The attraction of monoclonal antibodies lies in the potential availability of an unlimited supply of material, homogeneity, reproducibility, and the feasibility of manipulation for specific purposes. Their use, however, is limited because of structural and physicochemical idiosyncrasies, occasionally restrictive factors influencing affinity and avidity, unique cross-reactivities, lack of multivalency compared to polyclonal antibodies, and restricted Fc region-specific functions. Approaches to the isolation, purification and fragmentation of monoclonal antibodies derive from basic concepts of the four chain and domain structure of immunoglobulins that have developed from the study of human M-components and a large number of well-characterized mouse myeloma proteins. The specific antibody activity used to generate the hybridoma provides an additional property that can be utilized to advantage in selecting clones and purifying material. Each monoclonal antibody is unique and presents its own set of problems affecting isolation, stability and suitability for chemical modification and radiolabeling. Isotypic determinants correlate with biological activities such as binding to staphylococcal protein A, Clq or specific Fc receptors, that may be important for purification and therapy. Analysis of fragmentation patterns obtained with various proteolytic enzymes [F(ab′)2, Fab, Fab′, Fab/c] can be related to accumulating protein and DNA sequence data. Fragments generated need to be defined for each hybridoma and may present additional idiosyncrasies. These include instability in standard solvents, poor reconstitutability following reduction, and loss of affinity for antigen. To some extent, idiosyncrasies reflect the idiotypic specificities of the antibody. Structural correlates of idiotypy include the binding site for a specific antigen (hypervariable regions), framework determinants, and conformational antigens that may be lost on varying pH, exposure to denaturants, or separation of heavy and light chains. In spite of these problems, however, fragments are more interesting therapeutically, as well as for the design of specific assays, because of loss of effector functions and the potential for artificial constructs. These properties have been exploited in the use of somatic mutants, transfectomas and hybrid molecules. Injection of such fragments to another species may bypass most (though not necessarily all) of the host response to isotypic and allotypic determinants, and still permit various anti-idiotypic antibodies to develop, only some of which may in fact interfere with antibody binding.