ARTICLE

A Theory of Human Life History Evolution:Diet, Intelligence, and LongevityHILLARD KAPLAN, KIM HILL, JANE LANCASTER, A. MAGDALENA HURTADO

Our theory is that those four lifehistory characteristics and extremeintelligence are co-evolved responsesto a dietary shift toward high-quality,nutrient-dense, and difficult-to-ac-quire food resources.

The following logic underlies ourproposal. First, high levels of knowl-edge, skill, coordination, and strengthare required to exploit the suite of high-quality, difficult-to-acquire resourceshumans consume. The attainment ofthose abilities requires time and a sig-nificant commitment to development.This extended learning phase, duringwhich productivity is low, is compen-sated for by higher productivity duringthe adult period and an intergenera-tional flow of food from old to young.Because productivity increases withage, the investment of time in acquiringskill and knowledge leads to selectionfor lowered mortality rates and greaterlongevity. The returns on investmentsin development occur at older ages.This, in turn, favors a longer juvenileperiod if there are important gains inproductive ability with body size andgrowth ceases at sexual maturity.

Second, we believe that the feeding

niche that involves specializing onlarge, valuable food packages pro-motes food sharing, provisioning ofjuveniles, and increased grouping, allof which act to lower mortality duringthe juvenile and early adult periods.Food sharing and provisioning assistrecovery in times of illness and reducerisk by limiting juvenile time alloca-tion to foraging. Grouping also lowerspredation risks. These buffers againstmortality also favor a longer juvenileperiod and higher investment in othermechanisms to increase the life span.

Thus, we propose that the long hu-man life span co-evolved with lengthen-ing of the juvenile period, increasedbrain capacities for information pro-cessing and storage, and intergenera-tional resource flows, all as a result ofan important dietary shift. Humans arespecialists in that they consume onlythe highest-quality plant and animal re-sources in their local ecology and relyon creative, skill-intensive techniques toexploit them. Yet the capacity to de-velop new techniques for extractive for-aging and hunting allows them to ex-ploit a wide variety of different foodsand to colonize all of earth’s terrestrialand coastal ecosystems.

We begin with an overview of thedata on which the theory is based: acomparative examination of hunter-gatherer and chimpanzee life-historytraits and age profiles of energy acqui-sition and consumption. The datashow that hunter-gatherers have alonger juvenile period, a longer adultlifespan, and higher fertility thanchimpanzees do. Hunter-gatherer chil-dren are energetically dependent onolder individuals until they reach sex-ual maturity. Energy acquisition ratesincrease dramatically, especially for

Human life histories, as compared to those of other primates and mammals,have at least four distinctive characteristics: an exceptionally long lifespan, anextended period of juvenile dependence, support of reproduction by older post-reproductive individuals, and male support of reproduction through the provision-ing of females and their offspring. Another distinctive feature of our species is alarge brain, with its associated psychological attributes: increased capacities forlearning, cognition, and insight. In this paper, we propose a theory that unites andorganizes these observations and generates many theoretical and empirical pre-dictions. We present some tests of those predictions and outline new predictionsthat can be tested in future research by comparative biologists, archeologists,paleontologists, biological anthropologists, demographers, geneticists, and cul-tural anthropologists.

Hillard Kaplan is Professor at the University ofNew Mexico. His recent research and publi-cations have focused on integration of life his-tory theory in biology and human capital the-ory in economics, with specific emphases onfertility, parental investment, and aging in de-veloped, developing, and traditional settings.He has also conducted fieldwork with nativeSouth Americans and southern Africans. E-mail: hkaplan@unm.edu.Kim Hill is an Associate Professor at the Uni-versity of New Mexico. He studies humanbehavioral ecology with a focus on life historytheory, foraging patterns, sexual division oflabor, food sharing, and the evolution of co-operation. He has carried out fieldwork in fivedifferent South American hunter-gatherer ortribal horticulturalist populations in the past23 years. E-mail: kimhill@unm.edu.Jane Lancaster is Professor of Anthropologyat the University of New Mexico. Her re-search and publications are on human repro-ductive biology and behavior especially hu-man parental investment; on women’sreproductive biology of pregnancy, lactationand child-spacing; and on male fertility andinvestment in children. She edits the quarterlyjournal, Human Nature, which publishes re-search in human evolutionary ecology.E-mail: jlancas@unm.edu.A. Magdalena Hurtado is Associate Professorin the Department of Anthropology, Univer-sity of New Mexico. She has done researchon a wide range of problems in human be-havioral ecology and evolutionary medicineamong the Ache, Hiwi, and Machiguenga ofLowland South America. She is Co-Directorof the Native Peoples and Tropical Conser-vation Fund, University of New Mexico. E-mail: amhurtad@unm.edu.

