Some remarks on the symmetry algebras of soliton equations