Largely quantitative experiments by Pappas have indicated that the momentum imparted to an electrodynamic impulse pendulum was not balanced by an equal and opposite momentum change of field energy as required by the special theory of relativity. The authors repeated Pappas' experiment using discharge currents from a capacitor bank which contained a known amount of stored energy. It turned out that, for momentum conservation, the magnetic-field energy required would have been 1000 to 2000 times as large as the energy that was actually stored in the capacitors. In the second part of the paper the pendulum experiments are interpreted in terms of Ampere's force law. It is shown that the Ampere force exerted on the pendulum is almost exactly the same as the lorentz force, but it arises in different parts of the pendulum conductor. Furthermore, the Ampere reaction force does not reside in the field but in the stationary part of the circuit which supplies current to the pendulum. Hence in the Ampere electrodynamics the momentum is definitely conserved. The experimental and analytical findings confirm the work by Pappas. A new and important experimental fact emerged from the present investigation. The momentum imparted to the pendulum was found to be significantly smaller than the calculated mechanical impulse given by the Lorentz and Ampere force laws. The Ampere force distribution offers an explanation of this observation in terms of the elastic distortion of the pendulum structure. The Lorentz force distribution could not produce this distortion.