Extensions and the irreducibilities of the induced characters of cyclic p-groups
Katsusuke Sekiguchi
(Received July 21, 2001) (Revised January 4, 2002)
Abstract. Letfbe a faithful irreducible character of the cyclic groupCnof order pn, where pis an odd prime. We study the p-groupGcontainingCnsuch that the induced character fG is also irreducible. The purpose of this paper is to determine the subgroupsNGðNGðCnÞÞandNGðNGðNGðCnÞÞÞofGin the case when½NGðCnÞ;Cn ¼p.
1. Introduction
Let G be a finite group. We denote by IrrðGÞ the set of complex irre- ducible characters of G and by FIrrðGÞ ðHIrrðGÞÞ the set of faithful irreduc- ible characters of G.
Let p be a prime. For a non-negative integer n, we denote by Cn the cyclic group of order pn. A finite group G is called an M-group, if every fAIrrðGÞ is induced from a linear character of a subgroup of G.
It is well-known that every nilpotent group is an M-group. Hence, when Gis a p-group, for anywAIrrðGÞ, there exists a subgroupH ofG and a linear character f of H such that fG¼w. If we set N¼Kerf, then NkH and f is a faithful irreducible character of H=NGCn, for some non-negative integer n. In this paper, we will consider the case when N¼1, that is, f is a faithful linear character of HGCn.
We consider the following:
Problem 1. Let p be an odd prime, and f be a faithful irreducible character of Cn. Determine the p-group G such that CnHG and the induced character fG is also irreducible.
Since all the faithful irreducible characters of Cn are algebraically con- jugate to each other, the irreducibility of fG ðfAFIrrðCnÞÞ is independent of the choice of f, and depends only on n.
2000 Mathematics Subject Classification. 20C15.
Key words and phrases. p-group, extension, irreducible induced character, faithful irreducible character.
This problem has been solved in each of the following cases:
(1) CnkG ([2]),
(2) Ghas a subgroup H containingCn such thatCnkH and½G:H ¼p ([6]).
On the other hand, when p¼2, Yamada and Iida [4] proved the following interesting result:
Let Q denote the rational field. Let G be a 2-group and w a complex irreducible character of G. Then there exist subgroups HjN in G and a complex irreducible character f of H such that w¼fG, QðwÞ ¼QðfÞ, N¼ Kerf and
H=NGQn ðnb2Þ; or Dn ðnb2Þ; or SDn ðnb3Þ; or Cn ðnb0Þ:
Here, Qn, Dn and SDn denote the generalized quaternion group, the dihedral group of order 2nþ1 ðnb2Þand the semidihedral group of order 2nþ1 ðnb3Þ, respectively, and QðwÞ ¼QðwðgÞ;gAGÞ.
They considered the following:
Problem 2. Let f be a faithful irreducible character of H, where H¼Qn
or Dn or SDn. Determine the 2-group G such that HHG and the induced character fG is also irreducible.
Yamada and Iida [3] solved this problem in the case when ½G;H ¼2 or 4 and we have solved it when ½G;H ¼8 ([5]) for all H¼Qn or Dn or SDn.
Moreover, we have recently solved Problem 2 completely ([7]). In [7], we showed that
G¼NGðHÞ or NGðNGðHÞÞ;
for all H¼Qn orDn orSDn, if Gsatisfies the conditions of Problem 2. Here, as usual, NGðHÞ and NGðNGðHÞÞ are the normalizers of H and NGðHÞ in G, respectively. This means that, if we define subgroups of G by
M1¼NGðHÞ; and Miþ1¼NGðMiÞ; for ib1;
then
HHM1HM2¼M3 ¼M4¼ ¼G;
for all H ¼Qn or Dn or SDn.
In this paper, we consider Problem 1. We also define subgroups ofG by N1¼NGðCnÞ; and Niþ1¼NGðNiÞ; for ib1:
The purpose of this paper is to determine the groups N2¼NGðNGðCnÞÞ and N3¼NGðNGðNGðCnÞÞÞ in the case where ½NGðCnÞ;Cn ¼p and ½G;Cnbp3. As a consequence of the results, we will see that
CnYN1YN2YN3; in this case.
Throughout this paper, ZandNdenote the set of rational integers and the natural numbers, respectively. We will frequently use the word ‘‘respectively’’
so it is abbreviated to ‘‘resp.’’.
