The usefulness of antibody complexation, as a way of increasing the chances of crystallization needs to be re-evaluated after many antibody complexes have been crystallized and their structure determined. It is somewhat striking that among these, only a small number is a complex with a large protein antigen. The problem is that the effort of raising, cleaving and purifying an Fab is rewarded only by an extra chance of getting crystals; depending on the relative likelihood of crystallization of the complexed and uncomplexed protein. The example of the complex between HIV gp120, CD4 and an Fab fragment from a neutralizing antibody suggests that further complexation of an antigen–antibody complex with a third protein could, by increasing the number of possible combinations, improve the likelihood of crystallization. We propose the use of Ig-binding proteins as a way of extending the method from HIV gp120 to all proteins for which there are monoclonal antibodies. We discuss this technique, combinatorial complex crystallization (CCC), as part of a multi-component system for the enhancement of crystallization of macromolecular complexes. The method makes use of single Ig-binding domains from Staphylococcus aureus protein A (SpA), Peptostreptococcus magnus protein L (PpL) and the streptococcal protein G (SpG). The generality of the method depends on the ability of these domains to interact with a large repertoire of antibodies without affecting antigen binding. There is strong evidence to suggest that these Ig-binding domains bind outside the antigen-combining site of the antibody without perturbing antigen binding. It is clear from the crystal structure of the single SpG domain complexed with an Fab that the interaction involves mainly the immunoglobulin CH1 domain, a region not involved in antigen recognition. We have recently determined the structure of the complex between a human Fab and the domain D from SpA and found that steric hindrance is unlikely even for large antigens. We find that such binding involves only the well conserved framework region of the variable domain of the antibody heavy chain (VH) and does not affect the conformation of the hypervariable loops that define the antigen recognition site. Thus this domain could be used to complex to Fab or Fv fragments derived from a wide variety of antibodies. While protein A complexes with the VH domain, protein L recognizes the VL region of immunoglobulins. Our recent study of the interaction between an Fab and a domain of protein L shows that the situation is very similar. Indeed this domain binds to the VL framework region outside the antigen binding site. Since individual domains from each of these three multi-domain proteins bind to well separated and independent locations on immunoglobulins, they can be combined to search for a suitable crystalline lattice. This allows us to propose a combinatorial method as a rational way to exploit antibody complexation for the crystallographic structure determination of proteins that are otherwise difficult to crystallize. The overall method has strong parallels with other combinatorial methods used elsewhere in biology and chemistry, and we propose that together with stoichiometry variation screening (SVS), it may further enhance the probability of crystallization.