Chapter XV.
On
Soluble and Composite Groups.
T he most general problem of pure group-theory, so far as it is concerned with
groups of finite order, is the determination and analysis of all distinct types of
group whose order is a given integer. The solution of this problem clearly involves
the previous determination of all types of simple groups whose orders are factors
of the given integer.
Now there is no known criterion by means of which we can
say whether, corresponding to an arbitrarily given composite
integer
as order, there exists a simple group or not. For certain particular forms
of ,
this question can be answered in the affirmative. For instance, when
,
previous investigation enables us to state that there is a simple group of
order ; but
even in these cases we cannot, in general, say how many distinct types of simple group
of order
there are.
On the other hand, we have seen that, for certain forms
of ,
it is possible to state that there is no simple group of
order . Thus,
when is the
power of a prime, or is the product of two distinct primes, there is no simple group whose order
is ; and further, every
group of order is soluble.
Again, if is divisible by
a prime , and contains
no factor of the form ,
a group of order
cannot be simple; for, by Sylow’s theorem, it must contain a self-conjugate sub-group having
a power of
for its order.
We propose, in the present Chapter, to prove a series of theorems which enable
us to state, in a considerable variety of cases, that a group, which has a
number of given form for its order, is either soluble or composite. If the
results appear fragmentary, it must be remembered that this branch of
the subject has only recently received attention: it should be regarded
rather as a promising field of investigation than as one which is thoroughly
explored.
The symbols , ,
, … will
be used throughout the Chapter to denote distinct primes in ascending order of
magnitude; while distinct primes, without regard to their magnitude, will be denoted by
, ,
, ….
242. THEOREM I. If
is one of conjugate
sub-groups of a group
of order ,
and if is not
a factor of ,
cannot
be simple.
The group
is isomorphic with the group of substitutions given on transforming the set of
conjugate sub-groups by each of the operations
of . This is a transitive
group of degree ,
and its order therefore is equal to or is a factor
of . Hence, if
is not a factor
of , the isomorphism
cannot be simple; is
therefore not simple.
Corollary94. If a prime
divides , and if
is the greatest
factor of which is
congruent to unity ,
will be composite
unless is
a factor of .
There cannot, in fact, be more than
conjugate sub-groups whose order is the highest power
of that
divides ;
so that this result follows from the theorem.
In dealing with a given integer as order, it may happen that, though no single
application of this theorem will prove the corresponding group to be composite,
a repeated application of Sylow’s theorem, on which the theorem depends,
will lead to that result. As an example, let us consider groups whose order is
. Any such group
must contain either
or sub-groups
of order ,
and either or
sub-groups
of order .
Hence, if the group is simple, it must contain
sub-groups of
order , and
sub-groups
of order ;
and each one of the latter must be contained self-conjugately in a sub-group of
order . Now the
sub-groups of
order would contain
distinct operations
of order , leaving only
others; and therefore
a sub-group of order ,
if the group contain such a sub-group, must in this case be self-conjugate. This is,
however, in direct contradiction to the assumption that the group contains
sub-groups
of order ;
therefore the group cannot be simple.
243. THEOREM II. Every group whose order is the power of a prime is
soluble.
This has already been proved in § 54.
THEOREM III95. A group
whose order is ,
where is
less than ,
being the
index to which
belongs ,
is soluble.
If ,
contains a
self-conjugate sub-group
of order .
If ,
contains
either or
sub-groups of
order . In the
latter case, if is
the order of the greatest sub-group common to two groups of
order , and if
is such a sub-group,
must (§ 80) be contained
self-conjugately in a sub-group
of order ,
,
;
and this sub-group must contain more than one sub-group of
order .
Hence must
be ; and therefore
must be common to
all the sub-groups
of order , so that
is self-conjugate,
and is
. If
is zero, no two
sub-groups of order
have a common operation except identity, and the
sub-groups
contain
distinct operations, so that a sub-group of
order is self-conjugate.
Hence when is
equal to ,
contains either a self-conjugate
sub-group of order
() or else one
of order .
If ,
contains
or
sub-groups
of order .
If there is only one, it is self-conjugate.
If there are sub-groups
of order ,
and if
()
is the order of the greatest group common to any two of
them, it may be shewn, exactly as in the preceding case, that
contains a self-conjugate
sub-group of order .
If ,
the
operations, other than identity, of each of the
sub-groups of
order are all distinct.
Now is isomorphic with
a group of degree . If the
isomorphism is multiple, must
have a self-conjugate sub-group whose order is a power
of . If the isomorphism is
simple, we may regard as a
transitive group of degree ;
and every operation, except identity, of a sub-group of
order
will leave one symbol only unchanged. Hence
is equal to or is a factor
of . Moreover, when
is thus regarded, the
sub-groups of order
must be transitive; for all the sub-groups of
order
will be given on transforming any one of them by the operations of a sub-group of
order .
Consider now a sub-group of
order ,
which keeps one symbol fixed and contains a sub-group of order
self-conjugately.
It will contain
() sub-groups of
order . Any one of these
must keep symbols
fixed; for it is the sub-group of a transitive group of
order and
degree , which keeps one
symbol fixed. If , there
are operations of order
which transform one of these sub-groups into itself, and therefore interchange among themselves the
symbols unchanged by
the sub-group. Hence
must be divisible by .
This requires that , which
is impossible. Hence , and
the sub-group of order
contains sub-groups
of order .
It is obvious that no two of these can enter in the same sub-group of
order ;
for if they did, they would generate a group of
order
(),
a sub-group of this order cannot enter in a group of
order . Therefore
contains
sub-groups
of order .
Since the sub-groups of ,
of order ,
are transitive, their self-conjugate operations must (§ 106) displace all the symbols,
and they cannot therefore be permutable with any operation whose order is a power
of . Suppose now that every
operation of a sub-group ,
of order , is permutable
with an operation of
order , and that there is
no sub-group of order of
which this is true. If is
a self-conjugate operation of a sub-group of
order to which
belongs,
is transformed into
itself and into a
new sub-group
by . Hence the sub-group
which contains self-conjugately
has two, and therefore ,
sub-groups of order .
No two of these can belong to the same sub-group of
order ,
for if they did they would generate a group of
order
(). Now since
is permutable
with ,
it must leave unchanged the symbol which
leaves unchanged.
Hence must leave
every symbol unchanged; i.e. it must be the identical operation. The order of every operation
of is therefore
either a power of
or a power of .
Suppose now that is
the order of the greatest sub-group that is common to any two sub-groups of
order , and that
is such a sub-group. Then
(§ 80) must be permutable with
an operation of order . Since
no operation of is permutable
with any operation of order ,
it follows that is
divisible by , and
therefore that .
Let , of
order , be the greatest
sub-group that contains
self-conjugately. If ,
is common to each
of the sub-groups
of order ,
and it is therefore a self-conjugate sub-group
of .
If , let
be a sub-group,
of order ,
of .
In
there must be a group, whose order is divisible
by ,
containing
self-conjugately. Hence there must be operations
in , which transform
into itself and
into a conjugate
sub-group . If
and
have a common sub-group, it must be transformed into itself by every operation
of . This is
impossible, since
is not divisible by
when .
In the group ,
is one of
conjugate sub-groups, and
therefore no operation of
can transform into itself.
Hence if an operation
of transforms
into itself, it must
transform into
a group , which
is conjugate to
in . Hence,
being the greatest sub-group
that contains self-conjugately,
must be common to
and .
But containing two
sub-groups of order , must
contain such sub-groups; and
therefore the sub-groups
of order that enter
in are identical with those that
enter in . This is impossible if
and
are distinct; for the sub-groups
of , of
order , generate
. Hence no operation
of , except identity,
transforms
into itself, and the number of sub-groups in the conjugate set to which
belongs must
not be less than ;
so that
Now we have seen that
from these congruences, and the inequality
, it follows
that
This inequality is inconsistent with the previous one, which follows directly from the
supposition that is
less than .
Hence , and
is a self-conjugate
sub-group of .
It follows therefore that, on every possible supposition,
must
have a self-conjugate sub-group. If this sub-group be represented
by ,
the same reasoning may be repeated with respect to the groups
and .
Hence is
soluble.
Corollary. All groups whose orders are
,
,
,
,
,
are
soluble.
Since the congruence
is satisfied only by ,
, the
results stated follow immediately from the theorem, except for the cases
and
.
If in these cases there are
sub-groups of order
(), there must (§ 78) be
sub-groups of order common
to two sub-groups of order ;
and, in the groups of order or
,
such a sub-group, if not self-conjugate, must be one of either
or
conjugate sub-groups. The groups must therefore be isomorphic with groups of
degree
or ;
from this it follows immediately that they must be soluble (§ 146).
244. THEOREM IV. Groups of
order
are soluble.
