Recently A. Chenciner and R. Montgomery found a remarkable periodic orbit for a
three-body problem by variational methods.
On this orbit all masses chase each other along a figure-eight
circuit without any collision, and the solution curve is indeed
a minimizer of the action functional on a properly chosen path space.
One technical difficulty, where numerical integration had been used
in their proof, is to show that the minimizing orbit does not
experience any collision. In this paper a short analytical proof will
be presented.