ^ ,

^ 1
/z<a
/or every
l
i>.
s^ s a = s a
and if k
equal
to
is
be the
s/ s a = order of s a and
8
e
sp
s/ sp = s a Sp
is
sy\
s^ the lowest
power of
s$
which
is
a power of
equal
sp, s y is
s a,
which
to s^Sp",
power of s y sj then the group of lowest order which con

s a ", finally if
its
1,
the lowest
tains s a9
s a sp s y t
8
e
of order kqt, and

1,
.
substitutions are of the form

. .
.
(3
= 0,
1; e
= 0,
1; C
= 0,
1,
t1).,
The proof is simple and that we may omit it.
so clearly analogous to the
preceding
40
38.
With
this set of propositions belongs also the following
//
<7
= [1,
ft,,
s3
,
.s,.]
H=[l,t
,t 3
...*,. ]
be two groups of substitutions between which the relation

Sa$p
= tyS&
have holds for all values of a and ft, and if furthermore G and no substitution hi common- except identity, then all the combinations
8 a tfa
^/3Sa> (
= 1, 2,
... r;
/3
= 1, 2,
form a group of order
rr which contains
G and
H as subgroups.
= S a (s
e
For since
Safp
Syts
8a(tp8y)t&
t^}t S
=S
fji
tv
And we
other.
the substitutions s a tp form a group of, at the most, rr substitutions. will show also that all of these are different from one an
For
if,
for example,
Sa
tfi
= Sy
t&
then
we multiply both sides and tp~ at the right we obtain

if
l
of this equation
by
sy
at the left
Sy~
Sa
=tstft-
But
sy
s a is a substitution of
G and
ttp~~
if
a substitution of H, and
1.
consequently they can only be equal
both are equal to
Hence
The
substitutions of the
is
new group
.
are then
all different,
and the
order of the group
therefore rr
We
denote the group by
K= \G, H\.
We
add without proof, the following generalization of the
last
theorem.
Under
the
same assumption
s a tp
y ss,
if the
two groups
TT f
-y-
G and
which
H hare
/-
substitutions in
common
there is
a group of order
contains
39.
G and
as subgroups.* In later developments a group

Palermo Rend.
I.
will frequently
be required
*F. Giudice.
pp. 222-223.

the order of which
is
41
p.
a power of a prime
number
The
exist
ence of such a group will be demonstrated by the proof of the fol lowing proposition, from which the nature of the group will also be
apparent.
Theorem
number p
there
is
XII.
is
If
pf
be
the highest
!
power of
1
the
. . .
prime
n,
ivhich
a divisor of the product n

2
then
a group of degree n and of order p. In the first place suppose n < p so that n ap + b (a, b < p ). of the numbers ... Then, 1, 2, 3, n, only p, 2p, 3p, ... ap are divisible by p, so that / = a. We select now from the n elements a systems of p Zetters each, aud form from each system a cycle, as
,
follows
s1
x^x,^,A
x-i, s 2
=
from these is the group required
s2 ,
. . .
:
Then the group which
arises
s a \ .*
For every s\ with its various powers forms a subgroup of order p, and since no two of these a subgroups have any element in common, it follows from Theorem II that
Accordingly every possible combination of substitutions s^, s/, can be brought to the form belonging to
sfsfftf ....
v
fl
(a, ft ft
= 0, 1, 2,
p
n
1).
The group
it
K therefore contains at
For
if
Q S l
the most
substitutions.
And
actually contains this number, for all these p" operations are dif
ferent from one another.

Q S it
a.
l
Q BS Q y S
o v S a
<*
e y Q P S S 2 3
o 6
V
5
would follow that
and therefore, since
Sj
has no element in
common
with
s2
s3
we must have
Again,
if
=a
etc.
np\
we
shall
have/ = p + l>
,
-j
j-
since in the series
as distinguished from the bracket, *In the designation of a group the brace, [ ] indicates that the group referred to is the smallest group which contains the included substitutions. The bracket contains aHthe substitutions of the group considered, while the brace contains only the generating substitutions. The latter can generally be se lected in many ways. (!f. the notation at the close of the last Section.
3a
42
1, 2, 3,
.
.
,p~,
by p.
We
the numbers p, 2p, 3p, ... (p form now again the substitutions
l)p,
are divisible
000
1 >
O
jj )
2) 03 j
as before,
all
and
in addition to these the substitution sp + 1
which
affects
the
elements
Then the required group

KI
is
si
-2
si> 9 si> +
1 }
For
in the first place

S2> + 1 S 21
ff
we can
o Ar*
readily
show that
l
S l Sp + l Ao
g
A o
2
S aS1} + l
Sp +
o
2
S a+l)
a
2
Sj>
QA
2
.
A A
O A
=S
S AS
S AS
=S
s2 ,
. . .
2
S.
Accordingly every combination of the substitutions can be brought, as in 37, to the form
sfsfajf
. .
.
Sj
s2>
s/Sp-n"
(,
/?,
r,
7.
i,
0, 1, 2,
1).
But we must
also
show
x
in the present case that

th
we need only take

find that
the powers of sp +
as far as the (p
l)
We
which occurs in
can replace the highest power of sp + p by powers of s u s 2 .s^, and these can then be written in the order above. + f substitutions The question then remains whether the
Consequently,
if
l
&>p, we
k
l
sp +
2>
=p
thus obtained are
all distinct.
If
two of them were equal
we should have
But the substitution on the right does not affect the first subscripts while that on the left does, unless % = i, k, The proof then
. .
.
/.
proceeds as before.
If
w>p
we
select
but <j/, that is, if n = ap*-\-bp-\- c (a, b, c<p\ from the n elements x\ any a systems of p 2 elements each.
2
43
and any other b systems of p elements each. With the former we construct a groups K>, and with the latter b groups K^ The com
.
bination of these a-\-b groups gives the required group K^. the product of the numbers
For
(a-l) P
(a
is
I)p
(a-l)p + 2, -fp (a<p)
l,
a _i) p2+pj
divisible
by only the same power of

1,2, ...p,
as the product of
...p\
Again, account of the last
if
=p
then the exponent term of the series
of
p f is
increased by 1 on
(p
p partial groups 2 is In this case, exactly as for n = p 2 we add another 2 substitution which contains all the p p elements in a single cycle, th and the p power of which breaks up into the p substitutions as inr
not sufficient.
,
so that in this case the multiplication of the
the case of s2
above.
Then, as in that case, we can show that the
new group
the requirements. At the same time it is clear that the method here followed is perfectly general, and accord
satisfies all
ingly the theorem at the beginning of the Section is proved. 40. Since all the groups enter into the forma 2 3
K K K
l , ,
tion of the
group K, we have
// p
the following
Corollary.
sor of n\
,
then tve
power of p which is a divi can construct a series of groups of n elements

1,
f is the highest
K\
K\ KX +
,
i ,
Kf
which are of order respectively
i,p,p ...jp ,p
,
x+i
,
...y.
in the next following
Every group
is
contained as a subgroup
CHAPTEK
III.
THE DIFFERENT VALUES OF A MULTIPLE- VALUED FUNC TION AND THEIR ALGEBRAIC RELATION TO ONE ANOTHER.
41.
We
have shown in the preceding Chapter, that to every

l ,
function of
n elements x
ov,
x n there belongs
substitutions,
and that conversely
to every
a group of group of substitutions
there correspond an infinite
number
of functions of
the elements.
The examination of the relations between different functions which belong to the same group we reserve for a later Chapter. The problem which we have first to consider is the determination of the
connection between the several values of a multiple -valued function and the algebraic relations of these values to one another.
If ? (#1
,
0*2
>
is
t
Si
words,
if
the substitutions
longing to of the group G be s,. the possible substitutions (i. e. r < n!),
= 1,
We
proceed to construct a table, the
first line
of which consists
of the various substitutions of the

Sj
group G:
.
= 1,
S2 , S3 ,
.;
6r;
y^.
The second
plication of
all
line is obtained
from the
*7 2
.
first
by right hand multi
the substitutions s x by
<7
This gives us
S2
<T
S3
<7
...
8,.
ff.
(T
<.,
We
^"a<r
show
then, as in Chapter II,
35, 1) that all substitutions of
this line
convert
^V,
into
^2
for
since
<? tta
= <pu
it
follows
that
9V,=
then
2) that no other substitutions except those of this

<p.2
;
line convert ?, into

effect,
for
if
r is
a substitution which
has this
we
shall
have
MULTIPLE-VALUED FUNCTIONS
so
T(T2
ALGEBRAIC RELATIONS.
45
that
~~
l
~
Tff>2
leaves
r
;
tlie
function
= s\
and
= (rff
~
2
)ff.2
8\<T 2
c^ unchanged; consequently and therefore r is contained in

;
the second line

if
3) that all substitutions of this line are distinct
for
s a <r 2
= Spff
l
it
follows that s a
different
sp
4) that the substitutions
of
this
kne are
all
from those of the

all
first; for the latter all
leave ^ unchanged, while the former

If the
convert
into
>
2r substitutions
SA
and
s\ff.2
do not yet exhaust
all
<r
the pos
sible n\ substitutions,
vert ft into a
new
y>. 2
function
then any remaining substitution 3 will con = $r 3 for all the substitutions which <p ffs
;
produce
c^
and
are already contained in the
first
two
lines.
By
the aid of
^ we form
ff s
,
the third line of our table

,
S 2 ff3
S 3 <7 3 ,
s,.o- 3
<r.
cr
;
The
convert
substitutions of this lino have again the four properties just
discussed.
They
er
;
are
,
all
the substitutions and the only ones that

all
c^ into
and they are
different
from one another and

the possible nl,
from those of the preceding

If these 3r substitutions
lines.
do not exhaust
all
all
we
proceed in the same way, until finally
the nl substitutions are
arranged in lines containing r each.
We
kind.
shall frequently
have occasion to construct tables of
this
All these tables will possess the properties: 1) that all the
substitutions in any line will have a special character; 2) that only
the substitutions of this line will have this character

substitutions of
that all the

fre
any
line are different
from one another; and
that
quently, but not always, the fourth property will also appear: 4) all the substitutions of any line are different from those of any
other
line.
*
results
Summarizing the preceding
we have
the following

I.
If the multiple-valued function
(x l1
x<2
a?,,)
values
cr
,
cr
,
. ,
y ? and
is
if
,
of order r and contains the

possible n\
1, s 2
s3
.s,.,
we can arrange
substitutions as in the following table:

*Such
tables
were given by Cauchy:
Exercices d analyse et de physique matliemat-
ique, III., p. 184.
46
S,.ff^
GI
in irhlch
every line contains all

into the value
<? a
and only
those substitutions which
convert
y>
prefixed to the line.

$r a
,
The
several values c^
...
cr
of the function
<$
are called con
jugate values.
42.
From
the circumstance that
all
the substitutions of this
table are different
table exhaust all
from one another, and that the p lines of the the possible substitutions we deduce the following
theorems
Theorem
uiritior of n\
II.
The order r of a group The number

is
of
n elements
is
Theorem
III.
{>
of the values of an integral
function of n elements
a divisor of n\
/>
Theorem
is
IV.
The product of the number
of the values
of an integral function by the order r of the corresponding group
equal to n\
The third theorem imposes a considerable limitation on the pos sible number of values of a multiple valued function. Thus, for
example, there can be no seven- or nine-valued functions of five elements. But the limits thus obtained are still far too great, as the
investigations of Chapter
43.
to the
will show.
41 can be applied Precisely the same method as that of more general case where all the substitutions of the group G
belonging to v are contained in a group

function
</
j
belonging to another
so that
special case above
G G
is
a part or
subgroup of H, just as in the

of the entire or
was a subgroup
all
group.
We
see
at
once that
symmetric can be
arranged in a series of lines, each line containing r substitutions of the form .SA^ (^ And we pass directly from the pre 1, 2, ... r). to the case present ceding by reading everywhere for "all possible substitutions" simply "all the r substitutions of H". have
We
then
MULTIPLE -VALUED FUNCTIONS -- ALGEBRAIC RELATIONS.
47
Theorem
contained
of r
l
.
V.
among
If all the r substitutions of the group G are those of a group of order r then r is a divisor
x ,
Theorem
(,
VI.
if
Given two functions

the
(x lt
x2l
...
x n ),
& retains
which leave y unchanged, the

tiple of the total
(x 1? x z ... x u ) and same values for all substitutions total number of values p of y is a mul
<f>
,
number of values ^ of
n\
p
<f>.
For we have
n\
~-~-^
ft=
p
;
7T
7T
7*1
v-
Corollary. // a function <p(x l9 ... x n ) belongs to a subgroup of the group H, and if r is the order of G and r that of H, then <p on being operated upon by the substitutions of takes exactly
G
M
values.
r
still further extension of the subject we may include the case where two groups G and contain any substitutions in common. This case can be at once reduced to that of the prece
44.
By
ding Section.
tion
:
For
this
purpose we employ the following proposi
Theorem
VII.
The substitutions common
form a new group,

For
both
the order of
to two groups which is accordingly a divisor of the
orders of both the given groups.

if
ff
and
and r belong to both G and G., then G2 and occurs among the common
1
<r
r also
belongs to
substitutions.
<p l
2
The
same
result can also be obtained as follows.
If
and ^ be func
tions belonging to
and
G.,
respectively, then the function
where a and
substitutions
ft
are arbitrary constants, remains
unchanged
for those
is,
which leave both

1
and
y.2
unchanged, that
all
which
are
common to G and G>

^,
These, being
the substitutions which
belong to
form a group.
I.
Corollary
stitution.
Two groups
ivhose orders are
prime
to
each
other can have no substitutions in
common
except the identical sub
48
Corollary
all or
II.
The order of every group
which consists of
G<
is
a part of the substitutions common a divisor of both i\ and r.
to the
two groups
G and
l
45.
We
proceed next to determine and tabulate the groups

cr,
which belong to the various values
of
<?,
(i
1, 2, 3,
/>).
The group G = G
of
<?i
contained the substitutions

3)
. . .
Sl
= l,S.,,S
S,..
The value
<f.,
of c
l
was obtained from

tr.
c^
.
by the substitution
If then
first
0y,
c?
2
con
suc
2 sequently cessively the substitutions

cr.,
ff.r
converts
back into
ff.,~\
^
2
,
we apply
to
sa ,
ff
the
of these will convert

will
into
<f l
cr
n the second
cr.,.
will leave c^
It
unchanged, and the third

cr_,
con
vert
back into
every substitution of the form 2 8 a ff.,, (a second line of our table we take therefore
ff.
appears therefore that ~

l
is
. .
1, 2,
unchanged by For the r).

.
We
can then show that this line contains

.
all
the substitutions
For if be any substitution which belonging to the group of <p.2 T <j.r will convert c? 2 into 
Consequently the
c?
1
substitution
it
ff.r
a.,?a.r
n and we may write

T(T 2~
equal
l
to
sa
tf2
sa
it
follows r
is
(ff.,Tff 2
)ff 2
to the group of But from the equation as was asserted. ff,r s a
belongs
l
ff.>,
Again
easily seen that all the substitutions of the second
line are different
from each
other.
For
if
it
follows that
We may
note however that the substitutions of the second line

first.
are not necessarily different from those of the.

identical substitution is of course always
In fact the
common
(Cf.
to both
and
oft jr
substitutions
may
also occur in
common.
50).
the three properties of the second line obtained above, it follows that the r substitutions of this line form the group of ^ 2
.
From
We
will denote this
group can
also
group by be shown formally
G.,.
;
That these substitutions from a

for
<T (fff- S a 2 )
l
ALGEBRAIC EELATIONS.
)
(
.
~
(ff.2
. .
.
~
Sp
ff.
2)
ff.
S a (ffff.r
S ft
<T.
<r
-\S a Sp]
ff.,
so that,
if s a
S0,
form a group, as was assumed, the same

all
is
true
of the
new
substitutions.
Similar results hold for
the other values, ^ 3
$P 4
<p ft
of
95,
and we have
therefore
Theorem
function
c?
VIII.
<7
If the values 9^,

<p
proceed from
3
,
. . .
cr ., pp o/ a p-valued 2 by the application of the substitutions

,
.
l,*r.,,
<7
Mm
-
the groups
G G21
l ,
Gp
of
c,,
%,,...
cr
respective ly a re
S,
1, A,
Sj, ,
S,.J
^]
^P
Gi* P
46.
The
functions c^,
cr. 2 ,
<p p
are
of precisely
the
same
form and only

the
differ in the order of
enter into them.
arrangement of the x^s which Such functions we have called ( 3) similar or of
same
also
type.
Accordingly the corresponding groups

tJie
G lt
G.,,
...
must
produce the same system of rearrangement of the elements x\ and only differ in the order in which the elements are numbered. This is clear a
be simitar or of
same
type, that is they
priori, but
i
we can
6-
also prove
fact
it
from the manner of derivation of

that not only the groups
sa
~i~â 7 from
i l ,
and in
we can show
G>
but also the individual substitutions

is,
and G~
sa v
are
similar; that
cycles,
these two substitutions have the same
each containing the same number of elements.

1
number of The process
of deriving the substitutions
tfY~
-V7V from the

1
s a s is
called trans
formation.
The
<r i9
substitution
^"
,s
a ^.
is
called the
conjugate of s a
the conjugate of
uiith respect to
and similarly the group

<r,-.
is
GI with respect to
as above,
We
shall denote this latter relation frequently,
by the equation
l
To prove the
(-ic
similarity of s a
and
ff~ s a ff it let
sa
us suppose that
<r,.
n
,
a? 2 a? 2
,
a? a
is
any one of the cycles of

a?/.,,
.
and that
replaces
x
ff
... fa
by
a*,-,,
a"
/a
so that, in
the notation
A of
^ 22,
may be
written
50
Xi
X2
...-.!
Then the
ively
factors of the substitution

a-,, x.,,
o*/ a
ff
s a <T,- will
replace
a?
/3 ,
tl
success
by
. .
av, similarly
.
is
will
be replaced by
T,~ s a <r
l
aj,- 3
by
a?,4
and
(.r M
l
finally
,r/.,
.
obtained from the corresponding cycle ) by regarding this cycle, so to speak, as a function (x In the of the elements a\ and applying to it the substitution &,-.
x, a ),
is x.2
. .
.
by a*,^ and this
Accordingly
contains
the cycle
a? a
of s a
same way every cycle of fiT s a proceeds from the corresponding The two substitutions have therefore the same number cycle of s a
/J
/
.
of cycles, each containing the
same number
of elements, as
was
to
be proved.
Example:
The function of four elements
has,
as
8.
we have
The
seen, three values,

<r
and
and
<>
its
~=
group
2
is
of order
substitutions
= (x x
2
a)
?>
= (x x^
which are not
contained in
{ ,
convert
into
respectively.
By
2 ),
transposition with respect to
rr,
and
ff
we obtain
from
G = [1,(^!X
l
(x 3 x t ), (x
the two groups belonging respectively to c 2 and ~

6r 2
<f. A
ff.
= Giffz
[1, (ajâ),
(x 2 x 4 ),
(aj
[1, (x,a- 4 ),
(a? 2 a? 3 ),
(x
47.
Corollary
I.
// a
gr?-owp
o/ substitutions
is
trans
formed with respect to any substitution whatever, the transformed substitutions form a group.
Corollary
s a sp
II.
The
two,
generally different,
1
substitutions
and
Sps a are similar.
For
s a sp
^s/T
(sps a )sp..
s a sps a
51
Corollary III.
with respect to
sa
The substitution
is
conjugate to Sp
l
.
Corollary IV.
be such that
sa
,
its
If the substitytion s a is of order r and if Sp o^ and no lower power occurs among the powers of
1
furthermore the conjugate of Sp with respect to s a is a power of Sp, then the smallest group containing s a and Sp is of order
if
and
r.
(cf.
37, 38,
Chapter
II).
Corollary V.
Corollary VI.
group into itself
All substitutions which transform a given

,
substitution s into its potcers, s a
form a group.
All substitutions which transform a given
form a group.
If two substitutions, or two groups, are similar, substitutions can always be found which transform the one In the case of two substitutions the transforming into the other.
Corollary VII.
substitution is
found
at once
from
46.
In the case of groups,

other.
ice
have only
to
construct the corresponding functions
and determine
the substitution c i
which converts the one into the

a
Corollary VIII.
tution are similar when,
greatest
47.
Two ijowers s and s& of and only when, and

s.
the
ft
same
substi
have the same
common
divisor with the order of
We
turn
now
to a series of
developments relating to the
existence of certain special types of groups analogous to those of

39,
Chapter II. Given any group
of order
let
cr
t
g, let
HI
of order h
and K^ of order
all
&!
be subgroups of G, and
and
<!\
be functions belonging to H^
and
respectively.
These functions, on being operated on by
the substitutions of G, take respectively -~-= /j,,and-|

"i
A-
values
#1
in all
43, Corollary)
9"u
let
these be denoted by
9"/
<f<,
and
</
,,
^.
.
()
The group
where
<r
of
a is
a = (? a any one of the functions c a will be any substitution of G which converts ^ into c v
j
K^
<r
a,
We
form now the entire system of values a

c- A
+b
(;.
= 1,
2, ... /v,
/x
= 1, 2,
... *
),
52
-
where a and b are arbitrary parameters, and divide these

tions into classes, such that
all
func
/i
i
"i
the functions of each class proceed
from any one among them by the operations of G.

If
all
a <?
-j-
b 4 a be one of the values above,
and
if
we apply
to
it
the substitutions of G, the resulting values are not necessarily all In particular some of them may coincide with the given distinct.
value.
The number
of the latter
:
is
tutions
common
is
to
and
~
a
<r
K^ a
l
equal to the number of substi Let this number be d a

.
Then
all
the g values arising from a c

easily seen.
.
-|-
<
will coincide in sets of
d a each, as
obtained
is
The number
of
distinct values
thus
therefore
da
all
If these
do not exhaust
the possible values a
<f\ -f-
^^
let
#Pr-h& ?V be any remaining

this belongs also
value.
Then
in the
same
class with
<f ff(r
-f-
T(r
= a Vi 4" b
is
j3.
From
the latter
value
we
can, as before,
deduce a class contain

the
ing in this case

tutions
distinct values, to
where d$
1
number
of substi
common
and K$ =
<rp~
<rp
Proceeding in
this
way, we must finally exhaust

<f\ -j-
all
the
L
-.1
yI
val-
ue of the functions a
</v
Writing then the two numbers of

have, after dividing through by g
,
values equal to each other,
we
:
A)
where
_JL
h, k,
d,
^ +1+
d,
is
^d +1. m
a
<f
denotes the
number
G.
of classes of values of
ft,
\-\-b
,
</v
with
respect to the
write fa
group
Since
a multiple of every d a
we may
=/a
and consequently
^
where the/
--=/,+/,+ ...+/,
*
s are all integers.
if
From Theorem V of 43 it follows as a special case that 49. a group contains a substitution of prime order p, the order r of
("I.
* Formulas A) ;tud B) were obtained l>y t;. Frobenius, Crelle extension of a result given by the Author, Math. Annalen Xtlf.
p. 281. as
an
MULTIPLE -VALUED FUNCTIONS
53
the group is a multiple of p, and if a group contains a subgroup of order p a where pis a prime number, the order of the group is a mul
tiple of
p a By the aid of the results of the can now also prove the converse proposition:
.
preceding Section we
Theorem
ber
which
is
If p a be the highest power of the prime a divisor of the order h of a group H, then
IX.
num
con
tains subgroups of order
a
.
In the demonstration we take for the

tion the symmetric group, so that
of the preceding Sec
g = n\.
For
we
take the pres
ent group H, and for
the group
of order p f of
is
39,
Chapter
II,
being the highest power of
p which
a divisor of n\.
The
formula B) of
48 then becomes
=/,+/,
+
is
...+/..
no longer
divisible
The
left
member
of this equation
byp; con
also not di-
sequently there must be at least one fp

visible
=
,
^
a/s
which
is
a and Kp, by p and therefore the latter being a conjugate of K, have exactly p a substitutions in common, f These form the required subgroup of H.
by p
that
is
dp
is
divisible
Corollary. At the same time it appears that the group K con For tains among its subgroups every type of groups of order pv.
we need only take any group of order pi for
stration. 50.
H in the above demon

:
The
last
theorem admits of the following extension
Theorem X.
pP, then
If the order h of a group contains subgroups of order pP.

follows at once from
is divisible
by
The proof
proved
Theorem XI,
as soon as
we have
Theorem XI.
group of order
Every group
of order
pa
contains a sub
a~
l
.
The
corollary of the preceding Section permits us to limit the

Joe. cit.,
*Cauchy,
proved this theorem for the case a

it,
1.
The extension
to the
case of any a was given by L. Sylow, Math. Annalen V, pp. 584-594.
tFor every subgroup of order a power of p.
and consequently every subgroup of K$, has for
its
54
a as subgroups in proof to the case of groups of order p which occur the group of f there obtained 39, Chapter II. The group
K K
i
was constructed by the aid

1,
of a series of
. .
subgroups
40)
K.2
K\ KX + 15
,
Kf
..
of orders
1,
p\
ir,
...
A p\ p +
1 ,
p f ~\
every one of which is contained in the following one. If the group occurs in this series, the theorem is already proved; if not; then let be the lowest group which still contains H. K^ then does not -BÂ ^
i
contain
of
6?,
all
the substitutions of H.
We
apply now the formula B)

in the place of
48 to the groups
1 ,
K\, and H, taking these
and H.
We find
P-=/i
+/+
= p.
5
..+/.
l
This equation has two solutions, since

h
_//*,
are powers of p: either
the/ s, being divisors of we must have f =/ = ... =1 and

2
m ==p,
that h
or else
m=1
e.,
and /,
In the former case
it
would follow
is
=d
lj
i.
that
is
a subgroup of K\, which

*.
contrary to
hypothesis.
Consequently /i = p
e.
H and K\ have a common sub

1
group
of order
J
.
To Theorem
X we
the required group. * can now add the following
This
is
Corollary.
a Every group of order p pf -p 2 a2
canbe con
structed by the combination of subgroups, one of each of the orders
PMVSPA
A
smaller
...
number
of subgroups is of course generally sufficient.

is
further extension of the theory in this direction
not to be
Thus, for example, the alternating group of four ele ments, which is composed of the twelve substitutions
anticipated.
has no subgroup of order 6. 51. We insert here another investigation based on the con
struction of tables as in
41.
l ,
Let
H be a group of order h affecting the n elements x

p.98!i-4.
x2
xn
*G. Frobenius: Crelle CI.
MULTIPLE-VALUED FUNCTIONS -- ALGEBRAIC RELATIONS.
55
,
From these n
and
let
elements we arbitrarily select any fc, as x x>, xk be the subgroup of which contains all the substitu
1 ,
tions of the latter that

to be the order of
tions.
do not
affect a^,
t>,
.
a? 2
xk
and ^
=1
Suppose h
th
to
be
its
several substitu
We
proceed then to tabulate the substitutions of

sa
H as
Then
fol
lows:
Given any substitution
of
,
H, suppose that
this
converts
all
l ,
x.2
xk
into
x ai
Xa.2
alc
in the order as written.
the substitutions
o
/ o to
J &
,
C-30 a
/ l
G
ii
also convert
x x
1 ,
...
fc
into
x ai
a?o.,
^.
respectively,
this effect.
and these
are the only substitutions of
which have
We
take
these various sets of h substitutions for the lines of our table, which
is
accordingly of the form
T)
The
substitutions of the table are obviously all different,

/JL
and conse
contains
quently
=h
Again, suppose that v is any substitution of among its cycles one of order k, say
1
H which
(1)
M
all
=. (X-iXtfCz
X k ).
Then
(2)
the (necessarily distinct) substitutions

Vl
,
t 2 v ly t.v,,
...th >v
will contain the
tions of
H which contain this cycle. We wish now to determine how or any cycle many substitutions of H contain either the cycle
(1)
same cycle
(1)
and these
will
be the only substitu
obtainable from (1) by transformation with respect to the substitu

tions of
(la)
H, say w a = (x ai X a X a
.
3 afr).
Now
since all the substitutions of the
same
line of T) convert
the elements of the cycle (1) into one and the same system of ele ments, it follows that if we write
56
where each
tained in v
s a is
gates of the cycle (1)

l ,
the same as that of the table T), then all the conju are con which occur in the substitutions of
H
a
v.2
contained in v a not affect

jc
Suppose the notation so chosen that u> a is which do If now we denote the substitutions of
.
VM
ai ,x a 2
.
aic
i.
e.
those of the group s a

J
Hs =H
by
/ V (a)
&2
/ (a) 1 3 >
ia
/ 1
(a)
1>
and by right hand multiplication by

(3)
form the
.
-
line
va
(3)
t.^v a
* 8 (a) Va,
fj,-Ua,
then
(jj
contains
,UL
all
H which
.
involve the cycle
a.
The
lines of the following table

Ui,
to,
to,
..
f//^,
the required substitutions. The question then arises, how many of the lines of T ) give the same cycle; for example, in how many lines the cycle (1) occurs. If
therefore contain
all
v 8 T then S T must permute the elements and must therefore contain a power of (1) Xj,. cyclically, #!, as a cycle. Consequently we must have S T equal to one of the sub
this cycle occurs in v T
a? 2
,
.
. l
=s
stitutions
y u v*,v*,
v*~
v*,
or one of these multiplied
by some
t v.
But the substitutions v }1

table
"
belong lines of appears therefore that the sets of k each. We have then
. .
.u/
to different lines of the
o ).
It
fj.
T)
coincide in
Theorem
occurs in
rise
(2),
XII.
Every individual
cycle of
order k which
and consequently every one which

to all the
by transformation with respect

cycles.
H, gives h substitutions of H,
k which occur
oc-
occurs in
to
K
j
The h distinct conjugate
cycles of order
in
H therefore give rise to -K = K

is
cycles.
The number of letters
curring in these cycles

this it follows that the
is
therefore equal to h, the order of

letters
H. Prom
number of
in all the cycles of order k

is
a multiple of
sets of
the order of H. The multiplier non- conjugate cycles of order k.
the
number of
- ALGEBRAIC
elements
RELATIONS.
57
Corollary.
of H.
The
number of
which remain un
the order
changed in the several substitutions of

substitutions of
is
a multiple of
If every element can be replaced by every other one by the
H,
this
number
is
exactly equal to the order of
EXAMPLE.
Consider the alternating groups of four elements,

(x 4 ),
(x.x,) (x 3 x 4 ),
(x,) (x.2 ) (oc s ) (x,)

(x,)
(x,x
(a? 3 )
(x 2 x.A x,\
(x 2 XtX s ),
(x 2 ) (x^Xt), (x 2 ) (x&Xa),
of
(x 3 ) (XiXiXt),
(x,)
is
(x^x^).
12,
Here the
number
cycles with one
element
which
is
equal to the order of the group.

occur in cycles of the second order
The number
is
of elements
which
3,
also 12.
But, for k
the
number of elements is 24 = 2 shown that the group permits
12.
Correspondingly
it is
readily
of replacing any element
by any
other one; that the cycles of order 2 are all conjugate; but that the cycles of order 3 divide into two sets of four each
:
the second set being non-conjugate with the first. 52. return now to the table constructed in
We
45.
This
;
table did not possess the last of the four properties noted in
41 the
substitutions of one line were not necessarily all different from those of the other lines. For every group certainly contains the identical
substitution
1,
which therefore occurs
f>
times;
and again
in the
example of
46 three other substitutions

(ajjOJa) (a? 3 a; 4 ),
(x^)
(x2 x),
(x.x,} (x 2 x,)
occur in each of the three groups. general ivhen it is possible that one
We
and
have now to determine in

the
same
substitution shall
occur in all the groups

several values
just cited
G
.
,G 29
of
Gp
^2
(p p
<p.
We
belonging respectively to the shall find that the example
is a remarkable exception, in that there is in general no substitution except 1 which leaves all the values of a function un
changed. f
*
article
In connection with this Section C/. Frobenius, Crelle CI. by the Author, ibid. CIII p. 321.
p. 273,
followed by an
tL. Kronecker:
Monatsberichte
d. Berl.
Akad.
1879, 208.
58
If
we apply
<r,
to the series of functions
^2
y> p
any arbitrary
substitution
we. obtain
fai fa* d")
^0-30-)
Vp<r
These values must coincide, apart from the order of succession, with the former set, for $0, ^._,, ... ^ p are all the possible values of and
, >,
the
/>
values just obtained are
all
different
from one another.
The
groups which belong to the
latter
T~
6r 1
<7,
<7~
2 <T,
<T~*G3
(7,
<T~
Gp
ff
are therefore also, apart from the order of succession, identical with
6rj
,
6r 2
6? 3
6r p
that
is,
the system of the
6r s,
regarded as a whole,
is
whatever.
unaltered by transformation with respect to any substitution If now we denote by the group composed of those
substitutions which are
common
latter
to
G.2
Gp
then
}
H
l
is
also the
group
.
.
.
of those substitutions
Gp<r.
which are common

is
to <7~
<r,
v~
2 ff,
<r-
group H<r, consequently we have

<rl
ff~
But the
also of course expressed
by
Hff
= H\
that
is,
the group
H
it
is
any substitution;
includes therefore
it.
unaltered by transformation with respect to all the substitutions which
are similar to any one contained in
We
which
character.
of this proceed now to examine the nature of a group We consider in particular those substitutions of
affect the least
number
of elements, the identical substitution
excepted.
It is clear that these
can contain only cycles of the same
number
of elements, since otherwise
some
of their powers
would
contain fewer elements, without being identically 1. We prove with regard to these substitutions, first that no one of
their cycles can contain
more than three elements.
For
if
H con
tains, for example, the substitution

S
= (X X X X
l
...)...
l
and
if
we
take
<7
= (x x
z
t ),
then, since <T~
Hv = H,
the substitution
will also occur in
H.
Now
x^.
is
only differs from
s in
the order of the
two elements x3 and
also occur in H, since H
Consequently their product,

a group,
which must
MULTIPLE -VALUED FUNCTIONS
- ALGEBRAIC
,
RELATIONS.
59
s-s
will certainly not
=(x
3)
(x,x,. ..)...
affect the
it
element # 3 but cannot be the identical

. .
substitution, because
contains at least the cycle (x^x^
).
This
product therefore affects fewer elements than s l which

hypothesis.
is
contrary to
Secondly we prove that,

affects the least
cycle.
if
> 4,
the substitution of
which
form
number
of elements cannot contain
more than one

of the
.
.
For otherwise
Sa
( Xl
would contain substitutions

.
X 2 ) (XzX)
8 ft
= (X.X.X.) (X^XQ)
and therefore the corresponding conjugate substitutions with respect

to T
x4x 5
Consequently the corresponding products

l
SaT s*
= Oi)
(x- 2 )
(x 3 ...)..., SfT
V=
(a?,) (o- 2 )
(x a )
which are not

.ZJ,
1,
but affect fewer elements than
s,
must
also occur in
H H contains a substitution the second case H must contain

or
which would again be contrary to hypothesis. If then n > 4, either consists of the identical substitution
(oâ?,,),
1,
or a substitution (x^x^x^.
In
all
the transpositions, that
is
is
the symmetric group. In the third case must contain all the cir cular substitutions of the third order, that is is the alternating
group.
(Cf.
34-35).
G?j,
to the group (7, it appears that if Returning from the group 6r2 Gp have any substitution, except 1, common to all, then
,
. . .
either the second or the third case occurs.
H, which
is
contained in
the alternating group; G is therefore 6r, 1. either the alternating or the symmetric group, and /> 2, or /> 4 we might have, beside s If, however, n 1, another substi
includes in
either case
tution
in the group.
With this
its
conjugates, of which there are only two,
must
also occur.
The group
H cannot contain any further substitu
tion without
We
becoming either the alternating or the symmetric group. have then the exceptional group
60
H =[ =
Sl
l, So,.s 3
,s*],
and
this actually does transform into itself with respect to every
sub
stitution.
II, that
Beturning to the group
it
follows from
43,
Theorem
the order of
is
a multiple of that of
from Theorem II the order of The choice is therefore restricted to the numbers 4, 8, 12, and 24. The last two numbers lead to the general case already dis cussed where ^=2, or 1. The first gives G = H, = 6, and for
of 4; again
4!
H, that is, a multiple G must be a divisor of
= 24.
/>
example,
cj
cr
:
,
= (x,x. + x^ = (x x -f x x
2
}
(aj,a-3
4)
# 2 aj 4 ), c? 2 = (x,x 2 -f x x,) = (Xx3 -f o?2 a?4 ) (x rx 4 -f- ^ x 3 ), ? 4

-h
z 2
(x x,
}
+x x
2
3)
fax, -f o?8 a? 4 )
In the second
case, r
= 8, G contains H
to
G we
must add other substitutions
can be cyclical of the third order, for r = l2 or 24. If we select any other is included among those treated in of which 46, group
this
To obtain those of H. None of these in this case we should have substitution, we obtain the
as a subgroup.
39.
For
group
c\
f>
= 3, and,
-\s
for example,
,
= Xl x,
x x4
</> 2
= XjX
4-
a? 2
x4
03
= x^
-f- x.2
x3
If ^ 4 there is no function, except and symmetric functions, of which all the p values are unchanged by the same substitution (excluding the case of the identi cal substitution}. Ifn = 4:, all the values of any function belong ing to the same group ivith
the al
Theorem XI.
ternating
= (x x, -h x.x,)
}
(x,x 3 -h x,x,) or with $
= x,x
-f
xs x,
are unchanged by the substitutions of the group
H=[l,
53.
fax.)
(a? 3
ir 4
),
(aâjg) (x.2 x),
(x,x
We
have thus far examined the connection between the
values of a /^-valued function from the point of view of the theory of turn now to the consideration of the algebraic substitutions. relations of these values.
We
We
tem
saw
at the
of values y l1
<f> 2 ,
beginning of the preceding Section that the sys v p belonging to a function 9 is unchanged
.
. .
MULTIPLE- VALUED FUNCTIONS
- ALGEBRAIC
RELATIONS.
61
as a whole by the application of any substitution, only the order of All integral symmet succession of the several values being altered. cr are therefore ric functions of tr,, <p n 2 unchanged by any substi
,
.
.
tution,
and are consequently symmetric not only
in the
cr
but also
in the x^s.
They can therefore be expressed as rational integral

CA
functions of the elementary symmetric functions
of the x^s.
If
we
write then
(Vi)
?i
+ F* +
+V =
p
î fa
c,, ...
O,
i,
,,
.,,,
the
jfiJ
are the are
coefficients of
roots.
an algebraic equation of which
^ ne
Theorem XII.
CP
//ie
The
f>
values
cr._, ,
y- p
of
[i-
valued
/>
integral rational function y are the roots of
an equation of degree
R^P
+ ^^c
l
...
^ -^
P
coefficients
of which are rational integral functions of the

,
ele
.
mentary symmetric functions

51.
e3 ,
c,.
of the elewntN
,r,
,r.,
,
ar f4
As an example we determino the equation of which the
throve roots are

c-,
= (K&., 4,
,r ;r r,
c,
,r r r, -j-
,r,^ 4
,M + ^ x
4
2
where,
*c 1} .r 2
a- 3
.x 4
are themselves the roots of the equation

4
f(x)r=X
-C^ + CjX? -C 0?-(-C4 =p.

3 r
2
We
find at once
f
anci again,
+V
-f-
fa
S (x x z )
}
c.2
by
10,
Chapter
i
I,
6-c 4
V lâ
+ Wa + ^f f = S (xôXa) =
coefficients
,
/
-f ,-c^ -f
C2
2
.
The numerical
amples.
^,
y are readily
0.
found from special ex-
They
are
4,
/3
= 1, f
2 2
Hence
Finally
c^c..-.,
= So?
2
1
a?2 a;8 )
+
is
ar!a: 2 c a a; 4
S(a?1
Accordingly
tlie
required equation
62
/(cr)
<p*
ctf?
-f (0^3
4c 4 )^
(c,
c4
4c.2 e 4
+c =
2 3
0.
We
roots.
examine the discriminant of
this equation,
i.
e.,
of
its
three
To determine
this function
it is
not necessary to employ the

I.
the general formula obtained in
10,
Chapter
We
have
at
once
and consequently, if we denote the discriminant of the and that of the x\s by J,
J =(?i 9
2
>
by J^
^ 2)
2
fe
(a?!
^) 03
Pi)
=
We
(a?
a? 2
2 a)
(X
x4 )
(a? 2
a? 3 )
(a? 2
a? 4
(a? 3
x4 )
= J.
observe here therefore that the discriminant of an equation of the fourth degree can also be represented in the form of the dis
more important criminant of an equation of the third degree. consideration is that the special theorem here obtained can be gen
eralized in another direction, to
which we next turn our

41.
attention.
55
We
start out
from the table of

i.
If y
l
is
not single-
valued, the
first line
of this table,
does not contain
all
e., the group G belonging to ^ the transpositions. If a transposition (x a xp)
occurs in the second

Cj =
is
line, it results
<?%.
from the construction of the

If
table that (x a xp) converts c^ into

cr
2
therefore
-it
&a =
a?0,
then
also,
and consequently
oc a
?,
c2
since
vanishes for
(l
= Xp,
divisible
If,
by
xp.
of
<f l
then, any transposition (x^Xp) does not occur in the group G one of the differences ^ cr A (/ = 2, 3, ... (>) is divisible by
a factor of the form
xa
all
/.
xp.
t
Now
first line
there are in
transpositions of
n elements.
If the
of the table,
-
e.,
19
contains exactly q of these, then the

q.
other lines contain
The product
(<p l
<f^(<f\
f,-,)
(c^
cp)
is
therefore divisible by xp, and therefore by
^=
q different
factors
of
the form x a
their product.
c^
1 ,
have taken
Instead of starting out with the value cr 2 Since the group G z ^2~
.
we might
fT
equally well
y<>
belonging to
cr^,
is
- ALGEBRAIC
p p)
is
RELATIONS.
63
similar to the
pi)
group
(cr 2
l ,
it
also contains q transpositions,

(Pa
and the
q
i
product (p 2
factors of the
p3)
xp.
also divisible by
if
form x a
<PP
.
The same reasoning holds
we
start
with
p4
If
we multiply
the separate products together,
we
find that
_
^
JL\
-J
[^/^ ^
since J^
is
factors
â
xp.
But
xp
symmetric
in the X\B, the
presence of a factor x a
x$,
requires that of
every other factor
xy
and consequently of
Suppose that J
is
II (x a
#0)% the discriminant of f(x).

is
the highest power of J which
contained as a factor in Jo, then, as

contains
J contains n(n 1) factors x a Xp, and consequently J n(n T)t such factors, we must have
=
The number
f
n(w
1)
/
can be
only
.
when
y
q
is
= -~
Y)\ ) l
/
_ I
-,
that
is,
when
all
the transpositions occur in
Again q can be
group or one of
only
G when G
l
then symmetric and // = 1. contains no transposition. One of

is
the cases in which this occurs

its
that where
is
the alternating
subgroups.
//
<f
Theorem
x
l
,
XII.
is p -valued
function of the n elements
a- 2
xn
the
group of which contains q transpositions, the dis

j>
criminant J
of the
values of
<f
is divisible
by
If
of
<f
is is
not symmetric, the exponent of J
is
not zero.
is zero.
If the group
(f
contained in the alternating group, q
All multiple-valued functions therefore have some of their values coincident if tico of the elements x\ become equal.
04
perceive
We
functions
50.
now why it was impossible ( 32) to obtain w!- valued when any of the elements ,r A were equal (Cf. also 104). Returning now to the equation ( 53) of which the roots
tf
are
?iPi
? p_JJiV5 p-l 4. ft^p-4 -_-...
ft
= 0,
we endeavor
this equation
to
determine whether and under what circumstances

i.
<
j
can become binomial,

th
/>
..,
whether there are any
//-valued functions whose
powers are symmetric.
For
/>
= 2,
we
already
general case
know that cr \/ J we will assume
satisfies this condition.
that, if
,
In treating the the required function contains

<f>
any factors of the

outset.
form \f J these factors are

c
-,
all
removed
at the
If the resulting quotient is
so that
then, since
also.
<F~P
and (\/ J )~ p
n*
both symmetric,
"
is
symmetric
We
write then
be any root of this equation, and of unity, then all the roots are
If
</ ,
<r
if
M be a primitive
lb
(2//)
root
l,
"><! i,
<
W *P
l,
and consequently
j,^,"r
V
1
(w
"
(i--^(i-^)
2
. . .
(>-
a
,
r.
From Theorem XIII

is
this discriminant
must be
divisible
by
J,
unless
But the factors containing to are constant symmetric. and therefore not divisible by J, and by supposition t^ does not con
itself
tain \f J as a factor.
according as
is
Consequently odd or even,
</
is
symmetric, and
we
have,
XIV.
the
e/ ric
on.it/
//
//ze
?/
c.lnnc.ntx
.rj,.r.,.
y n are
inde-
unsymmetric functions of which a power can be
are I he alternating functions.
57.
On
add
in the present
account of the importance of this last proposition we and following Sections other proofs based on
entirely different grounds.
If
ut
- ALGEBRAIC
and
if
RELATIONS.
65
is
a primitive
th
/>
root of unity,
is
any root of the
equation
<f>
= 0,
then
all
the roots of this equation are
Since the w
are constants,
this
all
the values of
 n cases give /> All the values of a
function are of the same type, and consequently there are substitu tions which transform one into another. Suppose, in the case
p
= n\,
that
ff
converts the value
^
.
. .
into ^ 2
= w^;
then
^ converts
?i into
wVi- Accordingly the

1,
<7,
series
2
<T
,
tf"
-1
all
consists of distinct substitutions,
which therefore include
the
possible substitutions.
n and order n The case n

!
Such a substitution does not
must therefore be a substitution of degree exist if n > 2.

In
this case the
=4
furnishes no exception.
group
common
XIII)
to all the values of ^
might be the special group (Theorem
G=
<P
[1,
(x&t) (x 3 Xt), (x&s) (x 2 ^ 4 ),
(x^
4)
(# 2 a? 3 )];
tution n which converts î into coy
would then be a six-valued function, and there must be a substi and which is of order 6. But
l
there
is
no such substitution in the case of four elements.

Finally
58.
we
give a proof which
is
based on the most ele

to
mentary considerations and which moreover leads extension of the theorem under discussion.
In the
first
an important
place
we may
if
limit ourselves to the case
where
/>
is
a
it
prime number. follows from
For,
^^p.g,
where p
is
a prime number,
?*=:(?)=:
that there
is
also a function
^ of
which a prime power, the p th
is
symmetric.
If,
accordingly,
th
,
we denote by y
a function of which a prime

itself is
power, the p
is
symmetric while y
unsymmetric, then,
66
(
since the group of y cannot contain all the transpositions
34),
suppose that
it
ff
= (x a ocp)
converts ? into <?, where

1 7> 7>

is
a primitive
th
root of unity.
If
we apply
that

we
obtain the
ft
new equation
==?,.
cr,
Multiplying these two equations together and dividing by
cr^
we
have
and consequently, since p is a prime number, p = 2 and e> = S \/ J Having shown that only the alternating functions have the prop
erty that their prime powers can be symmetric,
we may next exam
ine whether there are any functions prime powers of which can be
two-valued.
Suppose that
c"
is
multiple-valued, while
its
th
power
is
two-val
Then there is some circular substitution of ued, q being prime. the third order ff (xaXpXy), which does not occur in the group of 4\ if this since, group contains all the substitutions of this form, it
must be the alternating group

but that
35).
Suppose, then, that c/v^^ n
since
cv/ ,
being a two-valued function,
stitution of the third order.
We
unaltered by a circular sub must therefore have

is
and must therefore be a primitive q th root of unity. If we apply to this last equation the substitutions and <r, and remember that 
is
not
ff
Multiplying these three equations together and removing the func tional values. \ve have
If
---
c
67
all
now we assume n
> 4,
then the group of

fifth
cannot contain
the circular substitutions of the

II).
order,
(Theorem X, Chapter
group of
(,
If T is one of those not occurring in the
then
and consequently,
if
<u
be a q th root of unity different from
1,
0T=01.
It follows
from
this, precisely as
above, that, since r
= 1,
5 = 5. and consequently w 1, g But this is inconsistent with the
first result.
It follows there
fore that
is
not greater than
4.
Theorem XVII.
independent quantities.
59.
If
n
is
>
4,
there
is
no multiple-valued
function a power of tvhich
two-valued,
if the elements
x are
We
conclude these investigations by examining for w5i4
the possibility of the existence of functions having the property dis cussed above.
In the case n 3 we requires no consideration. undertake a systematic determination of the possible functions of
The
case
n=2
the required kind.
We
begin with the type
and attempt to determine /5, Y so as For this purpose we make use tions.
,
to satisfy the required condi
of the circumstance that

so that
-
some
<T
= (x^x^x-^
converts
<f
^
-\-
into tay
3
l
(w
= 1)
v = axJ
a>a
fas
+ Y&I
=:
w(aV
-f-
^M =
;
The
of
.
last three
We
equations can be consistently satisfied for every value 1 and therefore may take a
Pi
=a?1r 4-Va^r
r a>a?8
is
a function of the required type.
This result
is
confirmed by actual calculation.
We find
68
(x*
+ x? +
*.
a-a
+ 6 aY>
<
3
2
a- 3
-y-
(xfxi + av ay + aY^ +;...)

r V ^3
-f-
/^/
3 ( V -> nr O^.Xj 1^2
v *
r nr r -\- /y 2 tX
^3
-2
w *^
2>
V ^1
4- ^3 T2
"^1
T"
T ^3
2;
> ^2
">
If r
=1
the result becomes simpler, in that the last parenthesis

to
becomes equal
whereas, in general, this parenthesis
is
only a rational function of
*J J.
If
we
write, as usual,
we
have, for r
= 1, ^=i
2 Cl
9 Cl c 2
+ 27 c
3J
Suppose now that

in
n = 4.
type axf -{- fixj -}- yxt
-}-
It is obvious that a function of the
dxf,
which every term contains only one element, cannot condition of being multiplied by at when operated on by a
satisfy the
We
type
(x^x^). enquire then whether the required function can be of the
where every term contains two of the four elements.

also fail to satisfy
If this type
the requirements we should have to pro should ceed to more complicated forms. shall, however, obtain a posi
We
tive result in the present case,
and
in fact
we may take
= 1,
so that
From
the condition
<f ff
= wa, Y
j t
= we have the series of = w-f = o a, a = wfi = =

w<f> lt
2
equations
3
<JD~Y
co
a.
,
All of these equations are satisfied independently of the values of a
and
! ,
and we have
But
is
again, the substitution r
= (aâvc
^
4)
converts
c>]
into ^ T
where
equal to the product of
by some cube root
of unity, since
(f,^
- ALGEBRAIC
is
1,
RELATIONS.
69
= (p*.
 1
a^Xz -f- X
c"
(toax 2
+ arax + oja^),
i
1
and since the terms
and
which contain x x i are
a^Xi and aô?^,

the cube root of unity by which
<y
2
.
ip l
is
multiplied must therefore be
From
the equation
(f r
=^
j
cr
it
follows,
by comparison
of coef
ficients that
j ,
These two equations are consistent, since w

arranging according to the
(fl
1.
Putting
= 1, and
).
powers
of w,
we have then
a? 2 ar 3
( Xl
+x
3 a? 4 )
-f atfaxs
+ ^2^4) + ^(a?!^ +
The function
a function which
c^ is therefore a
combination of the three values of

discussed.
2
we have already
(a? 3
a; 4
),
The group
of c^
is
G = [1, (XM)
That
3
c",
(x^ (x x),
(x^)
if
(x,2
x 3 )~\.
is
two- valued
is
also readily shown,
we
write
For then
case n = 3
^
;
coincides with the expression obtained above for the
and since y
c 2 2/
2
t ,
y 2 yA are the roots of the equation

, ,
y*
(ciC 3
4c 4 )
?/
(c x c 4
4c 2 c 4 -f
c3 )
= 0,
where the
a?
c s are the coefficients of the

o? 4
,
n x
aj 3
54),
we
equation of which the roots can translate the expression obtained for
the four
=3
a? 2
directly into
^3? ^4) since
a two-valued function of
elements
#!,
we have
54) J
;/
= 4r-
CHAPTER
IV.
TRANSITIVITY AND PBIMITIVITY. SIMPLE AND COMPOUND GROUPS. ISOMORPHISM.