156 Evolutionary Anthropology

males, until mid-adulthood and stayhigh until late in life.

We then present both theoretical andempirical tests of our theory. For thetheory to be correct, a model of naturalselection must show that mortalityrates, the length of the juvenile period,and investments in learning co-evolvein the ways predicted by the theory.Building on existing models of life-his-tory evolution,1–5 we develop such amodel. The results of our analysis con-firm the theory’s predictions. Those the-oretical tests are followed by empiricaltests. In order for our theory to be cor-rect, we must demonstrate that: Hu-mans do, in fact, consume more skill-intensive, difficult-to-acquire, high-quality foods than do chimpanzees andother nonhuman primates; Difficulty ofacquisition explains the age profile ofproduction for both humans and chim-panzees; Men play a large role in sup-porting human reproduction; The for-aging niche occupied by humanslowers mortality rates among juvenilesand adults relative to correspondingrates among chimpanzees and othernonhuman primates. We present strongevidence in support of the first threepropositions and suggestive evidence insupport of the fourth.

We then examine the evolution ofthe primate order to determine whetherthe same principles invoked in ourtheory of hominid evolution explainthe major primate radiations. We thenconsider the fundamental differencesbetween our theory and the “grand-mother hypothesis” recently pro-posed by Hawkes, Blurton Jones andO’Connell.6,7 We conclude with a list-ing of the new and unique predictionsderived from our theory.

Our theory is not the first to proposethat high-quality foods, extractive for-aging, and hunting are fundamental tohuman evolution. However, it is thefirst to do so with a specific model ofnatural selection that unifies the evolu-tion of life history, brain and intelli-gence, diet, and age profiles of food pro-duction and consumption. As a result, itorganizes existing data in a new wayand leads to a novel set of predictions.

DIFFERENCES BETWEEN THELIFE-HISTORY TRAITS OF

HUNTER-GATHERERS ANDCHIMPANZEES

Mortality, Fertility, and Growth

Figure 1 shows the differences be-tween the life spans of traditional hu-

man foragers and chimpanzees; Table1 compares a variety of life-historytraits of the two species. The hunter-gatherer data come from studies onpopulations during periods when theywere almost completely dependent onwild foods, having little modern tech-nology and no firearms, no significantoutside interference in interpersonalviolence or fertility rates, and no sig-nificant access to modern medicine.The chimpanzee data are compiledfrom all published and unpublishedsources that we are aware of. Becauseof small sample sizes at individualsites, mortality data were combined tocreate a single synthetic life table andsurvival function that combines alldata for wild chimpanzees.8

The data suggest that hunter-gath-erer children have a higher rate of sur-vival to age 15 (60% versus 35%) andhigher growth rates during the first 5years of life (2.6 kg/yr versus 1.6 kg/yr)than do juvenile chimpanzees. Chim-panzees, however, grow faster betweenages 5 and 10, both in absolute weightgain (2.5 kg/yr for chimps versus 2.1kg/yr for humans) and proportionalweight gain (16% per year for chimpsversus 10% per year for humans) (Table1). The early higher weight gain for hu-mans may be due to an earlier weaningage (approximately 2.5 years for hunt-er-gatherers versus 5 years for chim-panzees) and parental provisioning of

Figure 1. Survival curves for forager populations were derived from sources listed in notes fortable 1. Chimpanzee mortality is from a synthetic life table combining all mortality data fromBossou, Gombe, Kibale, Mahale and Tai.8

Despite the fact that thehuman juvenile andadult periods are longerthan those ofchimpanzees and thathuman infants are largerthan chimpanzee infantsat birth (about 3 kgversus 2 kg), hunter-gatherer womencharacteristically havehigher fertility than dochimpanzee females.

ARTICLE Evolutionary Anthropology 157

highly processed foods. Among hu-mans, the slow growth during middlechildhood is intriguing. According tothe allometric growth law, mammaliangrowth can be described by the equa-tion dw/dt 5 Aw0.75 (where change inweight per unit of time is expressed as afunction of a growth constant, A, andweight, w, to the 0.75 power). Mostmammals show a yearly growth con-stant of about 1, whereas the mean pri-mate value for A is about 0.4.9 Hunter-gatherer children between the ages of 5to 10 years are characterized by ex-

tremely slow growth, with A being ap-proximately 0.2.

