As I flew last week over the Mekong River, then the Ganges Delta in Bangladesh and on over Gujarat, I could not help reflecting on the scale of catastrophes, from the 1931 estimate of two million deaths in one flood in China to the 15000 in the 1979 collapse of a dam in Gujarat. They have been listed by Fryer & Griffiths (1979). The ability to kill many people in man-made accidents is limited by the containment, whether in a car, train, aircraft or factory. Containment possibly makes society tolerate these risks while questioning the smaller risks, made all-pervasive by the concept of dose-commitment in radiological protection. When the same concept is used by Dr Inhaber to predict risk from SO 2 and airborne fluorides, many part company with him, although accepting the general thesis that economic activity carries known risks and spending money brings death or injury. This Discussion Meeting has ranged widely and has shown up some gaps. It has especially focused attention on the differing perceptions of risk and the failure of quantitative risk assessment to make much impact on them. Lord Ashby suggested that insurers know how to assess risk from past experience and sell policies by judging the perception of that risk. On the whole, their do not offer better terms than10 ‒4 /year in familar areas like house insurance. When it comes to betting, a 100 to 1 outsider attracts little money, and most of the figures discussed here mean nothing to the average man. They tend to agree with Damon Runyon’s assessment that everything in life is 6 to 4 against.