The date shows when this volume was taken. All books not in use for instruction or re- search are limited to all borrowers. Volumes of periodi- cals and of pamphlets comprise so many sub- jects, that they are held in the library as much as possible. For spe- cial purposes they are give'n out for a limited time. Graduates and sen- iors are allowed five volumes for two weeks . Other students- may have two vols, from the circulating library for two weeks. Books not needed during recess periods should be returned to the library, or arrange- ments made for their return during borrow- er'sabsence, if wanted. Books needed by more than one person are held on the reserve list. Books of special value and gift books, when the giver wishes it, are not allowed to circulate. 8817 ON PRIMITIVE GROUPS OF ODD ORDER A THESIS PRESENTED TO THE UNIVERSITY FACULTY OF CORNELL UNIVERSITY IN CANDIDACY FOR THE DEGREE OF DOCTOR OF PHILOSOPHY HENRY LEWIS RIETZ BALTIMORE Cge £orb QgSaftimore (fittee THE FRIEDENWALD COMPANY I904 ^ Cornell University Library The original of this book is in the Cornell University Library. There are no known copyright restrictions in the United States on the use of the text. http://www.archive.org/details/cu31924032189692 ON PRIMITIVE GROUPS OF ODD ORDER A THESIS PRESENTED TO THE UNIVERSITY FACULTY OF CORNELL UNIVERSITY IN CANDIDACY FOR THE DEGREE OF DOCTOR OF PHILOSOPHY BY HENRY LEWIS RIETZ BALTIMORE Zfyt Boti> (gaUimott (P«ae THE FRIEDENWALD COMPANY I904 On Primitive Groups of Odd Order. By Henkt Lewis Rietz. Introduction. In his "Theory of Groups of Finite Order " (1897), p. 379, Burnside has called attention to the fact that no simple group of odd composite order is known to exist. Several articles* have recently appeared bearing on this question, in which, among other things, it was proved that no such group can be represented as a substitution group whose degree does not exceed 100. This result was obtained by showing that there is no simple primitive group of odd composite order whose degree falls within the given limits. Burnside determined all the primitive groups of odd order of degree less than lOO.f Since any primitive group of odd order is simply transitiye, a study of simply transitive primitive groups may throw light on the question of simple groups of odd order. Some important properties of simply transitive primitive groups have been given by Jordan, Miller, and Burnside. J The main objects of the present paper are ; first, to make a further study of primitive groups with special reference to those of odd order ; secondly, to extend the determination of the primitive groups of odd order to all degrees less than 243. It results that all groups arrived at in this determination are solvable. From this result it is evident that no simple group of odd composite order can occur * Miller, Proc. Lond. Math. Soc, Vol. 33, pp. 6-10. Burnside, Proc. Lond. Math. Soc, Vol. 33, pp. 162-185 ; 257-268. Frobenius, Berliner Sitzungsberichte (1901), pp. 849-858 ; 1216-1230. f At the time of the publication of this work, I had also made this determination with the same results. % Jordan, "Traite des Substitutions," pp. 281-284. Miller, Proc. Lond. Math. Soc, Vol. 28, pp. 633- 542. Burnside, loc. cit., pp. 162-185. 1 2 Rietz: On Primitive Groups of Odd Order. as a substitution group of degree less than 243, since if a simple group is repre- sented as a substitution group on the minimum number of letters, it is primitive. Part I contains a number of theorems, most of which apply to primitive groups whether the order is even or odd, but some use can be made of nearly all of them in determining all the primitive groups of odd order of a given degree. Part II contains the determination of the primitive groups of odd order whose degrees lie between 100 and 243. I desire to acknowledge my indebtedness to Professor G. A. Miller for help- ful suggestions and criticisms during the preparation of this paper. Part I. §1. — On the Number of Substitutions of Degree less than n contained in