Square-free class sizes in products of groups
M. J. Felipe, A. Mart\'inez-Pastor, V. M. Ortiz-Sotomayor

TL;DR
This paper investigates the structure of factorized groups, particularly when conjugacy class sizes of certain elements are not divisible by the square of a prime, with special focus on mutually permutable products.
Contribution
It provides new structural insights into groups factorized as $G=AB$, especially under conditions on conjugacy class sizes related to a prime $p$.
Findings
Structural properties of $G=AB$ when conjugacy class sizes are not divisible by $p^2$
Special results for mutually permutable products
Conditions influencing the group's composition based on class size divisibility
Abstract
We obtain some structural properties of a factorised group , given that the conjugacy class sizes of certain elements in are not divisible by , for some prime . The case when is a mutually permutable product is especially considered.
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Square-free class sizes in products of groups
M. J. Felipe A. Martínez-Pastor V. M. Ortiz-Sotomayor The first author is supported by Proyecto Prometeo II/2015/011, Generalitat Valenciana (Spain), and the second author is supported by Proyecto MTM2014-54707-C3-1-P, Ministerio de Economía, Industria y Competitividad (Spain). The results in this paper are part of the third author’s Ph.D. thesis, and he acknowledges the predoctoral grant ACIF/2016/170, Generalitat Valenciana (Spain). The authors wish to thank John Cossey for helpful conversations during his last visit to Valencia.
Instituto Universitario de Matemática Pura y Aplicada (IUMPA), Universitat Politècnica de València, Camino de Vera, s/n, 46022, Valencia, Spain.
🖂: [email protected], [email protected], [email protected]
Abstract
We obtain some structural properties of a factorised group , given that the conjugacy class sizes of certain elements in are not divisible by , for some prime . The case when is a mutually permutable product is especially considered.
Keywords Finite groups Soluble groups Products of subgroups Conjugacy classes
2010 MSC 20D10 20D40 20E45
1 Introduction
All groups considered thorogouth this paper are finite. Over the last years, many authors have investigated the influence of conjugacy class sizes on the structure of finite groups. In the meantime, numerous studies in the framework of group theory have focused in factorised groups. In this setting, a central question is how the structure of the factors affects the structure of the whole group, in particular when they are connected by certain permutability properties. The purpose of this paper is to show new achievements which combine both current perspectives in finite groups. More precisely, our aim is to get some information about a factorised group, provided that the conjugacy class sizes of some elements of its factors are square-free.
The earlier starting point of our investigation can be traced back to the paper of Chillag and Herzog ([6]), where the structure of a group in which all elements have square-free conjugacy class sizes was first analysed. Next, in [7], Cossey and Wang localised one of the main theorems in [6] for a fixed prime , that is, they considered conjugacy class sizes not divisible by , for certain prime . Later on, this study was improved by Li in [13], and by Liu, Wang, and Wei in [14], by replacing conditions on all conjugacy classes by those referring only to conjugacy classes of either -regular elements or prime power order elements, using the classification theorem of finite simple groups (CFSG). These authors also first obtained some preliminary results in factorised groups. This research was extended in 2012 by Ballester-Bolinches, Cossey and Li in [3], through mutually permutable products. More recently, in 2014, Qian and Wang ([15]) have gone a step further by considering just conjugacy class sizes of -regular elements of prime power order (although not in factorised groups).
In the context of factorised groups, and aiming to obtain criteria for products of supersoluble subgroups to be supersoluble, several authors have considered products in which certain subgroups of the factors permute (see [4] for a detailed account). In this scene, we are interested in mutually permutable products, factorised groups such that the subgroups and are mutually permutable, i.e., permutes with every subgroup of and permutes with every subgroup of (see also [5]). Obviously, if and are normal in , then they are mutually permutable.
We recall that, for a group , the set is the conjugacy class of the element , and denotes the conjugacy class size of . If is a prime number, we say that is a -regular element if its order is not divisible by , and that it is a -element if its order is a power of . Moreover, if is an integer, let denote the highest power of dividing . The th group of order in the SmallGroups library [9] of GAP will be identified by . The remainder notation is standard and is taken mainly from [8]. We also refer to this book for details about classes of groups.
