Two separate groups of dyslexia families were ascertained through probands attending special schools for dyslexic students. An additional control group of families was ascertained through randomly selected students attending public schools. The 3 groups were interviewed by questionnaire about the family's demographic characteristics, and about the incidence of reading and spelling disorders in all first and second order relatives. One group of dyslexia families was also examined by standardized intelligence and academic achievement tests. Developmental dyslexia was found to aggregate in families; there were 4-5 times as many affected males as females among clinically identified students attending the special schools, but the sex ratio of affected relatives after probands had been excluded was approximately 1.4 males for every female. Sibs were at greater risk for reading difficulties when one parent was affected than when neither parent was affected. Sibs were also at greater risk for academic difficulties, and affected sibs were more severely impaired, when the father rather than the mother was the affected parent. In dyslexia families with 2 affected parents, the sibs were at greater risk, and the affected sibs were more severely impaired, than in families where only one parent was affected. Moreover, in families with 2 affected parents, both of the parents were more severely impaired in reading and spelling than parents of the same sex in families with one affected parent. Some indirect evidence is presented that assortative mating may codetermine patterns of affectedness in dyslexia families.