1 Introduction

All groups considered in this paper are finite. \(G\) always means a group, \(|G|\) is the order of \(G\), \(\pi (G)\) denotes the set of all primes dividing \(|G|\) and \(G_{p}\) is a Sylow \(p\)-subgroup of \(G\) for some \(p\in \pi (G)\).

Let \(\fancyscript{F}\) be a class of groups. We call \(\fancyscript{F}\) a formation, provided that (1) if \(G\in \fancyscript{F}\) and \(H\unlhd G\), then \(G/H\in \fancyscript{F}\), and (2) if \(G/M\) and \(G/N\) are in \(\fancyscript{F}\), then \(G/(M\cap N)\) is in \(\fancyscript{F}\) for any normal subgroups \(M\), \(N\) of \(G\). A formation \(\fancyscript{F}\) is said to be saturated if \(G/\Phi (G)\in \fancyscript{F}\) implies that \(G \in \fancyscript{F}\). \(\fancyscript{U}\) will denote the class of all supersolvable groups. Clearly, \(\fancyscript{U}\) is a saturated formation.

A subgroup \(H\) of \(G\) is called \(s\)-quasinormal (or \(s\)-permutable, \(\pi \)-quasinormal) in \(G\) provided \(H\) permutes with all Sylow subgroups of \(G\), i.e, HP\(=\)PH for any Sylow subgroup \(P\) of \(G\). This concept was introduced by Kegel in [5] and has been studied extensively by Deskins [1] and Schmidt [11]. More recently, Zhang and Wang [14] generalized \(s\)-quasinormal subgroups to \(s\)-semipermutable subgroups. A subgroup \(H\) is said to be \(s\)-semipermutable in \(G\) if \(\textit{HG}_{p}=G_{p}H\) for any Sylow \(p\)-subgroup \(G_{p}\) of \(G\) with \((p,|H|)=1\). Clearly, every \(s\)-quasinormal subgroup of \(G\) is an \(s\)-semipermutable subgroup of \(G\), but the converse does not hold. Many authors consider minimal or maximal subgroups of a Sylow subgroup of a group when investigating the structure of \(G\), such as in [1] and [414], etc. For example, Han in [4] proves the following result.

Theorem 1.1

Let \(p\) be a prime dividing the order of a group \(G\) satisfying \((|G|,p-1)=1\) and \(P\) a Sylow \(p\)-subgroup of \(G\). Suppose there exists a nontrivial subgroup \(D\) such that \(1<|D|<|P|\) and every subgroup \(H\) of \(P\) with order \(|H|=|D|\) or with order \(2|D|\) (if \(P\) is a nonabelian 2-group and \(|P:D|>2\)) is \(s\)-semipermutable in \(G\), then \(G\) is \(p\)-nilpotent.

As another generalization of the \(s\)-quasinormality, Li et al. [7] introduce the following concept: A subgroup \(H\) of \(G\) is called an \(ss\)-quasinormal subgroup of \(G\) if there is a subgroup \(B\) of \(G\) such that \(G=HB\) and \(H\) permutes with every Sylow subgroup of \(B\). Many authors consider minimal or maximal subgroups of a Sylow subgroup of a group when investigating the structure of \(G\), such as in [69, 12], etc. In [13], Wei and Guo provide a result as follows.

Theorem 1.2

Let \(\fancyscript{F}\) be a saturated formation containing \(\fancyscript{U}\) and \(E\) be a normal subgroup of a group \(G\) such that \(G/E\in \fancyscript{F}\). Then \(G\in \fancyscript{F}\) if only if for every non-cyclic Sylow subgroup \(P\) of \(F^{*}(E)\), there is a subgroup \(D\) of \(P\) with \(1<|D|<|P|\) such that every subgroup \(H\) of \(P\) with order \(|D|\) or \(2|D|\) whenever \(|D|=2\) is \(ss\)-quasinormal in \(G\).

The aim of this article is to unify and improve above Theorems using \(s\)-semipermutable and \(ss\)-quasinormal subgroups. Our main theorems are as follows:

Theorem 1.3

(i.e., Theorem 3.5) Let \(\fancyscript{F}\) be a saturated formation containing \(\fancyscript{U}\), the class of all supersolvable groups and \(G\) a group with \(E\) as a normal subgroup of \(G\) such that \(G/E\in \fancyscript{F}\). Suppose that every non-cyclic Sylow subgroup \(P\) of \(F^{*}(E)\) has a subgroup \(D\) such that \(1<|D|<|P|\) and every subgroup \(H\) of \(P\) with order \(|H|=|D|\) or with order \(2|D|\) (if \(P\) is a nonabelian 2-group and \(|P:D|>2\)) is either \(s\)-semipermutable or \(ss\)-quasinormal in \(G\), where \(F^{*}(E)\) is the generalized Fitting subgroup of \(E\). Then \(G\in \fancyscript{F}\).

