This paper describes some general criteria for wear testing. Laboratory wear tests can be designed to meet the criteria but test results generally do not accurately predict wear under field conditions. Field wear testing is needed to optimize the final selection of materials for the given application. The concept of wear gradient is introduced. Wear gradients, with respect to time and location, are small and reproducible in laboratory tests but generally are significant and fluctuate under field conditions. Field tests are classified as sequential series, concurrent series, and concurrent group tests. The use of reference specimens for establishing time-averaged wear gradients is discussed and formulae for correcting the test specimen wear for non-uniform tribo- conditions over the test area are presented. A field test for abrasive wear of a linear array of specimens is described. The corrected wear values of duplicate specimens agree within 5% of the mean.