Human, as a species, is very variable, and his variability is at the basis of his social organization. This variability is maintained, in part, by the chance effects of gene assortment and the variation in these genes is the result of mutations in the past. If our remote ancestors had not mutated we would not he here; further, since no species is likely to he able to reduce its mutation rate substantially by the sort of selection to which it is exposed, we may regard mutations of recent origin as part of the price of having evolved. We are here: all of us have some imperfections we would wish not to have, and many of us are seriously incommoded by poor sight, hearing or thinking. Others among us suffer from some malformation due to faulty development. A few are formed lacking some essential substance necessary to metabolize a normal diet, to clot the blood, or to darken the back of the eye. We will all die and our deaths will normally be related to some variation in our immu-nological defences, in our ability to maintain our arteries free from occlusion, or in some other physiological aptitude. This massive variation, which is the consequence both of chance in the distribution of alleles and variety in the alleles themselves, imposes severe disabilities and handicaps on a substantial proportion of our population. The prospects of reducing this burden by artificial selection from counsel¬ling or selective feticide will be considered and some numerical estimates made of its efficiency and efficacy. Screening is a procedure by which populations are separated into groups, and is widely used for administrative and other purposes. At birth all babies are sexed and divided into two groups. Later the educable majority is selected from the ineducable minority; later still screening continues for both administrative and medical purposes. Any procedure by which populations are sifted into distinct groups is a form of screening, the word being derived from the coarse filter used to separate earth and stones. In medicine its essential features are that the population to be screen¬ed is not knowingly in need of medical attention and the action is taken on behalf of this population for its essential good. A simple example is provided by cervical smear examination, the necessary rationale for which must be the haimless and reliable detection of precancerous changes which can be prevented from becoming irreversible. Any rational decision on the development of such a service must be based on a balance of good and harm and any question of priorities in relation to other services must be based on costing. The balance of good and harm is a value judgement of some complexity. In the example of cervical smears anxiety and the consequences of the occasional removal of a healthy uterus must be weighed against the benefits of the complete removal of a cancerous one, and such matters cannot be costed in monetary terms. In fact, even such an apparently simple procedure as cervical screening is full of unknowns and many of these unknowns can only be resolved by extensive and properly designed studies. In genetic screening the matter is even more complicated, since the screening is often vicarious; that is, one person is screened in order to make a prediction on what may happen to someone else, usually their children, who may be un¬conceived or unborn. Further, the action of such screening may not be designed to ameliorate disease, but to eliminate a fetus which has a high chance of an affliction, or to prevent a marriage in which there is a mutual predisposition to producing abnormal children. These considerations impose very considerable dif¬ferences, since the relative values placed on marriage, on having children within marriage, and on inducing abortion, vary widely between individuals and between societies.