The analysis included counties in four states with substantial nonmetropolitan populations: Florida, Georgia, South Carolina, and Nebraska. The counties had total populations ranging from 560 to 98,000. The measure of delinquency was the number of arrests of juveniles between 11 and 17 years of age. Primary dependent variables were arrests for homicide, forcible rape, aggravated assault, robbery, weapon offenses, simple assault, and the Uniform Crime Reporting violence index. Explanatory variables associated with social disorganization theory, based primarily on 1990 census data, included ethnic heterogeneity, family disruption, poverty, and proximity to metropolitan counties. Data analysis used negative binomial regression because the small number of arrests in many counties meant arrest rates would be ill-suited to least squares regression. Findings supported the generality of social disorganization theory. Juvenile violence was associated with rates of residential instability, family disruption, and ethnic heterogeneity. Although poverty rates were not related to juvenile violence, this finding was also in accord with social disorganization theory because, unlike urban settings, poverty was negatively related to residential instability. Juvenile violence rates varied markedly with population size through a curvilinear relationship in which counties with the smallest juvenile populations had exceptionally low arrest rates. Data on partial correlations among explanatory variables are appended. 68 references, 7 footnotes, 4 tables, and 1 figure