Extensive clinical experience has been obtained in the use of opiates during the last decade in special units devoted to symptom control in advanced cancer. Important contradictions have emerged with the clinical pharmacological literature on opiates calling into question its relevance to the treatment of chronic pain. Specifically in the case of morphine it is clear that: it is a very effective analgesic given orally, dosage must be individualized, parenteral use or exotic analgesic 'cocktails' are usually unnecessary, and tolerance, dependence and respiratory depression are rarely common or serious problems which prevent effective pain control provided morphine is used appropriately in accordance with its pharmacological characteristics. Heroin is a suitable alternative to morphine (particularly for intramuscular administration) if differences in milligram potency are taken into account, but has no advantages in terms of either analgesic efficacy or side effects. This paper summarizes clinical experience in the use of oral morphine for cancer pain at St. Christopher's Hospice, any data from clinical investigations which support this approach, and comments on the areas of controversy which have emerged.