life history costs and benefits of encephalization: a comparative test using data from long-term...

Life history costs and benefits of encephalization: a comparativetest using data from long-term studies of primates in the wild

Nancy L. Barrickman a,*, Meredith L. Bastian a, Karin Isler b, Carel P. van Schaik b

a Department of Biological Anthropology and Anatomy, Duke University, Box 3170, Durham, NC 27710, USAb Anthropologisches Institut und Museum, Universitat Zurich, Winterthurerstrasse 190, CH-8057 Zurich, Switzerland

Received 7 December 2005; accepted 20 August 2007

Abstract

The correlation between brain size and life history has been investigated in many previous studies, and several viable explanations have beenproposed. However, the results of these studies are often at odds, causing uncertainties about whether these two character complexes underwentcorrelated evolution. These disparities could arise from the mixture of wild and captive values in the datasets, potentially obscuring real relation-ships, and from differences in the methods of controlling for phylogenetic non independence of species values. This paper seeks to resolve thesedifficulties by (1) proposing an overarching hypothesis that encompasses many of the previously proposed hypotheses, and (2) testing the predic-tions of this hypothesis using rigorously compiled data and utilizing multiple methods of analysis. We hypothesize that the adaptive benefit ofincreased encephalization is an increase in reproductive lifespan or efficiency, which must be sufficient to outweigh the costs due to growingand maturing the larger brain. These costs and benefits are directly reflected in the length of life history stages. We tested this hypothesis ona wide range of primate species. Our results demonstrate that encephalization is significantly correlated with prolongation of all stages of devel-opmental life history except the lactational period, and is significantly correlated with an extension of the reproductive lifespan. These results sup-port the contention that the link between brain size and life history is caused by a balance between the costs of growing a brain and the survivalbenefits the brain provides. Thus, our results suggest that the evolution of prolonged life history during human evolution is caused by increasedencephalization.! 2007 Elsevier Ltd. All rights reserved.

Keywords: Brain size; Brain growth; Comparative analysis; Primate evolution; Human evolution

Introduction

The proposition that brain size is related to life history vari-ables, such as fecundity, age at maturity, and lifespan, is notnovel. Ever since Sacher (1959) first pointed to an interspecificcorrelation between brain size and life history variables ina sample of mammals, numerous studies have revisited thissubject, especially for primates (Sacher and Staffeldt, 1974;Sacher, 1975, 1978; Harvey and Clutton-Brock, 1985; Harveyet al., 1987; Austad and Fischer, 1992; Allman et al., 1993;

Allman, 1995; Hakeem et al., 1996; Allman and Hasenstaub,1999; Barton, 1999; Ross and Jones, 1999; Judge and Carey,2000; Deaner et al., 2003; Ross, 2003, 2004; Leigh, 2004).The existence of this evolutionary relationship has substantialimplications for the study of human evolution. First, life his-tory is crucial to determining the duration of parental care,thus affecting the developmental trajectory of the offspringand the reproductive strategy of the parents and their socialgroup. It also affects population dynamics, which can haveconsiderable ramifications for social behaviors such as groupsize, male-male competition, alliance formation, escalationavoidance, and infanticide (e.g., Smuts and Smuts, 1993; vanSchaik and Kappeler, 1997; van Schaik et al., 1999, 2006;Janson, 2003; Pereira and Leigh, 2003). Second, brain size islinked to cognitive and cultural abilities across many taxa

* Corresponding author.E-mail addresses: [email protected] (N.L. Barrickman),

[email protected] (M.L. Bastian), [email protected] (K. Isler),[email protected] (C.P. van Schaik).

0047-2484/$ - see front matter ! 2007 Elsevier Ltd. All rights reserved.doi:10.1016/j.jhevol.2007.08.012

Available online at www.sciencedirect.com

Journal of Human Evolution 54 (2008) 568e590

(e.g., Lefebvre et al., 2004; Deaner et al., 2007). Given thegreat brain size expansion during the evolutionary history ofhominins, understanding the brain size/life history relationshipis essential to understanding some fundamental humanadaptations.

Several hypotheses have been proposed to explain the linkbetween brain size and life history (for reviews see Leigh,2001; Deaner et al., 2003; van Schaik et al., 2006). However,the various comparative studies used to substantiate the linkhave often found conflicting results. Table 1 presents the find-ings of the recent studies of brain size/life history relationshipsin primates. Although some studies show that developmentalperiods, such as age at first reproduction, are correlated withbrain size, others found that only lifespan shows a correlationwith brain size. For instance, Ross and Jones (1999) founda significant relationship between age at first reproductionand brain size, whereas Barton (1999) did not. Conversely,Ross and Jones (1999) and Barton (1999) found no significant

relationship between brain size and lifespan, whereas Judgeand Carey (2000) and Deaner et al. (2003) did. It is currentlyunclear which life history stages, if any, are associated withbrain size in primates. Before further investigation into themechanisms that maintain this link can proceed, it must be de-termined whether this evolutionary correlation is real or an ar-tifact of data and/or methods of analysis.

These studies, and by extension the proposed explanations,have been hindered by three difficulties. First, the quality ofthe data used by previous studies is uncertain. The compara-tive data sets have been compiled over time by several re-searchers, and used in amended and somewhat varyingforms by the various studies. These sets often contain an as-sortment of values from captive and wild populations. Becausemany life history variables show considerable phenotypic plas-ticity (Lee, 1999; Lee and Kappeler, 2003), such mixed datamight create enough noise to obscure true relationships. Un-tangling the discrepancies between the results requires a close

Table 1Results from studies of the correlations between brain size and life history variables in primates

Adult brain size studies

Life history variable PNIa Statistical procedureb r2 p-value nc Reference

Gestation Yes (all) RR 0.04 0.16 44 Deaner et al., 2003Yes (old) MR 0.25 18 Barton, 1999No RR 0.02 0.41 34 Allman et al., 1993

Lactation Yes (highest correlation) MR 0.88 25 Ross, 2003

Juvenility Yes (highest correlation) MR 0.01 25 Ross, 2003Yes (highest correlation) MR 0.59 0.01 23 Ross and Jones, 1999

Age at first reproduction Yes (all) RR 0.04 0.18 45 Deaner et al., 2003No MR 0.74

examination of the data used, because, as Martin pointed out,in the rush to conduct analyses, insufficient attention is givento data quality, and a lot will be gained through improvementin this area (2003, page xx). Second, the disparity of resultsfrom previous studies may arise from variation in the way theycontrolled for phylogenetic non independence. Correlationsbetween variables in comparative analyses may be spuriousbecause the particular combinations of characters are the resultof inheritance from a common ancestor rather than the out-come of a shared adaptive function (Felsenstein, 1985). Ana-lyzing species values treats them as if they are independentdata points and not potentially influenced by evolutionary his-tory. The method of independent contrasts is the most com-monly used method for controlling for phylogenetic bias incontinuous variables (Harvey and Pagel, 1991; Purvis andWebster, 1999). Not all of the studies listed in Table 1 performthese analyses, and the details of the procedure can differamong those that do. Some studies (e.g., Barton, 1999; Rossand Jones, 1999; Deaner et al., 2003) emphasize the use ofonly old contrasts, which excludes the contrasted values be-tween individual species from the analysis. This procedure isfollowed because a larger number of contrasts at the tips ofthe phylogenetic tree can have too great an influence andbias the results (Purvis, 1995). Also, species values are moresensitive to measurement and/or sampling error than averagesof higher taxonomic levels (Purvis and Rambaut, 1995; Purvisand Webster, 1999). Barton (1999) performed analyses on onlyold contrasts, whereas Deaner et al. (2003) report the results ofanalyses performed on all contrasts and old contrasts. Rossand Jones (1999) selected the contrasts that resulted in themost significant correlation. In most cases exclusion of theyounger contrasts increased the strength of the correlationsfound considerably, and the results of these analyses are re-ported. In a few cases, in which exclusion of younger contrastsdid not increase correlations, the results from the whole dataset are given (Ross and Jones, 1999:91). Such differencesin the methods for removing phylogenetic bias may thusalso explain some of the disparities between previous studies.

Finally, many of the hypotheses proposed to explain the re-lationship between brain size and life history do not effectivelyintegrate all of the phases of life history; they make predic-tions about certain phases of life history and are silent aboutothers. For instance, the cognitive buffer hypothesis arguesthat greater degrees of encephalization provide greater defenseagainst mortality, through such benefits as increased behav-ioral flexibility and problem-solving capabilities (Allmanet al., 1993; Hakeem et al., 1996; Allman and Hasenstaub,1999). This hypothesis emphasizes the adult phase of life his-tory. Conversely, the skill-learning hypothesis argues thatphases of growth and development are extended in specieswith larger brains because sufficient time is needed to acquirethe cognitively complex strategies utilized by the species(Ross and Jones, 1999). Both of these types of hypothe-sesdthose that emphasize the growth periods and those thatemphasize the reproductive perioddcan potentially explainthe variation in life histories. However, it is currently unclear,given the problems with the quality of data and differing

methods of analysis, which phases of life history are moststrongly correlated with brain size. Establishing this relation-ship will greatly improve subsequent research questions thatattempt to explain why these two character complexes are as-sociated, and whether a direct link can be substantiated.

This paper seeks to resolve the difficulties of previous stud-ies in two ways. First, we will propose a model for a direct re-lationship between brain size and life history. Second, we willtest the proposed model using primarily data from long-termstudies of wild populations available from the primate litera-ture. The objectives of this study will allow for further testingof the selective mechanisms that have been proposed to ex-plain the evolutionary relationship of brain size and life his-tory. There may be many pathways, or modes as Pereira andLeigh (2003) have proposed, by which these relationshipsare maintained, and this study provides a general frameworkfor testing these varying modes.

