We consider here small-length-scale crystal structures with two clearly different molecular components (e.g., hydrophobic and hydrophilic). Using a perspective developed by studies on large-length-scale block copolymers and liquid crystals, we focus on the crystalline interface between the two components. We examine four types of two-component crystals: aromatic ammonium carboxylates, aromatic oligo(ethylene oxides), cyclohexylammonium carboxylates, and ether-thioether compounds. Of the 111 crystal structures found in the Cambridge Structure Database (CSD), 108 adopt one of the five generic topologies found in diblock copolymers: spheres, columns, perforated layers, layers, and bicontinuous structures. As in diblock copolymers, a key factor controlling the interfacial topology is shown to be the volume ratio of the two components. When the volume fraction of one component is less than 30% of the whole, more than five-sixths of the examined crystal structures are of columnar or spherical type. For volume fractions between 40 and 50% more than three-quarters are of lamellar or bicontinuous type. We use this model to predict the topologies of small-length-scale two-component crystals. We predict the crystal topolgies of six new crystal structures: three are predicted to be columnar, and the other three, lamellar or bicontinuous. The crystal structures of these systems were then determined by single-crystal X-ray methods. Five of the structures form in topologies consistent with the predictions: three in columns and two in layers. The remaining one forms as a perforated layer instead of the predicted columnar structure. Such predictive accuracy is consistent with the statistics of the CSD investigation.