any Transitive Group of Degree n. Let G be any primitive group of composite order g on the elements a x , a z , , a n , and G, the subgroup leaving a given letter a s fixed. If n — A a is the degree of any substitution of G s , and fi a the number of substitutions of G e of this degree, then the total number of substitutions of degree less than n con- tained in G is m + mi + 7 VL+ .... +m? or n (j± + bl + j& + . . . . + j±\ , Aj Ag A3 A p \ Aj Ag A3 Ap / where p is the number of different degrees occurring among the substitutions of G s . Since ^ + [i z + [i 3 + .... + fi p = ^- , the above summation in the paren- n 1 theses may be considered as the sum of just -^- terms of the form — . "We may n A then rewrite the above expression for the number of substitutions of degree less than n in the form a a = 2- Let x denote the number of systems of intransitivity of G s , and look upon Cr 8 as having just one system of intransitivity when it is transitive. Then x + 1 * Jordan, Liouville's Journal, Vol. 17 (1872), p. 352. Rietz : On Primitive Groups of Odd Order. 3 is the average value of K since the average number of letters in the substitutions of an intransitive group is equal to the excess of the degree over the number of systems of intransitivity.* Hence we have 1 = l 2> n The Xs in this summation cannot all be equal, since identity is included among the substitutions of O, . Since the arithmetic mean of any number of positive quantities which are not all equal is greater than their geometric mean, it follows that a=£ n 2>» -l_>-C/a 1 X 8 ....^ (3) g n and « = £ Jl*Lt/JLJ_.. * " »/ 1 1 1 " > V^-^-----v (4) n From (3) and (4) it follows that 9_ n l= £ From (2) and (5) it follows that a = •Jordan, Comptes Rendus, Vol. 74 (1872), p. 977. Frobenius, Crelle's Journal, Vol. 101 (1887), p. 288. i = l , = £. (6) 4 Rietz: On Primitive Groups of Odd Order. From (1) and (6) we obtain Theorem 1. — In any primitive group G of degree n of composite order g there are more than — ^ — substitutions of degree less than n, where x is the number of x + 1 systems of intransitivity of the subgroup which leaves a given letter fixed. In particular, for a multiply transitive group, x = 1. Hence, Cor. 1. In a multiply transitive group of degree n more than one-half of the substitutions are of degree less than n . Cor. 2. If G is of degree kp (p a prime) and of order mp (m prime to p and p — 1), the subgroup G s has at least p + 1 transitive constituents. For a group of this order contains exactly m operators whose orders divide m.* But all the substitutions of degree less than kp would be of orders prime to p. Hence from the above theorem we have x >■ J_ or x ~^>p — 1 . Since mp must clearly be an odd number, x must be even. Hence, x>p+l. While it is not our object to treat imprimitive groups, the above theorem can at once be extended to any non-regular transitive group. The only change in the argument is the substitution of x + m in expression (l) for x + 1, where m represents the number of letters of the transitive group left fixed by the sub- group which leaves a given letter fixed. Hence, Theorem 2. — In any non-regular transitive group of degree n of order g there are more than —£- — substitutions of degree less than n, where x and m are defined as above. When applied to known groups, I find that in many eases this simple formula gives very nearly the actual number of substitutions of degree less than the degree of the group. * Frobenius, Berliner Sitzungsberichte (1895), p. 1035. Rietz : On Primitive Groups of Odd Order. 5 §2. — Restrictions on the Order of G, when G s has a Transitive Constituent of Degree p, p a , pm or pq (p and q primes and m
0) ; this must be the case
if the theorem is not true. It will be shown that this hypothesis leads to a con-
tradiction. In H s all the substitutions whose orders are powers of p would gene-
rate a group H' s of order 7Jp m invariant in G s . In the conjugate G r of G s , leav-
ing fixed an element of T lt there occurs just 1/t of the substitutions of G„.