In this paper, motivated by the above development, at first we focus on the case of -groups, extending for factorised groups the well-known Knoche’s theorem (see [12]).
Theorem A**.**
Let be a prime number and let be a -group such that does not divide for all . Then , is elementary abelian and .
Our next goal in the paper is to prove the following theorem, regarding mutually permutable products.
Theorem B**.**
Let be the mutually permutable product of the subgroups and , and let be a prime such that . If does not divide for any -regular element of prime power order, then:
(1)* is soluble.*
(2)* is -nilpotent.*
(3)* The Sylow -subgroups of are elementary abelian.*
In the particular case when , we recover [15, Theorem A] (see Section 3, Corollary 3.5). We remark that both results use the CFSG.
Moreover, in relation to the third assertion, we have found a gap in one of the statements in [7, Theorem 1], as it is reported in Remark 3.7 (a).
On the other hand, we point out that it is possible to find examples of groups factorised as a product of two (mutually permutable) subgroups which satisfy the hypotheses of Theorem B for some fixed prime and, however, there exist elements such that divides either or (see Remark 3.7 (b)).
The next theorem generalises the last assertion of [3, Theorem 1.3] regarding -soluble groups, by considering only prime power order elements:
Theorem C**.**
Let be the mutually permutable product of the subgroups and , and let be a prime. Suppose that for every prime power order -regular element , is not divisible by . If is -soluble, then is -supersoluble.
In the line of [6, Theorem 1] and [7, Theorem 2], if we consider all prime numbers, then we obtain some information about the structure of the derived subgroup of a factorised group .
Theorem D**.**
Let be the product of the subgroups and , and assume that is supersoluble. Suppose that every prime power order element has square-free conjugacy class size. Then:
(1)* is abelian.*
(2)* The Sylow subgroups of are elementary abelian.*
(3)* has Sylow -subgroups of order at most , for every prime .*
If we limit our conditions only to -regular elements, as a consequence of Theorems B and C, we obtain the following result which extends [3, Corollary 1.5] (see Corollary 4.3) for prime power order elements, and also a theorem of [13], for products of groups.
Theorem E**.**
Let be the mutually permutable product of the subgroups and . Suppose that for every prime and every prime power order -regular element , is not divisible by . Then is supersoluble, and has elementary abelian Sylow subgroups.
We remark that the first statement in Theorem D is not further true under the weaker hypotheses of the above theorem, even for arbitrary groups not necessarily factorised, as pointed out in [15]. Indeed, as a result of Theorem E, the supersolubility condition in Theorem D can be exchanged by the mutual permutability of the factors.
On the other hand, with the stronger assumption that all -regular elements of the factors (not only those of prime power order) have conjugacy class sizes not divisible by , we get extra information about the orders of the Sylow -subgroups of , extending partially [7, Theorem 2].
Theorem F**.**
Let be the mutually permutable product of the subgroups and . Suppose that for every prime and every -regular element , is not divisible by . Then the order of a Sylow -subgroup of is at most .
In summary, when dealing with mutually permutable products, the next corollary follows directly from the above theorems.
Corollary G**.**
Let be the mutually permutable product of the subgroups and . Suppose that is square-free for each element . Then is supersoluble, and both and have elementary abelian Sylow subgroups. Moreover, is abelian, and both groups and have Sylow -subgroups of order at most , for each prime .
In Section 3 we prove Theorems A, B and C, which refer to class sizes not divisible by , for a fixed prime . Theorems E, D and F, which consider square-free conjugacy class sizes (for all primes), are proved in Section 4. In both cases we will illustrate the scope of the results presented with some examples.
2 Preliminary results
We use the following elementary properties frequently, sometimes without further reference.