2 Basic definitions and preliminary results

In this section, we collect some known results that are useful later.

Lemma 2.1

Suppose that \(H\) is an \(s\)-semipermutable subgroup of \(G\). Then the following assertions hold.

  1. (i)

    If \(H\le K\le G\), then \(H\) is \(s\)-semipermutable in \(K\);

  2. (ii)

    Let \(N\) be a normal subgroup of \(G\). If H is a p-group for some prime \(p\in \pi (G)\), then \(HN/N\) is \(s\)-semipermutable in \(G/N\);

  3. (iii)

    If \(H\le O_{p}(G)\), then \(H\) is \(s\)-permutable in \(G\)

Proof

(i) is [14, Property 1], (ii) is [14, Property 2], and (iii) is [14, Lemma 3]. \(\square \)

Lemma 2.2

Let \(H\) be an \(ss\)-quasinormal subgroup of a group \(G\).

  1. (i)

    If \(H\le L\le G\), then \(H\) is \(ss\)-quasinormal in \(L\);

  2. (ii)

    If \(N\) is normal in \(G\), then \(HN/N\) is \(ss\)-quasinormal in \(G/N\);

  3. (iii)

    If \(H\le F(G)\), then \(H\) is \(s\)-quasinormal in \(G\);

  4. (iv)

    If \(H\) is a \(p\)-subgroup(\(p\) a prime), then \(H\) permutes with every Sylow \(q\)-subgroup of \(G\) with \(q\ne p\).

Proof

(i) and (ii) are [7, Lemma 2.1], (iii) is [7, Lemma 2.2], and (iv) is [7, Lemma 2.5]. \(\square \)

Lemma 2.3

([12]) Let \(G\) be a group and \(P\) a Sylow \(p\)-subgroup of \(G\), where \(p\) is the smallest prime dividing \(|G|\). If every maximal subgroup of \(P\) is \(s\)-semipermutable in \(G\), then \(G\) is \(p\)-nilpotent.

Lemma 2.4

([2] III, 5.2 and IV, 5.4). Suppose that \(p\) is a prime and \(G\) is a minimal non-\(p\)-nilpotent group, i.e., \(G\) is not a \(p\)-nilpotent group but whose proper subgroups are all \(p\)-nilpotent.

  1. (i)

    \(G\) has a normal Sylow \(p\)-subgroup \(P\) for some prime \(p\) and \(G=PQ\), where \(Q\) is a non-normal cyclic \(q\)-subgroup for some prime \(q\ne p\).

  2. (ii)

    \(P/\Phi (P)\) is a minimal normal subgroup of \(G/\Phi (P)\).

  3. (iii)

    The exponent of \(P\) is \(p\) or 4.

Lemma 2.5

([4]) Let \(N\) be an elementary abelian normal \(p\)-subgroup of a group \(G\). If there exists a subgroup \(D\) in \(N\) such that \(1<|D|<|N|\) and every subgroup \(H\) of \(N\) with \(|H|=|D|\) is \(s\)-semipermutable in \(G\), then there exists a maximal subgroup \(M\) of \(N\) such that \(M\) is normal in \(G\).

Lemma 2.6

([2], VI, 4.10) Assume that \(A\) and \(B\) are two subgroups of a group \(G\) and \(G\ne AB\). If \(AB^{g}=B^{g}A\) holds for any \(g\in G\), then either \(A\) or \(B\) is contained in a nontrivial normal subgroup of \(G\).

The generalized Fitting subgroup \(F^{*}(G)\) of \(G\) is the unique maximal normal quasinilpotent subgroup of \(G\). Its definition and important properties can be found in [3, X, 13]. We would like to give the following basic facts we will use in our proof.

Lemma 2.7

([3], X, [13]) Let \(G\) be a group and \(M\) a subgroup of \(G\).

