Using the data from five epidemiological studies of the total child population, the traditional distinction between general reading backwardness (i.e. reading below average for age, regardless of I.Q.) and specific reading retardation (i.e. a disability in reading relative to the child's age and general intellectual level) is put to the test. The concept and measurement of underachievement are discussed and it is shown that severe degrees of specific reading retardation occur at a rate above that predicted on a statistical basis, suggesting a “hump” at the lower end of the distribution. Reading retardation is shown to differ significantly from reading backwardness in terms of sex ratio, neurological disorder, pattern of neuro-developmental deficits and educational prognosis. It is concluded that the concept of specific reading retardation is valid, but evidence is discussed which shows that this does not support the view of a genetically distinct syndrome of dyslexia.