60.
The two
familiar functions
XiX-2
"|
XgX
XiX%
OCÔC^
from each other in the important particular that the group belonging to the former
differ
-(?
= [!,
feavj),
(ar 3
oj 4
),
(aâ?,) (a?3 o; 4 ),
contains substitutions which replace x^ by
o? 2
or
;r 4
while in the
group
belonging to the latter
G =
l
[1,
(avr 2 ), (0%), (a?^)
(a^)]
.
there
is no substitution present which replaces x^ by a? 3 or a? 4 The general principle of which this is a particular instance is the basis of an important classification. designate a group as transitive,
We
if
its
substitutions permit us to replace
any selected element x by

l
every other element.
Otherwise the group is intransitive. Thus 6r, above, is transitive, while 6^ is intransitive. It follows directly from this definition that the substitutions of
a transitive
element
s
xk
i
group permit of replacing every element X by every For a transitive group contains some substitution
L
= (x x tion =
l
x by a?/, and also some substitu which ...)... replaces x by x ,. Consequently the ~ also in which occurs the 1, product group, replaces x by x k The same designations, transitive and intransitive, are applied to
...)... which replaces
1
(x-^JCk
.s
functions as to their corresponding groups. It appears at once that the elements of an intransitive group For are distributed in systems of transitively connected elements.
example, in the group transitively connected.
above x and
l
x.>
and again,
a? 3
and
x+
are
Suppose that in a given
intransitive
l ,
group
transi-
there are contained substitutions which connect
x2
xn
GENERAL CLASSIFICATION OF GROUPS.

tively,
71
li
others which connect
a? rt
1 ,
x r(+
xn +
and so
on,
but
none which, for instance, replace x by x a + \ G^Ll), and so

l
maximum
tems
is
possible
,
b !,...,
if
on. The number of substitutions within the several sys and consequently the maximum number in the
. .
given group,
a, b,
are known,
is
a! b\
If only the
sum
b -f is known, the maximum number of substitutions in an intransitive group of degree n is determined by the following
=n
equations
(nl)\
(n
2)!
11
= ^^(n
A
2)! 2!
> >
(n
2)! 2!,
(n
>
3)
2!=-^p^(w 3)! Theorem I. The maximum

of degree n are
(n
1)!,
first
3!
(n
3)! 3!,
(n
> 5).
orders of intransitive groups
i(n
1)!, (n
2)!2!,(n
2)!, (w
3)!3!, (w
3)!2l,
The
the alternating groups of (n
two orders here given correspond to the symmetric and 1) elements, so that in these cases one
unaffected.
element
of the
is
The
symmetric group maining two elements. The fourth belongs either

tion of the alternating
of (n
third corresponds to the combination 2) elements with that of the re

to the
combina
group
of (n
2) with the symmetric group
of the remaining two, or to the symmetric
2) ele group of (n ments alone, the other two elements remaining unchanged; and so
on.
The
construction of intransitive from transitive groups will be

(
treated later,
61.
99).
We
proceed
tive
group in a table.
now to arrange The first line
the substitutions of a transi

of the table
is
to contain all
those substitutions
which leave the element x unchanged, each substitution occurring

l
From the definition of transitivity, there is in the given only once. For the second line group a substitution 2 which replaces x by x.2
<7. l
.
of the table
we take
72
We
#2
;
show
all
for every
that
unchanged the substitutions which produce
then, 1) that all the substitutions of this line replace x and ^ 2 converts x into x 2 s x leaves QL\
l
by
2)
this effect are contained in
this line; for if r replaces
x by
}
a? 2
r^"
leaves
x unchanged and
l
is
therefore contained
that r
for
sa
among
all
the s\s; but from
Tff. 2
= s\
it
follows
= sA
L)
/7
3) that
rr
the substitutions of the line are distinct;
if s a <r
s/3
;
obtain by right hand multiplication by ^.,~ *, S|8 4 that the substitutions of the second line are all different
2
,
we
from those of the

former do
not.
first;
for the latter leave
OL\
unchanged, while the
We
form
select
now any
substitution
^ which
converts
into
x and
:i
for the third line of the table
The
ties
all
then be shown to possess proper similar to those of the second. We continue in this way until
substitutions of this line
may
the substitutions of the group are arranged in n lines of
sub
stitutions each.
We have then
II.
l
Theorem
tive
If the number of substitutions of a transi group, which leave any element x unchanged, is m, the order r
is
of the group
mn,
i. e.,
always a multiple of
of this
,
n.
The following extension
theorem
.
is
readily obtained:
Corollary.
the group G,
If x
(l ,
x b xc
are
any arbitrary elements of

f
and if is the order of the subgroup of G ivhich does not affect these elements, then the order of G is mn where n is
,
the
number of
distinct systems of elements
x a ,xp.x y ,,.. by
,
.
.
ivhich
62.
.
G
is
can replace x a x b x c
, ,
group
said to be k-fold transitive
when
its
substitu
ones.
tions permit of replacing k given elements

It
by any k arbitrary
can be readily shown that any k arbitrary elements can then be The definition includes the case where replaced by any k others.
any number of the k elements remain unchanged. A AN fold transi tive group must contain one or more substitutions involving any
arbitrarily chosen cycle of the
fc
th
or any lower order.
Thus
in a
four- fold transitive

.Tj
and x2
group there must be substitutions which leave unchanged but interchange a*3 and # 4 and which are there,
73
also
;
form fa) fa) (x3 x) contain a substitution which leaves

fore of the
.
.
l ,
The same group must x a? x all unchanged

2
,
this
may
of course be the identical substitution.

three- fold transitive
If,
For an example of a
alternating group
stitution
group we may take the
of 5 elements.
x.2
for instance,
we
l
require a sub
which leaves
#3
respectively,
unchanged and replaces x and x- by x 5 and the circular substitution S = fax- x 3 ) satisfies these
two transpositions fax^) fax^), This same alternating group can not however, be four- fold transitive, for it must then contain a sub
conditions, and, being equivalent to the
belongs to the alternating group.

stitution
which converts x l9 x z
a? 3
into
x lt
x.2
xs x
,
respectively;
this could only
be the transposition fax & ), and this cannot occur in

ele
In general, we can show that the alternating group of n

ments
(n
is
always
(n
2) -fold transitive.
The requirement
that any
elements shall be replaced by (n 2) others may take any one of three forms. In the first place it may be required that 2) given elements shall be replaced by the same elements in a (n
2)
different order, so that
two elements are not involved.
Secondly,
the requirement elements.
may
involve (n
1) elements, or, thirdly, all the
In the
first
case suppose that

let r
tr
is
a substitution which satisfies
the conditions, and

elements.
be the transposition of the two remaining Then ar also satisfies the conditions, and one of the two
-,
substitutions
If (n
1)
element
t
is
,
XH
<rr belongs to the alternating group. elements are involved, suppose that the remaining so that neither the element which replaces x n nor that
The elements which are to replace is assigned. x n _i are all known with the exception of one. Suppose Then from the that it is not known which element replaces x n _ which substitution one x n _ we can construct elements x x x a X H _ x b ..)..., and from satisfies the requirements, say ( ...
which x replaces
x lt x
l ,
<t
the
fact
n elements
that
a second one, only distinguished from the

,
first
l
tt
in the
.
.
the alternating group. are two elements Finally, if all the n elements are involved, there for which the substituted elements are not assigned. Suppose these
<r
=a
x n _^ is followed by x n thus r x Then either or r belongs to (x b n ).
(.
xa x n _ x x b
.)
74
to
bex
ll
and
.r,
.
(
If
now
the elements are arranged in cycles in
the usual manner, there will be two cycles which are not closed, the one ending with a? _ n the other with x can then construct
rt
lt
.
We
two substitutions and which satisfy the requirements, the one being obtained by simply closing the two incomplete cycles, the other
ff
by uniting the
Theorem XI,
ting group.
it
then follows that either
latter in a single parenthesis. <t
From Chapter
II,
or
belongs to the alterna
The
alternating group of
It
n elements
is
therefore at least (n
it
2)-
fold transitive.
cannot be (n
1
l)-fold transitive, since

,x.2 ,
. .
.
contains
no substitution which leaves x

verts
tl
_.2
unchanged, and con
x _
ll
into
If
xn
is
63.
a AN fold transitive group, the subgroup
of
which does not
affect
will
be
(k
x.2
1) -fold transitive;
the subgroup
G"
and
of
G
.
which does not
affect
will
so on.
,
. .
k Finally the subgroup G(

l
be (k
l
^>
2)-fold transitive,
affect
which does not
a?!, a? 2
xk _
will
k
successively to G^
be simply
l
transitive.
,
Applying Theorem II
G",
G,
we
obtain
is
Theorem
equal to n(n
III.
1)
The order r of a k-fold transitive group

2)
.
(n
(n
fc-f-1)
m, ivhere
is
the order of
any subgroup
ivhich leaves k elements unchanged.
64. A simply transitive group is called non-primitive when elements can be divided into systems, each including the same number, such that every substitution of the group replaces all the
its
elements of any system either by the elements of the same system or by those of another system. The substitutions of the group can
therefore be effected by
units,
first
interchanging the several systems as
and then interchanging the elements within each separate sys

simply transitive group which does not possess this property
tem.
A
is
called primitive.
For example, the groups

G,
= [1,
j
(x
x.2 ),
OX),
(x
G,
1, (ajjaâ),
(xfaXs), (x 1 x B xg ),

are both non-primitive.
0-3,
l ,
75
G has two systems of elements, a\ x, and # 4 a? 5 a- 6 and a? 7 a? 8 a?9 has three G, x; systems, x lt x 2 x,, The powers of a circular substitution of prime order form a
, ,
primitive group,
e. g.
6r3
= [1,
(aj^scg),
(xîC)].
of
The powers
of
circular
substitution
composite
order
is
form a non-primitive group. n =p *i -p.2 a * -p 3 a3 where p

l
.
If the
}
degree of the substitution

3
,
. .
p.2
are the different prime
factors of n, the corresponding systems of elements can be selected
in [(!
-f- 1) (2 4~ 1) ( 3 2] different ways, as 1) For example, in the case of the group

r/
.
. .
is
readily seen.
G,
= [1,
(x
x.2
x x ix x
3 5
6) 9
(x.x.x,) (x 2 XiXR ), (x
we may take
a* 2
,
either
two systems of three elements each,
a^
a? 2
,
^5
a? 2
and
,
a? 4
o; 6
or three systems of two elements each,
a? t
,c 4
a? 5
and Xg,^.
theorem applies here, the proof which may be omitted on

its
account of
obvious character:
Theorem
IV.
is
I/,
of the elements into systems is possible in two different

that one division
third
for a non-primitive group, the division ways such
not merely a subdivision of the other, then a
mode of
division
new system the elements common and one of the second.

It
can also be obtained by combining into a to a system of the first division
must be observed that a single element

"
as a
" in the present system
sense.
not to be regarded Thus the group G^ above

is
admits of only two kinds of systems. 65. The elements of a non -primitive group
can be ar
ranged in a table, as follows.

those substitutions
8}
:
The
first
line contains all
and only
1, #2, #3,
S, n
which leave the several systems unchanged as
units,
and which
accordingly only interchange the elements within the systems. (The line will of course vary with the particular distribution of the ele
ments in systems.)
From
and "non-primitive" apply only
the definition of transitivity, (for the names "primitive" to simply transitive groups), there
76
in the given
must be
ment x a
of
group a substitution 2 which converts any ele one system into an element x b of another system, and
fr.
which consequently interchanges the several systems in a certain way. For the second line of the table we take
We
show
then, 1) that all the substitutions of this line produce the of the order of the systems as
all
<>._,;
same rearrangement
for every s\
leaves this order unchanged; 2) that

- is line; for if
substitutions which produce
the same rearrangement of the systems as

1
that
all
r 2 are contained in this ~ = s A so that T = SAT._.; 8) one of these, then rc- 2 the substitutions of the second line are different from one
,
another; and 4) that they are

line.
all
different
from those of the
first
If there is then
still
another substitution ^3 which produces a
new arrangement of
the systems, this gives rise to a third line which
possesses similar properties, and so on.
Theorem
group
If a non-primitive group G contains a sub of order in ivhich does not interchange the several si/stems
V.
of elements, the order r of

/JL
!
is
equal to mq, where q
is
a divisor of
JL
being the number of systems.

66.
If
we denote
the several systems, regarded, so to speak,
as being themselves elements,
tutions of any one line
A^, then all the substi of the table above, and only these, produce
by
A A
19
,
the same rearrangement of the A,
A^.
To every line
of the
table corresponds therefore a substitution of the
s,
the
first line,
for example, corresponding to the identical substitution, etc.
These
new
a
substitutions
that they form a
we denote by new group

.
and
to the elements
if
It is readily seen ... %,r 1, For the successive application of produces the same rearrangement of
t
the corresponding * a and the elements x. applied to Accordingly, since

these elements as
also 3 a
/3
^ were successively c^ = y we have

<r
,
where
<r
which contains
acteristic
corresponds to the line of the table above The system of s s therefore possesses the char
3y
We
and
(S.
property of a group. perceive here a peculiar relation between the two groups G To every substitution s of the former corresponds one sub-
77
of the latter, and again to every stitution of ($ corresponds either one substitution, or a certain constant number of substitutions s of G.
And
this correspondence is moreover of such a nature that to the product of any two s s corresponds the product of the two corres
ponding
s.
If to every
substitution, identity, in ($
unchanged.
this type.
only one which leaves the order of the systems A The two following groups may serve as an example of
s,
there corresponds only one
then there
is
Suppose that
(oô) (xzx) (ave),
G=
[1,
(aw)
(#2# 6 ) (â? 5 ),
(x&
Here the systems A 19 A 2 and A 3 are composed respectively of x andx2 x3 and a? 6 and o?4 and r 5 The corresponding substitutions
,
, ,
.
of the A* s
form the group

2
67.
= [1, (A A (AA (AAoA,), (A^s), (AA.A.OJ. We examine more closely the subgroup (7 = [s ,s ,s S ]
3 ),
2 ),
1 1
HI
65. Since 6^ cannot replace any element of group (7 of one system by an element of another system, it follows that G is intransitive. Any arbitrary substitution t of G transforms G into
of the
t~
G t= G
l
I.
The
latter is also a
subgroup of G:
it is
similar to
G^
and
that
it
evidently does not interchange the systems of G.
It follows
G\
= G,
,
Suppose that any system of a non-primitive group consists of the

elements x\ x
.
x\
The subgroup G
therefore permutes the
elements x
among
themselves.
We
proceed to examine whether
these elements are transitively connected with one another by the group GI or whether this is the case only when substitutions of G
,
added which interchange the systems of elements. Suppose x a and again x a + X Q, etc., are transitively that x\, a? 2
are
,
.
.
.
l ,
connected by
6?!.
Then G
contains
l 1
a substitution of the form

6r 1?
it
i
x a + ...)..., and since t~ G x places all the elements x\ x 2 consideration then shows that x\, x 2
(x
1
t=
t=
a
.
.
follows that
. . .
re
by x a + x a form
.
p.
Further
a system of non-
primitivity.
78
Accordingly
if
the systems of non-primitivity are chosen at the outset as small as possible, then the group G connects all the ele
l
ments of every system
transitively.
be thus chosen, we direct our attention to those cycles of the several substitutions of GI which interchange the elements x lt x 2 of any one system of non-primitivity.
Assuming the systems

.
. .
to
These form a
several substitutions of
Similarly the components of the group which interchange the elements x^ a \ x.{ a \ (a \ The groups of any second system form a group H", a are similar, for if t = (x xj ). ..)... is a substitution of G, then
transitive
.
the transformation t~
of
G t=G
{
will convert
/?
H
!
into
H( a \
//
The order
the
77/
is
a multiple of
,"-
and a
divisor of
/
-
where
is
number
of systems of non-primitivity.
68.
The following
easily
demonstrated theorems in regard
to
to primitive
and non- primitive groups may be added here:
Theorem
sitive
VI.
If from the elements
a? 1?
X H of
a tran
group
G any
system x\, x
can be
selected such that
which replaces any x a by an x & permutes the x"s only among themselves, then G is a non-primitive group.
every substitution of
Theorem
transitive
VII.
If
from
the elements
2
,
group
two systems x\, x
selected such that
any
substitution
x ,a%,...xM of a and x" lt x" 2 can be which replaces any element x a

}
...
by an x" $ replaces all the ,T"S by

group.
a?
"s, then
is
a non-primitive
contains substi
:
Theorem
by an element of
the system. *
69.
erties of
VIII.
the
Every primitive group
tutions which replace
an element x a of any given system x ,x 2 same system, and which at the same time replace any second element of the system by some element not belonging to
,
.
. .
discussion has led us to two general prop which, together with transitivity and primitivity, groups
The preceding
are of fundamental importance.
67 the subgroup G of G possessed the property of being reproduced by transformation with respect to every substitution / of
In
Radio: Ueber primitive Gnippen.
Crelle CI. p.
l.

(?,
79
conveniently
so that for every
we have t~ G t=
l l
6r t
We may
I
.
indicate this property of

I
I
by the equation
I
G- G G = G
It is to
or
G=GG
be observed however that this notation must be cautiously employed. For example, if G is any subgroup of G, we have always G ~^GG = 6r, and from this equation would apparently fol
l l
low
G GI = G
6r,
and consequently
G =G
l
G.
}
But
.
this
last
The reason equation holds only for a special type of subgroup G for this apparent inconsistency lies in the fact, that in the equation
G~~ GI
l
G=
S
l
G! the two
6r s
represent the same substitution and the
two GI ~
GG = G
Two
in general different substitutions, while in the equation
the reverse
is
the case.
We
1)
introduce here the following definitions.

substitutions
s^
and
s2
are commutative * if
2 Si
.
S rS 2 =: 2)
A substitution
Two groups
sl
and a group
8l
H are commutative if
l
.
H=Hs
3)
H and
to
arc commutative if
HG=GH.
The
of
last
equation
is
any substitution of
be understood as indicating that the product

into
product of some substitution of

that, if the substitutions of
any substitution of G is equal to the G into some substitution of H, so are denoted by s and those of by /,
then
S a tp
for every a
and
2) s l
ft.
Under
may
be a substitution of
Jff;
for s
and
are then
always commutative. Under 3) a case of special importance is that, an instance of which we have just considered, for which is a sub
group of G.
In
this case s a
and
s&
of the equation s a tp
y s$
are
always to be taken equal.
A
a
subgroup
H of
any group
G, for ichich
G~
HG = H,
is
called
self-conjugate
subgroup of G.
"
* German: " vertuuschbar"; French: "^changeable", retained as "interchangeable by Bolza: Ainer. Jour. Math. XIII, p. 11.
80
The following may serve as examples: The substitutions s = (x^x^) (x ix^x& ), 1)

:
s2
= (x^) (x
are commutative; for their product

of the order of the factors.
is
(x^x^x^),
is
independently
Every power
of any substitution s
commutative with every
other power s& of the same substitution.
Two
tive.
substitutions
which have no common element are commuta
The group [1, (x^) (x3 x^, (x^) (a? 2 a? 4 ), (x^) (x^c 3 )~\ is 2) commutative with every substitution of the four elements x x 2
l , ,
a58 ,o?4 .
The
alternating group of
n elements
is
commutative with every
substitution of the
3)
same elements.
The group
of the latter.
H of
2),
group group
of the four elements
being commutative with the symmetric x x.2 a?3 x is a self -con jugate sub
l ,
, ,
,
The
alternating group of n elements
is
a self-conjugate sub
group of the corresponding symmetric group.
Every group
of order
r,
which
is
not contained in the alterna

of subgroup the group G which are contained
ting group A, contains as a
self -con jugate
order Jr composed of those substitutions of

in
(Theorem VIII, Chapter

identical
II).
is,
The
substitution
by
itself,
a self-conjugate sub
group of every group.

70. may employ the principle of commutativity to further the solution of the problem of the construction of groups begun in
We
(
Chapter II
33-40).
All substitutions of n elements which are commutative with any

given substitution of the same elements,
form a group.
s, it
For
if
f]
t.
,
are commutative with
follows from
that
so that the product
t-f 2
also occurs
among
the substitutions
t.
All substitutions of n elements which are commutative with a

given group
81
of the same elements form a group which contains
as a self-conjugate subgroup.
For from
follows
and among the
are included
all
If two commutative groups

the identical substitution, in
have no substitution, except then the order of the smallest common,
G and
group
K=\G,H\
is
equal
71.
to the
product of the orders of
G and
H.
of order
If a group G of order 2r contains a subgroup is a self-conjugate subgroup of G. r, then
H
,
of
For
if
t,
if
H are denoted by
which
is
1, s 2
,&,,...
s,.
and
is
any substitution of
,
. .
.
not contained in H, then
tsz, ts 3
ts,.
s.2
are the remaining r substitutions of G.

s 3 t,
. .
But
in the
same way,
are also these remaining substitutions. t, t, s a t is equal to some tsp, that is, wr substitution Consequently every
.
.t t
HG = H. If a group G contains a self-conjugate subgroup H and any other subgroup K, then the greatest subgroup L common to H and K a self-conjugate subgroup of K. If the orders of G, H, K, L
have in every case t~ 8pt
l
,
= sa
and therefore
G~
is
are respectively
a
g, h, k,
I,
then
hi
is
a multiple of
l
k
.
For
s~ Ls
l
if
s is
any substitution of K, then s~ Ls

~
is
contained in K,
since all the separate factors s

is
\ L, s are contained in
K.
But
and H, subgroup as these s~ Ls is in contained L, and, Consequently two groups have the same number of substitutions, we must have s~ Ls = L, and L is a self -conjugate subgroup of K.
also contained in for
is
of
Hs = H.
between the orders of the four groups follows at once from the formula of Frobenius ( 48). We have only to take
relation
for the
The
K of
.
this
formula the present group H, and to put

to
1.
all
the
{ ,
d,2
d m equal
We
have then
self-conjugate
subgroup of a transitive group either

it
affects
every element of the latter, or else

tution alone.
consists of the identical substi
H= G~ HG a -conjugate subgroup of the transi H does not affect the element x then, since G tive group and
For
1
if
is
self
(7,
if
l ,
contains a substitution
that s^~
l
SA
which replaces x by x\,

l
it
is,
would follow
that
Hs^ = H
would
also not affect
,T A
that
H would
intransi
not affect any element.

If a self-conjugate subgroup of a transitive group
tive,
is
then
is
non-primitive
and
H only interchanges
to
the elements
within the several systems of non-primitivity.
For suppose that x and x\ belong Then intransitivity with respect to H.

1
l
}
two different systems of

SA
G
l
it follows that which replaces x by x\, and since s\~ s\~ Hs^ must replace x\ only by elements transitively connected re with ,T A with respect to H. But s^" replaces x\ by x and
1 1
Hs\ = H,
places
x by every element of the same system

l
of intransitivity with
Consequently the remaining factor s\ must replace every ele ment of the system containing x by an element of the system con
l
. l
taining
O?A.
The systems
of intransitivity of
H are
therefore the
systems of non-primitivity of G.
72. Another important property is that of the correspond ence of two groups, of which an instance has already been met 66. The two groups G and ($ of this Section were so with in
related that to every substitution s of

tion
of
,
G corresponded
one substitu
of s
s.
and
to every
corresponded a certain
number
S
s.
The correspondence was moreover such

two
.s
that to the product of any
corresponded the product of two corresponding
consider at once the more general type of correspond * where to ence, every substitution of either group correspond a certain number of substitutions of the other, and to every pro
We may
duct
s a.s0
vice versa.
corresponds every product S a S/s of corresponding S s and We may then readily show that to every substitution of
the one group correspond the same number of substitutions of the For if to 1 of the group G correspond 1, S 2 S 3 other. S 7 of
, , .
.
A. Capelli: Battaglini Gior. 1878, p. 32seq.
GENEKAL CLASSIFICATION OF GROUPS.

then,
if
83
,
corresponds to
s,
s,
all
the substitutions
3S 2 ,3 3
S3 v
correspond to
1
by
definition.
1
Conversely,
if
any substitution
therefore
corresponds to s, then g" * corresponds 3- * is contained in the series 1, ,, 3

series
,
tos^s^l, and
,
.
3*,,
contains
all
Consequently the the substitutions of which cor

. .
respond to
to every
and the number g is constant for every s. Similarly correspond the same number p of the substitutions s.
s,
It is evident at
once that
G ()
w/wc/i correspond to the identical substi
tution of
(G) form a group
(,))
which
is
self-conjugate sub-
group of
G ().
of two groups as just defined
is
The correspondence
morphism. of ($, and
called iso
If to every substitution of
to every substitution of (S
G
p
correspond q substitutions
substitutions of G, then
if
and
(S
are said to be (p-q)-fold isomorphic, or
p and
q are not
specified,
manifold isomorphic.
If
p = q = l,
the groups are said
to be simply isomorphic. *
EXAMPLES.
I.
The groups
are simply isomorphic, the substitutions corresponding in the order

as written.
For
if
any two substitutions of G, and the corres

are
ponding
substitutions of 1\
multiplied together, the resulting
products again occupy the same positions in their respective groups.

II.
The groups
G
beside
= [1, (x^l
1,
=[!,
(4M,) ($&), (,V.) (,V,),

1 of
.
(,*, ,*,)
(?A)]
take,
are (1 -2) -fold isomorphic.
Corresponding to
/
G wo may
any other arbitrary substitution of

itself in different
It follows that / is
simply isomorphic with

*( /. Camille
:
ways.
etlric"
Jordan Trait e etc., 07-74, where the nuincs "liolocdric" and *merlisomorphism are employed. These have been retained by 15ol/a: Ainer. Jour, Vol. XEII. The "simply, manifold, (p-Q)-fold isomorphic " Hbpve represent the "ein" of the (Jenmui edition. stulis, melirstulig, (p-Q)-stu.(ig isomorph
84
III.
The groups
G=
[1,
(XM)
(xz x},
(x^)
(x,x,\ (x.x,) (x,^)],
r=[l, (,vV8),
are (2-3)-fold isomorphic.
1, (rjfâ), (^1^3*2)
(*,*,), (4^ 2 ), ft*,), (*&)]
To
the substitution 1 of
to
1
correspond
of
s
^>
-
an d
conversely
of
F correspond
1,
fax?) (x z Xt) of
73.
If
G and F are (mn)-fold isomorphic^
then their orders
are in the ratio of
n.
a self-conjugate subgroup of G, and if A is the corres ponding subgroup of F, then A is a self-conjugate subgroup of 1\
If
is
For from
G~ LG = L
1
follows at once
it
F~ A F
1
A.
In the case
of (p-l)-fold isomorphism,
may however happen that the group

alone.
A consists of the identical substitution

74.
groups
in reference to transitivity, primitivity, commutativity,
Having now discussed the more elementary properties of and
isomorphism, we turn next to certain more elaborate investigations devoted to the same subjects.
The
of
in substitutions of a transitive
group
which do not
affect
the element
subgroup x 2 from a second subgroup 6r 2 and so on All these subgroups to the subgroup G H which does not affect x n are similar; for if (r a is any substitution of G which replaces x by
v l
x form
of G.
Similarly the substitutions

,
which do not
affect
a? a
we have
If
ff
Giâ=
G*>
The groups G a
number
are therefore
all
of
order m.
now we denote by
affect exactly
is
2
,
[g] the
of those substitutions of
which
q elements but leave the remaining (n

also the corresponding
3
,
. .
.
1)
unchanged, then [g]

of the other groups
number
for each
G G
Gn
.
It follows
then from the mean
ing of the symbol [g] that
m = |>-1] + [nZ] +
where the
Gr,, 6r 2
.
+ [q] +
+ [2] + [0],
= 1.
1] substitutions
all
symbol [1] does not of course occur, and [0]
Gn
therefore
possess
which
affect exactly (n
1) elements.
together n[n These are
different, for
any substitution which leaves

cannot also occur in Gp.
only x a unchanged occurs
in
Ga
but
But
this is not the case with substitutions
GENERAL CLASSIFICATION OF GEOUPS.
85
which
affect exactly (n
2)
elements
for
if
any one of these leaves

.
both x a and x$ unchanged, it will occur in both G a and G$ Accord ingly every one of these n[n 2] substitutions is counted twice,
and
therefore
contains
Jw[w
2]
substitutions
which
affect
exactly (n Similarly every one of the n[q] substitu tions of q elements which occur in 6r 1? 6r 2 G n is counted (n q)
2) elements.
,
.
times,
in
and there are therefore only affect exactly
[q] different substitutions
which
q elements.
We
have then for the

than n elements
total
number
of substitutions in G,
which
affect less
If this
number
is
subtracted from that of
all
the substitutions in G,
the remainder gives the number of substitutions in But from Theorem II exactly n elements.
r
which
affect
= mn = n[n
l]-fw[>
2]+ ..+
.
and consequently the required difference
is
No term
n ~~ l
in the parenthesis
=
is
negative.
The
last
one
is
equal to
since [0]
1.
Consequently
N^ (n1).
gromp contains n elements. If
at
least
Theorem IX.
(
Every
which
transitive
1) substitutions
affect all the
there are
contains substitu more than (n 1) of these, then the group also elements.* than tions which affect less 1) (n
Corollary.
which
A
.
.
k-fold transitive group contains substitutions
affect exactly
elements,
fc
and
others
ivhich affect exactly
(n1),
(n
2),
(n
+ 1)
elements.
Those substitutions which

substitutions
affect exactly
k elements we shall
call
of the
k tn
class.
We
,
have just demonstrated the

class.
existence of substitutions of the

If
n th or highest
we
group
consider a non-primitive group 6r, there is ( 66) a second isomorphic with G, the substitutions of which interchange
(2),
*C. Jordan: Liouville Jour.
XVII, p
351.
86
the elements
tions
H exactly as the corresponding substitu interchange the several systems of non-primitivity.
1 , ,
. .
.
A.2
of
G
is
Since
transitive,
is
also transitive.
From Theorem IX
fol
lows therefore
Theorem X.
75.
Every non-primitive group
contains substi
tutions which interchange all the systems of non-primitivity.
We
Gr,
construct within the transitive group

all
the subgroup
H of
lowest order, which contains
the substitutions of the high
est class in
and prove that
this
group
is
also transitive.
H
is
is
evidently a self-conjugate subgroup of G.
If
were
intransitive,
the case, let

19
2
, . .
.
must then be non-primitive (Theorem VI). If this 66 which affects the systems be the group of
as elements.
is transitive.
A A
Ap regarded
To
substitu
tions of the highest class in

est class in G.
$
correspond substitutions of the high

is
(The converse
not necessarily true).
Suppose that
the subgroup of the lowest order which contains all the substi or To $ then corresponds either tutions of the highest class in
is
.
a subgroup of H.
aj s.
If
is
transitive in the .A
s,
H
if
is
transitive in the
The question
and
).
)
therefore reduces to the consideration of the
groups
can be intransitive only
is
non- primitive
of non-
and
G accordingly
contains
more comprehensive systems
primitivity.
If this
were the
),
same way from and The is then complete. proof group.
case, we should again start out in the and continue until we arrive at a primitive
Theorem XI.
the highest class form
76.
In every
transitive
group the substitutions of
by themselves a transitive system.

transitive
Suppose a second
If then
group
to
have
all its
sub
stitutions of the highest class in
common
with
of the preceding
,
Section.
we
ing to the
in
subgroup H of
construct the subgroup

G,
for 6r
correspond
we have
H.
Moreover the number
of the substitutions of the highest class
is
where
[g],
has the same relation to
NI
is,
as
we have
just seen equal to the
H as [q\ to G. But the number N of 74, Consequently
87
But, since
is
entirely contained in
left
and therefore the

vanish
if
hand member
is 0.
6r, it follows that [g]2l[g]i, of the equation above can only
each parenthesis
Consequently
th
and
can only
differ in respect to substitutions of the (n
l)
class.
Theorem XII.
stitutions
If two transitive groups have
all their
sub
of
the highest class in
common, they can only
differ in
those substitutions which leave only one element unchanged.

77.
Let
of
subgroups These groups

sequently
6r2
,
G be any G which
are, as
. .
.
transitive
group and
G G
1 ,
. . .
Gn
those
do not
affect
x x2
l ,
we have
seen, all similar.
If
x u respectively. now G and con

1 ,
6r,,,
are /c-fold transitive, then
is at least (fc-f- 1)-
if it be required that the (k -f 1) elements be replaced by x x h x ik+l respectively, we can find in G some substitution s which replaces x lt x2 o?3 xx + by Xf^Xj^Xj,^. ,x hk+1 where o^ 2 ,a?, may be any elements
f old transitive.
For
X H x 2j
#*
+ shall
!
tl ,
l3 ,
whatever.
a?/, 2
,
Again
Gh
a;
contains some substitution

.
o? ;
...
by
aj
/2 ,
V3 ,
t which replaces st of the substitution Consequently
satisfies
the requirement.
From
and
this follows the
more general
If a group
Theorem
tive,
XIII.
is
at least k-fold transi
if the
subgroup of
which leaves k given elements un
changed
sitive.*
is still
h-fold transitive, then
G is
at least (k-}-h}-fold tran
Suppose that those substitutions of a AN fold group G, which, excluding the identical substitution,
78.
transitive
affect
the
smallest
number
of elements,
are of the q th class,
i.
e.,
that they
is
affect exactly q elements. The question arises whether there connection between the numbers k and q.
any
In the
first
place suppose &2lg, and let one of the substitu
tions of the g th class contained in
Then, on account of
ildeten Doppelmodul.
its
A;-
fold
G be s = (x^ ...)...(... x _ x, transitivity, G also contains a substituq

l
2 ).
*G. Frobenius: Ueberdie Congruenz nach einem aus zweiendlichen Gruppen gebCrelle CI. p. 290.
OO
tion
<T,
and which
which replaces x i9 x2 ,. ..x q _ is therefore of the form

ff
l ,
x by x x
tl
l ,
,...
# _n x *
ff
= (x )(x )...
l
(x q _ )(x q x K ...).
l
We have
then
and since
this substitution affects only 3 elements,
it
follows that
tf3.
The substitutions of the gth class Secondly, suppose k < q. then be of either of the forms
may
In the
first
case
we take
.
. .
<r,
(ojj) (a; 2 )
(x
and
in the second
It
is
evident
that both
are
possible,
if
in the latter case
it
is
remembered that n
> g.
We
obtain then
and
if
we form now
2fc
(î"
*!^)^"
at
if
the
first
(k
2)
elements are
k)
removed,
and there remain,
the
most,
(
ff
+ (g
the
(k
2)
= 2q
2.
Similarly,
we form
,f-
s 2 <T 2 )s.r \
first (k
+ 1)
k -\- 1) elements are removed, arid there remain, at the most, q-\- (q = 2q 2&-J-2. (k l) By hypothesis, this number cannot be
less
than
g.
Consequently
g^2fc
2.
Theorem XIV.
substitution,
If a k-fold transitive group contains
any
except the identical
than (2k 2) elements, it the most only three elements.
substitution, which affects less contains also substitutions which affect at
In this case, This theorem gives a positive result only if k > 2. by anticipating the conclusions of the next Section, we can add the
following
89
contains sub
Corollary.
stitutions,
If a k-fold transitive group
k>2
different
from
identity,
which
affect
not
more
than
(Ik
2) elements,
it is
either the alternating or the symmetric group.

this result with the corollary of
it
We may
IX.
(n
If
now combine
Theorem
the
is
&-fold transitive,
contains substitutions of the class
fc If Accordingly q^.(n 1). alternating nor the symmetric group, q > (2/c
k-\-l).
G
2).
is
neither
Consequently
Theorem XV.
sitive.
If a group of degree n
it is,
is
neither the alter

-j-
nating nor the symmetric group,
at the most,
1 \-fold
tran
occur
That the upper limit of transitivity here assigned may actually is demonstrated by the five-fold transitive group of twelve
elements discovered by Mat^hieu,
G=
(XX^X^) (X^X^XT),
(UiX) (aê) (^3^7)
(2/a2/2)
(xXsX.jX-,)
(2/22/0
(X^X^X^).
(#1*3)
0%),
(*W) (->W),
\-
(aw) feî) (aw)>
(yMz) (^1^3) (^4^5) (^0^7)
79.
Theorem XVI.
If a k-fold transitive group (k

it
>
contains a circular substitution of three elements,

alternating group.
contains the
Suppose that
since
G
= (x
jc. xd) 2
G
&)
is
at least
two -fold
occurs in the given group G. Then, substitution a it must contain transitive,
= (x
(x^^x^ ...}... and consequently also
In the same
way
it
appears that
contains
.
. .
(T rr 2 .T 5 ), (x,x,x^,
Consequently
35)
G contains
Theorem XVII.
tains
If a k-fold transitive group

is
(h>
1)
con
a transposition, the group
symmetric.
The proof is exactly analogous to the preceding. For simply transitive groups the last two theorems hold only
90
limitations, as
under certain
l
appear from the following instances
G = [1, (a^
Both of these are
transitive.
But the former contains a

and the
substi
latter a
tution of two elements, without being symmetric,
substitution of three elements without being the alternating group.
explanation of this exception in the case of simply transitive groups is obtained from the following considerations.
80.
An
If
it is
we
arbitrarily select
two or more substitutions of n elements,
to be regarded as extremely probable that the group of lowest order which contains these is the symmetric group, or at least the
alternating group.
in favor of the symmetric
In the case of two substitutions the probability group may be taken as about f and in
,
favor of the alternating, but not symmetric, group as about J. In order that any given substitutions may generate a group which
is
are necessary, and

substitutions
s,-
only a part of the nl possible substitutions, very special relations it is highly improbable that arbitrarily chosen
=
[
"
should satisfy these conditions. The

all
exception most likely to occur would be that

tions
the given substitu
were severally equivalent to an even number of transposi tions and would consequently generate the alternating group.
which
In general, therefore, we must regard every transitive group is neither symmetric nor alternating, and every intransitive
group which
cidedly
made up of symmetric or alternating parts, as de And we shall expect to find in such cases exceptional.
is
not
among the substitutions of the group, of such a nature as to limit the number of their distinct combinations.
special relations
Such relations occur in the case of the two groups cited above. Both of them belong to the groups which we have designated as In G the elements x x.2 form one system, and non-primitive.
l
l ,
a? 3
a? 4
another;
it is
therefore impossible that
G
7
should include, for

therefore
example, the transposition (x rT 3 ).

non-primitivity
In
there are three systems of

9
,
x l9 x ,x at
z
x^x^x^, and X ,x 8 ,x
(7 2
cannot contain the substitution (x^xx.

It
is,
91
then, evidently of importance to examine the influence of
primitivity on the character of a transitive group, attention now in this direction.

81.
and we turn our
With the
last
two theorems belongs naturally

If a primitive group contains either of
Theorem XVIII.
the two substitutions
ff
~ fax^),
= (x&t),
sym
it
contains in the former case the alternating, in the latter the
metric group.
The proofs in the two cases are of the same character. only that for the latter case.
We
give
From Theorem
tion
a? 3 d? 3
,
VIII, the given group must contain a substitu which leaves x unchanged and replaces x 2 by a new element or which leaves x 2 unchanged and replaces x by a new element or which replaces x by x.2 or x.2 by x and the latter element in
1 v l
t
either case
by a new element x 3
If then
we transform
with
respect to this substitution,

either
we
obtain a transposition r connecting

r
The presence of r and r in the group shows that the latter must contain the symmet ric group of the three elements x x 2 x$ From Theorem VIII
l
or
x 2 with x
for
example
= (xiX
, .
3 ).
l ,
there must also be in the given group a second substitution which replaces one of these three elements by either itself or a second one
among them, and which also replaces one of them by av Suppose this substitution to be, for instance,
S
(
,
new element
tJCÔC-^
QCftOO^
We
and
obtain then
T
it
"=s-
(x 2 x 3 )s
= (x x
1
4 ),
follows that the given group contains the symmetric group of

,
the four elements a^

82.
x2 x3 x
, ,
and so
on.
We can generalize the last theorem as follows: Theorem XIX. If a primitive group G with the
#!
,
elements
x.2
x n contains a primitive subgroup
of degree k
< n,
then
contains a series of primitive subgroups similar to
H,
92
such
ivhere
that every H), affects
the
elements
a^
o? 2
a?*
j ,
.r,,
+\_
l ,
x ly
x- 29
.x k _ may be
l
selected arbitrarily.
We take H^ = H
tutions of (rinto
and transform
lt
H with respect to
Now
let
all
the substi
H H
19
H"^
H\
be that one of the

.
x of transformed groups which connects the k elements a? u x.,, H-i with other elements, but with the smallest number of these.
.
f,
We
maintain that this smallest number

,
is
one.
For
if
several
new
elements lu I 2 ... occurred in H\, then from Theorem VIII there must be in the primitive group H\ a substitution which replaces
one I by another c and of the elements oj n x 2

,
at the
.
same time replaces

Suppose that
a second c
by one
xk
is
H =tH t~
rr
such a substitution, the case where ,5 = -f being included. l will still contain r 7 but will not contain f a 1 l \
.
Then
,
there
if
fore contains fewer

is
new elements
than
H\
Consequently
H\
l
.
properly chosen, it will contain only one new element, say x k + It will therefore not contain some one of the elements of say x a We select then from H^ a substitution u (. x a x,, ...)... and
form the group
u~ H\u =
l
This group contains

/.
*>
+n
affects
but not x k
only x ly
It
rily,
In the same way we can form- a group H^ which

.
x.2
Xn._ lt
+2l
and so
OD.
,
.
remains to be shown that x^ x 2

is,
jETj
xk _ can be taken
l
arbitra
that
that the assumption
H H
_a
is
always allowable.
in the series Il\
a
,
Sup
pose that
there
J/o
is
contains
a?,
aj 2
a? fr
Then
a group
(,
which
also contains
x.,,.
. .
H_
k
1 ,
Proceeding from
and the elements x lt
x ._ a +
A
we
construct a series of
//.
groups, as before, arriving finally at the

83.
group
Theorem XX.
// a primitive group
contains a primitive subgroup

(n
k-\- \}-fold transitive.
of degree k then
G of G is
degree
at least
From the preceding theorem H^

HI ments x l9
\
,
affects the
elements x
j
l ,
a? a
xk
7/2
the
.x,,
.
elements
. .
a? fr
x x x h+2 x/, +
l ,
. ,
xk X
,
H H
j
H^
the ele
] ,
and so
on.
All these groups are

transitive; consequently,
transitive,
93
,
from Theorem XIII,

three-fold transitive,
\H^,H
\H
and
2 \
is
two-fold
s \
finally
r=
is at least
\HD
>t
_ k+1 \
Therefore G, which includes
(n
k-\- l)-fold transitive.
is
also at least (n
&-f-l)-fold transitive.*
Corollary
(n
I.
circular substitution of the
If a primitive group of degree n contains a prime order p, the group is at least
p -j- ~L)-fold
transitive.
For the powers of the

degree
p.
circular substitution
form a group
H of
Corollary
II.
If a transitive group of degree
contains a
the
circular substitution of prime order
p <
it is
-^-, then,
if
group
does not contain the alternating group,
non-primitive.
is
From Theorem XV,

fold transitive
is
every group which
more than
-~- -j-
I -
either alternating or symmetric.
And
since the
presence of a circular substitution of a
prime order
in a primitive
group would require the

sitive, it
latter to
be
at least (n
p-}- l)-fold tran
would
follow,
if
p<
-Q-,
that the group
would be more than
-q-
+l
1-fold transitive
and must therefore be either alternating or
As these alternatives are excluded, the group must be symmetric. non- primitive.
84.
In the proof of
Theorem
XIX
the primitivity of the

substi
group H was only employed to demonstrate the presence of

tutions
kind.
which contained two successions of elements of a certain
The presence
if
of such substitutions
would
case.
also evidently be
assured
H were two-fold or many-fold transitive.

be valid in this
If a primitive group
is
Theorems
XIX
then
and
XX would therefore still

Theorem XXI.
The
latter
takes the form:
of degree n con-
* Another
proof of this theorem

1.
given by Rudio: Ueber primitive Gruppen,
Crelle CII, p.
94
tains
a k-fold
(7
transitive
subgroup
(k >.2) of degree q, then
is
at
least (n 85.
+ 2)
transitive.
If the requirement that the
subgroup
H of the preceding
We
Section shall be primitive or multiply transitive is not fulfilled, the the theory becomes at once far more complicated. * give here a of the few only simpler results.
Theorem XXII.
tains
If a primitive group
A
a subgroup
is
H of degree
< n,
then
G of degree n con G also contains a subgroup
whose degree
exactly n
1; or in other words:
transitive
group
of n elements, which has no subgroup of exactly n but has a subgroup of lower degree, is non-primitive.
1 elements,
Suppose that the subgroup

ments
a?], a? 2
,
. . .
H of
degree
if
<n
A
affects the ele
#*.
In the
first
place
<
then the group G,

s,
on account of
its
primitivity, contains a substitution

l ,
which replaces
one element of x
x^ by another element of the same system x \ by and at the same time replaces a second element of a?, a? 2
,
.
x2
>
some new element.

A
Then
H = s^ Hs
1
contains beside
1
some of the
\
H = \H,H more than elements, but less than n, since H and H affect at the most (2/ < n. If the degree ^ of H
old elements, also certain
new
ones, so that
affects
together
77
1)
is still
<
-^-,
77
we repeat the same
process, until
A, is
equal to or greater than
-^-.
are a? n a? 2 Then x\. Suppose that the elements of the last the primitive group G must again contain a substitution s2 which
l ,
. .
replaces two elements not belonging to
by two elements, one of
which does, while the other does not, belong to
.
j
H = Hs
}
Then the group

.fl",;
s.2
will connect
new elements with
those of
but,
from the way in which s, was taken, one new element is still not con That some of the old elements actually occur in H\ tained in H\
.
/j Jl J n. Accordingly H% = \H H\ con more elements than H^ but less than G. Proceeding in this way, we must finally arrive at a group K which contains exactly
follows from the fact that

tains
l ,
(n
1)
elements.
*C. Jordan: Liouville Jour. (2) XVI. B. Marggraf : Ueber primitive (iruppen mit transitiven Uutergruppen geriugeren Grades; Giessen Dissertation, 1890.

If so
95
l
,
is
transitive,
then
and consequently
H =\H
XIII,
on to K, are also
transitive.
From Theorem
H\, and G must

have
therefore in this case be at least two-fold transitive.
We
then the following
If a primitive group G contains a transitive subgroup of lower degree, then G is at least two-fold transitive.
Corollary.
86.
We turn
now
to a series of properties based
on the the
ory of self -con jugate subgroups.
Let
T= [1, ggyjfe,
of order
sm]
be a
self -conjugate
group
(
n = km.
The
substitutions of
subgroup of a can be arranged
41) in a table, the
first line
of
which contains the substitutions
of
H.
3 0fc,
.8m ffk
From
the definition of a self -con jugate subgroup
we have then
that
is,
the line of the table in which the product

ff
(s\<r a )
(s^p) occurs
tb
/5
depends only on
of the
th
and 0, or
in other words,
if
every substitution
line,
line is multiplied into every substitution of the
the resulting products all belong to one and the same line. If we denote the several lines, regarded as units, by z
l ,
z$
zk
then the line containing the product of the substitutions of z a into This symbol has then a defi those of Zp may be denoted by z a zp. Moreover, z azp cannot be equal to nite, unambiguous meaning.
z az y or to z y zp.
ff
For then we should have from the

ff
last
a ffft
ff
a ff y
or
a ff ft
ff
(T
ftj
that
is,
ffp
<r
or
<j
<7
paragraph Conse
.
quently
*a
=
among
the z
s,
denotes a substitution
and
this substitution corres

th
. . .
ponds
to
all
the substitutions s A r a
of the table.
m) of the a line The fs therefore form a group T, which is (1-m)-
= 1, 2, (A
96
fold isomorphic with the given group G.
are equal,
and both are equal
to
The degree and order of k (Cf. To the identical 97).
substitution of
corresponds the self-conjugate subgroup
H of
G.
We
87.
shall designate
as the quotient of
and.//, and write
accordingly
T = G:H.
different
A group G which contains a self-conjugate subgroup H, from identity, is called a compound group; otherwise G is
a simple group. If G contains no other self-conjugate subgroup is a maximal self- conjugate subgroup. which includes H, then
If
is
compound group, and

G,
if
the series of groups
G 19 G
...
G^
is so taken that every G\ is a maximal self -conjugate subgroup of the preceding one, then this series is called the series belonging to the compound group G, or the series of composition of G, or,
still
more
If the
briefly, the series
of G.
numbers
are the orders of the successive groups of the series of composition

of G, then e
l ,
e2
e^ + 1
1
are called the factors of composition of

. .
.
and we have
If,
r = e
e2 ey
e fL
in accordance with the notation of
86,
,
:
we
write,
,
G:G
=Î\,
G G =F
l
:
...
G> _
G^
= F^
G^-.I
F,
the order and the degree of every Fa is equal to e a (a 1, 2, ... // -f~ 1). r F For a is (l-r a )-fold isomorphic All the groups a are simple.
with G a _
?
1 ,
and
to the identical substitution in

if
ra
corresponds
Ga
in
contains a self-conjugate subgroup dif Consequently, ferent from identity, then the corresponding self -con jugate sub
G a -i-
Fa
The group of G a -i ( 73) contains and is greater than G a would therefore not be a maximal self -con jugate subgroup of
.
latter
G a -\*
the
The groups
groups
I\
which define the transition from every
G a to
following one in the series of composition, are called the factor of G. *

*O. Holder; Math. Ann.
XXXIV,
p. 30
ft.

88.
97
Given a compound group G,

if
it is
quite possible that the

It is
corresponding
series of composition is not fully determinate.
conceivable that,
a series of composition
G,
1 ,
6r 2
G^
has been found to exist, there

G,
in
may
also be a second series

. . .
G\,
,l
self -con jugate
which every
is
contained as a
maximal
in the
preceding one.
We
shall find
however
that, in.
subgroup whatever
way
the series of composition
may be
chosen, the
number
of groups
constant, and moreover the factors of composition are always the same, apart from their order of succession.
is
a.
Suppose the substitutions of G and G\ to be denoted by s a and Let r =r:e be the order of 6?!, and r j^rie respectively.
l
that of
G\
The
substitutions
common to G and G\ form

l
a
.
group
F(
44), the order
of
which
r
is
a factor of both
r,
and r\
We
write
1
=xy
= xy
ff
The
substitutions of
r we
1\
denote by
All the substitutions of
6?!
may then be arranged

the substitutions v a of
in a table, the first line of.
which consists of
We
obtain
where the
belonging to any line
is
any substitution of
not con
tained in the preceding lines.
The group G\ can be
treated in the
1?
2
same way.
We
,,
will
.
.
.
suppose that in this case, in place of
we have
2?
Every substitution of
or
G\
can, then, be
written in the form
Again, the product 8*-
it belongs to -p- 8 a s p which occurs in the second form of the product, is equal to s y and ~ the product itself is equal to s a S Y But, from the third form, this
l
.
*r *J* x *r (*t- ***d = (sa-Vp-^Xis G follows that s For, since G~ G G=G

l l l l
1 1
}
98
to
6r
,
s p~ 8 a product equal to quently the product belongs to the group r which is 8a
same product belongs ~ =

l r 1
since
G~ G
1
G=G
and therefore
s ys p.
8 y,
so that the
is
Conse
to
common
and G\
Hence
sa
A)
In
obtain
s p~ s a s p
ffy]
<r
sas p
= s ps a vy
s ps a
= sa s
p<r s
particular, since the
belong to both the
s s
and the
s"s,
we
B)
ffa
/3
= S^^;
**%?=%?*&
a/3=3/3S aT e
From
a
this
.
it
follows that the substitutions of the form $ a %
ff
form
group
For, by
repeated applications of the equation B),
we
obtain
The group
is
commutative with
6r
for
we have
The group
is
more extensive than
or
it is
contained in
consequently, from the assumption as to Gj and identical with G.
(5
must be
The order
is
of
@
a
is
equal to xy y
/9,
For,
if
Safi
^y
;&
it is
easily seen that

also ocyy
,
= a, 6 =
=r
Consequently the order of
(7
and since we have

r
= T& = xye
l ,
= r\e\ = xy e\
it
follows that
This
last result gives
us for the order of 1\
=
6j6
j
=~=
6
j
6j
F is a maximal self -conjugate subgroup and consequently occurs in one of the series of composition of either of these groups. For in the first place 1\ as a part of G 1} is commutative with G n and, as a part of G 1? is com
We
can show, further, that
of G,
and of
mutative with
G\
so that
we have
G
But
GI
,
~
l
TG = G M
l
G\" rG\
first
=G
since the left
member
of
the
equation belongs entirely to
the same
is
true for the right member, and a similar result holds
for the second equation.
Consequently
99
Gr FG^r,
l
G\- rG\^r.
Again there is no self -con jugate subgroup of tween G and / which contains the latter. For
group
G
if
intermediate be
there were such a
H with substitutions
l
ta
l
then
p
it
ta~ s fT taS p
= Ty
0~
t as
ta
would follow from A) that ~ = =

l
ta
p~
t as
t a <7
t&
that
commutative with G\. And since G and G\ is, must be commutative with together generate G, it appears that If now we add to the f a s the G. then the substitu 2 3
is
also
tions
form a group. we have from A)

a #0
(
For
since /
is
contained in
H and in G
19
a *p)
(V) =
& y tp* e
ts
*6
This group
nent groups
stitutions S
is
commutative with G, since

G\.
It contains
this is true of its
compo
H and
But
atfp.
It is
G\, which consists of the sub contained in G, which consists of the substitu
of G.
tions S aSpTy.
this is contrary to the
maximal self-conjugate subgroup

lowing
preliminary result:
assumption that G\ We have therefore the
is
fol
If in two series of composition of the group G, the groups next

succeeding
are respectively G^
take for the group next succeeding
and G\, then in both series we may G or G\ one and the same max
1
imal self-conjug ate subgroup

tutions
1\
.
which
is
l
composed of
all the substi
common
composition factors of composition, in the first series e\, in the second

89.
G and G\ If e and e\ are the factors of belonging to G and G\ respectively, then F has for its
to
l
We
can now easily obtain the
final result.
Let one
i)
series of composition for
G
9
be
G,
r,
Gy,0iv (%
r, = i\\e
2
...,
r3
i^
r-.e^
=r
:e 3
and
2)
let
a second series be
G,
r r
G\,G ,G
2
=
we can
construct two
Then from the
result just obtained,
more
series belonging to G:
100
3)
G,
G U /;J,#,
...
4)
G^G^J
and apply the same proof for the constancy of the factors of com position to the series 1) and 3), and again 2) and 4), as was employed
above in the case of the series
1)
and
2).
The
series
3)
and
4)
have obviously the same factors of composition.
The problem
agree to
is
now
reduced, for while the series 1) and 2) agree
only in their first terms, the series 1)
and
3),
and again
two new
2)
and
4),
two terms each.
The proof can then be

and
2) as before
carried another
series,
step by constructing from 1)

of
both
which now begin with G,

1
)
G, 6^,03,
r,
,,3,...,
?*!
:
i\
r.2
e2
=r =
r" 3
These
1)
series
and again 3
have again the same factors of composition, and V) and and 3) agree to three terms, and so on. )
We have then finally Theorem XXIII.

cases are
If a compound group
their order,
admits of two two
different series of composition, the factors of composition in the
identical, apart from groups in the two series is therefore the same.
90.
and
the
number of
From
6r
it
88 we deduce another
result.
Since
G~ FG
1
belongs to
1
n because
G~ G
i
G=G
G~
1
,
and
as a
also to
G\ because
of
G~ G\G= G u
l
appears that
TG,
common subgroup
is
is
G and G n must be identical with / so that 1 86 it follows that it subgroup of G. From

struct a
a self -con jugate

possible to con-
group & of order e^\ which
r
is (1
eie
-) -fold isomorphic
i
with G, in such a way that the same substitution of

to all the substitutions of
tt
corresponds
<r.
G
31
.
which only
differ in a factor
1,
._,,
We
5/j of
will
take now, to correspond to the substitutions
l ,
the substitutions
1, 5 ,, 5
3
,
. .
1,
ta.
21
ta
w/
t
of
t
2,
and, to correspond to the

1,
no case
is
% a = %p$y,
of G\, the substitutions

for the
<r
/2
/3
/^ of Q.
In
form the common subgroup of
101
/
s.
and G\.
Consequently the w
are different from the

:
Both
classes of substitutions give rise to groups

fil
and, since
= [1, "2, "3, %c& p = % y

ftZ a .<r
Vj,
,
<2.
= [1, "
i2
t
"
<J,
.
it
follows that
fl^
Sl
l
= fl^ ^
Moreover
every
s in (7 is
equal to %a& p ff yi that is
Q\.
We
obtain
&
therefore, by multiplying every substitution of Q^
by every one of Q\.