2. Statements of the results
For the rest of this paper, we assume that p is an odd prime. First, we introduce the following groups:
( i ) Gðn;mÞ ¼ha;bmi with apn ¼bmpm ¼1, bmab1m ¼a1þpnm, ðman1Þ.
( ii ) Gðn;m;1Þ ¼ha;bm;vi ðjGðn;mÞ ¼ha;bmiÞwithapn¼bmpm¼1,bmab1m ¼ a1þpnm, vav1¼a1þpnm1bmpm1, vp¼bm, vbmv1¼bm ð2man1Þ.
(iii) Gðn;1;1;1Þ ¼ha;b1;v;xi ðjGðn;1;1Þ ¼ha;b1;viÞ with apn¼b1p¼1, b1ab11 ¼a1þpn1,vav1¼a1þpn2b1,vp¼b1,vb1v1¼b1,xax1¼a1þpn3v, xp¼v, xvx1¼v, xb1x1¼b1 ð7anÞ.
We can see that Gðn;m;1Þ (resp. Gðn;1;1;1Þ) is an extension group of Gðn;mÞ (resp. Gðn;1;1Þ) by using Proposition 1 below:
Proposition1. Let N be a finite group such that GjN and G=N¼huNiis a cyclic group of order m. Then um¼cAN. If we put sðxÞ ¼uxu1, xAN, then sAAutðNÞ and (i) smðxÞ ¼cxc1; ðxANÞ (ii) sðcÞ ¼c.
Conversely, if sAAutðNÞand cAN satisfy (i)and (ii), then there exists one and only one extension group G of N such that G=N¼huNiis a cyclic group of order m and sðxÞ ¼vxv1 ðxANÞ and vm¼c.
Proof. For instance, see [8, III, § 7].
Theorem0.1 (Iida [2]). Let G be a p-group which contains Cn as a normal subgroup of index pm. Let fAFIrrðCnÞ. Suppose that fGAIrrðGÞ. Then man1, and GGGðn;mÞ.
In particular, when CnHG and ½G:Cn ¼p, Cn is always a normal subgroup of G. Hence we have:
Corollary 0.1. Let fAFIrrðCnÞ. Suppose that a group G containing Cn satisfies ½G:Cn ¼p and fGAIrrðGÞ. Then GGGðn;1Þ.
Theorem 0.2 ([6]). Let G be a p-group which contains Cn, and let fA FIrrðCnÞ. Suppose that ½G:Cn ¼pmþ1, fGAIrrðGÞ, and n3b2m. Fur- ther, suppose that there exists a subgroup H of G such that HjCn and
½G:H ¼p. Then
(1) GGGðn;mþ1Þ if Cn is a normal subgroup of G.
(2) GGGðn;m;1Þ if Cn is not a normal subgroup of G.
Corollary 0.2. Let G be a p-group which contains Cn and let fA FIrrðCnÞ. Suppose that ½G:Cn ¼p2, fGAIrrðGÞ and nb5. Then
(1) GGGðn;2Þ if Cn is a normal subgroup of G.
(2) GGGðn;1;1Þ if Cn is not a normal subgroup of G.
Our main theorem is the following:
Theorem. Let p be an odd prime. Let G be a p-group which contains Cn¼hai. We assume that ½G:Cnbp3. Define the subgroups of G by
N1¼NGðCnÞ; and Niþ1¼NGðNiÞ; for i¼1;2:
Let fAFIrrðCnÞ and 7an. Suppose that fGAIrrðGÞ, and ½N1:Cn ¼p.
Then
(1) N2=N1GC1 and N2GGðn;1;1Þ, (2) N3=N2GC1 and N3GGðn;1;1;1Þ.
Remark 1. Conversely, it is easy to see that the groups Gðn;1;1Þ and Gðn;1;1;1Þ satisfy the condition ðEX;CÞ, which is defined in section 3 of this paper. Hence these groups satisfy the conditions of Problem 1.
Remark 2. By results of Iida ([2], see Theorem 0.1. in this paper), we can see that N1GGðn;1Þ.
3. Some preleminary results
In this section, we state some results concerning the criterion of the irre- ducibilities of induced characters and others, which we need in section 4.
