Knowledge of rupture properties is required to characterize rubbers in laboratory evaluation adequately and to attempt to estimate service performance. The most commonly measured rupture property is tensile strength. Tear strength, although known to be important, is measured less often due to a general consensus that existing tests are inadequate. This paper describes a new, practical, tear test method based on theoretical work by workers at the Natural Rubber Producers' Research Association (NRPRA). Their work has shown that a characteristic tear strength may be measured independent of testpiece geometry. A testpiece has been devised that circumvents two deficiencies inherent in the “trousers” testpiece advocated by NRPRA. These are: appreciable leg extension, and failure of the tear to propagate down the central axis when tear tests are carried out on a wide variety of vulcanizates. These deficiencies are overcome by reinforcing the legs with fabric and by molding a longitudinal groove of a definite cross-sectional geometry into the testpiece. These alterations of the basic trousers test specimen allow almost any type of rubber to be tested successfully. Detailed descriptions of the testpiece, mold, method of test, and tear curve analysis are given, as well as results of a statistically planned evaluation program. Effects of crosslink density, carbon black content, processing oil content, and test temperature are determined for natural rubber and SBR. The influence of speed of testing in a rather restricted range is also included. When modified trousers tear strengths are compared with tear strengths from a common test, ASTM D-624, Die C, pronounced differences are apparent. These are pointed out and deficiencies in the Die C test discussed.