up | next | prev | ptail | tail |
H aving in the last chapter dealt in some detail with Abelian groups of order , where is a prime, we shall now investigate some of the more important properties of groups, which have the power of a prime for their order but are not necessarily Abelian. Besides illustrating and leading to many interesting applications of the general theorems of Chapter III, the discussion of groups, whose order is the power of a prime, will be found in many ways to facilitate the subsequent discussion of other groups, whose order is not thus limited.
52. If is a group whose order is , where is a prime, the order of every sub-group of must also be a power of ; and therefore (§ 25) the total number of operations of which are conjugate with any given operation must be a power of . The identical operation of is self-conjugate. Hence if the operations of , other than the identical operation, are distributed in conjugate sets containing , , … operations, the total number of operations of is ; and therefore
This equation can only be true if of the indices , , … are zero, being some integer not less than unity. Hence must contain self-conjugate operations, which form (§ 27) a self-conjugate sub-group14. Hence:—
THEOREM I. Every group whose order is the power of a prime contains self-conjugate operations, other than the identical operation; and no such group can be simple.
53. Let be the sub-group of of order , formed of its self-conjugate operations. Then is a group of order , and it must contain self-conjugate operations other than identity, forming a self-conjugate sub-group of order . Corresponding to this self-conjugate sub-group of , must have a self-conjugate sub-group of order , which contains . If , is a group of order ; and it again will have self-conjugate operations forming a sub-group of order . Corresponding to this, will have a self-conjugate sub-group of order , which contains . If the process thus indicated is continued, till itself is arrived at, a series of groups
is formed, each of which is self-conjugate in and contains all that precede it15. Moreover is the group formed of the self-conjugate operations of , and it is therefore an Abelian group.The series of sub-groups thus arrived at is of considerable importance. From the mode of formation, it is clear that a second group of order cannot be simply isomorphic with , unless and are simply isomorphic for each value of . It by no means however follows that, when this latter condition is satisfied, and are always simply isomorphic. This is indeed not necessarily the case; instances to the contrary will be found among the groups of order and in §§ 69–72. While in general there is no limitation on the type of any Abelian factor-group formed from two consecutive terms of the series, it is to be noted that the last factor-group cannot be cyclical16, a restriction that can be established as follows. If were cyclical, then, because is the group formed of the self-conjugate operations of , the operations of could be arranged in the sets
being the order of and being an operation of . But since is permutable with its own powers and with every operation of , these operations would form an Abelian group. Now the self-conjugate operations of form the group , and therefore cannot be Abelian. Hence the assumption, that is cyclical, is incorrect.One immediate consequence of this result is that a group of order is necessarily Abelian. For if it were not, its self-conjugate operations would form a sub-group of order , and being of order would be cyclical.
A second consequence is that if , of order , is not Abelian, the order of the sub-group , formed of the self-conjugate operations of , cannot exceed .
54. It was proved in the last chapter that an Abelian group has one or more sub-groups the order of which is any given factor of the order of the group itself. Hence, since is Abelian, must have one or more sub-groups which contain , of any given order greater than that of and less than its own.
In particular, if and are the orders of and , then must have a series of sub-groups of orders
each of which contains those that follow it, while each one is of necessity self-conjugate in . Hence between and , a series of groups , , …, of orders , , …, can be chosen, so that each of the series is self-conjugate in and contains all that precede it. We can therefore always form, in at least one way, a series of sub-groups of of orders such that each is self-conjugate in and contains all that precede it.It will be shewn in the next paragraph that every sub-group of order of is contained self-conjugately in some sub-group of order , and that every sub-group of order of a group of order is self-conjugate. Assuming this result, it follows at once that a series of sub-groups of orders
can always be formed, which shall contain any one given sub-group and which shall be such that each group of the series contains the previous one self-conjugately. It will not however, in general, be the case, as in the previous series, that each group of the series is self-conjugate in .55. Let be any sub-group of of order , and let be the first of the series of sub-groups
which is not contained in . Then is a sub-group of which does not contain all the operations of . Every operation of is self-conjugate in ; and therefore is self-conjugate in , a group of order greater than its own. Hence must be contained self-conjugately in some group of order , where is not less than unity. Moreover since must contain operations of order , there must be one or more groups of order which contain self-conjugately. Hence:—THEOREM II. If of order is a sub-group of , which is of order , then must contain a sub-group of order , , within which is self-conjugate. In particular, every sub-group of order of is a self-conjugate sub-group17.