If a group , of
order , contains
only sub-groups of
order , it cannot be simple,
for a group of order
cannot be expressed as a substitution-group of
symbols. Similarly,
if contains a single
group of order ,
it is not simple.
If contains
sub-groups
of order ,
and if these sub-groups have no common operations, except identity, there are
operations
in whose orders are
powers of ; and therefore
a group of order
is self-conjugate. If the operations of the
sub-groups of
order are not
all distinct, let
be the order of the greatest sub-group common to any two of them; and let
be such a sub-group.
Then (§ 80) must be
self-conjugate in a group
of order
(,
or ).
If ,
is self-conjugate
in ; and
if ,
must contain
sub-groups
of order , and
therefore
Now (§ 78)
and therefore
unless ;
but, if ,
then
Again, if ,
is one of
conjugate sub-groups;
as before, cannot
then be simple.
Suppose now that is
a sub-group of order ,
and that the self-conjugate operations
of form a
sub-group . This must be one of
,
or
conjugate sub-groups; in the first two cases,
cannot
be simple. Moreover, if any two sub-groups of the conjugate set to which
belongs have a
common sub-group, it must be self-conjugate in a group containing more than one sub-group
of order ;
again, cannot
be simple.
Let
be the
sub-groups (§ 80)
of order that
contain ; and suppose
that is not contained
in . Every operation of
each of the sub-groups
is permutable with every
operation of . But these
sub-groups, since they occur in
and not in , generate a
sub-group of , whose order
is divisible by . Hence there
must be an operation , of
order , which is permutable
with every operation of .
This operation would be permutable with every operation of a group of order
at least, and it would
therefore be one of conjugate
operations at most. Now, if ,
; and as a
group of order
cannot be represented in terms of less than
symbols,
would
in this case have a self-conjugate sub-group. If
, then
either or
. A group of
order cannot be expressed in
less than symbols; a cyclical
group of order cannot be
expressed in symbols;
and a group of degree ,
which contains a non-cyclical sub-group of
order ,
must (§ 141) contain the alternating group. Hence again, in this case,
must
have a self-conjugate sub-group.
Finally, suppose that contains
the sub-groups
If contains
further sub-groups of the conjugate set to which
belongs,
they must be permuted among themselves in sets
of when
is transformed by
an operation , of
order , belonging
to . Hence we
may assume that
is the complete set of sub-groups of the conjugate set to which
belongs, contained
in ; and that, when
transformed by ,
those in each line are permuted among themselves. If
is a sub-group96
of order , the
sub-groups
are
permuted transitively among themselves, when transformed by the operations
of . Hence
if change
into
, it must change
the set into the
set and therefore
is a set of
sub-groups of
order , which have the
common sub-group .
Now, since by supposition is
not a self-conjugate sub-group of ,
there must (§ 55) be in a
sub-group conjugate to and
permutable with . Let this
be the sub-group of order ,
common to
No
one of the sets
can have more than one group in common with this set, as otherwise
would not be the order of the greatest group common to two groups of
order . Hence, of the
set of groups at least
are distinct from the
groups, conjugate to ,
which enter in . Every
group of order occurring
in will have a similar
set of sub-groups of
the set , which do not
occur in . Hence, since
the operations of
are the only operations common to two sub-groups of
order
in , there
must be in
not less than ,
i.e. , sub-groups of the
set to which belongs.
In other words, must contain
a self-conjugate sub-group of .
On every supposition that can be made, we have shewn that
must
have a self-conjugate sub-group. If this sub-group
is , and if the
order of either
or
contains ,
we may repeat the same reasoning; while if the orders of
and
are both
divisible by ,
we have already seen that they must be soluble. Finally, then,
itself
must be soluble.
245. THEOREM V. A group of
order , in which the
sub-groups of orders
and
are Abelian, is soluble.
Suppose, first, that there are
() sub-groups
of order .
The group is then isomorphic with a substitution group of
degree . Now the operations
of a sub-group of order
transform the sub-groups
of order
transitively among themselves; and therefore, in the isomorphism
between the given group and the substitution group of
degree ,
there corresponds to an Abelian sub-group of
order a transitive
substitution group of degree .
But (§ 124) an Abelian group can only be represented as a transitive substitution
group of a number of symbols equal to its order. Hence the isomorphism between
the given group and the substitution group is multiple; and the given group is not
simple.
Suppose, next, that there are
sub-groups
of order , and
therefore (§ 81)
distinct sets of conjugate operations whose orders are powers
of . Let
be an operation whose
order is a power of ;
and let , of
order , be the greatest
group which contains
self-conjugately. Then contains
at least operations whose
orders are powers of , and
therefore at least distinct
operations of the form , where
and
are permutable and
Now is
one of
conjugate operations; corresponding to each of these, there is a similar set of
operations
of the form ;
while no two such operations can be identical. The group therefore contains
operations of
the form ; and
similarly, it contains an equal number for each set of conjugate operations whose order is a power
of . Hence, finally, the group
contains operations whose
orders are divisible by .
There is therefore in this case a self-conjugate sub-group of
order .
The group therefore always contains a self-conjugate sub-group. A repetition
of the same reasoning shews that it is soluble.
246. THEOREM VI. A group of order
, in which the
sub-groups of order ,
, …,
are
cyclical, is soluble97.
If the number of operations of the group, whose orders divide
, is
exactly equal to this number, it follows from Theorem VII, Cor. I, § 87, that the
same is true for the number of operations of the group whose orders divide
. Now,
when ,
this relation obviously holds; and therefore it is true for all values
of . The group therefore contains
just operations whose orders
divide ; hence a sub-group
of order is self-conjugate.
Any sub-group of order
must transform this self-conjugate sub-group of
order
into itself, so that the group must contain a sub-group of
order . If
there were more than one sub-group of this order, the group would contain more than
operations
whose orders divide this number. Hence the group contains a single sub-group of
order , and this
sub-group must be self-conjugate. By continuing this reasoning it may be shewn that, for
each value of ,
the group contains a single sub-group of order
, such
a sub-group being necessarily self-conjugate. The group is therefore soluble; for we
have already seen that a group whose order is the power of a prime is always
soluble.
The self-conjugate sub-group of order
must contain the complete set of sub-groups of
order . If then
there are of these
sub-groups, must
be a factor of ; and a
sub-group of order
must be contained self-conjugately in a sub-group of order
Hence if
and
are any two indices between
and , the group contains
sub-groups of order .
Now the above sub-group, which contains a sub-group of
order
self-conjugately, is of the same nature as the original group. Hence,
and
being any two indices less than ,
it contains a sub-group of order .
This process may be continued to shew that, if
is any factor of
which is relatively prime to
, then the group contains
sub-groups of order .
Ex. Shew that, when the groups of
order
are also cyclical, there are sub-groups whose orders are any factors whatever of the order
of the group.
247. A special case of the class of groups under consideration is that in which the
order contains no repeated prime factor. These groups have formed the subject of a
memoir by Herr O. Hölder98. He shews that they are capable of a specially simple
form of representation, which we shall now consider.
Let
be the
order of a group .
It has been seen above that a sub-group of
order is self-conjugate.
Suppose then that ,
, …,
are the
orders of those cyclical sub-groups of prime order which are contained self-conjugately
in . They generate a
self-conjugate sub-group
of order ,
say . Each of the
cyclical sub-groups ,
, …,
of , generated
by an operation of prime order, is permutable with all the rest; and therefore (§ 34) the
operations ,
, …,
are all permutable. Hence, since their orders are distinct primes, the
sub-group is
cyclical. Let now be
a sub-group of
of order , where
. No self-conjugate sub-group
of can be contained
in . For if it contained
such a sub-group of order
(),
this sub-group would contain a single cyclical sub-group of
order ; and
would contain a self-conjugate
sub-group of order , contrary to
supposition. Since then contains
no self-conjugate sub-group of ,
it follows (§ 123) that can be
expressed transitively in symbols;
when it is so expressed, an operation
of order , that
generates , will be a circular
substitution of symbols.
The only substitutions performed on the
symbols, which are
permutable with ,
are its own powers; and therefore no operation
of is permutable
with except its own powers.
Every operation of must therefore
give a distinct isomorphism of .
Now (§ 169) the group of isomorphisms of a cyclical group is Abelian. Hence
must be
Abelian; and since its order contains no repeated prime factor, it must be cyclical. The
group is therefore completely
defined by the relations
where belongs
to index .
248. The theorem of § 246 is a particular case of the following more general
one, due to Herr Frobenius99, from which a number of interesting and important
results may be deduced.
THEOREM VII. If is
a group, of order ,
where
and are relatively
prime, and where
and if sub-groups ,
, …,
, of
orders ,
, …,
, are Abelian,
while
and ,
and
, …,
and
, are relatively prime;
then contains (i) exactly
operations whose orders
divide , and (ii) a sub-group of
order which has a self-conjugate
sub-group of order .
