An extensive body of research documents elevated rates of alcohol and use among MSM and evidence links alcohol use with unsafe sexual behavior, condom failure, and risks for HIV acquisition. However, there has been very little research on alcohol use interventions in the context of HIV prevention in this population. While most HIV prevention interventions for MSM have focused on individual level interventions, it has become increasingly clear that these strategies are insufficient to halt the spread of HIV and that approaches that combine individual and structural interventions should be developed. In our preliminary research with MSM in San Francisco we have found that bar attendance is very common along with high rates of heavy alcohol use and that consumption of water in bars is the most frequent tactic used to decrease intoxication and pace alcohol use. However bar patrons report several structural barriers to availability and accessibility of water. We propose to develo and pilot test a multi-level intervention to alter the environment where alcohol consumption occurs by installing a self-service water dispenser facilitating patrons' self-regulation of alcoho intake coupled with a bar media campaign focusing on alcohol and HIV risk (structural level interventions) and to influence individual alcohol consumption through real-time, normative feedback to exiting bar patrons of blood alcohol levels as measured by breathalyzer (individual level intervention). To evaluate the potential effectiveness of the multi-level structural intervention we will compare data from an exit sample 1,200 bar patrons taken from 4 pairs of matched intervention and control bars in San Francisco and web-based follow-up alcohol and sexual behavior survey. See aims for specific measures and comparisons. The specific aims of this developmental research are: 1) To measure exposure to and uptake of a structural intervention of free, self-service water in gay bars and an associated bar media campaign and its impact on bar patron blood alcohol concentration (BAC). We will compare mean quantity/frequency measures of alcohol use, free water awareness and usage measures, and mean BAC among exiting bar patrons between intervention bars with installed free water taps and media messages and control bars without taps and media. 2) To measure acceptability and utility of normative feedback of BAC and alcohol and sexual HIV risk. We will provide individual normative feedback to exiting bar patrons of their BAC level in comparison to real-time average BAC all exiting patrons displayed graphically along with reference points for bar average and distribution and legal driving limits. 3) To estimate impact of the free bar water structural intervention and normative feedback on alcohol-associated HIV risk. We will calculate the estimated multi-level structural intervention effect size on alcohol associated HIV risk behavior by comparing the mean proportion of HIV transmission risk acts reported in follow-up between exiting bar patrons sampled from intervention and control bars. To our knowledge this will be the first study to test a novel combination of structural and individual level intervention components targeted at reducing alcohol associated HIV risks in MSM. Results from this study will be used to design a large-scale community-wide effectiveness trial of the multi-level intervention.