Chimpanzees spend less time as ju-veniles than humans do: Femalechimpanzees give birth for the firsttime about 5 years earlier than dohunter-gatherer women. In naturalhabitats, chimpanzees also have amuch shorter adult life span than hu-mans do. At age 15, chimpanzee lifeexpectancy is an additional 15 years,as compared to 39 more years for hu-man foragers. Importantly, womenspend more than a third of their adult

life in a postreproductive phase,whereas very few chimpanzee femalessurvive to the postreproductive phase.The differences in overall survival andlife span are striking (Fig. 1). Lessthan 10% of chimpanzees survive toage 40, but more than 15% of hunter-gatherers survive to age 70. These nat-uralistic observations are also consis-tent with data on maximum life spans.The maximum life span of humans isbetween 100 and 120 years, depend-ing upon how it is calculated, which isabout two times longer than the max-

TABLE 1. LIFE HISTORY PARAMETERS OF HUMAN HUNTER-GATHERERS AND CHIMPANZEES

Group

Probabilityof Survivalto Age 15

ExpectedAge ofDeath at15 (years)

Mean Age atFirstReproduction(years)

Mean Age atLastReproductionb

(years)

InterbirthIntervala

(months)

MeanWeightAge 5 (kg)

MeanWeightAge 10 (kg)

Humans

Ache femaled 0.61 58.3 19.5 42.1 37.6 15.7 25.9Ache male 0.71 51.8 15.5 27Hadza femalee 0.58 54.7 15.5 20Hadza male 0.55 52.4 14.2 21.2Hiwi femalef 0.58 51.3 20.5 37.8 45.1 18 29.8Hiwi male 0.58 51.3 16.4 33.6!Kung femaleg 0.6 56.5 19.2 37 41.3 14 19.5!Kung male 0.56 56.5 16 22.5Forager meanc 0.60 54.1 19.7 39.0 41.3 15.7 24.9

Chimpanzees

Bossou femaleh 51Bossou maleGombe

femalei0.545 32.7 14.1 64.6 10 21

Gombe male 0.439 28.6 10 24Kibale femalej 0.805 35.6 68Kibale male 0.408 40.6Mahale

femalek14.6 72

Mahale maleTai female 0.193 23.8 14.3 69.1Tai malel 0.094 24Chimpanzee

mean0.35 29.7 14.3 27.7** 66.7 10 22.5

a Mean interbirth interval following a surviving infant.b Age of last reproduction for chimpanzee females was estimated as two years prior to the mean adult life expectancy.c The forager mean values were calculated by weighting each forager study equally. The chimpanzee mean mortality is from a

synthetic life table using data from all five sites listed.8,138

d Ache: Demographic and weight data from Hill and Hurtado.10

e Hadza: Demographic data from Blurton Jones and colleagues.16 Weight data from Blurton Jones (personal communication).f Hiwi: Demographic data from Hill and Hurtado unpublished database collected on the Hiwi foragers from reproductive-history

interviews conducted between 1982 and 1991 using the same methodology published in Hill and Hurtado.10

g !Kung: Demographic and weight data from Howell.83

h Bossou: Data from Sugiyama.139

i Gombe: Data on mortality from Hill and coworkers,8 and Pusey and Williams (personal communication). Gombe data on fertilityfrom Pusey,140 Tutin,141 and Wallis.142 Weights from Pusey.140

j Kibale: All data from Wrangham (personal communication). Mortality data in Hill and coworkers.8k Mahale: Data from Nishida, Takasaki, and Takahata.143

l Tai: Data from Boesch and Boesch.20

158 Evolutionary Anthropology ARTICLE

imum adult chimpanzee life span (ap-proximately 60 years for captive pop-ulations).

Despite the fact that the human ju-

venile and adult periods are longerthan those of chimpanzees and thathuman infants are larger than chim-panzee infants at birth (about 3 kg

versus 2 kg), hunter-gatherer womencharacteristically have higher fertilitythan do chimpanzee females. Themean interbirth interval between off-spring when the first survives to thebirth of the second is more than 1.5times longer among wild chimpan-zees than among modern hunter-gatherer populations. These numberslead to an interesting paradox. Lifetables from modern human foragersalways imply positive growth (see Hilland Hurtado,10 chapter 14), whereasthe chimpanzee numbers presentedhere imply slightly negative popula-tion growth rates. Chimpanzee nega-tive population growth may be a real

feature of recent habitat destructionand other human intrusion, or “natu-ral” mortality rates may have beenoverestimated due to the inclusion ofdeaths from viral epidemics such asebola and polio (see Hill and cowork-ers8 for a discussion).

