The Chillingham herd of white cattle has been isolated and confined in a park in northern England for several centuries. Blood grouping confirmed the cattle to be remarkably homozygous. Random fixation of harmful alleles and consequent extinction have presumably been prevented by selection, and this paper discusses possible selective processes and the ways in which these have changed over the last hundred years. Herd records from 1862 to 1899 and 1953 to 1985 show that, in the former period, but not the latter, culling and castration took place. In both periods, breeding was not seasonal. Herd fertility (calves born per female) was higher in the latter period. Between 1953 and 1985, calves which survived for at least 12 months had a median date of birth (25 June) a month later than that of calves which did not survive. Conception intervals were rather longer and fecundity lower than those observed in commercial cattle. K-factor analysis showed mortality to differ in its causes between the sexes. A multiple regression model showed January-May rainfall, and population size on 1 January to influence mortality rate of the January-May period. The Chillingham cattle have evidently been, and continue to be, subjected to rigorous selection and this presumably underlies the survival of this herd.