The paper addresses the issues of testing modifications with handicapped students in Houston (Texas) Independent School District. Using descriptive data, the performance of secondary mainstreamed and nonhandicapped students on course proficiency field tests without modifications are compared. Approximately 65% of the mainstreamed students were learning disabled or speech handicapped. Comparisons of scores on criterion referenced tests (without adaptations) for 14 academic courses revealed that mainstreamed Ss performed as well or better than their nonhandicapped peers on five tests (Algebra I, geometry, world history, world geography, and U.S. history). Data further suggested that testing modifications are necessary for mainstreamed students on course proficiency tests in English and science, although testing without modifications seemed appropriate for the areas of mathematics and social studies. It is concluded that testing modifications are not the answer to the high failure rate for mainstreamed students in maximum competency programs, but that students with a low probability for success should not be mainstreamed. (CL)