91.
We
consider
is
now two
successive groups of a series of
the same thing, a group G and one of its maximal self -conjugate subgroups H. Suppose that s\ is a substi m tution of G which does not occur in H, and let s\ be the lowest
composition, or, what
power of 8\ which does occur in

a factor of the order).
If
(in is either
the order of s\ or
to
is
a composite
number and equal

s
}
p q, we put s
2
j
s1
and obtain thus a substitution
which does not
occur in H, and of which a prime power s^ is the first to occur in H. We then transform Sj with respect to all the substitutions of 6r, and
obtain in this
way
a series of substitutions s
l ,
s2
SA
No one
sa
of
these can occur in H.
For
if
this
were the case with
= ff~
H.
Si<T,
then
ffs a <r~
= 81,
being the transformed of a substitution

substitution <r~
l
s a of
with respect to a We consider then the group
of G,
would
also occur in
T=
This group contains
substitution of G,
\H, s n
is
s2 ,
S A|.
H and
s i^.
contained in G.
If
t is
any arbitrary
we have
= Hs
r is
il
,v=r.
These three properties of r -H" is a maximal self-con
therefore commutative with G.
are inconsistent with the assumption that
jugate subgroup of G, unless

If
r and
H are identical.
substitutions, as s 15 s 2
,
.
we remember
further that
all
SA ,
which are obtained from one another by we have
transformation, are similar,
Theorem XXIV.
group
Every group of
the
its
the series of composition

(or,
of any group G, is obtainable from is obtainable from any one of
next following
every
maximal
self-conjugate
102
subgroups) by the addition of a series of substitutions, 1) which are similar to one another, and 2) a prime power of which belongs to the smaller group. The last actual group of a series of composi
tion consists entirely of similar substitutions of
92.
prime
order.
The following theorem
is
of great importance for the
theory of equations:
The series of composition of the symmet n group of elements, consists, if n > 4, of the alternating group and the identical substitution. The corresponding factors of com The alternating group of more position are therefore 2 and ^n\
ric
Theorem XXV.
than four elements
is
simple.
is
We
to
have already seen that the alternating group
maximal
remains
self -con jugate
subgroup of the symmetric group.
It only
The the alternating group is simple. and the theorem there is that of to 52, proof perfectly analogous obtained, when expressed in the nomenclature of the present Chap
be shown
that, for
> 4,
ter,
becomes: a group which is commutative with the symmetric group is, for n > 4, either the alternating group or the identical sub
It will
stitution.
be necessary therefore to give only a brief sketch
of the proof.
is a maximal self-conjugate subgroup of the and which affect consider the substitutions of alternating group H, the smallest number of elements. All the cycles of any one of these
Suppose that
substitutions
must contain the same number of elements
52).
The
substitutions cannot contain
more than three elements
in
any
cycle.
For
if
H contains the substitution

= (x,x x.,x, ...)..., = (x x x^), s with respect to
s
2
<r 2
and
if
we transform
which of course
s.
occurs in H, then 8~
ff~*s<r
contains fewer elements than
Again the substitutions of H^ with the least number of the ele ments cannot contain more than one cycle. For if either
Sa
= (X,X
2)
(x 3 Xi) ...,*
= fax&z) (X^X.X,)
S^-^S^ff,
sa
,
occurs in H, and
if
we transform with
Sa<7~ S a *,
l
respect to a
(x^x^), the
products
,
will contain
fewer elements than
Sp respectively.
103
of elements
The
substitutions
which
affect the smallest
number
are therefore of one or the other of the forms
The
first
case
is
tain a transformation.
impossible, since the alternating group cannot con The second case leads to the alternating
group
If
itself.
obtain the following series of composition 1) the symmetric group; 2) the alternating group; 3) G.2 [1, (x^) (ave 4 ),
4,
:
we
(X&) (x&J,
93.
(x&t)
(x^
3 )];
3)
G =
3
[1,
(x,x 2 )
0%)];
5) G,
= 1.
The
exceptional group
G.2 is
already familiar to us.
We may add here the following theorems: Theorem XXVI. Every group G, which is
in the alternating group
position
is 2.
not contained
compound. The corresponding factor group

based on
35,
is
One of
its
factors of
[(1 z
,
com
.
is
z.2
)]
The proof
of
is
Theorem VIII.
The
substitutions
first
which belong
to the alternating
group form the
self-con
jugate subgroup of G.
a of order p p being a prime number, the factors of composition of G are all equal top.
Theorem XXVII.
If a group
is
The group K of order p f obtained in 30 is obviously, from the method of its construction, compound. It contains a self -conjugate
subgroup subgroup
of
of order
M of
p f and this order p f ~\ and so

l
again contains a self -conjugate

on.
The
series of composition
K consists therefore of
K,
L,
the groups
M,
l
...
Q,
k
R, ... S,
1,
of orders
p f,p f
,p
f~ 2
,
.
.p ,P*~\
-P,
1.
The last corollary of 49 shows that we need prove the present theo rem only for the subgroups of K. If G occurs among these and is
If G does is already complete. not occur in this series, suppose that is the first group of the We series which does not contain G, while G is a subgroup of Q.
one of the series above, the proof
apply then to
the
the second proposition of

of
71.
Suppose that
is
is
common subgroup
and
and G.
is
Then
subgroup of G,
its
order
a multiple of
pa
a self-conjugate
l
and
is
conse-
104
.
a-1 or a The latter case p quently either p G would be contained in R. Consequently the theorem is proved.
is
impossible since then a-1 and is of order p

,
Theorem XXVIII.
self-conjugate
If a group
of order r contains a
H of order T then no substitution of Q which does not occur in H can be of an order prime to *
subgroup
e.
.Hoi the order e. construct the factor group the identical of the substitutions of 1\ except substitution,
order prime to
in
e.
We
FG
No
is
one
of
an
To any
<r
substitution s of
is
which does not occur
H corresponds a
G
which
different
from
1.
On
account of the
isomorphism of
and
the same power ^ SK 1, then at the same time
1\ there corresponds to every power s* of s is the lowest power of s for which of <r. If
/.
<?
= 1.
is
x is to
therefore a multiple of
e.
the order of
If,
ff
and consequently
not prime to
in particular, e is a
prime number, then the order of every
substitution of
94.
which
is
not contained in
is
divisible
by
e.
Among
the various series of composition of a group G,
the principal series of composition, or briefly, the principal series, is of special importance in the algebraic solution of equations. This
principal series
is
obtained from any series of composition by re
taining only those groups of the series which are themselves self-
conjugate subgroups of G.
Suppose the resulting

L,
series to
be
G,H,J,... K,
The
series of
itself
M,
1.
G may be the principal series. the case, for example, as we shall immediately show, tors of composition of the series are prime numbers.
Assuming that the principal
series,
i7,
This will be
if
all
the fac
series is not identical with the given
suppose that the latter contains, for instance between
H and
other groups, as
HI,
H_
v
lt
is
therefore commutative with
H, but not with
G.
.
Consequently
H~ H^H "= HI
* A.
G~ HI G =
Ti
Kneser: Ueber die algebraische Unauflosbarkeit hoherer Gleichungen.
Crelle
CVL,
pp. 59-CO.
GENEEAL CLASSIFICATION OF GEOUPS.

Accordingly,
of G,
if
105
the substitutions
we transform
with respect to
all
,
we
shall obtain a series of
groups
l ,
H\
H\
All of
these are contained as self -conjugate subgroups in H, for if <r is any l l ff substitution of 6r, then <r~ is contained in <r~~ H<r H. l
H =H
Moreover
for
if
is
any substitution of H, then from

1
^-V-V^v-
(cf.
36).
Again J is contained in every one of the groups For 7 is contained in fi^, and consequently ff~ j<r =
l
H\ fT J is contained
l ,
,
1?
in ff~
H = H\j and so on.

l
ff
conjugate subgroup
of
H.
For
,
by transformation with to then the intermediate between and would respect group from transformation with that between and H\, by proceed In the series of 6r, the group respect to 0~ may therefore be
HI
group between In fact if

.
H and
H\
Finally H\, like I?!, is a maximal selfif there were any self-conjugate sub
H\
then the same would be true of

.fi^
H and
l
is
obtained from
<r,
any one of the several groups of the same type H!,H\, ... All of these belong to the same factor e of composition, e being the
quotient of the orders of
H and H
H
In accordance with the pre
88, we can then continue the series of G by liminary result of the substitutions common taking for the group next succeeding
and H\, or to and -ET"i, and so on. and \, or to From the same result the new groups all belong to the same factor of composition s. Every one of them contains J. We need of
to
l l
l
If there is course consider only the different groups among them. this entire J. For the must coincide with only one, system of groups
TT " !}
TT -"
-"
TTff
VT " -"
is
and consequently the group common to all of them, transformation with respect to G. The order of 2 obtained by dividing that of by
unaltered by
is
therefore
But
in the
if
there are several different groups,
same way. The substitutions example, form a group which in the
we can then proceed common to HI, H\, H \, for

series of
can succeed the
106
group composed of the substitutions common to

corresponding factor of composition
After order of
v
and
H\
The
is
again
s.
repetitions of this process
we
arrive at the
The by e by dividing that of last system before J consists of v groups which are v_ v_ all similar and all belong to the factor s, and which give
therefore obtained
v
.
J is
H
.
group
J.
The
l ,
H= \H _ H
v
l ,
&""_!,
}.
Theorem XXIX.
and J of
s,
not coincide with a principal series, but

the latter,
l
If a series of composition of G does if, between tivo groups
1 groups
2
,
. .
.
H ,H>
l
...
H_
v
of the
former are
inserted, then to
H ,H.
belong the same factors of composition
and
the order r of
tiplied
series
by
v
of
G is therefore equal to the order r" of J mul can be obtained from J by combining imth J a ivhich are all similar, and of v_ v_ groups
v
.
l ,
l ,
the order r"s
Corollary I. If the factors of composition of a group are not all equal, the group has a principal series.
Corollary
it
II.
contains
any
substitution except identity
Every non-primitive group is compound if which leaves the several

If the group con
systems of non-primitivity unchanged as units.

tains greater (including)
itivity, it
and
lesser (included)
systems of non-prim
has a principal
series.
The
instance of the group
G=
[1,
Oi^) (aw)
2
#6
(X.X.X,) (# aWj),
shows that non-primitivity may occur in a simple group. case the only substitution which leaves the systems x ,x 2
l
In this
,
,x$,
and x
XG unchanged
is
the identical substitution.
Corollary III.
commutative,
i.
The groups
H_ H
v
lJ
_ ,H" v _
l
l ,
are
e.,
the equations hold
For
in
a
the
series
H_
v
{Hv _^,H
preceding J we to occur. _ v &\,

(
may assume
the sequence
Accordingly we must have
107
or
Corollary IV.
series of
is
last actual group of the principal one or more composed of groups similar to one another,
The
which have no substitutions except identity in common, and which are commutative with one another.
95.
We
have now to consider the important special case
where
s is
a prime
number p.
_
l
Instead of
ient notation
,H" v _
l ,
...
we employ now
",
.
the more conven
H
Then
tion
tfj ,
H",
H
is
H( v \
latter a substitu
H
91)
is
th
obtained from
of
J by
adding to the
first to
the
p power
which
the
occur in J.
We may
1).
write
H = t^J
Since
J" is
.
.
H"=tfJ,
H "=tfJ
-a
a
...
(a
= 0, 1,
...p
a self-conjugate subgroup of every one of the groups
H",
we have
l
tl
~Jt a = J,
/
t2
Jt.2
= J, _
a
t3
~a
Jt3 a =
^
ay
,
J,
, .
and,
if
we denote
ay
I/ a
J by
/
i2
i3
I/
that
is,
of
H",
-H"
of
S ",
and so
on, are
com
mutative
Since
of
among themselves, apart from a factor belonging

we can return from J
to
to J.
by combining the substitutions

group
(
and H",
for example, into a single
88),
we have from
94, Corollary III
t^t& = tfi
IrV ^.*/*,
1
and consequently, by combination of these two
results,
The
left
member
of the last equation
is
a substitution of
the
108
right
member
a substitution of
H".
only the substitutions of

disappear.
J in common,
Since these two groups have the powers of /j and t.2 must
1,
Consequently
= 1,
/?
and
The
substitutions of
the group formed from J, ^, and
are
therefore commutative
among themselves, apart from a factor be longing to J. The same is true of the group formed from J, t lt and f3 or from J, ,, and f3 and consequently of the group
, ,
\Jjtu
t.
and so
on, to the
that Corollary III of
group 94 involves much
itself.
(It is to
this.
be noted
less
than
There
it
was
a question of the commutativity of groups, here of the single
substitutions.)
Every two substitutions of

a factor belonging to J.
H are,
will
then, commutative apart from
We
e is
tion
If
two substitutions of
a
H are
H
prove
the converse proposi commutative apart from a fac

this will
now
tor belonging to J, then
the case,
if
prime number. In fact have this property.
be
For, this
being assumed, if s were a composite number, suppose its prime fac tors to be g, g #", We select from H\ in accordance with
,
. . .
Theorem XXIV, The lowest power

t1
.
91, a substitution
which
is
not contained in J.
of
which occurs in
will
then be, for example,
Transforming, we have -
(FJ)H
t
=H -
t*H
and, since by assumption,
H =H
J,
The group
contains
\t,
J\
is
is
therefore a self-conjugate subgroup of
which
and
is
larger than J. smaller than For, if t
J and
H
is
Moreover,
is
it
is
contained in
J, then
commutative with
from
37-8 the order of

assumption that
series of G.
is r"q < r"e. This is contrary to the \ t, J\ the group immediately following in the
Theorem XXX.
tion of G, the order r of
If,
is
in the principal series of composi obtained from the order r" of J by

is the
multiplication by
position for the
substitutions of
factor of com intervening groups in the series of G, then the are commutative among themselves apart from
,
where the prime number p
factors belonging to J.
Conversely, if this
is the case, the
of
the
composition of the
groups between
factors
and J are
all
equal to
same prime number p.

96.
We
is
turn finally to certain properties of groups in rela
tion to isomorphism.
If
a maximal
self-conjug ate subgrourj of G,

1\
corresponding group of
then
is
also
and A the a.maximal self-conjugate
subgroup of
F.
For if F contained a self-conjugate subgroup 0, which con tained J, then the corresponding group Tof G would contain L.
The
series of composition of
G and F are
equal
to the
corresponds to that of F. If simply isomorphic, all the factors of the one group are
corresponding factors of the other.
But
if
is
mul
cor
tiply isomorphic to 1] then there occur in the series for G, besides
the factors of r, also
a factor belonging
to the
group
S which
responds
to the identical substitution of F.

is
The proof
readily found.
If G is multiply isomorphic to F, then group of the series of composition of G.

97.
is
compound, and S
is
Suppose that
G
,
.
is
. .
any transitive group of order
r,
affect
ing the n elements x lt x.2 valued function r of a?,, # 2 ,,

j ,
x
.
lt
We
construct any arbitrary n\its
a?,,,
denote
different values
by
c2
..;, and apply
to
any
one of these, as
all
the substitu
tions of G.
Let the values obtained from
in this
way be
M>
Z-2,
^3J
^/"
The
r substitutions of
will not
change
its
this
system of functions as
a whole, but will merely interchange
individual members, produ
cing r rearrangements of these, which we may also regard as sub These substitutions of the c s, as we have seen, form stitutions.
a
new group
1\
The group F
is
transitive, for
contains substi-
110
THEOKY OF SUBSTITUTIONS.
tutions which convert
into
any one of the values
$ 1?
f>,
and therefore the substitutions of

1,
2
>
replace
(7
by any element
order of
Again every substitution of

f r
is
alters the
*
1,
2
a w!-valued function.
Consequently every sub
stitution of / also rearranges the
c 2 ,...,..
The order
r.
of
is
therefore equal to
its
degree, and
both are equal to
G and
are simply isomorphic.

,
For
to every substitution of
corresponds one substitution of / and conversely to every substitu And in the latter case it tion of F at least one substitution of G.
can be only one substitution of G, since

order.
and
are of the
same
Theorem XXXI.
To any transitive group of order r cor

transitive group, the degree
responds a simply isomorphic of which are both equal to r.

98.
and order
Such groups are called regular.

Every substitution of a regular
Theorem XXXII.
group, except the identical substitution, affects all the elements. regular group contains only one substitution which replaces a given
element by a prescribed element.

consists of cycles of the
Every one of
its
substitutions
same
order.
If two regular groups of the
same degree are

i.
(necessarily simply) isomorphic, they are similar
they differ only in respect to the designation of the elements. * Every regular group is non-primitive.
e.,
greater part oE the the theorem is already proved in the preceding Section, and the remainder presents no difficulty.
The
We
need consider
in particular only the last

/
,
two statements.
z.2
,
.
. .
Suppose that
fft
<7
,
with elements
r1
and substitutions
ff n
is
1
isomorphic to
,s.2
s\.
,
. .
.
with elements
x^ x
x n and
substitutions S
<TA
sn ,
corresponds
Then
into
|A
the isomorphism being such that to every in pairs we arrange the elements x a and
as follows.
Any two
x
{
of them,
x and ^ form the

l
first
ff\
pair.
If
then
SA
converts
x^ and
if
the corresponding
converts
into CA, then x^
and
form a second
arise in this way, for there is
inconsistency can pair. substitution which converts one only

if s^
z/,
.
No
into
sion
xa We have now to prove that, x a x b then A contains the succession

.
contains the succes
<r
*A
regular group of prime degree
is cyclical.
Ill
We
have
.
= 8a
Xl X n
8b
X,X b
8a
Sb
Xa X b
and since there

follows that
is
only one substitution which replaces x
(t
by x b
it
If therefore S A contains a cycle
composed
of a given
number
of ele
ments x a then
.
<TX
contains an equal cycle composed of the corres

.
Therefore S A and o- A ponding elements c a and r are of the same type. The last part of the theorem is proved as
and consequently
follows.
1
6r
If a regular
.
group
then
G
it
t
= (x x
t
cannot also contain

l
(x^
,
.)
should then have st~

regular group.
= (x
X M ) (X M + x m + 2 ) x For we (x 4 m+2 ..)...

. . .
.
(x m + l x^ ...)...,
.
and
G would
not be a
x m i. e., the elements of any Consequently x l1 x.2 form a of arbitrary cycle, system non-primitivity. (The remaining systems however are not necessarily formed from the remaining
. .
cycles of the
99.
sitive,
same
substitution).
If the
if
groups
and F are isomorphic, and
if
is
intran
suppress all elements which are not transitively connected with any one among them, as x the remaining portions of the several substitutions form a new
then,
in every substitution of
l ,
G we
transitive
group
also isomorphic with /
It
l
may however hap

increased.
pen that the order

if a? 2 is
of
isomorphism of
to
is
Again,
l ,
any new
element, not transitively connected with

G.2 ,
x
/
we can
then form a second transitive group

taining x>,
isomorphic to
and con
and so
on.
The
intransitive
group
can therefore be decomposed into a
system of transitive groups isomorphic with 1 \ and conversely every intransitive group can be compounded from transitive groups
6r 1? Gr 2
,
.
.
In the case of simple isomorphism
it
is
only neces
sary to multiply the several constituents directly together.

100.
order r
V.
= mn^
G
Suppose that which is

,
G
A:
is
-fold
a transitive group of degree in and isomorphic with a second group

x.2
,
.
Let the elements of
be x l9
affect
group of
which does not
x m and The order

.
let
of
G G
be the sub
is
therefore
112
Wj
x.2
,
.
If
s 2 , s3
.
.
are
l
substitutions of
(7j
which convert
x^
into
XA
then
s.2
s3
comprise in each case all
and only
the substitutions which produce the same effects. Suppose that to G in G corresponds I\ in /
l
the order of
1\
being m^k.
<r
The order
/
of
c^
tion
belongs to
all
n then
Consequently, if the funcrk takes exactly values under the

is
rk.
Tfl ]K
operation of
tion
ff.
the substitutions of 1\
Suppose that the substitu
of
F which
all
corresponds
to s 2 of
G
4
converts c^ into
. 2
-
Then
Simi
F
<f>
ff.
contains
the substitutions which convert
into
lar considerations hold for #3
and p 3
<r
,
and ^ 4 and
,
so
on
to c m
and
If
we apply
all
?1,?2,
-
to the
system of values
?,,
we obtain rearrangements which can be regarded as substitutions of The order of the new group H is equal to the the new elements quotient of kr by the number of substitutions of / which leave all
tp.
the
s unchanged. which leave all the

>
These correspond
aj
to
tion.
To
and
this
e., unchanged, correspond k substitutions of

i.
to the identical substitu

/
,
order of
G
of
H H are, then, the same degree m

is r.
,
and consequently the
l ,
of the
if
same order
s is
r,
and
they are isomorphic and, in fact, similar.
For
a substitution
which replaces x$ by x a then

substitution of
The corresponding
stitution
H
15
~
<7 a
belongs to the system s a G sp. is obtained by applying a sub

l 1
, .
.
l\ ff p to the

^2
y>
Every one of these
substitutions replaces
by
only differ
from those of
Accordingly the substitutions of in the fact that the latter contain a? s

.
H
F
where the former contain the corresponding ^
s.
We
can therefore construct

all
all
groups
/
(or
isomorphic to
<?
by applying
/
-,
to
any function
belonging to
any arbitrary subgroup

If
of
and noting the resulting group of

>
substitutions of the elements pi, ? 2
<Pm/
,
is
a self -con jugate subgroup of
group H will be regular,

101.
the resulting isomorphic
as is easily seen.
In conclusion we deduce the following

Given any number of mutually mul elements of any one are
Theorem XXXIII.
tiply isomorphic grotqis, in ivhich the

all different
113
from
those of
any
other one, if
we multiply every sub-
titution of the one
group by every corresponding substitution of
every other group

is
an
intransitive group,
and form all the possible products, the result and conversely every intransitive group
this
can be constructed in
way.
is
The
first
part of the theorem
sufficiently obvious.
For the
second part we consider the special case of an intransitive group the elements of which break up into two transitive systems. The
general proof
obtained in a perfectly similar way. Suppose that the substitutions s\ of the intransitive group
is
divide into two components

SA
ff
\~\
where
I 1? f_
<r
affects only
*M-
the elements
x^
,oc<>,
...
tll ,
and
rA
only
It is possible
that T A occurs also in other combina
tions
Similarly
rA
may
occur in other combinations

Â T A
<>
?
ArA
&
A TA
We
coordinate
<r
,
now with
.
. .
<T A
all
the r A -\, r" A

,
and with
T A all
the T A
<r" A
stitutions S A of G.
T.
and proceed in the same way with all the sub The <>\s form a group - and the r A s a group
A
,
Suppose that
<7
<v are coordinated with r A -^

, , ,
Then there
are substitutions S A s^ s v
,
such that
S\Sfj,
=S =
v
ff v
~
v
,
and consequently
<r
is
coordinated with
rA
ja
=r
CHAPTER
V.
ALGEBRAIC RELATIONS BETWEEN FUNCTIONS BELONGING TO THE SAME GROUP. FAMILIES OF MULTIPLE-
YALUED FUNCTIONS.
102.
It
tion there belongs a
has been shown that to every multiple-valued func group composed of all those substitutions and
only those which leave the value of the given function unchanged. Conversely, we have seen that to every group there correspond an
infinite
is
number
of functions.
The question now

to the
to be considered
is
whether the property of belonging
same group
a funda
mentally important relation

this property implies
among
functions; in particular, whether
corresponding algebraic relations.

is
An
instance in point
that of the discriminant -L of the values
in Chapter III, It was there shown 55. cr, considered the of the consideration from group belonging to c>, that JA, simply and therefore the corresponding discriminant of any function belonging to the same group, is divisible by a certain power of the
of a function
discriminant of the elements x

103.
1 ,
tl
We
shall prove
now another mutual
relation of great
Importance.
Theorem
Suppose y
l
I.
Two functions belonging

to
to
the
same group
can be rationally expressed one in terms of the

and
r
v
i
other.
be two functions belonging to the same
group of order
and degree n
6r,
[<$!
=
t
1, S 2
S3
S,.].
If T, is
values which proceed from

.all
any substitution not belonging to G and if cr. and c cr and ^ by the application of
{
are the
,
then
the substitutions
,,
S,T,,
</-,
S a <7a
c -.,
8,.<7.2
:also
convert y and
l
c.,
into
and
respectively,
and these are the

values
c.2
-only substitutions which produce this effect.
The
and
vS
FUNCTIONS BELONGING TO THE SAME GROUP.

therefore again belong to one and the
115
6r.,
same group
/>
<r<
n.,
Proceeding in the same way, we obtain all the <> and v together with the /> corresponding
;
pairs of values of
groups.
For every integral value of

r AS
is
/,
the function
. .
.
+ rL.V
-f
.
.
-f
therefore, like
^ -f
?.,
+
all
<r
+ ?pVp = A or ^ -f +
0.2
,
. .
+
x
tt .
p,
an
inte
this
gral symmetric function of the elements x,
x.,
For
function
is
merely the sum of
the values which e^V can assume

i
and
is
accordingly unchanged by any substitution, only the order

Accordingly, if ,T A then A A
,
of the several terms being affected.
cr,
and
are
integral rational functions of the elements

rational function of
c,
is
an integral
lt c.,,
tt
Taking successively ponding equations:
= 0,
1, 2,
;>
1,
we
write
the corres
01+
^)
0+
ftV,+
01+-..
rf0.
0p
= 4a,
= ^,
fiV,+
..+
fpVp
If these equations are solved for as a rational function of cr u cr.,,

.
c -,, c\,,
.
^ P1 every c\
is
obtained
yp
/>
104.
We
multiply the
first
2 equations of the system S)

.
successively by the undetermined quantities y^ ?/ 1? y, yp _ 2 and the last equation by y p _ l and add the resulting products, wri
. .
ting for brevity
AVe obtain then

(1)
^ X (?i)
i
+ ^ X fe)
F
From
For
this equation
this
we can eliminate ^ ,^ we need only select the y s purpose

a
j
^p
and obtain
<}\
so that
we have simul
taneously
(to)
= 0,
/ (^ 8 )
= 0,
/ (<fp }
= 0;
116
III,
In Chapter
53,
/>
we have shown
V(c )
that
c?
^.2
c p satisfy an
equation of degree
= 0,
cn
c.,,
.
.
the coefficients of which are rational in
c,,.
Again, the
quotient
vanishes
if
<f>
c-,, c.,,
<?
But,
t)
-
if
=
? P)
cr
n we have
(<ri
?2)(?i
is
.
(?i
if
= A OO.
.
The
derivative A"(<0
,
not zero, for

.
x^x.,,
xn
are independ
ent, the values p,
?.2
c?
are
all different.
We
can therefore satisfy the requirements above by taking
that
is,
Or,
if
we
write
we have
and consequently
r
^/P-J=
Up-.i
= *\
0i
jî-h2
By
(2)
substitution in (1)
we obtain then
^(?,) =#(?,),
to ? simpler
as follows.
The value of c^ thus obtained can be reduced The product
form
X
is
()
a symmetric function of the

(cr,)
s,
and
in fact, as
appears from the
expression for A"

in algebraic sign.
(*)
is
above, only differs
from the discriminant J^
Moreover, the product
X (^X (c
)...X
symmetric function of the roots of the equation
117
and can therefore be rationally expressed

of this equation, that
is,
in terms of the coefficients

2
,
.
in terms of a n
c u c2
, .
.
pj
and
<f l
and con
sequently in terms of
cn
and ?
If
now we multiply
4\ in (2)
numerator and denominator of the expression for

product
(4)
(3),
by the
we
obtain
*=SM
of this last fraction
is is
The denominator
c,
,
rational
and integral
in c
l ,
in
cn
c2 ,
tl ;
the numerator
rational
and integral
c2
and
9",.
If the
numerator R^fr)
cr
is
of a degree higher than

is
(>
1 with
respect to
that
n a
still
further reduction
possible.
For suppose
where
Q(<p)
and
R (v)
2
are the quotient
and remainder obtained by
dividing R^(<p) by X(<?).

P
1.
The degree
cr
2
, . .
.
of
R (<p)
2
then does not exceed
Now
if
= 0. n -X"(?) RiW = R*M (* = 1,2, 8,cr
9"
<f p
Consequently
/")
and therefore
Jo
&,>.,. -..W
have therefore
We
II. If two p -valued functions <p^ and belong to same group G\, then v\ caw be expressed as a rational function of ivhich the denominator is the discriminant J^ and is therefore rational and integral in c,, C 2 while the numerator is an c,,,
<} *.
Theorem
the
integral rational function of y^, o/ a degree not exceeding p

iwYJz,
1,
coefficients
which are integral and rational in c n

of
c2
c,,*
105.
The converse
Theorem
is
proved
at once:
Theorem
III. If two functions can be rationally exone in terms of the other, they belong to the same group.
*Cf. Kronecker: Crelle9l, p. 307.
118
In
fact,
given the two equations
y
it
= R (f),
l
.R,(cr),
appears from the former that

c
<?
is
unchanged by
9?
all
substitutions
</-,
which leave
unchanged, so that the group of

it
contains that of
while from the latter equation contains that of cr. group of

<f
appears in the
same way that the

are therefore iden
The two groups
tical.
Apparently the proof of this theorem does not involve It must shall be rational functions. and the requirement that
REMARK.
<f>
</>
however be
be
fulfilled.
distinctly understood that this requirement
must always
For example,
in the irrational functions
the expressions under the square root sign are all unchanged by But it remains entirely uncertain the transposition a (x^x^). whether the algebraic signs of the irrationalities are affected by this
Considerations from the theory of substitutions alone cannot determine this question, and accordingly the sphere of appli
substitution.
cation of this theory

If,
is
restricted to the case of rational functions.
in the
last
two
irrationalities
above, the roots
are actually
extracted and written in rational form

(x l
it
x.2 ),
(#,
#.,),
<r
once that the transposition appears the former expression but leaves that of
at
first
while in the case of the undecided.

106.
changes the sign of the latter unchanged, irrationality this matter is entirely
Theorems I and III furnish the
basis for an algebraic
All on the theory of groups. rational integral functions which can be rationally expressed one in terms of another, that is, which belong to the same group, are
classification of functions resting
regarded as forming a family of algebraic functions.

l>
The number
is
of the values of the individual functions of a family
called the
The several families to which the different order of the family. values of any one of the functions belong are called conjugate
families.*
*L. Kronet ker: Monatsber.
d. Berl.
Akad.,
1879, p. 212.
119
The product of the order of a family by the order of the cor responding group is equal to n!, where n is the degree of the group.
Every function of a family of order

of degree
/>,
>
is
a root of an equation
cn ca
,
. .
the coefficients of
/>
which are rational in
c ti
The remaining
tions.
1 roots
of this equation are the conjugate func
The groups ivhich belong
to
conjugate families have, if
>
>
2,
n >4, no common substitution except
For
For
five
(>
=2
the two conjugate families are identical.
ft
= 6,
=4
there is
a family which
is
identical with its
conjugate families.
107.
In the demonstration of
104 the condition that

tions
which leave the value of

still

if
some
all
of the values of
should
coincide; but the values of
be different, as appears, for example, from the presence of the discriminant J in the denomi
nator of
4>
</
must
includes that of ^
.
Under the more general condition we have then the following
that the group of
If a function ^ is unchanged by all the sub group of a second function y>, while the converse is not necessarily true, then ^ can be expressed as a rational function of <p, as in Theorem II.
stitutions of the
Theorem
IV.
the family of the function $ is said included in the family of the function <p. can be rationally </ in terms of in but cannot be thus expressed c?, y expressed general in terms of ^. An including group corresponds to an included fam
to be
Under these circumstances
and vice versa. The larger the group the smaller the family, the same inverse relation holding here as between the orders r
ily
and
p.
From the preceding considerations we further deduce the lowing theorems:
fol
120
Theorem
expressed.
V.
It is
always possible
to find
a function
in
terms of which any number of given functions can be rationally

This function can be constructed as a linear combina
Its
tion of the given functions.
family includes
all the families
of the given functions.
Thus any given functions y

in terms of
to
/,
can be rationally expressed
= ae-\--ti!>-\-Y7.+
>
where
/J,
^,
are arbitrary parameters.
composed
changed,
<f
,
of those
substitutions which leave p,
For the group of w is ,/,... all un

<,
and which
.
are
therefore
of
(a
common
</
/,
The group
c,
is
therefore
the groups of contained in that of

to
every function
<!>
,/,..., and the theorem follows at once.

the
special case occurs when the group of w reduces to identical operation, ut being accordingly a n!- valued function.
this case
In
every function of the n elements
x\
x.,
X H can be
contained
rationally expressed in terms of w,

in that of
at.
and every family
is
The family
of
o>
is
then called the Galois family.
Theorem
elements
x lt x z
VI. Every rational function of n independent x n can be rationally expressed in terms of every
.
.
nl-valued function of the same elements; in particular, in terms
of any linear function ?
,a?i
2#2.
+,,#,
means
of expressing a
9
where
i ,
are arbitrary parameters.
108.
We
attempt
now
to find a
mul
,
tiple-valued function c in terms of a less valued function
the
group of the former being included in that of the latter. A rational solution of this problem is, from the preceding developments, im
possible.
The problem
in
is
an analogue and a generalization of that
treated
53, where a ^-valued function was Chapter III, expressed in terms of a symmetric function by the aid of an equa tion with symmetric coefficients of which the former was a root.
From
the analogy of the two cases
we can
state at once the
present result:
121
<?
Theorem
VII.
If the group of a
&,
mp -valued function
and
if
is
contained in that of a p-valued function
are the
values tvhich
<p
takes under the application of all the

these
substitutions which leave

the roots of
$ unchanged, then
values of
<p
are
an equation of degree m,
^.
the coefficients of
which are
rational functions of
In
fact, the substitutions of the group
G =
1
[X
= 1, a
a ,
83,
sj
=<p l
c?
,
of
</
are applied to any symmetric function of

is
c%. the
value of this function
unchanged, only the order of the several
terms being altered.

functions
In particular we have for the elementary
where the
V
/;
We
s are rational, but in general not integral functions of obtain therefore the equation
(A,)
 an d in general the equation

t
(A,)
A,
(^^-
H-
A (^^""
of which the roots are the in values of ^ which correspond to the
value
<r
of
c -.
The denominators
denominator
is
of the
Ax s
and, in fact, their least
common
always a divisor of the discriminant J^, as appears from the proof of Theorem II. If is a symmetric function, there
is
no longer a discriminant, and the denominator

53.
is
removed, as we
have seen in Chapter III,

109.
One
special case deserves
particular
notice.
If
the
included group
H of
of
the function
then,
if
<f>
is
commutative with the inclu
ding group
a single root of the equation

rationally determined.
(A
) is
known, the other roots are
all
For
if
122
if
are the roots of the equation (Aj), and
H
^
H.,,
H
<r
ltl
are the groups belonging to these values, finally,
if
= 1,
<r.,,
<T
w
l
are any arbitrary substitutions of
G
l
which convert c into

45)
.
<f l
^2
,
Hm = *- #>,.
6r,
But by supposition
fore
is
a self-conjugate subgroup of
and there
G H\ G = HI
~~
that
is,
and consequently
H = H. = H =
l
...
=H
al
The
??i
different values î
>
^2 >
same group H, and can consequently
^. therefore belong to one and the all be rationally expressed in

I.
terms of any one among them, in accordance with Theorem
The family
of
</
is
included in that of
l
When,
as in the
not merely contained in the present is G of a but </ group self-conjugate subgroup of G, the family of is called a self-conjugate subfamily of the family of ?, ^S
case, the
,
group H
of
<p l
is
VIII. In order that all the roots of the equation should be rationally expressible in terms of any one among (Aj) as it is necessary and sufficient that the family of vS y\ them,
,
Theorem
should be a self-conjugate subfamily of that of ^, *. e., that the should be a self-conjugate subgroup of that of </ The group of
(f-t
]
groups of
(^
y>.2 ,
<p
M are then coincident.
is a prime number. consider in particular the case where & G be the and of to of that Since every v\ Jf^ Suppose group
1 l
.
We
substitution of
Gr,
,
ues
v
s.
f,,
$r 2
p, M
the group
is
produces a corresponding substitution of the val G is isomorphic with a group of the

1
The
II,
latter
group
it
transitive
orem
Chapter IV,
its
order
is
divisible
and of degree m. From The by m, and from Theorem
X, Chapter III,
therefore contains a substitution of order m.
123
For
in
elements, where
is
prime, there
is
only one type of such
substitutions
The corresponding
i9<P2i
>
substitution r of
G
n
,
therefore
permutes
f", it fol
<Pm cyclically.
Moreover, since
the functions
corresponds to
...
<p tll
lows that
cordingly
T" T"
(
leaves
,
all
^2
unchanged.
Ac
stitutions of
and no lower power of

}
r, is
contained
among
the sub
Furthermore we readily show that
01=JH,,Tj.
For the substitutions
M
.#! is of
f
19
H,T,
H
!
~
f
are
all different
and, since
are
n\
all
order
there are
yr
- of
ft
them.
mfi essarily contained in

6r,,
m<>
They
order
nec-
which,
being
itself of
cannot
r
contain any other substitutions.

is
From
this
it
commutative with
H
v
appears again that
Theorem
and
M H^
if the
IX.
group
<J> 19
group
>*
of
p
If the equation (A^ is of prime degree m r of v is a self-conjugate subgroup of the then G contains a substitution r which permutes
cyclically.
This substitution
is
commutative with
;
its
tn
,
and no
lower,
power
l
.
is
contained in H^
together with
it
generates the group

110.
We
examine now under what circumstances (Aj) can
become a binomial equation, again assuming the degree prime number. If (A ) is binomial, its roots, ^ n wcr n >Vi>
t
m to
a) "
be a ~ Vi
evidently
that
It is therefore necessary belong to the same group. should be a self -conjugate subgroup of 6^.
all
We
proceed
now
to
show conversely
6r,
,
that, if the
/,
group
jff,
is
self-conjugate subgroup of
then a function
belonging to
l
.
H-^
can be found, the
mth
power
of which belongs to
Denoting any primitive
ra th root of
unity by w,
.
we
write
%\=<P\-\- M(Pi
If
+ W Vs +
+W
"
V;w
t
we apply
to this expression the successive
powers of
or
r,
we
obtain
124
x-2
v-2
^
4j 1
^
T"
4j
I
"V* 4o
[
4i
1
A2
-i
consequently
We
2) that
have now to prove

*
1) that
/ belongs to the group

t
.
and
/!*

first place,
In the
since ft,?a ,
is
. .
?,
l
are
.
substitutions of
the same
true of /
}
unchanged by all the Moreover if there were
any other substitutions which

example,
left
y unchanged
we should
have, for
and therefore
The
quently
latter equation
all its roots, in
would then have one
of its roots,
and conse
common ~ m * l *m
with the irreducible equation

..
-*+
.+1 = 0,
and we must therefore have
But we may assume the function method

of
Cj to
have been constructed by the

of the
31 as a
sum
n\ of
-
- terms
mp
form xfxf
with
.
.
undetermined exponents. The systems of exponents in cr M c 2 <? M will then all be different, and therefore, since the .r s are independ
,
.
ent variables, the equation
can hold only
if
^ = <p^
and
cr.>
<p
identically.
The function -/ A
this function is
therefore belongs to
It follows at
unchanged by
once that //" belongs to

}
For
and
r,
and consequently by
!
Qi={H
No
"
l9
T\.
other
substitutions can leave //" unchanged.

less
For otherwise
-/
/,
would take
than
;>
values,
and
its
m th root
A less than
mp
values,
which would be contrary
to the result just obtained.
125
Theorem X.
In order that
H
a
of
may
contain functions the

it is
m th
family belonging to a group power of which belongs to the

the
family of a group G,
self-conjugate
necessary and sufficient that
H should be
subgroup of 6r, or, in other words, that the family should be a self-conjug ate subfamily of that of H.
From Theorems IX and

ter is readily
the following special case of the lat
deduced
Theorem XI.
In order that
the
prime power
p
<f
of a
p<>-
valued function <f mayjiave f> values, it is necessary and sufficient that there should be a substitution r, commutative with the group f f which the p ih power is the first to occur in H.
>
Finally an extension of the last theorem furnishes the following
important result:
Theorem
is
XII.
If the series of groups

G,
G G
l ,
(7 3
Gv
the following
~
a
so connected that every
Ga _
can be obtained from a substitution
Ga
by the addition to the latter of

6r a
,
commutative
occur in
.
.
with
of which a prime power, the

it is
pa
th
,
is the first to
f>
6r a
then
and only then

a
possible to obtain a
p^
-p.,
p v -valued
function belonging to
the solution of
G v from
series of
a j>- valued function belonging to G by binomial equations. The latter are then
respectively.
of degree
p^p^p*, ...#,
In the expression of a given function in terms of 111. another belonging to the same family, we have met with rational fractional forms the denominators of which were factors of the dis
criminant of
r 1?
.r>,
.
the given
function.
If
xn
as independent quantities, as
we regard the elements we have thus far done,
the discriminant of any function a is different from zero, for the have different forms. But if any various conjugate values of
<f>
relations exist
among
Ihe elements
.r,
it is
no longer true that a
dif
is
It ference in form necessarily involves a difference in value. therefore quite possible that if the coefficients in the equations
f(x)
=x
lt
c,aj*"
+ Cj*"".
-...
c,,
are assigned special values, the discriminant
126
may become
zero.
If this
were the
case,
c"
could not be employed
in the expression of other functions of the
same
family.
And
it is
conceivable that
the
might vanish.
It is
discriminant of every function of a family therefore necessary, in order to remove this
uncertainty, to prove
Theorem
other relations
XIII.
exist
If only no two x*s are equal, then whatever
may
among
the
a? s,
there are in every family
functions the discriminants of
which do not vanish.
30. It is however proof might be given similar to that of more convenient to make use of the result there obtained, that
under the given conditions there are
still
w!- valued functions of the
form
tf
-f~ a^ji\ -f- a.2 x.,
-}-...+
s
,,a? w
We
suppose the
and the
,x
to
be free
to
assume imagin
This being the case, if the ary (complex) as well as real values. n\ values of c are all different, we can select the coefficient ,,, so
that the moduli of the values of
FA
<?
are also
(*
all different.
2, ... n),
For
if
- WA 4- ,A V^T
= 1,
then we can take
in such a
VA
way
that
all
the n\ quantities
= VA + <=
(WA
(WIA
+p)
+ (,A + q) V 11"!
For from
tf)"
=X
2,
n)
shall
have different moduli.
+ p)
-f
(,"-A
-f
= (w* + P)" -h (/^ -h qY

K
,
it
would
follow, if
p and g
WA
are entirely arbitrary, that
=W
In fact we can, for example, take
= /^. p = q and
//A
1
q so large that even
special values of q satisfy the conditions.
Suppose then that the tVs are arranged

tudes of their moduli
V i>
in order of the
magni
^ 3,
^ n\
e so
.
(mod.
^ > mod.
(/
<J
).
We take then the

fa
integer
great that
>
(tf A + ,
-K A+ +
-hv\/)
= 1, 2,
(Wl
1)).
THE NUMBER OF VALUES OF INTEGRAL FUNCTIONS.
127
From
</0
it
every equation of the form
= vV + vV 4- vY 4- ... b ^ c = follows accordingly that a = If now we apply

>V + <rV + vV +
.
.
. ,
,3
-f
the r substitutions of
G to
1
</
/>
and add the
results, the
sum
= # *+ 0* +iV+...+fc;
is
a function of the required kind.
For
in the first place
to
is evi
And in the second place the properties of dently unchanged by G. the equation / ) show that M has f> distinct values, and consequently J,., is not zero.
CHAPTER
VI.

112.
Thus
far
we have obtained only
ocasional theorems in
regard to the existence of classes of multiple-valued functions.
We
are familiar with the one- and two-valued functions on the one side
and the
n\ -valued functions
on the other.
But the possible
classes
lying between these limits have not as yet been systematically exam An important negative result was obtained in Chapter III, ined.
42,
where
it
was shown that
a divisor of n\.
to us.
take any value which is not // cannot Otherwise no general theorems are as yet known
can, however, easily obtain a great number of special the construction of intransitive and non- primitive groups. by But these are all positive, while it is the negative results, those
results
We
which assert the non-existence of classes of functions, that are pre

cisely of the greatest interest.
The general theory

morphism
Thus,
if
of the construction of intransitive groups

in
would require as we have seen

in its broadest sense.
101, a systematic study of iso
We
shall content ourselves there
fore with noting

if
some
of the simplest constructions.
there are
= a + &.+ c +
elements present, and
the symmetric or the alternating group of a of them, the symmetric or alternating group of b others, and so on, then on
we form
multiplying
all
these groups together,
we obtain an
intransitive
group
of degree
n and
of order
r
a 6 c
! !
where
1,
J,J,
g,
groups employed
all
in the construction is 0,
according as the number of alternating the rest being 1, 2, 3,

.
symmetric.
For the number of values
of
the corresponding
functions
we have then

.
129
. .
we can By distributing n in different ways between a, b, c, obtain a large number of classes of functions. For example, if
,
n=
5,
we may
take
e
e e
e
=5 a = 5; a = 4, b = 1 a 4, b = 1;
a
;
= 1, = 1 ^ = x^x.^x^. = i, p= 2; ? = (x x )(xi = 1, = 5; = x^x^x^ = J, = 10; c = (^ x (x

ft
;
x3 )...(x 4
a? 5
/>
y> 3
/>
2)
(x2
x3 ) (aj 2
a? 4 a? 5 .
= 3, b 2 a = 3,6 = 2;
;
= 1, = J, = |,
/>
/>
= 10; = 20; = 40;

= 20;
x ) (x x ) (x
3 4 3)
x) xt )
^3)
c^
^6
= x^x^ -f = (i ^2) fe = (x
x2) (x
a?4
(# 2
a
a
= 3, 6 = 2;
= 3, b = 1,
c
/>
^8
a? 3 )
(aJ 2
a? 3
+
=1
;
a? 5
= 1,
/>
v7
!*
= î
way
to the con
The imprimitive groups give

we may take any two systems
each, or
rise in a similar
struction of functions with certain values.
For example,
for
= 6,
of non-primitivity of three elements
any three systems of two elements each, and with these construct various groups, the theory of which depends only on that
of groups of degrees
113.
two and
three.
General and fundamental results are not however to be
obtained in this way.

different side,
We approach the problem therefore from a which permits us to give it a new form of statement.
<^ l
the familiar table of

9>l5
Given a />-valued function 41

:
]
with a group
G we
}
construct again
=1,
..r
G\
We
proceed then to examine the distribution of the substitutions of
a given type
among
the lines of this table.

1 transpositions
(a?j.r a ),
(
A)
There are n
9
= 2, 3, ...
n).
If
130
then
//
< n,
and
if
the group
of
<f> l
does not contain any trans
position of the form fax*), these (w
1) transpositions are distribu
ted among, at the most, (n

line after the first
2) lines of the table. at least
must contain
Accordingly some two of them. Suppose these
two are
Then
appears that a combination of the two (x&aXt) = (aj,a? ) fax*) s^^s^)- = sa <rA <rA
it
s j8
= s^- = s
!
occurs in
6?, /> w, 6r, contains either a transpo Consequently, sition or a circular substitution of the third order, including in
.
if
<
.
either case a prescribed element
The same
is
obviously true of
any prescribed element # A
B)
There are -
fyi
(ryi
_ =
1
J
transpositions of the form
(a53J/i),
(=f=^
of the
= 1,2,... n).
If therefore
//
<
and
if
the
first line
table does not contain

tains at least two.
If these
any transposition, then some other line con have one element in common, as (x a xp),
we have seen in A), their product (x axpx y ) occurs in have no element in common, as (x axp), (XyXs), then In either case G their product (XaXp) (x y xs) also occurs in G therefore contains a substitution of not more than four elements.
(x a x y ), then, as
If they
C)
(a--=-/3
There are (n
1) (n
2) substitutions of the 1) (n
2),
form
(x^XaXp),
if
=
no
2, 8, ... n).
If therefore f <^(n
and
con
tains
siibstitution of this form,
some other
line of the table
con
tains at least
two of them.
combination of these shows that
contains substitutions which
affect three, four, or five elements.
Proceeding in this way, we obtain a we present here in the following which of
series of results, certain
Theorem
is
I.
1)
If the number
1,
/>
of the values of a function
not greater than n
the
group of the function contains

yiifi v
a sub
stitution oj\
at the most, three elements, including
_
Q &
any prescribed
group of the
3)
~\\
element.
2)
If
/>
-is
not greater than

of,
-,
the
a substitution function contains

If
f>
at the most, four elements.

,
is
not greater than -
the
group of the function

greater
of,
131
at the most, six elements.

