It was hypothesized that race and gender would interact, such that the likelihood of a first arrest for a more serious and less serious offense would be highest for non-White males and non-White females, respectively. The Midwest Homeless and Runaway Adolescent Project, from which data were obtained for this study, was a cross-sectional study of 602 homeless and runaway youths in Iowa, Kansas, Missouri, and Nebraska. The youth were interviewed on the streets, in shelters, and in drop-in centers by outreach workers. Sixty percent of the 602 youths were female, and 60.1 percent identified themselves as White, 24.1 percent as African-American, 3.3 percent as Hispanic, 1.5 percent as Native American, and 10 percent as other. The youths averaged 16.27 years old. Youths were asked about the dates and types of offenses committed for their first arrest and for their four most recent arrests. Independent variables were under the general categories of deviant subsistence strategies, substance use, gang participation, prior arrests, age, prior physical abuse, age when leaving home, time on the street, and race. The findings show that non-Whites were more likely than Whites and males were more likely than females to be arrested for a more serious offense. White females were more likely than non-White females to be arrested for a less serious offense. 1 figure, 2 tables, 7 notes, and 44 references