This paper reports on an analysis of geographically based data from four communities conducted to evaluate relationships between measures of the physical availability of alcohol and rates of driving after drinking. From a review of the literature, it was expected that rates of driving after dnnking would be directly related to the availability of alcohol at on-premise establishments. Based on theoretical arguments regarding the hfe activities which underlie drinking and driving it was expected that the effects of availability upon these outcomes would extend significantly beyond the local areas of outlets. Taking into account the geographic variations in environmental characteristics (road network density, traffic flow, population density), and socioeconomic (age, gender, race, marital status, income, employment) and drinking characteristics (rates of abstention, frequency and quantity of use) of resident populations, a spatial analysis of drinking driving and alcohol-related crashes was conducted. The results of the analysis showed that physical availability was unrelated to self-reports of driving after drinking and driving while intoxicated and significantly related to rates of single vehicle night-time crashes. In the latter case, physical availability affected both local and adjacent area rates of crashing.