Hence the subgroup Hi would be one of a set of t conjugates transformed by G r
according to one of its transitive constituents T of degree t. In the invariant
subgroup HI of G r corresponding to identity in T, all the substitutions are com-
mon to G r and G s , since they transform H' s into itself. Now E' r would be of
order %"p m ( a," prime to p) . Since in T x all the substitutions whose orders are
not prime top are of degree t, all the substitutions whose orders are powers of
p common to G r and G s are contained in H, .
If H' r contained all the substitutions whose orders are powers of p which
occur in H s , the subgroup H[ would be invariant in G s and G r . But this is
impossible, since these subgroups are maximal. If Hi. contains only part of these
substitutions, let P be such a substitution not contained in Hi. The order of
* Miller, loo. cit., pp. 534, 535. t Jordan, loc. cifc, p. 284.
6 Rietz : On Primitive Groups of Odd Order.
\H' r , P\ would then be divisible byp m + 1 and there would be common to G s
and G r subgroups of order p m + 1 , which is impossible, since, by hypothesis, the
order of H s is not divisible by p m + \ Hence the theorem.
Cor. 1. If G s has a transitive constituent of prime degree p, the order of G s
is not divisible by p 2 .
Cor. 2. If any number of the transitive constituents of H s are of a given prime
degree p, the constituent group formed of all these transitive constituents is formed by
establishing a simple isomorphism between them.
Cor. 3. If in G 8 all the transitive constituents of a given degree p a are of class
p a — 1 , the order of G s is not divisible by p a + \
Cor. 4. If in G s all the transitive constituents of degree mp (p >■ m) have p sys-
tems of imprimitivity, the order of G s is not divisible by p 2 .
Lemma. When p and q are distinct primes each of the form 2 m + 1, there is
no imprimitive group of degree pq of odd order whose order is divisible by both p 2
and q* ; and there is no primitive group of degree pq involving in its order only
the primes p and q .
The part of this lemma which relates to the imprimitive groups follows at
once from the fact, that the only transitive groups of degrees p and q whose
orders are odd are the cyclical groups of orders p and q. Suppose there is a
primitive group of degree^ of order p a 'q°*. The maximal subgroup G u leaving
a given letter fixed, is then of degree pq — 1 and of order p ai ~ 1 q ai ~ 1 . Take
p^>q, then, since p 2 ^>pq — 1 , no transitive constituent can be of degree^> v (y ]>1).
The transitive constituents cannot all be of degree p, since p is not a divisor of
pq — 1. Since pq — 1 is not divisible by q, we may assume that some of the
transitive constituents are of degree p while others are of degrees equal to a
power of q. But the order of a transitive constituent of degree pis p, and would
therefore not contain q as a factor, but every prime which divides the order one
transitive constituent of G 1 divides the order of each of its transitive constituent.
Theorem 4. — If p and q are distinct primes of the form 2" 1 + 1 , and if G s is of
odd order, and has as a transitive constituent an imprimitive group of degree pq ;
then, according as T has p or q systems of imprimitivity, the order of G s is not
divisible by p 2 or q z .
To make the conditions definite, suppose that T has q systems of imprimi-
tivity. These systems are then permuted according to the cyclical group of
Rietz : On Primitive Groups of Odd Order. 7:
order q, and all the substitutions in the tail of T are of degree pq. Corre-
sponding to identity in T, there is in G a an invariant subgroup H s of degree
n — a (a <^ pq -\- I) . If we can show that the order of H 8 is not divisible by q %
our theorem is proved. Let G T be the conjugate of G s which leaves fixed an
element of T. Also let R s be the invariant subgroup of G s corresponding to the
head of T. In G r the subgroup H s is one of a set of pq conjugates transformed
by G r according to a transitive constituent T x of order p a 'q a *. According to the
lemma, T x is imprimitive and its order is not divisible by both p 2 and q 2 . The
subgroup T lt leaving a given letter fixed, would leave more than one letter fixed.
Hence in G r the subgroup H, is transformed into itself by some of its conjugates.