Life history costs and benefits of encephalization

Many life history theorists have argued that there is a bal-ance between the phases of life history (e.g., length of growthperiod relative to adult lifespan) across mammals (Kozlowskiand Wiegert, 1986; Stearns, 1992; Charnov, 1993; Charnovand Berrigan, 1993). This balance has been attributed to theoptimization of growth and reproductive effort in the face ofextrinsic mortality: higher mortality rates are offset by rapidattainment of maturity and high reproductive rates, resultingin a balance between the phases of life history. This scenariois supported by the consistent proportional relationships, or in-variants, between the length of pre-reproductive stages andlifespan across mammalian taxa (Purvis and Harvey, 1995),and more specifically, across primates (Hawkes et al., 1998;Alvarez, 2000). The effects of brain size on life history are fre-quently discounted (Charnov and Berrigan, 1993), althoughmore recent work (Charnov, 2004) does tentatively acceptthe possibility. However, if brain size affects the optimizationof growth and reproductive effort, then it should be linked tolife history and must correlate with all stages of life history,albeit not necessarily to the same degree with each. Our pointof departure is the juxtaposition of developmental costs andadult benefits of encephalization. The evolution of larger brainsize could entail delayed maturity to meet the needs of grow-ing a brain. Likewise, it could bring about certain benefits,such as decreased mortality or increased reproductive rateamong adults. Any organism could reap these benefits, pro-vided they outweigh the developmental costs. Correlationsof life history variables, such as age at maturity and lifespan,with brain size should reflect the balance between these costsand benefits.

The existence of these life history invariants has recentlybeen questioned because they may represent a statistical arti-fact (Nee et al., 2005) of regressing a proportion of a variableagainst itself. In this type of analysis, null models generatedfrom randomly distributed data can result in consistent rela-tionships (slopes and r2 values near to one). However, thenull models cannot account for the differing values of

570 N.L. Barrickman et al. / Journal of Human Evolution 54 (2008) 568e590

Christina Vo

invariants in differing taxonomic groups [e.g., fish havea higher slope than mammals (Savage et al., 2006)]. Also,the life history invariants generated from empirical data haveunimodal distributions and constrained ranges versus theeven distribution of invariants generated from the randomdata in the null models (Savage et al., 2006). Thus, the consis-tency across taxa of ratios between life history variables hasbeen demonstrated empirically. In addition, the present studydoes not generate dimensionless ratios (i.e., test for correla-tions between a variable that is a proportion of the original var-iable). Instead, this study compares mean values across taxa,which are not affected by this possible artifact (Nee et al.,2005).

The pre-reproductive stages of life history may be related tothe developmental costs of brain size. Brains are energeticallyexpensive to grow and maintain (Holliday, 1986; Leonard andRobertson, 1992, 1994; Aiello and Wheeler, 1995). In additionto the energetic costs, larger brains may take a longer timethan smaller brains to reach structural and functional maturityat a cellular level, and consequently at a behavioral level aswell, even after they have reached full volume (Gibson, 1970,1991; Rakic et al., 1986; Missler et al., 1993; Giedd et al.,1996, 1999; Bourgeois, 1997; Huttenlocher and Dabholkar,1997; Paus et al., 1999; Kwon et al., 2002; Gogtay et al.,2004; Casey et al., 2005). Consequently, the pace of braingrowth and development may constrain the growth rate of theentire soma, as originally suggested by Sacher and Staffeldt(1974). Martin (1996) extended this notion by proposing thematernal energy hypothesis, in which the size of the brain isdetermined by the metabolic turnover of the mother duringgestation and lactation (see also Martin, 1981, 1983). Thus,the energetic cost of encephalization is shouldered by themother, and her reproductive rate can be adversely affectedby a greater degree of encephalization in her offspring. Also,this developmental cost results in a trade-off for the offspring.The immature individual must balance the tradeoffs betweenstarting reproduction before it is fully competent, or dyingwhile waiting to reach reproductive maturity (Stearns, 1992;Harvey and Purvis, 1999; Stearns et al., 2000).

All these processes are expected to lead to a longer durationof immaturity in larger-brained organisms. In order for largebrain sizes to be favored by selection, the net reproductive fit-ness must increase, and thus the delay in maturation time mustbe more than compensated. From a demographic perspective,an increase can be accomplished in two ways: (1) extend thereproductive period (i.e., prolong adult lifespan), or (2) in-crease yearly reproductive output (i.e., grams of offspringweight per year, as measured by neonatal body size/interbirthinterval). Previous studies of the relationship between lifespanand brain size have produced ambiguous results (see Table 1).However, none of these studies compared maximum adult lifespan (maximum lifespan minus age at first reproduction) torelative brain size. Because the fitness benefits of large brainsize are not realized until the organism begins reproduction,the growth period must first be subtracted from total lifespanin order to test the effect on reproductive potential. Anothermeasure of the extension of the adult reproductive period is

life expectancy after reaching adulthood or the inverse of adultmortality rate (Harvey and Zammuto, 1985; Hawkes et al.,1998; Alvarez, 2000; Kaplan et al., 2000). There are no studiesthat compare life expectancy and brain size across primates.

There are very few studies of the relationship between brainsize and yearly reproductive output. Harvey and Clutton-Brock (1985) demonstrated a positive relationship betweenadult brain size and interbirth interval in primates (r 0.86),but they did not control for body size. No other study hasfound either a positive or negative relationship among pri-mates. It is well-documented that humans have a reducedinterbirth interval (e.g., Bogin, 1997, 1999; Hawkes et al.,1998; Kaplan et al., 2000; Kennedy, 2005), and this could bedue to a high degree of encephalization. For instance, Kennedy(2005) argues that large brain size requires high-quality, pro-tein-rich diets in early development, which is provided by so-phisticated foraging skills and large social networks among theadult caregivers. These supplemental foods and extra-kin sup-port allow for earlier weaning, and thus a reduced interbirthinterval. In addition, these behavioral complexes are enhancedand elaborated by larger brains in adults, creating a ratchetingeffect between the costs and benefits of brain size (Kaplanet al., 2000). However, this relationship between short inter-birth intervals and large degrees of encephalization has notbeen documented across primates.

Kaplan et al. (2000) also proposed a similar cost/benefit hy-pothesis, developed specifically to explain the extension of hu-man life history since the divergence from the last commonancestor with Pan. They argued that this extension resultedfrom complex food procurement strategies that require longperiods of growth and development to acquire the necessaryskills. In turn, this investment is compensated by (1) surplusfood output as an adult, which is bestowed on immatures,and (2) mechanisms for reduced mortality that are either inte-gral to the feeding ecology (social networks, high quality re-sources) or acquired during growth and development(strengthened immune function). These two compensatoryfeatures result in an overall increase in reproductive success,and their results demonstrate that the pattern holds for humanand chimpanzee populations. Their embodied capital hy-pothesis was expanded to include brain size as one of the in-vestments made during growth and development, whicheventually paid off with decreased mortality and thus greaternet reproductive success (Kaplan et al., 2003; see also Robsonand Kaplan, 2003). Their results demonstrate a strong relation-ship between both age of first reproduction and lifespan andbrain size in a large sample of primates (see Table 1), but theiranalyses did not consider potential phylogenetic bias and theextent to which primary field data were used is uncertain.Their hypothesis would be strengthened by further confirma-tion of their result. In addition, analyses that exclude humansfrom the sample would demonstrate that the relationship isa pattern across primates rather than an exceptional strategyamong humans.

If an organism invests the time and energy into attaininga larger brain, it must gain some reproductive benefit to offsetthese costs. Therefore, we hypothesize that larger brain size,

571N.L. Barrickman et al. / Journal of Human Evolution 54 (2008) 568e590

Christina Vo

Christina Vo

Christina Vo

Christina Vo

Christina Vo

holding body size constant, is correlated with longer develop-mental periods and longer reproductive lifespan and/or greaterreproductive rate, reflecting, respectively, the costs and bene-fits of evolving a larger brain. This hypothesis leads to severalexplicit predictions that will be tested in primates. With re-spect to the developmental costs of encephalization, we pre-dict that pregnancy is longer for species with larger brains,and mothers give birth to large-brained neonates. Previousstudies of mammals supported this prediction (Sacher andStaffeldt, 1974; Pagel and Harvey, 1988a; Deaner et al.,2003; Jones and MacLarnon, 2004) but studies that examineonly primates did not show a significant relationship (see Ta-ble 1 for results and citations). Second, we predict that the pe-riod of postnatal growth and development is longer in primatespecies with greater encephalization. Studies examining thisprediction have yielded conflicting results (see Table 1 for re-sults and citations). Lactation length has never been shown tohave a significant relationship with brain size, but the period ofjuvenility/adolescence was found to be significantly correlatedwith brain size. Tests for the correlation between the entirepostnatal developmental period, as measured by age at first

reproduction, and brain size have thus far produced contradic-tory results (Table 1). Finally, we predict that larger brain sizeis correlated with extended reproductive lifespan in primates,allowing for a net increase in reproductive output. Once again,previous studies have produced conflicting results so the exis-tence of this relationship remains uncertain (see Table 1 for re-sults and citations). If there is no positive impact of largerbrain size on the length of the adult lifespan, it is predictedthat the rate of offspring production will be higher.

Materials and methods

Variables

The raw life history variables, and their references, are pro-vided in Table 2. All variables were log-transformed prior toanalysis. The values used in this study were compiled fromlong-term studies of wild, unprovisioned populations. How-ever, gestation length (GL) was sometimes taken from captivepopulations, as indicated in Table 2. This variable is not easilycollected in the wild since the timing of conception is often

Table 2Life history data used for analyses, all data are from wild populations except those values that are italicizeda,b

Species GLc IBI Ld J/A AFR TI Le ALe

Microcebus murinus 2 12 3d 9 12 14 16.9Eulemur fulvus rufus 4 27.4 23.4 17.4 40.8 44.8 33.6Lemur catta 4.4 14.2 9.8 26.2 36 40.4 27Propithecus verreauxi verreauxi 4.6 20.5 15.9 56.1 72 76.6 8.1 21.6Propithecus diadema 5.9 21.6 15.7 48.3 64 69.9 9.1 nae

Leontopithecus rosalia 4.2 10.2 6 37.2 43.2 47.4 26.4Cebus capucinus 5.3 27.5 22.2 37.8 60 65.3 49.8Cebus apella nigritus 5 19.3 14.3 65.8 80.1 85.1 38.4Cebus olivaceus 6.3 26.8 20.5 63.5 84 90.3 nae

Lagothrix lagotricha 7.3 36.7 29.4 78.6 108 115.3 21Ateles geoffroyi 7.5 24.7 17.2 66.8 84 91.5 41Brachyteles arachnoides 7.2 35.6 28.4 81.2 109.6 116.8 20.9Alouatta palliate 6.1 22 15.9 31.1 48 54.1 21Alouatta seniculus 6.4 17.4 11 51 62 68.4 19.8Presbytis thomasi 6c 26.8 20.8 44 64.8 70.8 12.3 nae

