The colonizing potential of Escherichia coli K12 containing a vector coding for somidobove (bovine somatotropin) was determined. Treated male and female Fischer-344 rats were given a single oral gavage inoculum of sucrose with/without tetracycline (15 micrograms/ml). Untreated control animals received similar drinking water regimes. All animals survived until termination. There were no clinical signs of toxicity observed and no treatment-related effect upon body weight, food consumption, or efficiency of food utilization. Fresh fecal samples were collected from each rat every 24 h following inoculation and the population of the marked strain was quantitated until no bacterial colonies were observed for two consecutive days. While all inoculated rats were positive at 24 h, by 72 and 96 h all had become negative for the test (marked) strain, as were the corresponding control group throughout the test. The frozen stock of the marked strain used as the positive control demonstrated that the agar plates were selective for the test strain. Fourteen days following inoculation, all groups of rats were killed and the gastrointestinal tracts removed and treated to recover the marked strain. There was no evidence of the marked strain in the gastrointestinal tract of any from any group. Thus, the E. coli K12 host/vector system used in this experiment does not colonize the gastrointestinal tract of Fischer-344 rats.