  1. (i)

    If \(M\) is normal in \(G\), then \(F^{*}(M)\le F^{*}(G)\);

  2. (ii)

    \(F^{*}(G)\ne 1\) if \(G\ne 1\); in fact, \(F^{*}(G)/F(G)=Soc(F(G)C_{G}(F(G))/F(G))\);

  3. (iii)

    \(F^{*}(F^{*}(G))=F^{*}(G)\ge F(G)\); if \(F^{*}(G)\) is solvable, then \(F^{*}(G)=F(G)\).

Lemma 2.8

([10]) Let \(\fancyscript{F}\) be a saturated formation containing \(\fancyscript{U}\), the class of all supersolvable groups and \(G\) a group with \(E\) as a normal subgroup of \(G\) such that \(G/E\in \fancyscript{F}\). Suppose that every non-cyclic Sylow subgroup \(P\) of \(F^{*}(E)\) has a subgroup \(D\) such that \(1<|D|<|P|\) and every subgroup \(H\) of \(P\) with order \(|H|=|D|\) or with order \(2|D|\) (if \(P\) is a nonabelian 2-group and \(|P:D|>2\)) is weakly \(s\)-permutable in \(G\), where \(F^{*}(E)\) is the generalized Fitting subgroup of \(E\). Then \(G\in \fancyscript{F}\).

3 Main results

In this section, we will prove our main results.

Theorem 3.1

Let \(p\) be the smallest prime dividing the order of a group \(G\) and \(P\) be a Sylow \(p\)-subgroup of \(G\). If every maximal subgroup of \(P\) is either \(s\)-semipermutable or \(ss\)-quasinormal in \(G\). Then \(G\) is \(p\)-nilpotent.

Proof

Assume that the theorem is not true and let \(G\) be a counterexample of minimal order. We derive a contradiction in several steps. \(\square \)

By Lemmas 2.1 and 2.2, the following two steps are obvious.

Step 1. \(O_{p^{\prime }}(G)=1\).

Step 2. \(G\) has a unique minimal normal subgroup \(N\) and \(G/N\) is \(p\)-nilpotent. Moreover, \(\Phi (G)=1\).

Step 3. \(O_{p}(G)=1\).

If \(O_{p}(G)\ne 1\), then step 2 yields \(N\le O_{p}(G)\) and \(\Phi (O_{p}(G))\le \Phi (G)=1\). Therefore, \(G\) has a maximal subgroup \(M\) such that \(G=MN\) and \(G/N\cong M\) is \(p\)-nilpotent. Since \(O_{p}(G)\cap M\) is normalized by \(N\) and \(M\), we conclude that \(O_{p}(G)\cap M\) is normal in \(G\). The uniqueness of \(N\) yields \(N=O_{p}(G)\). Clearly, \(P=N(P\cap M)\). Furthermore, \(P\cap M<P\), and, thus there exists a maximal subgroup \(P_{1}\) of \(P\) such that \(P\cap M\le P_{1}\). Hence, \(P=NP_{1}\). By hypothesis, \(P_{1}\) is \(s\)-semipermutable or \(ss\)-quasinormal in \(G\). In either case, we have that \(P_{1}M_{q}\) is a group for \(q\ne p\). Hence

$$\begin{aligned} P_{1}\langle M_{p}, M_{q}|q\in \pi (M), q\ne p\rangle =P_{1}M \end{aligned}$$

is a group. Then \(P_{1}M=M\) or \(G\) by maximality of \(M\). If \(P_{1}M=G\), then

$$\begin{aligned} P=P\cap P_{1}M=P_{1}(P\cap M)=P_{1}, \end{aligned}$$

a contradiction. If \(P_{1}M=M\), then \(P_{1}\le M\). Therefore, \(P_{1}\cap N=1\) and \(N\) is of prime order. Then the \(p\)-nilpotency of \(G/N\) implies the \(p\)-nilpotency of \(G\), a contradiction. Therefore, \(O_{p}(G)=1\).

Step 4. The final contradiction.

