The purpose of the study was to explore the applicability of gateway theory to the development of substance use in homeless adolescents. Although the gateway theory predicts an orderly progression through substance use, it is not an inevitable progression. In this respect, the gateway hypothesis should be distinguished from more alarmist theories, which hold that any substance use necessarily leads to the use of more harmful illicit drugs. By contrast, the gateway hypothesis argues that individuals are more likely to initiate use of subsequent substances after having used the posited gateway drugs. Since there is no implied inevitable progression, there is a likelihood of abstinence or progression at each stage. In the current study, all participants were homeless at the time of contact, defined as not having lived with their parents within the past week, not being within the custody of the State, and not residing in a group home for more than 45 days. The majority reported that they made the decision to leave home themselves (60 percent); 13 percent reported a mutual decision between themselves and parents; and 21 percent said they were asked by their parents to leave; for 5 percent, the authorities removed them from the home. The study measured alcohol use, along with marijuana and other drug use. Sequential use of substances was analyzed by determining the particular pattern of progression for each individual, with drugs classified into the categories of alcohol, marijuana, and other drugs. Based on the gateway theory, participants were categorized into six profiles to describe the order in which they initiated use of the various substances, or they were classified as "nonprogressors" if they had not tried all three classes of drugs. Youth who progressed in the hypothesized gateway order (i.e., alcohol preceding marijuana, followed by other drugs) initiated their use at an earlier age than youth who had not progressed through all three substance classes; however, there was no relationship between a substance initiation profile and current substance use. Implications of these findings include the recognition that homeless street youth may follow different patterns of substance use than normative groups. Further, interventions geared toward youth who use substances heavily should include attention to contextual factors in addition to substance-use history. 5 tables and 41 references