A generalized variety problem (Q2663116)

A \textit{generalized variety} \({\mathcal V} \) of algebras is a class closed under morphic images, subalgebras, powers and finite direct products. The notations \({\mathcal L}({\mathcal V})\) and \(\mathcal {G}(\mathcal {V})\) refer respectively to the lattices of varieties and of generalized varieties contained in \(\mathcal{V}\). The finite members of generalized varieties form precisely the pseudovarieties in the given framework. The key examples here are the generalized varieties \({\mathcal Nil}\) and \({\mathcal N}\) consisting respectively of the nil and nilpotent semigroups. The author considers the following question raised by \textit{J. Almeida} [Finite semigroups and universal algebra. Singapore: World Scientific Press (1995)]: is the map \({\mathcal L}({\mathcal Nil}) \cup \{{\mathcal Nil} \} \longrightarrow {\mathcal G} ({\mathcal N})\), given by intersection with \({\mathcal N}\), an isomorphism? The author answered this question in the negative [\textit{G. Churchill}, J. Algebra 196, No. 2, 499--519 (1997; Zbl 0888.20035)], even when restricted to nil semigroups of index 5. The purpose of the present paper is to show that the negative result holds even for nil semigroups of index 2 (the least nontrivial case). The proof makes use of the square-free words introduced by \textit{J. Leech} [Math. Gaz. 41, 277--278 (1957; Zbl 0079.01101)].