We conducted a 4-year study of beavers Castor canadensis to compare the movements, survival and habitat of adults established in existing colonies to juveniles dispersing to new sites in a region with high beaver densities along a suburban-rural gradient. Estimates of annual survival were high for adult and juvenile beavers. Of nine known mortalities, seven (78%) were juveniles. Mortalities occurred during spring-summer, and none during fall-winter. There was a trend toward higher-to-lower survival along the suburban-rural gradient, respectively. Human-induced mortality (e.g. trapping and shooting) was higher in rural areas, whereas nonhuman-induced mortality (e.g. disease, accidents) was higher in suburban areas. Fifteen (14 subadults and one adult) beavers moved from natal colonies to other areas. The average dispersal distance for subadults was 4.5 km (SE = 1.0) along streams or rivers, or 3.5 km (SE = 0.7) straight-line point-to-point. Most dispersal movements were made in spring (April-June). In two cases, individual subadults made return movements from their dispersal sites back to their natal colonies. Dispersal sites tended to be in smaller, shallower wetlands or streams and in areas with higher overstorey canopy closure compared with natal colonies. Woody vegetation usually preferred by beavers for food tended to be less common at dispersal sites than at natal colonies. In regions with high densities of beaver, dispersing juveniles are likely to attempt to colonize lower quality sites. High densities of beavers also lead to more human-beaver conflicts and, in Massachusetts, the pest control management options in place during the past decade have been ineffectual at controlling population levels. Alternately, in regions with no beavers or very low densities and where reintroductions are being attempted, the landscape matrix surrounding release sites should include suitable sites for dispersing young to establish colonies.