Detecting patterns of spatial genetic structure (SGS) can help identify intrinsic and extrinsic barriers to gene flow within metapopulations. For marine organisms such as coral reef fishes, identifying these barriers is critical to predicting evolutionary dynam- ics and demarcating evolutionarily significant units for conservation. In this study, we adopted an alternative hypothesis-testing framework to identify the patterns and pre- dictors of SGS in the Caribbean reef fish Elacatinus lori. First, genetic structure was estimated using nuclear microsatellites and mitochondrial cytochrome b sequences. Next, clustering and network analyses were applied to visualize patterns of SGS. Finally, logistic regressions and linear mixed models were used to identify the predic- tors of SGS. Both sets of markers revealed low global structure: mitochondrial ΦST = 0.12, microsatellite FST = 0.0056. However, there was high variability among pairwise estimates, ranging from no differentiation between sites on contiguous reef (ΦST = 0) to strong differentiation between sites separated by ocean expanses ≥ 20 km (maximum ΦST = 0.65). Genetic clustering and statistical analyses provided additional support for the hypothesis that seascape discontinuity, represented by oceanic breaks between patches of reef habitat, is a key predictor of SGS in E. lori. Notably, the esti- mated patterns and predictors of SGS were consistent between both sets of markers. Combined with previous studies of dispersal in E. lori, these results suggest that the interaction between seascape continuity and the dispersal kernel plays an important role in determining genetic connectivity within metapopulations.