2)
. .
.
4) I/ p is not
than n(n
l) (n
(n
+ l)
~k~
he 9 rouP
f the func
a substitution of, at the most, 2k elements. 5) If p is not greater than (n 1) (n 2) ... (n &-f-l), the group of the a substitution contains at the 1 elements, function of, most, 2fc
tion contains
including any prescribed element, so that the group contains at least
r-
such substitutions.
the aid of these results the question of the number of values of functions is reduced to that of the existence of groups contain
By
ing substitutions with a certain

114.
minimum number of
elements.
In combination with
earlier theorems, the first of the
results above leads to
an important conclusion.
From Chapter
intransitive
IV,
is at
Theorem
I,
we know
that the order of an
the most (n 1)!. group Consequently, the num ber of values of a function with an intransitive group is at least
(nl}\
so that the
n\
^rr-
= n.
For such
a function therefore p cannot be less than n.
(n
number
least
V
J
~9~
eroup
is at
-n\
J
of values of a function with a non- primitive

.
For n
4, this
number
is less
than n:
2!*.!>!
"
but for n
> 4,
it
is
greater than
n.
n.
For such a function then,
if
Again for the primitive groups it follows from Chapter IV, Theorem XVIII, in combination with the
p cannot be less than
> 4,
the corresponding 2 or 1. The group is either alternating or symmetric, that is, p r is already known 8 n for which 4, p 4, non-primitive group
first
result of
Theorem
I,
113, that
if
<n,
to us,
46).
We have
II.
then
is
Theorem
less
If the number p of the values of a function

p = 1
than
n,
then either
or p = 2,
and
the
group of the func
tion is either symmetric or alternating.
An
exception occurs only
for n
ing tO
= 4, p = 3, r = 8,
X
the corresponding
group being that belong
132
115.
On
(f
account of the importance of the last theorem

different grounds.
/>
we
add another proof based on

Suppose
to be a function with the

?1
?:>
<
n values
?3
<?P-
If
we apply any
substitution whatever to this series, the effect will

//
be simply to interchange the
values
among
themselves.
If in
l
particular the substitutions applied belong to the group then the // values will be so interchanged that retains
of
 (n

1 values
(/>
<? p
Since
(>
<
n,
there are at the most,
only
the r
r,
_<_
(n
2)
such rearrangements.
Consequently among
>
(n
1)! substitutions of
there must be at least two, a and

<?.
, ,
.
.
.
which produce the same rearrangement of 2 y 3 y> p (?T~ is a substitution different from identity, which leaves
l
Then
all
the
s
,
unchanged, that
is,
which occurs n
in all the conjugate
groups
G-i
III.
no such substitution (Chapter Theorem XIII). Consequently /> _> n 1 16. Passing to the more general question of the determina
2
,
.
.
Gp
But
if
>4
there
is
tion of all functions

limit
whose number of values does not exceed a given

??,
we can dispose once for all of the less impor tant cases of the intransitive and the non primitive groups. For the purpose we have only to employ the results already obtained in
dependent on
Chapter IV. In the case of intransitive groups we have found for the maxi
mum
1)
2)
3)
orders:
r(n
r=
r
1)!.
Symmetric group
of
1 elements.
1 elements.
/>
= n. = 2 n.
2
Alternating group of n
2)!.
//
2!(n
Combination of the symmetric group of n

i>
elements with that of the two remaining elements.

4) r
= (n
2)
of
2 elements with the symmetric
Either the combination of the alternating group group of the two remaining
2 elements,
elements; or the symmetric group of n

f
,
In both cases
= n(n
* L.
1).
Etc.
d. Berl.
Kronecker: Monatsber.
Akad..
1889, j>.2ll.
133
For the non-primitive groups we have Cn

1)
=2
{-y-
Two
systems of non-primitivity containing each

is
elements.
The group
a combination of the symmetric groups
of both systems with the

selves.
two substitutions of the systems them-
p=
n\
For n
= 4, 6, 8,
we have
= 3,
10, 35, ...
3
.
(n -g-!j
is
Three systems of non-primitivity.
The group
a combination of the symmetric groups of the three systems with

!
the 3
substitutions of the systems themselves.

.
__
llT J
For
n=
6, 9, 12,
we have
.
= 15,
3!
280, 5770,
V -s-!j (n
we have
p
As
is
in 2), except that only the alternating
of the three systems
employed,
. .
group
.
For n
!
= 6, 9, 12,.
= 30, 560, 11540,

of
/>
8
.
ly
The values
increase, as
is
seen, with great rapidity.
In extension of the results of 117. 113 we proceed now to examine the primitive groups which contain substitutions of four, but none of two or of three elements.
Such a group
types
sl
must contain substitutions of one of the two
= (xax
b)
(xc xd ),
s2
2
of s 2 requires that of s 2 which belongs a c) b ), former type. Disregarding the particular order in which the elements are numbered, we may therefore assume that the substitu
to the
The presence
= (x x x x^. = (x x (x xd
a
b
c
tion
occurs in the group G. transform s 5 with respect to all the substitutions of G and obtain in this way a series of substitutions of the same type which
We
connect x x 2 x 3 a? 4 with all the remaining elements (Chapter IV, Theorem XIX). The group G therefore includes substitutions
1 ,
, ,
134
similar to
#1
>
s5
4,
which contain
besides some of the old elements

5
,
xi x z X
> >
other
new elements x
x6
a? 7
This can happen in three different ways, according as one, two, or three of the old elements are retained. Noting again that it is only the nature of the connection of the old elements with the new,
not the order of designation of the elements that we recognize that there are only five typical cases
is
:
of importance,
In the
first case, for
example,
it is
indifferent
whether we take
(a? 2 a? 5 );
(x&t)
(o? 3
a? 5
),
(XM)
(x&s), (xa x4 ) (x^), (x^x,}

l
, ,
and
is
replace x by x 2 x3 or a? 4 etc. cases are to be rejected, since their presence at once found to be inconsistent with the assumed character of
in the last
first
we may
fifth
The
and
the group.
Thus we have
the resulting substitutions in each case being inadmissible. There remain therefore only three cases to be examined, accord
ing as
contains, beside
s,
one or the other of the substitutions
A)
B)
C)
the
first
(XM)
(arj
( Xl
(x 2 x b ),
new elements
118.
case involving one each.
new
element, the last two cases two
A)
The
primitive group
contains the substitutions
and consequently
t
5
4
also
= s s = (x
a circular substitution of prime order
that
if
is 5, it
Since
83,
is
follows from
Then G must
w^.7, Corollary I, contain a substitution u, which does not affect
at
least three-fold transitive.
x but
l

replaces
x.2
135
by x& and x3 by
a? 7
If
we transform
s5
with respect to
this substitution,
we
obtain
s
= u~s u =
rj
XiX
r l
If
xa
s
f
is
contained
one element in
then
among x^x^x^x^, then s and s have only common and if x a is contained among x x9
s
, ,
.
and
s5
have only one element in common.
Both
alternatives
therefore lead to the rejected fifth case of the preceding Section.
If w _>7,
group.
becomes either the alternating or the symmetric There is in this case no group of the required kind.
For n 4 it is readily seen that there are two types of groups with substitutions of not less than four elements, both of which are
however non-primitive.
Groups of the type A) therefore occur only for n
=5
or
= Q.
For n
we have
first
the group of order 10,

|s 5 ,s*f
<r
#1=
If
we add
to
the substitution
= {s ,f. = (x x x x^)
5
l
we
obtain a second
group
of order 20
&*={**,
t,*\
= \t,*\.
G
l
The
latter
group
is
that given on p. 39.
and
6r 2
exhaust
all
the
types for n = 5. For n = 6 we
obtain a group
6r4
s ).
of the required type
by adding
II,
to 6r2 the substitution

transitive
(x^)
(x^x
Since
is
of order 10, the
group
must be
r
at least of order
60 (Theorem
Chapter IV).
And
again, since
\t,
<r,
(x^)
(x2
3 )ff
= r = (x^x^x^x^
we may
write
G =
t
We find then that = TV = ?<n*. rt <rV,

\.
Consequently from
37,
it
follows [that
6?4 is
of order 120.
The 60 substitutions
of 6r4
which belong
to the alternating
group
from another group of the required type
19.
B)
In this case
S5
contains
r
= (X.X,) (XzXt)
and consequently the combination
136
V
Ss-TSs
X
sufficient to
These three substitutions are not ments

#3,0-4
a;,
,
connect the six ele
a? 2
a?6
transitively, there
being no connection between

therefore
(x y x 8 )
(
and x lt x 2j x 6l x^.
The group must

the
83) contain
another substitution of
#i
type (x axp)
which
connects
x x *i
i->
ê with other elements.

t ,
If this substitution should con

, ,
tain three of the elements
would have three

the
x x a?5 a?6 and only one new one, it elements in common with v. This would lead
2
either to to the type
A) or
to the rejected first case of
117.
If
new
,
substitutions contained only

,
one of the
new elements
fifth
a?!, o?3
oj 5
Xf,
and three new ones, then we should have the

this is also to
case
of
117, and
be rejected.
two of the elements x

of one of the forms
There remains only the case where the new substitution connects x 2 a? 5 x with two others. It must then be
t ,
, ,
fi
(X&J
(x
2
(x 2 x b ),
(#!#) (x s x b ),
(x
2
(x^x a) fax,,), (x 5 xa ) (x6 xb ).
xa)
(x
x b ),
xa )
(x6 x b ),
Of these the
first,
T,
third,
fourth and sixth stand in the relation

first,
defined by C) to
while the
v.
second, fifth and sixth stand in
the same relation to

All the groups
B) therefore occur under either A) or
(7),
and
we may
pass at once to the last case.

C).
120.
In this case the required group contains
We
consider
first
the case
x^
= 6.
x
3
The elements x lt x 2l
are not yet connected with
,a?4 ,a?6 .
There must be a connecting substitution in the group of the type (XO.XP) (XyX$), where we may assume that x a is contained among the
the three elements
obtain,
x l9 x Z9 x 6
If
x a were x 2
ff
or
l
se 5
<r
then we should
,
by transformation with respect to

so that
or
a substitution
(âjft) (x,jx d )j
we may assume
(x&s) (x2 x m ),
(x,xm ) (x 6 x n ),
1.
The
possible cases are
then
(a)
(XM)
(x & x m ),
(x2 x5 ) fax,*)
m
m,
n,
3, 4, 6.
0?)
(r)
(x&n) (x n xp ),
(xjX m ) (x2 x n \
p = 3, 4, 6. m,n = 3, 4, 6.
137
The
<F
first
<r
and second
<r
lines are to
first
be rejected,
117, or
since their products with
I ,
lead to the
case in
directly to substitutions wtth only three elements.

for the different values of
There remain,
and
n, only the following cases:
(X.XQ) (xt
The second and

with
to
ff
l
since their substitutions have three elements in

ff
l ,
fourth lines and the third and sixth must be rejected, common, the former
.
the latter with T3

<r
The
first line
stands in the same relation

first
as the fifth to
We
need therefore consider only the

ff
l
line.
other.
The product of either of its substitutions by The required group therefore contains beside
gives the
<r
<r
l ,
<r 8
also
The group generated by
these four substitutions
is of
order 24.
J
It is non-primitive, the
being x l1 x 3
that there
tains
is
x2 x
,
and x 5 ,x6
It
systems of non-primitivity can also be readily shown
no primitive group
6r 5
of the required type which con
For n
=Q
there are only two groups of the required type.
These
are the groups

121.
(r
= 60)
and G,
(r
= 120)
of
118.
is
If the degree of the required
the indices m, n,
of the lines
), y9),
greater than 6, y} in the preceding Section
group
have a correspondingly larger range of values.

seen, however, that the three cases
(f)
)
It is
again readily
are inadmissible.
But
(/?)
both give rise to groups which satisfy the required conditions.

actual calculation shows that in every case a proper
The
combina
tion of the resulting substitutions gives a circular substitution of seven elements. 6) -fold Consequently the group G is at least (n
transitive
(
83).
138
If then n>.$,
is
at least three-fold transitive,
and therefore
contains a substitution which does not affect
l ,
but interchanges
and
o? 3
If
we transform
ff
*,
with respect to this substitution,

(oÔJg)
xz we
obtain
(x2 X a ),
and since
Section.
ff
has three elements in
common
with
<r
<r,
either
we have
the case A), or
xa
=x
and ^
is
equal to the
of the preceding
In the
latter case the
subgroup which
affects
x x
l ,
x6
is
itself at least
simply Combining with this group the cir cular substitution of seven elements we obtain a two -fold transitive
transitive*
group.
Consequently w>_9, at least 4 -fold

stitutions
(84) G
transitive.
is
at
least
(n
5)-fold,
and
for
contains therefore the sub
r
I
(aj,)
(x,xs ) (x&i
(o; 2 a? 5
),
.),
T~V T =
so that
(05,0-3)
we return
in every case to the type A).
For n >_Q there
is
therefore no group of the required type.
Theorem
S, the
III.
If the degree of a group,
which contains
substitutions of four, but none of three or of two elements, exceeds
group
is either
intransitive or
non -primitive.
113 and
116,
Combining
this result with those of
we have
Theorem
is
IV.
not greater than ^n(n
If the number /> of the values of a function = %n(n 1), 1), then if n > 8, either 1) />
and and
the function is symmetric in
2 elements on the one
in the two remaining elements on the other, or 2) /> 2n, the function is alternating in n 1 elements, or 3) p n,
the function is symmetric in
= =
2,
I elements, or 4)
/>
= 1 or
hand and and and the
function
is
symmetric or alternating in
all the
n elements.*
shall
122.
We
insert here a
lemma which we
need
in the
proof of a
more general theorem, f
From 83, Corollary II, a primitive group, which does not include the alternating group, cannot contain a circular substitution
"Cauchy: Journ. de 1 Ecole Polytech. X Cahier; Bertrand: lUd. XXX Cahier; Abel: Oeuvres completes I, pp. 13-21; J. A. Serret: Journ. del Ecole Polytech. XXXII Cahier; C.Jordan: Trait6 etc., pp. 67-75.
t
C.Jordan: Trait6
etc., p. 664.
Note C.

of a
139
prime degree
-~-
less
than
-^-
If
is
any prime number
less
than
and
if
is
the highest power of
p which
is
is
contained in
n\, then the order of a primitive group
not divisible by
pf
For otherwise
to the
group
would contain a subgroup which would be similar of degree n and order p f ( 39). But the latter
group by construction contains a circular substitution of degree p,
and the same must therefore be true of
G.
Consequently p
= nl
less
must contain the factor
at least once.
What has been proven

than
for
is
true of any prime
number
2n
-^-
and consequently
for their product.
We
have then
Theorem
two values
is
V.
If the group of a function with more than

is
primitive, the
multiple of the
number of values of the function product of all the prime numbers which are
less
than
2n
-^-.
123.
By
the aid of this result
we can prove
the following
Theorem VI. If k is any constant number, a function of n elements which is symmetric or alternating with respect to n k of them has fewer values than those functions ivhich have not this for small values of n exceptions occur, but if n exceeds property,
a certain limit dependent on
k,
the theorem is rigidly true.*
If 
1)
(n- 2)
l).
is neither symmetric nor alternating in A; or more be transitive with respect to n elements, may elements. But in the last case </ must not be symmetric or alterna
is
nk
a function which
it
ting in the transitively connected elements.
We proceed
*C. Jordan
:
to determine for both cases a

67.
minimum number of
Traits etc., p.
140
values of ^, and to
show that
if
is sufficiently
large, this
mini
mum
is
greater than the
maximum number
is
of values
A)
of ^.
124.
Suppose
at first that 4
transitive in less
than n
is
elements.
Then the order

A2
!
of the corresponding
group
(* a
a divisor
of
V
sible,
i.
A3 ! is
where ^
+ / + ^ H~
2
=n
of the
/ s
<n
k).
This product
e.,
maximum when one

k
1,
equal to n
i.
and a second
It
is
A is
is as large as pos then also as large
as possible,
k-\- 1
is
e.,
i.
equal to
fc-f-1.
>n
e.,
n> 2fc-fL
(n
The maximum order
further necessary that of the group
consequently
l)l
(fc
+ 1!,
4 is
2)
. . .
and the minimum number of values of

,
n\ (w
fc
_ n(n
l)(n
(n
k)
!)!(*? + 1!)~
1.2.3....(fc+l)
exceeds the
It
appears at once that the
minimum B)
maximum
A),
as soon as
This
is
therefore the limit above which, in the

of
first
case, the theo-
orem admits
125.
(%>_k),
no exception.
</
In the second case

it
is
transitive in
/.
elements
but
ments.
The
neither alternating nor symmetric in these ele group G of 4 is intransitive, and its substitutions
is
are therefore products each of two others, of which the one set
<T
I}
<r
while the other set r
connect transitively only the elements x ly x 3 x n _ K9 r2 connect only the remaining elements
,
.
.
l ,
The
forms
group
of 4 have, then, the product
where, however, one and the same <? It is easily seen that different r s.
may
all
occur in combination with

<r
the
occur the same

is
num
ber of times, so that the order of the group <r of the group 2 ,...]. [ff l
a multiple of that
S=
We
will
show that
is
neither alternating nor symmetric; oth-

erwise
141
which
it
be alternating or symmetric in n x elements, If the group 2 were alternating, This exceeds the maximum num would be of order %(n x)!.
is
G would
contrary to assumption.
ber
x\
of the order of the

as
group
T= [r,, r
9
of
elements,
ft ,
as soon
ff/B^jB,
n
l
>
ff
2k.
in
T
which
ft)~
ra
=
l
Consequently G contains substitutions <r a r and therefore substitutions Tp but ff*j**p
<?a r a (ffft
a ffp~
which
entire
affect only the elements #,
o? 2
.x n _ K
of the first set.
The
self-conjugate subgroup
H of
complex of these substitutions forms a

G.
This subgroup
is
unchanged by
,
transformation with respect to either G or 2, since ~ a r^, ... have is therefore a self no effect whatever on the substitutions of H.
H
2
and must accord conjugate subgroup of the alternating group is therefore a subgroup of G, ingly coincide with 2 ( 92).
,
H= S
and
(/
would, contrary to assumption, be alternating in n
ele
ments.
of the group G is therefore equal the maximum order of a non- alternating by x elements. We denote the latter order transitive group of n Then the minimum number of values of </ is x). by R(n
126.
The maximum order

*!
to the
product of
n\
_
still
(n
x)|
n(n ~
1)
(n
x -j1)
x\R(nx)R(nx)
We
have now
to determine
~~xT~
R(n
x),
the
maximum
f
order of
a non- alternating transitive
group of n
values of
n
;
elements, or
the
minimum number
x
of
a non- alternating transitive
function of n
elements.
is
If this function
non-primitive in the n
of values is
elements,
it
follows
that the
minimum number
Substituting this value in C)

values of
<,
we obtain
for the
minimum number
of
n(n
\)
(n
142
We compare this number with the maximum number A) and examine whether, above a certain limit for n, C\) becomes greater than A), i. e., whether
"
^./
j-y
\-v
For
sufficiently large
n we have
-j-
<n
A; -f-
1.
We have
therefore to prove that
.n
This
is
shown
at once, if
we
write the right
hand member
in the
form
For the first factor is constant as n increases, and the ratio left hand member to the second parenthesis has for its limit
of the
127.
itive,
Finally,
if
the function
of the
122.
<j>
elements
it
is
prim
we recur to the lemma of the minimum number of values of

numbers
less
From
this
follows that
is
than f (n
x).
We will
the product of all the prime denote this product by
Introducing
it
in C),
we have
l)
.
. .
(nx)-\ n(n
(n
We have then to
value A)
is less
show
that, for sufficiently large values of n, the

i.
than
2 ),
e.,
that
The right hand member if we replace every n

jc
of this inequality will be greatly increased

x
by the
x
first
factor
x.
There are
factors of the form
a.
These
will
be replaced by
THE NUMBEE OF VALUES OF INTEGRAL FUNCTIONS.

A
143
(n
x) ~~".
If
we
write then
v,
-,
we have only
to prove
that for sufficiently large
or
This can be shown inductively by actual calculation, or by the
employment of the theorem of Tchebichef, that 2. always a prime number between v and 2
For we have from
this
if
> 3,
there is
theorem
Now
whatever value the

t
first
quotient on the right
may
have,
we
can always take
so great that the left
hand member of
S>S
increases without limit,
is
if
only
is
taken greater than 2*
K
.
The
now complete. proof of the theorem The limits here obtained are obviously far too high. In every As we have, however, special case it is possible to dimmish them.
\n(n 1), it does not already treated the special cases as far as p of seem necessary, from the present point view^to carry these inves
tigations further.
CHAPTER
VII.

128.
We
recur
now
to the
results
obtained in
48,
and
deduce from these certain further important conclusions.* a Suppose that a group G is of order r=p m, where p is a prime number and m is prime to p. We have seen that G contains a sub
Let J be the greatest subgroup of G which of order p a group J contains H, and the order of J is there is commutative with H. a i is a divisor of m and is consequently prime to p. fore p i, where
.
Excepting the substitutions of If, J contains no substitution of For if such a substitution were present, its powers an order p
ft
.
would form a group

for
6rj
,
of order p&.
7,
But
if
in
A)
of
<r
48 we take
l
H K
l ,
the present groups
L, K, then since
H<r y
= H,
we should have
]jri
pf.lfi "~
d
l
"
p
The
left
~d
1
a
member of this equation is not divisible by p. Conse = _p, that is, L is con quently we must have in at least one case d y
tained in
<7
Hff y
= H.
of order p a which occurs in Again, every subgroup For if we replace 6r, obtained by transformation of H.
,
}
G
r
is
H K
,
of
A)
48 by G, H, M, we obtain
and
for the
same reason as before d y

~
<r
l
= pa
in at least one case,
and
therefore
is
M=
Hv y
Since H, as a self -conjugate subgroup of J,
transformed into
itself
by the
pa i
substitutions of J,
it
follows that
there are always exactly

into
substitutions of
which transform
any one of
its
conjugates.
.584-94.
There are therefore
of the latter.
*L. Sylow: Math. Ann. V.

Finally,
if
145
we
replace G,
r
JET,,
K,
of
A)
48 by G, J, H, we
have
p am __ p a i
~
Since
tains
is contained in J^ we must have d = p a and since J con no other substitutions of order pP, no other d can be equal to
l , 1
,
pa
It follows that
r
= p"i (kp + 1), m = i(kp -f 1)

-j-
The group
has therefore kp
1 conjugates
with respect to Gv
We have then the following Theorem I. If the order r of

but by no higher
the
a group
is divisible
power of
the
prime number p, and
if
is
by p one of
the
subgroups of order
contained in G, and
is
of order
p ai
largest
subgroup of
which
commutative with H, then the order
of
is
r=p a i(kp-\-l).
G is conjugate to H. Of conjugate groups there are kp -\- 1, and every one of them a can be obtained from by p i different transformations.
a Every subgroup of order p contained in
these
129.
sitive
In the discussion of isomorphism we have met with tran groups whose degree and order are equal. In the following
Sections
If
we
shall designate such

all
groups as the groups
LI.
we regard
therefore the orders r are
simply isomorphic transitive groups, for which all equal, as forming a class, then every
construction of
such class contains one and only one type of a group & ( 98). The all the groups LI of degree and order r therefore
all
furnishes representatives of
with the number of these classes.
the classes belonging to r, together The construction of these typi
cal groups is of especial importance, because isomorphic groups have the same factors of composition, and the latter play an impor
tant part in the algebraic solution of equations.
One type can be

type
is
established at once, in
its full
generality.
This
of this
formed by the powers of a circular substitution. A group Li type is called a cyclical group, and every function of n ele
to a cyclical
ments which belongs 10
group
is
called a cyclical function.
146
We
limit ourselves to the consideration of cyclical
groups of
th
prime degree p.
If s
(x
xp ), and
if
u>
is
any primitive p
root of unity, then
Is a cyclical function
belonging to the group

s a into
G = [1, s, s
2
,
.
s*"
For
(p
is
converted by
so that
c>
is
unchanged by the substitutions of
G.
x ta
Moreover, if for any substitution t, which converts every x a into we have 
^ ne
substitution
replaces
x lt x
...,
s,
by and
among + p, belongs to the group G. It is obvious that for group G furnishes the only possible type Q.
a7
1+j8 ,o
t is
therefore contained
the powers of
$P
r=p,
the
proceed next to determine all types of groups & of degree and order pg, where p and q are prime numbers, which for the present we will assume to be unequal, p being the greater.
130.
1)
We
It
is
already known to us. characterized by the occurrence of a substitution of order pq.
One
type, that of the cyclical group,
is
any other types, none of them can contain a The substitution of order pq, since this would lead at once to 1).
2)
If there are
only possible orders are therefore p, q, 1. A substitution s of order p is certainly present. This and its powers form within SI a sub
group
order
H of
p,
order p.
If
tt
contains any further subgroups

-/.p
of
their
number must be
4- 1,
where
> 0.
These sub
in
groups would have only the identical substitution They would therefore contain in all
(p
substitutions.
common.
l)(p*
+ l) + l=p[(p
l)*
+ l]>pg
x
This being impossible, we must have
= Q.
147
The subgroup
of order
q.
H contains
is
only
substitutions; the rest are all
Their number
pq
There are therefore
p=(ql)p.
g,
subgroups of order
and consequently from
Theorem
we must have
p-1
that be
is, q must be a any new type i2.
divisor of
1.
Only in
this case
can there
3)
The group
H
rt ,
is
quently every substitution

tion s of
a self-conjugate subgroup of Q. Conse t of order q must transform the substitu

also be equal to
1.
into s
l
where a might
.
We
write
= (Xj
Xp
Xi] ... (Xfxf
Xp *),
(where the upper indices are merely indices, not exponents). Then no cycle of t can contain two elements with the same upper index. For otherwise in some power of t one of these elements would
follow the other, and
of
if
this
power of
were multiplied by a proper

therefore take for one
s, power With a proper choice
one of the elements would be removed,

of notation,
we may
.
cycle of
(xfxfx*
It follows
xf).
then from
t- st
that
j
= s"
x 3 b by x 2a + b + \
l
.
.
t
b
replaces
l
,
.
x aa+ +
t
so
. .
xf by x a + that we have
Xi
)
.
b
l
1 ,
xa +
j>
l
by
= (Xi Xi
(X a
_|_ i
Xaa
_|_ i
X aa %
_j_ i
...
X aa g
_|_ i
...)...
If
now
the latter cycle
is
to close exactly with the element
x aa q-i +l q
we must have
aa 4 +l
=
.
-|-l,
a 9 EEEl
(mod. p).
The
solution
= 1 is to = (x^x* st = (x^xf
a
t
. .
be rejected, for in this case we should have

.
l xf) (x2 x*
l
x2 q)
l
.
.
(xp
x*
a;/),
x q q x q + x q+
l
...)...,
of
so that the latter substitution
would contain a cycle

s.
more than q
elements, without being a power of
148
It follows
THEuivtf
OF SUBSTITUTIONS.
then from the congruence a q (mod. p) that q is a as that a, belong we have further divisor of p 1, already shown; a> a a to has 1 values the _ n 2 q exponent g, q finally that ing
,
.
. .
~l
l ;
these values are congruent (mod. p} to the powers of any one l among them. From t~ st s" follows
all
t- st
so that,
if s is
= = s*\
t-*s?
= s\...
s a A,
. . .
transformed by t into any one of the powers s"? are also substitutions in LI which transform s into s
tt
i
there
s
a
v
Accordingly the particular choice of a A has no influence on the

tutions s
resulting group, so that if there is any type and #, there is only one.
generated by substi
The group formed by the powers

that formed by the powers of
stitutions gives rise to a
s,
of
being commutative with
the combination of these two sub
group exactly of order pq. group are the

t
The remaining
first
pq
g-j- 1 substitutions of the
powers
of the
1 substitutions
conjugate to
f
If
x?)
(?
= 2, 3,
.p).
typeLi.
p and
q are unequal,
tue
have therefore only one new

all
131.
Finally
.
we determine
types of groups
Li
of degree
2 and order p
1)
The
cyclical type, characterized
by the presence of a substi
tution of order
2)
2
,
is
already known.
If there are other types,
none of them can contain a sub

in every case If
,s
stitution of order
2
.
There are therefore

1.
s,
sub
stitutions of order
p and one of order
is is
any substitution
fully
of Q,
and
and
t.
by
any other, not a power of For all the products

x"t
l
then
Li
determined
(a,
= 0,1,2,
6
./>!)
are different,
and therefore
Li
= [V /*]
/
e
i
(a,
= 0, 1, 2,
.
...
1).
We
If
must have therefore

ts
= sA
fs
8s-
i%
.s
&
"-
1 .
now two
of the exponents J are equal,
it
follows from

t
149
)
=s
e
,
=s
l
t*
(ab,
(s
s
l
e=fe
er
that
a
b
(t
s)-\t
s)
= s~
t,
c t
t<)- (s
= P.
contains a
its
Since for
f we may
t
write
is
it
therefore appears that

s into
substitution
which
transformed by
one of
powers
t*.
The same result holds, if all the exponents 8 are different. For one of them is then equal to 1, since none of them can be 0, and
from
3)
t
= st
t
follows s
is
l
e
.
There
therefore always a substitution

l
= (x, x
xp ) (x?x?
l
x/)
(xfx/
ft.
x/)
in the pre
which
is
transformed by
s into a
power of
itself
As
ceding Section,
we may
take for one cycle of s
fa
xf).
Then from 3-^8

s
=
.
.
1?
follows
= faX
2
.
xf] (x?x a + ?x a *+*
x af -i +
*
l
...)...
If the second cycle is to close after exactly
elements,
we must
have a* -f IEEE 2,
This
is
a*=l
(modp).
possible only
s
if
= 1.
Accordingly
._.
= (xfx?
xf) (xjx?
./)-.. (x^x*
/).
The p -\- 1
substitutions
<? t O, t,
/
Ol/j
/ 06
2
,
.
<afp Ol>
are
all
different
first
and no one of them

1
&.
Their
is a power of any other one. powers together with the identical substitution
form the group
Summarizing the preceding
results
we have
Theorem
the degree
II.
and order are equal

The cyclical
type,
There are three types of groups Q, for which to the product of two prime num
2)
bers:
1)
one type of order
pq (p>q),
The first and third types are always pres the second 1. occurs ent; only when q is a divisor of p
3) one type of order
132. consider now another category of groups, character ized by the property that their substitutions leave no element, or
We
150
only one element, or all the elements unchanged. the groups we assume to be a prime number p.
The degree
of
Every substitution of such a group is regular, i. e., is composed of equal cycles. For otherwise in a proper power of the substitu
tion, different
from the
identity,
two or more of the elements would
be removed.
The
is
substitutions which affect all the elements are cyclical, for
prime number. From this it follows that the groups are tran sitive, and again, from Theorem IX, Chapter IV, that the number of substitutions which affect all the elements is p 1. We may
a
therefore assume that
and its first p 1 powers are the only substitutions of p elements which occur in the required group. The problem then reduces to the determination of those substi
tutions which affect exactly
p
s,
1 elements.
If
t is
any one of these,

all
then t~
st, being similar to ments, must be a power of s
and therefore affecting
the ele
where every index
is to
is
(mod p).
Since
,
it
be replaced by its least positive remainder merely a matter of notation which element is
that
not affected by
It follows that
we may assume
is
the unaffected element.
If
now g
is
a primitive root (mod. p), then

first
1
all
the remainders
(mod. p) of the
6r)
p
,
1
.
.
2
,
a*
powers of g p~2 p ~~ g g
. ,
=1
(mod. p)
are different,
and we may therefore put

ra
"
gr
(mod. p).
It
We will
denote the corresponding tbyt^.

A*
appears then that
t^
consists of
cycles of
- elements each.
For every cycle
of t^
closes as soon as
CEKTAIN SPECIAL CLASSES OF GKOUPS.
151
and
this first
happens when z
_
.
which leaves x unchanged and which replaces every x a + by x ag + then t^t^ replaces every If now we take x a + by x ag *p + ftv + and/? so that ap.-}-{3v is and v congruent (mod. p) to the smallest common divisor w of
If there
is
any further substitution

l
tv
1 ,
;j.
we have
in
t
= tflf
all
t v are powers. the substitutions which
t ff
a substitution of the group, of which both t^ and
Proceeding in this way,

leave
we can express
x unchanged
l
as
powers of a single one among them

t
where g*
is
the lowest power of g to which a substitution
of the
group corresponds.
The group
it
is
determined by
II,
and
tv
Since
is
of order
p -
_i
,
follows from
i
<T
Theorem
Chapter IV, that the group contains in

<t
all
substitutions.
1.
may be
taken arbitrarily among the
divisors of
p
To
133.
sidered,
obtain a function belonging to the group just con
we
start with the cyclical function belonging to s

fi
= (#1 +
>#2
+ " ^3
2
++
p ~ lx
p )*,
where &
is
any primitive
th
root of unity.
Applying
to
the
successive powers of f^,
we
obtain
The powers
quently
s all
group, leave
forming a self-conjugate subgroup of the given The powers of ^, and conse the ^ s unchanged. b a the substitutions s t,r of the group, merely permute the
of
s,
all
among
themselves.
Every symmetric function of
152
is
therefore
if
unchanged by every substitution

is
of the group.
Ac
cordingly
any arbitrary quantity, the function

(0
V = (tfft)
is
00
(0
^Wi)
<r
unchanged by
134.
all
these substitutions and by no others.
Wv
there
fore belongs to the given group.

If,
in particular,
1).
we
take *
t ff is
I.
the order of the group
becomes p(p
The
substitution
then of the form
containing only one cycle.

tive
The group
is
therefore two-fold transi
(Theorem XIII, Chapter IV). A group of this type is called a metacyclic group, and the corresponding function 9\ a metacyclic
function.
If
ff
= 2,
the order of the group
is
~^
and
t ff is
of the
form
The
indices in
the
first
cycle,
p)-,
each diminished
by
1,
are
the
quadratic remainders (mod.
is
any quadratic non-remainder
The group is in this case called half-metacyclic group (mod. p). and the corresponding function WT a half-metacyclic function.*
135.
the groups of in a two Sections, preceding prime in occur the which the changes simple way by expressing merely Thus the substitutions of indices of the elements x 19 x zy ...x n
all
We
can define
degree p, as well as those of the
& are defined by

sa
= \z z + a\
The symbol here introduced

z
in the substitution s a every element
its least
The
not negative remainder (mod. p) is to be taken. groups of the preceding Section contain then, in the
,
place, all the
and beside these, if we suppose substitutions s a diminished to be index by one, also those substitutions for every which every index is multiplied by the same factor, that is, for
= 0,l,2,...p--l
* L.
Kronecker;
cf.
F. Klein: Math. Ann.
_
substitutions of
.
(mod. p).
is
= 0, 1, 2,
.p
1),
to
be understood as indicating that
xg
is
replaced by a^ + a
and
for
first
XV,
258.
153
1).
ffp= z
fiz\
(mod. p)
(/9=1,2,3, ...jp
The symbol
t=z\g
ftz
+ a\
all
(mod.p)
= 0,1,
,
...p
1;
ft
= 1, 2,
1)
includes
the substitutions s a
00,
and
their combinations.
Since
|z
it
0*-|-a|.|2
t form a group of degree p and of This group therefore coincides with that of 134
follows that the substitutions
order
If
p(p !). we prescribe
that the
/5
s shall
take only the values
the products Pfa belong to the same series, and

of degree 136.
we
obtain the group
p and
of order
considered above.
The consideration
of the fractional linear substitutions
(mod. p) leads to groups of degree p-f-1 and of order (p-|-l)P These substitutions are of the form (p 1).
(mod. p),
where z
is
to take the values 0, 1, 2,

a?
,
.p
2
,
. . .
oo
the elements of
.
the group being accordingly

">
x ly x
xp _ lt X M
u,
The values
&>
î
determine a single substitution

results
but
it
may happen
a,
ft,
that one
and the same u
from different systems
^, 8.
To
D)
avoid, or at least to limit this possibility,
we make use
of the
determinant of u
ad
this is a quadratic
fr.
According as
remainder or non-remainder, we
-
divide numerator
ad.
and denominator of
coefficients
by \faS
/S^
or
by
For the new

ad
we have then
(mod. p).
coefficients
,
D
If
*i
i
PY^
now
Pi
,
for
î
two different systems of

#
/9,
YI
an(*
Y>
the relation
_
-y
J_
/?
(mod. p)
154
it would follow from the comparison and z\ 2, with the aid of D that if
, ,
were possible!
cients of z
real,
2
,
of the coeffi
,
/3, /3
are
=
If,
to /?, p, range of the values of 0,1,2, p 1, there are always two and only two different sys tems of coefficients which give the same substitution u.
therefore,
. .
.
we
restrict the
<5
With
tion only
D
if
z,
it is
assumed that ad
{3? is different
from
0.
This
restriction is necessary, for the
symbol u can represent a
substitu
different initial values of z give rise to different final

i.
values of
e.,
if
the congruence
az
,
is
(mod p)
impossible.
This
is
ensured by the assumption ad

are
We
E)
determine
now how many elements

*
unchanged by the
if
substitution u.
An
r2
index z can only remain unchanged by u
+ (d
)zp=Q
(mod. p).
There are accordingly four distinct cases: The two roots of E) are imaginary. This happens a)
2
is
if
} -r*
(ag-p r =
The corresponding xp _ X M
.
.
tutions affect
b)
a quadratic non-remainder (mod. p). all the elements X ,x 1 ^x 2
substi
l ,
The two
roots of
E) coincide.
This happens
if
The corresponding substitutions leave one element unchanged, The two roots of E) are real and distinct. This happens c)
if
is a quadratic remainder (mod. p). leave two elements unchanged.
The corresponding
identically.
substitutions
d)
The equation E) may vanish
This happens
if
r=0,
/5EEO,
(mod. p).
155
The corresponding
Finally
we observe
substitution leaves all the elements unchanged. also that

I
Q
I
M)
N)
(a0
,?
r)
(A
fa Yl )
= (aa,
If
We
proceed
now
and
to collect our results.
we take
not
=0
2 /? y arbitrarily, then for each of the (p l)p (mod. p), Since however resulting systems we obtain two solutions of D ). there are always two systems of coefficients which give the same
and
substitutions u,
tions.
we have in if we take Again,

{3
all,
in the present case,
substitu
^0
1, 2,
.
(mod. p) and
.
<5
arbitrarily,
then
)
restricting
to the values
,<?,/?
.p
p;
1,
we
obtain from
for
every system
two values of
u,
but as two systems of coef
ficients give the
same
we have
3
in this case
p(p
1) substitutions.
There are therefore in
all
linear substitutions (mod. p).
p From M)
/ ,.
P = (p-}-l)p(p
it
I
1)
fractional
appears that these

\\ Y)(
Y)
a group.
Among them
there are -
_-
form
I ^
substitutions
which correspond to the upper sign in D ). From M) and N) it is This latter group is called "the clear that these also form a group.
99 group of the modular equations for p .*
Both groups contain only substitutions which affect either p -f- 1, Those substitutions which 1 elements, or no element. orp, or p
x^ unchanged, for which accordingly rÔ? f rm 134 As the latter is two- fold transitive, the metacyclic group of it follows XIII, Chapter IV) that the group of order (Theorem
leave the element
(P
+ 1)
JP
(P
1) is three-fold transitive.
III. The fractional linear substitutions (mod. p) 1 and of order (p a !.) form 1) P (p group of degree p Those of which the determinants are quadratic remainders (mod. p)
Theorem
form a subgroup of order

lar equations for p.
)P\P~
a
_J
j
the group of the
modu-
If any substitution of these groups leaves more than two elements unchanged, it reduces to identity. The
first
*
of the two groups

Cf. J. Gierster:
is three
fold transitive.
p. 319.
Math. Ann. XVIII,
156
To
linear substitutions,
construct a function belonging to the group of the fractional we form first as in 133, a function *I ] of the
J
,
elements x
tutions
t
x lt x2
xp _
which belongs
to the
group of substi
= zfr +
fl
(mod.jp).
it,
(a
= 0,
l
1, 2,
...p
1;
,9
= 1,2,
...p
1)
The
substitutions
applied to
Iff 1
1 5
l\
produce
ifi 1
p -\-
1 values
IF l
p+1
which these substitutions merely permute among themselves. is any undetermined quantity, the function cordingly, if
Ac
z =(r
137.
,)
(*
2)
(*
^-H)
belongs to the given group.
We
have now
finally to
turn our attention to those
groups
all
the substitutions of which are commutative.
We
employ here a general method of treatment of very exten

ft ()
sive application.*
" are a series of elements of finite 6", Suppose that and of such that from any two of them a third a nature number,
,
.
.
one can be obtained by means of a certain definite process.

result of this process is indicated
If the
by
/,
there
is
to be, then, for
every two elements ", such that f(0 0"}
0",
which may
".
also coincide, a third element
We
will
suppose further that
but that,
if
0"
and
"
are different from each other, then
These assumptions having been made, the operation indicated by / possesses the associative and commutative property of ordinary
multiplication,
and we may accordingly replace the symbol /(#

0", if in
0")
by the product
sign oo
the idea of equivalence.

,
the place of complete equality we employ Indicating the latter relation by the usual
0" CND "
the equivalence
is,
then, defined by the equation
f(0 0")
9
()
".
1870, p.
for the
*L. Kronecker: Monatsber. d. Berl. Akad., most part verbatim from this article.
881.
The following
is
taken

Since the
is
157
denote by n,
number
to
of the elements
0,
which we
will
assumed
I)
be
finite,
these elements have the following properties:

there are always
Among
the various powers of an element

to unity.
some which are equivalent
powers are integral multiples of be said to belong.

II)
If
The exponents of all these one among them, to which may

()
any
belongs to an exponent
v
.
v,
then there are elements
belonging to every divisor of

III)
If the exponents
>
and
-,
to
which
and 0" respectively

(>
belong, are
prime
to each other, then the product
it"
belongs to the
exponent
/><T.
IV) If r&j is the least common multiple of all the exponents to which the n elements belong, then there are also elements which
belong to
n,
.
the greatest of all the exponents to which the various elements belong. Since, furthermore, n is a multiple of every one of these exponents, we have for every the equival
l
The exponent n
is
ence
fl">
CSD 1
may
of k
138. Given any element O belonging to the exponent n we extend the idea of equivalence, and regard any two elements
l
}
and 0" as "relatively equivalent" when
for
any integral value
We
retain the sign of equivalence to indicate the original
more lim
ited relation.
If now we select from the elements any complete system of elements which are not relatively equivalent to one another, this subordinate system satisfies all the conditions imposed on the entire
In particular there
that the
system and therefore possesses all the properties enumerated above. will be a number n.2 corresponding to ^ such
, ,
th
2
power
i.
of every
O
n
of the
k
.
new system
is
relatively equiv
o lt in the
tb
alent to unity,
e.,
*coHi
Again there are elements

n.2 is
new system
of
which no power lower than the

ft"-
relatively
Since the equivalence and element, consequently a fortiori every

equivalent to unity.
0" CND 1
is
holds for every
relatively equiva-
158
lent to unity,
it
follows from I) than
is
equal to n 2 or
is
a multiple
of ri-
now
and
k
if
both sides are raised to the power

the equivalence
n2
we
obtain, writing
n2
= m,
From
this
it
follows that, since O l belongs to the exponent

.
n l9
is
an integer and k is therefore a multiple of n 2 an element 2 denned by the equivalence

,
There ~m
is
therefore
CNO 0,,
is
or
oo O lf 0^1
of which the
?i 2
th
power
absolutely equivalent to unity.

tively
not only relatively equivalent, but also This element belongs both rela
and
absolutely, to the exponent

Of* CNO O n
n
*
n2
n
>-
for
CNO
we have
the relation
Off*
if
~ m "2
CNO
OfOl
~ m "2
"2
CNO 1.
Proceeding further,
0" as relatively equivalent
we now regard any two elements when
and
we
obtain, corresponding to
HZ,
an element
belonging to the expo
nent
where n 3
is
equal to
n.2
or a divisor of
n2
and so
on.
We
obtain therefore in this

\ y
way
a fundamental system of
elements
e i i #3
which has the property that the expressions

0**0,**0**...
(h,.= 1,2,.. .n)
include in the sense of equivalence every element

once.
once and only

19
2
,
The number
n.2
n3
to
which the elements

is
belong, are such that every one of

of
them
equal to or
. . .
is
a multiple
the next following. The product n^ n2 ns is equal to the number n of the elements 0, and this number n accordingly contains no other prime factors than those which occur in the first
entire
number
w,
139.
In the present case the elements
are to be replaced by
substitutions every two of which are commutative.

of the elements
The number n
becomes the order r of the group.
We
have
then
159
Theorem
mutative, there
IV.
is
// all the substitutions of a group are com a fundamental system of substitutions Sj s 2 s 3

,
, ,
.
which possesses the property that the products
SiXV
,
,
.
.
...
(h t
= 1, 2,
rO
include every substitution of the group once numbers r l5 r2 r3 are Me orders of s 1? s 2

.
and only once. s3 and are

,
.
The
SMC/Z,
/ia
ever?/
owe
is
equal to or
is divisible
,
by the next following.

egwaJ
to the order- r
The
product
group.
of these orders r n r2
r3
is
of the
The number r
is
determined as the
maximum of
the orders of the
the other hand the corresponding substi tution s is not fully determined, but may be replaced by any other substitution s/ of order ^ then as we start s or from According
several substitutions.
l
.
On
s/,
the values of
r2 ,r3 ,...
might be
if
different.
We
shall
now
show that
In the
to the
selves,
this is not the case.

first
place
it
is
r,
plain that
s
several successive s
same exponent
l
these s
may
without any change in the r s. T without any change in the r replaced by s/-s a+] s a+2 vided only that /A is prime to r a
.
. .
s belong be permuted among them Moreover, every s a can be

s,
pro
If
forms
8l
then
W
/>!
ff
l
now
the given group can be expressed in the two different
... (*
,
= 1, 2, ...
fT f
r,),
= rj
,,*
VG T MQ r 83 SA
(ft.-
= 1, 2,
/>,),
and
>\
Since
belongs to
v
.
t\
at least
one of the exponents

it
//.,
v,
must
be prime to r
this is
/j.,
From
the
first
remark above, we may assume that

follows that the group can also be
and from the second
expressed by
<0s 2 *as3
...
(hi
1, 2,
?-,-;
f)
r,).
Consequently the groups
s/Vs
so on.
(h
= 1, 2,
r,),
<7>3
3
.
(h t
= 1, 2,
/><)
are identical.
From
V.
this
it
follows, as before, that
(> 2
=r
and
Theorem
a given group.*
The numbers r n r2 r3
,
are invariant for
*This theorem is due to Frobenius and Stickelberger. <rruppeu mit vertauschbaren Elementen; Crelle, 86, pp. 217-262.
cf.
their article:
Uber
CHAPTER
VIII.
ANALYTICAL REPRESENTATION OF SUBSTITUTIONS. THE LINEAR GROUP.

In the preceding Chapter we have met with a fourth method of indicating substitutions, which consisted in assigning the
140.
analytic formula
by which the
final
determined from
is
its initial value.
Thus,
value of the index of every x is if the index z of every xz
converted by a given substitution into <p(z\ so that x z becomes the substitution is completely defined by the symbol Xp,)
,
<?(z)\.
<p(z),
Obviously not every function can be taken for

essential condition that the system of indices
shall all
for
it
. .
is
.
an
^(1), ^(2),
y>(n)
be different and shall be identical, apart from their order,

1, 2, 3,
with the system
... n.
On
the other
hand
it
is
readily
shown that every substitution can be expressed For if it is required that

?(!)=*!, ?( 2 )
in this notation.
*2,
...?(")
s
we can
from
construct,
by the aid of Lagrange
interpolation formula,
a function y(z) which satisfies the conditions, viz:
~-
_
ll
Fz
(
i_
Fz
hF
(
|-
~F
(z)(ziy
is
(z)(z2y
1 in
ln
F (z)(z-n)
also satisfy the
This function
of degree n
It is evident that there are

<p(z)
an
infinite
number
of other functions
which
required conditions.
141.
If
is
a prime
number
p,
we can on
the one hand
diminish the restrictions imposed on <p by permitting the indices .p to be replaced by any complete system of remainders 1, 2, 3,
.
(mod.
p), so that indices
greater than
are also allowable.
And on

the other
161
hand we can depress every form
of v(z} to the degree

z.
1,
since zp
^z
\
(mod. p) for
all
values of
this
In
particular,
i
we have
1.
in
case
F (z)-~-pz p For n = p,
F(z)=z p
and
the functions
<p(z)
which are adapted
for the expres

. .
sion of a substitution, for which therefore ^(0), ^(1), <f(p 1) form a complete system of remainders (mod. p), are denned by the
.
following theorem:
In order that \z <pz\ may express a substitu I. p elements, it is necessary and sufficient that after <f(z) and 2 powers have been depressed to the degree p its first p 1 by
tion of
Theorem
means of
the congruence z v
^z
1
(mod. p),
and
after all multiples of
have been removed, these p
powers of
<p(z)
should all reduce to
the degree
2.
Let
be any integral function (mod.

[?fe] EEE
"
p),
>}
and suppose that

-+.
A ;>
(
>
>
-}-
A,""
+ Ar
1
.
2
. .
-f A,,
j<
""?
>
(mod. p).
Since for every
<p
a
+l +2 +
we have
ZE
If
-HP
l)
~0
(mod.
p\
"
A^.
_!-
(mod. p).
now
<F
(f(.z) is
. . .
^(0),
(1),
<p(p
adapted for the expression of a substitution, then, as 1) form a complete system of remainders
that for
A,,
(mod.
p),
we conclude
m (z),
given function
it
m = 1, 2,
l>
2,
follows from
formula B) of
8 that
Hi)]
HP
/)
i)]-^
,5
az
p).
Ê(!
Hermite: Coinptes rendusde
1
(mod.
57.
Academic des Sciences,
11
162
Accordingly,
is
if
satisfied
=0, the linear congruence by thep integers ^(0), ?>(1),

. .
(1
<p(p
a)z
1).
/3^(mod. p) These must

is
therefore
all
equal to
0.
every one of them But in this case the congruence of degree p 2

is 0,
2
l
be equal, and, as their
sum
^(z)
= A + A z-\-A z +
2
...
+Ap _,z
/?
-*=Q
(mod. p)
1,
would have
impossible.
p 1 different roots z = 0, 1, 2, ...p Consquently = 1 and then = 0, that is
which
is
=z(z
and the values
^(0), ^(1),
<p(p
1)
(2-2)
[z
0, 1,
.
(p
.
1)],
1,
1) coincide with
.p
It appears therefore that the condition apart from their order. stated in Theorem I is both necessary and sufficient to insure that
z
|
<pz
defines a substitution.
|
142.
To
we may
also
distinguish the individual elements of any system employ several indices in each case instead of a single
index as heretofore.
indices z
For example,
,
in the case of
elements the
to
and u of xz
might each assume any value from
1.
Any
substitution
among
these
elements could then be
denoted by
8=\Z,U
where
140.
If
y>(z t
<P(Z,U\<! (Z,U)
satisfy conditions similar to those of
u)
k
and
</
(#,
u)
must
the elements could be denoted by
a^.o^,.
.x gk
(z,-
= 0,1,2,
...p
1),
and any
substitution s
?i(z,
by the symbol
,
8=\Z
which
,Z 2 ,...Z k
22
<
Zk), <P*(Z\
>Z*9-~
Z*)>
k(Zi
,Z 2 ,...Z k)\,
z, is
:
is to be understood as indicating that every index
to
be
converted into
k
^( n
>
z k ).
The
^2
functions c^,
$r 2
<p k
must
.
.
then be so taken that the

give rise to
different systems of indices z

<f l
, j
z^
zk
different systems
?*
These considerations could be further extended to include the case
where n contains several distinct prime factors, but as the theory then becomes much more complicated, we do not enter upon it here.
143.
=m
The simplest analytical expressions for substitutions of elements are those of the linear form
163
The
a s are arbitrary integers (mod. m).
They can
therefore be
of
selected in
this type.
2)
different ways, so that there are
mk substitutions
...
Again, since
S ai
,
a ,,,
... a/,
Sft
jB 2
... fo-
#04
+ ft
ou
(3,
a ft
+p
fc ,
these arithmetic substitutions * form a group of order and degree This group is transitive, since the a s can be so chosen that m*.
any given
element xzi
-^ or
x& &
&
is replaced ~, by any other element îs P ur pose we need only take

t ,
.
.
gje
a i ==
-si
>n
(l
->.
= *2
^2?
ak
*A--
Â"
There
result.
is
only one substitution of the group which produces this
In order that an arithmetic substitution
may
leave any element
x unchanged,
it is
necessary that
2
r/.j
= 0,
= 0,
all
ak
(mod. m)
Such a substitution leaves

fore reduces to identity.
in the
The Q considered in the preceding Chapter. groups By the continued application of the formula 2) we obtain
c? t?
,
the elements unchanged, and there present group is therefore included
âj
0-2
a;,
"1
...
aJ
i
c "0
a
?
.>
-...^0,0,...!
a z-
so that
O\ &)
we may
define the group
by
o,
i
,
.
_ ur
/nr
l<i,o,...o,
c 6
..
o?
o 6 o,o,
..
The
tive,
substitutions contained in the parenthesis are all
commuta
and the same property consequently holds
for all the substitu
tions of G.
144.
We
,...Z
lc
determine
now
the most general form of the substi
tutions
t=
,Z<2
^fo, 3 2 ,...
fc ),
?,(*?,, Sj,... 3*),... ft(*i
*.*)
|i
which are commutative with

*
(r,
for
which therefore
^-i.... *
= *&,..,
for
s ai
,...
It is obviously sufficient to
take
3).
the several genera
ting substitutions
Cauchy: Exercices
given
III., p. 232.
in
The
substitution
^X^...^
164
replaces
FA(ZI, 2 2
>
. .
.
*)
by
? A (2,
+ I,2
.
2,..).
Consequently,
taking
1, 2, 3,
k
.-
we have
*)
Fxfci
+ 1, a,
$p A (2i
2V,
Zk) -h
,st)
.
-A
(mod. m).
.
Similarly in the case of the substitutions
____
we
obtain
^ A(ZI 3a
,
+
,
.
1,
2*.)
EEE tf^z-i
,,
...
Zj.) -j-
5A
(mod. m)
^(Zj
z,
ZA
+ 1) ^^ Â(Î
2,.
Â-)
+C
(mod.
the
-HI).
From
linear
dx
these congruences it appears at once functions of the z s, having for their

0, 0, ... 0).
fact,
that
cr A
are
constant
terms
V A (0,
found.
In
The remaining we have

,
.
coefficients are
then readily
cr A
(^
z,
;,)
a K z,
+ Ma +".+ Â^- +
+
>A
and therefore
î
Conversely
itself.
b.2 z.,
+ c,z + X
k
.
j
all
substitutions of this type transform the

t
group
into
Thus, for example,
transforms slf0
\
>
,...
into
afa -f
M? +
~h
c\
%+
^(^ + 1) + M. +
i.
+
I
CJJB*
+ *....
a.,
.
.
e.,
into the arithmetic substitution

1
2)
*1
^l
~-
a
I
-2
-/
"~k
I
=: 5 a]
a/|
By
left
hand multiplication by
SSi
.
6,
S/,
8^.
S.,,...
- &k
we can reduce
to the
form
Such a substitution
is
called a geometric substitution. *
We
strated
proceed to examine this type.
We
have already demon
All geometric substitutions and their combi nations with the arithmetic substitutions, and no others, are com II.
Theorem
mutative
ivith the
loc. cit.
group of the arithmetic substitutions.
*Cauchy:

145. stants
165
We
&A,
.
.
have
-
first
of all to determine
arbitrarily.
whether the con

certainly
aA
CA
can be taken
They must
,.,
be subjected to one condition, since two elements x^ ^ and x &, &. .-.(* must not be converted into the same element unless the indices z lt z 2 z k coincide in order with C2 More -/,.
,
.
.
.
generally, given any system of indices r 1?
:,,
C*,
it
is
necessary
that from
aft
-f-
b,z,
. . .
+ c^. = d a^ -f 6 +
,
2 2,
. . .
+cz
2
A.EEE
C2 ,
. . .
(mod. w)
the indices
2j
,%,...
in
,,
In other words, the

equal number
conditions for this
...
be determined without ambiguity. systems of values z must give rise to an

shall
C".
of systems of values
is
The necessary and
sufficient
that the congruences
A + MJ + + c,z,.=0,
are equivalent to
o^t -f 62*2
l
. . .
+ 0^=0,
0,
.
.
(mod. m)
0.
shall admit only the one solution z 2 0, determinant of the coefficients is denoted by
zk
If the
J, these congruences
=Q,
s2
is
~0,
.2f fc
=0
(mod. m).
if
The required condition

prime
therefore satisfied
and only
if
is
We have Theorem III. t = \z ,Z ;... z k a^

to m.
1
then
In order that the symbol

b^z,
-f. . .
-f-
+ c^t,
a 2 Zi -f
b.2 z 2
-f
+ c^/,,
.
|
may
that
denote a (geometric) substitution,
it is
(mod m) necessary and sufficient
an &n
J
cl
c2
a2
&2 ,
a*, 6*,
Cfc
should be prime to the modulus m.

