Abstract Ecological comparisons of transplanted, natural (reference) and disrupted (unvegetated) marsh sites on the Fraser River estuary, British Columbia, were conducted between 1991 and 1994. The study examined vegetative biomass and cover, invertebrate abundance, fish abundance, fish residency, fish food, and submergence time for the three habitats. Standing crop biomass at three transplant sites was within the range of values for reference sites, but was much lower at an unstable site where sediment slumping had occurred. The percent cover of Lyngbyei's sedge (Carex lyngbyei) in eight transplant sites was 0.05) was observed among marsh sites when chum salmon (Oncorhynchus keta) and chinook salmon (O. tschawytscha) fry abundance were compared. However, chinook and sockeye smolt catches were significantly different (p>0.05) was observed among marsh sites when chum salmon (Oncorhynchus keta) and chinook salmon (O. tschawytscha) fry abundance were compared. However, chinook and sockeye smolt catches were significantly different (p<0.05) among marsh sites and were usually higher at disrupted sites. In nine sites in the North Arm and Deas Slough area chum fry residency was examined. At one transplant site (DEI) marked chum fry were caught up to 48 h after release. No fry were caught 1 h after release at a transplant site (Dll) and a disrupted site (DE4). At the remaining sites, fry were caught up to 1 and 3 h after release. At all sites, over 80% of the total number of food organisms examined in chum fry stomachs were harpacticoid copepods. Mean submergence time for reference marshes ranged from 33.2 to 50.7%, but for transplanted sites the value ranged from 26.4 to 60.1%. Our study shows that numerous factors need to be examined in determining if restored marshes will function as natural habitats. The development of a standardized set of reference criteria would assist in evaluating whether or not transplanted marshes are functioning as designed.