On groups that correspond to the simplest problems of classical mechanics

The following questions are discussed: 1) what is the maximum possible complexity of a finite-dimensional group $\mathscr{G}$ of “latent” symmetry? 2) does the existence of a complete set of single-valued integrals of motion always imply the existence of a nontrivial group $\mathscr{G}$? The impossibility of essential extension of the groups $\mathscr{G}$ for known examples is proved; a negative answer is given to the second question.