We denote by z¼zpn a primitive pnth root of unity. It is known that, for Cn¼hai, there are pn irreducible characters fn ð1anapnÞ of Cn:
fnðaiÞ ¼zni; ð1aiapnÞ:
The irreducible character fn is faithful if and only if ðn;pÞ ¼1. It is well- known that
AuthaiGðZ=pnZÞ GC Cn1
where ðZ=pnZÞ is the unit group of the factor ringZ=pnZandC is the cyclic group of order p1. Further, Cn1 is generated by the element 1þp in Z=pnZ.
First, we state the following result of Shoda (cf [1, p. 329]):
Proposition 2. Let G be a group and H be a subgroup of G. Let f be a linear character of H. Then the induced character fG of G is irreducible if
and only if, for each xAGH ¼ fgAGjgBHg, there exists hAxHx1VH such that fðhÞ0fðxhx1Þ. (Note that, when f is faithful, the condition fðhÞ0 fðxhx1Þ holds if and only if h0xhx1).
Using this result, we have the following:
Proposition 3. Let hai¼CnHG, and f be a faithful irreducible char- acter of Cn. Then the following conditions are equivalent:
(1) fG is irreducible,
(2) For each xAGCn, there exists yA hai Vxhaix1 such that xyx10y.
Definition. When the condition (2) of Proposition 3 holds, we say thatG satisfies (EX,C).
Let H be a group. For a normal subgroup N of H, and any g;hAH, we write
g1h ðmodNÞ
when g1hAN. For an element gAH, we denote by jgj the order of g.
4. Proof of Theorem
Let fAFIrrðCnÞ. Since fG¼ ðfN1ÞGAIrrðGÞ, we must have fN1 A IrrðN1Þ. Therefore, by Corollary 0.1, we can take an element b1AN1Cn¼ fgAN1jgBCng such that
N1¼ha;b1japn¼b1p¼1;b1ab11 ¼a1þpn1iGGðn;1Þ:
Proof of (1). Since Gis a p-group and ½G:N1bp2, by our assumption, we have
N1YNGðN1Þ ¼N2:
Take an elementvAN2N1¼ fgAN2jgBN1g such thatvpAN1. Denote by N10 the subgroup of G generated by v and the elements of N1. Then
½N10:N1 ¼p and N10jN1: Since Cn is not a normal subgroup of N10, we have
N10¼ha;b1;viGGðn;1;1Þ
by Corollary 0.2. Hence we may assume that the elementsa,b1, and vsatisfy the following relations:
apn¼b1p¼1; b1ab11 ¼a1þpn1; vav1¼a1þpn2b1;
vp¼b1; vb1v1¼b1: ðIÞ
Hereafter, we write b instead of b1 for the sake of simplicity.
Remark 1. More precisely, in [6], we have shown that there exist an integer s1;ðs1;pÞ ¼1 and vAN10, such that the elements a1 ¼as1;b, and v satisfy the same relations as (I). Since hai¼has1i, we can take a1 instead of a, and hence, we may assume that N10¼ha;b;vi and a;b and v satisfy the same relations as (I).
To prove the theorem, we need the following:
Lemma 1. For any integers i;j, the following equalities hold.
(i) ab1ba ðmodhapn1iÞ.
(ii) bapb1 ¼ap. (iii) ðaibjÞp¼aip. (iv) vap2v1¼ap2. (v) vjavj¼a1þjpn2bj.
(vi) ðaivjÞp1apivpj ¼apibj ðmodhapn1iÞ.
(vii) ðaivjÞp2 ¼ap2i.
Proof of Lemma 1. (i), (ii), (iii) and (iv) can be shown by direct cal- culations.
(v). Since nb7, by our assumption, we have vapn2v1¼apn2, by (iv).
Hence, vjavj¼a1þjpn2bj, for any jAZ.
(vi). By (i) and (v),
vjaivj¼ ða1þjpn2bjÞi1aið1þjpn2Þbij ðmodhapn1iÞ: Using this relation repeatedly, we can get
ðaivjÞp1apiaijpn2ð1þ2þþðp1ÞÞ
bijð1þ2þþðp1ÞÞ
vpj ðmodhapn1iÞ
¼apiaijpn2ðpðp1Þ=2Þbijðpðp1Þ=2Þ
vpj 1apivpj¼apibj ðmodhapn1iÞ;
since p is odd.
(vii) follows from (vi).
The assertion (1) follows from the following Claim I. N10¼N2.
Proof of Claim I. Suppose that N10YN2. Take an element wA N2N10¼ fgAN2jgBN10g such that wpAN10.