146.
From
the
number
this consideration it is now possible to determine r of the geometrical substitutions corresponding to a
given modulus m.
We
that any
denote the number of distinct systems of
/>
integers which
are less than
and prime
of the
;>
to
by [m, />].
It is to
be understood
number
integers of a system
may
coincide.
166
Suppose
1, t2
,
to
which leave the
be the number of those geometric substitutions first index z unchanged. If then r 2 is

l l
any
T2
substitution which replaces z

T2
T , fc} 2
-
by
a^
-f-
b^z*
-f-
+ 0,2*,
then
^ and these are

z
1
are
all
all
different
. . .
by a/^
on.
+ fc/Za-j-
the substitutions which produce this effect, from one another. Similarly, if r 3 replaces are all the C/2*, then r 3 J 2 r 3 f3 T 8
substitutions which produce this effect,
and these are
all different,
and so
tutions
We obtain
,
therefore the
number
r of all the possible

of substi
geometric substitutions by multiplying

1, To
N
.
. .
by the number
T3 ,
The choice of the systems a n 6 n ... c/; ... is Cjj a/, ft, limited by the condition that that the integers of a system cannot have a same common factor with ra. There are therefore [m, k] such systems, and an equal number of substitutions 1, T 2 ,r 8 ...
, ,
Consequently
The
\z 19 z29
substitutions
1
are of the form

2
...z k z
,a,iZ 1
+b Zi +
---- c 2 z &
a^ +
M
.
-f
..
(mod. m).
Since a 2 a 3
, ,
.
.
nant
J, these integers
ak do not occur in the expression of the discrimi k ~ can be

chosen arbitrarily,
*
e.,
in
dif
ferent ways.
The
6A
C A are subject to the condition, that
must be prime
r
,
to m.
If the
number
of systems here admissible

k
is
we have
r
= [m, k] m
The number
k
has the same significance for a substitution of 1 indices (mod. m) as r for k indices. Consequently
r r
= [m, k] m^-
L m,
A?
1]
m*~ V,
and so
on.
We obtain
r>
therefore finally
where r^
[m, 1]
.
corresponds to a single index, and therefore
Hence
ANALYTICAL REPKESENTATION OF SUBSTITUTIONS.

4)
167
1]
.
= [m, Afjm*-
[m,
k~ 1] m*~
m
2
.
. .
[m,
2]m[m,
evaluation of [m, fc] presents little difficulty. limit our selves to the simple case where is a prime number p, this being the only case which we shall hereafter have occasion to employ.
The
We
We
have then evidently

5)
[P, /"]=*>"
0,
,
1,
since only the combination
...
is
to
be excluded.
By
the
aid of
6)
5),
we
obtain from 4)
1
= (^-l)^- ^- -!)^= (p* 1) (p*p) (p* p

2
2
.
-l)p(p-l) (p*-^- ).*

.
(p
The entire system of the geometric substitutions a group the order of which is determined from 4) forms (mod. m) or from 6). This group is known as the linear group (mod. m).
147.
If the degree
is
to be particularly noticed,
we speak
of the linear
group of degree m*. It is however evident that

leave the element
the substitutions of this group unchanged. For the congruences

all
(mod. m)
have for every possible system of

ZjÊO,
32
coefficients the solution

.
^0,
3fc^0
(mod. m).
We
shall
have occasion to employ the linear group in connection
with the algebraic solution of equations.
Theorem
4).
IV.
The group
of
the geometric
substitutions
(mod. m), or the linear group of degree
m?
is
of the order given in

,
Its substitutions all leave the element
XQ
...
unchanged.
It
is
commutative with the group of the arithmetic substitutions.

*Galois: Liouville Journal
(l)
XI,
1846, p. 410.
PART
II.
APPLICATION OF THE THEOKY OF SUBSTITUTIONS TO THE ALGEBRAIC EQUATIONS.
CHAPTER
IX.
THE EQUATIONS OF THE SECOND, THIRD AND FOURTH DEGREES. GROUP OF AN EQUATION. RESOLVENTS.
148.
The problem
of the algebraic solution of the equation
of the second degree

1)
c,.r
+ =
c.j
can be stated in the following terms:

metric functions c :
to determine the
From
the elementary
of 1)
it
sym
and
c.2
of the roots x\ and

1
x.2
is
required
is
two -valued function x by the extraction of roots.*

to us (Chapter
I,
Now
it is
already known
13) that there
always
a two- valued function, the square of which, viz., the discriminant, In the present case we have is single-valued.
= (x = (x
2
2)
= (x + x. Y
l
4a?!X.2
.
=c
2
x
4c 2
x2 )
\/c
2
}
4c 2
Since there
is
only one family of two-valued functions, every such
function can be
in terms of rationally expressed linear two- valued functions we have

"l a
V^-
For the
-f
,
2
and
in particular, for
= 1, = 0,
a
and
for
= 0,
*C. G. J. Jacobi: Observatiunculae ad theoriam aequationum pertinentes. Werke, Vol. Ill; p. 269. Also J. L. Lagrange: Reflexions sur la resolution alg^brique des Equa tions. Oeuvres. t. Ill, p. 205.
ELEMENTARY CASES
149.
GROUP OF AN EQUATION
RESOLVENTS.
169
In the case of the equation of the third degree

x*
c^-^-c^x
c3
the solution requires not merely the determination of the threebut that of the three three-valued functions valued function x
l
,
a?!,
x x
2
,
With these the 3 .-valued function

^
=ax
l
-{-
-+-
a 3 jc
}l
is
also
in terms of c.
known, and conversely x x 2 x can be rationally expressed We have therefore to find a means of passing from
l , ,
:i
the one-valued functions

extraction of roots.
Cj
c 2 c 3 to a six-valued function
,
by the
In the
J =
first
place the square root of the discriminant

l
(a?!
x^Y (x
x$Y (x z
x3 ) 2
=
4c 2
s
27c 3 2
-f-
IScgCaCj
4c 3 Cj
Co
cf
furnishes the two-valued function
(x
in terms of
sible.
x2 )
(x.
x,) (x3
xB],
is
which
all
the two-valued functions are rationally expres
Tfre question then
becomes whether there
any multiple-
This question has valued function of which a power is two -valued. The six-valued func 59. already been answered in Chapter III,
tion
on being raised
to the third power, gives
?*
=i 2^90^ + 270, + 3 V
^
by changing the sign of
again,
if
^2
is
obtained from
V-
we have
?2
Accordingly
3
3
=(x
+ "x
4-"X>
= HSi
3V
3J).
3J
i (-S,
Combining with these the equation
170
and observing that

1
-Jco
-\- o)
0,
we
obtain the following results
The
solution of the equation of the third degree

150.
is
then complete.
In the case of the equation of the fourth degree
X4
it is
CjO! -}-
C 2 ,T
C 3 .X
-|
C4
,
=
,
again only the one-valued functions Cj c 2 c 3 c 4 that are known. From these we have to obtain the four four-valued functions
,
Xxxx.
and with them the 24-valued function
= a lXl -by the repeated extraction of roots. In the first place the square root of a rational integral function
of
Cj
,
c2
c3
c 4 furnishes
the two-valued function ^/ j
Again,
we
have met in
<p.=
59 with a six-valued function
(x^
-f
x 3 x4 )
-+-
^(aw -f ^ î) + ^
2
(a?i^ 4
+ x. Xs)
2
the third power of which
is
family of ^/ j
We
can therefore obtain

9?
two-valued and therefore belongs to the 
being
also
determined, every
is
known.
The group
r
is
G
To
this
[1,
(X&)
(x..x,\ (xjXg) (x 2 o? 4 ), (x,x,} (x 2 x 3 )~]
/>
= 6,
= 4.
same group belongs the function
which can therefore be rationally obtained from ^, while ^, which 2 can be obtained from by extraction of a square root, belongs to
</
the group
H=
[1,
(x,x z )
(x,x,)-]
= 12,
= 2.
^ may therefore be regarded as known:
Finally
ELEMENTAEY CASES
GKOUP OF AN EQUATION
</ ,
^RESOLVENTS.
*/
171
can be rationally expressed in terms of

function.
and
is
a 24-valued
All rational functions of the roots,
and
in particular the
roots themselves can then be rationally expressed in terms of 7.
To determine the
roots
~
we may,
i
for example,
combine the four

,
which and -j- fa 2 equations for In attempting the algebraic solution of the general 151. equations of the fifth degree by the same method, we should not be able to proceed further than the construction of the two- valued
in
/ and
fa =
functions.
For we have seen
in
58 that for more than four
independent quantities there is no multiple- valued function of which a power is two-valued. It is still a question, however, whether
the solution of the equation fails merely through a defect in the method or whether the impossibility of an algebraic solution resides
in the nature of the problem.
latter is the case.
It will hereafter
be shown that the
shal^ demonstrate the full sufficiency of the method by the proof of the theorem that every irrational function of the coefficients which occurs in the algebraic solution of an equation
All the steps leading from the to the one-valued function given required nl-valued functions can therefore be taken within the theory of the integral rational func
is
We
a rational function of the roots.
tions of the roots.
152.
We
turn our attention next to the accurate formulation
of the
in the solution of algebraic equations. tb all the that roots of an equation of the n degree Suppose
problem involved
1)
f(x)
If one of
=
1
are to be determined.
them x
is
only partially solved.
By
n
the aid of the partial solution

it
known, the problem is we can
however reduce the problem and regard

the determination of the the ^ degree f(x) reduced equation
2)
th
now
as requiring not
remaining roots of the equation of

1
0,
but that of all the n
roots of the
J^L=
1
fl ( X )
still
= ^^
n
ri
^-^ r^^-still
?V-1
= 0.
We
have then
to accomplish the solution of
2).
If one of its
roots
is
known, we can reduce the problem
further to that of
the determination of the
2 roots of the equation
172
l
THKOKY OF SUBSTITUTIONS.
-(
i
of
"
Proceeding n this way we arrive

degree.
It
finally at
an equation of the
first
th appears thoroforo that solution of an equation of the n
degree involves a series of problems. All of these problems are however included in a single one, that of the determination of a
xiiiyle root
of a certain equation of degree n\

the entirely independent roots
Thus,
equation constants
4)
is
if
l ,
x,
...
;/
of the
])
,,
are known, then the n!-valued function with arbitrary

._,,
.
u.
ti
,x,
4if
-(-...-(-
also
known.
is
Conversely
c is
known, every root of the equa

, ,
tion 1)
.r 2 known, can be rationally expressed in terms of 9 the n! valued function
for every rational integral function of
.r n
=.
The function
5)
$ satisfies
1
an equation of degree n\
fiI
Fft^-A^
. . .
.
+ ...A
--(,--*)G--^)...(^-^
= 0,
the coefficients of which are rational integral functions of those of This equation is, in distinction from 1), 2 ,, 1) and of
.
a very special one. For its roots are no longer independent, as was the case with .1), but every one of them is a rational function of
every other, since with respect to x,

that of 1)
is
all
,
the values
.
.
.
,
belong to the group 1

5),
./-,,
./.
The
solution of
and consequently
therefore complete, as soon as a single root of the
former
is
known.
The equation F(Z)

th<
(Jaloiti
=
of
()
is
called the resolvent equation,
and
insolvent
in a
1).
We
, ,
shall
presently
introduce the
name "resolvent"
l< )o.
more extended
sense.
c 2 ... c,, of the equation 1) are c, the Galois resolvent cannot break up into entirely independent, rational factors. if the Galois resolvent does not break Conversely,
If
the coefficients
up
into factors, then, although relations

*;,
may
exist
among
the coef
ficients
cation in the form of the solution.
they are not of such a nature as to produce any simplifi From this point of view the
general equation,
or,
equation
ecker,
it
1) is in this case a
according to Kron
has no
affect.
ELEMENTARY CASES
RESOLVENTS.
7H
On the other hand, if for particular values of the coefficients c the Galois resolvent F(;) breaks up into irreducible factors with
rational coefficients
then the unsymmetric functions A
T
(
(r),
which
in
the case of fully

rationally
independent
coefficients are irrational, are

1) is
now
known.
The equation
manner
Cu e 2
in
then a special equation, or according to Kronecker, an affect equation. The affect of an equation lies, then, in the
which the Galois resolvent breaks up

particular relations
./,.
.r_,,
.
or, again, in
the
fact that, as a result of

,
.
among
the coefficients
c.,
or the roots
)
,r /0
certain unsymmetric irreduci
ble functions F;( =
are rationally known.
The determination
is
of that
which
is
to
be regarded as rationally known in any caso

importance.
obviously
in this
of the greatest
As
a result of any
change
respect, an equation may gain or lose an affect. If the group belonging to any one of the functions
:
; (
is
</,-,
then every function belonging to G, is rationally known, being a rational function of F It is readily seen that the groups (j. all coin
;
.
cide.
For if the subgroup common
to
.
them
all is /
then the ration

/
,
ally
known function
n.l<\ -+-, i/<\
4-
/ 1<\ -j-
belongs to
and conse
quently every function belonging to / is rationally known. Accord ingly the factor of F, which proceeds from the application of the substitutions of / alone to any linear factor r "" /^.r/.,. -.
/,./-,.,
;
of
is
itself
rationally known.
/< ,,
This
/
is
inconsistent
G.2
with the
v
.
assumed
irreducibility of
/*
,
unless
,
. . .
= (}
G
All the functions

family,
F.2
h\ therefore belong to the same

or by any
to G.
and
</>
this
is
.
characterized by a certain group

.
.
function
(./,, ./._,, .r,,) belonging ing to (j is rationally known and conversely every rationally function belongs to G.
Every function belong
known
Theorem
izcd
Inj
I.
Every special or
IHJ
affect equation
/.s
r/
a group G, or
a xinyle relation between the
The
equation
.
(jrniifi (}
is
/.s called the Galoix f/roup of the equation. accordingly completely defined by the system
^ XA
=c
^>A.^V
=C
(x
x.
OC H )
= 0. *
*Cf. Kronecker; Grundzuge elner arithnietischen Theorie

10, 11.
<icr algebraist-lien (ir"s-
h rrri/
srii, ^
174
For example, given a quadratic equation
x2
x -f-
c2
= 0.
2
the corresponding Galois resolvent
is
2(a,
+ a )c^ + (!
z
2)
C2
+ 4a a c =
2
1
0.
In general the latter equation is irreducible, and the quadratic equa tion has no affect. But, if we take
2d
the equation in :
(c
= m-\- n,
2
c2
m n,
a l xl
affect.
becomes
a^n
a^m
fjL
n) (c
m) =
= (c
a 2 x2 ) ( - -
a^
î)
and the given quadratic equation has an 2 = 0, we have Cj Again, if c 2

(c
ttjd
2 2
c2)
= 0= (5
obtain
1
a^
2 a3 2
)(^
i^2
2î)-
But
if c 2
2C/
= 0,
C
2
we
-2(
+a
2)
Cl ,-
+ 2 (rV + a
i
2 2 ) Cl
- 0,
known, the equa becomes
(aj
and
this equation
has no
affect, so
long as we deal only with real
quantities.
If
however we regard
=
(a,
\/
1 as
tion has an
(
affect, for the Galois resolvent then

2)
(!
Cj
+ (,
2)
C,i) (I
2)
a 2 ) cj)
= 0.
It is clear that every unsymmetric equation y(x ly ... x n ) 154. between the roots produces an affect. On the other hand an r = produces an / c,,) (c 1? c 2 equation between the cofficients
is always the case if *P is the product of the conjugate values of an unsymmetric function,
when, as a This rational factors.

affect only
result, the Galois
resolvent breaks
up
into
all
*>i,
for
it
c2 ,
c,,)=
//"^
K,
is
/s ,
-^J;
follows then from
2 that one of the factors of the product

also necessary, is a ques
must vanish. Whether this condition tion which we will not here consider.
A special equation might also be characterized by several rela A direct consid tions between the roots or the cofficients or both.
eration of the latter cases would again present serious difficulty. From the preceding Section we recognize, however, that whatever
the number or the character of the relations
may
be, if
they pro-
ELEMENTAKY CASES
duce an
0(0?!,^,
155.
.
all
- RESOLVENTS.
175
affect,
they can
be replaced by a single equation
..X)=0.
We
will illustrate the ideas
and
definitions introduced
in the preceding Sections
by an example.
Suppose that
all
the roots of an irreducible equation f(x)
are rational functions of a single one
among them
x2 =
The two equations
have then one
all
<?<>
(x,),
xs
<? 3 (
root,
the other roots
x^ x X
2
,
in
B
, .
.
common, and since f(x) is irreducible, x n of the first equation are also roots
.
of the second.
Consequently f(x)
is satisfied
by
and
in general
by every
âOî)]
Again
for otherwise the
(,/5
2,3,
.n).
two equations
/0)=0,
<F*(x)
<Fp(x)
=Q
:
would have the root pyfa) and consequently all the roots of f(x) = in common. It would then follow that y> a (x ) = (o^), i, e., two
of the
roots
it
of f(x)
would be equal.
This being contrary to
assumption,
appears that the series
coincides, apart
from the order of the elements, with the
series
2)
x
_
l ,
^(ajO,
<p 3 (#i),
^n(î).
now what substitutions can be performed among the roots x x x n without disturbing the relations existing among them. If such a substitution replaces x by x y then it must also replace every x a by y>a.(x y ). The substitution is there There are fore fully determined by the single sequence x lt x y conditions. These the which n substitutions satisfy accordingly only and Q its form a group degree ( 129); for the system is transitive
determine
1 ,
We
and order are both equal
to n.
176
This group
tions
tt is the group of the given equation. For the rela which characterize the given equations are equivalent to the
single relation
</
.,3,
[sr,
PaOr,)]
+ &[>,
is
e (x
:t
]+
+?
[x n
vfa)]
= 0,
and
if
this
function &
,r
replaced by
every
,r a
remain unaltered, then when ,r, is must be replaced by <f a (^Cy) P[^v(^)}t
to
exactly as under the application of the substitutions of

156.
&
Without entering further
into the theory of the
group
of an equation
we can
still
give here two of the most important
theorems.
Theorem
transitive.
II.
The group of an irreducible equation
is
Conversely, if the
group of an equation
is
transitive,
the equation is irreducible.
Thus
if
the group
of the equation
t
f(x
is
:EE (jc
f,
.r,
(jc
xn
( )
intransitive,
suppose that it connects only the elements
ir,
jc.2
,r a
with one another.

(f
Then
(x
the function
x^
(x
.r.,)
(x
.r a
is unchanged by G, and consequently belongs either to the family In either case <f is rationally of G or to one of its sub -families.
known, that
is,
the coefficients of
cr(.r) is
cr
are rational functions of

)
known
above
quantities, so that
a rational factor of /(-/
Conversely,
if
f(x)
is
reducible,
factor
cr(.r)
of the
be rationally known, and which replaces any element .r, ,r._,,
form
will
G
. . .
can contain no substitution

,
the rationally
known function
x a by -r a+l c would not remain
for
otherwise
for
unchanged
all the substitutions of G.
Consequently
is
intransitive.
Theorem
group
III.
// all
the roots of
an
irreducible equation
are rational functions of any one
among
them, the order of the
of the e(>nalion is
n.
Conversely, if the group of
an equa
then all
tion is transitive,
the roots of
tJte
and
if its
order and degree are
equa.1,
equation are rational functions of airy one
among
them.
part of the theorem follows at once from proceed to prove the second part.
The
first
155.
We
ELEMENTARY CASES
RESOLVENTS.
177
From
the transitivity of the group follows the irreducibility of
the equation.
If we specialize the given equation by adjoining to it the family the group will be correspondingly reduced. It belonging to x will in fact then contain only substitutions which leave x un
1 , l
( changed. But as the group is of the type 129), it contains only one substitution, identity, which leave x unchanged. Accordingly
l
after the adjunction of a^
all
functions belonging to the group 1
or
to
a? 2
,
.
any
.
Xu
xn
In particular larger group are rationally known. are rationally known, i. e., they are rational functions
of #,.
From
proved in
this follows again the
theorem which has already been
155:
Theorem
VI.
If all the roots of an irreducible equation any one among them, they are rational
functions of every one

157.
among
them.
From Theorem
III, the
is
group of the Galois resolvent

of order n\
equation of a general equation
To obtain
it,
we
apply to the values

i-
Ci,,Ca,
...}
s
.
the resulting rear Since every substitution of the x s affects all of the r s, the group of the c s belongs to the The group of the x s and that of the c s are simply groups
all
1 ,
xn and regard
rangements as substitutions of the r
isomorphic ( 72). For an example we

the third degree. then
may
take again the case of the equation of

of f(x)
The groups G
and
/ of
F (?) =
are
G=
[1, (a?,a;2 ),
8)
(aw), (avs), Ow*s),

5)
UW
2 )]
r= [1, (^- (V) (ê), GVe) (^

i-::=\r=~-\
\^l~ô) V^2~3^G/i Vl*5*4^
iZ = Z
\
(V*), GV*) (^5)
(V),
(V~
- - \~\ J"! t)~3.U-
equation is an affect equation with a r s only group G of order r, then of the n substitutions among the r s which those those are to be retained which connect any => with
If however, the given
!
together with
F;(C) Of F(C).
c.
belong to one of the rational irreducible factors
12
178
158.
We
apply the name resolvent generally to every
/>-val-
ued function
f(x)
^(,T ]5
x2
xn )
of the
th
/;
roots
of
given equation
is satisfied
= 0.
its
is
The equation
of the
is
degree which
by
e>
and
conjugate values
called a resolvent equation.
This desig
nation
appropriate, in that the solution of a resolvent equation
reduces the problem of the solution of the given equation. Thus, for example, in the case of the equation of the fourth degree, we employed the following system of resolvents ( 150)
:
1) 2)
The 2-valued function The 6 -valued function
VJ =
^
</
(#1
l
-#2)
2
#3)
(# 8
#*)
= (x x
-f-
-\-x3 x^)-}-oj(x l x 3 -\-x 2 x^

2 s
3)
4)
The 12-valued The 24-valued
function
function
= (#1^2 / = a (x
l
^(x&i -f x x. ), X3 X *) (x-iXs -\-x2 x4 \ x ) -f- a (x a? 4 ).

l
After Originally the group of the equation was of order 24. the solution of the quadratic equation of which the two- valued
function \7 J was a root, the general, symmetric group reduced to The extraction of the alternating group of 12 substitutions.
a cube root led then to the group
150) of order 4; another
square root to the
group
of order 2;
and
finally
we
arrived at
the group 1, and the solution of the equation was complete, the function r being superfluous. The above reduction of the group of an equation to its /> th part
by the solution of a resolvent equation of degree For general equations and resolvents this reduction
/>
is is
exceptional.
not possible.
We
shall see later that
it is
possible for the biquadratic equation
and the particular series of resolvents employed above only because the family of every resolvent was a self-conjugate subfamily of that of the succeeding one.
159.
isfies
Given any
-valued resolvent <y(x lt

th
/>
a?2
a? w
),
this sat
an equation of the
P
degree
H-
A^P-
v<>-
...
A = (\
p
To determine
as in the preceding Section.
the group of this equation we adopt the same Suppose that

?1
r
method
?2J
<r
are the //-values of
c.
Every substitution of the group
of the
ELEMENTARY CASES
equation
jj
RESOLVENTS.
179
/(.*")
values.
The
latter substitutions
cr.
produces a corresponding substitution of these form the required group of the

This group
is
resolvent equation for

transitive, since
isomorphic to
6r,
l
and
it is
Â
If,
in particular, the
it
contains substitutions which replace <p by every group of y is a self -conjugate subgroup
l
of G, then
is
also the
group of
cr
cr
,
<p p
The
latter values
are therefore
all
rational functions of c^,

is
and the group of the
resolvent equation
160.
a group
i2,
as in
156.
* Following the example of Lagrange we might attempt
to accomplish the reduction
and possibly the solution

later that this
of the general
Galois resolvent equation

ents.
F()
It will, however, appear ceed in general for equations of a degree higher than the fourth.
by employing particular resolv method cannot suc
*Mem.
d. Berl.
Akad.
Ill;
and Oeuvres
III, p. 305
ff.
CHAPTER
X.

161. of unity
u>,
The equation which
is
satisfied
by a primitive
p th
is
root
(p, as usual, always denoting a It is of the
the cyclotomic equation.
prime number), form
called
and
2)
its
roots are
co,
w2
to
,
to*"
1
.
We
prove that the
left
member
of 1) cannot be expressed as a pro

coffi-
cients.
duct of two integral functions v (x) and d> (x) with integral For if this were the case, we should have for
x=l
KI)<KI)-P,
and consequently one of the two integral 1. must be equal to Moreover, since
root in
vanish.
factors, for
<f>
example
at least
a
<?
^(1),
(x)
has
one
common
with
1),
at least one of the expression
(to )
must
Consequently
ôy-i^K-i J.-sK-i
where
iu l
J-O
is
the series
may be any cu^w^w^
root of
.
1),
1
.
since the series 2)
identical with
w^"
The equation
has therefore
all
the quantities 2) as roots.

is
l
Consequently the
left
member
3)
of this equation
2
divisible
by
1).
1
Suppose that
2
.
<p(x)
?(x
v(x*~
= F(x) (x^- + x*- + )
+x-h 1),
From
where F(x) is an integral function with integral =1 3) we have for x
coefficients.
and therefore [^(1)]^

Accordingly
1)
is
which
is
equal to
1,
must be
divisible
by
p.
not reducible.
181
Theorem
ber
I.
The cyclotomic equation for the prime
num
is irreducible.
162.
If
now g
is
a primitive root (mod. p), then the series 2)
of the roots of the cyclotomic equation can be written

4)
1 *,*;;?.., a**.
Sinco
is
I) is
irreducible, the corresponding
group
is transitive.
ffS
.
There
therefore a substitution present which replaces ^ g by w

" to 9
is
Then
every
replaced by
and the substitution
is
therefore
8
= (w* at*
(a*
The p
tions in
this group,
powers of and from
form the group of 1). For they 156 the group contains only p
all
occur in
1 substitu
all.
We
form now the

(o
cyclical resolvent
sl
+ aa+a
ai*
+
1
2 -f a*-
^" )^2
1
,
in
which
a denotes a primitive root of the equation
z*-
= 0.
For brevity we write with Jacobi
From
129 the resolvent
p~
(a, u>}
is
unchanged by
and
its
pow
It can therefore be is, by the group of the equation. a of 1). the coefficients and function of as a rational expressed
ers, that
If
by
5)
TJ
we denote we have
a (p
th
l)
root of this rationally
known quantity
(a,
w)= T
is
It l)-valued function of the roots of 1). the substitution for the of group, changed by every substitution s converts it into
r x is
The quantity
a (p
(a,
= a-
(a, ai)
r1
182
It follows
from the general theory of groups that every function

l
.
of the roots can be rationally expressed in terms of however give a special investigation for this particular case. group of the cyclotomic equation leaves the value of
6)
We
will
The
(a\w)(a,w)
effect of the
p-
-*
is
unchanged; for the

function into
substitution s
to convert this
p-
-*
(a\w9)(a,w*)
= a-*(a\a>)a-*- + *(a,io)*= (a\ai)(a, o^" -^

1 1
]
-*
i.
e.,
,
into
itself.
If
jTA
where in particular
now we denote the rationally known value 6) by ~ = rp Tn we obtain for = 1, 2, .p 2,

/.
l
.
.
the following series of equations
Combining with these the obvious

coefficients of
1)
relation
among
the roots and the
(l,o*)=
-1,
we
obtain by proper linear combinations
It is evident that a
change
TJ
in the choice of the particular root a
or of the particular value of
__
\f
I\
only interchanges the val
ue
tu
among
themselves.
Theorem
the
II.
The solution of
the
cyclotomic equation for
prime number
1)" root of
requires only the determination of a primitive
root
of the equation
(p The cyclotomic equation therefore reduces

tions of degree
zp 1 the extraction of the 0, and an expression which is then rationally known.

to
two binomial equa
1.
THE CTCLOTOMIC EQUATIONS.

163.
plified.
The second of these operations can be still further sim The quantity T^ is in general complex and of the form
1
Since
now (, w)*- and (a- w- )*1

1 ,
are conjugate values,
it
fol
lows that
p-l
(a, at)
~
(a
~
\
a>
p~l
)
p (cos
ft
-f
i sin
ft
) />
(cos #
i sin
#)
/>
Again
it
can be shown, exactly as in the preceding Section, that

(a,
>)(a-\
w~
belongs to the group of the cyclotomic equation and is conse and of the coefficients of 1). If quently a rational function of
we denote
its
value by
U we
have
Accordingly for any integral value of k

cos "
--j-
+
.
sm
.
Since
and # are both known, we have then

III.
Theorem
requires
the
zp
is
determination of
The solution of the cyclotomic equation a primitive root of the equation
1=0, the division into p 1 equal parts of an angle which then known, and, the extraction of the square root of a known
The
latter quantity, U, is readily calculated.
quantity.
We
have
U=
To reduce
(t
-f
M 9 -f
product we begin by multiplying each pair of cor of the two parentheses together. The result is terms responding
this
+ 1 + !+.,. +l=-p-l
of
Again,
fc
if
we multiply every term
the
first
th
term to the right of the corresponding term
parenthesis by the in the second par
enthesis,
we
obtain
K)
-*(-*+
(ju- yk
-** +I +
Now
is
th
root of unity
+--^ +i +**+ ...) ^ different from 1; for
if
184
then
gr*+l
EE(),
(mod. p),
i.
e.,
or
l.
The
quantity
K)
is
therefore equal to
and consequently
U=p
(a-
+ a-*+
T/ie
4-
Theorem
square root
164.
is
quantity of Theorem III, of which the to be extracted, has the value p.
IV.
The resolvent 5) was (p l)-valued, and consequently the preceding method furnished at once the complete solution of
the cyclotomic equation. By the aid of resolvents with smaller numbers of values, the solution of the equation can be divided into
its
simplest component operations. Suppose thatpi is a prime factor of
1,
and thatp
^P\Q_\-
We form then the resolvent

(a>
+ a^ + a*a>9* +
3*1
2
. .
-f-a*-
^"
)*!,
where
is
a primitive root of the equation
= 0.
The
i
w^ are all different, and the higher powers of V, It follows that, if take the same values again.
values
a ,
the resolvent above can be written

(sô
+ Wi +
s
iV 2
i^
^-
1)*>,
or,
again in Jacobi
notation,
K,?)*.
By
is
the same method as before

if
we can show
ff
that this resolvent

it
unchanged group of 1) and
M
is
is
replaced by
ta
that
is,
that
consequently a rational function of
belongs to the ^ and of the

coefficients of 1).
185
We
denote
its
value by N,
= v,\
and have
accordingly
If then
we
write precisely as before,

(
a f>?)( a
i,v)*i~
-* =
Ni,
it
appears that
N^
is
rationally known,
and that
These several functions are

that
is,
all
unchanged
if
is
replaced by w" \
they are
unchanged by the subgroup
We
have therefore
Theorem
V.
The
p r valued
resolvents
cr ((
cr,
<
,
of
group formed by the power* oj ,sî. They can be obtained by determining a primitive root p 1 = and, extracting the p* h root of a of z quantity which /.s then rationally known.
i
the cyclotomic equation belong to the
If
is
a second prime factor of
1,
and
if
p
1)/>
=p
p.2 q>,
then the resolvent

(
w
is
Vl
_|_
a^ffVi
_j_
,."
,_,-
-f
-f-
/i
/(?1
)^
in
which
,,
a primitive p.2 -f
-h-i
th
root of unity,
is
reducible to the form
fa
where
-f- "o/r,
+..-:
/*" /ft-i)*a
,/)*s,
This resolvent
is
is
unchanged
if
<>
is
replaced by
\
w^
that
is, it
unchanged by the substitutions

12a
sp
Consequently
it
186
$PO
,
can be rationally expressed in terms of known. Again if we write

S X)
2 ,
if
is
regarded as
*
y>
,
then
A is
also a rational function of
and we have
n J
/=-
ir
L
?o
+ î-f M uW+
-Wo
>
PIJ-I of the cyclotomic equation belong to the group formed by the pow ers of 8 P P They can be obtained by determining a primitive root
}
Theorem
*
VI.
The
p p
-valued resolvents / ?7n
*.
of
l>
=
By
rational in this primitive root

165.
and extracting a and
p.?
root of a quantity
c,,.
which
is
in
continuing the same process
we have
.
.
finally
the solution of the z ps cyclotomic equation for the prime number p requires the determina
Theorem
VII.
If
= Pip
tion of
a primitive root of each of the equations

2*1
1=0,
z**
= 0,
z*V
= 0, ...
p
th
2
,
and
the extraction successively

is
of the p^,
p-f\
roots
of
expressions each of which
rationally
known
in terms of the pre
ceding
known
is
quantities.
c1 are rationally known, since they n cr 2 <p^ same group. Similarly the coefficients of
,
.
If f
all
given,
to the
belong
M) (x
(lt
^)
(x
afri)
(x
u* 81
*-
^)
= 0,
are all rationally known.
rem
is
tors 7).
Accordingly after the process of Theo has been carried out, the equation 1) breaks up into p fac Since the group belonging to each of these new equations
l
transitive in the corresponding roots, the factors 7) are again all
irreducible, so long as only the coefficients of 1)
and ^ are known.
187
Again, after the process of Theorem VI has been carried out, All the values of this function belong to the same is known.
all
group and they are therefore

the coefficients of
\<JC
rational functions of /
(l
Similarly
to) \3C
to-
~j (3C
i>J
~J
\3C
(O

ble,
The equations
7) are therefore
factors 8),
.
now
reduci
and each of them resolves into

until
p.>
irreducible within the

this
domain defined by /
We
which are again can proceed in
way
166.
we
arrive at equations of the first degree.
The
particular case for which the prime factors of

is
are
all
equal to 2
of especial interest.
Theorem
of
in
VIII.
m If 2
tomic equation belonging to
quadratic equations.
-f"
-(- 1 is a prime number p, the cyclobe solved by means of a series can p In this case the regular polygon of
p = 2"
In
1 sides
can be constructed by means of ruler and compass.
fact, for
one root of the cyclotomic equation we have

2.
,
2-
2,
2-
to=.COS
\-lSin
rrrCOS
lain
p
1
p p
,
-4- to
= 2 COS 2-
and consequently the angle -- can be constructed with ruler and

compass. In order that 2 m
-\-
1
/x
may be
ra ]
2M
,
a prime number,
it is
necessary that
2 A
m = 2M
For
if
m=2
where
1.
m
1,
is
odd, then 2
4~ 1
mi
)
(2
4- 1
would be
divisible
by 2
-f
,u
If
= 0,
2,3, 4,
the values of
are actually prime
numbers
p = 3, 5,
and
fore be constructed.
17, 257, 65537,
in these cases the corresponding regular polygons can there
But
for
/*
=5
we have
.
2* 4- 1 = 4294967297 = 641
6700417,
188
so that
it
remains uncertain whether the form 2 2
1 furnishes
-|-
an
infinite series of
prime numbers.*
cases
We add the actual geometrical constructions for the = 5 and p = 17. For p = 5, we take for a primitive root g = 2, and obtain accord
167.
ingly
= 1,
g,
= 2,
g-
= 4,
=3
(mod.
5).
Consequently
- -l+A/5
If
o>-
-1
c^
V5
and
Cj
is
substituted for w, the values
are interchanged.
The
algebraic sign of
\/5
made.
is
therefore undetermined unless a par

If
ticular choice of
is
we
r
prescribe that
sin
= cos
then
CF O
2*. -
-f- *
.2* -,
<^>
-j-
w4
= 2 cos 2-=
crj
<f r
or
-J-
w3
= 2 cos 4-^-
consequently cr n > 0, to be taken positive.
< 0,
and the \/5
in the expression above is
Furthermore
1=0;
1+ V5 + VlO-f 2 V5
i
1+ V5"
* Cf.
all
V 10 + 2
by Baltzer: Crelle
87, p. 172.
Gauss
the
Disquisit. arithm.,
22^
362.
The statement there made that Fermat sup

At
posed
numbers
2
25
2l2
+1
to be prime, is corrected
present the following exceptional cases are known:
+]
+i
divisible
divisible
by
641
(Landry),
(J.
. .
2 2
by
114689
167772161.
Pervouchine),
s23
divisible by
divisible
(J.
Pervouchine; E. Lucas),
),
T
Cf. P. Seelhoif
:
4- 1
by
274877906945
(P. Seelhoff
Schlomilch Zeitschrift,
XXXI.
pp. 172-4.

the sign of i being so taken that the imaginary part of w and that of w* negative.
is
189
positive
For the construction of

the regular pentagon it is sufficient to know the re2-.
solvent
er
2 cos "1 D
Suppose a
circle of radi
us 1 to be described about
as a center.
at
On
the tan
gent the horizontal radius

a distance
is
the extremity of
OA
AE
laid
off.
Then
OE =
If
now we
take
E F = E O,
we have
A F = 2 cos -ô
Finally
if
2-
AF
HJ,
H,
is
bisected in
OC
= cos
to
then
GHJ drawn to O A and since cosHOC = AG HOC=COJ^~, o

G and
//
O,
and
are therefore three successsive vertices of a
regular pentagon.
168.
For
p=Yl
we
take for a primitive root g
6.
Accord
ingly
we have
1,
6,
2,
12,
4,
7,
8,
14,
16,
11,
15,
5,
13,
10,
9,
3,
1;
190
THEOBY OF SUBSTITUTIONS.
find
r <-
To
V i5
we multiply every term

it
of ^
in y
l ,
ih into the k term to
the right of that immediately below k obtain then 0, 1, ...
taking successively
We
?o
<Pi
<Pi
+ ?o -f ?o + ?o + Pi + ?i -h Pi + ?o
4 Oo +
<Pi)
4.
Consequently
2
-f-fp
= 0,
-1
.2
-1+VT7
2
where the sign

to
of \/ 17 is If
undetermined until the particular root
is specified.
we
take
2* .2* Sin Tn> COSy^ -f~*

.
we have
for the determination of the sign
^=
W3
_|_
W U)
6:r
+
.
W5
+ W + (w + ^1!) +
!
2)
(<tf
+ wlU)
f = O 2
[^08^ +
lO^r
127r
COS
-yy
-H COS
+ COS^y J
14>T
and
V 17
above must be taken positive.
We
have further
*o,*i= 2
=:
-\/~~T~
/2 /3
^-y~~^
The
algebraic signs are again easily determined.
We have
Xa
(a,
+ (, +
=2
(cos
~ + cosL
> 0,
Consequently
With /
as a basis
we proceed
further:
Since
fore
now
2_
2cop^,
r ^ = 2cos^,
we have ^ >
and there
These results
of 17 sides.
as a center,
the construction of the regular polygon a circle of radius 1 to be described about O Suppose
suffice for
zontal radius
and a tangent to be drawn at the extremity of the hori OA. On the tangent take a length
AE
then
A/17
~T~
Further,
if
EF=EF
= E O,
we have
192
OF =
-- + 1,
OF =
Taking then
we have
AH = -AF + We bisect AH in T; then

f
We take now AS = 1,
eter; if this
and describe a semicircle on
H S as a diam
meets the continuation of
OA
in
then
Again
if
we take
LK AY
and
circle of radius
LK
KL = LM~LN,
obtain
and describe a
about i,
we
The greater
of the two lengths
AN,
AM
is
equal to ^
we may
write therefore
If
is
the middle point of
AM,
and
if
QP
AO, and Of)
A O,
then
Q and D
are two successive vertices of the required reg
ular polygon of 17 sides.

169.
We
is
consider
now
the case p
2,
under the assumption

gp
3
p>
2.
If g
a primitive root, then

[
flri*
4
[ j^
To determine ^
Sections,
^,
we use the same method
as in the preceding
and obtain
+ [/+ + /+"+...+
The exponents which occur here
in
any bracket
^+
4-
1,
flW +1 + 1),
0V
+1
+ 1),-
are plainly either all quadratic remainders, or all quadratic nonIn the first case the value of the remainders, or all equal to 0. and in the third corresponding bracket is /v
m.,
>
;
represent the
number
of brackets of the several
a+1
<7-
+ 1==0 (mod. p),

is
then 2(2a
+ l)=p
and accord
ingly
an odd number.
The
third case occurs therefore only

for
when p = 4fc-f~3, and then only

sequently
Since
once, viz.
=p
is
__ 3
.
Con-
m =
;J
or
is
w =
3
according as
p ^~
6
in
___ ^
even or odd.
cr
()
cr,
rational
1),
and integral
product
is
the coefficients of the
cyclotomic equation therefore take

r
<,
this
an integer, and we
may
<~\
Wl 3
^~
n=
//
cr
-f- cr
where n
is
an integer.
(
It follows
then from
(
/S)
that
w,
+n
cr
(1
-f-
in., -\-
<f
1
In this equation all the powers of u> can be reduced to powers lower than the p th and the equation can then be divided by M. The is of then 2 at the and still resulting equation degree p highest,
,
has the root

It
is
to
in
common
with the equation 1) of degree
-1.
therefore an identity and
Consequently we have for the values of
>w.
m.,
and
CP O
<r,
~ m = p,
^<f,=
p ~
1
,
(p
[
even
~^
1)
195
y-o
_V;
"2
-1
^
~~
!
j-i
(ulLL?
-i-V(--i)
2
>
~~
where the algebraic sign
of the square root is necessarily undeterm ined so long as the particular value of to is not specified.
170.
We
consider
<"
now
the two equations
z
;,
(tf
^.)
<^
(a-
"X" = 0,
3
-(x
w J)
(xta*)
(x-r~-u>
ii:
(x
~
jiJ
0.
The
roots

ta-
coefficients of these equations are
un
changed
if
is
replaced by
;/
".
If therefore,
when
the several
factors are
a
,
it
y>Q
must
or
multiplied together, any coefficient contains a term aJ also contain terms mat* 9 \ .... that is, it con
2 ,
mw
tains
9?j
according as a
p).
is
a quadratic
remainder or nonwill
remainder (mod. form
Accordingly every
coefficient
be of the
1)
and on introducing these values we have
ZQ
_X+Y
2
,
where
that
X and F are integral functions of x with integral coefficients.

obtained from z by replacing ut by w ? or ^ by a change of the sign of the square root, we have
l
Again since z
is,
is
by
<pi
,=
Hence
196
and we have
Theorem
4
where
IX.
For every prime number p
and
Y are
integral functions of
x with
integral
coeffi
cients. *
*The extensive literature belonging to this Chapter is found in Bachmann: Lehre von der Kreistheilung, Leipzig, Teubner, 1872. The present treatment follows in part the method there employed. The two figures are taken from Bachmann s work.
CHAPTER XL
171. The cyclotomic equation has the property that it is irre ducible and that every one of its roots is a rational function of every other one. turn now to the treatment of the more general prob
We
lem of those irreducible equations of which one root x\

function of another root
is
a rational
Among
these equations the cyclo
tomic equation
is
evidently included as a special case.
Suppose that
1)
is
/(*)
=
x
l
are connected
2)
the given irreducible equation, and that two of the roots x\ and by the relation
SB
,
=*(,),
Then
/[^(
1
where
is
a rational function.
/(,)
= 0,
)]=0,
x
l ,
so that the irreducible equation 1) has the root

all its roots,
and consequently
in
common
with the equation
3)
/[()] =0.
6 (a^)
In particular x\
must be a root
of 3), so that
Consequently 0[#(#i)]
\0\0(x$]
e [e
it
\
is
a root of
1)
and therefore of
3).
Then
is
a root of
,
1),
and so
on.
With the notation

[o (*)]
()] = p (x)]
all
[p ()]
<fl
= e* (x),
appears that
the
members
of the infinite series
are roots of the equation

finite
1).
Since however the latter has only a
number
of roots
it
follows from a familiar process of reason-
198
is
m ing that there must be in the series a function O (x^ which while all the preceding functions to the initial value x
l ,
equal
are different
from one another.
reproduces these same values in

0>"( Xl )
The continuation of the series then the same order, so that only
...
= 0* m (xi)
l
take the initial value
l ,
and that for k

k
)
<m
l
only
...
<r+ (x
are equal to #*(#,)
If the
.
m + tt
(x
)=
system of
roots thus obtained does not include all the

1),
roots of the equation
suppose that x,
a
is
Then
on.
x.t
also satisfies 3),
and therefore
0(x.2 )
any remaining root. is a root of 1), and so

/;.
Accordingly we have now
new system
of
different roots
But
4)
since the equations
0*(yyy =
x^
,
o*(z)
a?
2
=Q
with the irreducible
have each one root y
in
all
common
The latter. equation 1), by former equation of 4) is therefore satisfied by x, the latter by JL\ is a multiple of and // is a multiple of w, i. e. consequently
they are satisfied
the roots of the
,
.
/j.
m=
/Jt.
Again
of the
all
the roots of the second series are different from those
first.
For
if
6\x,}
then,
= d\Xl
=
on
~ the operation m b applying M x2 (x 2 )
(o,6<m), we should have

,
=0+-
(x l ),
and x would be contained

2
in the first series,
which
is
contrary to
assumption. If there
is
another root
cc3
not included
among
the
2m
already
found, the same reasoning applies again.
We have therefore
is
v
Theorem I. If one root of an irreducible equation f(x) = a rational function of another, then the roots can be divided into
systems of
roots each, as in the following table
199
5)
U,^),^),..
I
- w,
m
.
xv
0(x v ), 0\x v),
~\x v ).
The function
is
such that for every root x a
and
the equation
f(x)
is
of degree mv.
172. We can now determine the group of the equation 1). Since the equation is irreducible, its group is transitive ( 156); it therefore contains at least one substitution which replaces x\ by any
arbitrary
,r a
It follows then that all the roots of the first line of
th The group of 1) is there line. 5) are replaced by those of the fore non-primitive and has v systems of non-primitivity of ele
The number of admissible permutations of the > systems not as yet determinate; in the most general case there are v! of them. If any ,r a is replaced by # A (,r a ), then every (P(x a ) is replaced +A by 0* (sca); there are therefore possible substitutions within the
ments each.
is
single system.
of
v
The order
of
>
!
of the
v
.
group of
1) is
therefore a multiple
m and a divisor Theorem II.

contains
>
It
several lines
The group of the equation 1) is non-primitive. systems of non-primitivity, which correspond to the The order of the group is of 5).
r
= ^ m",
where r
is
173.
a divisor of vl. In the following treatment we employ again the notation
of Jacobi
where w
is
a root of the equation
&m
= 0.
Similarly
we
write
We
form then the following resolvents:

f,-= (1
,
o( Xl )
),
V*
= (1
ofa) ),
?v
= (1
e( Xv )
Sî
is
symmetric in the elements of the
first
system and
is
changed
it
in value only
when
the
first
system
its
is
replaced by another;
is
therefore a
v-
valued function,
values being
200
Every symmetric function of the

coefficients of 1)
.
<?
s is
a rational function of the
The
quantities
that
6)
is,
th
coefficients of the equation of the
>
th
degree
c-^which
?i,
p
2
, .
.
+ &*>
~2
--.
SV
= Q,
of
are the roots, are therefore rationally known.
Theorem
is
III.
The resolvent
a root of an equation of degree

in
the coefficients of
which are
equation
rationally fcnoirn
terms
of
the coefficients
of
the
/ 0*0 = 0.
174.
The equation
of
6)
has no affect (158), unless further
relations are explicitly assigned
among
the roots x\
x2
xn
If
however any root
<f a 6) has in any way been determined, the can be obtained algebrai values of the corresponding x a 0(x a ),
,
cally
by exactly the same method
as that
employed
in the preceding
Chapter.
0"
Thus, the equation ^

l
tf)
of which the roots are x, 0(x),
&\x\
(x) is irreducible.
Its
1
group consists of the powers of the sub

if
stitution (100*
". .
.
).
And
we
write
where
have
to
is
now assumed
to be a primitive
mth
<o0
,
root of unity,
we
"
[0(x)+w(P(x)+
that
...4-o"
^-(o;)]
=
=
m
<
K %)]"
\0(x)+
\x)-r...
<t>
m - Om (x}~]
l
is
!T,
is
2 unchanged by the substitution (100
Om
).
Conse
quently T, the primitive

quantity
l
is
a rational function of the coefficients of G) and of
mth
root of unity
r,
.
T we
denote by
The m th root Again if we write

t.
of this
known
(V(a;))
it
(a,,
#())"-
=T
can be shown by the same method that we have already frequently
201
coefficients of #).
employed that
2\ is
/
also rational in
w and the
1,
Taking successively 1, ing equations, we have then
= 0,
2,
and combining the
result-
The function
is
TJ
being ra-valued, x also admits only

a;,
,
these coincide with

substituted for
r,
0(#), # (#), ....
If
as
any other
m values, and w th root of T

l
the
values of
are permuted cyclically.
Theorem
are
IV.
If the
roots of
*i,%i), ^(i),
where 0(x)
is
m ... o "\x l ),
Om
(x^}
a rational function for which
then the
solution of the equation requires only the determination of a primi m th tive root of z Q and the extraction of the l root of a
known
quantity.
Theorem
algebraically.
V.
If one root of an equation of prime degree

root, the
is
a rational function of another
equation can be solved
For
in
this
v
case
we have mv=-p and
m>1
consequently
m =p
and
1.
175.
If all
the coefficients of f(x) are real, the process of
the preceding Section admits of further reduction.
The quantity
T
The
latter
= (u>,
m
e(x))
= p(cos # + i sin =
ft
can be rationally expressed in
terms of w and the
coefficients of /.
\/ being real, the occurrence of i Consequently. /entirely to the presence of w.
1 in
is
due
(a>
\ 0(x)
p(cos ft
sin ft]
^T/ =
7
/>;,
18a
202
THEORY OF SUBSTIT UTIONS.
where
is
.
again rationally known, since
it is
unchanged by the
group of &
We have then.
#
rt
=û( cos m
+ 2k-
VI. // all the coefficients of f(x) are real, the second operation of Theorem IV can be replaced by the extraction
of a square root of a known quantity and the division of a known angle into m equal parts.
176.
If the
Theorem
m of the Theorem IV is a compound number,
the
solution can be divided into a series of steps by the aid of special resolvents. If is any factor of ra^/i/, where 1 m, we arbitrary
m
.
take
0i
= x, + 0"
fa)
+ tfr fa) +
+ rt-
- ). fa)
X = <n-
fa)
+ <Pi-* fa) + P*-*fa) +... + <rV"i-
fa)
and consider the resolvent
in
which
is
a primitive
m^
root of unity.
This resolvent
is
equal to
n 2 replaced by ^(ajj), then ted, and the resolvent is unchanged.

