Understanding molecular recognition of supramolecules for solid substrates is essential for designing chemical sensors and molecular devices. The rules of molecular recognition are well established at the level of single molecules. However, during the transition from molecular-scale devices to macroscopic devices, issues concerning control over recognition that are well-established at the molecular level become much more complex. Hopefully, the conceptual and practical considerations reported here will clarify some of these issues. The immune system uses antibodies to identify molecular surfaces through molecular recognition. Antibodies are thus appropriate tools to study the rules of macromolecule-surface interactions, and this was done using crystal surfaces as substrates. Crystals can be formed or introduced into organisms and should be thus treated by the organism as any other intruder, by eliciting antibodies specific to their surfaces. A structure-recognizing antibody is defined here as complementary to a certain ordered supramolecular organization. It can be considered as a mold bearing in its binding site memory of the organization against which it was elicited. On the surface of a crystal composed of relatively small organic molecules, an antibody binding site would encompass an array of 10-20 molecular moieties. The antibody binding site would not detect one molecule, but rather a two- or three-dimensional molecular arrangement on the surface, similar to a macromolecular surface. The complementarity between antibody binding site and surface is supported by stereoselective supramolecular interactions to the repetitive structural motifs that are exposed at the surface. A procedure was developed in order to isolate monoclonal antibodies that specifically recognize a certain crystalline surface. The procedure was applied in particular to crystals of cholesterol monohydrate, of 1,4-dinitrobenzene, and of the tripeptide (S)leucine-(S)leucine-(S)tyrosine (LLY). A series of antibodies were selected and studied, three of which provided reliable specific antibody-antigen structural models. The three docking models show an astounding geometrical and chemical match of the antibody binding sites on the respective crystal surfaces. We also showed that antibodies are intrinsically capable of recognition at the length scale necessary for detection of chirality. Once the structural parameters determining the antibody specificity to the target surfaces are characterized, the antibodies may be conceivably used as reporters of the existence and location of target domains with similar structure in biological milieus. In this context, we developed and characterized monoclonal antibodies specific to crystalline mixed monolayers of cholesterol and ceramide, fundamental building blocks of lipid microdomains in cellular membranes. When used on cells, one antibody indeed labels cell membrane domains composed of cholesterol and ceramide. The fundamental contribution of the approach developed here may be in the antibody ability to report on the structural organization of paracrystalline domains that cannot be determined by other means. Alternatively, structure-recognizing antibodies may be conceivably used to carry information or build connections to specific targets, which may offer interesting developments in medicine or electronics.