Suppose now that , of order , is the greatest sub-group of which contains a given sub-group , of order , self-conjugately; so that is one of conjugate sub-groups. Suppose also that of order , (), is the greatest sub-group of that contains self-conjugately. Every operation of transforms into itself; and no operation of that is not contained in transforms into itself. Hence, in , is one of conjugate sub-groups and each of these is self-conjugate in .
The sub-groups conjugate with may therefore be divided into sets of each, (), such that any operation of one of the sets transforms each sub-group of that set into itself.
Similarly if , of order , is the greatest sub-group of that contains a given operation self-conjugately, and if , , is the greatest sub-group that contains self-conjugately, then must contain self-conjugately operations of the conjugate set to which belongs, and therefore any two of these operations are permutable. Hence the conjugate operations of the set to which belongs can be divided into sets of each, (), such that all the operations of any one set are permutable with each other. In particular, if is one of a set of conjugate operations, all the operations of the set are permutable.
56. If is an operation of , of order , which is conjugate to one of its own powers , there must be some other operation such that
From this equation it follows that
Again
and so on. Hence
If is the lowest power of which is permutable with , then in will be one of conjugate operations, and therefore must be a power of , say , less than . Now
and therefore whileFirst, we will suppose that is an odd prime. Then since
whatever integer may be, we may assume that , where is not a multiple of ; and then
Hence
andIn particular, we see that no operation of order can be conjugate to one of its powers. Hence if and are two conjugate operations of order , and have no operation in common except identity. Also, if be a self-conjugate sub-group of order , each of its operations is self-conjugate.
If is , we must take
where is odd, and we are led by the same process to the result57. If is an operation of which belongs to the sub-group (§ 53) and to no previous sub-group in the set
then is a sub-group of ; and therefore every operation of is self-conjugate in . The set of operations is therefore transformed into itself by every operation of , and hence every operation that is conjugate to is contained in the set . Suppose now, if possible, that all the operations conjugate with are contained in the set . Then every operation of transforms this set into itself, and therefore every operation of is self-conjugate in . This however cannot be the case, since by supposition does not belong to . Hence among the operations conjugate to , there must be some which belong to and not to . In particular, if is one of conjugate operations, no operation conjugate to can belong to the set . For there must be an operation such that belongs to and not to ; and this being the case, no one of the operations , (), can belong to . But these operations with constitute the conjugate set.It follows from the above reasoning that, if belongs to , then each operation of the set , where is any operation of , belongs to . If belongs to , (), the operation , regarded as the product of and , must similarly belong to . Hence unless every operation of is permutable with , there must be operations conjugate to , which belong to the set . Moreover those operations of which transform into operations of the set form a sub-group. For if
and then and when both and belong to , so also does .58. In illustration of the foregoing paragraph, we will consider a group of order where is an odd prime in which the group , is of type , while is an Abelian group of type .
Let be an operation of order that generates ; and let , , …, be operations no one of which is a power of any other, such that, with , they generate .
If and are not permutable, then
andNow belongs to , and therefore
so thatHence the only operations which are conjugate to are
and is one of a set of conjugate operations, which are transformed among themselves by . Similarly is one of a set of conjugate operations which are transformed among themselves by . Hence both and are permutable with each of the operations , , …, . Again, is not self-conjugate, and there must therefore be another, say , of the set , , …, which, with its powers, transforms into a set of conjugate operations. Then will similarly transform ; and , will be permutable with each of the set , …, , , . Hence finally the set of operations , , …, must be divisible into sets of two, such that each pair are permutable with all the remaining operations, but are not permutable with each other. The group can therefore only exist if is even18.The sub-group is a group which satisfies the same conditions as , when . Its order is , and it contains a self-conjugate operation of order . Now we shall see in §§ 65, 66 that such a group is necessarily of one of two types19, the generating operations of which satisfy one or the other of the sets of equations
orThe same is true for the sub-groups , etc.; and all the operations of any one of these sub-groups are permutable with each operation of .
