Phylogenetic bias also complicates the use of comparative data for inferring adaptive associations between brain size and other traits. Strepsir-hines, for example, not only have smaller brains than haplorhines, they also tend to have lower metabolic rates (Martin, 1996), shorter gestations, a different type of placentation and smaller neonates (Martin, 1990), and tend to be more nocturnal and live in smaller groups (Smuts et al., 1987; Kappeler and Heymann, 1996). Interspecific analysis of any of these variables with brain size would risk finding a spurious correlation as a result of the overall grade differences between the two suborders. The word 'spurious' is used here to mean a correlation that does not reflect a general adaptive association. In order to infer that two traits have such an association, it is necessary (though not sufficient) to show that they exhibit correlated evolution: that is, they can be shown to have covaried in a consistent way across multiple evolutionary events. As an example, an interspecific analysis of relative brain size and activity timing (diurnal versus nocturnal habits) in primates shows a statistically significant difference: with the effects of body weight partialled out, diurnal species have significantly larger brains (t = 4.3, df = 115, p < 0.0001). There should not, however, really be 115 degrees of freedom in this analysis, because there have only been a few evolutionary changes in activity timing against which to match changes in brain size. The result could largely reflect the fact that strepsirhines are small brained and predominantly nocturnal, whereas haplorhines are relatively large brained and predominantly diurnal. Indeed, using the BRUNCH procedure for categorical variables in Purvis' implementation of the independent contrasts method (see Chapter 3), there is no significant trend for activity timing and relative brain size to evolve together (t = 1.38, df = 4, p = 0.24). This remains true regardless of the allometric exponent used to generate residual brain size values. Other ecological variables do, however, correlate with overall brain size. A multiple regression analysis on 68 independent contrasts shows that, when body size is taken into account, brain size shows independent positive correlations with the percentage of fruit in the diet and social group size (Table 7.1). This confirms the well-known correlation in primates between diet and relative brain size (Clutton-Brock and Harvey, 1980; Foley and Lee, 1992). The size of one particular brain structure, the neocortex, is also correlated with both percentage frugivory and social group size (Barton, 1996). The neocortex size analysis controlled for the size of the rest of the brain, thus examining relative neocortical expansion rather than size relative to body size. Hence, the ecological correlates of relative neocortical expansion and of overall brain size are similar. This might be because brain size differences are, at least partly, a consequence of selection operating specifically on neocortical processing systems, a possibility explored further below.
Ecological correlates of brain and neocortex size have been interpreted as evidence for the idea that specific aspects of lifestyles have selected for specific cognitive abilities. The correlation with frugivory led to the suggestion that large brains are needed to memorise and integrate information on the location of fruit trees distributed patchily in large home ranges (Clut-ton-Brock and Harvey, 1980; Milton, 1988; see Eisenberg and Wilson (1978) for a similar explanation of large brains in frugivorous bats). Correlations between neocortex size and social group size have been taken as evidence for selection on social intelligence (Dunbar, 1992; Sawaguchi, 1992; Barton and Dunbar, 1997). Whereas diet and social group size have sometimes been seen as alternative explanations for the evolution of brain size, Table 7.1 suggests that both are independently important. Some authors, however, are more sceptical that brain size can be related to the specific information-processing demands of different ecological niches. Their ideas are explored in the next section.
Was this article helpful?