Encephalization should normally only occur if the benefits of enlarged brains exceed its costs. These benefits and costs will operate in specific lifestyles, which need to be specified in any evolutionary account (Johnston, 1982). The two major costs to encephalization appear to be developmental and metabolic (Bennett and Harvey, 1985b). All other factors being equal, bigger brains require a longer time to develop and are energetically more expensive to maintain. The metabolic cost of brains is particularly high in humans and other primates (Aiello and Wheeler, 1995; Aiello and Wells, 2002; Fish and Lockwood, 2003), but less clear in other taxa (e.g., bats; Jones and MacLarnon, 2004). It is important to note that both metabolism and development are related to body size and diet and that their relationship to brain size is thus likely to be complicated by these interactions. If slower development means fewer offspring per unit time and is an allometric correlate of large body size, then the relative importance of natural selection and behavioral flexibility as alternative mechanisms to track environmental change will be affected. If high metabolic rate is associated with the higher surface-to-volume mechanism of heat loss in small-bodied animals, then this might affect the amount of energy available for encephalization. If herbivores are on average characterized by precocial development, low metabolic rates due to low nutrient quality, large ranges required to collect large amounts of low-quality food and large body size due to selection for an enlarged digestive system as well as defense against predators, then all these factors are also likely to affect brain size.
The brain develops slowly and some researchers have proposed that it is the major developmental constraint on time to reproduction (Sacher, 1978). Apart from parental behavior, four life history traits will affect the number of descendants per unit time, which affects the selection probability of a mutation favoring adaptation to environmental change: length of the reproductive period (reproductive longevity, time to sexual maturity), number of offspring per reproductive event, time to sexual maturity of the offspring, and time intervals between successive reproductive events. An animal that lives for 25 years, takes 5 years to mature sexually, and has one offspring every 2 years will have far fewer descendants at the end of 100 years than an animal that has three offspring per yearly reproductive event, lives for 5 years and takes 1 year to mature. Changes due to selection will occur much more slowly in the first species because fewer generations per century mean both fewer mutations and less differential reproduction, thereby increasing the value of behavioral flexibility as a mechanism for change.
The major hypothesis for the explanation of ence-phalization is that bigger brains allow enhanced cognitive abilities, abilities useful in certain lifestyles. The problem then becomes the definition of enhanced cognitive abilities and the lifestyles they could be useful in. This hypothesis can be tested by specifying the lifestyles that could benefit from more complex cognition, operationalizing the complexity of cognition, then looking for a statistical association between lifestyle, brain size, and cognitive complexity. General cognitive complexity can be opposed to specialized cognitive skills associated with specific lifestyles in a restricted set of taxa. The spatial memory associated with food caching in corvids and parids or brood parasitism in cow-birds is thought to be one example of the latter, as is the acoustic memory associated with song repertoire size in oscines (DeVoogd et al., 1993). In these cases, specialized neural structures are studied and the cognitive variation is relatively easy to operationa-lize: memory for more cache locations or acoustic memory for more songs. The small size of the structures (HVC, RA, hippocampus) implies that they are unlikely to form the basis of encephalization. What we are looking for instead is variation in unspecia-lized cognition over potentially all species, based on the involvement of a large enough part of the brain that can address the issue of whole brain encephalization.
The difficulty is operationalizing unspecialized cognition in all species. Researchers usually look to abilities associated with complex cognition in humans. If we want to include as many taxa as possible in our tests, we have to look outside of tests that only a few nonhuman species can solve, such as learned sign language, episodic memory, fast-mapping, or understanding of the mental states of others (theory of mind). Associative learning is one obvious possibility. If we define the continuum of cognitive complexity as the latency of or the number of errors in learning, however, we face the problem of confounding variables and ecological validity. If a crow solves a learning test in the lab faster than does a kiwi, this might be because that the crow is tamer and less neophobic in the lab, the task favors visual rather than olfactory cues, or the task resembles situations crows encounter often in the field but kiwis do not. What we need is cognition that occurs spontaneously in the field, without the confounding variables of tests in captivity and in a situation that is natural to each animal. Tool use, play, and presumed deception and social learning are possible choices. We can also look to innovative behavior.
