Boys and girls within a family can be treated very differently.

It's not news that *overall* males tend to be closer to the extremes in all sorts of tests. If there's a genetic basis, it would presumably be due to the fact that unusual recessive genes on the X chromosome will usually be suppressed by dominant genes on the other X chromosome in a woman, and won't be expressed. A man has only one X chromosome, so unusual recessive genes are more likely to be manifested. This is why men are more likely than women to be colour-blind; and why haemophilia is almost exclusively found in men. There are some disorders that specifically affect the brain and cause low IQ, such as fragile-X syndrome, which are caused by an abnormal gene on the X-chromosome, and which are much commoner in males than in females. In theory, it's also possible that unusual genes on the X-chromosome could cause outstandingly *high* ability, and this would be commoner in males. However, there's no real evidence for this, and it seems a bit unlikely, as variations in ability in the normal and gifted range seem to be due to multiple genes interacting with environmental factors, rather than to any one single gene.

Two important reasons for males being more 'extreme' than females are likely to be that (1) it's more acceptable for males to stand out from the crowd, whereas females are more likely to wish to conform and not appear too noticeable; (2) men are more likely to take risks, which, in the case of tests, tends to mean that they are more likely to come up with a brilliant idea that raises their score *and* more likely to fail completely, whereas women tend to play safer.

As regards specific abilities, it is generally found that females are a bit better verbally and males are a bit better spatially, though the differences are not huge. There are very conflicting findings about gender differences in arithmetic, with lots of variations over time and place, which suggests a strong influence of social factors. Gender differences in arithmetic and other aspects of maths are much smaller now than 20 or 30 years ago. Elizabeth Spelke published an interesting article about this in 2005.