On the multilinear identities which hold in the Lie ring of a group of prime-power exponent (Q1082436)

It is well known that the associated Lie ring of a group of prime exponent \(p\) satisfies the \((p-1)\)th Engel identity. \textit{G. E. Wall} was the first to exhibit [in Proc. 2nd Int. Conf. Theory of Groups, Canberra 1973, Lect. Notes Math. 372, 667-690 (1974; Zbl 0286.20050)] the new polylinear identity of degree \(2p-1\) which also holds in the Lie ring of any group of prime exponent \(p\). However only with the aid of a computer and only for \(p=5\), 7, 11 this identity was shown to be really new -- that is not a consequence of the \((p-1)\)th Engel one and \(px=0\). In a recent paper \textit{M. R. Vaughan-Lee} [Bull. Lond. Math. Soc. 17, 113-133 (1985; Zbl 0544.20032)] found all polylinear identities which hold in \(L(\bar B(n,p))\). He conjectured that only for degrees \(k(p-1)+1\), \(k=1,2,...\), they are really new -- that is not a consequence of those of smaller degree. The free groups \(\bar B(n,p)\) of the Kostrikin variety of locally finite groups of prime exponent \(p\) and their Lie rings \(L(\bar B(n,p))\) are of exceptional interest as classical objects. Moreover, the new identites in \(L(\bar B(n,p))\) turned out to be connected with other problems about finite \(p\)-groups. Counterexamples to the Hughes conjecture discovered firstly by \textit{G. E. Wall} [Proc. Int. Conf. Theory of Groups, Canberra 1965, 357-362 (1967; Zbl 0189.31701)] exist when Wall's identity of degree \(2p-1\) is really new in \(L(\bar B(3,p))\). The reviewer [in Mat. Sb., Nov. Ser. 118(160), 567-575 (1982; Zbl 0484.20017)] constructed a 2-generated counterexample to the Hughes conjecture in connection with the new identity in \(L(\bar B(2,p))\) (with the aid of acomputer). In a recent article [Algebra Logika 25, No.2, 227-240 (1986)] the reviewer used the results from the preprint of the article under review to refute the conjecture of \textit{N. Blackburn} and \textit{A. Espuelas} [from Proc. Am. Math. Soc. 92, 478-484 (1984; Zbl 0556.20016)] and to refute the conjecture about the index of the generalized Hughes subgroup \(H_{p^ 2}(G)\). The present article continues the previous works of the author who developed a rich new technique for dealing with the Lie ring of \(\bar B(n,p)\). Here he proves the conjecture of \textit{M. R. Vaughan-Lee} (op. cit.). In fact theorems 1, \(1^*\) and 3 cover some more general situation including the case of groups of exponent \(p^ m\). Theorem 2: if the polylinear identities of degree \(k(p-1)+1\) which hold in the Lie ring of \(B(n,p^ m)\) for all \(k=1,2,...,r\) are not consequences of those of smaller degree then there exists a finite \(p\)-group \(P\) of class \(r(p- 1)+1\) in which \(| P| \neq | P: H_{p^ m}(P)| \geq p^ r\). Independently the reviewer also proved theorems 1 and 2 in the particular case when \(r=3\) and \(m=1\) [in Izv. Akad. Nauk SSSR, Ser. Mat. 50, No.6, 1308-1325 (1986)] and earlier also theorem 2 for \(r=2\), \(m=1\) [in Mat. Sb., Nov. Ser. 116(158), 253-264 (1981; Zbl 0473.20016)].