Background/Context: The United States is becoming increasingly racially and ethnically diverse, and increasingly racially isolated across race-ethnic boundaries. Researchers have argued that both diversity and racial isolation serve to undermine the social cohesion needed to bind American citizens to one another and to society at large. Focus of Study: Given the compelling and consistent findings relating desegregation to social inclusion, this research posits that the issue of declining social trust and social cohesion may be better understood as a consequence of segregation and social isolation within communities rather than as a consequence of variations in diversity across communities. Thus, this study examines the relationship between social cohesion (social distance) and social isolation (race-ethnic segregation) at the institutional level--in schools and neighborhoods. Thus, in the present study, social distance, which reflects both weak connections among ethnically diverse groups in society and limited "bridging capital," serves as our operational indicator of social cohesion. Participants: Participants in this study come from the National Longitudinal Survey of Freshmen, a national probability sample of approximately 4,000 first-time students entering selective colleges and universities in 1999. Equal numbers of African American, Latino, Asian, and White students were sampled from 28 participating institutions, which resulted in an oversampling of minority students to provide meaningful comparisons across each of the major race-ethnic groups. Research Design: This study examines the effects of early racial isolation in schools and neighborhoods on social cohesion (i.e., preference for same-race neighbors, preference for children to have same-race schoolmates, and social distance); as such, the measures of social cohesion are drawn from the baseline survey (Wave 1) conducted at the beginning of the first year, before college context and experiences could reasonably impact these outcomes. The models in this study are estimated by race-ethnic group using ordinary least squares regression. The social cohesion outcomes (i.e., preference for same-race neighbors, preference for children to have same race-schoolmates, and social distance) are estimated separately for each race-ethnic group as a function of early racial isolation in neighborhoods, early racial isolation in schools, high school type and context, and student demographics. Findings/Results: Results suggest that social isolation in schools plays a more significant role than neighborhood isolation in diminishing social cohesion among young adults, although both matter. Our overall findings relating social isolation in K-12 schooling and young adults' feelings of social distance, as well as preference for same race-neighbors, offer further support for perpetuation theory, which suggests that early school segregation leads to segregation across the life course and across institutional contexts. The findings also point to school segregation's intergenerational consequences and are consistent with the results of Crain's classic research using Office of Civil Rights data, which laid the foundation for later studies on the long-term effects of desegregation.