We examined post-spill trends (1989-1998) of marine bird populations in Prince William Sound (PWS) following the Exxon Valdez oil spill (EVOS) to evaluate recovery of injured taxa. Two criteria were employed. First, we examined population trends of injured taxa only in the oiled area of PWS using regression models. Second, we examined population trends of injured taxa in the oiled area relative to the unoiled area using homogeneity of the slopes tests. We considered a population recovering if there was a positive trend using either criteria. We considered a population not recovering if there was no trend using either criteria or a negative trend in the oiled area. A significant negative trend in the oiled area relative to the unoiled area was considered a continuing and increasing effect. Most taxa for which injury was previously demonstrated were not recovering and some taxa showed evidence of increasing effects nine years after the oil spill. Four taxa (loons Gavia spp, Harlequin Duck Histrionicus histrionicus, Bufflehead Bucephala spp, and North-western Crow Corvus caurinus) showed weak to very weak evidence of recovery. None of these taxa showed positive trends in both winter and summer. Nine taxa (grebes Podiceps spp, cormorants Phalacrocorax spp, Black Oystercatcher Haematopus bachmani, Mew Gull Larus canus, Glaucous-winged Gull Larus glaucescens, terns Sterna spp, murres Uria spp, Pigeon Guillemot Cepphus columba, and murrelets Brachyramphus spp) showed no evidence of recovery during summer or winter. Four taxa (scoters Melanitta spp, mergansers Mergus spp, goldeneyes Bucephala spp, and Black-legged Kittiwaka Rissa tridactyla) showed evidence of continuing, increasing effects. We showed evidence of slow recovery, lack of recovery, and divergent population trends in many taxa which utilize shoreline and nearshore habitats where oil is likely to persist. Potential lingering spill effects and natural variability appear to be acting in concert in delaying recovery of many PWS bird populations.