Lemma 2.1**.**
Let be a normal subgroup of a group , and let be a prime. Then:
(a)* divides , for any .*
(b)* divides , for any .*
(c)* If is a -element of , then there exists a -element such that .*
We need specifically the following fact about Hall subgroups of factorised groups. It is a reformulation of [2, 1.3.2] which is convenient for our purposes. We recall that a group is a Dπ-group, for a set of primes , if every -subgroup is contained in a Hall -subgroup, and any two Hall -subgroups are conjugate. In particular, all -separable groups are Dπ-groups for any set of primes , and all groups are Dπ-groups when consists of a single prime.
Lemma 2.2**.**
Let be the product of the subgroups and . Asume that , and are Dπ-groups for a set of primes . Then there exists a Hall -subgroup of such that , with a Hall -subgroup of and a Hall -subgroup of .
We collect here some results on mutually permutable products, which will be very useful along the paper.
Lemma 2.3**.**
Let the group be the product of the mutually permutable subgroups and . Then we have:
(a)* ([4, 4.1.10]) is the product of the mutually permutable subgroups and .*
(b)* ([4, 4.1.21]) If is a subgroup of , then is a subgroup, and and are mutually permutable. Moreover, if is a normal subgroup of , then is also normal in .*
Theorem 2.4**.**
([5, Theorem 1])*
Let the non-trivial group be the product of the mutually permutable subgroups and . Then is not trivial.*
The following lemma will be essential in the proofs of our theorems.
Lemma 2.5**.**
([3, Lemma 2.4])*
Let be a prime, and be a -group acting faithfully on an elementary abelian -group with , for all . Then is cyclic.*
The next result is the first assertion of Theorem A in [15], which uses the CFSG.
Theorem 2.6**.**
Let be a group. For a fixed prime with , if does not divide for any -regular element of prime power order, then is soluble.
Finally, the later lemma, which is a nice result due to Isaacs, will be very useful in the proof of Theorem D.
Lemma 2.7**.**
([11, 4.17])*
Let be an abelian normal subgroup of a finite group , and let be non-central. Then , where and is arbitrary.*
3 Class sizes not divisible by , for a fixed prime
The well-known Knoche’s theorem (see [12]) asserts that if is a -group, a prime, then the conjugacy class sizes of are square-free if, and only if, . We begin this section by proving Theorem A, which clearly extends it for factorised groups.
Proof of Theorem A. Since for each , it follows . Therefore, commutes with both and , so . Hence is elementary abelian, and for all . Since and (see [8, A - 7.3(a)]) for any , it follows that is elementary abelian. Now it remains to prove that .
Let be a generator of . Since and with and , then . Thus . Clearly, is elementary abelian. Suppose , and let and be two generators of , with and . We distinguish three cases in order to prove that :
i) Suppose first . Let . Since then . Moreover, implies that , where with . On the other hand, so , with . Hence . Finally, as , then with . Therefore , and recall that and both have order . Thus, .
ii) Now suppose . Then . There exists . Therefore, with and , so . Arguing analogously as in case i) with instead of , we conclude that too.
iii) Finally, suppose but . Let . Therefore, we have . If then , and applying case i) to both and we conclude that they generate the same cyclic group of order . On the other hand, if and , then we apply again the first case. Finally, if both then they generate the same cyclic group by case i) again. Thus . Since the last one has order , it follows . So we have , which is equal to by i) again.
In conclusion, if , then it has order . Analogously with . Hence and this establishes the result. ∎
Example 3.1**.**
The converse of the above result is not true in general, in contrast to Knoche’s theorem. Let be the group of the Small groups library of GAP with identification number , which is the product of a cyclic group of order and a quaternion group of order . Then its derived group is , and . Nevertheless, there are elements in the quaternion group with conjugacy class size equal to 4.
Example 3.2**.**
Let be the direct product of a quaternion group and a dihedral group of order . Then every element contained in each factor has conjugacy class size equal to either or , so Theorem A applies. However, there are elements in with conjugacy class size divisible by , and Knoche’s result cannot be applied.
Now we proceed with a key result in the sequel.