Assume that \(P\) has a maximal subgroup \(P_{1}\) which is \(ss\)-quasinormal in \(G\), then there exists a subgroup \(B\) of \(G\) such that \(G=P_{1}B\) and \(P_{1}\) permutes with every Sylow subgroup of \(B\). Noticing that \(H\cap B\) is \(s\)-quasinormal in \(B\), we have \(P_{1}\cap B\le O_{p}(B)\). Since \(p\) is the smallest prime dividing the order of \(G\) and \(H\cap B\) is a maximal subgroup of a Sylow \(p\)-subgroup of \(B\), we see that \(B/O_{p}(B)\) is \(p\)-nilpotent and so \(B\) is \(p\)-solvable. By [2, Chapter VI, Hauptsatz 1.7], we may assume that \(K\) is a Hall \(p^{\prime }\)-subgroup of \(B\), \(\pi (K)= \{p_{2},\ldots , p_{s}\}\) and \(P_{i}\in \) Syl\(_{p_{i}}(K)\). Since \(P_{1}\) permutes with every \(P_{i}\), \(P_{1}K=KP_{1}\) and therefore \(P_{1}K\) is normal in \(G\). Now let \(\{H_{j}: j=1,\ldots , t\}\) be the set of all maximal subgroup of \(P\). Then, by the above, there is \(K_{j}\) such that \(H_{j}K_{j}\) is normal in \(G\) for \(j=1,\ldots , t\), respectively. Let \(N=\bigcap \{H_{j}K_{j}: j=1,\ldots , t\}\). Then \(N\unlhd G\) and \(G/N\) is a \(p\)-group. Thus we have \(P\cap N=\bigcap _{j=1}^{t}((H_{j}K_{j})\cap P)=\bigcap _{j=1}^{t}((P\cap K_{j})H_{j})=\bigcap _{j=1}^{t}H_{j}\le \Phi (P)\). By Tate\(^{\prime }\)s theorem [2, Chapter IV, Satz 4.7], \(N\) is \(p\)-nilpotent and hence \(G\) is \(p\)-nilpotent, a contradiction. Now we may assume that all maximal subgroups of \(P\) are \(s\)-semipermutable in \(G\). Then \(G\) is \(p\)-nilpotent by Lemma 2.3, a contradiction.

Theorem 3.2

Let \(p\) be the smallest prime dividing the order of a group \(G\) and \(P\) be a Sylow \(p\)-subgroup of \(G\). If \(P\) has a subgroup \(D\) such that \(1<|D|<|P|\) and every subgroup \(H\) of \(P\) with order \(|H|=|D|\) or with order \(2|D|\) (if \(P\) is a nonabelian 2-group and \(|P:D|>2\)) is either \(s\)-semipermutable or \(ss\)-quasinormal in \(G\). Then \(G\) is \(p\)-nilpotent.

Proof

Suppose that the theorem is false and let \(G\) be a counterexample of minimal order. We will derive a contradiction in several steps. \(\square \)

Step 1. \(O_{p^{\prime }}(G)=1\).

If \(O_{p^{\prime }}(G)\ne 1\), Lemma 2.1 (ii) and Lemma 2.2 (iii) guarantee that \(G/O_{p^{\prime }}(G)\) satisfies the hypotheses of the theorem. Thus \(G/O_{p^{\prime }}(G)\) is \(p\)-nilpotent by the choice of \(G\). Then \(G\) is \(p\)-nilpotent, a contradiction.

Step 2. \(|D|>p\).

Suppose that \(|D|=p\). Since \(G\) is not \(p\)-nilpotent, \(G\) has a minimal non-\(p\)-nilpotent subgroup \(G_{1}\). By Lemma 2.4 (i), \(G_{1}=[P_{1}]Q\), where \(P_{1}\in \) Syl\(_{p}(G_{1})\) and \(Q\in \) Syl\(_{q}(G_{1})\), \(p\ne q\). Let \(X/\Phi (P_{1})\) be a subgroup of \(P_{1}/\Phi (P_{1})\) of order \(p\), \(x\in X\backslash \Phi (P_{1})\) and \(L=\langle x\rangle \). Then \(L\) is of order \(p\) or 4 by Lemma 2.4 (iii). By the hypotheses, \(L\) is either \(s\)-semipermutable or \(ss\)-quasinormal in \(G\), thus in \(G_{1}\) by Lemmas 2.1 (i) and 2.2 (i). First, suppose that \(L\) is \(ss\)-quasinormal in \(G_{1}\), then \(L\) is \(s\)-quasinormal in \(G_{1}\) by Lemma 2.2 (iii). Thus \(LQ\le G_{1}\). Therefore, \(LQ=L\times Q\). Then \(G_{1}=P_{1}\times Q\). Hence \(G_{1}\) is \(p\)-nilpotent, contrary to the choice of \(G_{1}\). Therefore, \(L=\langle x\rangle \) is \(s\)-semipermutable in \(G_{1}\) for every element \(x\in P_{1}\). Thus \(LQ\le G_{1}\). Therefore, \(LQ=L\times Q\). Then \(G_{1}=P_{1}\times Q\), a contradiction.

