Biological invasions in freshwater communities can result in great economic and ecological harm, including the loss of native species and ecosystem function; prediction of the course of invasions is important in targeting efforts to prevent or mitigate such damage. The aim of this research is to combine ideas from metacommunity theory with invasion ecology to develop a predictive model of the success of invasive fishes in stream systems. Principles of metacommunity theory and invasion ecology suggest that invasive species might be disproportionately successful in either well-connected or poorly-connected sites within stream networks; the former because of the ease of dispersal for invaders into well-connected sites, the latter due to poorly-connected sites having smaller, less diverse native communities which may be more vulnerable to invasion. This research will investigate how the balance between these two contrary effects for a given invader is related to the traits of that species and of the watershed it has invaded. Using fish community assessment data from thousands of streams in Minnesota, Iowa, Wisconsin, Illinois, and North Carolina, we will analyze how the effect of connectivity (measured as network centrality) on invasion success relates to selected species traits, using a generalized linear mixed model. We hypothesize that ecological traits related to the ability to disperse to many new sites (e.g. fecundity) and the ability to establish a foothold despite biotic resistance (e.g. habitat generalism) may mediate the effect of connectivity on invasion success. Our model will provide valuable insights into what types of invasive species are more likely to establish populations in different areas of the network.