Hence finally, since the equations to be satisfied by each pair of operations, such as and , may be chosen in two distinct ways, there are in all distinct types of group of order , for which is a cyclical group of order , and is an Abelian group of type .
59. It has been seen in § 55 that every sub-group of order of a group of order is self-conjugate. Suppose now that contains two such sub-groups and . Then since and are permutable with each other, while the order of is , the order of the greatest group common to them must (§ 33) be ; and since is the greatest common sub-group of two self-conjugate sub-groups of , it must itself be a self-conjugate sub-group of . The factor group of order contains the two distinct sub-groups and , which are of order and permutable with each other. Hence must be an Abelian group of type , and it therefore contains (§ 49) sub-groups of order . Hence, besides , must contain other sub-groups of order which have in common with the sub-group . If the sub-groups thus obtained do not exhaust the sub-groups of of order , let be a new one. Then, as before, and must have a common sub-group , of order , which is self-conjugate in . If were the same as , would be contained in , which by supposition is not the case. It may now be shewn as above that there are, in addition to , sub-groups of order which have in common with the group . These are therefore necessarily distinct from those before obtained. This process may clearly be repeated till all the sub-groups of order are exhausted. Hence finally, if the number of sub-groups of , of order be , we have .
60. The self-conjugate operations of a group of order , whose orders are , form with identity a self-conjugate sub-group whose order is some power of ; and therefore their number must be congruent to . On the other hand, if is any operation of of order which is not self-conjugate, the number of operations in the conjugate set to which belongs is a power of . Hence the total number of operations of , of order , is congruent to . Now if is the total number of sub-groups of of order , the number of operations of order is , since no two of these sub-groups can have a common operation, except identity. It follows that
and thereforeIf now is any sub-group of of order , and if is the greatest sub-group of in which is contained self-conjugately, then every sub-group of which contains self-conjugately is contained in . But every sub-group of order , which contains , contains self-conjugately; and therefore every sub-group of order , which contains , is itself contained in . By the preceding result, the number of sub-groups of of order is congruent to unity, . Hence the number of sub-groups of of order , which contain of order , is congruent to unity, .
61. Let now represent the total number of sub-groups of order contained in a group of order . If any one of them is contained in sub-groups of order , and if any one of the sub-groups of order contains sub-groups of order , then
for the numbers on either side of this equation are both equal to the number of sub-groups of order , when each of the latter is reckoned once for every sub-group of order that it contains. It has however been shewn, in the two preceding paragraphs, that for all values of andHence
Now it has just been proved that
and therefore finally, for all values of , We may state the results thus obtained as follows:—THEOREM III. The number of sub-groups of any given order of a group of order is congruent to unity, 20.
Corollary. The number of self-conjugate sub-groups of order of a group of order is congruent to unity, .
This is an immediate consequence of the theorem, since the number of sub-groups in any conjugate set is a power of .
62. Having shewn that the number of sub-groups of of order is of the form , we may now discuss under what circumstances it is possible for to be zero, so that then contains only one sub-group of order .
If this is the case, and if is any operation of not contained in , the order of must be not less than ; for if it were less, would have some sub-group of order containing and this would necessarily be different from . If the order of is , then is a cyclical sub-group of order ; and it must coincide with . Hence, if is the only sub-group of order , it must be cyclical.
Suppose now that contains operations of order (), but no operations of order ; and let be an operation of of order . Then must be contained self-conjugately in a non-cyclical sub-group of order .