Write waw1¼ai0bj0 for some i0;j0AZ, 0ai0apn1, 0aj0ap1.
Then
wapw1¼ ðai0bj0Þp¼api0 by Lemma 1 (iii). Therefore
wp2apwp2¼apip
2 0 :
Since wp2AN1, we must have
i0p211 ðmod pn1Þ:
So,
i011 ðmod pn3Þ:
Hence we can write as i0¼1þk0pn3, and waw1¼a1þk0pn3bj0
for some integer k0. Since n3b4, by our assumption, we have vapn3v1¼apn3;
by Lemma 1 (iv). Hence
vpj0waw1vpþj0¼vpj0a1þk0pn3bj0vpþj0¼a1þk0pn3þðpj0Þpn2;
by Lemma 1 (v). This means that vpj0wAN1, which contradicts the hypoth- esis that wBN10. Hence the proof of Claim I is completed.
Proof of (2). Since G is a p-group and ½G:N2bp, by our assumption, we have
N2YNGðN2Þ ¼N3:
Take an element yAN3N2 such that ypAN2. Denote by N20 the subgroup of G generated by y and the elements of N2. Then
½N20:N2 ¼p and N20jN2: First, we show the following
Claim II. We can write as
yay1¼a1þkpn3vj; yby1¼apn1db;
yvy1¼apn2sbdv;
for some k;j;d;sAZ, such that ðk;pÞ ¼ ðj;pÞ ¼1.
Proof of ClaimII. First, we consider the elements yby1. By Lemma 1 (vi), jaivjjbp2 when ðj;pÞ ¼1. So, we must have yby1¼ad0bt0 for some d0;t0AZ. But
1¼ybpy1¼ ðad0bt0Þp¼ad0p; by Lemma 1 (iii). Therefore
d010 ðmod pn1Þ:
Hence, we may write d0¼pn1d and
yby1¼apn1dbt0; for some d AZ.
Next, consider the element yay1. Since yBN2, we must have yay1¼aivj;
for some jAZ, ðj;pÞ ¼1. By Lemma 1 (vi), yapy1¼ ðaivjÞp¼apiþmpn1bj for some mAZ. Since yay1B hai, yapy1 B hai, and yap2y1¼ ðapiþmpn1bjÞp¼ap2iA hai, we must have
ap2i0ap2; by the condition (EX,C). Therefore
iB h1þpn2i;
where h1þpn2i is the subgroup of ðZ=pnZÞ generated by 1þpn2. But ypap2yp¼ap2ip and ypAN2¼Gðn;1;1Þ. Hence
ipA h1þpn2i: Thus we may write as i¼1þkpn3 and
yay1¼a1þkpn3vj; for some integers k;j such that ðk;pÞ ¼ ðj;pÞ ¼1.
Since n1b6, by our assumption, we have, yapn1y1¼ ða1þkpn3vjÞpn1¼apn1:
by Lemma 1 (vii). Taking the conjugate of both sides of the equality,bab1¼ a1þpn1, b y y, we get
ðapn1dbt0Þða1þkpn3vjÞðapn1dbt0Þ1¼ ða1þkpn3vjÞapn1: Hence,
að1þkpn3Þð1þt0pn1Þvj¼a1þkpn3þpn1vj:
Therefore,
t011 ðmod pÞ;
and hence
yby1¼apn1db:
Finally, we consider the element yvy1. Write yvy1¼as0vh0. Then yby1¼yvpy1¼ ðas0vh0Þp¼aps0þepn1bh0;
for some integer e, by Lemma 1 (vi). Therefore apn1db¼aps0þepn1bh0; and, we have
h011 ðmod pÞ; and s010 ðmod pn2Þ:
Write
h0¼1þpl and s0 ¼pn2s;
for some l;sAZ. Then we have
yvy1¼apn2sv1þpl ¼apn2sblv:
Taking the conjugate of both sides of the equality, vav1¼a1þpn2b by y, we get
ðapn2sblvÞða1þkpn3vjÞðapn2sblvÞ1¼ ða1þkpn3vjÞapn2ðapn1dbÞ;
since yapn2y1¼apn2. Hence, we have
a1þkpn3þpn2þpn1lbvj¼a1þkpn3þpn2þpn1dbvj: Therefore,
d1l ðmod pÞ; Thus the proof of Claim II is completed.