Let H Si be one of these conjugates such that H'~ 1 n,H ai = E 8 . H ti then occurs in
both G s and G r . Hence, it occurs in B s . If R s contains operators of order q,
they clearly occur in H s . Hence, H, and H ti have the same substitutions of
order q. But H g is invariant in G s and H Si in G Si . The substitutions of order
q in H s would then generate a group invariant in both G s and 6r Si . But this is
impossible, since G g is maximal. Hence the theorem.
•
§3. — On Certain Subgroups Contained in G.
Let p a be the highest power of a prime p which divides the order of G .
and suppose that the number p is prime to n, the degree of G. Let P be
any subgroup of order p a . It must be contained in some of the subgroups
(•?!, G 2 , .... , G n , leaving a given letter fixed, since its degree is prime to n. If
P is of degree n — a, (a, > 1), it is proved by Burnside ("Theory of Groups," p.
202), that the subgroup of G, which contains all the substitutions of G which
transform P into itself, permutes the /I elements not occurring in P transitively.
It is our object to consider the case % = 1 . Let P' be a subgroup of order p*
common to any two of the subgroups G it G z , G 3 , .... , G n such that there is no
subgroup of order p y (y >-/3) common to any two of these subgroups. We shall
first assume ft > . P' must be contained in subgroups P x and P 2 of order p a in
those subgroups which leave a given letter fixed in which it occurs. Since, in a
subgroup of order p a , any subgroup P' is transformed into itself by operators of the
group not contained in P', it follows that P' is invariant in a subgroup P" of
P x which is of degree n — 1 . Likewise in P 2 the subgroup P' is invariant in a
subgroup P'" of degree n — 1. Hence, the subgroup P' is invariant in \P", P"'\
8 Rietz : On Primitive Groups of Odd Order.
of degree n. Since n = 1 raod^j, the number of elements of G not occurring in
P is congruent to unity mod p . Also, since P" and P" are each of degree n — 1,
it follows that \P", P'"\ has a transitive constituent of degree 1 + kp (&>0),
formed of elements not occurring in P, and whose order is multiple of p.
When (3=0, the subgroup P is clearly formed by establishing a simple
isomorphism between regular groups. Hence,
Theorem 5. — If p a is the highest power of a prime p which divides the order of
G, and if a subgroup P of order p* is of degree n — 1 , then, unless P is a regular
group or is formed by establishing a simple isomorphism between regular groups of
order p*, G contains an intransitive subgroup of degree n having a transitive con-
stituent of degree 1 -f- Jcp (Jc > 0) and of order Ip.
Cor- lfp a is the highest power of a prime p which divides the order of G s , and
if the degree of each transitive constituent of G s is divisible by p 1 *, but at least one of
them is not divisible by p p + 1 , then either a = (3 or the group G contains a subgroup
of degree n having a transitive constituent of degree 1 -f- lep {Jc >• 0) and of order
equal to a multiple of p.
It may be observed that the theorem and corollary just stated apply to any
transitive group in which the subgroup which leaves a given letter fixed leaves
only one letter fixed, as well as to a primitive group.
§4. — On the Transitive Constituents of G s .
Theorem 6. — If G s has an invariant subgroup H s of degree n — a (a > 1) ,
then G s has at least one transitive constituent whose degree exceeds the degree of any
transitive constituent of M s .
Suppose, if possible, that H s has a transitive constituent T such that its
degree is equal to the degree of the transitive constituents of G s of largest
degree.
Consider a conjugate G r of G s , leaving fixed an element of G e not occurring
in B„. Since H 8 occurs in both G 8 and G r , these two groups have at least one
transitive constituent in the same elements, i. e., in the elements of T. The
group \G r , G,\ would then be intransitive. But \G r , G,\ must be identical
Bietz : On Primitive Groups of Odd Order. 9
with G, since G s is maximal. Hence the hypothesis that H s has the transitive
constituent T leads to an absurdity.
