Five rolling contact fatigue tests, Tests {1}–{5} have been conducted. In Tests {1}–{3}, when a fatigue crack was initiated on the surface of a follower, the test was halted. Then, in Test {1} the rotating direction was reversed. In Test {2} the follower and driver were interchanged, and in Test {3} the test was continued unchanged. In Test {3} the original crack grew to a pit. In Tests {1} and {2} the original crack immediately stopped propagating. In Tests {4} and {5}, mating with a harder roller, a softer roller was used as the follower in Test {4} and as the driver in Test {5}. A typical pit occurred in Test {4}. In Test {5}, surface damage substantially different from a typical pit was generated. Based on these experimental results, a 3-D crack analysis including the effect of frictional force on the contact surface and oil hydraulic pressure on crack surfaces, was conducted to elucidate the mechanisms of pit formation and surface damage in contact fatigue.