Schools use an array of strategies to match curricula and instruction to students' heterogeneous skills. Although generations of scholars have debated ''tracking'' and its consequences, the literature fails to account for diversity of school-level sorting practices. In this article, we draw on the work of Sørensen and others to articulate and develop empirical measures of five distinct dimensions of within-school cross-classroom tracking systems: (1) the degree of curricular differentiation, (2) the extent to which sorting practices generate skills-homogeneous classrooms, (3) the rate at which students enroll in advanced courses, (4) the extent to which students move between tracks over time, and (5) the relationship between track assignments across subject areas. Analyses of longitudinal administrative data following approximately 20,000 eighth graders enrolled in 23 middle schools through the 10th grade indicate that these dimensions of tracking are empirically separable and have divergent relationships with student achievement and the production of inequality.