EINSTEIN'S theory of gravitation is based upon a fundamental postulate which he calls thq principle of equivalence, and which asserts that gravitation and inertia are identical in nature, and hence indistinguishable. This, if true, is of the greatest theoretical importance, for gravitation has steadily refused to show any kinship to other physical phenomena.
Einstein's justification for this postulate was found in the gravity pendulum experiments of Newton and Bessel, and the torsion pendulum experiments of E6tv6s, the results of which established to a high degree of precision (about I part in 2oo,ooo,ooo) that the inert mass and the gravitational mass of a body were proportional, or, in other words, that gravitation is independent of the nature of the matter acted upon.
A still more delicate test of this postulate is possible in a crystal of one of the non-isometric systems; for in such a crystal every known physical property (except inertia, and possibly weight) varies with the axial direction in the crystal. It is therefore an interesting question whether, in such a crystal, gravitation will be found to align itself with inertia or will show some variability which will classify it with the great majority of physical phenomena. If, for example, such a crystal be weighed in different axial orientations with respect to the earth (which may be done with great precision) and any difference in weight be found in the different positions, Einstein would be wrong.
To test this point, crystals weighing a kilogram or more were thus weighed, the specimens coveringall five non-isometric crystalline systems. The precision reached, in nearly every case, was one part in a billion (lO9). To this degree of precision no difference in weight was detected; the results have failed to prove Einstein wrong.