To summarize, hunter-gatherershave a juvenile period that is 1.4 timeslonger than that of chimpanzees and amean adult life span that is 2.5 timeslonger than that of chimpanzees. Theyshow higher survival at all ages afterweaning, but lower growth rates dur-ing middle childhood. Despite alonger juvenile period, slower growth,

Figure 2. Daily energy acquisition data are recorded by individual among the Ache andHiwi. Thus, the age and sex of each acquirer is known for every day sampled. Meanproduction for 5- or 10-year age intervals (y value) was calculated from raw data bysumming all calories produced over the sample period by individuals in that age-sex classand dividing by the total sample of person days monitored for individuals in that category.This was plotted at the mean age of person days sampled (x value) in the categoryanalyzed. Hadza production levels are given for various juvenile age categories, for alladult men combined (no age breakdown), and for all reproductive women and all womenof postreproductive age combined (no age breakdown). All values are calculated asdescribed in the notes for Tables 2 and 3.

Adult men acquiremuch more food thando those in any otherage-sex category.Although the patterns formen seem consistent forall three societies,Hadza children andpostreproductivewomen appear toacquire substantiallymore food than do theirAche and Hiwicounterparts.

ARTICLE Evolutionary Anthropology 159

and a longer life span, hunter-gath-erer women achieve higher fertilityrates than do chimpanzee females.

The Age and Sex Profile ofEnergy Acquisition

Data on food acquisition by age andsex category exist for only three mod-ern foraging populations. Ache andHiwi food production was directlymonitored by weighing all food pro-duced by those in different age andsex categories throughout mostmonths of various years. (See Hill andcoworkers11 and Hurtado and Hill12

for definitions, methodology, andsampling plan). Hadza women’s andchildren’s plant-food acquisition wasestimated indirectly from samples ofin-patch return rates for different fruitand root resources over various age orsex classes during part of the wet sea-son and part of the dry season of var-ious years. (For details, see Hawkesand coworkers,6 and Blurton Jones,Hawkes, and O’Connell13,14). Thesedata were combined with sample esti-mates of time spent foraging and fre-quency across days to estimate dailyfood acquisition.13,14 Hadza men’sfood acquisition from hunting wasmeasured directly by weighing alllarge game brought to camp.15

Although there is some cross-cul-tural variation, all three societiesshow similar patterns. Hunter-gath-erer children produce little food com-pared to adults (Fig. 2). In the latejuvenile period, daily food acquisitionrates rise dramatically, especially formales. These rates continue to in-crease until mid-adulthood for malesin all three groups and even longer forHadza and Hiwi females. Adult menacquire much more food than dothose in any other age-sex category.Although the patterns for men seemconsistent for all three societies,Hadza children and postreproductivewomen appear to acquire substan-tially more food than do their Acheand Hiwi counterparts. But total food-consumption estimates for the Hadzamay be unrealistically high, since thedata suggest per capita consumptionof about 3,400 calories per day (Tables2, 3). That is 126% of the mean dailycaloric consumption of the Ache, de-spite the fact that 10-year-old Hadza

Figure 3. The mean expected daily energy consumption per individual for each age-sexcategory of each foraging group was estimated by first multiplying all the age-sex specificproduction rates for that foraging group times the proportional representation of thatage-sex category as expected from the survival curves and summing expected productionacross all age categories. This total expected production for the group was then divided bythe expected total number of individuals (as determined by the survival curve) at all agestimes their proportion of a standard consumer. This procedure assumes that all populationsare in steady state and that the proportional representation of each age category isdetermined by the probability of surviving to that age, and gives consumption per standardconsumer in each group. This gives the mean consumption for a standard consumer in thegroup. Daily expected consumption for individuals in various age-sex categories is esti-mated by multiplying the proportion of a consumer represented by each age class timesthe mean consumption of a standard consumer. Kaplan17 provides a detailed descriptionof these calculations and how proportional standard consumers were determined for eachage-sex category. A standard consumption rate of 1 is assigned to young adult males andfemales; children begin at a consumption level of 0.3 that of a standard consumer. Dailyenergy acquisition for age-sex categories is calculated as described for Figure 2 andaveraged across the Ache, Hiwi, and Hadza, weighting each group equally.