Proposition 3.3**.**
Let be a group, and let be a prime. Suppose that is an abelian minimal normal subgroup of , which is a -group. Then:
(1)* If is -nilpotent, and is not divisible by for every element , then .*
(2)* If has order , and is a Sylow -subgroup of , then .*
Proof. (1) Suppose that the result is not true, and let be a counterexample of minimal order. Since the hypotheses are inherited by quotients, we may assume by standard arguments that , and then also . Since is abelian, by a Gaschütz’s result ([10, 4.4]) is complemented, that is, with . We may assume that is not a -group, so by the minimality of . Let be a Sylow -subgroup of (so is a Sylow -subgroup of ). Hence it follows . Let be a minimal normal subgroup of . Since is -nilpotent, we get , where is normal in and . It follows that so . Note that is normal in . Consequently, by the minimality of , we have either or . If , then , which implies that , a contradiction. So we may affirm , for every minimal normal subgroup of .
Now let such that a Sylow -subgroup of , say , is contained in , so . By the hypotheses, is not divisible by , so it follows either or . The first case yields and then , a contradiction. Therefore, we may assume that , and so is normal in . In addition, since is abelian, by the minimality of , we have either or . The last case gives , a contradiction again. Hence, and it follows that and . We only need to see that to finish the proof.
Note that so it follows If , since it is normal in , we can choose a minimal normal subgroup of such that . But then , so , a contradiction. Therefore, we may assume that . On the other hand, we have If is a non-trivial subgroup of , since it is normal in , we have a contradiction with . Consequently we get the final contradiction . The first assertion is then established.
(2) Let , which has order , and let be a Sylow -subgroup of . Then . Moreover, is normal in and so, by the minimality of , we have either or . The first case leads to , and then , a contradiction. Thus we have , and by coprime action it follows . ∎
Note that every dihedral group of order (for an odd prime) verifies the hypotheses of the above proposition (take ).
Theorem B (3) is indeed an immediate consequence of the next more general result.
Theorem 3.4**.**
Let be a soluble group, which is the mutually permutable product of the subgroups and . Assume that is -nilpotent for a prime . If does not divide for any -regular element of prime power order, then has elementary abelian Sylow -subgroups.
Proof. Suppose that the result is false and let be a minimal counterexample. We may assume by the minimality of that , and therefore too. By Theorem 2.4, we can assume that there exists a minimal normal subgroup of such that . Moreover, is -elementary abelian, for some prime . Furthermore, since , by Gaschütz’s lemma we may write , with . Let be a Sylow -subgroup of (so it is a Sylow -subgroup of ). Let . By the minimality of we have , and by Proposition 3.3 (1) it holds . We may choose such that is a Sylow -subgroup of . Since by Proposition 3.3 (2), it holds that , , and . Hence . Finally, since which is elementary abelian by the minimality of , it follows that so is, and this leads to the final contradiction. ∎
Proof of Theorem B. Note that the quotients of satisfy the hypotheses. Moreover, if is a normal subgroup of such that , then also inherits the hypotheses. (Observe that this occurs, for instance, if either or .)
(1) We first see that is soluble by induction over . Since every group of odd order is soluble, we may affirm that because . By Theorem 2.4, we can assume that there exists a normal subgroup of such that . If , then is soluble by minimality. Analogously is also soluble, and then so is . If , we apply Theorem 2.6.
(2) Suppose that the result is false and let be a counterexample of minimal order. Since the quotients of inherit the hypotheses, the class of -nilpotent groups is a saturated formation, and is soluble, we may assume that possesses a unique minimal normal subgroup , with . If is a -group, since is -nilpotent by the minimality of , it follows that is -nilpotent, which is a contradiction. Thus, we may assume that . By Theorem 2.4, we can assume without loss of generality that , and that there exists a minimal normal subgroup of such that either or . In the first case, we have . In the second case, it follows Therefore, we have , where the factors are mutually permutable by Lemma 2.3 (b). In both cases, is normal in and verifies the hypotheses. Hence, if , then is -nilpotent by the minimality of . Since , we get that is -elementary abelian for some prime , so it follows that , with the normal Sylow -subgroup of . Hence, is normal in which implies that , a contradiction.