Cor. If all the transitive constituents of G 8 are primitive groups of the same
degree t, then G, is formed by establishing a simple isomorphism between these tran-
sitive constituents.
This follows readily from the theorem if we remember that every invariant
subgroup of a primitive group is transitive.
Theorem 7. — If G s has as a transitive constituent a regular group T of degree
t , and if the order of G s exceeds t , then G, has another transitive constituent of degree
t which has the property that its subgroup which leaves a given letter fixed permutes
all the remaining letters.
Consider the invariant subgroup H s of G, corresponding to identity in T.
In a conjugate H r of G s , leaving fixed an element of T, there occur just 1/t of
the substitutions of G s and the subgroup H s is one of t conjugates transformed
according to a transitive constituent T x . If, in the group T lt the subgroup which
leaves a given letter fixed, leaves more than one letter fixed, H s is transformed into
itself by some of its t conjugates under G r . But the substitutions of E s are the
only substitutions common to G r and G,. Hence, the transitive constituent T
has the property mentioned in the theorem.
Theorem 8. — If G s has X systems of intransitivity, and if H s is the invariant
subgroup of G s corresponding to identity in any transitive constituent T, while G s
transforms H Si , H H , .... , H s _ (denned as in §2) according to a constituent group
having u systems of intransitivity, then H s has more than — systems of intransi-
tivity, excepting when /x = 1 , and then it has at least X .*
If H, has as few as — systems of intransitivity, the [i conjugate sets under
G s into which H s , H $ , , H s are divided could, at most, contain elements
from f l)(i + l = a, — [i + l of the /I systems of intransitivity of G s , since
each of the subgroups H Si , H H , ^ a -i must contain at least a cycle from T.
* Of. Miller, loc. cit., p. 535
10 Kietz : On Primitive Groups of Odd Order.
These subgroups could not then generate a group of degree n — 1 unless (i = 1 .
Hence, by means of 1, §2, the theorem follows.
Cor. If all the transitive constituents of G„ are primitive groups,
H Si , H H , . . . . , H a cannot be a single conjugate set under G g .
§5. — On the Transitive Constituents of G s when the Order of G is Restricted
to be an Odd Number.
Burnside recently proved the interesting theorem* that, if G is of odd order,
G s has its transitive constituents in pairs of the same degree.
Let a Si , a H , a H , , a st be the elements of any transitive constituent of
degree t. The above theorem was proved by considering the quadratic function
8 = n
/= 2 a * K + a n + •■■■ + «J,
s = l
which is transformed into itself by all the substitutions of G. In this summa-
tion, a s occurs in the parentheses exactly t times. Hence the function /may
also be written
S = 71
/=2 (°« + a ** + ' • • • + °o a *>
s=l
and it is shown in the proof of the above theorem that the elements
a si , a ai , a K , .... , a H are elements of a transitive constituent T of G s distinct
from T. The constituents T and T' will be spoken of as a " pair of transitive
constituents." Use will be made of the two ways in which / is written to prove
some theorems in reference to the transformation by G, of its subgroups
E Si , H 8i , . . . . , E s _ (defined as in §2) when H s corresponds to identity in T. It
is known (p. 5) that these a — 1 subgroups are transformed by G 8 according
to one of its constituent groups. But it is not known whether this constituent
group ever contains elements occurring in E s . Form the conjugate G Sa of G s ,
leaving fixed an element of T. From the two ways of writing /, it is seen that
in G Sa the element a s occurs in the transform of T' ; i. e., in B~ 1 T'B, where B
is such that B~ l G s B = G Sa . But H s is transformed by G in the same manner
as a 8 is replaced. Hence,
* Loc. cit., p. 163.
Eietz : On Primitive Groups of Odd Order. 11
Theorem 9. — Some of the subgroups E H , E H , .... , E s _ are transformed
according to T 1 when H s corresponds to identity in T.
