The variability of anisometropi in a sample of 310 children with astigmatism at the age of 1 year was longitudinally studied during a 3-year period between 1 and 4 years of age. The prevalence of anisometropia of 1 D or more at each year level was rather stable. When individual cases were examined we found that between the first and the last test session 19 of the 33 children with anisometropia at the first test session had become non-anisometropic and were substituted with 14 new cases which were non-anisometropic at the age of 1 year. In general, less than half of the cases, at all levels of anisometropia, remained anisometropic throughout the whole test period. We also found that children with anisometropia persisting through the whole test period were at considerable risk, about one out of four, of developing amblyopia. There was no simple relationship, however, between anisometropia at a certain age level between 1 and 4 years and amblyopia and/or strabismus. Non-persisting anisometropia in an emmetropizing eye is in most cases a benign sign and not connected with an increased risk for developing amblyopia.