Introduction Since the original work on regional perfusion in the therapy of cancer,3,12-14 leakage of perfusate has been described or alluded to by almost every author concerned with the subject.2-4,15,16 Further, some articles16 have had as one of their main theses the discussion of escape of perfusate to be expected when given areas are being perfused. Such concern and emphasis are understandable, since an important advantage to perfusion of cancer is the ability to treat intensively the isolated tumor with anticancer agents while sparing the body from otherwise lethal doses of the drug. Uncontrolled or excessive leakage of perfusate would defeat the very aim of this technique of cancer therapy. It appears that, although many authors allude to the problem of leakage,2-4,7,15,16 in the clinical application of perfusion arbitrary criteria are frequently used. Because of many unknowns it is often not clear whether the selected criteria