Introduction and Aims: Despite the well-documented comorbidity between disordered alcohol use and problem gambling, little is known about the co-occurrence of drinking and gambling in gambling venues. This paper appears to be the first to investigate the association between drinking and gambling behaviour among a large sample of gamblers during a specific, non-laboratory gambling episode. Design and Methods: We conducted a mail survey of all available households in the Northern Territory of Australia, including questions on drinking and gambling behaviour on the last visit to a gambling venue. We estimate the effect of moderate (1-4 standard drinks) and risky (>4 standard drinks) alcohol consumption on gambling participation and gambling duration for both problem and non-problem gamblers using regression analysis of 7044 survey responses. Results: The probability of participating in electronic gaming machine (EGM) gambling decreased with alcohol consumption for non-problem gamblers, while the probability of participating in TAB (Totalisator Agency Board, off-course totalisator) gambling increased with risky alcohol consumption for all gamblers. Alcohol consumption was not associated with EGM gambling participation for problem gamblers. Moderate alcohol consumption was negatively associated with EGM gambling duration, with a stronger effect observed for problem gamblers. Discussion and Conclusions: Moderate alcohol consumption is inversely correlated with both the duration of play and probability of participation for EGM gambling. Current laboratory studies do not predict the drinking-gambling behaviour of the general population in non-laboratory settings. Future research on alcohol and gambling co-occurrence must explicitly consider the drinking and gambling environment in order to produce policy-relevant findings.