We then plot the taxonomic distribution of our cognitive measure from the field and see if, as predicted by encephalization theory, the distributions correlate positively with relative brain size, taking out as many confounding variables as possible. If the results of field and laboratory tests are positively correlated, this would support the assumption that they are both valid estimates of cognition. Research on birds and primates, based on re-analyses of published data as well as new tests (Webster and Lefebvre, 2001; Reader and MacDonald, 2003; Lefebvre et al., 2004), suggests that tests in captivity are indeed positively correlated with field measures. We can further test our assumption that unspecia-lized cognition exists by predicting positive correlations between the distribution of the different cognitive measures. All of these relationships have been tested in birds and primates (see below). Overall, the analyses conducted up to now on the two taxa suggest convergent evolution of interspecific variation in cognitive abilities (see also Emery and Clayton, 2004). Similar positive correlations between innovation rate, tool use rate, and reversal learning performance have been found in birds and primates, perhaps suggesting that these cognitive abilities are nonmodular (Lefebvre et al., 2004). The only exception for the moment seems to be the relationship between food storing and innovation in 22 species of birds; a negative relationship is found in New World corvids and Old World parids, suggesting a (possibly modular; Lefebvre and Bolhuis, 2003) trade-off between storing and innovativeness.
It is important to note that correlations between brain size and cognitive variables do not demonstrate the survival value of having a large brain, nor are they evidence for natural selection on enlarged brains. The correlations suggest that large brains are on average present in tool using, innovative, playful, social taxa that develop slowly, but they also suggest that small-brained taxa can do well, provided their lifestyles do not include these attributes. A survey of long-term population trends (1968-95) in 40 British bird species provides evidence for selection on large over small brains, with larger declines observed in small-brained species than in large-brained ones (Schultz et al., 2005). Sol et al. (2002, 2005) examined colonization success of introduced birds in different parts of the world; some species succeed almost everywhere (e.g., sparrows and blackbirds), while others are extinct after only a few years. Relative brain size (and innovation rate in the zone of origin) significantly predicts variance in colonization success. Contrary to natural invasions, where unsuccessful cases are seldom documented, introductions allow good coverage of the entire spectrum of responses. If introductions are an unbiased estimate of all invasions, then establishment in new areas might be one of the key selective forces that affect encephalization trends (Sol et al., 2005) and the allopatric divergence that often follows invasion, a key mechanism in the association between speciosity and relative brain size in birds (Nicolakakis et al., 2003).
To express ecological theory in a very simplified way, the distribution of abiotic factors drives the distribution of vegetation, which in turn drives the distribution of animals. Lifestyles (diet, sociality, and sexual selection) are then driven by the distribution of animals and plants. If biotic and abiotic resources are spatially and temporally predictable and in relatively low-density clumps, a specialized, conservative, territorial polygynist with monoparental care may do better than a generalist, opportunistic, invasive, gregarious monogamist with biparental care. The reverse would apply to spatially and temporally unpredictable resources found in abundant patches. We would then expect selection to act on cognition to provide the information-processing capacity that best suits each lifestyle, with accompanying selection on encephalization (Bennett and Harvey, 1985a). Testing the idea that omnivory should be associated with brain size is thus not an ecological prediction on cognition, but a dietary prediction on encephalization with two missing links: how does resource distribution favor omnivory and how does omnivory require more complex cognition or lift dietary constraints on brain size? The use of an ecological framework is all the more important because the same resource distribution may lead to similar predictions on lifestyle differences that are sometimes viewed as independent pressures for complex cognition and encephalization. For example, social and diet breadth pressures on the evolution of cognition and brains are often seen as alternatives (see Forebrain Size and Social Intelligence in Birds). If, however, one type of resource distribution favors gregarious generalists and another favors territorial specialists, then the two pressures go in the same direction. Whether or not the lifestyle differences are independent is a matter of empirical test (with multivariate techniques, for instance), not a logical a priori.
Diet (Eisenberg and Wilson, 1978), sociality (Dunbar, 1992; Dunbar and Bever, 1998), sexual selection (Madden, 2001), and parental care (Gittleman, 1994) have all been shown to be associated with encephalization. In some of these tests, we do not know to what extent the apparent co-evolution of the traits is due to common ancestry or repeated independent events, given that independent contrasts have not been conducted. Larger brains have been found in omnivorous and frugivorous groups (Allman et al., 1993) compared to folivorous or herbivorous animals. For sexual selection, there is interspecific evidence for an association between brain size and bower building (Madden, 2001) and intraspecific evidence for an association between tele-ncephalon size and song repertoire size in zebra finches (Airey and DeVoogd, 2000). Monoparental versus biparental care, which is a consequence of sexually selected mating systems, has also been implicated in brain size differences (Gittleman, 1994).
Was this article helpful?