Hemodialysis is accepted as a routine treatment now, but its success still depends on making a connection to the patient's circulation. Arteriovenous shunts were the first practical way to do it; they made chronic dialysis possible, and, in acute renal failure, ended the race between return of renal function and exhaustion of cutdown sites that was a feature of early dialytic therapy. The original Scribner shunt has been supplemented and, to some extent, replaced by other prostheses, arteriovenous fistulas, and grafts of autogenous or animal origin. Despite them, the access problem is far from solved. The number of patients receiving dialysis increases every year, this growth recently stimulated by federal funding of renal failure therapy, and the composition of the dialysis patient group is changing. There are more old patients, more with vascular disease, more who have spent years receiving dialysis or are once again receiving it after transplant failure.