From very limited data, it has been generalized that agriculturalized populations, regardless of climate do tend to have higher intelligence by about 13 IQ points compared to non-agricultural populations (LINK). The explanation given was that the population growth induced by agriculture produced greater genetic diversity, which would have included the development of some gene variants which give rise to greater intelligence. I would add that individuals with slightly higher intelligence would have proven fitter to survive and reproduce in agricultural societies even in warm climates, and this is because the food surplus created by intensive agriculture can allow for economic niches which more intelligent individuals can utilize to their reproductive advantage by obtaining higher social status than ordinary peasants. Agriculture also produces more easily tradeable essential goods (grain) than hunting and gathering, and this produces a greater incentivize for cooperation. Cooperation must be maintained by laws, which often criminalize behaviors resorted to by less intelligent individuals (theft, murder, etc.), removing less intelligent individuals from the gene pool.
However, one common argument in favor of the hypothesis that agriculturalization increases intelligence may be incorrect. The observation that arctic Native Americans (genetically similar to East Asians) have significantly lower intelligence than agriculturalized East Asians fails to account for a possible non-genetic negative influence on IQ in arctic populations. The mercury pollution of fish and whale stocks can decrease the measured IQ of extant hunter-gatherer populations significantly; this includes arctic Native Americans (LINK). The influence of mercury needs to be controlled for in order to gain an accurate understanding of arctic peoples’ intelligence. One should also account for the reported intelligence of Mongolians, a historically nomadic pastoralist population, which according to Lynn and Vanhanen has an average IQ of 101, greater than that of arctic Native Americans, but less than that of the Chinese (105), Japanese (105), and Koreans (106). This Mongolian average IQ is also greater than that of the long agriculturalized population of Vietnam (94). This probably indicates that pastoralism has some of the eugenic effects associated with agriculture and that both agriculturalization, as well as climate, significantly influence the selection for intelligence.
Variations in climate likely affect several other characteristics besides intelligence. Cold weather can reduce child mortality by killing pathogens and pathogen vectors which cause illness disproportionately in children, removing a selective pressure which would be compensated for with higher fertility and earlier maturation. If a high birth rate is less necessary and maturation can be delayed, there can be greater energy investment in each individual child, which can allow for the development of greater intelligence. Higher intelligence in humans; as well as lower fertility, longevity, and delayed maturation; does appear to co-occur in human populations which have evolved in colder climates (Rushton, 1985) (Lynn and Vanhanen) (Rushton, 2004). There is some recent research indicating that the negative relationship between IQ and fertility exists even when controlling for other “potential mediator variables” (Boutwell, 2013). As usually implied by proponents of the cold-winters theory, living in northern latitudes in pre-modern times may also have its own reasons for positively selecting for intelligence as it would have been needed to plan ahead to cope with the limited food availability induced by cold winter weather.
The variation in life history strategies induced by climate cannot be attributed to the existence of greater “r selection” in tropical latitudes than in colder climates at more polar latitudes because cold weather itself is a seasonal, density-independent, and therefore “r-selective” pressure on northern populations (Anderson, 1991). The infectious diseases which plague the tropics are also more density-dependent and therefore more “K-selective” than seasonal cold weather, making the selective pressures in the tropics more “K-selective” than those in northern latitudes (Anderson, 1991). (r/K selection theory categorizes selective pressures by the degree to which these pressures depend on population density to exist (Pianka, 1970)). r/K selection theory on its own it does not appear to be the best way to understand the origins of life-history strategies, compared to demographic models examining the relationship between age-specific mortality and life-history (Reznick et al., 2002). J.P. Rushton noticed important data trends, but attributing of these trends to r/K selection was flawed.
There is some evidence, from measured reaction times, that intelligence has decreased by about 12 to 14 points among Europeans since the late 19th century (Woodley et al., 2013). The authors suggest that this could be due to a decline in infectious disease which previously limited the reproduction of the underclasses. However, one might also suspect that the underclasses could have eventually adapted to this selective pressure; perhaps this occurred, but it was only after the underclass population had been decreased significantly by disease, making it difficult for the underclasses to automatically rebound to their previous demographic percentage.