Nil radicals in groups (Q1191170)

According to a well-known theorem of Plotkin, any group \(G\) contains a unique maximal normal, locally nilpotent subgroup -- the Plotkin locally nilpotent radical \(P(G)\). In the Kourovka Notebook Sh. S. Kemkhadze posed the question on the existence in an arbitrary group of a binary-nilpotent radical: ``Is a binary-nilpotent group a product of two normal binary- nilpotent subgroups?'' In the present paper it is shown, by means of the groups of Golod, that the condition of local nilpotency of normal subgroups in Plotkin's theorem cannot be weakened to nilpotency of \(n\)- generator subgroups, and it is obtained a negative answer to Kemkhadze's Question 1.39.