We will take first the case in which is an odd prime. Then (§ 55) must contain an operation which does not belong to , such that
If were unity, would be an Abelian group of order containing no operation of order . Its type would therefore be , and it would necessarily contain an operation of order not occurring in . It has been shewn that this is impossible if is the only sub-group of order , and therefore cannot be unity.We may then without loss of generality (§ 56) assume that
Moreover if were not divisible by , the order of would be , contrary to supposition. Hence we must have the relationsBy successive applications of the first of these equations, we get
for all values of and ; and from this it immediately follows thatHence if
the order of , an operation not contained in , is . This is impossible if is the only sub-group of order . If then , must contain operations of order greater than ; and is therefore a cyclical group. Hence:—THEOREM IV. If , of order , where is an odd prime, contains only one sub-group of order , must be cyclical.
63. When , the result is not so simple.
Let be an operation that transforms into itself, and suppose that the lowest power of that occurs in is . Then must be defined by
where Moreover must be a multiple of , as otherwise the order of would be greater than . A simple calculation now gives and since and can always be chosen so that When is thus chosen, is an operation of order which is not contained in . But this is inconsistent with being the only sub-group of order ; hence must contain .If now is not a self-conjugate sub-group, there must (§ 55) be some operation of order , conjugate with and not contained in , which transforms into itself; and then is defined by
where is or , and is odd.If
then and can be chosen so that is of order . Hence this case cannot occur.If is either or , the defining relations of are not self-consistent unless is ; for they lead to
andIf is , and if is an operation that transforms into , so that
then sinceHence would be of order ; and therefore this case cannot occur.
Hence, finally, must be self-conjugate.
If , must be transformed into itself by some operation not contained in , so that
and where and are each either zero or unity. If both are zero or both unity, is permutable with ; and if one is zero and the other unity, In either case, the group contains an operation of order that does not belong to , and this is inadmissible. Hence, lastly, must not be less than .Suppose now that is equal to , and that of the operations of , not contained in , has as small an order as possible, say (). Then of order is contained in ; without loss of generality we may assume
Moreover
and hence Now is permutable with , one of its own powers, and therefore is an operation of order . Hence is and is unity. If then is greater than unity, must contain operations of order and must therefore be cyclical. Moreover if is unity, the relations lead to and they are therefore inadmissible.On the other hand, the relations
give and every operation of , not contained in , is of order . Also these relations are clearly self-consistent, and they define a group of order . Hence finally:—THEOREM V. If a group , of order , has a single sub-group of order ,
, it must be cyclical; if it has a single sub-group of order , it is either cyclical or of the type defined by
64. We shall now proceed to discuss, in application of the foregoing theorems and for the importance of the results themselves, the various types of groups of order which contain self-conjugate cyclical sub-groups of orders and respectively. It is clear from Theorem V that the case requires independent investigation; we shall only deal in detail with the case in which is an odd prime, and shall state the results for the case when .
The types of Abelian groups of order which contain operations of order are those corresponding to the symbols , , , and . We will assume that the groups which we consider in the following paragraphs are not Abelian.
65. We will first consider a group , of order , which contains an operation of order . The cyclical sub-group is self-conjugate and contains a single sub-group of order . By Theorem IV, since is not cyclical, it must contain an operation , of order , which does not occur in . Since is self-conjugate and the group is not Abelian, must transform into one of its own powers. Hence
and since is permutable with it follows, from § 56, thatSince the group is not Abelian, cannot be zero; but it may have any value from to . If now
then and therefore, writing for , the group is defined byThese relations are clearly self-consistent, and they define a group of order .
There is therefore a single type of non-Abelian group of order which contains operations of order , because, for any such group, a pair of generating operations may be chosen which satisfy the above relations.
From the relation
it follows by repetition and multiplication that and therefore that andHence contains cyclical sub-groups of order of which and () may be taken as the generating operations. Since and are permutable, also contains an Abelian non-cyclical sub-group of order It is easy to verify that the sub-groups thus obtained exhaust the sub-groups of order ; and that, for any other order , there are also exactly sub-groups of which are cyclical and one is Abelian of type .
The reader will find it an instructive exercise to verify the results of the corresponding case where is ; they may be stated thus. There are four distinct types of non-Abelian group of order , which contain operations of order , when . Of these, one is the type given in Theorem V, and the remaining three are defined by
When , there are only two distinct types. In this case, the second and the fourth of the above groups are identical, and the third is Abelian.
