The mechanism and site(s) of action of volatile anesthetics are unknown. In all organisms studied, volatile anesthetics adhere to the Meyer-Overton relationship--that is, a ln-ln plot of the oil-gas partition coefficients versus the potencies yields a straight line with a slope of -1. This relationship has led to two conclusions about the site of action of volatile anesthetics. (i) It has properties similar to the lipid used to determine the oil-gas partition coefficients. (ii) All volatile anesthetics cause anesthesia by affecting a single site. In Caenorhabditis elegans, we have identified two mutants with altered sensitivities to only some volatile anesthetics. These two mutants, unc-79 and unc-80, confer large increases in sensitivity to very lipid soluble agents but have little or no increases to other agents. In addition, a class of extragenic suppressor mutations exists that suppresses some altered sensitivities but specifically does not suppress the altered sensitivity to diethyl ether. There is much debate concerning the molecular nature of the site(s) of anesthetic action. One point of discussion is whether the site(s) consists of a purely lipid binding site or if protein is involved. The simplest explanation of our observations is that volatile anesthetics cause immobility in C. elegans by specifically interacting with multiple sites. This model is in turn more consistent with involvement of protein at the site(s) of action.