The hypothesis that polymorphisms in the gene for nicastrin (NCSTN) are associated with differences in cognitive level and ageing was tested in 462 relatively healthy surviving participants of the Scottish Mental Survey 1932. None had a history of dementia. They were tested on the Moray House Test of verbal reasoning at age 11 in 1932 and at age 79 between 1999 and 2001. At age 79 they also took tests of non-verbal reasoning, short- and long-term verbal declarative memory, Verbal Fluency, and a short screening test for dementia. Subjects who possessed at least one copy of the NCSTN B haplotype (Hap B) had higher scores on the Moray House Test (a test principally of verbal reasoning) at age 11 (p=0.036) and age 79 (p=0.027). The effect of Hap B on cognition at age 79 was non-significant after adjusting for the effect at age 11. Therefore, the effect of Hap B in this sample is on the life-long stable trait of cognitive ability, and not on age-related cognitive change. The possibility that this result might be a selection effect was not supported by the samples being in Hardy-Weinberg equilibrium with respect to the distribution of NCSTN genotypes.