Cor. If the subgroups E Si , E ti , .... , E t _ are a single conjugate set under G e ,
they are transformed according to T 1 when H, corresponds to identity in T.
Suppose, next, that G s has only two transitive constituents T and T. If, cor-
responding to identity in one of these constituents, say T, there is in G t an inva-
riant subgroup E s , the subgroups E 8i , E H , .... ,E Snl (Cor., Theor. 9) are trans-
formed by G s according to the elements of E s . Then, for any two of the n
subgroups Si, E % , .... , E n , which are conjugate under G, one of two relations
Hf^H f -H. or H7 l E^=H, (1)
holds, but both cannot hold for any two of the subgroups. Let x be the number
of elements common to E a and E fi ; then x is clearly the number of elements
common to any two of the H's. Also, let x + y be the degree of H s . Let
a lt a 2 , . . . . , a y , b 1: b 2 , . . . ■ , b x be the elements of H s , and b 1: b it . . . . , b x ,
c i> c 2> ••'•'• > c y * ne elements of H tl , one of the subgroups E h , E H , . . . . , H, n _ 1 con-
tained in G s . Since E Si must be transformed according to an element of E e not
contained in E h , it must be transformed according to one of the a's. There
must be substitutions in E„ which do not transform E !t into itself. If iS is such
a substitution, then S~ 1 E S] S contains all the a's, since E Si and /S~ 1 E h S have just
x elements in common. But since a Sl , according to which E Si is permuted,
occurs in S~ 1 E Sl S, this latter subgroup cannot transform E Sl into itself. By
exactly the same reasoning E h cannot transform S~ 1 E H S into itself. But this
is contrary to relations (1). Hence,
Theorem 1 0. — If, in a primitive group G of odd order, the subgroup G s has
only two transitive constituents, G„ is formed by establishing a simple isomorphism
between them.
Theorem 11. — If, in a primitive group G of odd order, G s has not more than
four transitive constituents, and if these are all primitive groups, then it is formed by
establishing a simple isomorphism between them.
Since G s has an even number of transitive constituents, we need consider
only the cases where it has two or four transitive constituents. Since any inva-
riant subgroup of a primitive group is transitive, and since a simply transitive
12 Rietz : On Primitive Groups of Odd Order.
primitive group of degree n cannot have a transitive subgroup of degree less
than n, the theorem follows at once when G, has only two transitive constituents.
If G s has four transitive constituents,- and is not formed according to the
theorem, there corresponds to identity in some transitive constituent T of degree
t an intransitive subgroup H s invariant in G s . It has two or three systems of
intransitivity. Suppose, first, that H, of degree n — a has three systems. Then
a — 1 = t. In the conjugate of G s , leaving fixed a letter of T, the subgroup H
is one of t conjugates. But these a — 1 subgroups cannot be conjugate (Cor.,
Theor. 8). It remains to consider the case where H s has two systems of intran-
sitivity ; then n — a (the degree of H s ) is an even number. Hence a — 1 is an
even number and the a — 1 subgroups E x , ff Si , , H s _ could only be trans-
formed according to a group T having two transitive constituents. But by
Theor. 8 this is impossible. Hence the theorem.
§6. — Certain Primitive Groups of Odd Order contained in the Hohmorjph of the
Abelian Group P of Order p m (p an odd prime) of Type (1, 1, , l).
Represent P as a regular group. Suppose that the order of its group of
isomorphisms L is divisible by q n (q an odd prime). To any subgroup of order
q n in L there corresponds in the holomorph of P a transitive group of degree p m
and of order p m q u . The subgroup of this transitive group, which leaves a given
letter fixed, is of order q n , and is clearly maximal, if m is the index to which p
belongs mod q. Hence,
Theorem 12.— If p m = 1 mod q n (n <£ l), m being the index to which p belongs
mod q, there is a primitive group G of order p m q n contained in the holomorph of the
abelian group of order p m of type (1, 1, , 1).