66. Suppose next that , a group of order , has a self-conjugate cyclical sub-group of order , and that no operation of is of higher order than . We may at once distinguish two cases for separate discussion; viz. (i) that in which is a self-conjugate operation, (ii) that in which is not self-conjugate.
Taking the first case, there can be no operation in such that is the lowest power of contained in , for if there were, would be Abelian and, its order being , it would necessarily coincide with . Hence any operation , not contained in , generates with an Abelian group of type , and we may choose and as independent generators of this sub-group, the order of being . If now is any operation of not contained in , must occur in this sub-group, and therefore
If were not zero, would be the lowest power of that occurs in , and we have just seen that this cannot be the case. Hence
and is again an Abelian group of type . If and are independent generators of this group, the latter cannot occur in . Now since is not self-conjugate, and since is permutable with so thatHence
where is not a multiple of . If finally, be taken as a generating operation in the place of , the group is defined byThere is therefore a single type of group of order , which contains a self-conjugate operation of order and no operation of order .
67. Taking next the second case, we suppose that contains a cyclical self-conjugate sub-group of order , but no self-conjugate operation of this order.
If contains an operation such that is the lowest power of occurring in , it follows (§ 56), since is self-conjugate, that
Moreover, since the order of cannot exceed , we have
A simple calculation, similar to that of § 65, will now shew that
whileHence is an operation of order , none of whose powers, except identity, is contained in . If this operation be represented by , is defined by
If is a multiple of , the group is Abelian.
If is , and if is taken for a new generating operation, where is chosen so that
we have a single type defined byIf is not a multiple of , we may take it equal to , where and are both less than , and is not zero. (This case cannot occur if .) Then
when ; while if , the coefficient of , viz. of , in the index of must be increased by .If now we choose so that
and then so that (with the suitable modification when ), then transforms into . We therefore again in this case get a single type of group, which, if is represented by , is defined by As already stated, this type exists only when .If every operation of is such that its th power occurs in , must be one of a set of conjugate operations; for if is one of conjugate operations, there must be an operation which transforms into , where is not a multiple of , and then is the lowest power of which is permutable with . Since is one of conjugate operations, it must be self-conjugate in an Abelian group of type . Again, since and are conjugate operations, must be contained in a group of order , defined by
If is not a self-conjugate operation, it must be one of conjugate operations and (§ 57) these must be
Now if
then Hence, if is not self-conjugate, is a self-conjugate operation of order : and the group is that determined in the last paragraph. Therefore must be self-conjugate, and the group is defined by
It is clear from these relations that the group thus arrived at is the direct product of the groups and 21.
It is to be expected, from the result of the corresponding case at the end of § 65, that the number of distinct types when is much greater than when is an odd prime. There are, in fact, when , fourteen distinct types of groups of order , which contain a self-conjugate cyclical sub-group of order and no operation of order . They may be classified as follows.
Suppose first that the group has a self-conjugate operation of order . There is then a single type defined by the relations
Suppose next that the group has no self-conjugate operation of order , and let be a self-conjugate cyclical sub-group of order . If is cyclical, there are, when , five distinct types. The common defining relations of these are
and the five distinct types areIf , then (iv) and (v) are identical, and (vi) is Abelian; so that there are only three distinct types. If , there is a single type; it is given by (ii).
When is not cyclical, the square of every operation of is contained in . If all the self-conjugate operations of are not contained in , there must be a self-conjugate operation , of order , which does not occur in . If is any operation of , not contained in , then is a self-conjugate sub-group of order , which has no operation except identity in common with . Hence is a direct product of a group of order and a group of order . There are therefore, for this case, four types (vii), (viii), (ix), (x), when , corresponding to the four groups of order of § 65. If , there are two types.
Next, let all the self-conjugate operations of be contained in ; and suppose that is one of two conjugate operations. Then must contain an Abelian sub-group of type , in which occurs; and it may be shewn that, when , there are two types defined by the relations
When , there is, for this case, no type.Lastly, suppose that is one of four conjugate operations. Then must contain sub-groups of order , of the second and the third types of § 65, and a sub-group of order of either the first or fourth type (l.c.). In this last case, there are two distinct types defined by
These two types exist only when .68. We shall now, as a final illustration, determine and tabulate all types of groups of orders , and . It has been already seen that when the discussion must, in part at least, be distinct from that for an odd prime; for the sake of brevity we shall not deal in detail with this case, but shall state the results only and leave their verification as an exercise to the reader.
