I bet the score dispersion within the racial groups are larger than the difference in means between groups. Also your stats are not controlled for education, social conditions, health etc.

Is it a surprise that bushmen score the lowest when they're also the least plugged into the modern world?

I'd say its practically impossible construct an experimental design to fully normalise for education, environment, health, etc given that the invididual experience is so varied. Apart from that how could any research design factor for individual traits like determination or sloth?

on the observed stereotypes e.g. indians = doctors, rubber tappers, paraquat suicides OR chinese = accountants, businessmen, ahlong victims: a lot of these arise from social conditions where the people from the group share common ethnic experiences/ influences. Indians more likely to be laywers & doctors because these are prestigious professions for that ethnicity (social) and their parents are more likely to be doctors/lawyers (family).

there's no reason to suspect genetic traits.

Ifthere is a link from genes -> traits -> behaviour -> IQ, how strong is causal link between each link in the chain?

Even if this is true (& I have reservations over "Behaviour are strongly inherited bloodline-wise") how far can these genes be applied to summarise racial IQ? Bear in mind also that even those who strongly identify with a certain race may not genetically be of pure race anyway.