Step 3. \(|P:D|>p\).

By Theorem 3.1.

Step 4. \(P\) has a subgroup \(D\) such that \(1<|D|<|P|\) and every subgroup \(H\) of \(P\) with order \(|H|=|D|\) or with order \(2|D|\) (if \(P\) is a nonabelian 2-group and \(|P:D|>2\)) is \(s\)-semipermutable in \(G\). Assume that \(H\le P\) such that \(|H|=|D|\) and \(H\) is \(ss\)-quasinormal in \(G\). Then we may assume \(G\) has a normal subgroup \(M\) such that \(|G:M|=p\) and \(G=HM\). Since \(|P:D|>p\) by Step 3, \(M\) satisfies the hypotheses of the theorem. The choice of \(G\) yields that \(M\) is \(p\)-nilpotent. It is easy to see that \(G\) is \(p\)-nilpotent, contrary to the choice of \(G\).

Step 5. If \(N\le P\) and \(N\) is minimal normal in \(G\), then \(|N|\le |D|\).

Suppose that \(|N|>|D|\). Since \(N\le O_{p}(G)\), \(N\) is elementary abelian. By Lemma 2.5, \(N\) has a maximal subgroup which is normal in \(G\), contrary to the minimality of \(N\).

Step 6. Suppose that \(N\le P\) and \(N\) is minimal normal in \(G\). Then \(G/N\) is \(p\)-nilpotent.

If \(|N|<|D|\), \(G/N\) satisfies the hypotheses of the theorem by Lemma 2.1 (ii). Thus \(G/N\) is \(p\)-nilpotent by the minimal choice of \(G\). So we may suppose that \(|N|=|D|\) by Step 5. We will show that every cyclic subgroup of \(P/N\) of order \(p\) or order 4 (when \(P/N\) is a non-abelian 2-group) is \(s\)-semipermutable in \(G/N\). Let \(K\le P\) and \(|K/N|=p\). By Step 2, \(N\) is non-cyclic, so are all subgroups containing \(N\). Hence there is a maximal subgroup \(L\ne N\) of \(K\) such that \(K=NL\). Of course, \(|N|=|D|=|L|\). Since \(L\) is \(s\)-semipermutable in \(G\) by the hypotheses, \(K/N=LN/N\) is \(s\)-semipermutable in \(G/N\) by Lemma 2.1 (ii). If \(p=2\) and \(P/N\) is non-abelian, take a cyclic subgroup \(X/N\) of \(P/N\) of order 4. Let \(K/N\) be maximal in \(X/N\). Then \(K\) is maximal in \(X\) and \(|K/N|\)=2. Since \(X\) is non-cyclic and \(X/N\) is cyclic, there is a maximal subgroup \(L\) of \(X\) such that \(N\) is not contained in \(L\). Thus \(X=LN\) and \(|L|=|K|=2|D|\). By the hypotheses, \(L\) is \(s\)-semipermutable in \(G\). By Lemma 2.1 (ii), \(X/N=LN/N\) is \(s\)-semipermutable in \(G/N\). Hence \(G/N\) satisfies the hypotheses. By the minimal choice of \(G\), \(G/N\) is \(p\)-nilpotent.

Step 7. \(O_{p}(G)=1\).

Suppose that \(O_{p}(G)\ne 1\). Take a minimal normal subgroup \(N\) of \(G\) contained in \(O_{p}(G)\). By Step 6, \(G/N\) is \(p\)-nilpotent. It is easy to see that \(N\) is the unique minimal normal subgroup of \(G\) contained in \(O_{p}(G)\). Furthermore, \(O_{p}(G)\cap \Phi (G)=1\). Hence \(O_{p}(G)\) is an elementary abelian \(p\)-group. On the other hand, \(G\) has a maximal subgroup \(M\) such that \(G=MN\) and \(M\cap N=1\). It is easy to deduce that \(O_{p}(G)\cap M=1\), \(N=O_{p}(G)\) and \(M\cong G/N\) is \(p\)-nilpotent. Then \(G\) can be written as \(G=N(M\cap P)M_{p^{\prime }}\), where \(M_{p^{\prime }}\) is the normal \(p\)-complement of \(M\). Pick a maximal subgroup \(S\) of \(M_{p}=P\cap M\). Then \(NSM_{p^{\prime }}\) is a subgroup of \(G\) with index \(p\). Since \(p\) is the minimal prime in \(\pi (G)\), we know that \(NSM_{p^{\prime }}\) is normal in \(G\). Now by Step 3 and the induction, we have \(NSM_{p^{\prime }}\) is \(p\)-nilpotent. Therefore, \(G\) is \(p\)-nilpotent, a contradiction.