If
a?,
is
^
It
,
are cyclically
permu
denote
can therefore be ration

quantities.
ally expressed in
terms of
and known
root
We
this expression
by
and
2
0i
+
2
1
i02
m^ by ^ + iVa + + aT its
1
so that
0-,
=
.
",
If then, as before,
(01
it
we
3
write
)
lV 2
+ V+
is
(01
02
+lV +
3
-A
)"
= Â,
appears that 7VA
rationally
known, and that
208
Theorem VII. The m r valued resolvent 4^ can be obtained 1 and extracting the by determining a primitive root of z "\
m^ root of
By
a known quantity.
continuing the same process, we obtain
Theorem
degree are
VIII.
// the roots of an equation of the
m*
and
if
m=
?7i 1
mm
2
the solution of the equation requires only
the determination of
a primitive root of each of

m J2 2
the equations
2V-1-0,
and
1=0,
Zwa
-1-0,
th
,
...
3
,
.
. .
the successive extraction of the mj
m th ,m th
2
roots of
the
expressions, each of which is rationally
known in terms of
pre
ceding
results.
177.
The
solution can also be accomplished by a
still
different
method.
Suppose that
m=m m
l
mu = m
= m. n.
2
:
...
=m M n u
.
Then we can form the following equations

l
}
^" 1)wf (a;,), g (x)=Q, with the roots x 19 ^(x ), (P^fa), and with coefficients which are rational functions of a resol.
/\
^
1
vent
x\
~ x + 0m
\
(*i)
+
m
l
.
7.i
is
a root of an equation
hi (/)
of degree
A.
"(a5,), 2(Xi)> g 2 (x) 0, with the roots x^ 0"X#i)> and with coefficients which are rational functions of a resol vent / 2 = Xy is a root of an equation "-(î)
. . . i
2"
=
^(a?)
of degree m.2
= 0, with the roots

cofficients
ff
,,
0(x^0^(x,\
.
(11
.
.
^(x^
and with
vent *
fe
= *,+
select
which are rational functions of a resol mu / is a root of an equation , (x )

1
+
.
.
w (^) =
of degree
m^
. .
If
now we
mn w
so that they are prime to each
other, then the equations
204
have only the one root x
method
of the greatest
common to any two of themcommon divisor this root can be

}
By
the
rationally
expressed in the coefficients of g

cients of f(x)
g2
gw
that
is,
in the coeffi
and
in
/2
/.
The
solution of f(x)
depends therefore on the determination
of one root of each of the equations
of degrees mj
m.2
ww
respectively.
If
where Pi,p 2
are to take
.p u are the different prime factors of m, then we
m
If for
a
1
=--p
i,
m = pa
2
2,
the exponent A is any one of the equations than of then method recourse must be had to the earlier 1, greater solution to determine a %\.
m w =p = h^(*/)
.
a
ta
<.
In illustration of this type of equations we add the two In the one case we take following examples.
178.
where
that
o.d
A=
/3, Y\
= T,
8i
=$
?/?
are real quantities.
We assume
yjîQ; otherwise we should have

0(x)
= = -.
Y
.
The
functions
(x]
(x)
will also
be linear in x with real
coefficients.
We may
write
Again
^rr i
<r
" r^MI -
^
"
"
and a comparison of the two expressions
for 0" (x)
shows that
205
From
7)
these equations
we
obtain at once the characteristic relation
A.r
= (Mi
An)
(-X-i
ir-.-i)
By
a5 # Pr or by V/V dividing every coefficient of 0(x) by a<? /3y is positive or negative, we can arrange that for according as
the
k
new
coefficients of
O (x), the relation shall 8)
0(x\ and consequently for those of every hold

ad
fr=l.
it
We
determine now under what conditions

tr*(x)
can happen that
= x.
ff
The values
equation
of
x which
are
unchanged by the operation
satisfy the
For these
a
fixed values
we have
therefore, according as ad
fr
1,
A
and consequently
A)
that
is
We
that
assume
in the first instance that
a?
and x" are
distinct,
Nl. We have then

x
(T(x)x"
The necessary and
fore that
x"
sufficient condition that tr(x)
=x
is
there
206
the quantity in the bracket.

braic sign
This condition can be satisfied by complex or by real values of In the former case the upper alge
must be taken, and further
so that
we may
write
+~ = cos
<5
 = 2k~

>S
-
i sin
2 m<f>.
we must have
a
ki:
,
2ra^
<p
and
+d_
k-
~2~
where k
is
~^
is fulfilled,
2
any integer prime to ra. If the condition 9) m the function 6 (x) will be the first of the series 0(x\
take the
initial
(#), ... to
value x.
it
If the quantity in the bracket is real,
must be
either -f 1 or
1,
-f powers is to be equal to 1. The case 1 gives be rejected, since then x = x". The case JV =
since one of
its
1 is to
and
6 (x)
= x,
which agrees with the condition

x".
9),
B)
It, remains to consider the case x
since m = 2. We have then
= 0.
=+
The lower sign must be

a
It follows that
taken, and accordingly

2,
afi
+d=
pr
1.
-/^ _
1)
If
now
O m (x)
= x,
=F
(m
1)]
we have
207
that
is,
we must have already had
that, as
m increases,
O (x)
And, again, approaches the limiting value

0(x)
x.
it is
clear
We
have shown therefore that

a
d
-f-
= 2 COS
uL
ad-fr=+l
where k
that
initial
is
(x) shall
prime to m, are the sufficient and necessary conditions be the first of the functions 0\x} which takes the
value x.
For
m=2
the second condition
is
not required.
179.
For the second example we take
for 0(x)
any integral
divisible
rational function of
x with
For every integral x. For if 0(x)
constant coefficients. m the

difference
(x)
is
by
0(x)x= (xz )(xz,)

l
(xz
2
v ),
then for every z a 0(z a )
=za
and consequently
(z a)
0*(z a )
(z a)
= za
Moreover
and consequently
ga
a? a?
o\x v )-z v = x z
where
of
0,
is
since
it is
a rational integral function of symmetric in the roots z l9 za

...
x and
,
.
of the coefficients
.
,z v
If
now we
take
k successively equal to 0, 1, 2, equations, we have as asserted
1,
and add the resulting
where
is
a rational integral function of x.
From
this equation
it
follows that for every root of
Q(x)
=Q
we have
m
(x)
= x,
and conversely that every root of
which
is
not contained
among
1 ,
z2
z v also makes Q(x) vanish.
Every root c of Q(x)
therefore gives
and consequently
also
so that 0(),
and likewise
is
(c),
(|), ...
are
4: ,
all
roots of Q(x)
0(r) ==~z a
.
= 0.
Again since
different
from the z s, 0().==

is
and
Theorem
of degree
v
,
IX.
If 6(x)
a rational integral function of x

(v
then the roots of the equation of degree
l)m
can be arranged as in Theorem I, 5). 1 rows of 5) contains each of the v

*,(*),
180.
(*),...
if
If
is
a prime number then
roots
(*)
[r(f)
= f].
[
Conversely
the equation f(x)

.
= Q has fhe roots

l
X,
= ^(X
),
= 0*(X \
Xm _, =ff"- ( X^
every one of these roots
will also satisfy the equation
*(aj)aj = 0,
but no one of them will satisfy
o(x)
= Q]
consequently f(x)
is
a divisor of the quotient
The
tial.
restriction that e(x) shall

if
be an integral function
is
unessen
For
(x) is fractional
where g and
}
gr 2
are integral functions, then in

N
0( ^
_
flf 2
(b) ^(Ô^fe)
flfefcm-l)
the denominator, being a symmetric function of the roots of 0, is a rational function of the coefficients of /(#); and the /(a;)
THE ABELIAN EQUATIONS. second factor of the numerator, being symmetric in x ly x 2

is
,
. . .
209
xm _
is
integral function of XQ Consequently 0(x ) rational integral function of o; which can be reduced to
rational
the
(m
fch
l)
We
174
degree by the aid of f(xQ ) have therefore
= 0.
Theorem X.
is
Every polynomial of the equations treated in a factor of an expression
where O^x)
is
an
integral function of the

if
(m
,
th
~L)
degree.
For example,
by the
0*(x)
we
take O
-f-
bx -f- c
linear transformation y
6
=x
5 2
-f-
a to the
form
we may reduce this = x 2 -j- a. Then

1
x-=(0
(x)
+ (3a
+ * + (3a+lK + (2a + lX + (a + 2a + a + 1) -f 3a + l)x* + (a + 2a +

*)[*
3
l)a;
].
The discriminant
of the second factor on the right
is
If
now we
take
the second factor breaks
up
into two,
effected.
and
this is the only
way
in
which such a reduction can be
We
have then
[8x
3
[8^
or, for
fc
+ 4(1 + A?X + 4(l-Â;X

2/i
2(9
2(9
x
2
+ 2^ + ^)0;
A
(1
7fc
k2
-f
1,
==
2,
In this way we obtain the general criterion for distinguishing those equations of the third degree the roots of which can be expressed
by x, form
0(x), 0*(x).
In the
first
place
must
[be reducible
to
the
=x
14
210
that
is.
b
2
4c must be of the form 4
(x
-f
-f 2)
(2/1
-f I) -f 7
then to every
there correspond two equations of the required type

(A*
+
It appears at once, however, that
(A -j- 1),
(A
+ 2/ + 1) = + 2/ + 3/ + i) = 0.
-f
/
2
is
unchanged
and that the

substitution.
first
equation
is
if / is replaced by converted into the second by
this
same
It is sufficient therefore to retain
only one
of the two equations.

181.
We
introduce
now
the following
DEFINITION.
tions of
If
all the roots
of an equation are rational func
a single one among them,
Ô (x,\0,(x \...0
l
l
tl
^ (x \
l l
then, if these rational relations are such that in every case
M00*,)
the equation is called
= VaO*l),
that, if the roots of
an "Abelian Equation" *
an equation are
We
have already seen (173)

5),
defined by
the resolvents
),
<p*
?1 = (1, Ofa)
= (li
0iO* 2 )
^
Vv
(1
i(*)
the coefficients of which are an equation 6) of degree further that this equation is solvable noted We known. rationally
satisfy
only under special conditions.

present case.
These conditions are realized in the
We
proceed to prove
Theorem
ally.
XI.
Abelian equations are solvable algebraic
**
first
~
In the
#1(^1),
.
. .
place
we
observe that since
c\
is
symmetric
in
1 ,
#i"
<
1
î), every symmetric function of these quantities
is
rational in
and the
?2
= X, +
coefficients of 1).
If
now we
consider
+ ^(X
x,
2)
+.;
..+ 0- -\**\
and assume that
= Ôr,),
we have
* C. Jordan Trait6 etc. 402. ** Abel: Oeuvres completes, I, No.

:
XT;
p. 114-140.
211
For from
ft
l
0.
(x)=
#.//,(>)
follows also
(f
l
^(jc) = OÔ^x)-] =
O.[O
x) ]
O*(x)
The equation ?> R(<F i) shows that 6) has in the present case also the property that its roots are all rational functions of a single one.
We
Then
write
now
accordingly
Pa
= #.(>"l)
0(* a )
= 1, 2,
2
V)
(.r a )
M) + OOJipfa)) + ^(^(a-O) + .,
so that the operations
is
( <
>
are again commutative, like the
s.
The
therefore itself an Abelian equation, the degree of equation 6) which is reduced to the ??i th part of that of 1 ).
We can
182.
then proceed further
in the
same way
174.
until
we
arrive at
equations of the type treated and solved in
The character
of a
group of an Abelian equation

s
is
readily determined as follows:
Suppose that any two substitutions an arbitrary element jc by

l
and
of the
group replace
*
respectively.
= *,(!),
and
= ^(a;
x by
l
),
Then
st
ts replace
respectively.
But
since
by assumption
st
#.A(.r,)
= OA(&i)t
it
follows
that
= ts.
All the substitutions of the group are therefore commutative. If conversely the group G of an equation f(x) consists of
commutative substitutions, we consider first the case where G is intransitive and f(jc) is accordingly reducible ( 150). Suppose
that
where
If
/i(.r ),/*(*),
are rationally
roots of fi(x)
known
irreducible functions.
we consider the
=0
alone, every rationally
known
212
function of these
is
unchanged by the group

l
G
,
and conversely.
l
obtain the group G belonging to f (x)=Q by of G those simply dropping from all the substitutions s s 2 s 3 elements which are not roots of f (x) =0, and retaining among the
Accordingly we
resulting substitutions ^1,^2,^3,... those which are distinct.

clear that
It is
from
s asp
= Sps a
follows also
ff a ff
B ==ff B ff a-
every irreducible factor of f(x) is therefore itself mutative substitutions, and is moreover transitive.
The group of composed of com
We may
a?!
therefore confine ourselves to the case where

If
is
transitive group.
now G
contains a substitution s
2
,
t ,
which leaves
any substitu-
unchanged and replaces jc. by a- 3 then if we we have of G which replaces x by x tution

s.,
}
select
s,s2
(X.X2 .;..),
s 2 s,
(a-,0-3
).
This being inconsistent with the commutative property of the group,

it
follows that every substitution of

is
either affects all the elements,
or
If
now any
1
arbitrary root
of f(oc)
is
regarded as known,
the group of the equation reduces to those substitutions which Conse leave oc unchanged, i. e., to the identical substitution.
quently every function of the roots

particular
is
then rationally known; in
x2
x^ .
r2
>
(aO,
Again,
if
sa
and
s ft
replace
= ^(a i), byx a = a (x

>
and
Xfi
= 0p(x )
l
respectively, then
s a sp
,
and
s^s a replaces
x by
l
OO.OB(XI)
and
9fftJ(Xi) 9
and, as s a sp =ys^s a
the operations
are also commutative.
Theorem XII.
The substitutions of the group of an Abe-
lian equation are all commutative. Conversely, if the substitutions of the group of an equation are all commutative, the irreducible
factors of the equation are all Abelian equations.

183.
The
group of an Abelian
fulfill
equation,,
as well as the relations between the roots, therefore
the con
ditions assigned in the investigations of
137-139.
In particular we can arrange the roots in the following system

0*rt**0**
. . .
^>Oi)
njWoWj,
.
.
(h
.
nk
= 0, 1, 2, = n,
... n,
1),

in which every root occurs once
n^
,
213
n.2
n k are such
that every one of
and only once. The numbers them is equal to or is con
tained in the preceding one,, and that they are the smallest numbers
for which
Of. (or,)
= x,
is
<Vt(*i)
= *i,... <V**0*i) = x,
:
respectively.
There
only one substitution of the group of the

sa
,
equation which converts x^ into O a (x ). Denoting this by <can arrange the substitutions of the group also in a system
we
s^V
*
.
. .
8fc
( hi
.
= 0,
.
1, 2,
...
1).
n^ntfiz
nk
= n,
once,
where again every substitution occurs once and only

responding
to the properties of the O
s,
and cor
8^ = 1,
2
,
. .
sa
"
= l,
...** = !.
as those for the
s.
The numbers n^ n n k are the same To form a resolvent we take now

.
1>2,
3,..
.fife
and construct the

/i(*i)
cyclical function
= IV ifo) +
Wj is
iWi(*i)
+ "iXV
,0*,)
+ î"!- ^- V
1
i(a;i)]"
where
a primitive w, th root of unity.

of the equation.
Then ^(x,)
is
unchanged
by the group
group
For the substitutions

.
of the sub
GI=
i^,s ,,
:
.s k \
sl
leave 4 $^\} unchanged, and the powers of
convert
v i(î)
into
Vi(î),
0iVi(i),...^-V,(aj),
l
.
Conse respectively, so that these do not affect the value of / the Abelian of the coefficients l8 in rational equa given quently x\
tions
and
in
o>.
From Theorem
IV, 4
is
}
.
therefore a root of a
With c\ all the func "simplest Abelian equation" of degree n tions which belong to the subgroup 6r 2 of G are also rationally
known.
Again,
if
we
take
214
cyclical resolvent
Zi
t
and form the
*(*i)=
(*i) 4-
[Vi
,,(*,)
+ "A fi
w._,
&Vi
11
,.C*-i)
*""
V- V
(^)]\
is
in
which
is
a primitive
ft/
,
root of unity, the function /,
.,(JL$
unchanged by the group G 2 and is therefore a rational function of vV For the substitutions of the group
G,=
leave ^
i ?
\s3 ,8t,
.s k \
s.2
unchanged and the powers of
convert
into
Applying Theorem IV again, we obtain </ 2 from ^ the of a second simplest Abelian equation of degree solution by
respectively.
i
.
w2
In general,
if
we
write
>
...**
the value of
tion
Jf2( ...v
is
determined from that of the similar func
...
by the aid of a simplest Abelian equation, as denned by Theo*
rem IV.
By
a continued repetition of this process
we
obtain finally *
Theorem
XIII.
If the n roots of an Abelian equation are
defined by the system
iX*=
0A(*,)
?ij
2
(ht
. .
=- 0, 1, 2, ... ni
.
1)
n wa
?ZA-
w,
solution of the equation can be effected by solving successively k "simplest" Abelian equations of degrees
//ie
W M w 2 ,w 3
*L. Kronecker: Berl. Ber., Nachtrag
z.
nk
Dezemberheft,
1877; pp. 846-851.
215
184. The solution of irreducible Abelian equations can also be accomplished by another method, to which we now turn our
attention.
Theorem XIV. The solution of an Reducible Abelian where p p. equation of degree n = pftp,** are the differ.
ly
2l
ent prime factors of n, can be reduced to that of k irreducible lian equations of degrees p^pf*,
. . .
Abe
based on the consideration of the properties of a the group of the equation. For simplicity we take \pf*. Since the order of the group is r n, the order of every one of the substitutions is a factor of w, and is therefore of the form
The proof
* is
n=p
p*p"*.
6
Every substitution of the group can accordingly be con

its
power, (which is of order of order p/2 ). Consequently we Pj i) (p^)" power (which can obtain every substitution of the group G by combining all the
(p 2
)
structed by a combination of
a2
th
and
its
is
/ fc
l5
/ t/
/ 25 C
3?--
i Z
t-
>
the orders of which are a power of
l ,
with
all
the
the orders of which are a multiple of p 2 Since the f s are mutative, the substitutions of G are, then, all of the form
.
all
com
/j/
j.t
j.f
(t a t
(n
j.rf
j.rr
pt
.) (t
s t
$.
.).
The order
of the product in the first parenthesis
is
a power of
l ,
and therefore a
factor of pî.
t
For we have
f ft
j^t f^... =(t a t
...)^=:l.
Two
substitutions
r
(t a t
p...)(t",t" f ...\
t b (t a
...}(t" d t" e ...}
are different unless the corresponding parentheses are equal each to

each.
For
(t
if
t
the two substitutions are equal,

l
ft
we have
l
...)
(t a
t\ ...)
left
= (t"
t"
.)- (t" d t".

is
a
2
-,
.),
and since the order of the

that of the right sors is 1.
* O.
hand member
a divisor of
a divisor of
p^, and
hand member
each of these divi
Jordan; Traits
etc.
405-407.
216
The number
group
(
of substitutions s
t
since the substitutions

is
is equal to n = p^pf*. And form a group and every substitution of this

1 ,
43).
the order of the group itself must be Pi*" Similarly the order of the group formed by the t" a is
of order
p^,
m equal to p 2 2
It follows
then from
l
n=p *pf*=p m ip m
l
that
m = aj m = a
1
,
Suppose now that y

the t"
is
a function belonging to the group of
s. Then <p has p,*! values, and is the root of an equation of degree p^ the group of which is isormorphic with the group of the fa. This is therefore an Abelian equation.
Similarly the function
belonging to the group of the
s is
root of an Abelian equation of degree

If
now
<p
p^. and ^ have been determined, the function

*=
V+0V
roots,

is
1.
Every function of the
and
in particu
lar the roots themselves, are rational functions of 7,
and the theorem
proved.
185.
Theorem XV.
p
a
The solution of an irreducible Abe

to that of a series of contain which only substitutions of
lian equation of degree
can be reduced
Abelian
order
equations the groups of
p and the identical substitution. If G is the group of such an equation, the stitution of G is a power of p. Suppose that
order of the substitutions of G.
of orders not exceeding p^
p*- is
order of every sub the maximum
Then those
substitutions which are
and
of G. form a subgroup For if are two of these substitutions, then from the commutative
,
property
it
follows that
K~
l
(t l
t,Y
= ^- t^- = i,
l
tl
so that
t^
is
also of order not exceeding
p*~
l
.
If the
will take
pa
~a
.
group ~a
degree
since rp
pa
is
is of order p any function 
will take only
values,
contained in H.
The
substitutions
among the values
of
<p
217
are produced by the substitutions of G, and which form a
group isomorphic
to G, are therefore all of order p.

it
From
the iso
morphism
of the two groups

<p
follows, as in the preceding Section,
an Abelian equation. If <p is known the group G of the given Abelian equation reduces to H. We denote the group composed of those substitutions of ~ which are of order p A 2 or less by If ^ is a function belong lt is to then determined from ing by an Abelian equation of ~a i, the group of which again contains only substitutions degree p
that the equation for
is
>
l ,
cr-j
rt
of order p, and so on.

186.
Theorem XVI.
a
,
Abelian equation of degree

substitutions of order
p p and the
The solution of an irreducible the group of which contains only

identical substitution, reduces to
that of a irreducible Abelian equations of degree p.
Although
obtained in
last
this
Theorem is contained as a special case in that 183, we will again verify it by the aid of the method
employed. &i be any substitution of the group Again equation; then the order of s is p.
Let
G
if
of the given Abelian

s 2 is
any substitution
1 ,
S 2 s the of Grnot contained among the powers of s n then since SiS 2 2 It will s 2 [ contains at the most p substitutions. group | Si
H=
contain exactly this number,

that a
if
the equality s^ s^
= a,b =
fi.
But
if
s/ s/ = s^s./,
then
sl
~a
= sfs/ = s/~
requires
,
and
for
every value of ra such that

It follows
ft
b different
from
we can determine
(
number
m( ft
then that
s
b) E= 1
~~ l
)
mod.
"l
(
p).
= s m (P
=s
a~a
which
a b and a contrary to hypothesis. Accordingly ft not contained G of substitution s is a If a > 2, suppose that 3
is
.
among
the
substitution sfsf.
\
t ,
,
Since 8^3
= 8^
and
3
s 2 ss
=Ss
s
2>
s s 2 s 3 f contains at the most p the group s And it contains exactly this number, for if Si s.fs^
c ss v~
H=
b
s.2
substitutions.
= sf&fsj,
then
=s
a
1
~a
~P,
and so
on, as before.
all
<JT
Proceeding in this way, we perceive that can be written in the form
218
li
where every substitution occurs once and only once (cf. 183). now we take for the resolvents and the corresponding groups
then every resolvent depends on an Abelian equation of degree p~
The
of
roots of the given equation of degree

V-21
-
pa
are rational functions
Vn
<fa,
for the function
*=A?i+A?i+ ...+0r.
belongs to the group l(c/.
187.
177)
The p a
roots of such an equation
may be denoted by
Suppose that x^ ,$,...
Then
the root by which s/is 2 & the substitution

$a
is
s a^
replaces;
by virtue of the left hand form a by which s^s.^ __sj replaces
will replace x,
#&,$>,... a
$a
-
l<2
û^
.
r a by the root from the right

.
hand form
this
root
.
.
is
#+,,., + &...
a
+f a
Consequently every
.
. .
substitution sfi s.J*
sj
replaces any element x^ ^

,
<a
by
that
is
the substitutions of the group are denned analytically by the
formula
,
. .
.
za
-\-
ka
(mod. p).
The group of an Abelian equation of degree p a the substitutions of which are all of order p, consists of the arithmetic substitutions
,
of degree
188.
pa
(mod. p).
Finally
we
effect the transition
from the investigations
of the present Chapter to the
more
special questions of the preced
ing one.
2r
Let n be any arbitrary integer and
let
the quotient
be denoted
by
a.
Then, as
is
well known, the
n quantities

cos a
,
219
cos 2a
cos
3a
cos na
satisfy an equation, the
coefficients of
which are rational numbers,
-. ^_ 4lMff +A
If
fe;=5)^-._;^ =a
now we
write
jc
= cos a
then for every integer
m
if
cos
ma
we
ti(cos
a)
where
cosmâ
is
is
a rational integral function.

1
Similarly
the value
denoted by
(co.9a),
obtain,
by replacing a by w,a in
the last equation, the result
cos(m Wjfl)
Again
if
in the equatioL
= ^(cosrnâ) = 06 = eos(w (cosa)

1
(cos a)
a) the
argument a
is
replaced by
ma,
the result
cos (m
l
is
ma)
(cos
ma)
0$ cos a.
them
Consequently the roots of C) are so connected that every one of is a rational function of a single one among them, or, and that
1
0(;r)
= 00! (x)
(x
cos a)
The equation C)
is
therefore an Abelian equation.
Accordingly
= cos a = cos
can be algebraically obtained.
189.
n
181.
-)- 1.
We
is
have here an example of
Suppose now that n
an odd prime number, n
= 2v
Then the
roots of the equation C) are the following:
Since the last root
is
equal to
1,
the equation C)
is
divisible
by x
1.
The other
roots coincide in pairs
2m-
cos (2v 4-1
m)2-
coefficients, the roots of
Consequently we can obtain from (J) an equation with rational which will be the following
cos
This equation
is
of the form
220
With the notation

2cos 7:
Zv
-f-
*
1
= cos a = x
we have then
cos ^ uV
2m- = = cos m a, 6(x)

j~
J.
so that the equation
d) has
also the roots
that
is
cos a,
If
cos ma,
cosm 2 a,
cosm 3 a,
2v
-f-
cosmâ,
then the
now g
is
any primitive root (mod.

2
1)
terms of
the series
RI)
cosa,
cosga,
coso: a,
cosg
v ~~
a,
are distinct,
For from the equation

a cosg a
= cosgaa
gf
(a
>
er,
<
v)
it
would follow that

a
=
2;:
-f-
or,
replacing a by
its
value
1
>
1)
= ^(2.+ 1).
g^,
Dividing both sides of this equation by
and multiplying by
$-&
-j-
1,
we
obtain the congruence
this congruence But, since since 2(a /9) < 2v, a sequently cosg a is different from cosgaa.
is
impossible.
Con
Again
v cosg a
= cos a.
-\-
For
since #
2l"~"
1
is
= (0"
l)(g
1)
-f-
is
a multiple of 2v-(-l, one

so that
of the
two factors
divisible
by
2v
relation holds

cos g v a
It follows
221
= cos [
J2,),
1 -f k (2v
v
-f-
1)]
= cos
-\-
2k~)
cos a.
then that the
roots of the equation C\) are all contained
in the series
or again in the series

x,
0(x),
o\x),
Cj)
tf"-
(X),
while
cally.
= 1. (x) We have
v
The equation
. . .
can therefore be solved algebrai

cir
If
we have
an example of 174. = n^ n 2 n^ it appears that we can divide the

,
cumference of a
circle in 2>
, ,
-|. .
1
.
equal parts by the solution of w

.
w w are prime to If n^ n 2 w equations of degrees n-^ n 2 each other, the coefficients of these equations are rational numbers.
,
,
. . .
In particular
if
2W
we have
the theorem on the construction
of regular polygons
by the aid
of the ruler
and compass.
CHAPTER
XII.
EQUATIONS WITH RATIONAL RELATIONS BETWEEN THREE ROOTS.

190.
cases.
The method employed

will
in
188
is
also applicable to other
suppose group G are obtained by combination of the two substitu and s.,, which satisfy the conditions 1) that the equation tions a = s/ holds only when both sides are equal to identity, and 2) î that s a s, .S/S.J (Cf. If, then, the orders of s and s.> are HI 37).
transitive
,Sj
1
We
for
example that all the substitutions of a
and
n.2
all
are represented, each once
and
only once by
8^sf
Suppose now
that
(fe,=
is
c
()
0,l,2, ...n,
1).
construct a resolvent
=
</
the group of an equation /(x) belonging to the group 1, Sj Sf,

,
= 0. We
"
.
s/
and denote the functions which proceed from 2 Then of s.j, ,. a/ by ,, c\,,
s., 2
* ,
.
</
on the application
all
^_
these
</
belong
to the
same group with
</
. ()
For from
.s^ = s/k,
we have
from which
it
follows at once that the powers of

of G.
|V o
*,
form a
self
con
jugate subgroup
The
\
resolvent
a%\>
/
is
+ w^ + w
+
ff
,
"V
V$s2
"*
therefore unchanged by every
.s,
and since
permutes the
cyclically, %
remains unchanged by
V i,
all
the substitutions of the group,

th
and can be
rationally expressed in terms of the coefficients of /(#).

</
,
.
We
can therefore obtain
by the extraction of an
itself
n.,
root,
as in the preceding Chapter.
The group
of the equation then re
duces to the powers of Abelian equation.

191. Again, three substitutions
if
s,
,
sn
and the equation
becomes
a simplest
a transitive group consists of combinations of

s2
,
sa
for
which
1)
the equations
KATIONAL RELATIONS BETWEEN THREE ROOTS.
228
rare satisfied
Identity,
and
only when both sides of each equation are equal to 2) the relations hold
then
all
the substitutions of the group are represented, each once
and only once by
^V
(^ = 0,1,2, ...n,l),
.
where
wô,
/?
;5
are the orders of s 1? So,s 3
If
now G
is
the group
of an equation,
we can show by
precisely the same method as before,
that the equation can be solved algebraically.
Obviously we can proceed further in the same direction.
groups
actually arise in this
those treated in the last

p. 89,
where
192.
s.2
s
l
That which are not contained amonoway Chapter is apparent from the example on
3
.Sj
*,
Returning
it
to the
example of
If
190,
s.2
we examine more
to be replaced
s.,~ s 1 s s
l
closely the
its
group there given.
we suppose
by
. l
reciprocal,
sl
follows from the second condition that
=sk
From
46.
we can
We
therefore obtain every possible *., by the method of have only to write under every cycle of Sj a cycle of sf
order,
of the
same
and
to
determine the substitution which replaces
every element of the upper line by the element immediately below it. This substitution will be one of the possible s a s.
We
or
consider separately the two cases 1) where

cycles,
consists of
two
more
and
2)
where
s1
has only one cycle.

x._,.
In the former case the transitivity of the group is secured by Consequently every cycle of s,* must contain some elements different
from those of the cycle of

ments.
s1
l
under which
it
is
written.
It is clear
also that all the cycles of s
must contain the same number of
ele
Otherwise the elements of the cycles of the same order

intransitivity.
would furnish a system of
The order
of the cycles
can then obviously be so taken that the elements of the second cycle stand under those of the first, those of the third under those of the
second, and so on, so that with a proper notation the following order
of correspondence
is
obtained
224
s,
= (x.x-2
) )
(y,y 2
ys
k
8i
=(yiyi+*yi + &.
(ziz l+k z l
It follows
then that
The group is being aj n x

2
,
therefore non-primitive, the systems of non-primitivity 2 U .,, The substitutions Sj a leave 2/ 1? 2/ 2

.
the several systems unchanged, the substitutions s
a
l
s.2
permute the
systems cyclically one step,
s s.? two steps, and so on. Accordingly every substitution of the group except identity affects every element.
}
The group is, in fact, a group fl ( 129). The adjunction of any arbitrary element x^ reduces the group
to
Consequently all the roots are rational functions of any one among them. The following may serve as an example:
sl
=
^
193.
In the second case, where

is
s l consists of
a single cycle,
the transitivity
ter VIII,
already secured.
81
We may write
a
Sj
then, as in
Chap
=
s
\Z
-j-
1
|
-f-
a
|
To
construct the s 2
we proceed
as before
. .
and obtain from
s,
= (x,x
==
x,
.),
)
sl
(^^; + fr^,- + 2fc
the series of substitutions

kf*
1
1
Now,
ments.
in the first place,
it is
easily
shown that the group contains

all
substitutions different from identity, which do not affect

l
the ele
For among the powers of s there is certainly one sf which has a sequence of two elements in common with s. Then sfs^
2
.
does not affect

Again,
tity
it
all
the elements.
is
can be shown that there
which leaves the elements^ unchanged.
no substitution except iden For we have
z
rC
_L
RATIONAL RELATIONS BETWEEN THREE ROOTS.
225
and
if
,T A
and
a,-
were
not affected by the
substitution
we
should have
and consequently
**=!.
The
substitution then becomes
sfs.f
l
= \z
-\-
and since x^ and a\ + are unchanged, a_^0, and the substitution

is
identity.
The following
is
an example of
this type:
From
the preceding considerations

I.
we deduce
is
Theorem
defined in
of,
// the group of an equation
of the kind
190, all the roots of the equation are rational functions
at the most, two
among
them,
and
the equation is solvable alge
braically.
194.
We
turn
now
to the
converse problem and consider
those irreducible equations, the roots of which are rational func

tions of
two among them:

x*
- ? Oi
3
# 2 ),
a*
= ? (#i
4
^2),
...=
<? n (x\
x2 ).
x^
If any substitution of the group G of such an equation leaves and a? 2 unchanged, it must leave every element unchanged..
if
Again,
sa
and
s a
are any two substitutions of

a;,
which have the

or,
same
effect
on both
changed: consequently
s a sa
and x 2 ~
l
then
s a sa s a
leaves
a
.
and x 2 un
= 1,
5
and
=s
Suppose now that the substitutions of

S
l
are
S 2 5 S3
Sr
There are n(n 1) different possible ways of replacing x and x2 xn If any one of these ways is from the n elements x lt x 2l not represented in the line above, let t 2 be any substitution which
l .
.
produces the new arrangement. 15
Then the
substitutions of the line
226
x and x 2 by pairs of elements which are all different from one another, and none of which correspond to the first line.
will replace
l
is still < n(n 1), then there are other pairs of elements which do not correspond to either line. If f3 is any substitution which replaces ,r, and x., by one of these pairs, we can construct a
If 2r
third line
and so on
until all the
n(n
1)
possibilities are exhausted.
We
have therefore
Theorem II. The order of the group of an irreducible equation of the n degree, all the roots of which are rational func tions of two among them, is a divisor of n(n 1).
t}>
of the preceding Section are not yet with those 190. identified, however, previously considered in The alternating group of four This will be clear from an~example.
195.
The equations
elements contains no substitution except identity which leaves two elements unchanged. For such a substitution could only be a trans
Consequently the roots of position of the two remaining elements. the corresponding equation of the fourth degree are all rational
functions of any two
among them.
But the group cannot be writ
ten in the form
iV*,
for
it
(;r,# 2 )
?i]W.j
contains only substitutions of the two types (x x.2 x 3 ) and aJ so that the orders n and ?i 2 can only be 2 and 3, while (a? 3 4 ),
l
t
must be equal
196.
to 12.
If however, the
degree of the equation of
194
is
prime number
p,
we have
precisely the case treated at the begin
ning of the Chapter.
To show
this
we observe
that by
Theorem
is
II,
the transitive group
(JT of the equation is of an order which Since the transitive group is of degree p,
a divisor of
p(p
p
th
1).
its
order
s,
is
also a multi
ple of p.
It contains, therefore, a substitution
of the
order,
and
consequently also a
subgroup of the
same
order.
If
now
_
in
128,
Theorem
I,
we take
1,
and put
for r the order

it
227
follows that k
= 0,
that
is,
contains no substitutions of order

l
Consequently we must have except the powers of .s,. 188. and this is the assumption made in
s,_,,svsy
s/",
Equations of this kind were first considered *by Galois,* and have been called Galois equations. We do not however employ this
designation, in order to avoid confusion with the Galois resolvent equations, /. e., those resolvent equations of which every root is a
rational function of every other one.
If a substitution of the
group
x.,
of an equation of the pres
ent type 193
is
to leave
any element
unchanged, we must have from
Since k&
1 is either -i:0 (mod. p) or is

;r
prime
to p,
is
it
follows that
1
,
either every
is
unchanged and the substitution

most remains unchanged.
III.
equal to
or
one element
at the
Theorem
of prime degree p group of the equation contains, besides the identical substitution, p I substitutions of order p and substitutions which affect p 1
elements.
If all the roots of an irreducible equation are rational functions of two among tliem, the
The solution of the equation reduces
to
that of two
Abelian equations.
The simplest example of the equations of this type 197. furnished by the binomial equation of prime degree p xp
in the case
is
A=
domain of those quantities which
is
where the
real
th
root of the absolute value of the real
quantity
does not belong to the
we regard as rational. The roots of this equation,

a?!,
if
JL\
one of them, are

(w*
wxn
"Ar,,
.u*-^
= l).
is
The quotient
of
any two roots of the equation
therefore a power
of the primitive
th
root of unity w.
to
A
xa
properly chosen power of

if
this quotient is equal
itself.
Consequently
is
any two roots
.^ and
.Journal
,T
are given, every other root
defined by an equation
11 of
in Vol. Evariste Galois: Oeuvres mathematiques, edited by Liouville

tie
the
matliematiques pares ct appliquoes,
184G.
pp. 381-
I.
228
<A
y
.
that
is,
xa
is
a rational function of Xp
it is
and x y
As
soon, therefore, as
shown that the equation
is
irreducible,
If the
it is
clear that
it
belongs to the type under discussion.
p polynomial x
were factorable
then, since
of the
is
a prime number, the several factors could only

if
be
same degree,
be rational.
they were
all
of degree
1.
The
is
roots
would
then
all
Consequently this possibility

<PI(X)
is
Suppose then that
be rejected. of higher degree than Let 2 (x).

to
<f.
the roots of &i(x) and v<(x) be
Then the
last coefficient of
*
each of the polynomials
tr,
and
= ^*
(
>3 l>
and consequently
their quotient
wv/",
is
(in
>
0)
rational within the rational domain.
Since
is
a prime
number r
it is
possible to find an integer
//
such that the congruence
w#=l
or the equation
my.
p -\-
shall be satisfied.
(
Then the quantity

T
.r"
u>
Y=
,
X VP + M* T =
I
x ^T
i
,r
From the reducibility of the and consequently x is rational. follow the rationality of a root, which is equation would therefore
certainly impossible.
For if we leave The group of the equation is of order p(p 1). one root x unchanged, any other root MX, can still be converted w " ^,. ^x 1 roots tax lt u?x into any one of the p
l
229
Theorem
A
IV.
The binomial equation
xp
th
A = 0,
in which is not the p power of any quantity belonging to the rational domain, belongs to the type of 196. Its group is of order
p(p
1).
198. REMARK. By Theorem III every irreducible equation the roots of which are rational functions of two among them is alge At present we.have not the means of proving the braically solvable.
converse theorem.
It will
however be shown
in the following
Chapter
by algebraic considerations, and again
at a later period in the treat
ment
of solvable equations by the aid of the theory of groups, that

irreducible and algebrai an equation of the type above considered, or Before we pass to such general considera
is
every aquation of prime degree, which

cally solvable,
is
either
an Abelian equation. tions, we treat first another

relations
special case, characterized
by rational
among
the roots taken three by three.
199.
An
xy
equation
is
said to be of triad character, or
it
is
if its roots can be arranged in tri that a such any two elements of a triad deter Xp, way mine the third element rationally, i. e., if x a and xp determine y Xp and x y determine x ai and x y and x a determine xp.
called briefly a triad equation,*
ads
xa
in
-jc
Thus the equations
of the third degree are triad equations; for

#1
+ #2 + #3 =
<?,
Of the equations
of the higher degree, those of the seventh degree
may
be of triad character.
,
In this case the following distribution

is
of the roots x^
Xi,
x.,,
x.,
possible
x^
,.r
, ,
X3
If the
Xt, XT
x J" 4 X-; X 3 X5 X 6
{ , ,
,j*
X x, x
2
,
( t
x.>,
.-,.
<";,
degree of an equation
is n,
there are
pairs of roots
.
x^xp.
With every one of these pairs belongs a third root xy such triad occurs three times, according as we take for the Every
original pair of roots

fore
-
x a ,Xp; x$,x y
and since
p. 89.
or
x y ,x a
There are there

integer,
it
triads,
this
number must be an
*Noether: Math. Ann. XV,
230
is
follows that the triad character
= ftm
-f-
The
case u
Qm must
6w 1 or only possible when n be excluded, because n must

if
be an odd number, as appears at once
we combine
n
x^ with all the
other elements, which must then group themselves in pairs.
The general question furnishes a triad system
whether every
= fan
-f-
1,
= 6w +3
however
we do
not here consider.
It is
easy to establish processes for deducing from a triad system of n elements a second triad system of 2n -f- 1 elements, and from two
triad systems of n
r
and
n.2
From
the existence of the triad character for
elements a third system of H, n., elements. n 3 follows therefore
that for
n = 7,
15, 31,
all
9, 19, 39,
21, 43, ...
These do not
however exhaust
systems for n
200.
tioned.
.r,, j- 2
,
. .
possible cases.
There are for example triad
13, etc.
We
In the
.
proceed to develop the two processes above men first place suppose a triad system of n elements
.r,,
given.
,,.
To
these
we add
/
w+1
1 \
other elements
(t
x\, x
<>,
ments, and
retain the lti(n O also construct from these

.r
We
1) triads of
the former ele-
^~
a? s.
new
triads
by accent
ing in each case every two of the three

further triads
,r
,
Finally
we form n
in all
a;,, ar ,;
() ,
x
.
and have then
4w(n
1)
(2w
+ l)2n
to the
"T"
triads,
~6~
2n
-f-
which furnish the system belonging

3.
1 elements.
For example, suppose n tem
We
x
x
,
obtain then the following sys
xxx
,-
x
J
x
**
.A
<
,
ro
r "_>;
T *-0
*t
*1
which agrees, apart from the mere notation, with the triad system
for seven elements established in the preceding Section.
201.
be given.
The
Again, suppose two triad systems of degrees i^ and indices of the first system we denote by a b c
, ,
j/.,
to
.
those of the second by
,/?,? ,..
We may
designate a triad
by
the corresponding indices. system are
Suppose that the triads of the
first
231
a,6,c;
a,d,e;
6,d,#;
...
and those of the second

^2)
/
,r;
,#,
*,*/;
...
We
and form
denote the elements of the combined system by xaa x a p,x ba for these a triad system as follows. In the first
,
.
place,
we
write after every index of T,} the index
In
this
way
there arise
~g~
index of
~ triads of
elements with double indices.

ft
In the same way we write

T,).
then r
then
d,
after every
in all
We
obtain then in every case
^^l and
,
^(n, 1) ~~6~~
triads of the elements
x ((a x aft xay

, ,
x ba xbft
,
...
All of these
are different from one another.

f
They
are
aa,
/
6,
&
/
C a;
c/S;
aa, da, ea;

a/S,
rf/9,
6a, da,
6/3,
#;...
/
3
>
s
,
e/9;
d/5
Again, we write every index of the system Z ,) before every index of T2 ), and obtain
7
triads
among
the same
n^
elements with double indices.
These
are also different from one another
and from those of
jT
3)
They are
a a, a/3
6 a
.
a-f\
a a, a 5, a=;
6a
,
aa, aT, a>j;...

6a
,
6 /5
6^
65,
?
;
~
,
6 //
Finally
we combine every
triad of T,) with every triad of 7 a )
by
writing after the three indices of a triad of T^ the three indices of a triad of T2 ). With any two given triads this can be done in six For * ^ we have ways. example from 6 d g and
, ,
,
6</, r/r,
(/>>
6,
dfh g*
6r, da,
b
f
t
,
grr/,
/>r,
dr/? /a;
f
6//5 da,
t/Z]
dr,
gra.
232
obtain therefore from
ft
We
T^ and T )
2
^(M!
1)
n2 (n.2
6
1)
6
anch combinations.
H,W,-
n^
n,
Br
n2
+1
and from those

,
of
T
cf
;
These are again all different from one another 3 ) and T" 9 ). They are the following:
6 /5
6Y
ft
a /5
;...
ay
a;
6/5.
"
)
a.
a d/5, e^;
a,
a,
dy,
e/5;
a/5,
d,
e;
c?/5,
ea;
We have
,j
therefore
2
now
constructed in
all
--
*H~"
among
Wl
different triads
the elements
The
three tables
T)
3
therefore form a possible triad system for
HÎ.,
elements.
202.
is
The
triad
group
for
=8
demands no
special notice.
It
simply the symmetric group of the three elements. To determine the group of the triad equation for n
=7
we pro
ceed as follows, restricting ourselves to irreducible equations of this

type.
group of 7 elements is transi by 7, and it consequently con th tains a circular substitution of the 7 order, which we may assume
this restriction the resulting
tive.
With
Its order is therefore divisible
to
be
s,
= (#iXs#3#
4ic6a?6
c7 ).
conversely the arrangements of the 7 elements in These must be triads, which are not disturbed by the powers of s,
.
We
determine
now
such that
Xp -\tion
i,
if
x a x$ x y form
,
a triad, the
6).
same
is
true for every
x a -{-
x y H- i
=1 (i
Again with a proper choice of nota
X a = x l9 Xft = x29 since a proper power of s, will If now we apply the contain the two elements x a x in succession.
we may
of s
x>
}
take
ft
powers
to the system
a?
aj
x x x
x.
xn x
x>>
./

it
233
rise to a
appears that only the second and the fourth cases give
triad distribution of the required character, viz.

TI)
a?,
,
a? 2 a? 7
a; 2
,T 4
.TO
a? 2
a? 3
X5
a?3
x, ,
x4 x&
,
#4 r5
,
<r 7
.r,
.r,
,r,
X$
a? 7
Xi
T^
xu
^6
?
,x 6
5
a?,,x 3 ,x 7
^7
)
oââ*,;
.r 4
a? 5
x.,;
rt
x,
,r 3
^7
^4
*^1
^5 i
The two
distributions are not essentially different, each being ob

,
tained from the other by interchanging x 2 ,r 7 jr 3 r and .r,. .-,. may therefore assume that T,) is given, and that s, belongs
;
,
(i
We
to the corresponding group.
If there are other substitutions of the
7 th order belonging to the group, a proper these will contain a^ and aV in succession.
stitution therefore
power
of every one of
We may
r> (i
write the sub
(x^xâ ^x^x^.jc^) = (12

To
triad systems,
a4 a a a
7)
this substitution correspond, as in the case of s
t ,
only tw o
.
which proceed respectively from 1, 2, 4 and 1, 2 a 6 The indices ,, 6 must be so taken that the new systems coin cide with Tj). In this way we obtain seven new substitutions s.
. . .
:
For example,
1, 2,
if
the seven triads

7
;
a6
2, Og,
a3 a 4
,
1;
a4 a 5
,
2;
a3
(i
a,;
fi
7 ,
</.
a7
1,
a5
are to coincide respectively with
1,2,4;
2,3,5;
a-,
3,7,
8,
1;
7,6,2;
7
,
0,4,8;
4,5,7;
5,1,6,
we must have
ingly
$9
S*-
a4
6).
rj
= 6,
4*^*7
-^
a6
= 4,
^4 ~
= 5,
and accord
= (x XyX^x x^x x
1
Jt
Similarly we
~
;>
obtain for the seven

;J"^*5 /,
new
ft
- (XlXQ
JCfyXjQufflE rQts&Jm
^S
X^X>)^ ^ ^
^X^X.^X^X-jX^X^X^fj
Sg
(XiX^XrflCfflC^X^X^J^
St
Beside the powers of

substitutions of
the
it
further proof, that
s2 s 8 there can obviously be no other We note, without order in the group. follows from this by the aid of 76, Theorem
,Sj ,
,
.
. .
r th
/
XII, that the required group
is
{s,,s2 ...,}
,
The same result has been obtained by Kronecker from an

different point of view.
entirely
234
V. The roots of the most general irreducible triad equation of the 1" degree can be arranged as follows:
Theorem
group of the equation

defined by
z
is
the
Kronecker group of order 168 r

aO(z
.
az
6,
(<r=l,2,4;
+ b\, \z c = 0, 1, .,
Those of
+ b) +
=
<
6;
its
0()
-a^ + l))
/
.T
,
/s
.TJ
doubly
bi/
transitive.
-r
:!
substitutions which replace
are
a7,so
replace
*
x by x
s
Consequently we have also

^1
~ "l\^l? ^8/j
,r 4
),
==
^l(^*:!5
^0/5
similarly
,r,
= ^(xj,
x,
>,(.r 4
a*,); eic.
JC Q
/ê substitutions of the

)
group which interchange

(a-o^,)
(
;
and x are
(dr 2 ;r a; 5 a^),
fi
(J^,),
C^
,r o-r,) (.x 4 6 ),
r 2tr 4 x 5 ^c ),
it
(a?^)
and
.st
nce these all leave x s unchanged,
follows that
1
x9
and
the
= ^ (x
l
Qy
x )=^ (xl9 sc9)

l
same property
11
olds for all the other triads.

is
Every sym
metric function of the roots of a triad

203.
1- valued resolvent.
We
examine also the triad equations for w
= 9.
In the
construction of the triads it is easily recognized that there is only one possible system, if we disregard the mere numbering of the ele ments. We can therefore assume the system to be that constructed
in
201,
and designate the elements accordingly by two indices

00,10,20;
00, 01, 02:
each
01,11,21;
10, 11, 12;
02,12,22;
20, 21, 22;
00,11,22:
00, 12, 21;
01,12,20;
01, 10, 22;
02, 10,21;
02, 11, 20;
A
is
characteristic property of every such triad
pq,
the condition
pq
p"q"
235-
P+P +P"^q + q
this
it
+q";
(mod.:-})
From
follows that every substitution

*
p, q
ap + bq -f
a,
ap
+ b q -f
,/
|
transforms the triad system into p", q" become respectively

-fa,
itself.
For the indices p,q
p g
,
a p
+ -fa, ap" + ty" -f

op
&</
,
+ /A/ -f a a> +6V -fa p" + &y -f a"

;
and
if
satisfied
the condition B) is satisfied by by the new indices.
p,p ,p";
</,
,/, ,/
it
is
also
changed can be written

coefficients a, b,
;
,
Conversely, every substitution that leaves the triad system un in the form * by a proper choice of the
triad
For if t, //, which group replaces the index (0,

.
is
any substitution of the

by (,
a
),
0)
then
does the same, and consequently t, the group, leaves (0,0) unchanged. b ), then ( b,
ti
sl
which also belongs

t,
to
If
now
replaces
(0, 1)
by
where a and a are

replace
(0, 1)
arbitrary, will leave

.
(0, 0)
by
6,
Consequently
t,
= Uy
p+q
unchanged, and will will leave both (0, 0)

(r,
.and
(0, 1)
unchanged.
*x
Again
if
A,
replaces (1,0) by
cp, c
),
then
p,q
will leave (0, 0)
and
(0, 1)
l
consequently
t4
= t.^-
r unchanged, and will replace (1, 0) by (c,c ) which belongs to the group of the triad
equation, will leave
(0, 0), (0, 1)
and
(1,
0)
t4
the triad system shows that

ingly that
we must have
t}
tS
.
= 1,
unchanged. A glance at and it follows accord
a.So.Sj
Consequently
further that
tl
is
actually of the assigned form,
llemembering
145 the necessary and sufficient condition that this form shall actually furnish a substitution, we have
established in
we have
the following
286
VI. The group G of the irreducible triad equa tion of degree 9, consists of all the substitutions
SEE;
j
Theorem
j),
ap
ah
-j-
bq -f
dp -f- 6
-}-
(mod. 3)
a b z=:0
(mod. 3)
T/ie
order of
(i
/.s,
/?*o??i
145
2
r
T/ie
= 3^
:
1)(3
3)
= 27- 16
roo/s o/
xi/sff)H,
the equation
are connected, in accordance with the
triad
a* follmvs
>H.r,
1(
#ao
.<,),
2}
V(.r (>1
,r n
),
x,,
V(x02
.r 12 );
All the substitutions of are of the form

::
irliich
replace
x w and x
l(i
by x w and
x.M
P,
</
j"
+ bL + 1
l
.r.,
?y</
mod
3)
an//
,s/>ic
///e^-e
a/Z convert
"*
/>ifo ,r 00
-^
it
follows that
^ oo J
?
we hare also
~
()0
X ldi "^20/J
in
^
(.
^ -20?
All the xnbstitutioHx of the group which interchange
,r 00
and x lo are
of
the
form
*"
Pi
<l
-P Hx.M
b<l
+1
b q
(mod.
3),
since thene all leave
unchanged, we have, again,
204.
The arrangement
in triads given at the
beginning of the
preceding Section possesses a peculiarity, which we can turn to The triad system is so distributed in four lines that the account.
three triads of every line contain all the 9 elements. Evidently every substitution of the group permutes the several
lines as entities stitutions
among
themselves.
We
-j-
determine
If
<>
\
now
those sub
which convert every

p^
(i
line into itself.

