Endomorphisms of free solvable groups preserving primitivity of systems of elements. (Q904320)

Let \(G_r\) be the free group of rank \(r\) in a given variety of groups \(\mathcal M\). An element \(g\in G_r\) (resp., a system of elements \(\{ g_1,\ldots,g_m\}\)) is said to be \textit{primitive} if it can be complemented to a basis for \(G_r\). And an endomorphism \(\phi\in\mathrm{End}(G_r)\) \textit{preserves primitivity} (resp., \textit{preserves primitivity of systems of elements}) if the image of any primitive element (resp., any primitive system) is primitive (resp., a primitive system). Typically, automorphisms of \(G_r\) of course preserve primitivity of systems of elements, and there is a general open question whether the converse is true or not: is any endomorphism preserving primitivity (or primitivity of systems of elements) necessarily an automorphism of \(G_r\)? There are several partial answers to these questions in the literature. First, both answers are positive for the case of the variety of Abelian groups. For absolutely free groups \(F_r\), D. Lee proved that every endomorphism \(\phi\) satisfying \(\phi(\mathcal O)\subseteq\mathcal O\) must be an automorphism, where \(\mathcal O\) is any given automorphic orbit (like, for example, the set of primitive elements). The same result was shown to be true in rank two metabelian groups by Roman'kov and Timoshenko. Also, Gupta-Timoshenko proved that every \textit{inner endomorphism} (see the paper for a precise definition) of a free metabelian group preserving primitivity must be an automorphism. In the paper under review, the author proves in general that any endomorphism of a free metabelian group of rank \(r\) preserving primitivity of systems of elements must be an automorphism; and that any endomorphism of a free solvable group of rank two preserving primitivity must also be an automorphism.