Step 8. The minimal normal subgroup \(L\) of \(G\) is not \(p\)-nilpotent.

If \(L\) is \(p\)-nilpotent, then it follows from the fact that \(L_{p^{\prime }}\) char \(L\lhd G\) that \(L_{p^{\prime }}\le O_{p^{\prime }}(G)=1\). Thus \(L\) is a \(p\)-group. Whence \(L\le O_{p}(G)=1\) by Step 7, a contradiction.

Step 9. \(G\) is a non-abelian simple group.

Suppose that \(G\) is not a simple group. Take a minimal normal subgroup \(L\) of \(G\). Then \(L<G\). If \(|L|_{_{p}}>|D|\), then \(L\) is \(p\)-nilpotent by the minimal choice of \(G\), contrary to Step 8. If \(|L|_{p}\le |D|\). Take \(P_{*}\ge L\cap P\) such that \(|P_{*}|=p|D|\). Hence \(P_{*}\) is a Sylow \(p\)-subgroup of \(P_{*}L\). Since every maximal subgroup of \(P_{*}\) is of order \(|D|\), every maximal subgroup of \(P_{*}\) is \(s\)-semipermutable in \(G\) by hypotheses, thus in \(P_{*}L\) by Lemma 2.1 (i). Now applying Theorem 3.1, we get \(P_{*}L\) is \(p\)-nilpotent. Therefore, \(L\) is \(p\)-nilpotent, contrary to Step 8.

Step 10. The final contradiction.

Suppose that \(H\) is a subgroup of \(P\) with \(|H|=|D|\) and \(Q\) is a Sylow \(q\)-subgroup with \(q\ne p\). Then \(HQ^{g}=Q^{g}H\) for any \(g\in G\) by the hypotheses that \(H\) is \(s\)-semipermutable in \(G\). Since \(G\) is simple by Step 9, \(G=HQ\) from Lemma 2.6, the final contradiction.

The following corollary is immediate from Theorem 3.2.

Corollary 3.3

Suppose that \(G\) is a group. If every non-cyclic Sylow subgroup of \(G\) has a subgroup \(D\) such that \(1<|D|<|P|\) and every subgroup \(H\) of \(P\) with order \(|H|=|D|\) or with order \(2|D|\) (if \(P\) is a nonabelian 2-group and \(|P:D|>2\)) is either \(s\)-semipermutable or \(ss\)-quasinormal in \(G\), then \(G\) has a Sylow tower of supersolvable type.

Theorem 3.4

Let \(\fancyscript{F}\) be a saturated formation containing \(\fancyscript{U}\), the class of all supersolvable groups and \(G\) a group with \(E\) as a normal subgroup of \(G\) such that \(G/E\in \fancyscript{F}\). Suppose that every non-cyclic Sylow subgroup of \(E\) has a subgroup \(D\) such that \(1<|D|<|P|\) and every subgroup \(H\) of \(P\) with order \(|H|=|D|\) or with order \(2|D|\) (if \(P\) is a nonabelian 2-group and \(|P:D|>2\)) is either \(s\)-semipermutable or \(ss\)-quasinormal in \(G\). Then \(G\in \fancyscript{F}\).

