This paper is a review of opinion and research concerning the objectivity and relevance of grades and grade averages as measures and as predictors of success. As measures they are ambiguous, reflecting differences in sex, basic temperament, instructors, departments, institutions, as much as levels of competence. And as a predictor of "success," grade point average has not been particularly valid for either graduate school or occupation. Criticisms of current practices--classified as ethical, rational, and pragmatic--and possible alternatives are discussed. Four suggested innovations are (1) pass-fail grading, (2) credit by examination, (3) criterion-referenced teaching and evaluation, and (4) procedures for evaluating creative extracurricular achievements. Research on the revision of evaluation procedures at the University of Delaware over a ten-year period is reviewed; and a study of the senior class of 1969 is reported in a comprehensive set of tables. The results suggest that the university has been discriminating against students who are (a) male, (b) enrolled in the sciences and traditional academic disciplines, and (c) academic nonconformists. (CK)