The theorem will be proved inductively, by shewing that, if it is true for any similar
group of order ,
where is a
factor of , it
is also true for .
Suppose that the greatest group, which contains
self-conjugately,
is a group
of order ,
where is a
factor of . Let
be any operation of this
sub-group, whose order
is a factor of .
Then (§ 177) since
and are relatively prime,
is permutable with every
operation of . Hence every
operation of , whose order
divides , is permutable
with every operation of .
Let now be any operation
of ; and suppose that the
order of , the greatest
sub-group which contains
self-conjugately, is .
If the order of
is , the
order of
is ,
say . We have seen
in § 176 that, if is
a sub-group of ,
then is equal to
or is a factor of .
Hence if is a
sub-group of order
of ,
and
are relatively prime;
therefore, a fortiori,
and are relatively
prime. Similarly and
are relatively prime,
and so on. Moreover is
a factor of .
We may therefore assume that the theorem holds
for . Hence this group contains
exactly operations whose orders
divide , and it also contains
a sub-group of order in
which a sub-group of order is
self-conjugate. Therefore contains
a sub-group of order in
which a sub-group of order
is self-conjugate, and exactly
operations of the form , where
and
are permutable and the order of
divides .
Now is one
of conjugate
operations in ;
corresponding to each of these, there is such a set of
operations
of the form ,
while (§ 16) no two of these operations can be identical. Hence
contains
operations of the
form , arising from the
conjugate set to which belongs.
If then we sum for the distinct conjugate sets of operations
of whose orders are
powers of , the number
of operations of ,
whose orders divide
and do not divide ,
is .
Now since is the
greatest group that contains
self-conjugately, the greatest common sub-group of
and
is the greatest sub-group of
that contains
self-conjugately. The order of this group certainly contains
as a factor; and
if the order is ,
then is one of
conjugate operations
in . Every operation
of however, whose order
divides , is permutable
with every operation of ;
and therefore . Hence the
number of operations of ,
excluding identity, whose orders are powers
of , is
,
where the summation is extended to all the distinct conjugate sets of operations
of whose orders are
powers of . It has
been shewn (Theorem IV, Corollary, § 81) that the number of distinct conjugate sets of operations
in , whose orders are
powers of , is the same
as the number in .
Hence the symbols
and
represent summations of the same number of terms.
Finally, in accordance with the induction we are
using, we may assume that the number of operations
of whose orders
divide is exactly equal
to . For the order
of may be separated
into the factors
and where
is a factor
of ,
while the conditions of the theorem hold for this separation.
Hence the number of operations
of whose orders
divide , is equal
to
while at the
same time
Unless ,
, …,
the former of these numbers is negative, which is impossible; these conditions
must therefore be satisfied. The number of operations, whose orders
divide , is therefore exactly
equal to ; and the order of
the greatest group that
contains self-conjugately
is . The factor-group
has for its order
, and it satisfies the same
conditions as . Hence it
contains a sub-group of order ;
and therefore contains
a sub-group of order .
The free use of the inductive process that has been made in the preceding proof may
possibly lead the reader to doubt its validity. He will find it instructive to verify the truth of
the theorem directly in the simpler cases. In fact, the direct verification for the case in which
is a
power of a prime is essential to the formal proof; it has been omitted for the sake of
brevity.
Corollary I. If , where
and
are relatively prime, has
a single conjugate set of sub-groups of
order .
That contains
groups of order
may be shewn inductively at once. For since it is clearly true when
contains
only one, or two, distinct prime factors, we may assume it true when
does not
contain as a
factor. Now has
a sub-group of
order , which contains
a sub-group of
order self-conjugately.
The factor-group contains
then a sub-group of order
where is any factor
of which is relatively
prime to ; and
therefore contains a
sub-group of order .
If now and
are two sub-groups of
of order ,
they may be assumed to contain the same
sub-group of
order . For the
sub-groups of
of order
form a single conjugate set; and therefore, if
contained a
sub-group of this
set, would contain
, which may be
taken to be . Now
is a self-conjugate sub-group of both
and ,
and therefore the factor-groups
and of
order are sub-groups of the
factor-group . Hence if the
result is true when does
not contain the factor ,
it is also true when does
contain .
But when is
equal to ,
the result is obviously true; and therefore it is true generally.
Corollary II. If the sub-groups
,
, …,
of contain characteristic
sub-groups ,
, …,
of orders
,
, …,
; then
contains a sub-group of order
, which has a self-conjugate
sub-group of order .
Assuming the truth of this statement when
does not contain the
factor , the factor-group
must contain a
sub-group of order ;
and therefore contains
a sub-group
of order
which has
for a self-conjugate sub-group. Hence, since
is a characteristic
sub-group of ,
contains a sub-group of
order and a self-conjugate
sub-group of order .
It therefore contains a sub-group of order
, which has a self-conjugate
sub-group of order .
249. The reader will have no difficulty in seeing, as has already been stated,
that Theorem VI is a direct result of the theorem proved in the last paragraph.
We shall proceed at once to further applications of it.
THEOREM VIII. A group
of order
in which the
sub-groups of every order
are
Abelian, with either one or two generating operations, is generally soluble; the special
case p1 = 2,
p2 = 3, may
constitute an exception100.
If an Abelian sub-group Pr
of order prαr
is generated by a single operation, it is cyclical and
𝜃(Pr) is (§ 176)
equal to pr − 1.
If Pr is
generated by two independent and permutable operations,
𝜃(Pr) is equal
to (pr − 1)(pr2 − 1). Now no prime
greater than pr can
divide (pr − 1)(pr2 − 1) unless
pr + 1 be a prime; and this is
only possible when pr is
equal to 2.
Hence unless P1 = 2,
p2 = 3,
𝜃(Pr) and
N
p1α1p2α2…prαr are relatively prime
for all values of r
from 1
to n − 1. Omitting
for the present this exceptional case, the conditions of Theorem VII are satisfied
by G; it therefore contains
exactly pnan operations whose
orders are divisible by pn.
Therefore G has a
self-conjugate sub-group Pn
of order pnαn,
and hence also one or more sub-groups of
order pn−1αn−1pnαn. There are
however only pn−1αn−1pnαn
operations in G
whose orders are divisible by no prime smaller
than pn−1, and therefore
the sub-group of order pn−1αn−1pnαn
must be self-conjugate. This process clearly may be continued to shew that
G has a self-conjugate
sub-group of every order prαrpr+1αr+1…pnαn
(r = 2, 3,…,n); from which it follows
immediately that G is
soluble.
Returning now to the exceptional case, suppose that a group of
order 2α1
is contained self-conjugately in a maximum group of
order 2α13βm. If every operation
of the group of order 2α1
is self-conjugate within this sub-group,
G contains
2α1 − 1
distinct conjugate sets of operations whose orders are powers
of 2; hence it follows that
there are just N
2α1 operations
in G whose orders are
not divisible by 2.
In this case, the conditions of Theorem VII are satisfied
by G;
and it is still soluble.
If, lastly, the operations of the sub-groups of
order 2α13βm, whose orders
are powers of 2,
are not all self-conjugate, there must be an
operation B, whose
order is a power of 3,
in this sub-group which is not permutable with every operation of the sub-group of
order 2α1. Let now
A,A1,…,Ar,Ar+1,…
be a characteristic
series of the group A
of order 2α1; and
suppose that Ar+1 is
the greatest of these groups with every one of whose
operations B is permutable. Then
(§ 175) B is not permutable
with every operation of Ar
Ar+1.
Hence Ar
Ar+1 is
a quadratic group; and its three operations of
order 2
must be permuted cyclically when transformed
by B. It
follows that {Ar,B}
is multiply isomorphic with a tetrahedral group. Hence finally, under the conditions of the
theorem, G is
certainly soluble unless it contains a sub-group which is isomorphic with a
tetrahedral group.
Corollary. A group of order p1α1p2α2…pnαn
in which no one of the indices α1,
α2, …,
αn−1 is greater
than 2, is
soluble unless it has a sub-group which is isomorphic with a tetrahedral group. For a group
of order pr2
is necessarily an Abelian group which is generated by either one or two
independent operations.
250. We shall next consider certain groups of even order in
which the operations of odd order form a self-conjugate sub-group. If
G is a group
of order N,
where
N = 2αn, (n odd),
we shall suppose, in this and the following paragraphs, that a
sub-group 𝒬 of
order 2α is Abelian,
and that 𝜃(𝒬)
and n
are relatively prime. Let us take first the case in which
𝒬 is cyclical,
so that 𝜃(𝒬) is
unity and the latter condition is satisfied for all values
of n. If
such a group is represented in regular form as a substitution group of
2αn letters,
the substitution corresponding to an operation of
order 2α will
consist of n cycles
of 2α symbols
each. This is an odd substitution; and therefore
G has a self-conjugate
sub-group of order 2α−1n.
In this sub-group there are cyclical operations of
order 2α−1.
