Previous research has shown that burglaries tend to cluster in particular areas, at least in the short-term. The behavioral ecology perspective known as optimal foraging theory likens burglaries to animals foraging for food. Animals optimize foraging strategies to increase the rate of reward (food) while minimizing the time searching for food and the chances of being eaten by another animal. According to this perspective, offenders work under the same premise in that they seek to optimize the rewards of burglary while minimizing the time spent and the chances of being caught. As such, the theory predicts that in relation to residential burglary, offenders would target a neighborhood rather than just one home in order to increase the likelihood of rewards while minimizing the chances of being caught. However, it is also expected that the clusters of burglaries would shift in space over time in an effort to minimize the possibility of being caught. Thus, it is hypothesized that a burglary event would serve as a predictor for an elevated rate of burglary for properties within a range of 300 to 400 meters from the originally burgled home for a period of 1 to 2 months following the initial event. In order to examine the accuracy of this prediction, data on burglary for the county of Merseyside (United Kingdom) were analyzed for evidence of space-time burglary clusters moving in geographical space. Results of statistical analysis found that patterns of victimization tended to be similar in successive time periods, but also tended to change with time so that clusters of burglaries shifted, moving to nearby locations. Thus, optimal burglary foragers target potentially profitable homes and then move along when the rewards become too low or the risk of capture becomes too high. The findings have implications for law enforcement in that identified burglary hot-spots may move so that targeting hot-spots from last month may no longer be an effective police strategy as the optimal foraging offenders have already moved on. However, now that a pattern has been discerned, it may be possible to predict future hot spots based on data about current or past hot spots. Figures, tables, references