The spatial relationship between activities which generate trips and motor vehicle accidents is examined and applied to the City and County of Honolulu for 1990. A method is described for examining spatial variations in motor vehicle accidents, aggregated into small geographical areas. A spatial lag model is developed which examines the zonal relationship of motor vehicle accidents to population, employment and road characteristics. The resulting model is tested for each hour of the day, weekdays and weekends separately. The predictors of accidents fluctuate according to different trip generating activities and change considerably over the day. The method can generate expected accidents, allowing for an identification of areas which have higher than expected levels. It is argued that this method focuses attention on characteristics of neighborhoods and areas, and not just on the road system.