160 Evolutionary Anthropology ARTICLE

children weigh only 78% of the weightof 10-year-old Ache children, andadult Hadza women weight only 89%of the weight of Ache women. Thoseestimates are derived by assuming anage-structure consistent with theHadza life-table.16 If the dependencyratio in the camps studied by theHadza researchers was greater thanexpected by the life table, as BlurtonJones (personal communication) be-lieves this to be the case, the per-cap-ita consumption estimates would bereduced accordingly and would bemore realistic.

Figure 3 shows the mean daily en-ergy consumption and acquisitionrates for all three hunter-gatherer so-cieties as compared to the rates forchimpanzees of the same age and sex.The food-consumption rates of for-ager children and adults is estimatedfrom body weight and total group pro-duction.17 Chimpanzee energy acqui-sition, while not measured directly,can be estimated from body size andcaloric requirements, since very littlefood is transferred between age-sexcategories after weaning. Daily foodacquisition and consumption are vir-tually the same for chimpanzees fromthe juvenile period onward. The hu-man consumption-acquisition profileis strikingly different from that ofchimpanzees, with chimpanzee juve-niles acquiring considerably more en-ergy than forager children do untilabout the age of sexual maturity. Nochildren in any forager society pro-duced as much as they consumed un-til they reached their mid- to lateteens. Thus, human juveniles, unlikechimpanzee juveniles, have an evolu-tionary history of dependency onadults to provide their daily energyneeds. This can be appreciated by re-alizing that by age 15 the children inour forager sample had consumedover 25% of their expected life-timeenergy consumption but had acquiredless than 5% of their life-time energyacquisition.

The area in Figure 3 where foodacquisition is greater than consump-tion (where the solid line for eachspecies is above the dotted line) rep-resents surplus energy provided dur-ing the later part of the life span.These averaged data imply thathunter-gatherer men provide most

of the energy surplus that is used tosubsidize juveniles and reproduc-tive-aged women. Although based onaveraging only three societies, thistrend can be confirmed by compar-ing the food-acquisition rates ofadult males and females from a sam-ple of ten hunter-gatherer societiesin which food acquisition has beenmeasured with a systematic sample(Table 2).

A THEORETICAL TEST: WOULDNATURAL SELECTION ACTUALLY

PRODUCE THE CO-EVOLUTIONARY EFFECTS

PROPOSED BY THE THEORY?

Our proposal is that the shift to cal-orie-dense, large-package, skill-inten-sive food resources (Fig. 4) is respon-sible for the unique evolutionarytrajectory of the genus Homo. The key

element in our theory is that this shiftproduced co-evolutionary selectionpressures, which, in turn, operated toproduce the extreme intelligence, longdevelopmental period, three-genera-tional system of resource flows, andexceptionally long adult life charac-teristic of our species. We envisiontwo important effects of the change infeeding niche.

First, a long developmental period,parental provisioning, and a largebrain are necessary foundations of theskill-intensive feeding niche, andtherefore are products of selection asa result of entry into that niche. Ourview is that human childhood is elon-gated by including a period of veryslow physical growth, during whichthe brain is growing, learning is rapid,and little work is done. This is fol-lowed by adolescence, during whichgrowth is accelerated so that the brainand body can function together in thefood quest. Early adulthood is a timefor vigorous work during which re-source acquisition rates increasethrough on-the-job training. Thus, in-vestment in this life history involvesthree important costs: low productiv-ity early in life, delayed reproduction,and a very expensive brain to growand maintain. The return from thoseinvestments is delayed, with ex-tremely high productivity occurring inthe middle and latter portions of theadult period. That return increaseswith lengthening of the adult life spanbecause the return is realized over agreater period of time. Second, wepropose that the shift in the humanfeeding niche operated directly tolower mortality rates because it in-creased food package size, which, inturn, favored food sharing, provision-ing, and larger group size. Another in-direct effect was that the added intel-ligence and use of tools associatedwith the feeding niche also loweredpredation rates.

Underlying our theory is the hy-pothesis that these two effects pro-duce co-evolutionary processes oflarge magnitude. Holding all else con-stant, ecological changes that increasethe benefits of a long developmentalperiod and a concomitant increase inlater adult productivity not only pro-duce selection pressures to delay theonset of reproduction, but also pro-

Our proposal is that theshift to calorie-dense,large-package, skill-intensive food resourcesis responsible for theunique evolutionarytrajectory of the genusHomo. The key elementin our theory is that thisshift produced co-evolutionary selectionpressures, which, in turn,operated to producethe extremeintelligence, longdevelopmental period,three-generationalsystem of resourceflows, and exceptionallylong adult lifecharacteristic of ourspecies.

