A study has been made of the effect of ureteral ligation on the susceptibility of the kidney to pyogenic infection. In most experiments a strain of E. coli was employed as the test organism, being injected intravenously in varying quantity either before or after ureteral ligation. A few experiments were also carried out with S. marcescens. Preliminary observations were made on the distribution and persistence of E. coli following its inoculation into the blood stream of normal rats. Rapid reduction in number of bacteria in the circulation occurred during the first 30 minutes, but bacteriemia persisted at a comparatively low level for at least 48 hours. Large proportions of the inoculated bacteria were arrested and apparently destroyed in the liver, spleen, and lungs. Comparatively small numbers were deposited in the kidneys; nevertheless, these continued to be demonstrable during the 1st week, without notable tendency to increase or decrease, then disappeared during the 2nd week. There was no acceleration in rate of disposal of the bacteria in the kidney when a second injection was made 1 week after the first. In rats with one ureter ligated the number of bacteria lodging in the kidneys after intravenous inoculation did not differ from that found in normal animals. It appears, therefore, that the increased susceptibility of the obstructed kidney to infection via the blood stream is not attributable to an increased trapping of circulating bacteria. 4 to 6 hours after the intravenous injection, however, an increased number of bacteria could be demonstrated in the obstructed kidney, apparently due to local multiplication, and by the end of 24 hours purulent infection was usually obvious. A comparatively large number of bacteria was required to cause infection, even in the kidney with obstruction. This appeared to be related to the small proportion of the intravenous inoculum which lodged in the kidney initially. Although bacteria could be demonstrated in the normal kidney for a week or more following intravenous injection it was not possible to induce active infection with equal regularity by ligating the ureter throughout this time. During the first 3 days the majority of obstructed kidneys developed infection, but after 5 or more days this occurred in only a small proportion of animals so treated. The reason for the difference, in relation to interval between intravenous injection and time of ligation, is not apparent. When the ureter was ligated but no intravenous injection of bacteria was given, staphylococcal infection developed in the obstructed kidney within 2 weeks in about one-third of the animals. Reasons are given for the belief that this was blood-borne infection, and not the result of contamination at the time of operation. Staphylococci were not recovered from the normal rat kidney. These "spontaneous" staphylococcal infections seldom developed when E. coli was injected intravenously at the time of ureteral ligation.