Presbytis entellus 6.9 28.8 21.9 58.5 80.4 87.3 18.3Erythrocebus patas pyrrhonotus 5.4 12.8 7.4 28.6 36 41.4 20.9Chlorocebus aethiops 5.4 17.1 11.7 49 60.7 66.1 26.5Macaca fuscata yakui 5.7 26.9 21.2 52 73.2 78.9 26.9Macaca fascicularis 5.5 20.1 14.6 48.1 62.7 68.2 16.1 32.8Macaca mulatta 5.5 18 12.5 47.5 60 65.5 31Papio cynocephalus 5.7 21.3 15.4 56.3 71.9 77.6 13.6 39Papio hamadryas 5.7 22 16.3 56.9 73.2 78.9 29.5Hylobates lar 6.9 41.2 34.3 85.7 120 126.9 30Pongo pygmaeus abelii 8 84 76 86 162 170 24.9 45.5Gorilla gorilla beringei 8.4 47 38.6 81.4 120 128.4 44Pan troglodytes 7.5 72.5 65 92.2 157.2 164.7 19 46.3Homo sapiens 8.9 38.4 29.5 204.5 234 242.9 41.3 85.5a All life history data are given in months (except AL and Le, given in years). References are provided in Appendix 1. GL gestation length; IBI interbirth

interval between surviving offspring; L lactational period, IBI-GL; J/A juvenile/adolescent period, AFR minus La; AFR age at first reproduction for fe-males; TI total immaturity, AFR plus GL, Le life expectancy at AFR, or the inverse of mortality rate; AL adult lifespan (maximum lifespan from captivitye AFR).b Other variables are affected by the captive values of gestation. Specifically, lactational period (IBI e GL) and total period of immaturity (AFRGL).c For Presbytis thomasi: estimated gestation length (GL), mean of brackets for related species, see Sterck (1999).d For Microcebus murinus: the value for lactation (L) is the period from birth until seasonal torpor.e For Propithecus diadema, Cebus olivaceus, Presbytis thomasi: reliable maximum lifespan is not available (na) for these species because of lack of individuals

maintained in captivity.


Christina Vo

unknown. Fortunately, its timing has been shown to be tightlyconserved across a range of conditions (Martin and MacLar-non, 1988). Using primarily wild data is potentially advanta-geous because conditions in the wild are most similar tothose conditions in which a species life history traits evolved.Also, avoiding the mixture of wild and captive values withina given species increases the chances that reliable relationshipsbetween variables will be detected. These stringent require-ments limited the size of the sample to 28 species.

Age at weaning is difficult to pinpoint in primates becauseit is a gradual process rather than a precise landmark. There-fore, lactational period (L) was calculated by subtractinggestation length from interbirth interval (after a survivinginfant). This value provides a measure of the time periodan offspring is able to gain exclusive maternal energeticresources before these resources are diverted to the next off-spring. The juvenile/adolescent period (JA) in females wascalculated by subtracting the lactational period from theage at first reproduction. This variable represents the time pe-riod between initial independence from maternal resourcesand reproductive maturity. Total immaturity (TI) in femaleswas calculated by adding gestation length plus age at first re-production. This variable represents the entire temporal costof development. The period of gestation must be includedin the total length of the pre-reproductive period because ges-tation can vary relative to the length of postnatal immaturityfrom species to species (Leigh and Park, 1998; van Schaiket al., 2006).

Reproductive lifespan is represented by life expectancy atage of first reproduction (Le) in females, which was calculatedfrom the life tables of studies of wild populations followingthe method detailed in Hill et al. (2001). Life table datawere only available for eight species. Analyses with a smallsample size runs the risk of committing Type I and Type II er-rors, so all of the regression models were bootstrapped (1,000iterations with replacement) to verify their p-values. Anotheranalysis was performed using maximum lifespan recordedfrom captive records. This analysis is separated from the anal-yses of wild data because these values may not represent thereproductive lifespan in the wild. Instead, they represent thephysiological limit of life-sustaining maintenance inherent ina given species. Age at first reproduction was subtractedfrom maximum lifespan in females in order to arrive at a mea-sure of maximum reproductive lifespan (AL), and the valuesare provided in Table 2. In addition, there are sufficient datapoints to conduct an analysis on the tradeoffs of reproductivelifespan and yearly reproductive output, as measured by neo-natal body mass divided by interbirth interval. In the presentstudy, the post-reproductive period in female humans is in-cluded in the analyses of reproductive lifespan. Many re-searchers have argued that the long post-reproductive periodis crucial to the reproductive success of the individual eventhough she is not giving birth to her own offspring. Thegrandmother hypothesis argues that these females contrib-ute to the success of their daughters offspring through paren-tal care and resource procurement (Hawkes et al., 1998). Thisstrategy offsets some of the costs of parental investment and

allows the daughter to shorten her interbirth interval. This phe-nomenon has been modeled mathematically for humans(Shanley and Kirkwood, 2001; Lee, 2003), demonstratingthat there is an increase in Darwinian fitness resulting fromthe combination of (1) eliminating the mortality risk that arisesfrom continuing to bear offspring and (2) the benefits of in-vesting in grandchildren. It is also evident in analyses of pri-mate patterns (Alvarez, 2000; Judge and Carey, 2000),showing that the relationship between the length of the pre-re-productive period and adult lifespan in primates predicts thelong post-reproductive period in humans. Finally, this effecthas been demonstrated empirically in some populationsthrough higher survival and fertility of grandchildren in thepresence of grandmothers (Hawkes et al., 1997; Sear et al.,2000; Jamison et al., 2002; Lahdenpera et al., 2004; Gibsonand Mace, 2005).

Data on the life history of Homo sapiens may vary consid-erably. This study used data from the Ache, a hunter-gatherpopulation (Hill and Hurtado, 1996). Their life historiesclosely resemble the pattern that had evolved shortly beforeintense and relatively recent domestication and industrializa-tion, which could have substantially altered the average tim-ing and rates of life history events. Kaplan et al. (2003)compiled several life history variables (including adult mor-tality rate, age at first reproduction, and interbirth interval)from four hunter-gather populations, including the Ache.The variables have coefficients of variation of 0.09 or lessacross these four populations, and thus we are confidentthat these data from studies of the Ache adequately representthe human hunter-gather life history pattern. To ensure thesignificance of the relationships is not being unduly influ-enced by humans, all of the analyses were repeated ex-cluding humans from the sample. Any discrepancies arereported in the results.

The data on brain and body sizes are provided in Table 3.Obtaining data on brain sizes strictly from sources of wilddata is difficult because often only data from captive animalsare available, particularly for neonatal brain size. Data onadult brain sizes of females (Br) and neonatal brain size ofboth sexes (NBr) were gathered from several sources, sowithin a given species, a weighted mean was calculated ifmore than one source was utilized. Values for neonatal brainsize were only available for 12 species. Because of this smallsample size, all of the regression models were bootstrapped(1,000 iterations with replacement) to ensure that the p-valueswere not subject to Type I or Type II errors.

Some studies measured actual brain weight, whereas othersmeasured cranial capacity. Using cranial capacity as a proxyfor brain size can be problematic because the relationship ofbrain weight to the volume of the cranium has a slight negativeallometry (i.e., species or individuals with larger brains havea smaller brain weight relative to their cranial capacity). How-ever, the relationship is consistent, and Martin (1990) deviseda correction factor to ameliorate this problem, which wasapplied to the cranial capacity values to convert them to an ap-proximation of brain weight [brain weight (cranial capacity/0.94) 0.9804].


The values for adult body weight of females (W) were ob-tained from two different sources. The first value, provided inTable 3, was obtained from the field study in which the lifehistory data were collected, if available. Alternative sourcesare listed in Appendix 1. The second value was compiledfrom the data provided in Smith and Jungers (1997). Themethodological reasons for using two values are discussed be-low under analysis. Multiple sources for neonatal body weight(NW) are not available for most species, so only one value wascompiled for this variable. Smith and Leigh (1998) have dem-onstrated that many primate species show a significant degreeof sexual dimorphism in neonatal body weight, which is cor-related with their adult levels of sexual dimorphism. The cur-rent study used female neonatal body weights, wheneverpossible, for two reasons: (1) amount of postnatal body growthis calculated relative to adult female body weight; and (2)measures of reproductive success are often expressed as thenumber of daughters (Stearns, 1992; Charnov, 1993).

Analysis

The effect of body mass must be considered since bodymass and brain weight are themselves related. A simple test

for the relationship between a life history variable and brainsize could instead be detecting a relationship between that var-iable and body mass (Economos, 1980; Harvey and Bennett,1983; Harvey and Pagel, 1991). The prolongation of growthperiods could be a simple effect of growing a larger bodydittakes more time to grow larger. Among mammals, this patternhas been confirmed by studies showing that larger body sizesare correlated with slower life histories (Western, 1979;Eisenberg, 1981; Western and Ssemakula, 1982; Millar andZammuto, 1983; Blueweiss et al., 1987; Purvis and Harvey,1995). Researchers have attempted to control for this problemthrough residuals analysis (e.g., Harvey et al., 1987; Allmanet al., 1993; Deaner et al., 2003) and multiple regression(e.g., Sacher, 1975, 1978; Barton, 1999). Both techniquesare able to first account for the variation in life history ex-plained by body mass, and then examine the amount of re-maining variation explained by brain size. However, bodymass estimates are far more variable and prone to measure-ment error within a given species than are brain size estimates(Economos, 1980; Pagel and Harvey, 1988b; Smith andJungers, 1997). If the average body mass for a given speciesis over- or underestimated, then the residuals for both brainsize and the life history variables will be biased in the samedirection. For instance, a negative residual may result fromboth regressions because the mean body size is overestimated(as when captive animals are used). Thus, the test may producespuriously significant results (Harvey and Krebs, 1990; Bar-ton, 1999). One approach that attempts to alleviate this bias in-volves using separate body mass estimates to calculate theresiduals for brain size versus the residuals for the life historyvariable (Harvey and Krebs, 1990; Barton, 1999; Deaner et al.,2000, 2003; Ross, 2004). While this method does not com-pletely solve this problem, since body mass is still highly vari-able, it is a marked improvement over methods that do notconsider the intraspecific variation in body mass. To calculateresiduals, the present study used a reduced major axis regres-sion because, unlike least squares regression, it does not as-sume that the x-variable (body size) is measured withouterror (Sokal and Rohlf, 1995). Multiple regression is an advan-tageous technique for controlling for body size because it candetect the relative amount of influence of body size versusbrain size on the variation of a given life history stage. How-ever, it cannot use more than one estimate of body size, andthus cannot ameliorate the spurious correlation problem dis-cussed above. In addition, multiple regressions suffer fromthe problem of collinearity because of the strong relationshipbetween body and brain size in primates (e.g., Martin,1990). To avoid spurious correlations from error in bodymass estimates and to avoid the problem of collinearity, allof the analyses were also performed using residuals analysis.Only the results of multiple regressions are reported unlessthere is some discrepancy between the direction of the rela-tionship or level of significance between the two types ofanalyses.