"
i
ff^=
is
ap -J- bq -f-
ofp
b q
-j-
(mod. 3)
to convert the first line into itself, the
new value of q must depend

Consequently second
solely
on the old value of
q,
but not on the value of p.

for the
we must have a
The
substitution
ff
= 0.
is
If the substitution is to convert the
line also into itself,
we must again
ap +
same reason have b
= 0.
therefore of the form

t ,
\p q
b q
+a
mod.
3).
237
That conversely
is
all
these substitutions satisfy these two conditions

is
obvious.
Their number
3 .2
2
,
since ab
(mod.
3).
It is further required that * shall also leave the third
lines
unchanged.
( l
-
(P (li
and fourth third line has the that in property every triad P"<]"} the three sums
The
P"
have respectively the values
0, 1,
"
2 (mod. 3); the fourth that
p + q==p
If
+q =p"+q"
2a
(mod.
)
3).
now we apply
(,
;
a to the triad (00, 12, 21
of the fourth line,
we
obtain
+
r
2 b -f
/>
,
+</)
(mod.
3),
and consequently we must have
+
that
is,
+ 26 + +
a
ft
."./.
2a
+ b +a
3).
-f-
(mod.
3),
:&
(mod.
The
rinal
form of
T
ff
is
therefore
Pi<l
</
/
|
Conversely
all
the substitutions of this type convert every one of

itself.
the four lines into
The
substitutions form a subgroup
a can only take the values 1 and 2. 3", self -conjugate subgroup of 0. For if - is any substitution
of order 2
since
H H
of
is
of G,
then
Hr leaves
The group
every line unchanged,

3
2
,
i.
<?.,
H- = H.
of order 2
being a self -con jugate subgroup of

construct,
which
is
of order 2*
3
3
4
3
,
T= G .H
of order 2
= 24
the four lines of triads).
by 86, the quotient 4 of degree (corresponding to This group is of course the symmetric
we can
and
group of 4 elements.
If therefore
we
construct a function
<s
of the 9
elements
a?,
which belongs
is
to the
group H,
this four-valued function is the
root of a general equation of the fourth degree, the group of which

T.
If this equation of the fourth degree has
been solved, the group

3
2
,
of the triad equation reduces to J/, of
order ?
as
is
readily
is
apparent.
The
systematic discussion of this class of questions
however reserved for Chapter XIV.
238
205.
.//
We
consider further the subgroup of

first
which leaves
every single triad of the
line of
our table unchanged.
In
order that
;
/v/
ap-\-a,
the values
a
0,
ct(j
/
I
may have
q
0, 1, 2.
this property,
q
1,
= 0,
1, 2,
must give again
Consequently
and we must take
The
We
subgroup of /of h of order 3. U is of order 8-2 construct by $80 the quotient U = H:I.
r s
form again a
self -conjugate
U is therefore and of degree 3, corresponding to the three triads. If. then, we construct a the symmetric group of three elements. function c of the 9 elements .r, which belongs to the group 7, this
1
(after adjunction
of
cr)
three-valued function depends on a general
equation of the third degree.

If the latter
has been solved, the group of the triad equation
reduces to
I.
Accordingly the symmetric functions of
are known, and therefore these three values

of the third degree, the coefficients of
ible In
depend on an equation
which are rationally express
terms of
since
its
This equation is. in fact, an Abelian equation, group only permutes the roots cyclically. We have then
<.
-.
Theorem
9 can
20f).
VII.
The irreducible triad equations of decree
he solved algebraically.
In close relation to the above stands the following
Theorem
equation of the
.r, ==. "( .r,
VIII.
W
.
If three of the root** of an irreducible degree are connected by the equations

"(>,,
:!
.>,)
,r,),
.r,= ^.r
in
,,r
= 0(.r,, x,) = = ^(.r n

jc,
()(.-*,
2),
,r
;),
which
/.s
a rational function of
its tiro
elements^ then the e<jua-
tion
can be solved algebraically.
It is transitive, and it consider the group of the equation. three others between which the three ;r, j"., j^ roots, by replaces
, ,
We
the same relation must exist
as
;*:
j ,
between
.r
.,.
a^
,r 2
;r
;i
themselves.
.Suppose the
new
roots
to
be
.r
,,
If
the two
systems

...TJ ,
,
23U
u and ,r n # .r have two roots in common, then they .r, have also the third root in common. For, if ,r, = .r .r .r, it
.j
._,,
.,,
.,
follows that
y.,
= o(.r\
.r 3 )
= %,
,*-,)
,r,
and
if
.r
.,
and
,r
:!
are not the
same
root, the
given equation, having

a substitution in
equal
If
roots,
.r 4 is
would be reducible.
a root different
from
.r,
.*_,, ,r :l
there
is
If this substitution leaves no the group which replaces .r, by ,r 4 element unchanged, we obtain an entirely new system .r .r-,, .-,,.
.
But
if
one element, for example

,r.,, ,r 4
,
r L>,
remains unchanged, we have for

in this way,
it
anew system
possible
,r-.
Proceeding
and examining the

all
effects of the
substitutions,
is
seen that
the roots
arrange themselves in the triad system of 9 elements.

this
Comparing
is
result
with Theorem VI,
it
appears that the equation
exactly one of the triad equations just treated.

It is
known *
that the nine points of inflection of a plane curve

lie
of the third order
by threes on straight lines. These lines are twelve in number, and four of them pass through every point of inflection. Any two of the nine points determine a third one. so that
The abscissas the points form a triad system, as considered above. triad equa a or the ordinates of the nine points therefore satisfy
tion of the 9th degree,
and
this
equation, belonging to the type

are the abscissas or
above discussed,
is
algebraically solvable.
It can, in fact,
be shown that
points
of
if
.r,,
/,,
.r
:!
the ordinates* of three

straight line, then
.r,
inflection
lying on the
same
= 0(.r
1?
.n),
,r,
0(oc.,,
o- 3 ),
x.,
= 0(x^
r,),
where
is
a rational
The
discussion of this matter belongs
and symmetric function of its two elements. however to other mathemati

here.
p.
l!)i.
cal theories
and must be omitted

p. 08;
*O. Hesse: Crelle XXVIII,
XXXIV,
Salmon: Crelle
XXXIX.
p. wr>.
CHAPTER
XIII.

207.
treated for which certain relations

specified,
In the last three Chapters various equations have been among the roots were d priori
in consequence admitted the application of the
and which
theory of substitutions.
In general questions of this character, however, a doubt presents itself which, as we have already pointed out, must be disposed of
first of all, if
eral algebraic questions is to
the application of the theory of substitutions to gen be admissible. The theory of substi
tutions deals exclusively with rational functions of
the roots of
equations.
If therefore in the algebraic solution of algebraic
equa
tions irrational functions of the roots occur,
we
enter
gion in which even the idea of a substitution fails. mental question thus raised can of course only be settled by alge braic means; the application to it of the theory of substitutions
upon a re The funda
would beg the question.
To
cite
a single special example, proof
of the impossibility of an algebraic solution of general equations
above the fourth degree can never be obtained from, the theory of
substitutions alone.
208.
first
In the discussion of algebraic questions
it
is
essential
of all to define the territory the quantities lying within
which
are to be regarded as rational.
We adopt the definition* that

coefficients
all
rational functions with integral

91
,
of
certain quantities
(9t, 9ft" 9ft"
,
.
. .
9ft",
9ft
",
constitute
the
mlional domain
).
If
among any
functions of this
domain the operations of addition, subtraction, multiplication, divis and involution to an integral power are performed, the result ing quantities still belong to the same rational domain.
ion,
The
*!,.
extraction of roots on the other

I
hand
will in general lead
Kronecker:
.erl.
Ber. 1879, p. 205
if.; cf.
also: arithni. Theorie d. algeb. Grossen.
THE ALGEBBAIC SOLUTION OF EQUATIONS.

to quantities
241
which
lie
outside the rational domain.
We may
limit
ourselves to the extraction of roots of prime order, since an (mn) th root can be replaced by an m th root of an ?i th root.
All those functions of
9ft
,
9T,
91 ",
which can be obtained

a
from the rational functions of the domain by the extraction of

single root or of
ively, as
any
finite
number
of roots are designated, collect

,
. .
the algebraic functions of the domain (91 91", 9ft ", .). In proceeding from the rational to the algebraic functions of the domain, the first step therefore consists in extracting a root of prime
order p y of a rational, integral or fractional function v ($t 9f, 91 ) which in the domain (9t 9f, 9t ", .) is not a perfect power.
,
. . .
"
p^
Suppose the quantity thus obtained
to
be
so that
We
it
will
now extend
the quantity
91
,
V
v
,
the rational domain by adding or adjoining to so that we have from now on for the rational
, .
.
domain (TV,
functions of
9ft", 9ft"
3ft
,
.), i. e., all
rational, integral or fractional
91", 91 ", ...
are regarded as rational.

one.
The
present domain includes the previous
With
this extension goes
a like extension of the property of reducibility.
Thus the function

it
9ft
xp
F C31
V
91",
reducible and has,

tional factor
.) was originally irreducible: in the extended domain (Vv

;
.
has now become

,
9ft",
.),
the ra
x Vv The new domain can be extended again by

order.
the extraction of a
second root of prime which is not a perfect (p v ^-i) th power within denote its (p v _ ) th root by Fr _ n so that
1
We
construct any rational function
(Vv ;$l
9ft",
.,
and
It is not essential
here that
should occur in
F ,_
l
If
now
we
adjoin
we obtain the further extended rational domain V _ 15

,
(V V _ 19
TV,
91
91",
.).
Similarly
we
construct
Ff
10
^(F
2?
9T,
9T,
.)>
242
where the
theses,
F B denote rational
,
.
. .
functions of the quantities in paren
and Pi,p 2
Pv
are prime numbers.
algebraic expression can therefore be represented in with the preceding scheme, by treating it in the same conformity in the which calculation of such an expression involving only way numerical quantities is accomplished.
Any given
209.
The
.Fa s are readily
integral in the corresponding
V
,
reduced to a form in which they are r T a +n Fa+2 "P",, s, that is

. .
.
are fractional only in the 91
9T
Thus, suppose that
where
G
,
G,,
Cr 2
H ,H H
1?
Vv
9t
91", ...
If
now &
is
are rational in a+2 2 a+3 ,... a primitive (p aiH ) lh root of unity, the
,
.
product
is
Va + Fa+3
2
,
9t",
For on the
it
one hand the product

integral
is
rational in the
H
TT-
s,
and on the other
is
and symmetric
-[7-7-0
in the roots of
Tfi
a+J
(rr * a + 2i -ô + lV
vi
J A.
Sft" Jl ,
\
. .
.;
and
is
therefore rational in the coefficient lâ +

if
2
of this equation.
Again,
V a+l +
we omit from the product P) the factor the resulting product

. .
.
-\-H
Va +
.]
is
integral in
to
Va+l
and
rational in
Va+2l
9t
9?",
More
it
over, since
does not occur in P) or in the omitted factor,
does
not occur in Pj).

a by Pj), multiply numerator and denominator of the resulting denominator is a rational function of a +2, v $ 91", alone, while the numerator is rational in these quan*
If
now we
. .
.
THE ALGEBRAIC SOLUTION

titles
l
.
Otf
EQUATIONS.
243
and in Va + Dividing the several terms of the numerator by the denominator, we have for the reduced form of F*
where the
"Fa+2?
coefficients
J Jn J2
,
are
all
Jt
31", ...
On
account of the equations

-â+l â +
KT
rr
we may assume
power of
that the reduced form of

lh i)
.
than the (p a+1 The several coefficients J can

l
Va +
contains no higher
in the
now be reduced
Fa
same way as
above.
By
multiplying numerator and denominator of their
fractional forms
by proper
factors, all the
can be converted into
of a degree not exceeding p _ 2 1, a4 with coefficients which are rational in 9ft v a+3 ... 9T, ...
2
integral functions of
Va +
and
In
this
way we can continue

2
1
to the end.
0.
We
theorem which
will
have now at the outset to establish a preliminary bo of repeated application in the investigation of
I. // //,, .fp ,,- p are functions within a domain, the simultaneous existence of the two equa
. .
the algebraic form peculiar to the roots of solvable equations.*
Theorem
definite rational
tions
B)
iv*
=0,
same rational
requires either that one of the roots of B) belongs to the
domain with
/ ,/,,
,/],
fp _
l \
t\ or that
= 0, /o
If all
/,
= <),... /,_, =
0,
().
the/
>fp-~\
are not equal to

iu
the equations
A) and
B) have
at least
one root
common.
In the greatest common
divisor of the polynomials
A) and B)
the coefficient of the highest
power
of
is
unity,
from the form of B).
Suppose the greatest
common
C)
divisor to be
<f o
-f <f,w -f
*
cr
wr
-f
wv
Equated
to 0, this furnishes
roots of B).
If
one of these
11,
is
lii>.
denoivruu-
"This theorem w n originally iîveii by And: Ojuvies conipieit-.s eeker was the first to establish it iu the lull importance: JJurl. lier. 187<J,
p. 200.
244
a>,
ted bj tTj, and a primitive p ih root of unity by rojts of C) can be expressed by
then
all
the
Apart from
its
algebraic sign,
cr
is
the product of these roots
Now
since
t\
u and
for
is a prime number, which
it is
possible to find two
numbers
and consequently
One root,
/*&<,,
of the equation
5) therefore belongs
to the
given
rational domain. 211.
We
apply Theorem I
first to
the further reduction of
If
JK
is
anyone
a
of the coefficients </n J.2
which doos not vanish.
we determine
î)
new quantity
TFa +
x
W+
a
by the equation
^FJ + Ô,
F^,
annex
to this the equation of definition for
and
R)
iix
for the rational
domain
.,.
;
(^4.i.;F.+i,F. +t>
that, since the possibility
,-...),
It follows then, if
A) and 5) of Theorem
x ),
W +i
a
;
I are replaced
0,
JK
is
by A,) and excluded, we must
have
CO
where
to
-F
is
+l
{
= ^(â +1
P.+ 2
r,, ;
r,^".
...
a (p a +
il1
root of unity.
/ a in
_
We
place of
can therefore introduce into the expression for
the
root
Va+1
<",
the function
provided we
adjoin the (p a
,)
th
)
of unity,
to the rational domain.

2
,
From
A,) and C
.,,
it is
.
clear that
9i",
.
OP. +
1 ;
Fa +
...;
K .K",...)and(Fa41 ,F a +
...: ?R
define the
same
rational domain,
and the equation
THE ALGEBRAIC SOLUTION OF EQUATIONS. can be taken in the scheme of

of definition for
245
208 in place of BJ.

.
The equations
Va Va _
,
l ,
are not essentially affected by this
We have only to substitute in the functions a a _^ F change. The expression for in the place of a +i the value taken from C\). a then becomes simplified
, . .
.
F F
We may suppose this reduction to have been effected in the case of every Fa We pass now to the investigation of the form of the roots 212.
.
of algebraically solvable equations.

1)
6riven
an algebraic equation
/(aO =
of degree n, the requirement that this shall be algebraically solv from the rational domain able can be stated in the following terms:
($
$",...)?
which includes
at least the coefficients of 1),
we
are to
arrive at the roots of 1)
by a finite
number
of algebraic operations,
viz. addition, subtraction, multiplication, division,
raising to powers,
and extraction of roots of prime
orders.
One
of
the roots of
1)
can therefore be exhibited by the following scheme:
where
G ,G lt G
of ,... are integral functions

",
.
V F
t
,
and ra
211).
tional functions of 9T,
and
(?,
may
be assumed to be
1(
and reducing in every case those powTaking the powers of XQ th we obtain for every v ers of F, above the ( Pl l)
,
XJ
If these
= cff + o^ v, + 0w vt+.
+ ^,;L
F?
substituted in powers of # are
1),
we have
A)
where the
cients of
H
1).
Joining
s and the coeffi are formed additivily from the ^" of definition of the with
r
A)
equation
246
-THEORY OF SUBSTITUTIONS.
*)
F^-/^^, F
to
is
...):=0,
B),
and applying Theorem I

ties:
91
,
A) and
either a root of
B)
rational in the
we have only two possibili domain (F2 F3 V

, ,
.
Ut",
...),<*
Both cases actually occur.
In the former the scheme
2),
we passed from the
original rational
domain
to the root
by which x9t can
be simplified by merely suppressing the equation
Vfi
= Fi(V ,V ,...),
a
and adding the pf* root

213.
of unity to the rational domain.
As an example
f(x)
of this case
we may
26
0,
take the equation of
the third degree
=x
3ax
the rational domain being formed from the coefficients a and
b.
By
Cardan
formula
;r
+ |/ b*û*
|/6*_ tt 8.
This algebraic expression can be arranged schematically as follows:
The expression
becomes
for J/(aJ
formed as
in the preceding Section, then
J/Oi-o)
olT,
+ F
(
a)
-f
F F, = 0.
2 2
Comparing
this with
VV
and determining Vj from the
(6 last
T = 0,
a)
two equations, we obtain
v =
so that Fj
Ii
is
we now
already contained in the rational domain into an integral function of transform
(F2
F^; a,
6).
by the pro-
ceos of
20U,
we
obtain from the relations
THE ALGEEKAIO SOLUTION OF EQUATIONS.
247
2 2
F )a>F
3
2 3 2
<WF
2 2
[a
= 26(6+F )(a+F F -aF F +(6-f [a )F ][26(6 + )(a + F )]=[26(6-f- F )] [a(6 + F -a F + (b- F F [26(6 + F (a + F )] = 4a6 (6+F )F
3 2 8 2 2
+ (6 + F KF -aa>F
3 2
]
2
2
),
3)
3)
3)
where w
is
a primitive cube root of unity, the simpler form
F
Kemoving
_+F
)F _ aF
2 2
the fraction by 6
from the denominator by multiplying both terms of Fa, we have finally
and herewith the reduced form of x
,,=>,+
214.
^5
7,
F, can therefore be suppressed in the scheme above.
We
return
now
In
to the results of
212 and examine the
second possible case.
suppose that V is not rational in the domain Then from Theorem I

l
F F
2
,
<R",
. .
.).
If now, in analogy to 3),
we form
the expressions
.
**=
which w
l
6?
+ Gj^F, + (W*F!+
212, that
t^-y
(^
it
= 0,],...^
1)
in
is
a primitive pj th root of unity,
follows, with the
same
notation and process as in
f(xk )
Since the
to 0,
i.
= H + HM V,
is
I.
H
xk
e.,
is
vanish identically, the latter expression a root of f(x) for k 1, 2, .pi
also equal
For example, ia the case
of the equations of the third degree
248
where the
first
of the
two
possibilities
above has been excluded by
reducing XQ
to the
form
the other two roots are
r
215.
If
now we make
6?!=
its
1,
(whereupon
o;
V may
l
the allowable assumption ( 211) that possibly take a new form different from
l
original one),
a?,
.
.
we
.
tions for
x_
obtain by linear combination of the p equa

l
the value of F,
jpi-i
The
irrational function
Fj of the coefficients
is
therefore a rational
.
. .
and, in fact, a linear function of the roots

as the primitive Pi th root of unity w^
is
ar ,cci,
a^-i
as soon
adjoined to the rational
domain.
210.
In the construction of the scheme

1?
2) it is not
2
,
. . .
intended
to assert that
Fa necessarily contains Fa _ Fa _
another arrangement of 2)
is
If
Va _
is
missing in
the order
Fai
possible;
we can
replace
by the order
It is therefore possible, for example, that different
occur at the
end of the
series 2).
In this case different constructions 3) for the
THE ALGEBRAIC SOLUTION* OF EQUATIONS.

root
249
xQ
tion holds for the last
are possible, and the theorem proved in the preceding Sec of 2) in every case.
To prove the same theorem

last,
for all
VB
(
but in the next to the
last place in 2),

a
.
that
actually contains
The proof
which occur, not in the we will simply assume 215) of the theorem for
was based on the
fact that
an expression
satisfied
same property
an equation with rational for an expression
coefficients.
We
demonstrate the
If
1
)
to
we suppose all the permutations of the roots of the equation be performed on
the product of the resulting expressions is an integral function of //, with coefficients which are symmetric in the x*s and are therefore
If
9ft
,
3ft",
we denote
this function
by
<p(y),
the coefficients of
the
equation
of the roots, with possi belong to the domain (9ft 3ft", .), and one an unessential of the modification 211) is bly meaning of V2 (cf.
,
. .
It is therefore essential that
should actually occur in

the root
l
We
as
is
can now apply to <f(y) we applied above to f(x)

allowable, that the series
is
with the root y the same process which
= with V ,V _
v
If
we assume,
,...V
,V
is so chosen that V.2
not rational in the preceding "F
s, it
follows that
where w 2
is
a primitive p./ h root of unity.
Every y k
is
produced
16a
250
,r
,
from y by certain substitutions among the

quently
V.,
xn
a:,
1 ;
conse
is
a rational integral function of the p, th degree of the

0,
1
roots of f(x)
provided the quantities w to the rational domain.
and
>.
are adjoined
In the same way every

the series
V can be treated
which occurs
in the next
to the last but not in the last place in 2).
Proceeding upward in
we have
finally.
Theorem
isfies
II.
a solvable equation f(x)
The explicit algebraic function x which sat = 0, can be expressed as a rational

,
integral function of quantities
V V V8) v
It
*
11
V v
vi
with
$t
,
coefficients
which are rational functions of the quantities

the one
9ft".
The quantities V\ are on
functions of the roots of the equation

roots of unity,
hand rational integral f(x)=Q and of primitive,
and on
the other
hand they are determined by a

TT y
.
series of equations
TTpa y
a
*-
T?(TT * a
\
1 5
TT v
a.
9fr l
<fcf>
5
"*-
In
l ,
these equations the _pi,p 2 >_P3j

2
,
.
>Pv
are prime numbers,
and
F F
are rational integral functions of their elements

the quantities ^R
,
and rational functions of

ine the rational domain.
217. of
9ft",
ivhich determ
This theorem ensures the possibility of the application
the theory of substitutions to investigation of the solution of It furnishes further the proof of the fundamental prop equations.
osition
:
Theorem
For
if
III.
The general equations of degree higher than
the fourth are not algebraically solvable.
the
quantities
1 ,
x.2
xn
which in the case of the
general equation are independent of one another, could be algebra then the first introduced ically expressed in terms of 9t 9t",
,
. .
irrational function of the coefficients,
9ft
SR",
Since,
would be the p v th root of from Theorem II, Vv is a
rational function of the roots,

tion of
.Tj
,
it
appears that
r
as a
pv
valued func
symmetric, is either the square root of the discriminant, or differs from the latter only by
x.2
,
.
.
.
XH
the
pv
th
pow er of
which
is

a symmetric
If
factor.
251
Consequently we
must have p v
we
adjoin the function
latter
then includes
If
all
v \/ J to the rational domain, the the one-valued arid two-valued functions of

l
V =S
56).
the roots.
essary
if is
we
2,
i)
are to proceed further with the solution, as
is
nec
>
there must be a rational function

valued,
Fv "_, of
power
4
(
the roots,
is
which
ued.
(2/v_
and of which the(p

if
th
I/
._
two- val
But such
a function does not exist
>
58).
Conse
quently the process, which should have led to the roots, cannot be continued further. The general equation of a degree above the fourth therefore cannot be algebraically solved.
218.
We
**
return
now
to the
form of the roots of solvable alge
braic equations
3)
-Go+Vi + G, Ff +
+G
itl
Vp ~
h
,
(1
l .
We
adjoin to the rational
unity,
domain the primitive jV P and assume that the scheme which leads to ,r
\
is
roots of
reduced as
far as possible, so that for instance
Va
S
is
not already contained in

.
the rational
domain (Va _
TV,
<R
<R",
w 1? w 2
).
We
have seen that the substitution of
for in 3) produces again a root of f(x) the generalized theorem:
= 0. We proceed to prove
which leads
io the
Theorem
sion 3) for
Tâ-i,
,
IV. If in
is
l
-
the
scheme
2),
expres
XQ any Va
i,
.
multiplied by any root of unity, the values
V
.
a.
V V
2J
will in general be converted into neiv
quan
If the latter are substituted in the place the in the the result is (ujaur a root of former expression for x
tities r a
r a _,,
.r.2
.v
of f(x)
= 0.
may, without
loss of generality,
;
We
assume that f(x)

(
is
irredu
cible in the
domain
(9i
9i",
).
Starting
now from
construct
3),
arid
denoting by
"T
a primitive r th root
of unity,
we
(x
,r A
=
A
=-
[x
- (G.
-I-
<^ V, -1- G,
m* V? +..-)]
T
Jn
this
product
certainly vanishes,
Possibly other
vanish
252
also.
THEOBY OF SUBSTITUTIONS.
Suppose that
(a>.2)
is
the lowest
V that
actually occurs.
Then
The
which occur here belong
to the
domain
Fa
,,...).
We
construct further
where
PI -
V
i
(6>.a-|-l)
is
again the lowest
that actually occurs.
Similarly let
in,FM-i,.^
and
finally,
supposing the series to end
at this point.
c(x;
V , ve+
c
S
=/, (#; 91
;r
9t",
where f
with
rl
is
rational in
,R
^)l",
all
the T r+1
disappearing
V We
assume now, reserving the proof for the moment, that the
functions
/(;
F.,
.),
/6 (ar;
.),
/ (ar;
c
.),
^(.r;
ft
.)
are irreducible in the domains
(F.,F. +1
respectively.
...),
(F6 ,F, +1
1
...),
.
(Fr F
,
+ 1? ...), (* ,".....)*
Then
. .
S
;
/^(a
.)
=
in
and f(x)
have in the do
main
as a
(JR
9t",
.)
one root x
= x^
common
factor of
fd (x) and
it
f(x).
x occurs common, since x Both these functions being
by
assumption irreducible,
follows that
If
V
now we
PV
V
1
assign to
.
any arbitrary value

1
i- v
consistent with
V = F (W, F^" = ^-1(^5

. .
and
R
;
to
)>
Vv _ any value t^_, consistent with and continue in this way, we have the
seres
253
Hh
SW +
being obtained from 3) by putting the
-r a
in place of the l"
The product
=
will only differ
from those obtained above by the introduction
of the
)
andrrs in place of the

2).
Gs
this
arid
F s,
since in
is
all
the reductions 2
(
.
replaces
Consequently
product
equal
to/,(jr; r ,,r,,. M ,.
and similarly
This furnishes the proof that I,, is a root of /(#) = 0. We have still to prove the irreducibility of /,,(>), /fr(t ... in r the rational domains ("F T^ + ], ...), (F/,, T + ,,...)..., respect
).
f/ , /)
ively.
Assuming the irreducibility of f (x) in the domain Vn F we proceed to demonstrate that of fb (x) in the domain ( V,,. Vh The method employed applies in general.
tt
If
<f(x\
y
it
.)
is
one of the irreducible factors of
fb(x), so
chosen that
the domain
8)
contains
,
/
. .
ff
(a?;
F, M
.)
as a factor, then
we have
in
Va V
+
,
, ,
the equation
...).
0(-r;
v(x;Vb
.)=/(*; F.,
in the
),
which can be rewritten
form
254
where the /
B)
it

F;;
V
V
n
tl
j_
.
,
Since moreover
^
all
(FW +
the /
if
,...)
= 0,
is
follows from
is
Theorem
I that either
s
which
to
be excluded, or
still
.
+ vanish, so that A) and with

n
1 ,
.
rational in
V
M
it
.
.
the equation 8) above
hold,
n is
replaced by M U
.
~
(
ft ,
Again, /,(*
write
<"V
f<-,
is
different
from/,(r; wF,,,
.).
For
if
we
/(*; F.,
it
.)
= + *,F,, + s*V,? +
*o
would follow from the equality of the two functions / that
A,)
fiK*-r-/) F, +
,(.. .
.*)Fa
...
=0,
and consequently, from the equation of

#,)
definition
.)
f^*<F.+i,
= o,
+
1
,
.
that
since
must be rational
in the
domain (F
(x\
u> n
91
...*,,...),
^=/S.
ffl
Accordingly /
<r.
(a?; F,,,
.),
(l
...),... are all divisors of
All these functions are different from one another,
and they are

tr
all
irreducible in the
domain
(F,,,
F+i,
).
Consequently
con
tains their product, which, on account of the degrees of

.r,
arid/,, in
is
1),
possible only Since the foregoing proof holds for every irreducible factor of it still holds if we drop the assumption of irreducibility.
219.
if
<r
and//, coincide.
At the beginning of the preceding Section we remarked
that in the product construction with
other
Fs
might vanish.
This possibility is however excluded in the case of certain we shall now show.
F
the
s,
as
We
lowing
designate any
of 2) as
i.
r,n
external radical
when
fol
.
do not contain FT FT + 11 FT + 2, radical can such external be the last position of to Every brought ,r of the as in and can be arranged in 2), expression given 8),
T
+ 2 ...
,
e.,
(1
terms of every external radical present. We shall see that in the product construction with F, no other external radical can be
missing.
Thus,
if
is
missing in
255
then
cannot be changed radical expression by w T *V r
if
we
replace
in the
fundamental
a result, the
Gs
are converted into the #
without thereby changing If, as s. we should then have also

l
.
Every linear factor in x of this last expression must therefore be equal to some factor of the preceding expression
Taking into account the -equation of

B)
it
1
definition
follows
Vp F (V,,...) = 0, from Theorem I that either V is rational in

l
V.2 ,V
:i
<U T
which may be excluded, since otherwise ther on the adjunction of W TJ or that
2) could be reduced fur
00=
In some one of these equations
ing this equation according to
g2
=G
T
2,
must actually
occur.
Develop
powers of
we have
and combining with

B,)
this
V^ F (V
r
+ ,...)
l
= n,
the impossibility of both alternatives of Theorem I appears at once. Consequently F T could not have been missing in the product con
struction.
ending with V can also contain external radicals, in fact possibly such as are not These also cannot vanish external in respect to the entire series.
If
we
consider only
/ (218,
4)), the series 2)
tl
in the further product construction.
borne in mind, the proof
is
exactly
The irreducibility of / being the same as the preceding.
product construction of the preceding The Section no external radicals can disappear from / except V
Theorem
is
V.
In
the
same
true for
fh
. .
.
in respect to the external radicals occurring

.
among
V V
v
,
and
so on.
in x or in one of the ex If several external radicals occur all the corresponding expo pressions fa fh ,fc ,...,the product of
,
nents
is
a factor of
n.
256
Theorem
p
is
If an irreducible equation of prime degree algebraically solvable, the solution will contain only one exter
VI.
nal radical.
primitive
The index of the
latter is
equal to p, and if w
As
is
t}<
root of unity, the polynomial of the equation
Theorem
3)
is
VII.
// the algebraic expression

V,
x,
=G +
.
.
+ G, IV +....<?_, V*
0,
a root of an equation f(x)

(9T, SR",
.
luliich
is
irreducible
in the
domain
).
and
if
we construct
the product of the

. .
factors, in
which
is
replaced by w, 1^, ^\V\
where
of -the
Va is the lowest V present, and again the product f (x; V p factors f (x\ u> a*Va1 ...), and so on, we come finally
b 6
.
fl
(t
to
the equation f(x)
= 0.
the degree of
which
is
The functions fn fb
,
are irreducible in the domains
V V
a
,
220.
We
examine now further those radicals which vanish
in
the
first
product construction.
The remaining
V V
a
,
_ .
15
are
We not altered in the* product construction. these to the rational domain, or, in other words,
irreducible equation
may therefore add we may consider an

rational
(F.,F..H,... Here all the

construction.
f(x)=f ,",...)
,
f(
(x;
fl
.)
in the
domain
V V
}
already vanish in the
first
product
We
examine now what
is
the result of assigning to

its
Va
__
any
its
arbitrary value consistent with

this basis assigning
equation of definition, then with
equation of definition,
any arbitrary value of Va _ 2 consistent with and so on. Suppose that the functions
F_i, V*-*,
are thereupon converted into
V F G G
2
,
l5
6?,_i
257
"-iV-ii
.v 2
*>,;
9Q ,g2,
._,.
The new value assumed by
a?
is
then
From
M
2
.
. .
218
.
o>\,
c is again a root, and this together with the system when Uj is replaced by wv lf I^_ 15 which arises from w p -\, gives again the complex of all the roots. We can
therefore take
where
a>
,
o>",
at
"
, .
.
.
are the
th
roots of unity w,
w2 w3
,
apart
from their order.
By
addition of these equations
we
obtain
7o=
so that
(7
(7 is is
GO,
Also p
unaffected by the modifications of a _ lt a_2 a . the sum of all the roots, and is therefore a rational
,
.
.
:
function in the domain
(9t
91",
.).
Again we
obtain from the system above the equation
Here the first term on the right vanishes. theses in the following terms briefly by p^^p^ 2
9)
v,
We
,
denote the paren

> -
Ps,
.
and write
= fl,F + S
s
2
P7 + .0
W+
On
raising this to the
th
power
and annexing the equation of
definition
it
follows from2
,
Theorem
8
I that either F,
a
is
rational in
ffi
F F ,... F
or that
_
1
V 2 ,v3 ,...^_ i;
2
,!R",..
1
F/ = ^
17
A = 0, A = 0, ...^_ = 0.
258
We
v2 v 3l
,
.
.
consider
now
the
first
of these alternatives.
2
,
In the rational
9ft
expression of
.
V
l
in terms of
F F
v2 va
,
91"
... all the
va _ cannot vanish; otherwise

2).
pressed in
If then
we
define
system of successive radicals,
some
s.
F V
If
1? K
should have been sup as in 208 and 212 by a

2
will occur last
among
the v s
and some V\
for
in
a?
last among we have
the u
we
substitute the expression
Here
from
all
the v
the
reason
all
.
But
T s cannot vanish, since we might V and v\ are two external radicals,

K
cannot vanish, as we have just seen. For the same have started out
and the product
of their exponents
must therefore be a
factor of
is
p (Theorem
V).
This being impossible, the first alternative Accordingly we must have in A)
excluded.
The
its
question
th
now
arises
what the form of
9)
power may
take thp form
Vf = A
must be in order that The equation A) is
The
If
result just obtained

l
shows that the
left
member
is
unchanged
is
replaced by w V
l ,
^ V^
Consequently
and on the extraction
of the
lh
root
we have
. .
JVFj-h^XF^-f
But from
9) follows also
=yu
.
Q^
V,
+C
left
0,^+
a>* Vl
and equating the two usual, we have

fl 1
2
members and applying Theorem

_
I as
= 0, ^G = 0,...^
GK _
= Q,
C, +J (? +1
= 0,...,
that
is,
9) reduces to the single
term
V^Q&Vf.

Substituting this result, together with
ft ,
259
gr
=G
in the expression for
we have
On the
other
hand the
root I
which
is
contained
among XQ x
,
l ,
can also be expressed in the form
and, comparing the two right members, it follows from Theorem I that terms with equal or congruent exponents (mod. p) are identical.
In particular we have
and therefore
fl&TV=0X
fi,
Ftt
=
f
*,
10)
ta
"G
lt
K
.
Theorem
^r
VIII.
I/,
m the explicit expression 3)
o/ the root
o/ cm
irreducible equation of prime degree p, the irrationalities
are modified in
any way
consistent with the equations of defini
tion,
then T/Y
221.
is
converted at the same time into (G\V^) J\.

relations arising
The
e in
from such a transformation are

of
p.
most readily discussed by the introduction

ence root
the place of the prime
a primitive congru
number
e,
We
write
,
V,
#^
Fn
V,
= #5;, G Vf ==
to
VW
2
where however every remainder (mod. p 1).
e a is
be reduced to
its least
not negative
Then the
V*,
quantities
...
G*V*,
G^_,F/-
coincide, apart from their order, with the quantities
and we have
The changes
replace
in the values of the radicals, considered above,

?
which
tt
V
x
where
by (G^Vfy and consequently [GaF^] by [G ax F 
x]
(mod.p),
have therefore the effect of replacing every
R a by
260
tf.-M
(
= 0,l,...;p
(mod. p).
2),
where
* is
<p
and
is
defined by the congruence
e*=A
Consequently the quantities
I)
_R
,
R
+
lt
_R 2
-Rp__ 2
are converted in order into
R
and,
if
-R*
-it*
+2
*** +jp
the same operation

RO.KI
-^ttAC
is
performed
-âK + 25
a times, I)
25
is
replaced by
+M
+ -â/c+j;
where the indices are of course
to
be reduced (mod.
1).
If there is another modification of the radicals,
which converts
into
Rp,
this
on being repeated
,3
times converts the series I) into
Finally
if
we apply
the
first
operation
times and the second
ft
times, I)
becomes
Here
and
/?
can be so chosen that

x
ax
-I-,?/;,
gives the greatest

is
common
other
divisor of
is
and
/-/-.
Consequently
if
Rk
the -R of lowest
index which
obtainable from jR by alteration of the radicals, every obtainable from J^ in this way will have for its index a
multiple of k, so that the permutations of the within the systems
Rs
take place only
R Rk R2k
,
,
R*=l_i\ k
Here k
is
a divisor of
1.
There are then alterations

produce the substitution
(RO
222.
in the
meaning
of the radicals
which
R Rk
k
2
(î
Rk +
-R^fr+i
...)
to
The preceding developments enable us
determine the
group of the irreducible solvable equations 1) of prime degree p. Every permutation of the jc s can only be produced by the alter-

ations in the radicals V,,
261
permutations of the
alterations of the "F
a? s
s.
and consequently only such 2 a_ can occur in the group as are produced by
,
.
From
do not change we have

1
V V
the result of the preceding Section
can be converted into
^G^vf,
and the possible
alterations in
this form.
Substituting this in the table of
215,
We
this transformation.
examine now whether any root x^ can remain unchanged in In that case we must have
as a necessary
and from the method which we have repeatedly employed and sufficient condition, that
(a*
ek
it
follows,
=w
j-
p-e
=fj.-\-T
(mod.
p).
If
e*=l
IJL
therefore no root
every
tity.
(mod. p), then for r=^=0 there is no solution //, and But for T x^ which remains unchanged. 0, satisfies the condition, and the substitution reduces to iden
If e* =j= 1
then for every
r there is a single solution
<).
and the
corresponding substitution leaves only one element unchanged.
Theorem
of prime degree
groups.
223.
IX.
is the
The group of a solvable irreducible equation

metacyclical group
(
134) or one of
its
sub
Since now, as
we saw
in
221,
.
all
the substitutions of the
group permute the values Q ,Rk ,R.2k ,.. only among themselves, the symmetric functions of these values are known, and the values themselves are the roots of
_^
The
,
latter is
an
.
Abelian equation since the group permutes the values

only cyclically. function of R Q
R Rk ,R
is
2k ,
Consequently every
.
Jf4, _R2A., jR3A., ...
a rationa
But the same

end of
is
true of every
Ra
For the form
of the substitution at the
221 shows that after the adjunc-
0*
262
tion of
to 1.
all
the other
Ra s
if
are known, since the group reduces
Finally
it
appears that
âthen
+ *n
= Fa(R*.).
For the application of properly chosen substitutions of the group converts the first equation into the second.
now all the substitutions of the group of f(x) = = which leave R Vf unchanged and accordingly can only convert V Then x is replaced by x But since R is a Vi into some rational function of R it appears that R = G p V is also unchanged,
We
consider
u>
ep
so that
G
,
Vi
is
converted into some

;
^ V*.
The power
a>*
can
be determined from x v for the expression for x v contains the term e ve Gr e F w and this must be identical with G^Vf. Consequently
1
fj.
= ve,
and
G Vf
e
becomes
Q.fVft
while at the same time
V* becomes
so that the factor
tion of the group,

also.
That is, every substitu e remains unchanged. which leaves jR unchanged, leaves G e unchanged The same is Accordingly G e is a rational function of R
.
true of
12)
all
the other
(7 s.
We
can therefore write
XQ =
<? 2
,
where
(9ft
,
^3
.
are rational
this
it
functions of
that
Vf
in
in
the domain
appears ^0 n ot admit of multiplying every term by an arbitrary root of unity, as indeed is already evident a priori since otherwise X Q would have not p, but p p values.
9t",
. .
).
From
11) the radicals
V-^QJ ^-Ru.
A still
From
further transformation of 12)
is
possible.
We have
which convert
221 there are alterations in the
V ,V ,...Va _
l
The form of and consequently \f R into w^X/R,,. At the same the exponent of ^ evidently involves no limitation.
into
time the X Q becomes
263
and
since
becomes
Rk +
lt
it
follows that
^/E, becomes
If
now we apply
these transformations to the equations above,
we
obtain
We
can therefore also write
Theorem X.
degree
p can
be written in either of the two

,
.
The roots of a solvable equation of prime forms 12) or 13). In

~k
13) V^K*j tyR-2 k
\^R
The values
U
is
are roots of a simplest Abelian equation, the group of which
com
posed of the powers of
Its roots are connected
by the relations
)
14)
VBt =f(R
G jcf(E
e
V R\ WK
v
2k
=f(Bk )
where
Q ).
224.
The form
. .
E R
,
is
13), together with the relations 14) between not only necessary but also sufficient for a root x of
an irreducible solvable equation of prime degree p.

that
of
ff.
For
14)
shows
all
the possible permutations

If
among
the
,
Es
are simply powers
now
<r
converts _R
into
R ak
then V-ô will become
some w^/v/^, and from 14)

becomes
etc.
264
Consequently 13) becomes
That
is,
x has only the p values
a?
oc lt
xp _
225.
We
will
examine further the Abelian equation of degree
p
k
which
is satisfied
by
R%,
U(,,_,)i..
From
14)
we have
ft,
^(ft)-/"* (ft), ft*
ft, =/"(ft.)
and since
JS W&
= -R
it
follows from these equations that
Now
ep
the primitive congruence root e for p can be so chosen that 1 is divisible by no power of p higher than the first. For if
e*~
then
l
=l
(mod.p
2
(P
e)*-
= e*e)
(p
l)pe*-
~e*- +pe p - E=l+pe*l
so that (p
e in
place
p ~ l is divisible only ~ of e. In e p 1
t
(mod.
),
by
p,
and we can therefore take

then, q
is
l=p.#,
prime
to p.
Consequently we can determine

*g
so that
+ l=0
g = sp
(mod.p).
1.
Suppose
that
Substituting this in equation 15),
and taking the
p th
root of the
ih
power, we have
=*
-[/
(ft)
" -
/*
(ft.) ...]

Again,
13)
if
265
we
write
/ (*..)= a.
root
and take again the p th
we have
Since by equations 16) the a a an Abelian equation, the a a

equation.
s s
are rational functions of the roots of

are themselves roots of an Abelian
The
substitution
a
= (JR Ek E
(a Q
2k
of the former corresponds to the substitution
r=
of the latter.
If the roots
aâ z
.)
a 0l a ly a 2l
...
are
different
from one
another, then
R ak
is
a function of a a
and
this function
is,
in fact, the
same
for all values of
189).
to the
Theorem
form
XI.
The quantities ty~R can be reduced
35=
This
,
.
.
equation.
a B ,_, o/ a simplest Abelian form contains the roots a a u <P is an The form 17) is not only arbitrary function.
.
necessary but also
sufficient.
statement remains to be proved. In the first place the 7^ s, as rational functions of the (distinct) roots of an Abelian equation, are themselves roots of an Abelian
last
The
group
equation with the group 2 m~ r r

l
1,
<r,
**,...
o\
first
which corresponds
to the
1, T,
Again the
two equations of 17) give
so that
we
are brought back to the characteristic equations 14).*
*Cf. Abel: Oeuvres, II pp. 217 ff. (Edition of Sylow and Lie); and Krouecker: Monatsber. d. Berl. Akad., 1853, June 20.
CHAPTER
XIV.
THE GROUP OF AN ALGEBEAIC EQUATION.

226.
We
have already seen in Chapter IX,

f(x) =
is
153 that every

its roots.
special, or affect, equation
completely characterized by
or between
is
a single relation between
its coefficients
Sup
pose that in any particular case the relation
More
accurately speaking, it is not the function <p itself, but the family of <p and the corresponding group G, which characterize the
equation.
Only those substitutions among the roots are permissi which belong to G. For this group we have the fundamental theorem
ble,
:
Theorem
of the equation
I.
Given an equation f(x]

2ft,
and a correspond
ing rational domain
all rational integral

9ft
which are rational within
functions of the roots are unchanged by the
group
of the equation, i. e., they belong to the family of G or to an included family. Conversely, all integral functions of the roots
which are unchanged by
are rational within
9ft.
The
algebraic character of a given equation, for example one

coefficients, is therefore
;
with numerical
by no means determined by the knowledge of the cofficients alone but, as was first indicated by Abel, and then systematically elaborated by Kronecker, the bounda
ries of the rational
domain must
2
also be designated.
The
solution
of the equation x*
example, requires means, according as \f 2 is or is not included in the rational domain. The rational domain can be defined on the one hand by assigning
the elements
9ft
,
= 0,
for
very different
9ft",
from which
it is
constructed.
Or we may
construct the Galois resolvent equation and determine one of its irre The latter does not, to be ducible factors in the rational domain.
sure, entirely replace the
assignment of
9ft
9ft",
but
it
furnishes
THE GROUP OF AN ALGEBRAIC EQUATION.
267
everthing which is of importance from the algebraic standpoint for the equation considered.
The determination
group of the equation
;
of the irreducible factor gives at once thein the n factors

\
(u h a?!
-f-
u x 2 -fiz
-f-
u in x n )
of which the Galois resolvent is composed, we have only to regard the M S as undetermined quantities, and to form the group of the u s. which" permute the factors of the irreducible factor among them
selves.
It
must be always borne
in
mind
that from the algebraic stand
point only those equations have a special character, according to Kronecker an affect, for which the Galois resolvent of the (n!) th
degree
is
factorable.
227.
tion
On
and
its
account of the intimate connection between an equa group, we may carry over the expressions "transitive,""
"primitive" and "non-primitive," "simple" and "compound" from, the group to the equation. Accordingly we shall designate equations as transitive, primitive or non- primitive, simple or compound, when
their groups possess these several properties.
the term "solvable," which
is
Conversely, we apply taken from the theory of equations,,
also to groups, and speak of solvable groups as those whose equations are solvable. Since, however, an infinite number of equations belong to a single group, this usage must be justified by a proof that the
solution of all the equations belonging to a given group
is
furnished
by that of a single one among them.
This proof will be given
presently (Theorem Y). In the first place we attempt to reproduce the properties of the
groups in the form of equivalent algebraical properties of their equa tions. We have already ( 156)
Theorem
equation
228.
II.
If an equation
is irreducible,
its
group
is
transitive; conversely, if the group of

is irreducible.
an equation
is transitive,
the
To determine under what form
the non-primitivity of the
group reappears as a property of the equation, we recur to the treat ment of those irreducible equations one root of which was a rational
function of another.
The equation
of degree
reduced to
equa-
268
tions of degree m, the coefficients of which were rationally expressi ble in terms of the roots of an equation of degree v ( 174).
arrive in the present case at a similar result.
We
G of the equation f(x) = is non-prim then the roots of the itive; equation can be distributed into v sys tems of in roots each
Suppose that the group
such that every substitution of the group which converts one root of any system into a root of another system converts the entire former system into the latter. We take now for a resolvent any
arbitrary symmetric function of
1)
yi all
the roots of the

l2
,
.
first
system
= S(x ni x
, .
. .
.,)
Since G is non-primi converted either into itself
v
and apply
to
all
the substitutions of G.
tive, the entire system x n oj 12 or into one of the other systems.
x lm
is
There are therefore only

,
values
of y
=S(x n x
& (3*211
12
.x ]m),
&Zm)l
2/2 2/3
^225
a? 32
,
.
.
= S(X = S(x
31
X 3m ),
vill ).
yv Consequently y
3)
is
vl
x v2
.x
a root of an equation of degree

K2/)
= 0,
the coefficients of which are unchanged by all the substitutions of If G, and which are therefore, from Theorem I, rationally known.
has been solved, i. e., if all its roots y y z (p (y) y v are known, then all the symmetric functions of every individual sys tem are also known. For each of these functions belongs to the
l , ,
.
same group
as the corresponding y,
expressed in terms of the latter
and can therefore be rationally and of the coefficients of f(x). If

of
we
Xai
,
denote, in particular, the elementary symmetric functions

XOT>
,
. .
a>ll
uy
SM,
then the quantities x al x a2
,
,
S
.
(y a),
.
Sm (y a),
x am
are the roots of the equation
4)
(y a )x
269
4),
Consequently f(x) can be obtained by eliminating y from 3) and and we have
...
Conversely,
if
we
start
from the
last expression, as the result of
eliminating y from 8) and 4), then the group belonging to f(x) is nonFor primitive, if we assume that 3) and 4) are irreducible.
we form
first
a symmetric function of the roots of
4).
This
is
rational in
5) for
all
ya
We
denote
it
by F(y a ).
Again we form the product
O
the roots of
4).
This product
in
coefficients are
symmetric
all
,U-i
is rationally known; for its ,...y ,, and are therefore ration

l
ally expressible in the coefficients of 3).
Accordingly
5)
remains
the substitutions of the group, i. e., every substi unchanged by tution of the group interchanges the linear factors of this product
If therefore F(y a can be only among themselves. expressed in terms of the x"s in only one way, it follows that the group converts the symmetric functions of x a x a2 r a into those of another
)
,
;i
system.
The group
is
therefore non-primitive.
But
if
the roots of
f(x) are different from one another, the assumption in regard to F(y a ) can be realized by 111.
Theorem
which
is
The group of an equation of degree niv, III. obtained by the elimination of y from the two irreducible
equations
3)
?(y)
=y
l
4)
is is
(y)x
+... + S,( y )x>"- >
A iy
A V = Q,
.
S m (y)
non-primitive
229.
non primitive; and conversely every equation, the group of which is the result of such an elimination.
The properties of an equation the group of which is com do not pound present themselves in so apparent a form as in the case We can however of the transitivity or non-primitivity of the group.
replace the problem of the solution of the equation by another equiv alent problem in which the compound or the simple character of the
group has an
easily observed effect
on the equation
itself.
270
For
this
purpose we have only to take in the place of the general
equation
6)
its
/C) =
Galois resolvent equation
F(z}
is
irreducible.
We
have
first
to
examine more closely the
latter equation
and
its
properties.
Given a general equation 6), there is a linear function of the roots of 6), formed with n undetermined parameters
g\
__ n
g,
I
a
all
y,
which has n\ values; so that

.group
the substitutions of the symmetric
belonging
to 6) convert
into n\ different values
= The permutations among the ^ r H produced by G form a new r which we denote F is simply the n\ elements group among by / F has the isomorphic to 6r, and is the group of the equation 7).
,
property that its order is equal to its degree, as appears either from the method of its construction, or from the fact that every I is a rational function of every other one. The equation 7), which is
identical with
therefore requires for

of a single root.
its
complete solution only the determination

equivalent to that of 6). these relations are modified, if we pass
is
The
arises,
solution of 7)
The question
equation
Q\
/
how
from the general equation

is
6) to a special equation. Every special characterized by a single relation between the roots
<t
oc } (x x 25**"w/ <\l5
t
belongs to a group G of the order ?*, then only the substi G can be applied to the roots. For if a sub stitution were admitted which converted y into
If
<?
tutions belonging to
where
is
different
from
<f
then
all
271
would be
The rational domain thus determined rationally known. would be more extensive than that derived from <f. Consequently
9)
would not represent
all
the relations which exist between the
roots.
We
8)
can
now
obtain a resolvent of our special equation in either
of two ways.
Either we proceed from

Cl
l^l-f-2^2-h
6r,
-\-
a nXn->
apply to
4^
all
obtain
M
and
10)
^"25 ^"3
">>
form the resolvent of the r th degree

*,(*)
= (**,) (**,)
the expression
(?-,.)
= 0;
in 7)
or
we proceed from
F(z\ already given
and 7
),
and observe that F(z) becomes reducible on the adjunction of 9) and that f\(z) is one of the irreducible factors. The other fac
tors are, like
l ,
of the r il1 degree.

