Patterson (1998) argues that the family structure of African Americans has been harmed by the legacy of slavery. As a result of slavery, males were not able to protect their families from the devastating effects associated with slavery nor could they prevent the break-up of slave families that occurred when members were sold to other slave owners. We test this argument by determining if the percent of African American households that are female headed, as measured in the 2000 census for U.S. Counties, is higher in parts of the South that had more slaves in 1850. We extend Patterson's argument by testing for the direct effects of slavery on county level homicides committed by African American and indirect effects of slavery that operate through female headed households. The analysis is replicated for whites. Results indicate that percent slave in 1850 is negatively related to white female headed households in the year 2000 and are not related to white homicides; there, slavery produces a positive indirect effect on African American homicide that operates through African American homicide that operates through African American female headed households.