On a Problem in the Calculus of Variations