Hand preference and hand skill were examined in relation to sex and age from 3 1/2 to 50+ years. The data were drawn from several samples which avoided volunteer bias, either by selecting children according to birth-date or by taking complete class groups. The distribution of hand preference and of asymmetry of skill gave no evidence of systematic changes with age. Younger children were more variable and showed larger absolute differences in favour of the right hand than older children but L-R times were similar over most of the age range. Time for peg moving decreased with increasing age to a minimum in the late teens, remained stable for the next three decades and increased slightly in the 50+ age group. Males were faster than females with the left hand in almost all age groups. Females tended to be faster than males with the right hand up to 10 years of age but males then equalled and surpassed females, to be faster in most older groups. Correlations between hands did not vary systematically with age or sex. Left-handers were faster than right-handers and right mixed-handers were intermediate for peg-moving time by the non-preferred hand, in both sexes and both sets of samples. Differences for the preferred hand were less clear but still favoured left-handers in several comparisons. The findings raise the possibility that left hemisphere specialization for language is achieved through a right hemisphere handicap and that left-handers escape this risk to cerebral efficiency.

The lateral preferences and L-R skill of 109 male and 20 female dyslexics were as expected if the distribution of lateral asymmetry is shifted less far to the right in dyslexics than in controls. Several aspects of the data were consistent with Annett's hypothesis that some dyslexics lack the left hemisphere speech-organising factor postulated by the right shift theory of handedness and that this would be sufficient to account for the proportion of affected relatives. Some dyslexics were strongly dextral and these differed from the less dextral cases in several ways which resembled the distinction between backward and retarded readers.

The distribution of differences between the hands (L-R) in skill, as measured by a peg-moving task, was examined for several samples in which volunteer bias was absent or minimal. After comparing the main samples for hand preference and L-R times, they were combined to give 617 males and 863 females, aged 12-63 years. There was also a smaller sample of 122 males and 156 females, aged 6-15 years. The L-R distributions were negatively skewed and leptokurtotic. They were not compatible with the sum of two normal distributions, one of right-handers and one of left-handers. They were compatible with the sum of two or three normal distributions when one was unbiased to either side and the others were shifted to the right, as expected for the dominant or additive versions of the single gene interpretation of the right shift theory of handedness (Annett, 1978b, 1979). Sex differences in L-R scores confirmed that females tend to be more dextral than males. This was true of right-handers but not of left-handers, as expected if the sex difference is due to stronger expression of the rs+ gene in females than males and if this gene is absent in the majority of left-handers. There were trends suggesting that L-R asymmetry differs with educational status, undergraduates being less shifted toward dextrality than the general population.