Although gains in generational intelligence test scores have been widely demonstrated around the world, researchers still do not know what has caused them. The cognitive stimulation and nutritional hypotheses summarize the several diverse potential causes that have been considered. This article analyses data for a sample of 499 children tested in 1930 and one equivalent sample of 710 children tested 72 years later, the largest gap ever considered. Both samples comprised children aged between 7 and 11 who were assessed by the Draw-a-Man test in the city of Belo Horizonte, Brazil. Further, one additional sample of 132 children was assessed in 2004 in a rural area very similar in several diverse factors to the 1930 urban sample. The results are consistent with both the cognitive stimulation and the nutritional hypotheses. © 2006 Cambridge University Press.