Another potentially confounding factor is phylogeneticnonindependence. As discussed in the introduction, correla-tions between variables could be the result of inheritance

Table 3Body and brain weightsa

Species W Br NWb NBr

Microcebus murinus 62 1.73Eulemur fulvus rufus 2,210 13.35Lemur catta 2,200 22.58 65 8.78Propithecus verreauxi verreauxi 2,800 26.75 107Propithecus diadema 5,900 41.04Leontopithecus rosalia 600 13.00 55.3Cebus capucinus 2,540 63.85 250 32.4Cebus apella nigritus 2,070 68.81 232Cebus olivaceus 2,520 55.14Lagothrix lagotricha 7,650 92.68 450Ateles geoffroyi 7,290 105.84 485b 62.6Brachyteles arachnoides 8,380 115.5Alouatta palliata 5,670 49.18 460b 30.8Alouatta seniculus 4,500 50.86Presbytis thomasi 6,700 57.17Presbytis entellus 12,300 102.1 500Erythrocebus patas pyrrhonotus 5,750 84.42 625Chlorocebus aethiops 3,530 60.69 430Macaca fuscata yakui 8,030 84.00 503Macaca fascicularis 3,050 63.16 375Macaca mulatta 5,180 89.1 466.3 57.3Papio cynocephalus 15,900 151.96 803 77.3Papio hamadryas 11,860 132.19 695 75Hylobates lar 5,300 97.86 407 63.4Pongo pygmaeus abelii 37,800 350.87 1,653 161.3Gorilla gorilla beringei 95,000 429.75 1,996 260.5Pan troglodytes 35,200 351.27 1,814 142.2Homo sapiens 56,700 1212.72 3,334 359.4a W All body sizes and brain sizes are given in grams. Female body

weight from wild studies, references provided in Appendix 1; alternativebody weights for residuals analyses are from Smith and Jungers (1997);Br adult female brain weight; NW neonatal body weight, females;NBr neonatal brain weight, mixed sex.b Sex-specific values were not available for Ateles geoffroyi and Aloutta

palliata.


Christina Vo

Christina Vo

Christina Vo

from a common ancestor rather than the result of a functionalrelationship (Felsenstein, 1985). Calculating independentcontrasts is the method most commonly used to control forphylogenetic bias in continuous variables (Harvey andPagel, 1991; Purvis and Webster, 1999; Nunn and Barton,2001). In this study, we used the PDAP module of theMesquite program (Garland et al., 1999; Garland and Ives,2000) to calculate independent contrasts for each of the vari-ables. To ensure that the results represent a functional processand not an artifact of phylogeny, all of the analyses were per-formed using contrasted values. However, Martin et al. (2005)have raised some concerns about the methods that attemptto correct for phylogenetic inertia. They argue that (1) thesemethods may obscure functionally relevant grade-shifts inthe relationship between two variables; (2) most of thevariation at higher taxonomic levels is accounted for by vari-ation in body size, thus analyses that account for body size(through residuals or multiple regression) have already ac-counted for most of the phylogenetic inertia; and (3) differenttypes of phylogenetic inertia may be acting on the relationshipbetween any two variables (e.g., inertia in the scaling relation-ship or inertia restricted to only one variable). All of theseproblems are pertinent to this study, so all analyses wereperformed on both log-transformed species values and inde-pendent contrasts. Any disparities in the level of significancebetween these two sets of analyses are noted in the results.Otherwise, only the results of the analyses of contrasts arereported.

The branch lengths on the phylogenetic tree used to gener-ate the contrasts can be set to 1.0, or each branch can be set toa specific length representing the time since divergence (thisstudy used branch lengths provided in Purvis, 1995; Smithand Cheverud, 2002). The best method depends on the confi-dence that the distances, or divergence times, between speciesand nodes on the tree are correct. To determine which set ismost suitable for the data, the relationship between the abso-lute contrasts and the square root of the sum of the branchlengths was calculated. A slope that is significantly differentfrom zero indicates that the value for a given contrast is relatedto its divergence time, which would cause more weight to begiven to contrasts that are either more distant (positive rela-tionship) or more recent (negative relationship). In this study,branch lengths of 1.0 showed no significant relationships,while specific branch lengths had slopes that were significantlydifferent from zero in several instances. Therefore, branch

lengths equal to 1.0 were used for the analyses. In addition, ab-solute contrasts were standardized by dividing them by thesquare root of the sum of the branch lengths. This was donebecause the further back on the roots of a tree, towards themost primitive character states, the contrasts are more andmore removed from the observed values and are estimatedthrough an averaging process. Thus, the estimated primitivecharacters states were given less weight than the topmoststates (Garland et al., 1999).

Results

Developmental costs

The prediction that greater encephalization entails a devel-opmental cost is supported by the analyses of all of the pre-re-productive life history stages except length of lactationalperiod. The results of the multiple regression analyses aresummarized in Table 4 and plots are provided in Figure 1.With regards to gestation length, there are some disparities be-tween analyses. Gestation length against brain size is not sig-nificant in the residuals analysis (p 0.09) of independentcontrasts. The results are significant in the residuals analysisof species values (p 0.03), and both of the multiple regres-sion analyses (species values and independent contrasts) showthat brain size and body size have significant effects on gesta-tion length. However, two contrasts in the multiple regressionare on the extreme ends of the range of variation, and are prob-ably driving the relationship between gestation length andadult brain size, as illustrated in Figure 1a. These results leavesome ambiguity about whether a correlation between brainsize and gestation length exists. However, the link is confirmedby the analysis of neonatal brain size and gestation length (seebelow).

Lactation length is the only developmental period thatshows no correlation to brain size. Instead, the multiple regres-sion shows that the lactation length is positively correlatedwith body size, as shown in Table 4 and Figure 1b. Similar re-sults were found in analyses with species values and analysesof residuals. Thus, species with a large body size have longerperiods of lactation regardless of their brain size.

The juvenile/adolescent period shows a strong positivecorrelation with adult brain size but shows no significant cor-relation with body size (Table 4, Fig. 1c). Likewise, age atfirst reproduction (Fig. 1d) and total period of immaturity

Table 4Results of analysis of the effects of adult brain size and body size on developmental life history variables using multiple regression; all analyses use independentcontrasts (n 27 contrasts). Significant p-values in boldLife history variable Model summary Brain size Body size

r2 F-stat p-value coefficient (std error) p-value coefficient (std error) p-value

Gestation lengtha 0.84 64.27 p< 0.0001 0.11 (0.05) 0.006 0.12 (0.04) 0.05Lactation length 0.50 12.56 p< 0.0001 #0.22 (0.22) 0.34 0.53 (0.18) 0.006Juvenile/ adolescent period 0.69 22.72 p< 0.0001 0.69 (0.14) 0.0001 #0.22 (0.22) 0.06Age at first reproduction (AFR) 0.69 27.34 p< 0.0001 0.44 (0.13) 0.003 0.004 (0.11) 0.97Total immaturity (GestationAFR) 0.70 29.43 p< 0.0001 0.41 (0.12) 0.003 0.004 (0.09) 0.99a Not all analyses demonstrated a significant relationship between gestation length and adult brain size, see text for details.


Christina Vo

Christina Vo

(Fig. 1e) are correlated with adult brain size to the exclusionof body size. The same results were also found in analyses ofspecies values and residuals. These developmental life his-tory periods are directly related to each other given thattwo of the variables (the periods of juvenility/adolescentand total immaturity) are calculated from the remaining vari-able (age at first reproduction), so similar results are ex-pected. These analyses demonstrate that the length of the

juvenile/adolescent period is the primary developmental stagethat is correlated with brain size, and consequently this rela-tionship strongly affects the length of the entire period of im-maturity. However, the same two contrasts that are drivingthe relationship between gestation length and adult brainsize are also affecting the analyses of these three variables,as labeled on the respective figures. The points for thesethree variables (juvenile/adolescent period, age at first

Fig. 1. Plots from the multiple regression analyses of life history variables on body and brain size. The y-axis represents the residual variation in the life historyvariable after the body size effect has been removed. The x-axis plots the brain size effect. All variables in the analyses are independent contrasts. Details of theresults of statistical analyses are provided in Table 4.


Christina Vo

Christina Vo

reproduction, and total period of immaturity) are much closerto the fit line than is the case with gestation length. Thus, thepattern is evident, although the strength of the model (re-flected by the r2 values) is undoubtedly increased by thepresence of these contrasts.

The results of multiple regression analyses of neonatalbrain size, neonatal body size and life history variables pro-vide further confirmation that brain size is related to the lengthof growth and development. All the reported results are fromanalyses of independent contrasts. The results of analyses ofspecies values have the same levels of significance. Often,only one value for average neonatal body size was available,so residual analyses were not performed.

The results of neonatal body and brain size against gesta-tion length are summarized in Table 5 and illustrated inFigure 2. They demonstrate that neonatal brain size, but notneonatal body size, is significantly related to the length of ges-tation. In other words, these results indicate that most of thevariation in gestation length is determined by the amount ofbrain growth that occurs in utero (but see Leigh and Bernstein,2006). Unlike the analyses of adult brain size and gestationlength, these results are not affected by two outliers on the

extremes of the distribution. The lack of a correlation betweenneonatal encephalization and adult encephalization (Martin,1983, 1996; Harvey et al., 1987; Leigh, 2004) explains the dis-crepancy in results between the analysis of gestation lengthand adult brain size versus gestation length and neonatal brainsize. Some species grow their brains primarily in utero,whereas other species grow their brains primarily during thepostnatal periods. This discrepancy is particularly salient inthe distinction between strepsirrhines, which are born withsmall brains after a relatively short gestation, and anthropoids,which have longer gestation periods and larger neonatal brainsizes (Martin, 1983, 1996).

