Testing the Equivalence Principle : Why and How ?

Part of the theoretical motivation for improving the present level of testing of the equivalence principle is reviewed. The general rationale for optimizing the choice of pairs of materials to be tested is presented. One introduces a simpliied rationale based on a trichotomy of competing classes of theoretical models. 1. Why testing the equivalence principle ? Einstein introduced in 1907 1] what he called the \hypothesis of complete physical equivalence" between a gravitational eld and an accelerated system of reference. He used this \equivalence hypothesis" 1], 2] as a heuristic tool to construct a physically satisfactory relativistic theory of gravitation. The resulting theory, general relativity, has been very successful, both in renewing completely our description of the universe, and in passing with ying colours all the experimental tests it has been submitted to. For instance, the universality of free fall (the experimental basis for the equivalence principle) has been veriied at the 10 ?12 level 3], 4], the quasi-static weak-eld (\post-Newtonian") predictions have been checked in solar-system experiments at the 10 ?3 level 5], the radiative structure (propagation of gravity with the speed of light as an helicity-2 interaction) has been veriied by binary-pulsar data at the 10 ?3 level 6], and the quasi-static strong-eld predictions have been checked in binary-pulsar experiments at the 10 ?2 level 7]. In view of this impressive record, should one apply Ockham's razor and decide that Einstein's theory must be 100% right, and then stop testing it any further? My answer is deenitely, no! Indeed, one should continue testing a basic physical theory such as general relativity to the utmost precision available simply because it is one of the essential pillars of the framework of physics.