Hence the same reasoning will apply to it, and it contains a self-conjugate sub-group of
order 2α−2n. Now (Theorem VII,
Cor. I. § 87) G contains
exactly 2α−2n
operations whose orders are not divisible
by 2α−1. Hence the sub-group of
order 2α−2n must be self-conjugate
in G. Proceeding thus, it may
be shewn that G contains
self-conjugate (and characteristic) sub-groups of every
order 2α−rn
(r = 1, 2,…,α)101.
251. Suppose, secondly, that, under the same conditions,
G contains a
self-conjugate sub-group 𝒬
of order 2α.
Since 𝜃(𝒬) and
n are relatively prime, every
operation of 𝒬 is permutable
with every operation of G
(§ 177). If 𝒬′ is a
sub-group of 𝒬
of order 2α−1,
𝒬′ must therefore be
self-conjugate in G. Now the
order of the factor group G
𝒬′
is 2n;
hence, by the preceding result, it has a self-conjugate sub-group of
order n. It follows that
G has a self-conjugate
sub-group of order 2α−1n.
This group, again, has a self-conjugate sub-group of
order 2α−2n, and so on.
Hence G has a sub-group
of order n. Now by
Theorem VII (§ 248), G has
just n operations
whose divide n. Hence
the sub-group H, of
order n, must be self-conjugate;
and G is the direct product
of the two groups 𝒬
and H.
252. Suppose next that 𝒬 is
not self-conjugate in G;
and let I, of
order 2αn′, be the greatest group
that contains 𝒬 self-conjugately.
Every operation of I is permutable
with every operation of 𝒬; and
therefore (§ 81) G contains
2α−1
distinct sets of conjugate operations whose orders are powers
of 2. The case,
in which 𝒬 is
cyclical, has been already dealt with and will now be excluded; we may thus assume that
𝒬 contains
2β − 1
(β ≮ 2) operations
of order 2, and
that G contains
an equal number of sets of conjugate operations of
order 2.
If possible, let no two sub-groups of
order 2α
have a common operation except identity. Then if two operations
S
and T, of
order 2,
are chosen, belonging to distinct conjugate sets and to different sub-groups of
order 2α, they
cannot be permutable with each other. They will therefore generate a dihedral sub-group
of order 2m. If
m were odd,
S and T
would be conjugate operations. Hence
m must
be even; and the dihedral sub-group must contain a self-conjugate
operation U of
order 2. Since
U is permutable with both
S and T,
it must occur in at least two different sub-groups of
order 2α;
and therefore the supposition, that no two sub-groups of
order 2α
have a common operation except identity, is impossible.
Let now 𝒬′ of
order 2α′ be a sub-group
common to 𝒬
and 𝒬1, two
sub-groups of order 2α;
and suppose that no two sub-groups of
order 2α have a common
sub-group which contains 𝒬′
and is of greater order. Since every operation
of 𝒬 is
self-conjugate in I,
𝒬′ must be
self-conjugate in I.
If then J, of
order 2αn′r, is the greatest
sub-group of G that
contains 𝒬′ self-conjugately,
J contains
r sub-groups of
order 2α. Hence
the factor-group J
𝒬′
contains r sub-groups
of order 2α−α′,
and no two of these have a common sub-group; for if they had, some two sub-groups
of order 2α
contained in J
would have a common sub-group greater than and
containing 𝒬′, contrary to
supposition. The r sub-groups
of order 2α−α′
of J
𝒬′
must therefore be cyclical.
253. We will apply the results of the last paragraph to the case in which all the
operations of 𝒬
are of order 2.
Then 𝒬′ must be
of order 2α−1, since
𝒬
𝒬′ is cyclical. Suppose now
that R is an operation
of J of odd order, which
is permutable with 𝒬′
but not with 𝒬.
If R were
self-conjugate in a sub-group whose order is divisible
by 2α, this sub-group would
contain 𝒬′ and therefore one
or more sub-groups of order 2α
containing 𝒬′. But
R is not permutable with any
sub-group of order 2α that
contains 𝒬′; and therefore
the highest power of 2
that divides the order of the group in which
R is self-conjugate
is 2α−1, so
that R is
one of 2μ
(μ odd) conjugate
operations. If now A is
an operation of order 2
that belongs to 𝒬 and not
to 𝒬′, no operation conjugate
to R can be permutable
with A. Hence in the
substitution group of degree 2μ,
that results on transforming the set of operations conjugate
to R among themselves by
all the operations of G, the
substitution corresponding to A
is an odd substitution. This substitution group has therefore a
self-conjugate sub-group whose order is half its own, and therefore
G has a self-conjugate
sub-group of order 2α−1n.
In the same way it may be shewn that this sub-group has a self-conjugate sub-group of
order 2α−2n; and so
on. Hence G has
a sub-group H
of order n.
But it follows from Theorem VII, § 248, that
G has exactly
n operations whose
orders divide n;
and therefore H is
a self-conjugate sub-group. Moreover, since
G
H is an Abelian
group of order 2α,
it must contain self-conjugate sub-groups of every
order 2r
(r = 1, 2,…,α − 1); and
G therefore has self-conjugate
sub-groups of every order 2rn.
In general however these sub-groups are not characteristic, as is the case when
𝒬 is
cyclical.
254. We will consider next the case where
𝒬 is generated by two operations
A and B,
of orders 2α−1
and 2. There are
in 𝒬 three operations
of order 2, namely,
A2α−2
, A2α−2
B,
and B; and
in 𝒬 there are no operations,
of order greater than 2,
of which the two latter are powers. Suppose that
G contains an
operation B′, conjugate
to B, with which
A2α−2
is not permutable.
Then {A2α−2
,B′} is a dihedral
group, and if R is
an operation of odd order of this group, no power
of A is permutable
with R.
Since A2α−2
and
B′ are not conjugate
operations in G, there must
be an operation of order 2,
conjugate to A2α−2
B, and
permutable with R. The
sub-group, in which {R}
is self-conjugate, therefore contains representatives of each of the three sets of conjugate operations
of order 2 that
belong to G; and
in this sub-group these representatives form three distinct conjugate sets. Hence no operation
conjugate to A2α−2
can be
permutable with R;
and therefore no power, except identity, of an operation conjugate
to A is permutable
with R.
Hence R is
one of 2α−1μ
(μ odd) conjugate operations,
with none of which is A
permutable. It follows that, in the substitution group of
degree 2α−1μ, with which
G is isomorphic,
A is an odd substitution. Hence
G contains a self-conjugate
sub-group of order 2α−1n; and since
this is of the same type as G,
the same reasoning may be applied to it. Exactly as before,
G will contain self-conjugate
sub-groups of every order 2rn
(r = 0, 1,…,α − 1).
We have assumed that there is an
operation B′,
conjugate to B,
with which A2α−2
is not
permutable. If A2α−2
were
permutable with every operation that is conjugate
to B, the sub-group
in which A2α−2
is permutable would contain a self-conjugate
sub-group G′
of G whose order is divisible
by 2. In this case, we may
deal with G
G′ exactly as we
have been dealing with G.
The results of §§ 250–254 may be summed up in the following form:—
THEOREM IX. If in a group G
of order 2αn,
where n is odd,
a sub-group 𝒬 of
order 2α is an Abelian
group of type (α),
(α − 1, 1), or
(1, 1,…, 1), and if
𝜃(𝒬) and
n are relatively prime, then
G contains self-conjugate
sub-groups of each order 2βn
β = 0, 1,…,α − 1
.
255. Still representing the order
of G
by 2αn,
where n is
odd, there are two cases, in which the sub-groups of
order 2α
are not Abelian, where it may be shewn without difficulty that
G contains a self-conjugate
sub-group of order n.
THEOREM X. If the order of G
is 2αn, where
n is odd and not
divisible by 3, and
if a sub-group 𝒬
of order 2α
is of the type
(i) A2α−1
= 1, B2 = A2α−2
, B−1AB = A−1,
or (ii) A2α−1
= 1, B2 = 1, B−1AB = A−1,
G contains a self-conjugate
sub-group of each order 2βn
β = 0, 1,…,α − 1
.
For each of these types, 𝜗(𝒬) is
3 or 1;
and if 𝒬′ is any
sub-group of 𝒬,
𝜗(𝒬′) is either
3 or 1.
Hence if any operation of G
of odd order is permutable with
𝒬 or 𝒬′,
it must be permutable with every operation of
𝒬 or 𝒬′.
In each type, the self-conjugate operations form a cyclical sub-group of
order 2,
namely {A2α−2
}.
