A classification of some unsoluble varieties of groups of exponent four (Q1375350)

This substantial paper on varieties of groups of exponent 4 continues work begun by the author in another paper [``Varieties of groups of exponent four'', J. Lond. Math. Soc. (to appear)]. The results are worthy of full description, even though some preliminary definitions are needed. Let \({\mathfrak B}_4\) and \({\mathfrak E}_4\) be the varieties defined by the law \(x^4=1\) and the 4-Engel law \([x,y,y,y,y]=1\), respectively. Define \[ \ell_3(x,y,z)=[y,x,x,x,z][y,x,x,y,z,x][y,x,x,z,x,z,z][ ^2,x,y^2,x,x], \] and, for \(n\geq 3\), let \(\widetilde{\mathfrak L}_n\) be the variety defined by the laws \(x^4=1\) and \([\ell_3(x_1,x_2,x_3),x_4,\dots,x_n]=1\), and let \({\mathfrak L}_n=\widetilde{\mathfrak L}_n\cap{\mathfrak E}_4\). Define \[ k_4(x_1,x_2,x_3,x_4)=\ell_3(x_1x_2,x_3,x_4)\ell_3(x_1,x_3,x_4)\ell_3(x_2,x_3,x_4), \] and, for \(n\geq 4\), let \(k_n(x_1,\dots,x_n)=[k_4(x_1,x_2,x_3,x_4),x_5,\dots,x_n]\) and \[ d_n(x_1,\dots,x_n)=k_n(x_1,x_2,x_3,\dots,x_n)k_n(x_1,x_3,x_2,x_4,\dots,x_n)\cdots k_n(x_1,x_n,x_2,x_3,\dots,x_{n-1}). \] For each even integer \(n\), where \(n\geq 4\), let \(\widetilde{\mathfrak D}_n\) be the variety defined by the laws \(x^4=1\) and \(d_n=1\), and let \({\mathfrak D}_n=\widetilde{\mathfrak D}_n\cap{\mathfrak E}_4\). The author's main result is that every proper subvariety of \({\mathfrak B}_4\cap{\mathfrak E}_4\) which contains \({\mathfrak L}_3\) is either equal to \({\mathfrak L}_n\) for some \(n\) or is equal to \({\mathfrak D}_n\) for some even \(n\). From this he obtains a classification of all subvarieties of \({\mathfrak B}_4\) which contain \({\mathfrak L}_3\): the varieties in question are \({\mathfrak B}_4\), \(\widetilde{\mathfrak L}_n\) (\(n\geq 4\)), \(\widetilde{\mathfrak D}_n\) (\(n\) even, \(n\geq 6\)), \({\mathfrak B}_4\cap{\mathfrak E}_4\), \({\mathfrak L}_n\) (\(n\geq 3\)), \({\mathfrak D}_n\) (\(n\) even, \(n\geq 4\)). Finally, the author obtains a reduction theorem for the finite basis problem for subvarieties of \({\mathfrak B}_4\): if \({\mathfrak B}_4\) has a subvariety which is not finitely based then so does \({\mathfrak L}_3\).