The results of comparing the amount of postnatal brain andbody growth (defined as adult size minus neonatal size)against postnatal life history stages are summarized in Table 6and illustrated in Figure 3. The length of lactation shows norelationship to either the amount of postnatal brain growthor body growth. However, the length of the juvenile/adolescentperiod has a highly significant relationship with the amount ofbrain growth, and not with body growth. This result indicatesthat postnatal growth of the brain rather than the body ac-counts for a great deal of the variation in the length of juve-nile/adolescent period (r2 0.65). Moreover, age at firstreproduction is also related to the amount of postnatal braingrowth, and not to the amount of postnatal body growth.Thus, the length of the entire postnatal growth period is pri-marily determined by the amount of postnatal brain growthrather than postnatal body growth.

Reproductive benefits

The tests of the benefits of encephalization demonstrate thathigher reproductive fitness is achieved through greater adultlifespan and not through increases in yearly reproductive out-put. The results of these analyses are summarized in Table 7.Only the analysis of neonatal body size/interbirth interval andbrain size using contrasts and including humans produced sig-nificant results. These results indicate that brain size has no ef-fect on yearly reproductive output across primates.

The results of analyses of life expectancy at age of firstreproduction in the wild demonstrate a clear relationship be-tween decreased mortality and encephalization. As shown inTable 7 and Figure 4, the multiple regression analysis ofindependent contrasts shows a positive effect of brain size re-gardless of whether contrasts or species values are used, orwhether humans were excluded from the sample. However,the effect of body size is not as clear. The negative effect ofbody size is significant in the analysis of contrasts but not

Fig. 2. Plots from the multiple regression analysis of gestation length on neo-natal body and brain size. The y-axis represents the residual variation in ges-tation length after the neonatal body size effect has been removed. The x-axisplots the brain size effect. All variables in the analysis are independent con-trasts. Details of the results of the statistical analysis are provided in Table 5.

Table 5Results of an analysis of the effects of neonatal brain size and body size on gestation length using multiple regression; the analysis uses independent contrasts(n 11 contrastsa). Significant p-values in boldLife history variable Model summary Neonatal brain size Neonatal body size

r2 F-stat coefficient (std error) p-value coefficient (std error) p-value

Gestation length 0.77 14.84 p 0.001 0.22 (0.09) 0.04 #0.03 (0.09) 0.94a Due to small sample sizes, the regression models were bootstrapped with 1,000 iterations to avoid Type I and Type II errors in the p-values of the coefficients.


Christina Vo

Christina Vo

Christina Vo

Christina Vo

Christina Vo

Christina Vo

Christina Vo

species values. This same pattern is evident in the residualsanalyses, in which the contrasts reach significance (p 0.01)and the species values do not (0.07). This is an enigma, giventhat most studies show a decrease in significance from speciesvalues to independent contrasts. It is possible that one of thespecies has a body size value with considerable error, andthe degree of error is dampened by independent contrasts. Itis likely that the lack of significance is also caused by the lim-ited power of the test due to small sample size (n 8 species).Nonetheless, large brain size has a consistent relationship withgreater life expectancy.

The results of the analysis of maximum adult lifespan areprovided in Table 8 and illustrated in Figure 5. The correlationbetween brain size and adult lifespan is significant in all of theanalyses that utilize multiple regression. However, the rela-tionship was not evident in the analyses of residuals that ex-clude humans. The relationship of between encephalizationand adult lifespan, using data on maximum lifespan from cap-tivity, may not be robust to the exclusion of the genus Homo.Conversely, there may be a high degree of error in the bodysize estimates, given that the body size coefficients were

Table 6Results of analyses of the effect of postnatal brain and body growth on several postnatal life history periods, using multiple regression; all analyses use independentcontrasts (n 11 contrastsa). Significant p-values in boldLife history variable Model summary Amount of postnatal brain growth

(adult e neonatal brain size)Amount of postnatal body growth(adult e neonatal body size)

r2 F-stat coefficient (std error) p-value coefficient (std error) p-value

Lactation length 0.09 0.31 p 0.74 0.02 (0.26) 0.95 0.13 0.25 0.61Juvenile/adolescent period 0.65 9.01 p 0.007 0.52 (0.14) 0.005 #0.14 (0.14) 0.32Age at first reproduction 0.65 8.56 p 0.008 0.42 (0.12) 0.007 #0.08 (0.12) 0.50a Due to small sample sizes, all regression models were bootstrapped with 1,000 iterations to avoid Type I and Type II errors in the p-values of the coefficients.

Fig. 3. Plots from the multiple regression analysis of life history variables on postnatal body and brain growth (adult size minus neonatal size). The y-axis rep-resents the residual variation in the life history variable after the effect of postnatal body growth has been removed. The x-axis plots the postnatal brain growtheffect. All variables in the analysis are independent contrasts. Details of the results of the statistical analyses are provided in Table 6.


Christina Vo

significant only when Homo was included in the multipleregressions.

To ensure that larger brain size is correlated with longeradult lifespan to the exclusion of any effect from variationin yearly reproductive output, a multiple regression was per-formed that considered the effect of both life history variableson brain size. The results are provided in Table 9, and demon-strate that larger body size and longer adult lifespan arepositively correlated with larger brain size when yearly repro-ductive output (neonatal body size/interbirth interval) iscontrolled. Thus, species with greater encephalization havea longer adult lifespan, and consequently would have morecycles of offspring production than small-brained species

regardless of yearly reproductive output. These results wereevident whether species or contrasts were analyzed, andwhether or not humans were included in the analysis.

Discussion

Developmental costs

The results of this study confirm that large brain size hasa developmental cost in primates: the duration of all the pre-reproductive stages, except lactation, are positively correlatedwith brain size. However, this developmental cost does not ap-pear to be caused by the brains high energy requirements, assuggested by both Sacher and Staffeldt (1974) minimax hy-pothesis and Martins (1996) maternal energy hypothesis,because only gestation length was positively correlated withencephalization. Earlier studies of brain growth in a limitedsample of primate species showed that growth occurs primar-ily during gestation and lactation (Schultz, 1941, 1965; Count,1947; Dobbing and Sands, 1973; Holt et al., 1975; Martin,1983; Vrba, 1998; Herndon et al., 1999). However, more re-cent studies of brain growth in a larger sample of primatesshowed that it is highly variable in rate and duration, andmay complete either before or after weaning (Pereira andLeigh, 2003; Leigh, 2004). This finding may explain whythe results of the current study did not show a relationship be-tween adult brain size or postnatal brain growth and the lengthof lactation. It appears that different primate species use differ-ent strategies for meeting the metabolic costs of growinga large brain; some species grow most of the brain prenatallyand others grow most of it postnatally (Harvey and Clutton-Brock, 1985; Leigh, 2004). Thus, the amount of brain growthcompleted at birth is reflected in the length of gestation. Neo-natal brain size is correlated with gestation length in the pri-mates included in this study, supporting the findings ofprevious studies of primates (Harvey and Clutton-Brock,1985; Harvey et al., 1987; but see Leigh and Bernstein,2006), and the findings of studies that include all mammals(Sacher and Staffeldt, 1974; Pagel and Harvey, 1988a). In con-trast, the length of lactation appears to reflect the amount of

Table 7Results of analyses of the effects of adult brain size and body size on variables that increase lifetime reproductive fitness, using multiple regression

Life history variable Model summary Brain size Body size

r2 F-stat p-value Coefficient (std error) p-value Coefficient (std error) p-value

Neonatal body size/IBIn 21 species 0.66 17.61 p< 0.0001 0.50 (0.26) 0.07 0.06 (0.22) 0.78n 20 contrasts 0.54 11.30 p 0.0007 0.54 (0.22) 0.02 0.05 (0.18) 0.77n 20 species Homo excluded 0.56 10.64 p 0.001 0.38 (0.33) 0.27 0.13 (0.25) 0.60n 19 contrasts Homo excluded 8.35 p 0.003 0.14 (0.27) 0.61 0.29 (0.19) 0.14Life expectancya

n 8 speciesb 0.95 51.61 p 0.0005 0.68 (0.13) 0.003 #0.35 (0.14) 0.06n 7 contrasts 0.81 22.79 p 0.003 0.71 (0.13) 0.003 #0.39 (0.13) 0.03n 7 species Homo excluded 0.90 18.64 p 0.009 0.71 (0.19) 0.02 #0.36 (0.18) 0.11n 6 contrasts Homo excluded 0.80 5.63 p 0.07 0.69 (0.25) 0.05 #0.38 (0.21) 0.14a Due to small sample sizes, all regression models were bootstrapped with 1,000 iterations to avoid Type I and Type II errors in the p-values of the coefficients.b The residuals analysis of species values did not reach significance: r2 0.45, F 4.89, coefficient 0.66 (0.30), p 0.07.

Fig. 4. Plot from the multiple regression analysis of life expectancy at adult-hood on body and brain size. The y-axis represents the residual variation inlife expectancy after the effect of body size has been considered. The x-axisplots the brain size effect. All variables in the analysis are independentcontrasts, and Homo is included in the sample. Details of the results of the sta-tistical analysis are provided in Table 7.


Christina Vo

Christina Vo

Christina Vo

Christina Vo

Christina Vo

Christina Vo

Christina Vo

Christina Vo

postnatal somatic growth, as suggested by its significant corre-lation to adult female body size.

The total amount of postnatal brain growth is correlatedwith the total period of postnatal immaturity (as measuredby age at first reproduction), and more specifically the juve-nile/adolescent period. The strong relationship between the ju-venile period and brain size is somewhat of a conundrum.Adult-sized brain volume is frequently attained by weaningor shortly afterwards (Schultz, 1941, 1965; Count, 1947;Dobbing and Sands, 1973; Holt et al., 1975; Martin, 1983;Vrba, 1998; Leigh, 2004). The juvenile/adolescent period maytherefore reflect a period of catch-up growth, in which theenergy that was devoted towards brain growth during the earlystages is now diverted to somatic growth (Vinicius, 2005). Anexamination of somatic growth trajectories supports this

notion. In primates, longer duration of growth is associatedwith prolongation of early stages of the growth trajectory,which are characterized by slower growth rates, and a reduc-tion in the relative length of later stages that have high growthrates (Leigh and Shea, 1996; Leigh and Park, 1998; Leigh,2001; see also Vinicius, 2005). In addition, the long juvenileperiods and slow growth rates of primates has been suggestedto be a strategy to reduce the risk of metabolic deficiency, par-ticularly in the face of elevated feeding competition stemmingfrom living in large social groups (Janson and van Schaik,1993). However, the precise relationship between the corre-sponding somatic and brain growth trajectories remains to betested on a range of primate species.

