I wonder if the distribution of intelligence is really is normal -- gaussian. In many cases people assume a normal distribution because the mathematics are well understood for that distribution -- not because data (or their measurement errors) are normally distributed.