An ecological survey was conducted from March, 1995 to February 1996 to determine sources of Salmonella colonization in two flocks of turkeys reared consecutively in a newly constructed facility. Sampling was conducted prior to placement of poults, at Day 0, and again at 2, 10, 14, and 18 wk. Samples were collected at comparable times for the second flock except that final sampling occurred at 22 wk instead of 18 wk. Poult box liners, birds, new litter, drinkers, and air were sampled. Feed was collected from each shipment upon arrival at the facility. Feeders, drinkers, and used litter were monitored to evaluate horizontal transmission. Conventional methods for isolation and serological screening were used to analyze samples for the presence or absence of Salmonella. Prior to placement of the first flock, litter, drinkers, and air samples were negative for the presence of Salmonella, whereas drinkers were positive prior to placement of the second flock. Following placement of poults, 51.1, 63.8, and 22.8% of all litter, drinker, and air samples, respectively, were positive. Salmonella was isolated from 13.6% of poult box liners, 25.0% of yolk sac samples, and 53.8% of ceca, excluding Day 0. Salmonella was isolated from 14.8% of feed shipments and 39.1% of feeder contents. Frequency of Salmonella detection was higher P < 0.05) in Flock 1 than Flock 2 for cecal and air samples. Salmonella colonization of turkey flocks and the spread of Salmonella within the environment was extensive once initial contamination of the production house occurred. Drinkers, feeders, litter, and air were critical sources of horizontal transmission within each pen as well as between pens.