Cor. 1. If q n is the highest power of q which divides p m — 1, there is only one
group G satisfying the above conditions.
Cor. 2. If p — 1 (p zp, 3) is not divisible by 3, there exists a primitive group G
of degree p 2 and of order Zp % . Furthermore, if p — 1 is divisible by 3 , there is no
primitive group of degree p* whose order is 3p z .
The first part of this corollary is merely a special case of the general
theorem.
The second part may be proved as follows :
By Sylow's theorem, a group of this order contains a single subgroup P of
Rietz : On Primitive Groups of Odd Order. 13
order p*. Since G is primitive, an invariant subgroup P must be transitive.
The subgroup is, therefore, regular, and it must be the non- cyclical group of
order p 2 . P would contain p + 1 subgroups of order p, these would have to
occur in conjugate sets of three, since G cannot contain an invariant intransitive
subgroup. But p + 1 is not divisible by 3 when p — 1 is divisible by 3 .
§7. — On the Class of Primitive Groups G of Odd Order.
By the class of a substitution group is meant the smallest number of ele-
ments in any one of its substitutions besides identity.*
Let n — [i represent the class of G. For all odd values of fi less than 7
there exist groups G of odd order of class n — p. Thus :
For fi = 1 , in any non-cyclic invariant subgroup of a metacyclic group.
For fi = 3, in the primitive group of degree 27 of order 27.39.f
For (i = 5, in the primitive group of degree 125 of order 125.93.|
It will now be shown that there is no primitive group of odd order in which
(i is even and less than 6. G s has an even number of transitive constituents in
pairs of the same degree (p. 10), and is clearly formed by establishing a simple
isomorphism between its transitive constituents. If p = 2 , at least two of the
transitive constituents must be non-regular, since they are in pairs of the same
degree. Suppose that t x is the degree of one of these non-regular transitive con-
stituents. It must clearly be of class t x — 1 . In the constituent of degree 1t y
formed by combining these two, every substitution of degree t x — 1 would corre-
spond to a substitution of degree t x , or G would contain substitutions of degree
n — 3. But this is clearly impossible, since a transitive group of degree t t and
of class t y — 1 , the order of the substitutions of degree t x is prime to the order of
those of degree t t — 1 .
It remains to consider the case when (i = 4. Here again G s must be formed
by establishing a simple isomorphism between transitive constituents, not all of
which can be regular. If t x is the degree of any non-regular transitive constitu-
ent, this constituent must either be of class t t — 3 or t x — 1 . Suppose that all
* Jordan, Liouville, Vol. 16 (1871), p. 408. t Burnside, loc. cit., p. 180.
% See p. 30 of this paper.
14 Rietz: On Primitive Groups of Odd Order.
the non-regular transitive constituents are of class one less than their degrees.
Since there must be an even number of such transitive constituents, it readily
follows that G 8 cannot have more than two such transitive constituents or G
would contain substitutions of degree less than n — 4. But in this case G would
have no substitutions of degree less than n — 3 . Hence G s must have at least
one transitive constituent T of some degree t s of class t z —3. Now, G„ must
have at least one more transitive constituent T' of degree t t . This constituent
must be of class t z — 1 or t 3 — 1 . In combining T and T' into a constituent of
degree 2t z , all the substitutions of degree t 2 — 3 in one must correspond to sub-
stitutions of degree t z in the other or G would contain substitutions of degree
less than n — 4. From this it is easily seen that substitutions of degree t z — 3
and those of degree t z to which they correspond must be regular. Hence all the
substitutions of degree t % — 3 would be of order 3 . Consider the subgroup P s of
G 8 corresponding to a subgroup of To? degree t 2 — 3 of order 3° such that there
is no subgroup of T of order 3 a + 1 which is of degree t z — 3 . P 3 would then be
invariant in a subgroup of G of degree n , and the 4 letters of G not occurring in
P s would be transitively connected so that the order of G must be an even num-
ber. Hence there is no primitive group of odd order of class n — 4.