Proof

Suppose that \(P\) is a non-cyclic Sylow \(p\)-subgroup of \(E\), \(\forall p\in \pi (E)\). Since \(P\) has a subgroup \(D\) such that \(1<|D|<|P|\) and every subgroup \(H\) of \(P\) with order \(|H|=|D|\) or with order \(2|D|\) (if \(P\) is a nonabelian 2-group and \(|P:D|>2\)) is either \(s\)-semipermutable or \(ss\)-quasinormal in \(G\) by hypotheses, thus in \(E\) by Lemma 2.1 (i) and Lemma 2.2 (i). Applying Corollary 3.3, we conclude that \(E\) has a Sylow tower of supersolvable type. Let \(q\) be the maximal prime divisor of \(|E|\) and \(Q\in \) Syl\(_{q}(E)\). Then \(Q\unlhd G\). Since \((G/Q, E/Q)\) satisfies the hypotheses of the theorem, by induction, \(G/Q\in \fancyscript{F}\). For any subgroup \(H\) of \(Q\) with \(|H|=|D|\), since \(Q\le O_{q}(G)\), \(H\) is \(s\)-permutable in \(G\) by Lemmas 2.1 (iii) and 2.2 (iii). Since \(s\)-permutable implies weakly \(s\)-permutable and \(F^{*}(Q)=Q\) by Lemma 2.7, we get \(G\in \fancyscript{F}\) by applying Lemma 2.8. \(\square \)

Theorem 3.5

Let \(\fancyscript{F}\) be a saturated formation containing \(\fancyscript{U}\), the class of all supersolvable groups and \(G\) a group with \(E\) as a normal subgroup of \(G\) such that \(G/E\in \fancyscript{F}\). Suppose that every non-cyclic Sylow subgroup of \(F^{*}(E)\) has a subgroup \(D\) such that \(1<|D|<|P|\) and every subgroup \(H\) of \(P\) with order \(|H|=|D|\) or with order \(2|D|\) (if \(P\) is a nonabelian 2-group and \(|P:D|>2\)) is either \(s\)-semipermutable or \(ss\)-quasinormal in \(G\). Then \(G\in \fancyscript{F}\).

Proof

We distinguish two cases:

Case 1. \(\fancyscript{F}=\fancyscript{U}\).

Let \(G\) be a minimal counter-example.

Step 1. Every proper normal subgroup \(N\) of \(G\) containing \(F^{*}(E)\) (if it exists) is supersolvable. \(\square \)

If \(N\) is a proper normal subgroup of \(G\) containing \(F^{*}(E)\), then \(N/N\cap E\cong NE/E\) is supersolvable. By Lemma 2.7 (iii), \(F^{*}(E)=F^{*}(F^{*}(E))\le F^{*}(E\cap N)\le F^{*}(E)\), so \(F^{*}(E\cap N)=F^{*}(E)\). For any Sylow subgroup \(P\) of \(F^{*}(E\cap N)=F^{*}(E)\), \(P\) has a subgroup \(D\) such that \(1<|D|<|P|\) and every subgroup \(H\) of \(P\) with order \(|H|=|D|\) or with order \(2|D|\) (if \(P\) is a nonabelian 2-group and \(|P:D|>2\)) is either \(s\)-semipermutable or \(ss\)-quasinormal in \(G\) by hypotheses, thus in \(N\) by Lemma 2.1 (i) and Lemma 2.2 (i). So \(N\) and \(N\cap H\) satisfy the hypotheses of the theorem, the minimal choice of \(G\) implies that \(N\) is supersolvable.

Step 2. \(E=G\).

If \(E<G\), then \(E\in \fancyscript{U}\) by Step 1. Hence \(F^{*}(E)=F(E)\) by Lemma 2.7. It follows that every Sylow subgroup of \(F^{*}(E)\) is normal in \(G\). By Lemmas 2.1 (iii) and 2.2 (iii), every non-cyclic Sylow subgroup of \(F^{*}(E)\) has a subgroup \(D\) such that \(1<|D|<|P|\) and every subgroup \(H\) of \(P\) with order \(|H|=|D|\) or with order \(2|D|\) (if \(P\) is a nonabelian 2-group and \(|P:D|>2\)) is \(s\)-permutable in \(G\). Applying Lemma 2.8 for the special case \(\fancyscript{F}=\fancyscript{U}\), \(G\in \fancyscript{U}\), a contradiction.

Step 3. \(F^{*}(G)=F(G)<G\).

If \(F^{*}(G)=G\), then \(G\in \fancyscript{F}\) by Theorem 3.4, contrary to the choice of \(G\). So \(F^{*}(G)<G\). By Step 1, \(F^{*}(G)\in \fancyscript{U}\) and \(F^{*}(G)=F(G)\) by Lemma 2.7.

Step 4. The final contradiction.

