Data are presented on the prevalence of current left-handedness and prior left-handedness (switched) in 2787 subjects from 21 to 101 years of age. In addition, data on sex differences, familial sinistrality, hand posture when writing, and education were recorded. Two hypotheses were tested. The elimination hypothesis states that reduced frequency of left-handers in old age is due to reduced longevity. The modification hypothesis states that differences in the number of left-handers between older and younger persons are due to changing patterns of social norms. The results showed a decreasing prevalence of left-handedness across the age span, with 15.22% in the youngest group (21-30 years), but only 1.67% in subjects older than 80 years. There was however a corresponding increase in the number of subjects who had switched hand for writing, 2.69% in the youngest group to 6.75% in subjects 80 years and above. This supports the modification hypothesis and questions the elimination hypothesis. However, the mean percent score was still lower in subjects above compared to below age 40 after correction for hand switching. Thus, although changes in social norms towards left-handers seem to be the most likely explanation, we have not empirically disproved the elimination hypothesis.