Now, we consider the element yp ðAN2¼Gðn;1;1Þ ¼ha;b;viÞ. Write yp¼ ar0vh. Then
ar0vh¼yp¼yypy1¼yðar0vhÞy1¼ ða1þkpn3vjÞr0ðapn2sbdvÞh: Therefore
vh1vjr0vh ðmodN1¼ha;biÞ:
Since ðj;pÞ ¼1, we have r010 ðmod pÞ. Hence we can write as r0¼pr, and
yp¼aprvh; for some rAZ.
We show the following:
Claim III. There exists an integer e such that ðaeyÞp¼vpt0þh, for some integer t0. Further, ðh;pÞ ¼1.
Proof of Claim III. It is easy to see that hapn3i is a normal subgroup of N20. By Claim II, the following equalities hold:
yay11avj ðmodhapn3iÞ;
yby11b ðmodhapn3iÞ;
yvy11bdv ðmodhapn3iÞ:
Using these relations repeatedly, we can get
ypayp1abdjð1þ2þþðp1ÞÞvpj¼abdjðpðp1Þ=2Þvpj ¼avpj¼abj ðmodhapn3iÞ;
since p is odd. Hence we can write as
ypayp¼a1þbpn3bj;
for some bAZ. On the other hand, since yp¼aprvh, we have ypayp¼ ðaprvhÞaðaprvhÞ1¼a1þhpn2bh;
by Lemma 1 (v). Thus we have bj ¼bh, and, in particular, ðh;pÞ ¼1.
Define the subgroup H of N20 as
H ¼hapn3ihbi:
It is easy to see that N20jH, and the following equalities hold:
va1av ðmodHÞ; yay11avj ðmodHÞ;
yvy11v ðmodHÞ:
Using these relations repeatedly, we can get
ycamyc1amvmcj ðmodHÞ; for any m;cAZ. Using this equality, we have
ðamyÞp1apmvjmð1þ2þþðp1ÞÞyp¼apmvjmðpðp1Þ=2Þaprvh¼apðmþrÞvh ðmodHÞ; for any mAZ. Therefore we may write as
ðamyÞp ¼apðmþrÞþypn3btvh;
for some integers y and t. Note that y and t are not independent of the choice of m. If we set y1¼ary, then
y1p¼ay0pn3bt0vh¼ay0pn3vpt0þh;
for some integers y0 and t0. Further, set e¼ y0pn4r, and y2¼aey¼ay0pn4ry¼ay0pn4y1:
Since n4b3, by our assumption, we have y1apn4y11 ¼apn4: Hence,
y2p¼ ðaeyÞp¼ ðay0pn4y1Þp¼ay0pn3y1p¼vpt0þh: This completes the proof of Claim III.
Since ðpt0þh;pÞ ¼1, there exists k0AZ, such that ðpt0þhÞk011 ðmodp2Þ. Hence
y2k0p ¼vðpt0þhÞk0¼v:
Therefore
y2vy12 ¼v; and y2by12 ¼b:
Further, we have
y2ay12 ¼ay0pn4ryay1ay0pn4þr
¼arða1þkpn3vjÞar
¼ara1þkpn3ða1þjpn2bjÞrvj
1ara1þkpn3arbjrvj ðmodhapn2iÞ 1a1þkpn3bjrvj ðmodhapn2iÞ
¼a1þkpn3vpjrþj: Therefore
y2ay12 ¼a1þkpn3þdpn2vpjrþj;
for some integer d. If we set k1 ¼kþdp, then ðk1;pÞ ¼1, and y2ay12 ¼a1þk1pn3vpjrþj:
Since n3b4, by our assumption, we have y2apn3y12 ¼apn3: Thus,
y2payp2 ¼a1þk1pn2vpj¼a1þk1pn2bj: On the other hand, since y2p¼vpt0þh, we have
y2payp2 ¼ ðvpt0þhÞaðvpt0þhÞ1¼a1þðpt0þhÞpn2bpt0þh¼a1þðpt0þhÞpn2bh; by Lemma 1 (v). Therefore we have
k11pt0þh ðmod p2Þ;
and
j1h ðmodpÞ:
Summarizing the results, we have
y2ay12 ¼a1þk1pn3vjprþj¼a1þk1pn3bhrvj; y2p¼vk1;
y2by12 ¼b;
y2vy12 ¼v:
There exists an integer l1, such that
l1k111 ðmodp3Þ:
Since
k11h1j ðmod pÞ;
we have
l1k11l1j11 ðmod pÞ:
Hence, we may write as
l1j¼1þps2; for some s2AZ. Set y3 ¼y2l1. Then
y3ay13 ¼y2l1ayl2 1 ¼a1þpn3k1l1vl1ðjprþjÞ