Therefore, we can assume that . So we have , where is an abelian Sylow -subgroup of . By Lemma 2.2, we may assume that , with either or . Suppose first that , and take . Let , which is normal in . Hence, we have If , then and we can choose , so that is normal in . Thus, in both cases, we have that inherits the hypotheses and, if , it follows that it is -nilpotent. Therefore is a normal Sylow -subgroup of , which is again a contradiction. Consequently, we may assume that , for some -element .
Note that is normal in , since is abelian. By the minimality of , it follows that either or . The second case leads to , a contradiction. Hence, it follows that Then and so , by the hypotheses. Now, we get that is isomorphic to a subgroup of , the cyclic group of order . Hence, divides both and , which contradicts the fact that . This finishes the proof of the -nilpotency of .
(3) It follows from Theorem 3.4. ∎
In the particular case when we recover:
Corollary 3.5**.**
([15, Theorem A])*
Let be a group. For a fixed prime with , if does not divide for any -regular element of prime power order, then is soluble, -nilpotent and has elementary abelian Sylow -subgroups.*
Note that if is the direct product of two symmetric groups of degree 3, then verifies the hypotheses of Theorem B for , but not those of Corollary 3.5. Moreover, the asumption that is necessary, which can be seen by considering , the alternating group of degree 5, and the prime .
We include here a theorem due to Cossey and Wang [7], which was the initial motivation for our results, to notify a gap that we have found in one of the statements.
Theorem 3.6**.**
([7, Theorem 1])*
Let be a finite group, and be a prime divisor of such that if is any prime divisor of , then does not divide . Suppose that no conjugacy class size of is divisible by . Then is a soluble -nilpotent group, and has a Sylow -subgroup of order at most . Further, if is a Sylow -subgroup of G, then has order at most , and if , then is abelian.*
Remark 3.7**.**
(a) The statement “ has a Sylow -subgroup of order at most ” in the above theorem (and so the corresponding one in [14, Theorem 6]) is not true.
To see this, consider the semidirect product (where is a symmetric group of degree 3), which is the group of the Small groups library of GAP with identification number , and the prime . Then verifies the hypotheses of Theorem 3.6 but and . We reveal that this example has been communicated to us by John Cossey.
(b) The same example shows that the hypotheses in Theorem B for the conjugacy class sizes of the elements are not necessarily inherited by the factors, unless they are (sub)normal in . The above group can be factorised as the mutually permutable product of and (we checked this using GAP). It is clear that satisfies the hypotheses of Theorem B for , but there are elements with divisible by .
Remark 3.8**.**
A natural question is how to extend the last assertion of Theorem 3.6 for (mutually permutable) products. Concerning this, we show the following example:
Let be a dihedral group of order 8 and , and consider the prime . Then is a mutually permutable product of and , and is -nilpotent. Moreover, does not divide any conjugacy class size of elements in . However, is not abelian.
Regarding the claim “ has order at most ” in Theorem 3.6, we get the next extension for factorised groups, as an immediate consequence of Theorem A:
Corollary 3.9**.**
Let be the product of the subgroups and . Assume that is -nilpotent, and that for all -elements in the factors, does not divide . If is a Sylow -subgroup of , then , with elementary abelian of order at most .
In particular, from this fact and Theorem B, we get [14, Theorem 7] as a corollary, taking .
Finally, we prove Theorem C, which is motivated by [3, Theorem 1.3].
Proof of Theorem C. Suppose that the result is false and let be a counterexample of minimal order. Note that cannot be simple. Since the class of -supersoluble groups is a saturated formation, we may assume that there exists a unique minimal normal subgroup of , and that . By the minimality of , we get that is -supersoluble. Since is -soluble, it follows that is either a -group or a -group. In the second case, since verifies the thesis by minimality, we get a contradiction. Consequently, we may assume that is -elementary abelian and we must show that . As and is -soluble with , by [8, A - 10.6] it follows that , and also . Applying Theorem 2.4, we may assume that there exists a minimal normal subgroup of such that , so Since is -soluble, it follows that is either a -group or a -group. The first case leads to , a contradiction. Hence, we may assume that is a -group.
