Given a continuously differentiable vector field $\bf a$, demonstrate the equivalence (iff) between the requirement that it satisfies ${\bf a}\cdot(\nabla \times {\bf a})=0$ and that it has the representation ${\bf a}=\lambda \nabla \phi$, where $\lambda, \phi$ are scalar functions. Stated in another way, why is ${\bf a}\cdot(\nabla \times {\bf a})=0$ a necessary and sufficent condition for vector fields to have normal congruences.
Many reference cites Lord Kelvin as the source of this theorem, but his proof, though classical, is too magnetism-oriented to readily understand. Thanks!