ARTICLE Evolutionary Anthropology 161

TABLE 2. PRODUCTION OF ENERGY BY MEN AND WOMEN IN FORAGING SOCIETIES

Daily Adult Production in Caloriesa

% Total AdultCalories

% Total AdultProteinMeat Roots Fruits Other

Mean DailyTotal

Ongeb men 3919 0 0 81 4000 79.7 94.8women 0 968 1 52 1021 20.3 5.2

Anbarrac men 2662 0 0 79 2742 70.0 71.8women 301 337 157 379 1174 30.0 28.1

Arnhemd men 4570 0 0 8 4578 69.5 93.0women 0 1724 37 251 2012 30.5 7.0

Achee men 4947 0 6 636 5590 84.1 97.1women 32 0 47 976 1055 15.9 2.9

Nukakj men 3056 0 0 1500 4556 60.4 98.6women 0 0 2988 0 2988 39.6 1.4

Hiwig men 3211 2 121 156 3489 79.2 93.4women 38 713 83 82 916 20.8 6.6

!Kung1,h men 2247 974 3221 45.5 44.7women 0 348 348 3169 3864 54.5 55.3

!Kung2,i men 6409 6409 @50women

Gwif men 1612 800 0 0 2412 43.0 78.7women 0 0 0 3200 3200 57.0 21.3

Hadzak men 7248 0 0 841 8089 64.8 94.1women 0 3093 1304 0 4397 35.2 5.9

a Edible portion and caloric values were taken from individual studies when available. Otherwise we assumed vertebrate meatat 85% edible, the Ache measured average for animals, and used the following conventions for calories/100 g edible: mammals150; roots 150; fruits 70; fish 120. When not specified, protein was assumed at 20% by weight for meat and 2% for roots and fruits.

b Onge: Data come from Bose.144 We assumed that all food is produced by adults and that men and women make up equalpercentages of the reported population. Caloric values (p. 156) and edible portions are taken from Meehan.145 We assumedthat males got all pigs, turtles, fish, and honey, whereas females acquired all crabs, bivalves, and plant products. Total caloricintake seems very low, but the Onge are the smallest foragers in this sample and had very low fertility.

c Anbarra: Data come from Meehan145; diet is found in Tables 29–32. It is assumed that women collected 85% of shellfish (p. 125)and that men obtained only birds, fish, mammals, and some shellfish (p. 149). Total person days of consumption are in eachtable. Women’s production days come from Table 27. We assumed an equal number of production days for men.

d Arnhem: Arnhem land data are from McArthur150 (pp. 127–128 and p. 138). It is assumed that adults acquired all food, that menobtained only vertebrate meat and honey, and that women acquired all other resources.

e Ache: Data come from all observed foraging trips between 1980 and 1996 on which KH, HK, and/or MH were present. Data priorto 1984 were published in Hill and coworkers.11 Subsequent data come from forest trips between 3 and 15 days long when nearlyall foods consumed were acquired from the forest. All foods were weighed on site and the edible portion was calculated fromrefuse samples collected after consumption. Caloric values were determined as previously published. Total production of fruitswas estimated by multiplying measured collection rates for different age categories times the time spent collecting by eachindividual. We have made two important modifications of 1984 data because of new field measures: 1) We now estimate theedible portion of wild honeycomb to be only 35% by weight; 2) The edible portion of palm starch is estimated at only 6% byweight, with the caloric value of the edible portion being 3,920 cal/kg. These corrections and new production data havelowered previously published estimates of daily caloric intake.

f Gwi: Meat production per hunter day is averaged from Silberbauer’s146 one-year observations of a band including 20 men andTanaka’s147 180-day observations (p. 111) of 10 men. We estimate Silberbauer’s band to contain 20 men and 24 womenbecause there were 80 individuals, 46.5% of whom were male and 55% were adult (p. 286, 287). For live weight meat, we assume85% edible weight containing 1,500 cal/kg. Plant production for adult women is estimated at the observed per-capitaconsumption, 800 g/consumer day times 80/24 (the ratio of the total population to adult women) times 1,500 cal/kg raw plant,times 80% collected by women147 (p. 70). Men are assumed to have produced 20% of the plant calories. Meat consumption percapita is the average from Tanaka147 (p. 70) and Silberbauer146 (p. 446). We assumed that Tanaka’s raw weights are 85% edible;we also assume 1,500 cal/kg edible meat for both studies. Plant consumption is reported to be 800 g/person in both studies147