Let be a Sylow -subgroup of , where is a prime (so is a Sylow -subgroup of ). Therefore is a Sylow -subgroup of , which acts faithfully on . If , then . By the hypotheses, since does not divide , then neither divides , so either or . The first case leads to , a contradiction. Thus, Lemma 2.5 yields is cyclic. Since this is valid for all primes , we get by [11, 5.15] that is soluble. By the minimality of , it follows that and is abelian with cyclic Sylow subgroups. Consequently , where and the order of is , for some prime .
We may assume that and . Hence . By the minimality of , we have either or . The first case leads to , a contradiction. Therefore, since , it follows , and this final contradiction establishes the theorem. ∎
Example 3.10**.**
Let be the symmetric group of degree . Then is a mutually permutable product, where denotes the alternating group of degree 4 and is a Sylow 2-subgroup of , which verifies the hypotheses of Theorem C, for .
4 Square-free class sizes
We begin this section with the proof of Theorem D.
Proof of Theorem D. (1) Suppose that the result is false and let be a counterexample of least possible order. Since supersoluble, is abelian, and so . Moreover, the quotients of inherit the hypotheses and the class of metabelian groups is a formation, so we may assume that there exists a unique minimal normal subgroup of with , for some prime divisor of . Hence, , a Sylow -subgroup of . Since is abelian, . Hence , where and are Sylow -subgroups of and respectively, by Lemma 2.2. Applying Theorem A, we have , and is elementary abelian of order at most . Note that , because .
By Lemma 2.2, we may consider a Hall -subgroup of , such that , where and are Hall -subgroups of and respectively. Moreover, is abelian. Let be a prime power order element. Since , it follows by the hypotheses that , and so . Thus , and analogously for . Consequently we get . Since , it follows . In particular, .
If , then and , a contradiction. Hence we have either or . Assume . Let be a Sylow -subgroup of , for some prime . Note that , because is a -group. Let , which acts on , which is elementary abelian because . Suppose . Then for all , so . Let denote the order of . Thus . It follows for all , so . Then and the action is faithful. Let . Therefore , with . Moreover, which divides , because is normal in . Since and is a non-central prime power order element in , it follows , and so . Applying Lemma 2.5, we get that is cyclic. Since this is valid for each prime divisor of , we deduce that has cyclic Sylow subgroups, but it is abelian, so is cyclic. Analogously, if , then it is cyclic.
Let . Assume first that is a -element, for some prime , and that . By the above argument, acts faithfully on , and . Let . If , then . Hence , so and , a contradiction. Then , and since , it follows . Therefore we have so , and then which is abelian, a contradiction.
Hence, we may assume that, for every prime , if is the -part of , then . Note that is normal in , and . Therefore, by the minimality of , it follows that is abelian. Notice that , since is normal in . Let , which is an abelian normal subgroup of , and let . Then Lemma 2.7 leads to . If divides we get a contradiction, because . Hence, and , for each . By coprime action, . Thus . If is a generator of , then with , so . Since this is valid for each prime divisor of the order of , we get:
[TABLE]
Analogously, if , then . Since , we get . This final contradiction establishes statement (1).
(2) Suppose that the second assertion is not true and let be a counterexample of minimal order. We point out that the hypotheses are inherited by every quotient group of and, by (1), is abelian. There exists a prime divisor of such that does not have any elementary abelian Sylow -subgroup. By the minimality of , we may consider that . Moreover, since is supersoluble, then is abelian, and is a normal Sylow -subgroup of such that . Using Lemma 2.2 and Theorem A, we obtain respectively that , and that is elementary abelian with .
