Fish faunas across the continental United States have become more similar through time because of widespread introductions of a group of cosmopolitan species intended to enhance food and sport fisheries. On average, pairs of states have 15.4 more species in common now than before European settlement of North America. The 89 pairs of states that formerly had no species in common now share an average of 25.2 species. Introductions have played a larger role than extirpations in homogenizing fish faunas. Western and New England states have received the most introductions, which is a reflection of the small number of native fishes in these areas considered desirable gamefish by settlers.
Disruption of movement patterns due to alterations in habitat connectivity is a pervasive effect of humans on animal populations. In many terrestrial and aquatic systems, there is increasing tension between the need to simultaneously allow passage of some species while blocking the passage of other species. We explore the ecological basis for selective fragmentation of riverine systems where the need to restrict movements of invasive species conflicts with the need to allow passage of species of commercial, recreational, or conservation concern. We develop a trait‐based framework for selective fish passage based on understanding the types of movements displayed by fishes and the role of ecological filters in determining the spatial distributions of fishes. We then synthesize information on trait‐based mechanisms involved with these filters to create a multidimensional niche space based on attributes such as physical capabilities, body morphology, sensory capabilities, behavior, and movement phenology. Following this, we review how these mechanisms have been applied to achieve selective fish passage across anthropogenic barriers. To date, trap‐and‐sort or capture‐translocation efforts provide the best options for movement filters that are completely species selective, but these methods are hampered by the continual, high cost of manual sorting. Other less effective methods of selective passage risk collateral damage in the form of lower or higher than desired levels of passage. Fruitful areas for future work include using combinations of ecological and behavioral traits to passively segregate species; using taxon‐specific chemical or auditory cues to direct unwanted species away from passageways and into physical or ecological traps while attracting desirable species to passageways; and developing automated sorting mechanisms based on fish recognition systems. The trait‐based approach proposed for fish could serve as a template for selective fragmentation in other ecological systems.