The use of linear regression to identify variables that predict grades of black students in precalculus and calculus I was studied at Xavier University of Louisiana, a historically-black institution. The predictive power of the variables for black males and females was also assessed. These two mathematics courses are prerequisites for students wishing to study health fields. Entering freshman majoring in biology, chemistry, or pre-pharmacy in the fall semesters of 1981 and 1982 were studied. Of this population, 135 females and 81 males studying precalculus and 100 females and 64 males studying calculus I were evaluated. The following variables showed some potential for predicting grades in both precalculus and calculus at Xavier: high school grade point average (GPA), composite score on the American College Testing (ACT) program, vocabulary scores on the Nelson-Denny Reading Exam, and English and mathematics subsections of ACT. The ACT composite score was the single best predictor for precalculus, high school GPA was the best for all students taking calculus, and grade in precalculus was the best predictor of calculus. The variables predicted the grades of both sexes equally well, in contrast to other research that indicates that traditional criteria have more predictive power for black females than for black males. (SW)