It has been shewn (§ 53) that all groups of order are Abelian; and hence the only distinct types are those represented by and .
For Abelian groups of order , the distinct types are , and .
If a non-Abelian group of order contains an operation of order , the sub-group it generates is self-conjugate; hence (§ 65) in this case there is a single type of group defined by
If there is no operation of order , then since there must be a self-conjugate operation of order , the group comes under the head discussed in § 66; there is again a single type of group defined by
These two types exhaust all the possibilities for non-Abelian groups of order .
69. For Abelian groups of order , the possible distinct types are , , , and .
For non-Abelian groups of order which contain operations of order there is a single type, namely that given in § 65 when is put equal to .
For non-Abelian groups which contain a self-conjugate cyclical sub-group of order and no operation of order , there are three distinct types, obtained by writing for in the group of § 66 and in the first and the last groups of § 67. The defining relations of these need not be here repeated, as they will be given in the summarizing table (§ 73).
It remains now to determine all distinct types of groups of order , which contain no operation of order and no self-conjugate cyclical sub-group of order . We shall first deal with groups which contain operations of order .
Let be an operation of order in a group of order . The cyclical sub-group must be self-conjugate in a non-cyclical sub-group of order , defined by
If is any operation of , not contained in , then since is not self-conjugate, we must have (§ 57)
and thereforeNow
and therefore the th power of every operation of order in is a self-conjugate operation.First let us suppose that contains other self-conjugate operations besides those of . Every such operation must occur in the group that contains self-conjugately; hence in this case must be self-conjugate.
We now therefore have
These equations give
Hence, if , is an operation of order . Denoting this by and by , the group is defined by
On the other hand, if is not zero, is an operation of order such that transforms it into a power of itself. This is contrary to the supposition that the group contains no cyclical self-conjugate sub-group of order . Hence cannot be different from zero: we therefore have only one type of group.
70. Next, let contain all the self-conjugate operations of ; and as before, let be the group that contains self-conjugately. If contains an operation of order which does not occur in , there must also be a non-cyclical sub-group of order which contains self-conjugately. Now and must have a common sub-group of order ; since this is self-conjugate in , it cannot be cyclical. The only non-cyclical sub-groups of orders that and contain are and . Hence these must be identical, and therefore must occur in . If now and were both Abelian, would be permutable both with and with , and would therefore, contrary to supposition, be a self-conjugate operation. Hence either (i) must contain a non-Abelian group of order , in which is an operation of order ; or (ii) the Abelian group , in which is an operation of order , must contain all the operations of of order .
In the case (i), the group is defined by
These equations give
and therefore, except when , is an operation of order . Hence, if , we may, without loss of generality, put zero for . In this case a simple calculation gives whereHence if , , , where , and are not multiples of , then
If now we take
thenDropping the accents, the defining relations now become
where is either zero, unity, or any given non-residue. Since the sub-group contains all the operations of order of the group, it follows at once that these three cases give three distinct types.
When , it will be found that the defining relations again give three distinct types. The operation may always be chosen so that and are zero. If it is so chosen, the three types correspond to the values , ; , ; and , ; of and .
In case (ii), the group is defined by
with the condition that all operations of , not contained in , are of order .
The formulæ for and enable us to calculate directly the power of any given operation of . Thus they give
If , this gives
so that, if is not a multiple of , the order of is . Hence the type of group under consideration can only occur when .In this case
Hence if and , we obtain a new type. A reduction similar to that in the previous case may now be effected; and taking unity for , the group is defined by
71. It only remains to determine the distinct types which contain no operation of order .
Suppose first that the self-conjugate operations of form a group of order . This must be generated by two independent operations and of order .
