The oligomeric assembly of the lactose repressor has been studied by a series of deletion and extension mutants at the C-terminal sequence. Deletion of the presumed mini leucine zipper sequence at the C-terminus (-5 aa, -11 aa, -18 aa, and -32 aa) has resulted in a family of dimeric repressors, while deletion of the last four amino acids (-4 aa) does not affect the oligomeric state or function of the repressor. Furthermore, extension of the C-terminus with leucine zipper-forming sequences (G359V+5 and GCN/C) has been deduced to strengthen a subunit interaction in tetramers. The role of the presumed mini leucine zipper in tetramer formation for the lac repressor is therefore confirmed. In addition, a long-axis dimer (GCN/C-Y282D) is obtained from increasing the leucine zipper length and disrupting a presumably distinctive interface composed of Tyr\sp282 and Lys\sp84. The generation of this altered protein confirms the postulate that there are two distinctive subunit interfaces for the assembly of tetrameric repressor. Detailed characterization of these mutant repressors has provided insight into the relationship between structure and function in the lac repressor. The subunit communication upon inducer-binding in a tetrameric repressor is between the two side-by-side monomers and does not appear to involve monomers in end-to-end orientation. The lower operator-affinity observed for the deletion dimers compared to the tetramer is derived from a difference in the association process.