We will consider first the case where
𝒬 is of type (i). In
this case, {A2α−3} and
{B} are sub-groups
of G of
order 4, the former
being self-conjugate in 𝒬
while the latter is not. If they are conjugate sub-groups
in G, we have seen (§ 82)
that G must contain an
operation of odd order S,
such that the sub-groups S−n{B}Sn
(n = 0, 1, 2,…) are
permutable with each other. This is impossible, since every operation that is permutable with
a sub-group 𝒬′
is permutable with all its operations. Hence
{A2α−3
} and
{B} are not conjugate sub-groups; and
A2α−3
and B
are not conjugate operations.
Now A2α−3
is self-conjugate
in {A} and is one of two
conjugate operations in 𝒬;
therefore in G it must be one
of 2μ conjugate operations,
where μ is odd. When these
2μ conjugate operations
are transformed by A2α−3
,
two only, namely, A2α−3
and A−2α−3
,
remain unchanged. Let us suppose that, in the resulting substitution, there are
x cycles of
2 symbols
and y cycles
of 4 symbols.
When the 2μ conjugate
operations are transformed by B,
none can remain unchanged; and we may suppose that the resulting substitution has
x′ cycles of
2 symbols
and y′ cycles
of 4 symbols.
Now since
A2α−2
= B2,
y = y′,
and therefore
1 + x = x′.
Hence one of the two substitutions
A2α−3
and
B must be odd;
G has therefore a self-conjugate
sub-group of order 2α−1n.
The groups of order 2α−1
contained in this self-conjugate sub-group are either of type (i), with
α − 1 written
for α: or they are
cyclical; for, like 𝒬,
they can only contain a single operation of
order 2. If
they are cyclical, the reasoning of § 250 may be applied; and if they are of
type (i) the same reasoning will apply to the self-conjugate sub-group of
order 2α−1n that has been
used for G. Finally, then,
G must contain sub-groups of orders
n, 2n,
22n, …,
2α−1n,
each of which is contained self-conjugately in the next; and therefore the
sub-group H
of order n
must be self-conjugate.
If 𝒬 is of
type (ii), A is one
of 2μ conjugate
operations in G,
where μ is
odd. It may be shewn, as in the previous case, that
B and
A2α−2
cannot be
conjugate in G,
so that no one of the operations conjugate to
A has B
for one of its powers. If now B were
permutable with any one of these
2μ conjugate
operations, the group would contain an Abelian sub-group of
order 2α,
which is not the fact. Hence the substitution, given on transforming the
2μ operations
by B, consists of
μ transpositions
and is an odd substitution. Therefore
G contains a self-conjugate
sub-group of order 2α−1n, in which
the sub-groups of order 2α−1
are cyclical. Hence, again, there is a self-conjugate sub-group of
order n.
256. The only non-Abelian groups of
order 23
are those of types (i) and (ii) of § 255, when
α = 3. The Abelian groups of order
22 and 23
are all included in the types considered in Theorem IX, § 254. For an Abelian
group 𝒬,
of order 23
and type (2, 1),
𝜃(𝒬) is 3;
and for one of order 23
and type (1, 1, 1),
𝜃(𝒬) is 21.
Hence Theorems IX and X shew that, if
2, 22
or 23 divide the order
of a group, but not 24,
then the operations of odd order form a self-conjugate sub-group, with possible exceptions
when 12 or
56 is a
factor of the order. Hence:—
THEOREM XI. A group of even order cannot be simple unless
12, 16
or 56 is
a factor of the order102.
257. In further illustration of the methods of the preceding
paragraphs, we will deal with another case in which the sub-groups of
order pα
are not Abelian.
Let G be a group
of order pαm, where
m is relatively prime
to p(p −1)…(pα −1); and suppose
that a sub-group H
of order pα
is such that within it every operation is either self-conjugate or one of
p conjugate operations103.
Let h be a sub-group
of G whose order is
a power of p; and
let S be an operation
of G whose order is
not divisible by p.
Then since m is
relatively prime to p(p −1)…(pα −1),
if S is permutable
with h, it is permutable
with every operation of h.
Hence it follows from § 82 that no operation
of H, which is not
self-conjugate in H, can
be conjugate in G to a
self-conjugate operation of H.
Suppose now that S1 and
S2 are two operations
of H, each of which
in H is one of
p conjugate operations;
and that while S1 and
S2 are conjugate
in G, they are not
conjugate in I, the greatest
group that contains H
self-conjugately. Let H1 and
H2 be the
sub-groups of H,
of orders pα−1, which
contain S1 and
S2 self-conjugately.
Since S1 and
S2 are conjugate
in G, all the
sub-groups of order pα−1
which contain either of them self-conjugately belong to the same conjugate set. If then
H1 and
H2
are identical, there must be an operation, of order prime
to p, which will
transform S1
into S2 and
H1 into itself. This, we have
seen, is impossible. If H1 and
H2
are not identical, there must be an operation which will transform
S1 into
S2 and
H1 into
H2. Now
H1 and
H2 both contain the
sub-group h formed of the
self-conjugate operations of H; and
since no self-conjugate operation of H
is conjugate in G with
any operation of H
which is not self-conjugate, the operation in question, which transforms
S1 into
S2 and
H1 into
H2, must transform
h into itself.
Hence S1 and
S2 are conjugate operations
in that sub-group, G′,
of G which contains
h self-conjugately. Now in
G′
h the operations, that correspond to
S1 and S2,
are self-conjugate operations of H
h .
Since then these operations are conjugate
in G′
h , they must (§ 82)
be conjugate in I
h.
This however is impossible, since every operation
of I
h whose order
is a power of p
is self-conjugate. Finally, then, no two operations
of H, which are not
conjugate in H, can
be conjugate in G;
and the number of distinct sets of conjugate operations
in G, whose orders are
powers of p, is equal
to the number in H.
From this it follows, as in previous cases, that the number of operations
of G whose
orders divide m
is equal to m.
258. THEOREM XII. The only simple groups, whose orders
are the products of four or of five primes, are groups of orders
60, 168,
660,
and 1092:
and no group, whose order contains less than four prime factors, is simple.
Groups, whose orders are
p1, p1p2,
p12,
p13,
p12p2,
p1p22
or p1p2p3
are all proved to be soluble by previous theorems in the present chapter.
If the order of a group contains four prime factors, it must be of one of the forms
p14, p13p2,
p12p22,
p12p2p3,
p1p23,
p1p22p3,
p1p2p32
or p1p2p3p4. If
p1 is an odd
prime, groups of any one of these orders have already been shewn to be soluble; while if
p1 is 2,
the only case which can give a simple group
is 22p2p3. If
a group of this order is simple, it follows from Theorem VIII, Cor. (§ 249), that
p2 must
be 3; and the order of
the group 12p3. A cyclical
sub-group of order p3 must
be one of 1 + kp3 conjugate
sub-groups. Hence 1 + kp3
must be a factor of 12,
so that p3 is either
5 or 11.
If p3 were 11,
the 12 conjugate
sub-groups of order 11
would contain 120 distinct
operations of order 11,
and the tetrahedral sub-group, which the group (if simple) must contain, would be self-conjugate.
Hence p3 must
be 5 and the order
of the group is 60.
We have already seen that a simple group of
order 60
actually exists, namely, the icosahedral group; and that there is only one type for
such a group (§ 85).
If the order of a group contains five prime factors, it must be of one of the
forms:—
p15,p
14p
2,p13p
22,p
13p
2p3,p12p
23,p
12p
22p
3,p12p
2p32,p
12p
2p3p4,p1p24,
p1p23p
3,p1p22p
32,p
1p22p
3p4,p1p2p33,p
1p2p32p
4,p1p2p3p42,p
1p2p3p4p5.
If p1 is
an odd prime, it follows from previous theorems that none of these forms, except
p13p2p3 and
p1p23p3, can
give simple groups.
Taking first the order p1p23p3,
let us suppose (without limitation to the particular case) that a group of
order p1p2mp3 is simple. It
must contain p3 or
p1p3 conjugate
sub-groups of order p2m.
In either case, the operations of these sub-groups must be all distinct, or else they must
all have a common sub-group, for the order of the group contains no factor congruent
to unity (modp2),
except p3 or
p1p3. Now the group
contains (§ 248) just p2mp3
operations whose orders are not divisible
by p1; and if there
are p3 sub-groups
of order p2m
whose operations are all distinct, there are
(p2m − 1)p3 operations whose orders are
powers of p2. Hence, in this
case, a sub-group of order p3
is self-conjugate. On the other hand, if there are
p1p3 sub-groups of
order p2m, their
operations cannot be all distinct; and the group has a self-conjugate sub-group whose order
is a power of p2.
A repetition of this reasoning shews that a group of
order p1p2mp3
is always soluble.
We consider, next, a group whose order
is p13p2p3. If
p1 is odd,
p12 cannot be
congruent to unity, (modp2)
or (modp3). The same
is true, when p1 = 2,
if p2 is not
equal to 3.
We shall therefore first deal with a group of
order p13p2p3 on the
supposition that p2 is
not 3.
