Understanding the connectivity of marine populations is vital for conservation and fisheries management, particularly for the strategic design of reserve systems. A recent proliferation of molecular and statistical tools allows increasingly sophisticated integration of genetic and geographic data (e.g., Manel et al., 2003). Such advances have fueled considerable hope that the challenging problem of tracking movement of individuals within the vast ocean will soon be solved. Here, we focus on some of the inherent limitations of genetic approaches to inferring connectivity, particularly in marine environments. More optimistically, we also point to a number of situations where genetic approaches have been particularly successful in the past, as well as newer integrative approaches that deserve further attention.