NMR residual dipolar couplings have great potential to provide rapid structural information for proteins in the solution state. This information even at low resolution may be used to advantage in proteomics projects that seek to annotate large numbers of gene products for entire genomes. In this paper, we describe a novel approach to the structural interpretation of dipolar couplings which is based on structural motif pattern recognition, where a predefined gapless structural template for a motif is used to search a set of residual dipolar couplings for good matches. We demonstrate the applicability of the method using synthetic and experimental data. We also provide an analysis of the statistical power of the method and the effects of order tensor frame orientation, motif size, and structural complexity on motif detection. Finally, we discuss remaining problems that must be overcome before the method can be used routinely to identify protein homologies.