If neither p2 nor
p3
divides p13 − 1,
it follows from §§ 248, 257 that there are just
p2p3 operations in the group
whose orders divide p2p3.
When this is the case, the group clearly cannot be simple.
If p3 divides
p13 − 1, and if there are more than
p2p3 operations whose orders
divide p2p3, a sub-group of
order p13 must be self-conjugate
in a sub-group of order p13p3,
so that the group is isomorphic with a group of
degree p2.
In this case, again, the group cannot be simple.
Finally, then, we have only to deal with the case in which a sub-group of
order p13 is self-conjugate
in a sub-group of order p13p2,
while in this sub-group an operation of
order p2
is one of p13
conjugate operations. Now the congruences
p13 ≡ 1(modp
2),
p2 ≡ 1(modp1),
are inconsistent; therefore, if a sub-group of
order p2 is self-conjugate
in a sub-group of order p1αp2,
the latter must be Abelian. Hence the group contains
p13(p2 − 1)p3 operations whose orders
are divisible by p2. Now if
the sub-groups of order p13
have a common sub-group, it must be common to all of the sub-groups of
order p13, and it is
a self-conjugate sub-group. If, however, no two have a common sub-group, the group contains
(p13 − 1)p3 operations of
order p1; and there
remain only p3 operations
whose orders divide p3.
The group is therefore in any case composite.
If now p2 = 3, a
group of order 233p3
has, by Sylow’s theorem, a self-conjugate sub-group of
order p3
unless p3 is
5, 7, 11
or 23. If
p3 = 23, the group (if simple)
would have just 24 operations
whose orders divide 24;
it is thence easily seen to be non-existent. If
p = 11,
the group (if simple) could be expressed as a doubly transitive group of
degree 12.
In this form, however, each operation, which transforms an operation of
order 11
into its own inverse, would be a product of
5 transpositions
and therefore an odd substitution. This group therefore cannot be simple. If
p3 = 5,
the group could be expressed as a doubly transitive group of
degree 6. The sub-group of
order 4 which transforms
a sub-group of order 5
into itself would be cyclical; and the corresponding substitution being odd, the
group could not be simple.
Hence, finally, the only possibility is a group of
order 168.
That a simple group of this order actually exists is shewn in § 146; also, there is
only one type of such group.
When p1 is
equal to 2,
the only case, that requires discussion in addition to those we have dealt with, is
p12p2p3p4.
A group of this order can only be simple (§ 249) when it
contains a tetrahedral sub-group. In this case, the operations of
order 2
form a single conjugate set, and the group contains just
3p3p4 operations whose orders
are divisible by 2. If a
sub-group of order 2α3q
is the greatest that contains a sub-group of
order 3 self-conjugately, the
group will contain 22−α(3 − 1)p3p4 operations
whose orders are divisible by 3
and not by 2.
Hence the group must contain either
p3p4, 5p3p4,
or 7p3p4, operations
whose orders divide p3p4.
On the other hand, the number of these operations may be expressed in the form
x(p3 − 1)p4 + y(p4 − 1) + 1,
where
x is a factor
of 12, and
y is a factor
of 12p3 which is
congruent to unity, (modp4).
Hence
x(p3 − 1)p4 + y(p4 − 1) = zp3p4 − 1,(z = 1, 5, 7).
The case z = 1 leads
to y = 1, so that the
sub-group of order p4
is self-conjugate.
The case z = 5 gives no
solution; but when z = 7,
it will be found that there are two solutions, namely,
x = 6, y = 12, p3 = 5, p4 = 11,
andx = 6, y = 14, p3 = 7, p4 = 13.
That simple groups actually exist corresponding to the two orders
660 and
1092 thus
arrived at, is shewn in § 221. There is also in each case a single type; the
verification of this statement is left to the reader.
259. As has already been stated at the beginning of the present chapter, the
solution of the general problem of pure group-theory, namely, the determination of
all possible types of group of any given order, depends essentially on the previous
determination of all possible simple groups. A complete solution of this latter
problem is not to be expected; but for orders, which do not exceed some given
limit, the problem may be attacked directly. The first determination of this kind
was due to Herr O. Hölder104, who examined all possible orders up
to 200: he proved that none
of them, except 60 and
168, correspond
to a simple group. Mr Cole105 continued the investigation, by examining all orders
from 201
to 660,
with the result of shewing that in this interval the only
orders, which have simple groups corresponding to them, are
360, 504
and 660. The existence of a
simple group of order 504
had not been recognised before Mr Cole’s investigation.
The author106 has carried on the examination from
661
to 1092, with the result
of shewing that 1092 is
the only number in this interval which is the order of a simple group.
As the limit of the order is increased, such investigations as these
rapidly become more laborious, since a continually increasing number
of special cases have to be dealt with. There is little doubt however
but that, with the aid of the theorems proved in the present chapter,
the investigation might be continued without substantial difficulty up
to 2000.
We shall here be content with verifying the result of
Herr Hölder’s and Mr Cole’s investigations for orders up
to 660.
The method used in dealing with particular numbers may suggest to the reader
how the determination might be continued.
260. It follows from Theorem XII that, if an odd number can
be the order of a simple group, it must be the product of at least
6 prime
factors. Moreover, by previous theorems, we have seen that
36, 35p
and 34p2
cannot be orders of simple groups. Hence certainly no odd number less than
34 ⋅ 5 ⋅ 7
or 2835107 can be
the order of a simple group. Therefore, by Theorem XI, we need only examine numbers less
than 660 which are divisible by
12, 16,
or 56. When each number less
than 660 which is divisible by
12, 16,
or 56 is written
in the form p1α1p2α2…,
it will be found that, with eleven exceptions, Theorems II to XII of the present chapter
immediately shew that there are no simple groups corresponding to them. The exceptions are
60, 168,
240,
336,
360,
480,
504,
528,
540,
560, 660.
We will first deal with such of these numbers as do not actually correspond to
simple groups.
240 = 24 ⋅ 3 ⋅ 5.
A simple group of this order would contain
6 or
16 sub-groups of
order 5. If it contained
6 sub-groups
of order 5,
the group could be expressed as a transitive substitution group of
degree 6; and there is
no group of degree 6
and order 240. If there
were 16 sub-groups
of order 5,
each would be self-conjugate in a group of
order 15,
which must be cyclical. The group then would be a doubly transitive group of
degree 16
and order 16 ⋅ 15.
Such a group (§ 105) contains a self-conjugate sub-group of
order 16.
336 = 24 ⋅ 3 ⋅ 7.
A simple group of this order would contain
8 sub-groups of
order 7, each self-conjugate
in a group of order 42.
We have seen (§ 146) that there is no simple group of
degree 8
and order 8 ⋅ 7 ⋅ 6.
480 = 25 ⋅ 3 ⋅ 5.
Since 25 is not a
factor of 5!, a group of
order 480 must, if simple,
contain 15 sub-groups
of order 25. Now
25 is not congruent
to unity, (mod 4),
and therefore (§ 78) some two sub-groups of
order 25 must have a common
sub-group of order 24. Such
a sub-group, of order 24,
must (§ 80) either be self-conjugate, or it must be contained self-conjugately in a sub-group
of order 25 ⋅ 3 or
25 ⋅ 5. In
either case the group is composite.
528 = 24 ⋅ 3 ⋅ 11.
There must be 12 sub-groups
of order 11,
each being self-conjugate in a group of
order 44. A group of
order 44 necessarily contains
operations of order 22,
and such an operation cannot be represented as a substitution of
12 symbols.
The group is therefore not simple.
540 = 22 ⋅ 33 ⋅ 5.
There must be 10 sub-groups
of order 33, each self-conjugate
in a group of order 2 ⋅ 33. There
must also be 36 sub-groups of
order 5, each self-conjugate
in a group of order 15.
Now when the group is expressed as transitive in
10 symbols, the
substitutions of order 5
consist of 2 cycles
of 5 symbols
each, and no such substitution can be permutable with a substitution of
order 3.
Hence the group is not simple.
560 = 24 ⋅ 5 ⋅ 7.
There must be 8 sub-groups
of order 7. A transitive
group of degree 8
and order 8 ⋅ 7 ⋅ 10
does not exist (§ 146).
That there are actually simple groups of orders
60, 168,
360,
and 660,
has already been seen; the verification that there is only a single type of simple group
of order 360
may be left to the reader. It remains then to consider the
order 504.
504 = 23 ⋅ 32 ⋅ 7.
A simple group of this order must contain
8 or
36 sub-groups of
order 7. There is no
group of degree 8 and
order 504 (§ 146). Hence
there must be 36 sub-groups
of order 7.