Let be the nilpotent residual of . Note that ; in other case, is a -group and then , a contradiction. Since , it follows that is abelian. By using [8, III - 4.6, IV - 5.18], we have that is complemented in , and its complements are precisely the Carter subgroups of . Accordingly, with a nilpotent subgroup of and . These facts yield , and . On the other hand, the minimality of implies that is elementary abelian, and thus is not so. If , since is normal in , by the minimality of we have is an elementary abelian group, but
[TABLE]
which is a contradiction. Hence, . In particular, we deduce that .
By Lemma 2.2, there exists a Hall -subgroup of such that . Let be a non-trivial element of prime power order. Since , it follows that , and so . Thus , and analogously for . Consequently, we get . Since and , it follows , which implies that is elementary abelian, the final contradiction.
(3) Assume that the result is false and take a counterexample of minimal order. Consider a prime such that . By minimality, we can affirm that . Since is supersoluble, we get that is a normal Sylow -subgroup of . Then, we apply both Lemma 2.2 and Theorem A to get the final contradiction. ∎
Example 4.1**.**
Let be the direct product of two symmetric groups of degree 3. Then is supersoluble, and every element contained in each factor (not only those of prime power order) has square-free conjugacy class size, but neither the derived subgroup nor are cyclic, in contrast to [7, Theorem 2].
Example 4.2**.**
In view of [7, Theorem 2], it is natural to wonder if we can affirm in the above result that the Sylow -subgroups of have order at most . This fact is not further true, as we show:
Let , where is a dihedral group of order 14, and is the direct product of such a dihedral group and a semidirect product of a cyclic group of order 7 and a cyclic group of order 3 ( has identification number in the Small groups library of GAP). Then is supersoluble, and satisfies that all prime power order elements contained in each factor have square-free conjugacy class size, but has order .
Now we proceed with the proof of Theorem E.
Proof of Theorem E. Considering the smallest prime divisor of and Theorem B, we conclude that is soluble. Hence, it is -soluble for each prime divisor of . Applying Theorem C, we get that is -supersoluble for each prime that divides , so it is supersoluble.
Now we prove the second assertion by induction on . Let be an arbitrary prime, and be a Sylow -subgroup of . We want to show that is elementary abelian. Since is supersoluble, we have that is abelian. Moreover, we may assume by induction that . Therefore, we have that where is a Hall -subgroup of . Consequently, is normal in and is -nilpotent. Finally, by Theorem 3.4 the result is established. ∎
When considering in the above theorem all -regular elements in the factors, we get as a corollary:
Corollary 4.3**.**
([3, Corollary 1.5])*
Let the group be the mutually permutable product of the subgroups and . Suppose that for every prime and every -regular element , is not divisible by . Then is supersoluble.*
Example 4.4**.**
Consider , where is a symmetric group of degree 3, and is the direct product of such a symmetric group and a dihedral group of order 10. Then satisfies the hypotheses of Theorem E. However there exists some -regular element in , not of prime power order, such that divides its conjugacy class size, so Corollary 4.3 cannot be applied.
In the particular case when and are normal in , we obtain [14, Proposition 9].
Corollary 4.5**.**
Let and be normal subgroups of such that . Suppose that is square-free for every element of prime power order of . Then is supersoluble.
This development has its origins in the contributions of Chillag and Herzog [6, Theorem 1], and Cossey and Wang [7, Theorem 2]. Our next result Theorem F and Theorem D can be considered somehow extensions of the ones above for (mutually permutable) products. In fact, Theorem F provides further information on the Sylow subgroups of .
Proof of Theorem F. Suppose the result is not true and let be a counterexample of least order possible. Then if is a Sylow -subgroup of , we have . We can assume by the minimality of that , so . By Lemma 2.2 we can choose , with and Sylow -subgroups of and respectively. By Theorem E, we have that is supersoluble and has elementary abelian Sylow subgroups. In particular, is normal in . Hence Lemma 2.3 (b) asserts that is normal in , and it is a mutually permutable product. Moreover, If we suppose , by the minimality of it follows , a contradiction. Thus, we may assume , and so is -nilpotent.