Part II.
§8. — On the Primitive Groups of Odd Order of Degree less than 243.
It is known* that all transitive groups of odd order of prime degree are
invariant subgroups of the metacyclic group. Inasmuch as these groups are
well known, we shall consider only those primitive groups whose degrees are
not primes. As already stated in this paper (p. 1), the primitive groups of odd
order have been determined for degrees not exceeding 100. It is the object of
this part to extend this determination to all degrees less than 243. "We are then
concerned with the groups whose degrees lie between 100 and 243. It is stated
without proof by Burnside (loc. cit., note, p. 185) that any transitive group of odd
order of degree 3/> (p a prime) is imprimitive. We have examined all the com-
* Burnside, loc. cit., p. 177.
Rietz : On Primitive Groups of Odd Order. 15
posite numbers within the given limits for the degrees of primitive groups of odd
order, and the results agree with this statement. We shall, therefore, for the
sake of brevity, omit these degrees.
Represent by G n a primitive group of odd composite order of degree n . As
in Part I, let G, denote the subgroup of G n containing all the substitutions which
leave a given letter a a fixed. The method is, briefly, as follows :
For each odd degree n (n not a prime nor 3 times a prime) it is assumed
that a group G n exists. The degree of any solvable primitive group is a power
of a prime.* Hence, in order to prove that no group G n exists, for a given value
of n which is not the power of a prime, it is sufficient to prove that there is no
simple group G", provided there is no simple group of odd composite order of
degree less than n . As the latter condition is satisfied, it is further assumed
that G n is simple when it is not a power of a prime.
We write down for examination all those, and only those, systems of intran-
sitivity of G s which are not excluded by the conditions, 1°, that every prime
which divides the order of one transitive constituent divides the order of every
transitive constituent ;f 2°, that the transitive constituents occur in pairs of the
same degree ;J 3°, that when G n is simple, there can be no transitive constituent
whose degree is a prime of the form 2 m + 1 ;§ 4°, that if the degree of one tran-
sitive constituent is a prime of the form 2™ + 1 , all the transitive constituents
are of this degree.
This method of excluding transitive constituents of G s depends, of course,
on a knowledge of the primes which occur in the orders of transitive groups of
odd order of degree less than w/2. The following table shows the primes which
may occur in the orders of transitive groups of odd order of degree less than
120. The primes written under the degree are the primes which occur in
the orders of some transitive groups of odd order of the given degree in addition
to those primes contained in the degree itself.
* Lettre de Galois, a M. Auguste Chevalier, Liouville's Journ. (1846), p. 41.
t Jordan, loc. cit., p. 284. t Burnside, loc. cit., p. 165. § Miller, loc. cit., p. 6.
16
Rietz : On Primitive Groups of Odd Order.
Degree :
3
5
7
9
11
13
15
17
19
21
23
Primes :
3
5
3
3
11
Degree :
25
27
29
31
33
35
37
39
41
43
45
Primes :
3
13
7
3, 5
5
3
3
5
3, 7
Degree :
47
49
51
53
55
57
59
61
63
65
67
Primes :
23
3
13
29
3, 5
3
3, 11
Degree :
69
71
73
75
77
79
81
83
85
87
89
Primes :
11
5, 7
3
3, 5
3, 13
5, 13
41
7
11
Degree :
91
93
95
97
99
101
103
105
107
109
111
Primes :
3
5
3
3
5
5
3, 17
53
3
Degree :
113
115
117
119
Primes :
7
11
3.
We shall use &, S 2 , . • • • , S t to represent the sets of systems of intransi-
tivity of G s which are not excluded by the conditions above stated. The tran-
sitive constituents thus obtained are either shown to lead to impossibilities or
the orders and number of the groups are determined.
\