(p. 70),146 (p. 199). We assume that this is equally split between roots and melons, with a mean caloric value of 1,500 cal/kg rawweight. Man days hunting are reported for both studies, but calculations of the sample size of women’s production andper-capita consumption are not specified in either study.

g Hiwi: Data come from a sample of days between 1985 and 1988 when KH and MH resided with the Hiwi and weighed all foodproduced by a sample of camp members. Details of calculations of edible portion and food value are published in Hurtado andHill.12,148

h !Kung1: All data on adult production and per-capita consumption are from Lee42 (pp. 260–271). Women’s plant production(non-mongongo) was assumed to be evenly split between roots and fruits.

i !Kung2: Data are from Yellen43 as calculated in Hill149 (pp. 182–183). Edible portion and caloric value are the same as in Lee.42

Only hunting data are recorded. In order to estimate per-capita consumption, adult men and women are assumed to compriseequal percentages of the band members. The percentage of the diet from meat is calculating assuming total consumption of2,355 calories per person day, as per Lee.42

162 Evolutionary Anthropology ARTICLE

duce selection pressures to investmore in survival during both the juve-nile and adult periods. At the sametime, ecological changes that lowermortality rates during the juvenileand adult periods also produce selec-tion pressures that favor a longer ju-venile period if it results in higheradult productivity. If both types ofchange occur (increased payoffs fortime spent in development and lowermortality rates), great changes in bothmortality rates and time spent in de-velopment may result. Furthermore, ifthose changes are accompanied bylarge increases in productivity afteradulthood is reached, we expect addi-tional increases in time spent in devel-opment and in survival rates. Our pro-posal is that the skill-intensive feedingniche, coupled with a large brain, isassociated with a significant amountof learning during the adult period.

To test this hypothesis, we devel-oped a model to determine whether ornot natural selection would actuallyresult in co-evolution of the develop-mental period and the life span. Thismodel builds on two bodies of theory,life-history theory in biology and hu-man-capital theory in economics.Life-history theory is based on thepremise that organisms face trade-offs in the allocation of their time andeffort. Gadgil and Bossert2 offered the

first explicit treatment of allocationstrade-offs with respect to reproduc-tion and longevity. They postulatedthat during the life course selectionacts on the allocation of energy to

each of three competing functions: re-production, maintenance, and growth.Energy allocated to reproduction willnecessarily reduce the quantity avail-able for maintenance and growth.

Maintenance and growth may be seenas investments in future reproduction,for they affect both the probabilitythat an organism will survive to repro-duce in the future and the amount ofenergy it will be able to harvest andtransform into reproduction. Thus,one fundamental trade-off is betweencurrent and future reproduction.

Human-capital theory in econom-ics18,19 is designed to analyze invest-ments in education and trainingthrough the course of life. Central tothis theory is the notion of foregoneearnings: Time spent in education andtraining reduces current earnings inreturn for increased earnings in thefuture. The economic trade-off be-tween current and future earnings isdirectly analogous to the trade-off be-tween current and future reproduc-tion in biology.

Charnov,1,9 building on earlier workon optimal age at first reproduction,developed a mathematical model ofthe trade-off between growth and re-production for mammals. His modelis designed to capture determinategrowth, in which an organism has twolife-history phases after attaining in-dependence from its parents. Theseare a prereproductive growth phase inwhich all excess energy, remaining af-ter maintenance requirements havebeen met, is allocated to growth and a

. . .we developed amodel to determinewhether or not naturalselection would actuallyresult in co-evolution ofthe developmentalperiod and the life span.This model builds on twobodies of theory, life-history theory in biologyand human-capitaltheory in economics.

TABLE 2. (CONTINUED)

j Nukak: Data come from Politis151 (chapter 4). We assume that all food was produced by adults and that men and women makeup equal percentages of the population. Edible portions and caloric values for foods come from similar Ache resources. Fruitsshow edible portions varying from 21% for fruits brought in and weighed with the stalk to 40% for fruits without the stalk collectedin baskets. Caloric values of fruits ranged from 600 cal/kg for sweet pulpy fruits to 1,430 cal/kg for oily palm fruits. Other resourceswere equivalent to common Ache and Hiwi resources.

