A humanized IgG antibody to CAMPATH‐1 antigen (CDw52) is known to be lympholytic both in vitro and in vivo. So as to improve therapeutic potency through protein engineering strategies, we wish to define the structural motifs underlying some of the documented differences in function between human (h) IgG1 and IgG4 forms of the antibody. By the creation of heavy chain domain‐switch and intra‐domain recombinant antibodies we have established an important role for the carboxy‐terminal half of the CH2 domain in determining differential behaviour in antibody‐dependent cytotoxicity (ADCC) and in complement lysis. If this same region were necessary for the effector mechanisms that operate in vivo, then it might be possible to improve antibody effector functions by construction of novel antibodies that possess within the one molecule multiple copies of the crucial hinge‐CH2 associated structures. Although our previous work suggested that the hIgG4 CAMPATH‐1 antibody was ineffective at ADCC, we found this to be so only in some individuals. In others, IgG4, and indeed all the IgG subclasses were able to mediate ADCC. Overall, though, hIgG1 remains the best choice isotype for lytic therapy in vivo.