Evidence that GAP is an effector of ras action can be summarized as follows: GAP interacts at a site on p21 defined genetically as the effector binding site. Regions of p21 that are nonessential for biological activity are nonessential for GAP interaction. GAP interacts with all known types of p21. (Upstream factors are expected to be specific for individual types). GAP interacts with p21 proteins (normal and mutant) in a GTP-dependent fashion. None of these constitute proof. It remains possible that GAP simply regulates p21-GTP levels, and binds to the same site as the true effector without transmitting a downstream signal. If indeed GAP mediates ras action, the question immediately arises as to the biochemical function of GAP itself. The requirement of ras proteins for membrane localization to exert their effects may be a valuable clue in the search for this function. Perhaps GAP is an enzyme (or is bound to an enzyme) that acts on membrane components in a p21-GTP-dependent manner and in doing so transmits signals to other downstream effectors. The ability of GAP to interact with many members of the ras family would allow many upstream signals to feed into this downstream pathway. Clearly, proof (or disproof) that GAP is downstream of ras is the next step toward clarification of this aspect of ras action; identification of biochemical activities associated with GAP (or the true ras effector) will, we hope, follow soon.