.
They
differ
from each other
only in the constants a

plying together all tion of the group
j?l (z)
Every one of them is obtained by multi c a which arise from the applica the factors
,
to a single one
=Q
r.
regarded as a
It is
The group of among them. is of degree and the s, group among
order
simply isomorphic to the group

<p.
of degree
n and
differ
order r belonging to
The groups
of all the factors
F^\
of
F(; ) therefore
from one another only
in the particular designation of their elements.
Theorem
9)
IV.
If a special equation f(x)
=Q
is
charac
terized by the family of
VfaiX*,
#n)
= 0,
with a group
of order
n\
r,
then the general Galois resolvent decom-
poses into p
io)
j actors
*i(0=o,
resolvent of the special every one of which can serve as the Galois All the roots of 10) are rational functions of every one equation.
among
to F^) = The transition from f(x) rationally expressed. has its counterpart in the transition from G to the simply isomor
them,
and
in terms of these all the roots of f(x)
can be
phic group Q
129) of f\(;).
272
Since the construction of 10) depends only on the group G, and not on the particular nature of 9), this same resolvent belongs to all equations which are characterized by functions of the same family
with
9).
If
and consequently and with

it
one of these equations has been solved, then x x 2j xn are The is known. therefore Cj solved, equation 10)
l ,
. . .
proof of the theorem stated in
every other equation of this 226:
sort.
We
have then the
Theorem
which belong
either c^
V.
2
Given an equation f(x)
= 0,
the coefficients of
of the roots
any arbitrary rational domain, the adjunction of or cr = 0, where c^ and ^ 2 belong to the same family x n leads, as regards solvability, to the same x lt x
to
2
,
.
.
special equation.
230.
relations
We have treated in earlier Chapters cases where such between the roots either were directly given or were easily
as involved in
recognized
the data.
Frequently,
<J
however,
(x
l
,
the
.
conditions are such that, instead of a
known function,
x u)
being directly designated as adjoined, </ presents itself implicitly as For example, a root of an equation which is regarded as solvable.
problem of the algebraic solution of equations the auxiliary equation is of the simple form
in the
Here y
of f(x)
of
is
regarded as known,
adjoining to
it
i.
e.,
we extend
the rational domain
by which any power belongs

It is a natural step,
every rational function of the roots
to the
domain.
The
actual solution of
the auxiliary equations does not enter into consideration.
when an
irreducible auxiliary equation

</
,
is
roots to the
but all of its regarded as solvable, to adjoin not one of roots 0. These roots are*the different val domain of /(.r)
{
x ) assumes within the rational domain. For ues which 4 (x x 2 to find the auxiliary equation which is satisfied by ^ we apply to all the r substitutions of the group G and obtain, for example, <!
,
,
. . .
tt
distinct values
fe,
11)
*,*,.
^.
and therefore the
coeffi
The symmetric functions

cients of the equation
of these values,
THE GEOUP OF AN ALGEBRAIC EQUATION.

12) are
273
g(*}
= (*-^} (<!>&)
(0-0.)
the rational domain of f(x) 0, and 12) is the required auxiliary equation, the solution of which is regarded as
known within
known.
Now given
it
the equation f(x)
= 0,
.
characterized by the group

to
6r,
or by any function
y(x lt
x2
x n ) belonging
to the
G,
we
adjoin to
all the roots of 12), or,
what amounts
same
thing, a linear
combination of these
m roots = 101 4X
=
m 0m,
where the
are undetermined constants.
The question then
arises,
what the group of f(x)
becomes under the new conditions.
The adjoined family

it is
of functions
was originally that of
<p.
Now
that of
? +7.
= +
01
02
+ *m 0mthat subgroup of
(7,
The group was

is
originally G.
Now
it is
which
also contained in all the
groups
HU
f
Hm
0n
If
02?
0m-
Suppose
that
of these
m groups.
is
Then
all
the greatest common subgroup belongs to the function /.

is
now we apply
0m are
G to the series ^,
new
... </
,,.
the result
0i? 02?
in every case
all
the same series in a
order;
for
the values which
2 quently the series JET,, tion with respect to (7; and

,
produces from
0j.
Conse
Hm
is
is
also reproduced
by transforma
therefore unchanged by transform
ation with respect to G.
We
have then
tained in K.
Again we denote by P the greatest subgroup of G which is con F therefore belongs to ^-f-/, and accordingly char
which belongs
to f(x)
is
acterizes the family
after the adjunction
of all the roots of 12). like K, also commutative with G\ for on transforming T with respect to G, the result must belong to both G and K, and is therefore F itself. F is, then, a self-conjugate sub
9
group of G, and
are also
in fact
to
l ,
is
2
,
common
18
H H
the most comprehensive of those which

.
. .
Hm
If / does not reduce to the identical operation,
is
compound
274
group.
jf (a:)
If
is
is
simple,
is
necessarily identity,
and the group of

1,
i.
reduced by the solution of 12) to

all
e.,
after
the
the roots of /(#) are known; or, in other also. have words, the solution of 12) furnishes that of f(x)
solution of 12)
We
then the following
Theorem VI.
the
Given any arbitrary equation f(x)

to it all the roots
with
group G, if
we adjoin
11)
*1$& .&.
m
th
of an irreducible equation of the m g(^ )=<P 12)

the coefficients of
degree
l
A^~ + ... =0,

x
l ,
which are rational in the rational domain of f(x),

a? 2
, .
.
and
all
the roots rational functions of
x n then G reduces
,
to
the largest self-conjugate
unchanged.
is it
If
to
is
subgroup a simple group,
P of G which
leaves
v i,
4m
is
P= 1.
Only in case
compound
possible by
the solution of
different
reduce the group
a subgroup
an auxiliary equation to from identity, and conse
quently to divide the Galois resolvent equation into non-linear fac

tors.
231.
We
consider these results for a moment.
If the general
n th degree f(x)=Q is given, the corresponding = n\. This group is compound, the only group G is of order r actual self con jugate subgroup being the alternating group (92).
equation of
the
If we^take for a resolvent
fc=Vî
where J denotes, as"usual, the discriminant of /(a?), then the resolv
ent equation becomes
12
= 0,
After adjunction of the two roots the alternating group. irreducible Galois resolvent equation divides the previously ) of into two [conjugate factors degree Jn!, and only such substitu
is
and P
of 12
tions can be applied to the resolvent
as leave
V^
>4
unchanged and therefore belong

the alternating group
is
to the alternating
group.
For n
simple.
If there
is
an m-

Tallied resolvent
4
,
275
its
values
</ 2 ,
4 m are obtained
by the
solution of an equation of the
mth
degree.
On
tho adjunction of
these values, or of
the group of the given equation reduces, by Theorem VI, to the

identical substitution.
for all functions are
other group. no reduction of the degree of the equation to be solved can be of the values of v effected in this way, since if n > 4, the number
The equation /(o?) = is therefore solved; known which belong to the group 1 or to any The investigations of Chapter YI show, however, that
if it
exceeds
if
?i
2,
is
greater than n or equal to

<f>
n.
In the
latter case,
6,
the function
is
\
always symmetric in n
\ >
1 elements, so
is
that we can take directly 4 = x and the resolvent equation tical with the original f(x) = 0. Theorem VII. The general equation of the n
iden
th
degree
(n
>
4)
is solved,
as soon as
any
arbitrary resolvent equation of
degree higher than the second
is solved.
resolvent equations the degree of ichich is
There are, however, no greater than 2 and less
than n.
ri
h
no resolvent equation of the For n = 6 there is 0. degree essentially different from f(x)
Moreover, if ra===6, there
is
distinct resolvent equation of degree 6.
One other
result of our earlier investigations, as reinterpreted
from the present point of view, may be added here:
Theorem
.has
VIII.
The general equation of the
fifth degree
a resolvent equation of the sixth degree.

232.
ceding
return now, from the incidental results of the pre the group of the Section, to Theorem VI, and examine
We
equation
12)
g (<! )
c
=
.
(<!
vV) (^
</
*)
(0
all
J = 0,
the roots
.
ô,
4 m of
which were
adjoined to the equation
The order
that
it is
of the
group of 12)
is
equal to the
of degree of the irreducible equation
most easily found from the fact which
is a root.
We
must therefore apply
to
">
all
276
G.
The
values thus obtained
may
partly coincide.
The number of
distinct values gives the order of the
group
c
-
of 12).
all
Now if we retain
substitutions of
the designations of
#j,
r2
,
.
.
Theorem VI, r includes

leaves
.
x u which
is
all
the
unchanged.
Suppose that
the order of /
Then
the required
if
number
is
v=r:
From
this
we perceive
1, cordingly of order r equation is the same as that of f(x) can be effected in this way.
the group G is simple and r ac that, the order v of the group of every resolvent
= 0,
so that no simplification
We
it
actually obtain the

all
contains
group of 12) by the consideration that and only those substitutions among the a? s which
f(x)
;
do not
alter the nature of

/
= 0.
(
,
If therefore
we apply
to
Tl5V25V3>
Vm
,!,
all
the group required. however necessarily different the elements

c>
(7, the resulting permutations of the </ s form, All the r substitutions thus obtained are not
for all the substitutions of
leave
all
unchanged.
From
is v
this, again,
.
it
follows that the
order of the group

nize that
86,
K of 12)
= r:r
In the same way we recog
K
is
is (1 With the notation of r) fold-isomorphic to G. the quotient of G and F; G:F.
K=
G
Theorem
IX.
// the group
of
f (x)
=
of
is
of order
r,
and contains a self-conjugate reduces to F on the adjunction of

12)
subgroup F
of order /,
and
if
all the roots 11)
fl^)
= 0,
=r: r
.
then the group

the quotient of
K of
the latter equation is of order

is (1
is
G and F and
r)-fold isomorphic
<J>
to G.
By
a proper choice of the resolvent
we can
give the equation
12) a very special character.
choose as a resolvent a function / belonging to the self-con Then / is a root of an equation of degree jugate subgroup / which are rationally expressible in terms of of roots the v r:r all
.
We
anyone among them; for / n ^2? /> The group T, ( 109, Theorem VIII).
for
it is
all
belong to the same group

is
of 12)
therefore a group
fl;
transitive, since g(%)
is
irreducible.
We
have therefore
is
Theorem X.
If the group
of the equation f(x)
277
of order
r,
and contains a
self-conjugate
a^,
subgroup
. . .
F
f
of order r
and
if 7l
is
a function of the roots
o?2 ,
x, belonging
v
to F,
then an irreducible resolvent equation of degree
= r\r
can
be constructed, the roots of which are all rational functions of
single one
among
them,
its
and which
possesses the property that the
adjunction of one of
233.
roots to f(x)
=
=
reduces the group
to F.
XI. // is a maximal self-conjugate then the is a transitive, simple subgroup of G, group of h(x) group. Conversely, if F is not a most extensive self-conjugate sub is compound. group of G, then the group of h(%)
Theorem
We
denote the group of h(/)
by
Its order is
,
= r: r
We assume that G
From Theorem IX
of
contains a self -conjugate subgroup / of order r r G is r -fold fsomorphic to G. From the results
73
to the
v
follows that the subgroup J of G, which corresponds group F is a self-conjugate subgroup of G and is of order
it
,
is,
then, like
F,
self -conjugate
1
subgroup of
G,
is
(
and
con
232),
all
their orders are respectively
and
We
show that F
tained in J.
This follows directly from the construction of
in accordance with which the substitution 1 of
corresponds to
which leave the
series 11) unaltered.
F in G
therefore corresponds to the one substitution 1 of

if
Accordingly
is
compound, then
is
not a maximal self -con jugate subgroup
of G.
The converse theorem

isomorphic groups.
is
similarly proved
from the properties of
In these last investigations we have dealt throughout with the group of the equation, but never with the particular values of the
coefficients.
If therefore
two equations of degree n have the same
group, the reductions of the Theorem

the
coefficients
are entirely independent of

coefficients of h(/) will of
of the equations.
The
course be different in the two cases, but the different equations all have the same group, and every root of any one of these h(%)
equations
is
a rational function of every one of
its roots.
This com
mon
property relative to reduction, which holds also for the further
278
investigations of the present Chapter, is the chief reason for the

collection of all equations belonging to the
same group
into
a family.
234. observe further that with every reduction of the group there goes a decomposition of the Galois resolvent equation, while the equation f(x) need not resolve into factors.
We
Collecting the preceding results
we have
the following
Theorem
compound, and
a
XII.
if
If the group
of an equation f(x}
1
so that every
is
G,
is
G
is
(?a ,...(?,
series of composition belonging to G,
one of the
groups
l ,
G^,
a maximal self-conjugate subgroup of

groups of the
the preceding one, further if the order of the several

series are
1*1
>
^2)
TV>
-L)
then the problem of the solution of f(x) can be reduced. as fol lows. We have to solve in order one equation of each of the de
grees
L
the coefficients of
r*
^-i
which are rational in the rational domain deter

These equations
that all the
mined by
roots of
the solution of the preceding equation.
are irreducible
and
simple,
and of such a character
root
any one of them are expressible rationally in terms of any the same equation. The orders of the groups of the equa of
r
tions are respectively

r,
r r3 rv
r
j
r2
The groups are the quotients
G:G
lt
G,:G2
G,,
G v _,: G
G V :L
was
The equations being
solved, the Galois resolvent equation, ivhich
originally irreducible
and of degree
r
IT
r,
breaks up successively into

r ~
rv
r
r~ ?
2
rl
r
s
factors.
After the
last
operation f(x)
=Q
is
therefore completely
solved

235.
is
279
The composition
of the group
of an equation f(x)
tion into factors.

tion,
therefore reflected in the resolution of the Galois resolvent equa turn our attention for a moment to the ques
We
when
a resolution of the equation f(x)
itself occurs.
It is
readily seen that, in passing from
Ga
to
Ga+l
in the series of
com
position of G, a separation of f(x) into factors can only occur when G a+l does not connect all the elements transitively which are con
nected transitively by
Ga
The
resulting relations are determined
by
71.
Ga
is
non-primitive in respect to
the transitively con
nected elements which
G a+l separates into intransitive systems. 0, suppose now that Starting from G, with an irreducible f(x) G a are transitive, but that G a+l is intransitive, so that G!, 6r 2
by
71
Ga
is
non- primitive.
Then
the
at this
point f(x) separates

6r a+1
,
into as
many
factors as there are systems of intransitivity in

71),
But (again from
all
elements occur in
Ga+1
We
arrange, then, the substitutions of
Ga
in a table based
on the sys
for the first
,
tems of intransitivity of
line of the table
all
Ga +
Suppose that there are n such sys

//
tems, so that f(x) divides into
factors.
Then we take
and only those substitutions

first
of 6r a
which do
not convert the elements of the

those of
system of intransitivity into
The substitutions of this line form a another system. group, which is contained in G a as a subgroup. Its order is there The second line of the table consists of all the substi fore kr a+
i
.
tutions of
Ga
which convert the
first
is
the second.
lines,
The number of
these
also
system of intransitivity into kr a+1 There are such

.
,
and
they include all the substitutions of 6ra
Consequently
i.
e.,
the
number
p.
of the factors into which f(x) divides

/y*
is
a dim-
sor of the
number
r a+l
of the factors into which the Galois resolv-
ent equation divides at the
same
time.
similar result obviously
occurs in every later decomposition.
The decomposition can therefore only take place according scheme of Theorem III. The several irreducible factors are the same order.
236.
to the
all
of
Thus
far
we have adjoined
to the given equation
f(x)=0
280
the root
<,
of a second irreducible equation only
when
the
were
We
13)
will therefore
x lt x 2 now adjoin
,
.
xn
This seems a strong limitation.
to the equation f(x)
all
the roots
of an irreducible equation
g(z)
The only case of interest without making this special assumption. is of course that in which the adjunction produces a reduction in the
group G of f(x) = 0. In the first instance we adjoin only a single root z of g(z) = 0. If the rational Suppose that G then reduces to its subgroup H^ x n ) belongs to function <p (x x.2 then the same reduction lt
1
.
l ^
of G can be produced by adjoining ^ instead of z Suppose that under the operation of (7, the function c^ takes the conjugate values with the groups ... These ? m respectively. 2 fit 21 lt
1
.
H H
values satisfy an irreducible equation
fc(EEEO
l
?,)
? 2 ).
(<f>
?,)
has,
= o.
by the mediation
l
The adjunction of z to the made î rational, so of
rational
domain
is
l ,
that y
,
.
a rational function of z
^1(0?!, ar2
x n )=00i)-
It appears therefore that, in order that the adjunction of z i
may
,
.
pro
duce a reduction of G, it is necessary and sufficient that there should be a rational non-symmetric function of the roots o>i(x l9 x 2 XH )
. .
which
2 1?
is
rationally expressible in terms of z
are Suppose that the roots of the irreducible equation g (z) = = / is and its Since is that satisfied Zo, 2,4, group k[<! (z)]
.
by
z n all the
<f
(z t )
is
(^1,2,3,.../,)
^(z
}
are roots of values of
k(<?)
.
= 0;
that
),
^(z 2 ),
vK 2^) are the conjugate
On
the other
hand the
coefficients of the
product
are symmetric functions of the roots of g(z) == has rationally known; and the equation k 1
and are therefore

all its
roots in
com
mon
with
k = Q.
Consequently k
(^) is a
power of
k(<?)
and the
//
values ^ (z,), 0(z2 ),

;
0(2>) coincide in sets of q each.
281
With a
write
slight
change in the notation

=(.
(*;)
for the z s
we can therefore
ft
It
is
Since 4>(z*M)-?(ZfiM) Q, this quantity is rationally known. therefore unchanged by all the substitutions of 1\ i. e., P inter -
and therefore gives rise to a group T of the changes the lines of ), -elements <p which is isomorphic to P. To the substitution 1 in T Correspond in / the substitutions of the subgroup A of order d which
l
<only interchange
z\,z 2 .among themselves, and so

,
among
themselves, z" u z" z

(1
z" q
on.
P and T are
cQ-fold isomorphic.
If
we
coordinate
all
and T which leave
<f l
unchanged, and again one substitution each from converts <p into $r 2 one which converts ^ into 9? 3
l
,
G
,
and T which
on,
and so
an
isomorphism
tion 1 in
is
also established
between
G and
T.
To
the substitu
correspond in
is
the substitutions of the subgroup
D
.
of
.
order d which
the maximal
common subgroup
of
H H
lt
Hm
/
Accordingly
and
are also isomorphic, and in fact their iso
morphism is (d and again (r

Consequently
dj-fold, as
shown by the preceding
considerations,
r^-fold, as appears
from the orders of
and
If
now we
then
adjoin to the equation

$? 2
,
. . .
f(x)
all
the roots of
= 0, ,g(z)
subgroup
(x l
X.2
,
function
. . .
XH )
=
.
o.
l
(f l
-j-
^ 2 -f
+ a m m = ta(z
<f>
,*,... ^ M ),
and
x n which can be rationally expressed every function of Xi, x z to an inclu in terms of z z 2 z^ belongs to the family of /> or
,
.
l ,
ded family.
as the
For every such function

,
.
is
rationally
known, as soon
Zu z2
z^ are adjoined to the equation f(x)
0.
of f(x)
-15)
Conversely, if we adjoin to the equation g(z)=Q there 0, it follows by the same reasoning that
"ofcl,^,
2 M ) =/> (#l, ^2, ..- in),
all
is
the roots
a function
282
l ,
such that every function of z expressed in terms of x^, x 2

,
. .
z^ which
to the
can be rationally
family of
cn
x n belongs
s, it
or to
an included family. is rational Since now

/>
in the z
follows from the above prop
erty that
16)
P*
= R(p\
and since w
is
where
is
a rational function;
rational in the
a? s, it
follows that
17)
or,
w
which
is
=R
(io )
the same thing,

P
17)
= R*(f>
).
and belong to the same ) of of all the roots The family. /=0 to = there adjunction fore gives rise to the same rational domain as the adjunction of all
16) and 17
it
From
follows that p
/?
the roots of g
to
/=
0.
it
It is obvious at once that the first adjunction, since

î
2
>
made
.
,
so that
reduces to J
But the proof just given was necessary any further reduction.
If
to exclude the possibility of
we
write
=
a
a\
=v,
it
follows that
v
if
the second adjunction

first
reduces the order r of

also reduces the order
r^
to its
th
part,
v
then the
adjunction
of / to
its
th
part.
of the adjunction of all the roots of any arbitrary equation 13) on the reduction of the group
effect
Theorem
=
1 ,
XIII.
The
of f(x)
all
can be equally well produced by the adjunction of the roots of an equation 12) which is satisfied by rational func
tions of
x2
xn
In spite of removal of apparent limitations, we have therefore

not departed from the earlier conditions, where only the adjunction of rational functions of the roots was admitted.
237.
Theorem XIV. If = f(x) = 0, g(z)
are two equations, the coefficients of which belong to the same rational domain, and which are of such a nature that the solution
283
of the second and the adjunction of all its roots to the first reduces the group of f(x] = to a self-conjugate subgroup contained in it
times as small, then conversely the solution of the reduces the group of the second to its th part. The first equation
v
->
of an order
group of f(x) 0, like that of g (z) 0, is compound, and v is a Those rational functions of the roots of factor of composition. one of the two equations, by which the same reduction of its group
is
accomplished as by the solution of the other equation are rational

latter.
in the roots of the
As we
see, the
group of f(x)
(z)
=
.
can be reduced by the solution

It is only necessary that there
.
.
of an equation g
0,
although the roots of the latter are not

x,2
,
.
l ,
xn
t
should be rational functions of z ,zzi

functions of
z^.
which are also rational
x2
xn
From
the preceding
I.
Theorem follow
at
once the Corollaries
Corollary
simple, the equation
is // the group G of the equation f(x) = can only be solved by the aid of equations with
groups the orders of tvhich are multiples of the order of G.
For
since
reduces to
1,
the
of
Theorem XIV must be taken
equal to the order of G.
can be reduced Corollary II. // the group G of f(x) z^ are by the solution of a simple equation g(z) = 0, then z z 2
1
rational functions of the roots of f(x)
= 0. = 0.
For
in this case
v is
equal to the order of the group of g (z)

is
After the reduction this
equal
1.
Consequently
where
9\ is the Galois resolvent of g(z)
= 0.
Q
0,
Corollary III.
is to
}
If the adjunction of the roots of g(z)
produce a reduction of the group of f(x) r and r cannot be prime to each other.
then the orders
Sections contain a proof and an extension, rest to the theory of substitu ing entirely on considerations belonging the where of Theorem subject was treated purely 216, II, tions,
The preceding
arithmetically.
For
if
we
retain the notation of
216,
it
follows
284
that since
Vp = Fv
is
a simple equation,
is
a rational function
of the roots of f(x)
= 0,
and so
on.
The proof
Considerations.
238. to
of the impossibility of the algebraic solution of the
equations of higher degree
might therefore be based on the present
= /(a?)
As the adjunction of the roots of a new equation g(z} leads to nothing more than the adjunction of rational
functions of the roots of f(x) 0, so no new result is obtained, if the roots of both equations are connected by a rational relation.
We
prove
Theorem XV.
If
/Or)
= 0,
g(z)
are two irreducible equations, the roots of which are connected with each other by rational relations
?i(tfi, a? 2
,
.
.
XH
zu
... z^)
= 0,
,z^\
the latter
can
all be
obtained from a single relation of the form

.
ffajXt,
..xH) = x (zl9 zt ,.
in which the roots of the tivo equations are separated.

For,
if
we denote the corresponding Galois
resolvents
and the irreducible resolvent equations of f(x)
by c and C and g (z) = by
^() = o,
the degrees r and r
respective groups.
of
0(:)
= o,
and
are equal to the, orders of the
Now
z
l ,
#,
.
x n can be
T,
rationally expressed in terms of
and
z^
Zf,,
in terms of
p(a?i,
so that
Zt,
xn
z,
zj
The expression
that
its
degree becomes
can be so reduced by the less than r in c and
= 0. aid of F =
$(;, less
and
C.
G=
Then
than r in
the two equations

#(,-,3
= 0,
^(,-)
=o
have a common
root. Consequently, if we add C to the rational the resolvent domain, F(=) becomes reducible, since otherwise the irreducible equation of the ? tth degree would have a root in common

with an equation of a degree less than occurs when 0(1, C) is identically 0.
r.
285
The only exception

g(z)=
If this does not happen, the adjunction of all the roots of
or that of
breaks up the resolvent of /(*) have therefore the case of the last Section.
= into factors, and we We can effect the same

/of
*V).
a?
reduction by the adjunction of a rational function
a?2 ,
.<
,.
and we have
X(X,.
^2
...
3C n
<^
(Z
Z2
If several such relations exist, they can all be
deduced from one and

if
the same equation. function / such that
The
all
latter
can be easily found,
we
select
On
of :
the others belong to an included family. the other hand if </ (, ^) is identically 0, it follows that the
( } l
coefficients in the rjolynomial
arranged according to powers vanish, so that we have equations of the form

(=. ~)
/i(-0=-^(^,2
and
similarly,
if
</
zM)
= 0,
,
(r,
is
arranged in powers of
1
cM,.r ,a- 2
,...^)=0.
to
But this amounts 0. These equations can actually make $ of the roots of only an apparent, not an actual dependence
/(cc)~0 and g(z) = Q. of g(z) = 0, and c 2 (^,
The
o- 2
. .
function / 2
.
x n ) belongs
to the
group of f(x)
= 0.
CHAPTER XV.
ALGEBRAICALLY SOLVABLE EQUATIONS.
239.
In
234 we have established the following theorem.

the equation f(x)
G,
If
the group
1)
G of
=
,
.
has the series of composition

. .
G ly G
Gv
1,
and
if
the orders of these several groups are
then the solution of f(x)
can be effected by solving a series of
simple, irreducible equations of degrees

r
the
first
of
which has for
its
coefficients functions
6r,
,
belonging
on.
to
and
for its roots functions belonging to
the second coefficients
belonging to
equations
and roots belonging
to
Gr 2
and so
All these
*i
= 0,
72
= 0,... / =
<),
A +1 =
have the property that the roots of any one of them are all rational functions of one another, so that the order of the corresponding
group
is
equal to
its
degree
i.
e.,
the group
is
of the type
129).
We
have now
to
examine under what circumstances
all
these
equations y v
become binomial equations of order p A
where H\
is rational in the quantities belonging to the family of In other words, we have to determine the necessary and GTA-I. shall be algebraically solvable. sufficient condition that f(x)
For
r
,
this result it is necessary that the factors of composition

. . .
TZ
r3
--,
should
all
be prime numbers, Pi,p 2 ,p 3

/j
For these
0, ...
quotients give the degrees of the equations
= 0,
^2
0> /s
287
been shown in Not that every function belocging to G A on being raised to the (p A ) th power gives a function belonging to G A _ I; but some function can always be found which
^
This condition
is
also sufficient, as has already
110, 111,
Theorems
and XII.
has this property, as soon as the condition above We have then
is fulfilled.
Theorem
may
I,
In order that
be algebraically solvable,
its
factors of composition of
240.
the algebraic equation f(x) necessary and sufficient that the group should all be prime numbers.
it is
By the
aid of
Theorem XII,
110 we can give
this theo
rem another form
Theorem
may
II.
In order that
it
the algebraic equation f(x)
group should consist of a
series
25
necessary and of substitutions

is
tv
sufficient that its
1? ^1?
35
,t v +
which possess the two following properties: 1) the substitutions of the group <7 A / A _ n #A are commutative, except those jl, f n # 2 which belong to the group G^, = 1, \, t 2 and fe_ lf
,
.
.
t^_,\,
2)
power of prime number (cf. also

fA
,
the lowest
which occurs in
91,
GA _
(
has for
its
exponent a
Theorem XXIV).
J4 enable us to state The Again the investigations of orem I in still a third form. It was tkere shown that if the prin
cipal series of
2).
241.
G
G,
H,
J,
K,
does rot coincide with the series of composition, then 1) can be obtained from 2) by inserting new groups in the latter, for example
between
H and J the groups H H",

,
jffW.
Then
from
the factors of composition which correspond to the transitions

to
r
,
from
H
95)
to
H",
from
H ^ to
are
all
equal.
Accordingly, if all the factors of composition belonging to 1) are are not equal, then G has a principal series of composition 2).
We
H, H
saw further
,
that,
if,
in
passing successively from
H",
...
to the following group, the corresponding factors of

all
-composition were
prime numbers, (which then, as we have
just
288
seen, are all equal to each other),
tions of
H are commutative,
III.
and only
in this case, the substitu
except those which belong to J.
From*
this follows
Theorem
may
In order that the algebraic equation f(x)

it is
=
its
necessary and sufficient that
principal series of composition

Gr,
H)
J,
K,
... 1
should possess the property that the substitutions of every group are commutative, except those which belong to the next following group..
The
substitutions of the last
group of the
all
series, that
which pre
cedes the identical group, are therefore
commutative.
242. Before proceeding further with the theory, we give a few applications of the results thus far obtained.
Theorem
From
of F.
IV.
If a group
is
simply isomorphic with a
solvable group G, then
is
also
a solvable group.
96 the factors of composition of
coincide with those
Consequently Theorem IV
follows at once from
Theorem
L.
Theorem
the solvable
V.
// the group
is
multiply isomorphic with
group G,
and
if to the substitution I
of
corresponds
the subgroup - of I\ finally if - is
a solvable group, then
V is
also
solvable.
The
and those
factors of composition of r consist, from 96, of those of Reference to Theorem I shows at once the of .
validity of the present theorem.
Theorem
are also solvable.
VI.
If a group
is
solvable, all its
subgroups
We
write as usual
$j
f/
}
x^
-f-
-j-
-f-
Xn
apply to
r, all
the substitutions of G, obtain c 15 c 2

flf
...$,.,
and form
()
,) (,-
2)
,).
It is characteristic for the solvability of
that g
(I)
can be resolved
into linear factors by the extraction of roots.

If
now
of order r
is
a subgroup of G,
and
if
the applica-

tion of
289
.
to
gives rise to the values
,,
are
all
contained
among
*
,
*
a
,
.
.
^,
then these
*
.
rm
Consequently
(^
\) \*
*a/
?nJ
a divisor of g($). linear factors, i. e.,

is
Then
/i
(I) is also
resolvable
algebraically into
is
a solvable group.
this
We
might
also
of composition of
H
the
have proved
occur
by showing that
all
the factors
among
those of G.
Theorem
prime number p,
VII.
If the order of a group

is solvable.
is
a power of a
group
The group G is of the same type as a subgroup of the group which has the same degree n as G and for its order the highest 39 and 49). That the latter power pf which is contained in n (cf.
!
group
is
solvable follows from
its
construction
39), all of its fac

p.
tors of composition being equal to the
prime number
It follows
then from Theorem
VI
that
is
also solvable.
Theorem
where p ,p 3j p 3
}
VIII.
// the group
is
of order
are different prime numbers such that

,
Pi
> Pfpjpf
P > PJP?
2
P > pf
3
then
is
solvable*
use of the theorem of
128,
We make
then P!
If
and write
= pfq,
where
>
q.
G
G
:
contains at least one subgroup

kp^
-f-
H of
the order pf.
we denote by
1 the total
number
of subgroups of order
pf
and by pfi the order of the maximal subgroup of G which is commutative with H, then r = pf i(kp -\~ 1). Since r pfq and q<pu we must take k = and r = pfi. That is, G is itself
contained in
t
By the solution of an auxiliary equation of of order q, we arrive therefore at a function with a degree group to the of H, and the group G reduces to H( 232), family belonging
q,
commutative with H.
Theorem X).
Accordingly,
solvable also.
if
From Theorem VII
the latter group

is
is
solvable.
the auxiliary equation
solvable, the
group
is
The group of the auxiliary equation with the order q=pfpj.. admits of the same treatment as G. Its solvability therefore follows
.
*L. Sylow: Math. Ann, V, p.
585.
19
290
from that of a new auxiliary equation with a group of order and so on.
243.
p^pf
We return
p
TI
}
to the general investigations of
241.
The
transition
from
to
G
The
decomposes the Galois resolvent

transition from
equation into
factors.
to G.2
A*
decom2
poses each of these previously irreducible factors inte

factors,
=p
and so
on.
Since
f(x) =
into linear factors,
it
was originally irreducible, but is finally resolved 235 that once or oftener a reso follows from
lution of f(x) or of its already rationally
known
factors will occur
simultaneously with the resolution of the Galois resolvent equation

or of
into
its
already
known
rational factors.
The number
of factors
which f(x) = resolves, which is of course greater than 1, must from 235, be a divisor of the number of factors into which the
In the case of solvable equa number Pi,p 29 p3, Conse
-
Galois resolvent equation divides. tions the latter is always a prime
0. quently the same is true of f(x) n the solvable equation f(x) degree of
All prime factors of the
are factors of composi
tion of the group
6r,
and
it
in fact each factor occurs in the series of
composition as often as
occurs in
it
n.
if
To avoid
from
a natural error,
must be noted that
in passing
G\ the polynomial f(x) resolves into rational factors one of which is f \(x), this factor does not necessarily belong to the
to It
group G\.
may belong
to a family included
is
in that of
G\.
The number
to r r A
: .
of values of f\(x)
therefore not necessarily equal
It
f
may be
a multiple of this quotient.
And
the product
x(x) -f
\(oc)
to f(x),
but
will
We
so that
... of all the values of f \(x) is not necessarily equal may be a power of this polynomial. now assume that n is not a power of a prime number p,
n includes amoog its factors different prime numbers. Then different prime numbers also occur among the factors of composi tion of the series for G, and consequently ( 94, Corollary I) G has
a principal series G, If, J, K,
.
.
M,
1.
Suppose that in one of the series of composition belonging to

other groups

3)
291
H",
HW
among
its
occur between
least
and
J.
Since n includes
factors at
two different prime numbers, /(a?) must resolve into factors at least twice in the passage from a group of the series of composition
to the following one.
Since the
number
and
of the factors of /(a?)

since the latter
is
is
the
same
as the factor of composition,
the
same
for all the intermediate groups 3), the two reductions of /(x) cannot both take place in the same transition from a group of the prin It is to be particularly cipal series to the next following group J.
noticed, that all the resolutions of f(x) cannot occur in the transition
from the
last
group
M to
1,
that
is,
within the groups

>,
M ,M",
tions
first
... If (-
1,
following M in the series of composition.
At
least
one of the resolu
must have happened before M. Suppose, for example, that the and H". Then it follows from resolution occurs between
235 that
is
non-primitive in those elements
which
it
connects
transitively,
and that
H"
is
intransitive, the systems of intransitivity
coinciding with the system of non-transitivity of transitivity then occurs in all the following groups
likewise in the next group
tion
is
H H
The same
",
.
. .
in
H&\
and
of the principal series,
which by assump
different
from
1.
Suppose that
distributes the roots in the intransitive systems
these systems being taken as small as possible.
Then the expression
becomes a rationally known factor of f(x\ which does not contain any smaller rationally known factor. Since from the properties of
the groups of the principal series
all
the values of
$x)
belong to the same group
fore all rationally
known with f\(x).
J. They are there Of the values of /\(.r) we
know
already
/ * (x)
J
jfr
= (x
/^ (X
//v
/~A \ (X)
x ,) ^n \ x )
1
(x
(
o/ 2 )
(X
\
( rv*
T" 2 \ x )
(x (T
a?
,-),
^
I
r" L
f Y*
\
,-;,
/xv\
/(fft\ )(X)
(X
X?
/v
(***")
(X
X
/v
w*)\
>)
V* (X
\^j\ X? )-
292
If there
some /^ a) (x).
were other values, these must have roots in common with Then /A a) (x) and consequently f\(x) would resolve
^
into rational factors.
This being contrary to assumption, f\(x) has
only
m=
7?
values,
and
is
therefore a root of an equation of degree
m.
4)
If this equation is
y(y)=(yf$ (y-f
is
\]
(y-f^ m} ) = o,
-2
...
then f(x)
5)
the result of elimination between 4) and
f\(x)
=x
-^ X-
-f
vW)*
- 0,
where
so that
(J
are rationally expressible in terms of

5),
it
f\.
Since
f(x)
is
the eliminant of 4) and
follows from
228 that the
is non-primitive. group of f(x) = These conclusions rest wholly on the circumstance that
to the principal series of G,

is
and that accordingly
J belongs G~^JG = J. It
This shows
only under this condition that all the values of
f \(x) which occur
in the rational
itself
domain
of f(x)
very strikingly in an example to be presently considered.
equation
If the degree n of an irreducible algebraic by two different prime numbers, then n can always be divided into two factors n = im, such that the given equa tion f(x] = resolves into m new ones
is divisible
Theorem IX.
A(*)=o,/ \(*) = o,... /<ru*)=o,

which are
able
all
of degree
i,
and
the coefficients of which are obtain
from known
quantities by the solution of

the-
m.*
The group of
equation f(x)
=Q
is
non-primitive.
For the purpose
of comparison
we
consider the solution of the

2
2
,
general equation of the fourth degree, to which, since 4
the
preceding results are not applicable.

in the
It
appears
at
once that both
of the resolutions of the polynomial into linear factors take place
M,
M",
domain belonging to the last group of the principal series ... 1. The series of the equation consists of the follow
II, p. 191.
ing groups:
*Abel: Oeuvres completes

1) 2)
293
the symmetric group;
the alternating group;
4)
5)
Il,(x x 2 )(x.x,)l 4 )[l,(,aj 8 )(a?2 ar 4 )], or 4") the group 1.

1
[1
(x, x,) (x, x,)]
The
from
principal series consists of the groups 1), 3) to 4) and that from 4) to 5) both
is
2), 3), 5).
The passage
2.
give the prime factor
The group 4)
the
first
intransitive one.
the two factors (x x^ (x x2 ) the group 4) does not belong to the principal series, all the six val uesof (x x,2 ) are not known. If we had chosen the a-j) (x
For this/(r) resolves into and (x x 3 ) (xx,). But since

group
s
we should have had the two factors (x x^) (x x ) and (x x )(x a? 4 ), and so on. The product of these six values x ) (xx. ) (xx,) (x x^). give the third power of f(x)^(x
)
instead of
2
4),
can therefore, to be sure, resolve f(x) into a product of two factors of the second degree. But the coefficients of every such
factor are the roots not of an equation of degree
-^
We
= 2,
but of an
equation of degree 6. If we consider further the irreducible solvable equations of the sixth degree, it appears that these are of one of two types, accord ing as
we
eliminate y from
or from
*"
/iOK +f*(y)xf*(y) = o,
The preceding
p.
if
c,
y+% =
is
0.
244.
results enable us to limit our considera
tion to those equations /(x)
the degree of which
power
of a
prime number
For otherwise the problem can be
simplified
by
regarding the equation as the result of an elimination. Further more we may assume that such a resolution into factors as was con
sidered in the preceding Section does not occur in the case of our A present equations of degree p since otherwise the same simplifica
,
tion
would be
is
possible.
We
assume therefore that the group of the

to the investigation of the
is
equation
primitive, thus excluding both the above possibilities.
With
group.
this
assumption we proceed
is
Suppose that the degree of the equation
p A and
that
its
principal series of composition

2)
G,
H,J,K,... M,
1,
294
In passing from G through ,/,... to M, no resolution of Otherwise we should have the case of f(x) into factors can occur.
the last Section, and
G would
be non- primitive.
The passage from
to
"prepares" the equation f(x) for resolution, but does not
as yet resolve f(x) into factors.
The
in
resolutions of the equation
of degree
therefore occurs in passing from the last group of the

is,
principal series to identity, that
M,
M", ...M*~\
1.
Accordingly we must have x^/.

lary
The
application of
94, Corol
IV shows
that
all
are
is
commutative.
therefore
The equation characterized by the family

Abelian equation of degree
every transition
A
_p
(
of
M
M
an
to
182).
From
94 there belongs
is
from one group
to the next in the last series
the
.
factor of composition p,
so that the order of
K equal to p
can be obtained by combining x groups which have only Again the identical operation in common, which are similar to each other,
and are of order
p.
Suppose that these are

0c-i)
?
M
From
is
(K
-i)
Mf-i)
M(^
the above properties
it
appears that every one of these groups
composed
of the powers of a substitution of order
and that on account
of the commutativity of the groups
(cf.
95)
we must
also
have
S<f 8 ft"
= S^
Sa
(a, /?
Consequently every substitution of
= 0, 1, * M can be expressed by
.
1).
*f /-
*V-i*
and from the same commutative property
Every substitution of
the
group
is
of order p.
Our Abelian
186,
equation therefore belongs to the category treated in

substitutions are there given in the analytic form
and
its
= |z,,z
.z*
&i4-0i, 3*
of
+a
+
l ,
(mod. p).
z2
.
The symmetric occurrence
all
the indices z
Z K already
295
shows that in the reduction of

polynomial f(x) will occur, as
M
is
to 1 exactly
resolutions of the
if
also recognized
we
write for
example
lz,
z2 za
,
ZK
z 19
z, -j-
z,
3
M". = \z lt z
-f o 8
,
.
ZK -f
K
;
(mod. p),
(mod. p),
...ZK
^,2,2:3-4-
.s.
+ a|
Accordingly
= and we have as a first result Theorem X. The last group of the principal
*
/,
series
of a
arith
primitive, solvable equation of degree
pK
consists of the
metic substitutions
t=\z jZat ...z

l
lt
z -\-a 1J
l
+2j
ZK-\-" K
(mod. p),
the roots of the equation being denoted by
a*,i......
(*A
= 0,l,2,...p
is
1).
Since
<?,
the group of the equation,
commutative with
if,
it
follows from
144 that
is
a combination of arithmetic
and geo
metric substitutions.
We
have therefore as a further result
Theorem
tion of degree
XI.
K
The group
G of every solvable primitive equa

group of the arithmetic substitu
consists of the
tions of the degree
p combined
,
ivith geometric substitutions of the
same degree
(mod. p).
Before proceeding further with the general investigation, we consider particularly the cases x = l and = 2, the former of
245.
/.
which we have already treated above.
We
gree p.
is
consider
first
the solvable, primitive equations of prime de the term "primitive,

since
We may omit
non primitivity
impossible with a prime number of elements. The group of the most general solvable equation of degree then coincide with or be contained in
\
p must
G~\z az-\-a\ (a We prove that the former is the case,

1, 2,
...p
1;
0, 1,
1)
(mod.p).
by constructing the groups of

that
all
composition from
to
M and showing
the factors of com-
296
position which occur are prime numbers.
We divide p
1 into its
prime factors
=q
and construct the subgroup
z CI& z
then the subgroup

,|
(
2
= 1,2,...?
all
1
--.
,
= 0,1,.
if
.p-1),
and so
on.
Then
for
If. J,
belong to the principal series of G.

For,
Thus we have,
example
G~ JG = J.
z
we
take
az
then
z az
-\-
a
so that the transformation of a substitution of
with respect
to
any
substitution of
leads to another substitution of J.
Evidently the
principal series coincides here with the series of composition.

factors of composition g,
is
,
The
are
all
prime numbers.
The proof
then complete.
If a substitution of
leaves
it
leaves
all
the roots unchanged.
two roots x\ and x^ unchanged, then For from A = cu -j- a, // = a// -f- a
follows necessarily
tical
:
a~l,
u^0, and the substitution becomes iden
=
|
verts
o"x
+ into
1
,
unchanged and if it con For from x^, then every x v becomes a?^- A, (_*) + Aleaves one root x\
A=a / +
/Jt-zi"a(/4is
l)-f-, follows a=/A

z
j
/,
-f-
/(/
//
1), arid
.
|
the substitution
of the form
(//
A)^
/(/-
,a-{- 1)
leaves no root unchanged, and

is
fli
if
it
con
verts
converted into x v + _^. For only M / in this case is there no solution of the congruence /iâ/-f-,
;r A
into
a*
then every x v
when an^
These
1.
If /
,a
-f-
1 is to
become
//,
then we must have

<>
//
= /-(-.
.
This gives a
= / z z-\/, and the substitution is am pisipisely the same results which the earlier
us.
algebraic
method furnished
297
Theorem
degree
XII.
The general solvable equations of prime

196.
are those of
Their group
is
of order p(p
1)
and
consists of the substitutions of the

SEEE
form
1; a
z az
+ a\
= l, 2, (a
..p
= 0, 1,
..p
1)
(mod. p).
Its factors
tor ocurring as
of composition are all prime divisors of p 1, each fac many times as it occurs in p 1, and beside these
itself.
246.
We
.
pass to the general solvable primitive equations of

substitutions
degree
As a starting point we have the arithmetic
<=
which form the
last
z lt z 2
Zi
i,z 2
(mod.jp),
group
of the corresponding principal series.
To
arrive at the next preceding group,
we must determine
Its
a substi-
tution s which has the following properties.
form
is
SEE
|
^Zg a^ + 61^0, *!+&,
(mod.p),
and the lowest power of s which occurs in M, and is therefore of must have a prime number as exponent. Since now all /, the powers of s are of the same form as s itself, the required power
the form
must be \z^z. z lt Z = 1. s must be a prime number.

2 2
\
That
is,
the order of the substitution
these and other similar considerations we arrive at the fol * lowing results, the further demonstration of which we do not enter
From
upon.
Theorem
2
XIII.
is
The general
solvable, primitive equations
of degree p are of three different types.
The first type
characterized by a group of order
2p (p
I)
the substitutions of which are generated by the following :
(l,2 =0,
(a a
,
a2
1,2,
...#
. .
1), 1),
l, 2, 3,
.p
The groups belonging
to the
second type are of order 2
(p
1),
and
their substitutions are generated

Jordan
:
by the following :
* C.
LiouvJlle, Jour, 4e Math. (2) XIII, pp. 111-135.
20
298
62^
-f"
az 2
(a,6=0,
1,
...
1; but
not a,
6=0),
*a
*i
(mod. p).
e s cm?/ quadratic remainder (mod. p). The groups of the third type are of order 24p 2 (p
form of
(mod.
4).
their substitutions is different,
according as p=l or p = 3
1).
The
In
the
former case the group contains beside the two

i
|
substitutions
i
r\
-i
o
1),
az ,az2
l
(a^l,Z,6...p
(mod. p)
also the following four:
where
(mod.
is
a
the
root of the congruence
(mod. p).
If
p~3
4),
group
contains the
first
two substitutions above,
together with the following four:
where
and
satisfy the congruence s -\-t

first
For p
8 the
and second
types,
and for p
(mod. p). 5 the second
type are not general.
These types are then included as special
cases in the third type, which is always general.

247.
We
return from the preceding special cases to the more

in the case of
general theory.
The same method which we have employed above

2
p can be applied in general to determine the substitutions of the in the principal series of composition. group L which precedes
M
l
is
obtained by adding to the substitutions

t
l ,
Z2
ZK
a,
Z 2 -f
Z K -f a K )
(mod. p)
of
& further substitution
(mod.
p\
299
where the
nent.
first
power of
s s
s to
occur
among
the V s has a prime expo

s itself,
Since
all
the powers of s are of the same form as
any power of
Consequently
the group
248.
stricted,
L = \t,
From
which occurs among the f s must be equal must be of prime order. It is further
s\
to 1.
necessary
should not become non-primitive.
the form to which the substitutions of

at
are re
we have
once
All the substitutions, except identity, which
Theorem XIV.
belong to the group
Z 19 Z 2
,
...Z K
2i
i,
affect all the elements.
The converse
proposition, which
in the general case.
was true For the element xzi

>
for x
SK
1, is
does not hold
3I>
unaffected by
only in case the
congruences
S)
(mod. p)
are satisfied.
Consequently, as soon as the determinant

1 1
*
*
=0
aK
the
]
(mod.p),
bK
.C K
can be so chosen that the congruences S) are not

l
satisfied
by any system z
consider
,z 2
ZK
We
now
all
the substitutions of the group
G
1
.
which
leave one element unchanged.
Since the distinction between the

. .
as merely a matter of notation, we may regard x Then the substitutions which leave this element the fixed element.
elements
is
unchanged are
If
we
adjoin
o?
0>0
...
to the equation,
the group
reduces to
300
Since
of
all
,
. .
.
are obtained
cc
by appending
s
to those
r
tt
the constants a 1} a 2
it
ways,
can be chosen in p K follows that the adjunction of a single root reduces G to its
and since the
(p)
part.
249.
of
We
>S)
will
now
leaves
x -\- 1
elements
consider the possibility that a substitution xzi eo Then the con ZK unchanged.
,
gruences
of
, ,
the preceding Section are satisfied by

. . .
+1
.
systems
of values z l z 2
ZK
riX)
2
i
-
rW? Z
<.f
,
ZK
r f A) *K ! i
(I
V*
1 9 V, 1, 6 9
X).
We
will
howeyer regard not the

CI
(A)
,
coefficients a, 6, ... c; a
.
.
of the
substitution but the values
(A)
2
C K (/V) as
to determine the substitution
from these
data.
known, and attempt If now the determ
inant
E=
is
*
not
~h 1
TK ) each of (mod. p), then the * systems T^, T2 )> congruences with the unknown quantities a, 6, ... c a
.
.
.
=0
=o,
have only one solution each,
A)L
i
viz:
= 1,
6,
= 0,
G!
= 0;
at
= 0,
and these solutions furnish together the identical substitution 1. We designate now a system o/ x -j- 1 r*oofe o/ cm equation for which (mod. p) as a system of conjugate roots.
E=Q
We
have then
Theorem XV.
group of degree p*
If
leaves
a substitution of a primitive solvable unchanged x -j- 1 roots which do not
form a conjugate
system, the substitution reduces to identity.

If therefore
301
1 such roots to the equation, the reduces to those substitutions which leave x-j-1 roots unchanged, i. e., to the identical substitution. The equation is then
we] adjoin
group
solved.
Theorem XVI.
tion of degree
All the roots of a solvable primitive equa can be rationally expressed in terms of any x -f 1
among
If
,o,
. . .
them, provided these do not
form a conjugate
system.
x -f- 1
we choose
o,
the notation so that one of the
roots
is
the determinant becomes
E=
If
fit
i
ft!
a
COO c,(>
the roots are not to form
(mod. p).
The number
dition is determined in r
K
a conjugate system, then J r of systems of roots which satisfy this con 146. found
We
= (p l) (p p) (P P Theorem XVII. For every root

K K
2
(p
pK
).
o^
,..,.,
we can
deter
mine
(p
K
l)
(P
P)
(P
p*-
1, 2, ... x
systems of
roots each such that these
* -f- 1
roofe do
no /orm a con
jugate, system, so that all the other roots
can be rationally expressed

x
in terms of them.
The system composed of the

*^1
,
,
.
.
+ 1 roots
<^0
...
^0,0,0,...!
is
appropriate for the expression of all the
roots.
These results throw a new light on our earlier investigations in regard to triad equations, in particular on the solution of the Hes
sian equation of the ninth degree
(cf.
203-6).
It is plain that
we can
construct in the same

on.
way quadruple equations
of degree
3
,
and so
AU BOOKS MAY BE
R ECA
a EOAFTER7 DAYS
U.C.
BERKELEY LIBRARIES
0037^03^40
;*>
db
UNIVERSITY OF CALIFORNIA LIBRARY

Netto Theory Substitutions

Hochgeladen von

Dokumentinformationen

Originalbeschreibung:

Copyright

Verfügbare Formate

Dieses Dokument teilen

Dokument teilen oder einbetten

Freigabeoptionen

Stufen Sie dieses Dokument als nützlich ein?

Sind diese Inhalte unangemessen?

Copyright:

Verfügbare Formate

Netto Theory Substitutions

Hochgeladen von

Copyright:

Verfügbare Formate

LIBRARY

PROFESSOR OF MATHEMATICS IN THE UNIVEHSITY OF GIESSEX.

Revised by the Author and Translated with his Permission

Assistant Frofessor of Mathematios in the University of Michicjan.

THE REGISTER PUBLISHING COMPANY.

deals, the possibility of its

lacking, but not defective, generality. a smaller sphere of action.

A greater efficiency is attained in

the employment of integral functions from the outset,

proper to mention expressly here that

proofs and pro^

of his entire work.

THEORY or SUBSTITUTIONS AND OF INTEGRAL FUNCTIONS.

ALTERNATING AND TWO-VAL-

13. 14. 15.

Two-valued functions; substitutions. Decomposition of substitutions into transpositions.

Treatment and group of the two-valued functions.

MULTIPLE-VALUED FUNCTIONS AND GROUPS OF SUBSTITUTIONS.

Groups of substitutions. Correlation of function and group. Symmetric group.

THE DIFFERENT VALUES OF A MULTIPLE-VALUED FUNCTION AND THEIR

Relation of the order of a group to the of the corresponding function.

to the different values of a function.

The Cauchy-Sylow Theorem.

SIMPLE AND COMPOUND GROUPS.

self -con jugate

The alternating group

Principal series of composition. The factors of composition equal prime numbers.

The degree and order

Construction of isomorphic groups.

FAMILIES OF MULTIPLE-VALUED FUNCTIONS.

functions in terms of the

THE NUMBER OF THE VALUES OF INTEGRAL FUNCTIONS.

form of the question.

Functions whose number of values

than their degree.

and non-primitive groups.

Groups with substitutions of four elements. General theorem of C.Jordan.

CERTAIN SPECIAL CLASSES OF GROUPS.

129. 130. 131.

Groups of commutative substitutions.

ANALYTICAL REPRESENTATION OF SUBSTITUTIONS.

THE LINEAR GROUP.

Condition for the defining function. Arithmetic substitutions.

the constants of a -geometric substitution.

Order of the linear group.

APPLICATION OF THE THEORY OF SUBSTITUTIONS TO THE ALGEBRAIC

of the second degree,

THE CYCLOTOMIC EQUATIONS.

Definition and irreducibility. Solution of cyclic equations.

The regular pentagon. The regular heptadecagon.

THE ABELIAN EQUATIONS.

One root of an equation a rational function of another. Construction of a resolvent.

Employment of special resolvents Second method of solution.

Solution of the simplest Abelian equations. for the solution.

EQUATIONS WITH RATIONAL RELATIONS BETWEEN THREE ROOTS.

Triad equations. Constructions of

of the triad equation for of the triad equation for

Hessian equation of the ninth degree.

THE ALGEBRAIC SOLUTION OF EQUATIONS.

Representation of the roots of a solvable equation.

Solvable equations of prime degree.

THE GKOUP OF AN ALGEBEAIC EQUATION.

Galois resolvents of general and special equations. Composition of the group.