¼a1þpn3vl1jðprþ1Þ¼a1þpn3vðps2þ1Þðprþ1Þ
¼a1þpn3vpðrþs2Þþ1¼a1þpn3br1v;
where r1¼rþs2. Further
y3p¼y2pl1¼vk1l1 ¼v; y3vy13 ¼v; and y3by13 ¼b:
If we set a0¼a1pr1 ð¼a1þðpnpÞr1Þ, then we have a0pn¼1 and ba0b1¼a01þpn1: Further, we have
va0v1¼va1pr1v1¼vaapr1v1¼ ða1þpn2bÞða1þpn2bÞpr1
¼ ða1þpn2bÞað1þpn2Þðpr1Þ¼að1þpn2Þð1pr1Þb¼a01þpn2b;
by using Lemma 1 (iii). We also have y3a0y13 ¼y3aapr1y13
¼ ða1þpn3br1vÞða1þpn3br1vÞpr1
1ða1þpn3br1vÞðapð1þpn3ÞvpÞr1 ðmodhapn1iÞ
¼ ða1þpn3br1vÞapr1ð1þpn3Þbr1;
1a1þpn3apr1ð1þpn3Þv ðmodhapn1iÞ
¼að1þpn3Þð1pr1Þv¼a01þpn3v:
Hence we may write as y3a0y13 ¼a01þpn3agpn1v, for some integer g. But a0gpn1 ¼agpn1, so
y3a0y13 ¼a01þpn3þgpn1v:
Finally, we set y4¼bgy3. Then we have
y4p¼y3p¼v; y4vy14 ¼v; and y4by14 ¼b:
Further
y4a0y14 ¼bgy3a0y13 bg¼bgða01þpn3þgpn1vÞbg
¼að1þp0 n3þgpn1Þð1þpn1Þgv¼a01þpn3v:
Therefore, the relations of the elements a0;b;v, and y4 are the same as that of Gðn;1;1;1Þ. So, the group N20¼ha0;b;v;y4i is clearly isomorphic to Gðn;1;1;1Þ.
We will complete the proof of (2), by showing the following:
Claim IV. N20¼N3.
Proof of Claim IV. Since the proof of this claim is similar to that of Claim I, we state only an outline of the proof. Suppose that N20YN3. Take an element gAN3N20 such that gpAN20. Then we can write as
ga0g1¼a01þk2pn4vj1
for some integers k2 and j1, by the same way as in the proof of Claim I.
Since n4b3, by our assumption, we have y4a0pn4y14 ¼a0pn4: Thus
y4p2j1ga0g1yp4 2þj1 ¼a01þk2pn4þðp2j1Þpn3:
This means that y4p2j1gAN1, which contradicts the hypothesis that gA N3N20. Hence the proof of Claim IV is completed.
References
[ 1 ] C. Curtis and I. Reiner: ‘‘Representation theory of finite groups and associative algebras’’, Interscience, New York, 1962.
[ 2 ] Y. Iida: Thep-groups with an irreducible character induced from a faithful linear character, Preprint.
[ 3 ] Y. Iida and T. Yamada: Extensions and induced characters of quaternion, dihedral and semidihedral groups, SUT J. Math. 27 (1991), 237–262.
[ 4 ] Y. Iida and T. Yamada: Types of faithful metacyclic 2-groups, SUT J. Math.28(1992), 23–46.
[ 5 ] K. Sekiguchi: Extensions of some 2-groups which preserve the irreducibilities of induced characters, Osaka J. Math. 37(2000).
[ 6 ] K. Sekiguchi: Irreducibilities of the induced characters of cyclic p-groups, Math. J. of Okayama Univ. 41 (1999).
[ 7 ] K. Sekiguchi: Extensions and the irreducibilities of induced characters of some 2-groups, Hokkaido Math. J. 31 (2002).
[ 8 ] H. Zassenhaus: ‘‘The theory of groups’’, Chelsea, New York, 1949.
Department of Civil Engineering Faculty of Engineering
Kokushikan University 4-28-1 Setagaya Setagaya-Ku
Tokyo 154-8515, Japan