Greater learning opportunities and more time to becomeskilled at difficult tasks are other explanations for the relation-ship between postnatal brain growth and the length of juvenil-ity/adolescence (Dobzhansky, 1962; Gould, 1977; Joffe, 1997;Bogin, 1999; Ross and Jones, 1999; Kaplan et al., 2000). How-ever, this idea has not always stood up under empirical scru-tiny. Several studies have suggested that many of the skillsnecessary for survival reach adult levels well before age atfirst reproduction (Janson and van Schaik, 1993; Pereira andFairbanks, 1993; Blurton-Jones et al., 1999). Also, small sizeand limited strength rather than lack of experience may pre-vent juveniles from acquiring an optimum diet.

Nonetheless, it is possible that the acquisition of skills incomplex foraging niches requires time and experience. Theconnection between the length of the juvenile period, the ac-quisition of complex skills, and brain size is due to protractedbrain maturation at a cellular level, and consequently a behav-ioral level as well. Deaner et al. (2003) introduced the mat-urational constraints hypothesis, which links the slowing oflife history with the behavioral and cellular development of

Table 8Results of analyses of the effects of adult brain size and body size on adult lifespan using multiple regression. Significant values in bold

Response: adult lifespan Model summary Brain size Body size

r2 F-stat p-value Coefficient (std error) p-value Coefficient (std error) p-value

n 25 species 0.60 16.41 p 0.0001 0.53 (0.15) 0.0014 #0.26 (0.13) 0.05n 24 contrasts 0.41 8.16 p 0.002 0.56 (0.16) 0.002 #0.30 (0.13) 0.02n 24 speciesaexcluding Homo

0.45 8.61 p 0.002 0.41 (0.16) 0.02 #0.18 (0.13) 0.18

n 23 contrastsbexcluding Homo

0.25 3.51 p 0.05 0.41 (0.20) 0.05 #0.20 (0.15) 0.19

a The residuals analysis of species values excluding Homo did not reach significance: r2 0.15, F 3.76, coefficient 0.44 (0.23), p 0.07.b The residuals analysis of contrast values excluding Homo did not reach significance: r2 0.12, F 3.06, coefficient 0.34 (0.16), p 0.09.

Fig. 5. Plot from the multiple regression analysis of maximum adult lifespanon body and brain size. The y-axis represents the residual variation in adultlifespan after the effect of body size has been removed. The x-axis plots thebrain size effect. All variables in the analysis are independent contrasts, andHomo is included in the sample. Details of the results of the statistical analysisare provided in Table 8.

Table 9Results of a multiple regression of the effects of body size, adult lifespan, andinterbirth interval on brain size, using independent contrasts; r2 0.85,F-stat 33.82, p< 0.0001. Significant values in bold

Coefficient (std error) p-value

Body size 0.51 (0.10)

a large brain. They propose that immature brains do not func-tion at adult competency levels, particularly for complexskills, until reproductive maturity. Though the brain has al-ready reached its adult volume long before age at first repro-duction, it is not structurally mature. Synaptogenesis, theprocess by which the brains network of connections isformed, is ongoing until about the age at first reproductionin primates (macaques: Rakic et al., 1986; marmosets: Missleret al., 1993; humans: Bourgeois, 1997; Huttenlocher and Dab-holkar, 1997), and myelination of these connections also con-tinues until approximately the age at first reproduction(macaques: Gibson, 1970, 1991; humans: Giedd et al.,1996). Extended development in humans (the only primatestudied thus far) is further suggested by changes in neocorticalareas that govern particular cognitive tasks until they stabilizeat maturity (Gogtay et al., 2004; Casey et al., 2005). The net-work of connections in the neocortex is built and maintainedthrough external stimulation, a process of fine-tuning thebrains circuits through interaction with the social and physicalenvironment, [i.e., the learning environment necessary forlearning complex skills (e.g., Edelman, 1987; Greenoughet al., 1987; Elman et al., 1996; Quartz and Sejnowski,1997; Krubitzer and Kahn, 2003)]. These processes are timeconsuming, and given that a larger brain entails a larger andmore complex network, the necessary period of environmentalinput increases with brain size. These developmental processescontinue on a cellular level during the juvenile/adolescent pe-riod, and provide a potential explanation for the correlation be-tween the length of this stage and brain size.

This hypothesis of prolonged brain maturation in more en-cephalized species is supported by studies of the relationshipbetween life history and neocortex size, which also show a par-ticularly strong correlation with the period of juvenility (Joffe,1997; Kaplan et al., 2003), as well as age at first reproduction(Walker et al., 2006). In addition, the findings of Walker et al.(2006) are robust to the inclusion of potentially confoundingfactors such as diet and home range in the model.

In humans, this phenomenon is illustrated by hunting skills,which do not reach peak efficiency until at least reproductivematurity (Hill and Hurtado, 1996; Blurton-Jones et al., 1999;Kaplan et al., 2000; Bock, 2002; Kramer, 2002; Stout, 2002;Walker et al., 2002; Robson and Kaplan, 2003; but see Birdand Bliege Bird, 2002; Blurton-Jones and Marlowe, 2002).Humans are not the only species to exhibit this pattern.Whereas some studies did not show lower foraging efficiencyin juveniles relative to adults (reviewed above), others sug-gested that competency is not achieved until sexual maturity,and that this pattern was not necessarily due to small bodysize (birds: Ricklefs, 1983, 1984; Marchetti and Price, 1989;Wunderle, 1991; carnivores: Caro and Hauser, 1992; ceta-ceans: Baird, 2000; primates: Boinski and Fragaszy, 1989;Byrne and Byrne, 1993; Matsuzawa, 1994; Boesch andBoesch-Achermann, 2000; Johnson and Bock, 2004; see alsovan Schaik et al., 2006). In addition, some juvenile primatesand birds have diets that exclude foods that are difficult to ob-tain or process (Pereira and Altmann, 1985; Marchetti andPrice, 1989; Yoerg, 1994). These behavioral studies strongly

suggest that the brain does not function at full competence un-til adulthood in at least some species for at least some skills.

These findings suggest that the connection between enceph-alization and long life history may be the result of adaptationsto a complex foraging niche (see Kaplan et al., 2000, 2003).Extended dental development has been shown to correlatewith slow-paced life history and large brain size in anthropoidprimates (Smith, 1989a,b; 1991; Smith et al., 1995), and theseconnections may be explained as ecological/dietary adapta-tions. Godfrey et al. (2001) demonstrate that protracted dentaldevelopment is correlated with large brain size, slow acquisi-tion of foraging independence, and high-quality food resources(e.g., large percentage of fruit versus leaves).

Adult benefits

The benefits of encephalization are demonstrated by greaterlife expectancy at age at first reproduction, and thereby an in-crease in net reproductive fitness. Encephalization can beadaptive because it provides more complex cognitive skillsthat may serve to decrease mortality (e.g., Byrne, 1996;Rumbaugh, 1997; Beran et al., 1999; Gibson and Jesse, 1999;Gibson et al., 2001; Gibson, 2002; Kaplan et al., 2003). Largebrains improve the ability to find innovative solutions to eco-logical problems and are associated with greater behavioralflexibility (birds: Lefebvre et al., 1997, 1998, 2002; Timmer-mans et al., 2000; Sol et al., 2002, 2005a,b; Lefebvre andBolhuis, 2003; Shultz et al., 2005; primates: Reader andLaland, 2002; birds and primates: Reader, 2003; Lefebvreet al., 2004). They are also associated with more effective re-source mapping and food acquisition (Gibson, 1986; Milton,1988; Sawaguchi, 1990; Sawaguchi and Kudo, 1990), as wellas more complex social strategies (Dunbar, 1995; Barton,1996; Whiten and Byrne, 1997; Pawlowski et al., 1998; ParkerandMcKinney, 1999;Burish et al., 2004;Byrne andCorp, 2004).

Humans, in particular, use their complex social system tobuffer injured or ill individuals from potentially life-threateningsituations through provisioning and healthcare (Kaplan et al.,2000; Sugiyama, 2004; Hawkes, 2006). All of these cognitiveskills have the potential to increase the chances of survivaland reproductive success. In addition, neocortex size has beenshown to be correlated with many measures of cognitive ability,such as rates of innovation, social gregariousness, and complexforaging techniques (Dunbar, 1992; Barton, 1996; Joffe, 1997;Lewis, 2000; Nicolakakis and Lefebvre, 2000; Kaplan et al.,2003; Reader and MacDonald, 2003; Walker et al., 2006).

Maximum lifespan is derived from captive records, and onemight argue these artificial conditions do not replicate the pres-sures faced in the wild. However, it represents a physiologicallimit on lifespan, which is probably set by natural selection tobalance energy invested in cellular maintenance with the long-term mortality rate (Kirkwood and Austad, 2000). Evidencefor maintenance is particularly salient in the brain, where selec-tion for proteins that regenerate and remyelinate axons is foundin long-lived organisms (Finch and Sapolsky, 1999; Finch andStanford, 2004; Allen et al., 2005). Maximum lifespan couldrepresent the full realization of the investment in cellular


Christina Vo

Christina Vo

Christina Vo

Christina Vo

Christina Vo

Christina Vo

Christina Vo

Christina Vo

Christina Vo

maintenance. In contrast, life expectancy is an estimate based oncurrent conditions in the wild when the data were collected. Formany reasons, including human interference, these conditionsmay deviate from the long-term average. Thus, a large sampleof species and careful considerations of the conditions at thefield site are necessary to better determine the effect of brainsize on a species demography.

The cognitive buffer hypothesis proposes that large brainsprovide enhanced cognitive capacity that facilitates behaviorsthat can extend the life expectancy and reproductive success ofthe species, such as difficult extractive-foraging techniques,feeding innovations, problem-solving, learning ability, socialgregariousness, and other strategies of predator avoidance(Allman et al., 1993; Hakeem et al., 1996; Allman and Hasen-staub, 1999; Kaplan et al., 2000; Judge and Carey, 2000; Careyand Judge, 2001). The relationship between these skills and re-duction in mortality still requires more testing, but the connec-tion between longevity and the ability to procure moreresources, attract more mates, avoid escalation of social con-flicts, and avoid predation is highly likely.

