1. It is suggested that the incomplete antibody is essentially bivalent; one valence is nonreactive, and the other can become reactive under certain conditions.
2. The nonreactive valence of the incomplete antibody is believed to fix the antigen following trypsinization, or to fix nontrypsinized antigen with the aid of the colloid plasma factor.
3. The data obtained from our patients with acquired hemolytic anemia, further support the hypothesis presented; in addition, the data suggest the possibility of a special type of antibody, which acts essentially as a "doubly incomplete antibody," that is to say, that it possesses two nonreactive valences identical to the nonreactive valence of the incomplete antibody.