Clinical experience with transplantation of kidneys infected prior to implantation led to a dog experiment designed to delineate the best mode of therapy. Five groups of dogs were homotransplanted and immunosuppressed. The donor kidneys were perfused for 4 hr. Dogs in group 1 were controls, received no gentamicin, and the perfusate for their transplant was sterile. The animals in group 2 received kidneys perfused with gentamicin-sensitive Escherichia coli-contaminated perfusate and were not protected with gentamicin. Average death in this group was within 4 days, secondary to anastomotic rupture or generalized sepsis. The dogs in group 3 were treated preoperatively and thereafter with gentamicin and received a perfusatecontaminated kidney. They did as well as the controls, with no evidence of sepsis at autopsy 10 days post-transplantation. The animals in group 4 received a perfusate-contaminated kidney but were protected with gentamicin beginning 1 day post-transplantation. Average death was at 6 days; most showed evidence of E. coli sepsis at autopsy. The dogs in group 5 received a kidney, the perfusate of which contained E. coli plus gentamicin. They did as well as the controls. To evaluate the effect of gentamicin in the perfusate, two subgroups of five dogs each were autotransplanted after 4 hr of perfusion with and without gentamicin. Comparison showed no detectable differences in serial serum creatinine levels and in microscopic changes of the transplanted kidneys after 10 days.