Since \(F^{*}(G)=F(G)\), each non-cyclic Sylow subgroup of \(F^{*}(G)\) has a subgroup \(D\) such that \(1<|D|<|P|\) and every subgroup \(H\) of \(P\) with order \(|H|=|D|\) or with order \(2|D|\) (if \(P\) is a nonabelian 2-group and \(|P:D|>2\)) is \(s\)-permutable in \(G\) by Lemmas 2.1 (iii) and 2.2 (iii). Applying Lemma 2.8, \(G\in \fancyscript{U}\), a contradiction.

Case 2. \(\fancyscript{F}\ne \fancyscript{U}\).

By hypotheses, every non-cyclic Sylow subgroup of \(F^{*}(E)\) has a subgroup \(D\) such that \(1<|D|<|P|\) and every subgroup \(H\) of \(P\) with order \(|H|=|D|\) or with order \(2|D|\) (if \(P\) is a nonabelian 2-group and \(|P:D|>2\)) is either \(s\)-semipermutable or \(ss\)-quasinormal in \(G\), thus in \(E\) Lemma 2.1 (i) and Lemma 2.2 (i). Applying Case 1, \(E\in \fancyscript{U}\). Then \(F^{*}(E)=F(E)\) by Lemma 2.7. It follows that each Sylow subgroup of \(F^{*}(E)\) is normal in \(G\). By Lemmas 2.1 (iii) and 2.2 (iii), each non-cyclic Sylow subgroup of \(F^{*}(E)\) has a subgroup \(D\) such that \(1<|D|<|P|\) and every subgroup \(H\) of \(P\) with order \(|H|=|D|\) or with order \(2|D|\) (if \(P\) is a nonabelian 2-group and \(|P:D|>2\)) is \(s\)-permutable in \(G\). Applying Lemma 2.8, \(G\in \fancyscript{F}\). These complete the proof of the theorem.

The following corollaries are immediate from Theorem 3.5.

Corollary 3.6

Let \(\fancyscript{F}\) be a saturated formation containing \(\fancyscript{U}\). Suppose that \(G\) is a group with a normal subgroup \(E\) such that \(G/E\in \fancyscript{F}\). Then \(G\in \fancyscript{F}\) if and only if every maximal subgroup of any Sylow subgroup of \(F^{*}(E)\) is \(s\)-semipermutable in \(G\).

Corollary 3.7

Let \(\fancyscript{F}\) be a saturated formation containing \(\fancyscript{U}\). Suppose that \(G\) is a group with a normal subgroup \(E\) such that \(G/E\in \fancyscript{F}\). Then \(G\in \fancyscript{F}\) if and only if every cyclic subgroup of any Sylow subgroup of \(F^{*}(E)\) of prime order or order 4 is \(s\)-semipermutable in \(G\).

Corollary 3.8

([6], Theorem 3.3) Let \(\fancyscript{F}\) be a saturated formation containing \(\fancyscript{U}\). Suppose that \(G\) is a group with a normal subgroup \(H\) such that \(G/H\in \fancyscript{F}\). Then \(G\in \fancyscript{F}\) if and only if every maximal subgroup of any Sylow subgroup of \(F^{*}(H)\) is \(ss\)-quasinormal in \(G\).

Corollary 3.9

([6], Theorem 3.7) Let \(\fancyscript{F}\) be a saturated formation containing \(\fancyscript{U}\). Suppose that \(G\) is a group with a normal subgroup \(H\) such that \(G/H\in \fancyscript{F}\). Then \(G\in \fancyscript{F}\) if and only if every cyclic subgroup of any Sylow subgroup of \(F^{*}(H)\) of prime order or order 4 is \(ss\)-quasinormal in \(G\).

Corollary 3.10

([8], Theorem 3.4) Let \(\fancyscript{F}\) be a saturated formation containing \(\fancyscript{U}\). Suppose that \(G\) is a group with a normal subgroup \(E\) such that \(G/E\in \fancyscript{F}\). Then \(G\in \fancyscript{F}\) if and only if every maximal subgroup of any Sylow subgroup of \(F^{*}(E)\) is \(s\)-quasinormal in \(G\).

Corollary 3.11

([9], Theorem 3.3) Let \(\fancyscript{F}\) be a saturated formation containing \(\fancyscript{U}\). Suppose that \(G\) is a group with a normal subgroup \(E\) such that \(G/E\in \fancyscript{F}\). Then \(G\in \fancyscript{F}\) if and only if every cyclic subgroup of any Sylow subgroup of \(F^{*}(E)\) of prime order or order 4 is \(s\)-quasinormal in \(G\).