The scientific problems in the development of nuclear power divide conveniently into the physical, chemical and metallurgical problems. The physicist is concerned with prediction of the nuclear performance of reactors. He has to be able to predict the critical size and the amounts of fuel and the degree of enrichment required to enable it to operate at the designed heat output. He has further to be able to predict the course of the reactivity - how much heat can be extracted per ton of fuel before the fuel has to be changed. In order to do this he must know a great deal about the interaction of neutrons with the important heavy element isotopes inducing plutonium 240 and 241. He must know how the number of neutrons per fission changes with neutron energy and the relative chances of a neutron being captured with and without producing fission. He must know the chances of neutrons being absorbed by the fission products formed in the fuel elements, and any non-fissile materials used by the metallurgists and engineers. Different data are required for fast and thermal reactors. The physicist must also keep a watchful eye on the safety of reactors - he must be wary lest changes in temperature or operating conditions induce additional reactivity and cause power increases. He must also study the effects of high speed, high intensity neutron bombardment on the properties of important structural materials in the reactor such as graphite and steel. The chemists and metallurgists have the problem of providing all the nuclear materials - fissile and non-fissile - in a high degree of purity. The chemist must devise processes for separation of fertile materials, fissile materials and fission products. The metallurgist has to understand the effect of radiation on uranium and thorium metal at the operating temperature and has to devise methods of counteracting deleterious effects. To do this he has to study the various possible alloys or cermets bearing in mind the prohibitions of the physicist who dislikes neutron absorbing materials and the chemist who dislikes the job of processing highly alloyed metals. The chemists and metallurgists have also to study the compatibility of all the materials used with the coolants specified by the engineer. Thus the reaction of graphite with carbon-dioxide gas at 400° C in the presence of reactor radiation must be measured; the compatibility of liquid sodium with zirconium and stainless steel at 550° C must be determined for a reactor of the sodium-graphite type; solutions of uranyl sulphate in water must be compatible with structural materials used in homogeneous reactors. All these problems require prolonged study in test rigs or reactor experiments. The scientist must also plan ahead, for any new type of reactor is likely to take at least 10 years to come into full scale use from the commencement of work. During this period he has to carry out research and development on any new materials which may be required such as heavy water, thorium, zirconium and beryllium. Work has to be started many years before full scale use can be certain. The crystal ball is never more required than in the development of atomic energy.