A nearly ideal two-dimensional scotch yoke mechanism was constructed to test a model of wear depth as a function cycle number.  Model variables include the reciprocating mass, a two dimensional wear-rate, crank radius, and angular velocity.  The model originally developed by T.A. Blanchet (1997), was non-dimensionalized and simplified under conditions of large numbers of cycles, to predict the importance of including coupling based solely on a ratio of maximum allowable wear depth to the crank radius.  Experiments show a linear progression of wear over two distinct regions, suggesting a sudden transition in wear modes just after 1.5 million cycles.  The need for cycle or time dependent wear rates in analysis, which is a potentially far more significant source of error, is clearly illustrated by the experiment and discussed.