Again, there must be 7 or
28 sub-groups of
order 32; and since there
is no group of degree 7
and order 504 (l.c.), there
are 28 sub-groups of
order 32. If two of these have
a common operation P
of order 3,
it must (§ 80) be self-conjugate in a group containing more than one sub-group of
order 32. If the order of this
sub-group is 22 ⋅ 32, it must
contain a sub-group of order 22
self-conjugately; and therefore this sub-group of
order 22 must be self-conjugate
in a group of order 23 ⋅ 32
at least. Such a group would be one of
7 conjugate
sub-groups, and this is impossible. Similarly,
P cannot
be self-conjugate in a group whose order is greater
than 22 ⋅ 32. Hence no two
sub-groups of order 32
have a common operation, except identity.
There are therefore in the group
216 operations of
order 7, and
224 operations whose
orders are powers of 3,
leaving just 64
other operations.
Suppose now that A is an
operation of order 2, self-conjugate
in a sub-group of order 23.
If A is self-conjugate
in a group of order 23x
(x being a factor
of 32 ⋅ 7), this group must contain at
least x operations of odd order
and therefore at least x operations
of the form AS, where
S is an operation of odd
order permutable with A.
There are also x operations
of this form, corresponding to each of the
63
x operations conjugate
to A; so that the whole
group contains 63 operations
of the form AS, where
S is of odd order and
permutable with A, while
A is one of a set of conjugate
operations of order 2,
each of which is self-conjugate in a sub-group of
order 23.
Hence taking the identical operation with these
63 operations
of even order, all the operations of the group are accounted for;
and there are therefore no operations whose orders are powers
of 2
except those of the conjugate set to which
A belongs. The
sub-groups of order 23
are therefore Abelian groups whose operations are all of
order 2; and since
the 7 operations of
order 2, in a group of
order 23, are all conjugate
in the group of order 504,
they must (§ 81) be conjugate in the sub-group within which the group of
order 23
is self-conjugate. Hence, finally, there must be
9 sub-groups of
order 23, each self-conjugate
in a group of order 23 ⋅ 7, in
which the 7 operations
of order 2
form a single conjugate set.
A simple group of order 504,
if it exists, must therefore be expressible as a triply transitive group of
degree 9, in which the
sub-group of order 56
that keeps one symbol fixed is of known type. From this, several inferences can be drawn. Firstly,
all the 63 operations
of order 2 of the
9 sub-groups of
order 23 must be distinct, since
each operation of order 2
keeps just one symbol fixed. There is therefore a conjugate set of
63 operations
of order 2,
and there are no other operations of even order in the group. Secondly, the
224 operations whose orders
are powers of 3 cannot all be
operations of order 3. For, if they
were, there would be 8 operations
of order 3
containing any three given symbols in a cycle; and therefore the
group would contain operations other than identity which keep
3 symbols
fixed. This is not the fact; the sub-groups of
order 32
are therefore cyclical.
Conversely, we may now shew that such a triply transitive group of
degree 9,
if it exists, is certainly simple. The distribution of its operations in conjugate sets
has, in fact, been determined as follows. There are:—
3 conjugate sets of operations of order 7 each containing 72 operations
3 ” ” 9 ” 56 ”
1 ” ” 3 ” 56 ”
1 ” ” 2 ” 63 ”
A self-conjugate sub-group, that contains a single operation of
order 7,
must contain all of them; and such a sub-group, that contains a single operation of
order 9,
must contain all the operations of orders
3 and 9.
Hence if a self-conjugate sub-group contains operations of
order 9, its order must be of the form
1 + 224 + 216x + 63y;
and if it contains no operations
of order 9, its order must be of
the form
1 + 56x′ + 216y′ + 63z′;
each of the symbols
x, y,
x′, y′, z′
being either zero or unity. The order must at the same time be a factor
of 504. A
very brief consideration will shew that all these conditions cannot be satisfied, and
therefore that no self-conjugate sub-group exists.
That there is a triply transitive group of
degree 9
and order 9 ⋅ 8 ⋅ 7,
has already been seen in § 113; and the actual formation of the substitutions
generating this group will verify that it satisfies the conditions just obtained.
We may however very simply construct the group by the method of § 109; and
this process has the advantage of shewing at the same time that there is only one
type.
There is no difficulty in constructing the sub-group of
order 56 as a doubly transitive
group of 8 symbols. If we denote by
s and t
the two substitutions
(1254673) and(12)(34)(56)(78),
{s,t}
is, in fact, a group of the desired type. If now
A is any substitution
of order 2 in
the symbols 1,
2, …,
8, 9
which contains the transposition (89),
and satisfies the conditions
AsA = S1,AtA = S2AS3,
where
S1, S2, S3
belong to {s,t}: then (§ 109)
it follows that {A,s,t} is a triply
transitive group of degree 9
and order 504.
Now every operation of order 2
in the group, which interchanges
8 and 9,
must transform s into its
inverse. Also from the 9 symbols
only 7 operations
of order 2,
satisfying these conditions, can be formed; and if one of them belongs to the
group, they must all belong to it. Hence there cannot be more than one type of
group satisfying the conditions.
Finally, if
A = (89)(23)(46)(57),
then
AsA = s−1,
and
AtA can be expressed
in the form
(1635842)A(1685437),
where
the two operations (1635842)
and (1685437) belong to
{s,t}. It follows, then, that
{A,s,t} is a triply transitive
group of degree 9
and order 504,
which satisfies all the conditions, and that there is only a single type of such
group.
261. The order of a non-soluble group must be divisible
by the order of some non-cyclical simple group. Moreover, if
N is
the order of a non-cyclical simple group and if
n is
any other number, there is always at least one non-soluble group of
order Nn; for,
G being a simple
group of order N
and H any group of
order n, the direct product of
G and H
is a non-soluble group of order Nn.
Herr Hölder108 has determined all distinct types of non-soluble groups whose orders are
less than 480.
He shews that, besides the three simple groups of orders
60, 168,
and 360, there are
22 non-soluble
groups whose orders are given by the following table.
|
|
|
|
|
|
|
120 | 180 | 240 | 300 | 336 | 360 | 420 |
|
|
|
|
|
|
|
3 | 1 | 8 | 1 | 3 | 5 | 1 |
|
|
|
|
|
|
|
|
For the proof of all these results except the first, and for the very
interesting and suggestive methods that lead to them, the reader is
referred to Herr Hölder’s memoir. The case of a non-soluble group of
order 120
is susceptible of simple treatment; and we will consider it here as exemplifying
how composite groups may be constructed when their factor-groups and the
isomorphisms of the latter are completely known.
262. The composition-factors of a non-soluble
group G of
order 120 are
2 and 60.
If these may be taken in either order, then
G contains a self-conjugate
sub-group of order 2 and a
self-conjugate sub-group of order 60.
The latter, being simple, must be an icosahedral group and cannot contain the former. Hence, in
this case, G must
(§ 34) be the direct product of an icosahedral group and a group of
order 2.
Next, suppose that the composition-factors can only be taken in the order
2, 60.
Then G contains a
self-conjugate sub-group H
of icosahedral type and no self-conjugate sub-group of
order 2. Hence if
S is any operation
of G which is not contained
in H, the isomorphism
of H which arises on
transforming its operations by S
must (§ 165) be contragredient. It follows that
G is simply isomorphic with a
group of isomorphisms of H.
Now H is
isomorphic with the alternating group of degree five, and its group of isomorphisms is
therefore (§ 173) simply isomorphic with the symmetric group of degree five. Hence, in this
case, G is
simply isomorphic with the symmetric group of degree five.
Lastly, suppose that the composition-factors can only be taken in the order
60, 2.
Then G has a self-conjugate
sub-group of order 2 and no
self-conjugate sub-group of order 60.
Hence (§ 35) G has no
sub-group of order 60. Suppose
that A is the self-conjugate
operation of order 2, and
arrange the operations of G
in the sets
1,A;S1,S1A;S2,S2A;…;S59,S59A.
Since G
{A}is
simply isomorphic with an icosahedral group, it must (§ 203) be possible to choose three
sets
S′,S′A;S′′,S′′A;S′′′,S′′′A;
such that either operation of the first, multiplied by either operation of the
second, belongs to the third set; while at the same time the cube of either
operation of the first set, the fifth power of either operation of the second
set, and the square of either operation of the third set, belong to the set
1, A.
Now if
S′3 = 1,
then
(S′A)3 = A.
Hence we may assume, without loss of generality, that
S′3 = 1,S′′5 = 1,S′S′′= S′′′orS′′′A.
If S′′′ is of
order 2, so
also is S′′′A, and
G would then contain an
icosahedral sub-group. Hence
S′′′4 = 1,
and
S′′′2 = A.
Now
S13 = 1,S
25 = 1,(S
1S2)2 = 1,
is a complete set of defining relations for the icosahedral group; so that, from the symbols
S1 and S2,
exactly 60
distinct products can be formed. Hence from
S′, S′′,
and A,
where
S′3 = 1,S′′5 = 1,(S′S′′)2 = A,A2 = 1,
A being permutable
both with S′
and S′′,
exactly 120
distinct products can be formed. It follows that the relations
just given are the defining relations of a non-soluble group of
order 120, which has a self-conjugate
sub-group of order 2 and
no sub-group of order 60;
and that there is only one type of such group. We have already seen (§ 221)
that there must be at least one such type of group; viz. the group defined by
x′ ≡αx1 + βx2,
x′ ≡γx1 + δx2,
αδ −βγ ≡1,
(mod5).
