The physiological breakdown of C3 has been studied using monoclonal anti-C3 antibodies, and it has been found that the later stages of this process--the breakdown of C3bi--is more complex than had previously been recognized. C3bi is the reaction product produced from C3b by the action of factor I which, in the presence of factor H, produces a double cleavage in the alpha chain of C3b. It is here reported that, both on cells and in the fluid phase, the breakdown of C3bi in serum gives rise to two products: C3c and the product previously described as alpha 2D, which we now propose to designate C3d,g. Alpha 2D differs from C3d in that it contains an additional fragment of approximately 8,000 mol wt that carries the antigenic determinant for the clone 9 monoclonal anti-C3 antibody. C3g cannot be precipitated by anti-C3 antisera and therefore behaves as a uni- or bideterminant antigen. The cleavage of C3d,g to C3d and C3g does not occur in sterile serum. It is also still uncertain what enzyme cleaves C3bi to C3c and C3d,g in plasma. Plasmin can do so in vitro, but plasminogen-depleted serum can still produce the cleavage. The antigenic determinant recognized by clone 9 in C3 is not exposed in C3 or C3b, but appears as a neoantigen in C3bi (and in C3d,g). Anti-C3g therefore is a potentially useful ligand for detecting complement-activation products. C3g represents a new, highly anionic C3 fragment and seems not to be identical with the C3e fragment described by others.