If now is any other operation of the group, must be an Abelian group of type . If again is any operation not contained in , it cannot be permutable with ; for if it were, would be self-conjugate. There must therefore be a relation of the form (§ 57)
Since any operation of may be taken for one of its generating operations, we may take or for one. If then is an independent operation of , will be defined by
in addition to the relations expressing that and are self-conjugate. There is then in this case a single type. That all the operations of in this case are actually of order follows from the obvious fact that every sub-group of order is either Abelian or of the second type of non-Abelian groups of order .72. Suppose, secondly, that the self-conjugate operations of form a sub-group of order , generated by . There must then be some operation which belongs to a set of conjugate operations; for if every operation of which is not self-conjugate were one of a set of conjugate operations, the total number of operations in the group would be congruent to . It follows that must be self-conjugate in a group of order ; and since is also self-conjugate in this group, it must be Abelian. Let , , be generators of this group and any operation of not contained in it. We may now assume that belongs to the sub-group (§ 53), and therefore that
whileIf were zero, would be a self-conjugate operation not contained in ; and therefore must be different from zero. We may now put
and the group is then defined by the relationstogether with the relations expressing that is self-conjugate. There is thus again in this case, at most, a single type. It remains to determine whether the operations are all actually of order .
The defining relations give
where and therefore whereHence
If is a greater prime than , the indices of , , are all multiples of ; hence is of order , being any operation not contained in . If however , then
so that, if is not a multiple of , is an operation of order . Hence this last type of group exists as a distinct type for all primes greater than ; but for , it is not distinct from one of the previous types containing operations of order .73. In tabulating, as follows, the types of group thus obtained, we give with each group a symbol of the form
indicating the types of , , , …, where is the series of self-conjugate sub-groups defined in § 53. This symbol is to be read from the left so that is the type of .Moreover in each group there is no operation of higher order than that denoted by .
I. , two types.
(i) ; (ii) .
II. , five types.
(i) ; (ii) ; (iii) ;
(iv) , , , ;
(v)
III. , fifteen types.
(i) ; (ii) ; (iii) ;
(iv) ; (v) ;
(vi) , , , ;
(vii)
(viii) , , , ;
(ix)
this group (ix) being the direct product of
and ;
(xi), (xii), and (xiii) ,
where for (xi) , for (xii) , for (xiii) any non-residue, ;
(xi), (xii), and (xiii) ,
where for (xi) , , for (xii) , , for (xiii) , .
(xiv)
this group (xiv) being the direct product of
and
;
(xv) ,
(xv) ,
74. To complete the list, we add, as was promised in § 68, the types of non-Abelian groups of orders and ; the possible types of Abelian groups being the same as for an odd prime.
Non-Abelian groups of order ; two types.
(i) identical with II (iv), writing for ;
(ii) , , , , .
Non-Abelian groups of order ; nine types.
(i), (ii), (iii), (iv) and (v) identical with III (vi), (vii), (viii), (ix) and (x), writing for ;
(vi)
this group (vi) being the direct product of
and
;
(vii) , , , ;
(viii) , , , ;
(ix) , , , , .
75. In the following examples, is a non-Abelian group whose order is the power of a prime ; and the sub-groups , , …, referred to are the series of self-conjugate sub-groups of § 53.
Ex. 1. If and are any two operations of , the totality of operations of the form generate a self-conjugate sub-group identical with or contained in .
Ex. 2. Shew that, if every sub-group of is Abelian, is an Abelian group of type ; and any two operations of , neither of which belongs to and neither of which is a power of the other, generate .
Ex. 3. If , , , … are of types , , , …, shew that can be generated by two operations.
Ex. 4. If , of order , is not Abelian, and if every sub-group of is self-conjugate, shew that must be . (Dedekind.)
Ex. 5. If is of order , and of type , and if , is the direct product of two groups. If , there is one type for which is not a direct product, viz.
Ex. 6. A group , of order , (), where is an odd prime, contains an operation of order , and no cyclical self-conjugate sub-group of order . Shew that, if is one of a set of conjugate operations, the defining relations of the group are of the form
and that, if is one of a set of conjugate operations, the defining relations are of the form
Determine in each case the number of distinct types.
up | next | prev | ptail | top |