Einstein's Two-and-a-Half Tests
Different as Einstein's and Newton's theories are, within the solar system their results are almost identical. Only on a cosmic scale, or near extremely dense objects such as black holes, does general relativity bring large changes. Einstein in 1916 could only think of three potential manifestations of general relativity, all minuscule.
Perihelion precession:Mercury's orbit around the Sun should gradually turn in its plane through an angle minutely different from Newtonian prediction -- an effect called perihelion precession.
Starlight deflection:Stars observed near the edge of the Sun should appear slightly displaced outward from their normal positions.
Gravitational redshift:Light leaving a star should change color slightly, shifting toward the red.
For over forty years, these three effects -- weak both in what they tested and in how well they tested it -- were all there was. Starlight deflection proved frustratingly difficult to measure. Mercury's orbit, though better, was subject to Newtonian disturbances. Least satisfactory was the redshift, which was observationally messy and hinged on the assumption (the "Einstein equivalence principle") far short of general relativity. This was at most a half-test.
Worse, competing theories soon appeared giving the same predictions for Einstein's tests of general relativity.
http://einstein.stanford.edu/content/st ... html#tests