Let be a minimal normal subgroup of . We can assume without loss of generality that it is contained in by Theorem 2.4. Note . By Proposition 3.3 (1), it follows , and by the minimality of we have Since , we may assume . As , [[8], A - 10.6 Theorem] leads to . If is the unique minimal normal subgroup of , then which is isomorphic to a subgroup of so is cyclic and elementary abelian, which implies that its order is , a contradiction.
Now we denote by the product of all minimal normal subgroups of distinct of , so and . It follows . We denote and . Since (and ), by the minimality of we may affirm . On the other hand, since , we have (analogously ). In addition, it follows so . Let be a Sylow -subgroup of such that . Hence . In addition, since where is a -element contained in , by the hypotheses it follows that does not divide . Moreover, since , we may assume , and since , we have . Accordingly , and since is abelian, we have necessarily . This leads to
[TABLE]
Suppose , and let . Let . Then since and is abelian, we have that is a -regular element, so by the hypotheses does not divide . As is a direct product, we have . In addition, , and therefore divides which is a -number, so (recall that ), and analogously , a contradiction. We conclude .
If with normal in and a minimal normal subgroup of contained in , by similar arguments we can deduce . This means, in particular, that neither nor can contain two distinct minimal normal subgroups of .
On the other hand, since is the unique -Hall subgroup of , Lemma 2.2 leads to Moreover, since , it follows . Note that with so we distinguish two cases: either or . In the first case . Thus there exists another minimal normal subgroup contained in and distinct of , a contradiction. Hence we conclude so .
Now suppose that is a -group. Then, since , it follows that is -elementary abelian. In addition, acts faithfully over . Let , then , with since and is abelian. By the hypotheses, does not divide , and therefore it does not divide . Thus we may affirm . By Lemma 2.5 we conclude that is cyclic, and analogously is cyclic too. So they are both cyclic and elementary abelian, that is, they both have order . Thus a contradiction.
Hence we may suppose that there exists a prime such that divides . Let be a Sylow -subgroup of (so it is a Sylow -subgroup of ). Then , and since , necessarily we have that is the product of the minimal normal subgroups of with order . Let be one of those minimal normal subgroups. Arguing exactly in the same way as with , it follows . But so with both minimal normal subgroups of , and . Let be a Sylow -subgroup of such that . Then by the hypotheses we have . Since , it follows . However, is normal in so , and by Proposition 3.3 (1) we have . This final contradiction establishes the theorem. ∎
Example 4.6**.**
Under the hypotheses of Theorem E (even under those of Theorem D), it is not possible to assure that has Sylow -subgroups of order at most , as the following example shows:
Let be a finite set of pairwise distinct odd primes, and let be the direct product of dihedral groups of order , . Then is a mutually permutable product of and , and each prime power order element contained in the direct factors has square-free conjugacy class size. However, has order .
The reference list from the paper itself. Each links out to its DOI / PubMed record.
- 1[1]
- 2[2] B. Amberg, S. Franciosi and F. de Giovanni: Products of groups. Oxford University Press, New York, 1992.
- 3[3] A. Ballester-Bolinches, J. Cossey and Y. Li: Mutually permutable products and conjugacy classes. Monatsh. Math. 170 (2012) 305–310.
- 4[4] A. Ballester-Bolinches, R. Esteban-Romero and M. Asaad: Products of finite groups. Vol. 53 of de Gruyter Expositions in Mathematics, Berlin, 2010.
- 5[5] J. C. Beidleman and H. Heineken: Group classes and mutually permutable products. J. Algebra 297 (2006) 409–416.
- 6[6] D. Chillag and M. Herzog: On the length of the conjugacy classes of finite groups. J. Algebra 131 (1990) 110–125.
- 7[7] J. Cossey and Y. Wang: Remarks on the length of conjugacy classes of finite groups. Commun. Algebra 27 (1999) 4347–4353.
- 8[8] K. Doerk and T. Hawkes: Finite Soluble Groups. Vol. 4 of de Gruyter Expositions in Mathematics, Berlin, 1992.