Ex. If p is
an odd prime, shew that a group, whose composition-factors are
60 and p,
must be the direct product of an icosahedral group and a group of
order p.
(Hölder.)
263. Let G be
a composite group; and suppose that the factor-groups of the composition-series
of G
are the two non-cyclical simple groups
H′ and H
in the order given. Then (§ 165) if
H1 is the group formed
of all the operations of G
which are permutable with every operation
of H, the direct
product {H,H1}
of H and
H1 is a self-conjugate
sub-group of G. Now the
composition-factors of {H,H1} are
composition-factors of G.
Hence H1 must either
be isomorphic with H′,
or it must reduce to the identical operation. The latter alternative can only occur if
L
H has a sub-group simply
isomorphic with H′,
L being the group of
isomorphisms of H.
In the cases of the simple groups whose groups of isomorphisms have been investigated
in Chapter XI, i.e. the alternating groups, the doubly transitive groups of
degree pn + 1 and order
1
2pn(p2n − 1), and the triply transitive
groups of degree 2n + 1
and order 2n(22n − 1),
L
H was found
to be an Abelian group. (These groups include all the simple groups whose orders do not
exceed 660.)
If H and
H′ belong to these classes
of groups, then G must
be the direct product of simple groups simply isomorphic with
H and H′.
For instance, a group of order 3600,
whose composition-factors are
60 and 60,
must be the direct product of two icosahedral groups.
This result may clearly be extended to the case of a group, the factor-groups
of whose composition-series are a number of non-cyclical simple groups
H1,
H2, …,
Hn,
such that no two of them are of the same type, while for each of them
L
H is
soluble. Such a group must be the direct product of
n simple groups which
are isomorphic with H1,
H2, …,
Hn.
If, however, several of these groups are of the same type, the inference is no longer necessarily
true. Thus if G is
the direct product of five icosahedral groups, and if
L is the group of
isomorphisms of G,
the order of L
G
is 25 ⋅ 5!,
and this group contains icosahedral sub-groups. A group of
order (60)6,
the factor-groups of whose composition-series are all of
order 60,
is not therefore necessarily the direct product of
6 icosahedral
groups.
Note to § 257.
The property stated in the footnote on p. 1050 may be proved as follows. Let
S1 and S2
be two operations of the group which are not permutable with each other; and suppose that
S2−1S
1S2 = S1Σ.
The sub-group of
order pn−1, which contains
S1 self-conjugately, also
contains S1Σ self-conjugately
(§ 55); therefore it contains Σ
self-conjugately. Similarly the sub-group, which contains
S2 self-conjugately,
contains S2Σ−1 and therefore
Σ self-conjugately.
Hence Σ,
being contained self-conjugately in two distinct sub-groups of
order pn−1,
is a self-conjugate operation. From this it follows at once that
H
h is
Abelian.
Note to § 258.
The statement, on p. 1056, that the congruences
p13 ≡1(modp2) and
p2 ≡1(modp1) are inconsistent, is subject
to an exception if p12 + p1 + 1 is a prime.
When this is the case and when p2 = p12 + p1 + 1,
a group of order p12p2p3
can only be simple, under the conditions assumed in the text, if it contains
(p13 −1)p3 operations
of order p1,
p12(p2 −1)p3 operations
of order p2,
p12p2(p3 −1) operations
of order p3,
and no other operation except identity. The relation
p13p
2p3 = (p13 −1)p
3 + p12(p
2 −1)p3 + p12p
2(p3 −1) + 1
cannot however be satisfied; the group is therefore composite.
Note to § 260.
No simple group of odd order is at present known to exist. An investigation as to the
existence or non-existence of such groups would undoubtedly lead, whatever the
conclusion might be, to results of importance; it may be recommended to the reader as
well worth his attention. Also, there is no known simple group whose order contains
fewer than three different primes. This suggests that Theorems III and IV, §§ 243, 244,
may be capable of generalisation. Investigation in this direction is also likely to lead to
results of interest and importance.
Appendix.
T he technical phraseology that has been used in this book is borrowed almost entirely
from French or German; and far the greater number of important memoirs on the
subject of finite groups are written in one or the other of those languages. To enable
the reader to refer, with as little trouble as possible, to the writings of foreign
mathematicians, a table is here given of the French and the German equivalents
of the more important technical terms. Even abroad, the phraseology of the
subject has not yet arrived at that settled state in which every writer uses
a technical term in the same sense; but the variations of usage are not very
serious. In his recently published Lehrbuch der Algebra, however, Herr Weber
has introduced or adopted several deviations from ordinary usage; the chief of
these are noted, by the addition of his name after the term, in the subjoined
table.
Group |
Groupe |
|
Gruppe |
Abelian group |
Groupe des opérations
échangeables |
|
Abel’sche Gruppe |
Alternating group |
Groupe alterné |
|
Alternierende Gruppe |
Complete group |
Vollkommene Gruppe |
Composite group |
Groupe composé |
|
Zusammengesetzte Gruppe |
Factor-group |
Groupe facteur |
|
Factorgruppe |
Primitive or imprimitive group |
Groupe primitif ou non-primitif |
|
Primitive oder imprimitive Gruppe |
Transitive or intransitive group |
Groupe transitif ou intransitif |
|
Transitive oder intransitive Gruppe |
Simple group |
Groupe simple |
|
Einfache Gruppe |
Soluble group |
Groupe résoluble |
|
Auflösbare Gruppe |
|
Metacyclische Gruppe109 (Weber) |
| |
Substitution group |
Groupe
des substitutions or système des
substitutions conjuguées |
|
Substitutionengruppe |
Symmetric group |
Groupe symétrique |
|
Symmetrische Gruppe |
Group of isomorphisms |
Gruppe der Isomorphismen |
Sub-group |
Sousgroupe |
|
Untergruppe |
|
Theiler (Weber) |
Conjugate sub-groups |
Sousgroupes conjugués |
|
Gleichberechtigte Untergruppen |
|
Conjugirte Theiler (Weber) |
Self-conjugate sub-group |
Sousgroupe invariant |
|
Ausgezeichnete Untergruppe or
invariante Untergruppe |
|
Normaltheiler (Weber) |
Maximum self-conjugate sub-group |
Sousgroupe invariant maximum |
|
Ausgezeichnete or invariante
Maximaluntergruppe |
|
Grösster Normaltheiler (Weber) |
Characteristic sub-group |
Charakteristische Untergruppe |
|
|
Composition-series |
Suite de composition |
|
Reihe der Zusammensetzung |
|
|
Compositionsreihe (Weber) |
Composition factors |
Facteurs de composition |
|
Factoren der Zusammensetzung |
Chief series |
Hauptreihe |
Characteristic series |
Lückenlose Reihe charakteristischer
Untergruppen |
|
|
Simple isomorphism |
Isomorphisme holoédrique |
|
Holoëdrischer or einstufiger
Isomorphismus |
Multiple isomorphism110 |
Isomorphisme meriédrique |
|
Meriëdrischer or mehrstufiger
Isomorphismus |
Isomorphism of a group with itself |
Isomorphismus der Gruppe in sich |
Cogredient or contragredient
isomorphism |
Cogredient oder contragredient
Isomorphismus |
|
|
Degree (of a group) |
Degré |
|
Grad |
Order ” ” |
Ordre |
|
Ordnung |
|
Grad (Weber) |
Genus ” ” |
Geschlecht |
|
|
Substitution |
Substitution |
|
Substitution or
Buchstabenvertauschung |
Circular substitution |
Substitution circulaire |
|
|
Cirkularsubstitution |
Even and odd substitutions |
Substitutions positives et négatives |
|
Gerade und ungerade Substitutionen |
Regular substitution |
Substitution régulière |
|
Reguläre Substitution |
Similar substitutions |
Substitutions semblables |
|
Ähnliche Substitutionen |
Transposition |
Transposition |
|
Transposition |
|
|
|
|
|
|
|
One other term, introduced by Herr Weber, for which no simple equivalent is used in
this book, must be mentioned. The greatest sub-group which is common to two groups
he calls the “Durchschnitt” of the two groups.
The ratio of the order of a sub-group H,
to the order of the group G
containing it, is called by French writers the “indice,” by German the “Index,”
of H
in G.
The phrase is most commonly used of a substitution group in relation to the symmetric
group of the same degree.
The smallest number of symbols displaced by any substitution, except identity, of a
substitution group is called by French writers the “classe,” by Germans the “Klasse,” of
the group.