These results support the embodied capital hypothesisproposed by Kaplan et al. (2000), in which the costs of learn-ing complex foraging skills are compensated by surplus foodproduction and reduced mortality in adults. Hawkes (2006) de-tails an alternate hypothesis the embodied capital, or hunt-ing hypothesis. She emphasizes the life history trade-offs infemales rather than the investment in hunting ability in males.Specifically, slow aging and increased lifespan results in post-reproductive females who have the skills and strength necessaryto provide the extra provisions to the expensive immatures.This extra provisioning decreases childhood mortality, andallows reproducing females to shorten their interbirth inter-vals. The present study supports the notion that there is a heavyinvestment required to grow large-brained offspring, and thatthese costs are offset by long-living adults. However, it issilent as to whether grandmothers, fathers, or others are doingthe provisioning (see also van Schaik et al., 2006).

Conclusions

This study sought to clarify the relationship between lifehistory and brain size. Many previous studies had found am-biguous and contradictory evidence, weakening the ability togenerate hypotheses addressing the nature of the evolutionaryrelationship. This study relied on data from wild populationsalthough gestation length, which has been shown to be almostinvariable, was sometimes taken from captive records. Maxi-mum lifespan data were also taken from captive studies. How-ever, life expectancy in the wild formed the crux of ouranalysis on the reproductive benefit of encephalization.

Our perspective on the relationship between brain size andlife history argues that the evolutionary link between thesetraits depends on a balance between costs and benefits. Wefound that the benefits are manifested in primates through adecrease in adult mortality (i.e., increase in reproductivelifespan), achieved through the more complex foraging tech-niques, predator avoidance, and social skills that a larger brain

provides. These benefits must outweigh the developmentalcosts accompanying the growing of a bigger brain, specificallythe long period required to gestate a large-brained offspring,and the offsprings extended period of juvenility. Becausethe analyses demonstrate that these patterns exist across pri-mates, even when humans are excluded, the unusually slowpace of human life history and our extremely large brainsize fit the general primate trend in which increased brainsize causes extension in life history stages.

Acknowledgments

We appreciate the access to skeletal collections granted bythe following museum curators and staff members: R. Thor-ington, L. Gordon: National Museum of Natural History; R.MacPhee, E. Westwig: American Museum of Natural History.We would like to thank E. Chris Kirk for generously allowingus to use his data on cranial capacities. Also, we would like tothank Matt Cartmill, Steven Leigh, and two anonymous re-viewers for comments on previous drafts. Funding for portionsof this research was provided by Sigma Xi, Ruggles-GatesFund for Biological Anthropology, and AMNH.

Appendix 1. References for life history data from Tables 2e4. See below for abbreviation definitions

Species Study site for wild data References

Microcebusmurinus

Kirindy Forest,Madagascar

IBI, AFR: Kappeler andRasoloarison, 2003GL, W: Kappeler and Pereira,2003Br: Stephan et al., 1981; E.C.Kirk, unpublished dataAL: Hakeem et al., 1996

Eulemur fulvusrufus

Kirindy Forest,Madagascar

GL: Ostner and Heistermann,2003IBI: Overdorff et al., 1999AFR: Overdorff et al., 1999Kappeler, 1990E.C. Kirk, unpublished dataAL: Carey and Judge, 2000

Lemur catta Berenty Forest,Madagascar

GL: Kappeler and Pereira,2003IBI, AFR: Koyama et al., 2001W: Sussman, 1991Br: Stephan et al., 1981NW: Hick, 1976Ruempler, 1993NBr: Sacher and Staffeldt,1974AL: Carey and Judge, 2000

Propithecusverreauxiverreauxi

Beza Mahafaly Reserve,Madagascar

GL: Kappeler and Pereira,2003IBI, AFR, Le, W: Richardet al., 2002Br: Stephan et al., 1981NW: Harvey and Clutton-Brock, 1985AL: Carey and Judge, 2000


Christina Vo

Christina Vo

Christina Vo

Christina Vo

Christina Vo

Christina Vo

Christina Vo

Appendix 1 (continued)


Propithecusdiadema

Ranomafana NationalPark

GL: Wright, 1995IBI, AFR, Le: Pochron et al.,2004W:GoodmanandBenstead,2003Br: E.C. Kirk, unpublisheddata; cranial capacitiescollected by N.B.

Leontopithecusrosalia

Poco das AntasBiological Reserve,Brazil

GL: Kleiman, 1977IBI, W: Dietz et al., 1994AFR: Bales et al., 2001Br: Stephan et al., 1981;E.C. Kirk, unpublished dataNW: Harvey and Clutton-Brock, 1985AL: Carey and Judge, 2000

Cebuscapucinus

La Pacifica (Santa Rosa),Costa Rica

GL, W, AFR: Fedigan andRose, 1995IBI: Fedigan, 2003Br: Schultz, 1941; Stephanet al., 1981; E.C. Kirk,unpublished dataNW: Glander et al., 1991NBr: Leutenegger, 1970, 1973;Schultz, 1941AL: Carey and Judge, 2000

Cebus apellanigritus

Iguazu National Park,Argentina

GL, IBI, AFR, W: Di Bitettiand Janson, 2001Br: Stephan et al., 1981;cranial capacities collected byN.B.NW: Harvey and Clutton-Brock, 1985AL: Carey and Judge, 2000

Cebus olivaceus Hato Masaguaral,Venezuela

GL: Crockett and Sekulic, 1982IBI: Valderrama andSrikosamatara, 1990AFR, W: Robinson, 1988Br: cranial capacities collectedby N.B.

Lagothrixlagotricha

Parque Nacional NatlMacarena-Tinigua, Meta,Colombia

GL, IBI, AFR: Nishimura, 2003W: Kappeler and Pereira, 2003Br: Stephan et al., 1981, E.C.Kirk, unpublished dataNW: Harvey and Clutton-Brock, 1985AL: Carey and Judge, 2000

Ateles geoffroyi La Pacifica (Santa Rosa),Costa Rica

GL, IBI, AFR: Fedigan andRose, 1995W: Kappeler and Pereira, 2003Br: Schultz, 1941; Stephanet al., 1981; E.C. Kirk,unpublished data; cranialcapacities collected by NBW: Schultz, 1941; Miles, 1967;Leutenegger, 1973NBr: Schultz, 1941AL: Hakeem et al., 1996

Brachytelesarachnoides

Minas Gerais, Brazil GL: Strier and Ziegler, 1997IBI: Strier et al., 2001AFR: Martins and Strier, 2004W: Plavcan and van Schaik, 1997Br: E.C. Kirk, unpublished dataAL: Judge and Carey, 2000



Alouattapalliata

La Pacifica (Santa Rosa),Costa Rico

GL: Glander, 1980; Fediganand Rose, 1995IBI: Glander, 1980; Clarke andGlander, 1984AFR: Clarke and Glander,1984W: Fedigan and Rose, 1995Br: Schultz, 1941; E.C. Kirk,unpublished data; cranialcapacities collected by N.B.NW: Sacher and Staffeldt,1974; Harvey and Clutton-Brock, 1985NBr: Leutenegger, 1970AL: Carey and Judge, 2000

Alouattaseniculus

Hato Masaguaral,Venezuela

GL: Crockett and Sekulic, 1982IBI: Crockett and Rudran, 1987AFR, W: Crockett and Pope,1993Br: Stephan et al., 1981; E.C.Kirk, unpublished dataAL: Carey and Judge, 2000

Presbytisthomasi

Ketambe ResearchStation, LeuserEcosystem, SumatraIndonesia

GL: Sterck, 1999IBI, AFR, Le, W: Wich et al.,2007Br: E.C. Kirk, unpublished data

Presbytisentellus

Ramnagar, Nepal GL, IBI: Borries et al., 2001AFR: Ziegler et al., 2000;Borries et al., 2001W: C Borries, pers. comm.Br: Stephan et al., 1981AL: Carey and Judge, 2000

Erythrocebuspataspyrrhonotus

Segera Ranch, LaikipiaPlateau, Kenya

GL: Sly et al., 1983IBI, AFR: Chism et al., 1984W: Isbell, 1998Br: E.C. Kirk, unpublisheddataAL: Carey and Judge, 2000

Chlorocebusaethiops

Amboseli, Kenya GL: Bramblett et al., 1975IBI, AFR, W: Cheney et al.,1988Br: Stephan et al., 1981; Bolterand Zihlman, 2003; E.C. Kirk,unpublished data; cranialcapacities collected by N.B.NW: Harvey and Clutton-Brock, 1985AL: Carey and Judge, 2000

Macaca fuscatayakui

Yakashima Island, Japan GL: Kappeler and Pereira, 2003IBI, AFR: Takahata et al., 1998W: Plavcan and van Schaik,1997Br: Bauchot and Stephan, 1969NW: Harvey and Clutton-Brock, 1985AL: Carey and Judge, 2000

Macacafascicularis

Ketambe, Sumatra GL: Jablonski et al., 2000IBI, AFR: van Noordwijk, 1999W: Plavcan and van Schaik,1997Br: Stephan et al., 1981; ECKirk, unpublished dataAL: Carey and Judge, 2000

(continued on next page)


References

Aiello, L., Wheeler, P., 1995. The expensive-tissue hypothesis. Curr. Anthro-pol. 36 (2), 199e221.

Allen, J.S., Bruss, J., Damasio, H., 2005. The aging brain: the cognitive reservehypothesis and hominid evolution. Am. J. Hum. Biol. 17 (6), 673e689.

Allman, J., 1995. Brain and life span in Catarrhine primates. In: Butler, R.,Brody, J. (Eds.), Delaying the Onset of Late-life Dysfunction. Springer,New York, pp. 221e241.

Allman, J., Hasenstaub, A., 1999. Brains, maturation times, and parenting.Neurobiol. Aging 20, 447e454.



Macaca mulatta Dunga Gali, Pakistan GL: Jablonski et al., 2000IBI, AFR: Melnick, 1981W:PlavcanandvanSchaik, 1997Br: Schultz, 1941; Kerr et al.,1969, 1974; Stephan et al.,1981NW: van Wagenen, 1972;DiGiacomo et al., 1978;Ruppenthal, 1979; Martin,1984; Silk et al., 1993NBr: Schultz, 1941; Kerr et al.,1969, 1974; Harvey andClutton-Brock, 1985AL: Carey and Judge, 2000

Papiocynocephalus

Mikumi National Park,Tanzania

GL: Kappeler and Pereira,2003IBI, AFR, W, Le: Rhine et al.,2000Br: E.C. Kirk, unpublisheddata; cranial capacitiescollected by N.B.NW: Vice et al., 1966NBr: Hendrickx and Houston,1971AL: Carey and Judge, 2000

Papiohamadryas

Erer, Ethiopia GL: Kappeler and Pereira,2003IBI, AFR, W: Sigg et al.,1982Br: cranial capacities collectedby N.B.NW: Kaumanns, 1975;Kullmann, 1978NBr: Sacher and Staffeldt,1974AL: Carey and Judge, 2000

Hylobates lar Khao Yai National Park GL, IBI: Reichard, 2003AFR: Jablonski et al., 2000W: Plavcan and van Schaik,1997Br: Schultz, 1944; Stephanet al., 1981; E.C. Kirk,unpublished dataNW: Geissmann andOrgeldinger, 1995NBr: Schultz, 1944, 1965;Harvey and Clutton-Brock,1985AL: Carey and Judge, 2000

Pongopygmaeusabelii

Suaq Balimbing, Sumatra GL: Jablonski et al., 2000IBI: van Noordwijk and vanSchaik, 2005AFR, Le: Wich et al., 2004W: Plavcan and van Schaik,1997Br: Schultz, 1941; Tobias,1971; Stephan et al., 1981;E.C. Kirk, unpublished dataNW: Smith and Leigh, 1998NBr: Schultz, 1941, 1965;Rudder, 1979; Harvey andClutton-Brock, 1985AL: Carey and Judge, 2000



Gorilla gorillaberingei

Karisoke ResearchStation, Rwanda

GL: Kappeler and Pereira,2003IBI, AFR: Watts, 1991W: Plavcan and van Schaik,1997Br: Schultz, 1941; Tobias,1971; Stephan et al., 1981;E.C. Kirk, unpublished dataNW: Smith and Leigh, 1998NBr: Schultz, 1965; Harveyand Clutton-Brock, 1985AL: Carey and Judge, 2000

Pan troglodytes Mahale National Park,Tanzania

GL: Kappeler and Pereira,2003IBI, AFR, Le: Nishida et al.,2003; Hill et al., 2001W: Plavcan and van Schaik,1997Br: Schultz, 1941; Tobias,1971; Stephan et al., 1981;E.C. Kirk, unpublished dataNW: Grether and Yerkes, 1940;Nissen and Riesen, 1949;Keeling and Riddle, 1975;Smith et al., 1975; Smith andLeigh, 1998NBr: Gaul, 1933; Schultz,1940, 1941; Herndon et al.,1999AL: Carey and Judge, 2000

Homo sapiens Ache GL, IBI, AFR, W: Hill andHurtado, 1996Br: Schultz, 1941; Stephanet al., 1981; Ricklan and Tobias,1986; Ho et al., 1980a, bNW: Arbuckle and Sherman,1989NBr: Blinkov and Glezer,1968; Jordaan, 1976;Coppoletta and Wolbach,1933; Schultz, 1941, 1965AL: Carey and Judge, 2000

GL gestation length; IBI interbirth interval between surviving offspring;L lactational period, IBI-GL; J/A juvenile/adolescent period, AFR minusLa; AFR age at first reproduction for females; TI total immaturity, AFRplus GL, AL adult lifespan, maximum lifespan minus AFR; Le life expec-tancy at AFR; W female body weight from wild studies; alternative bodyweights for residuals analyses are from Smith and Jungers (1997); Br adultfemale brain weight; NW neonatal body weight, mixed sex; NBr neonatalbrain weight, mixed sex.


Allman, J., McLaughlin, T., Hakeem, A., 1993. Brain weight and life-span inprimate species. Proc. Natl. Acad. Sci. 90, 118e122.

Alvarez, H., 2000. Grandmother hypothesis and primate life histories. Am. J.Phys. Anthropol. 113, 435e450.

Arbuckle, T., Sherman, G., 1989. An analysis of birth weight by gestationalage in Canada. Can. Med. Assoc. J. 140, 157e165.

Austad, S., Fischer, K., 1992. Primate longevity: its place in the mammalianscheme. Am. J. Primatol. 28, 251e261.

Baird, R., 2000. The Killer Whale: foraging specializations and group hunting.In: Mann, J., Connor, R., Tyack, P., Whitehead, H. (Eds.), CetaceanSocieties: Field Studies of Dolphins and Whales. University of ChicagoPress, Chicago, pp. 127e153.

Bales, K., OHerron, M., Baker, A.J., Dietz, J.M., 2001. Sources of variabilityin numbers of live births in wild golden lion tamarins (Leontopithecusrosalia). Am. J. Primatol. 54 (4), 211e221.

Barton, R., 1996. Neocortex size and behavioural ecology in primates. Proc. R.Soc. B-Biol. Sci. B263 (1367), 173e177.

Barton, R., 1999. The Evolutionary Ecology of the Primate Brain. In: Lee, P.(Ed.), Comparative Primate Socioecology. Cambridge University Press,Cambridge, pp. 167e194.

Bauchot, R., Stephan, H., 1969. Encephalisation et niveau evolutif chez lessimiens. Mammalia 33, 225e275.

Beran, M.J., Gibson, K.R., Rumbaugh, D.M., 1999. Predicting Hominid Intel-ligence from Brain Size. The Descent of Mind: Psychological Perspectiveson Hominid Evolution. Oxford University Press, New York. 88e97.

Bird, D., Bliege Bird, R., 2002. Children on the reef: slow learning or strategicforaging. Hum. Nat. 13, 269e297.

Blinkov, S., Glezer, I., 1968. The Human Brain in Figures and Tables. PlenumPress, New York.

Blueweiss, L., Fox, V., Kudzma, V., Nakashima, D., Peters, R., Sams, S., 1987.Relationships between body size and some life history parameters. Oeco-logia 37, 257e272.

Blurton-Jones, N., Hawkes, K., OConnell, J., 1999. Some current ideasabout the evolution of human life history. In: Lee, P. (Ed.), ComparativePrimate Socioecology. Cambridge University Press, Cambridge, pp.144e166.

Blurton-Jones, N., Marlowe, F., 2002. Selection for delayed maturity: does ittake 20 years to learn to hunt and gather? Hum. Nat. 13, 199e238.

Bock, J., 2002. Learning, life history, and productivitydchildrens lives in theOkavango Delta, Botswana. Hum. Nat. 13 (2), 161e197.

Boesch, C., Boesch-Achermann, H., 2000. The Chimpanzees of the Tai Forest:Behavioural Ecology and Evolution. Oxford University Press, New York.

Bogin, B., 1997. Evolutionary hypotheses for human childhood. Yearb. Phys.Anthropol. 40, 63e90.

Bogin, B., 1999. Patterns of Human Growth, second ed. Cambridge UniversityPress, Cambridge.

Boinski, S., Fragaszy, D.M., 1989. The ontogeny of foraging in squirrel-mon-keys, Saimiri oerstedi. Anim. Behav. 37, 415e428.

Bolter, D.R., Zihlman, A.L., 2003. Morphometric analysis of growth and de-velopment in wild-collected vervet monkeys (Cercopithecus aethiops),with implications for growth patterns in Old World monkeys, apes and hu-mans. J. Zool. 260, 99e110.

Borries, C., Koenig, A., Winkler, P., 2001. Variation of life history traits andmating patterns in female langur monkeys (Semnopithecus entellus).Behav. Ecol. Sociobiol. 50 (5), 391e402.

Bourgeois, J., 1997. Synaptogenesis, heterochrony and epigenesis of mamma-lian neocortex. Acta. Paediatr. 422 (Suppl.), 27e33.

Bramblett, C.A., Pejaver, L.D., Drickman, D.J., 1975. Reproduction in captivevervet and Sykes monkeys. J. Mammal. 56, 940e946.

Burish, M.J., Kueh, H.Y., Wang, S.S.H., 2004. Brain architecture and socialcomplexity in modern and ancient birds. Brain Behav. Evol. 63 (2),107e124.

Byrne, R., 1996. Relating brain size to intelligence in primates. In: Mellars, P.,Gibson, K. (Eds.), Modelling the Early Human Mind. The McDonald Ar-chaeological Institute, Cambridge, pp. 49e57.

Byrne, R.W., Byrne, J.M.E., 1993. Complex leaf-gathering skills of mountaingorillas (Gorilla g. beringei): variability and standardization. Am. J. Pri-matol. 31 (4), 241e261.

Byrne, R.W., Corp, N., 2004. Neocortex size predicts deception rate inprimates. Proc. R. Soc. Lond. Ser. B-Biol. Sci. 271 (1549), 1693e1699.

Carey, J.R., Judge, D.S., 2000. Longevity Records: Life Spans of Mammals,Birds, Amphibians, Reptiles, and Fish. Odense University Press, Odense.

Carey, J.R., Judge, D.S., 2001. Life span extension in humans is self-reinforc-ing: a general theory of longevity. Popul. Dev. Rev. 27 (3), 411e436.

Caro, T.M., Hauser, M.D., 1992. Is there teaching in nonhuman animals? Q.Rev. Biol. 67 (2), 151e174.

Casey, B.J., Galvan, A., Hare, T.A., 2005. Changes in cerebral functional or-ganization during cognitive development. Curr. Opin. Neurobiol. 15 (2),239e244.

Charnov, E., 1993. Life History Invariants: Some Explorations of Symmetry inEvolutionary Ecology. Oxford University Press, Oxford.

Charnov, E., 2004. The optimal balance between growth rate and survival inmammals. Evol. Ecol. Res. 6, 307e313.

Charnov, E., Berrigan, D., 1993. Why do female primates have such longlifespans and so few babies? or life in the slow lane. Evol. Anthropol. 1,191e194.

Cheney, D.L., Seyfarth, R.M., Andelman, S.J., Lee, P.C., 1988. ReproductiveSuccess in Vervet Monkeys. Reproductive Success: Studies of IndividualVariation in Contrasting

life history costs and benefits of encephalization: a comparative test using data from long-term...

Documents