k Hadza: Data on the daily caloric production of children are from Blurton Jones, Hawkes, and O’Connell.14 We assumed that 61%of the calories produced come from fruit and the remainder from roots, as for youngest girls.6 Daily production of women takenfrom Hawkes, O’Connell, and Blurton Jones.6 We multiplied in-patch rates by time foraging for each season (both in Table 1), andthe proportion of time in patch (60% root, 66% berry) (Hawkes personal communication and Hawkes, O’Connell, and BlurtonJones,6 p. 350), equally weighting production in dry and wet seasons. Ekwa roots are calculated as 88% edible (Hawkes, personalcommunication) and 850 cal/kg edible (Hawkes, O’Connell, and Blurton Jones,6 p. 691). Fruits are assumed to be 50% edible andhave a caloric value of 2,500 cal/kg edible42 (pp. 481, 484 for grewia sp. berries). Meat acquisition is 4.89 kg/day for adult men(over age 18) and assumed to have a caloric value of 1,500 cal/kg with the discounting for the edible portion.7 Honeyproduction was assumed to be 0.78 kg/man-day for males over age 1815 (p. 86), with 35% edible and 3,060 cal/kg (as for theAche). All food production by age and sex category was weighted by the probability of survival to that age for the Hadza,16

then divided by the total of all survival probabilities to obtain the expected per-capita consumption. The estimate of totalper-capita consumption is very high, probably in part because actually sampled camps contained more juvenile consumersthan the life table implies (Hawkes, personal communication). However, we cannot correct the estimate of daily consumptionwithout a complete age-sex breakdown of the sampled camps, which currently is not available.

l Chimpanzee diet: We use the Gombe diet from Goodall76 (Fig. 10.1). The absolute amount of meat in the diet is from Wranghamand Riss.22 Kibale plant percentages are from Wrangham, Conklin-Brittain, and Hunt.97 The Kibale meat percentage is fromWrangham and coworkers.152 The Mahale diet is taken from Hiraiwa-Hasegawa.25 The absolute amount of meat in the diet wascalculated from Uehara,153 assuming adult prey at 13 kg and juvenile prey at 6 kg, on average, the percentage of adult preywas taken from Stanford.154 For Tai forest chimpanzees, the absolute amount of meat in the diet was calculated from Boeschand Boesch20 (Table 7.4).

ARTICLE Evolutionary Anthropology 163

reproductive phase in which all excessenergy is allocated to reproduction.By growing, an organism increasesits energy capture rate, and thusincreases reproductive rate. Duringadulthood the fundamental trade-offis between the expected length of thereproductive span (which is shorterwith each additional year spent grow-ing, because of the increased proba-bility of dying before reproducing)and the adult reproductive rate (whichis higher with every year spent grow-ing because of increased energystored in the form of adult bodymass). The model predicts the amountof time mammals will grow beforeswitching to the reproductive phaseby selecting the time that maximizesexpected energy for reproduction overthe life course.

Here we extend Charnov’s model inthree ways. First, we broaden the con-cept of growth from body size alone toinclude all investments in develop-ment. Development can be seen as aprocess in which individuals and theirparents invest in a stock of embodiedcapital, a term that generalizes theconcept of human capital to all organ-isms. In a physical sense, embodiedcapital is organized somatic tissue. Ina functional sense, embodied capitalincludes strength, immune function,coordination, skill, knowledge, andsocial networks, all of which affect theprofitability of allocating time andother resources to alternative activi-ties such as resource acquisition, de-fense from predators and parasites,mating competition, parenting, andsocial dominance. Because such stockstend to depreciate with time due to

physical entropic forces and direct as-saults by parasites, predators, andconspecifics, allocations to mainte-nance efforts, such as feeding, cell re-pair, and vigilance, can also be seen asinvestments in embodied capital. Inour model, the energy capture rate in-creases with embodied capital (that is,time spent in development).

Second, in Charnov’s model, organ-isms have no control over mortalityrates, which are exogenously deter-mined; the only variable of choice forthe organism is age at first reproduc-tion, or time spent in development. Inour model, the organism can exercisecontrol over mortality rates by allocat-ing energy to mortality reduction. Weinclude this second choice variable todetermine if time invested in develop-ment and energy allocated to mortal-ity reduction co-evolve. Our theorypredicts that increased investment oftime in development due to the exploi-tation of difficult-to-acquire, high-quality resources selects for increasedlongevity or lower mortality rates.Third, our model allows learning tocontinue after physical growth hasceased to analyse the effects on theincrease in return rates from foragingduring the adult period (see Figs. 2and 3) as a consequence of “on-the-job” training.

Thus, the model has two choicevariables upon which selection canact: age at first reproduction, a proxyfor time spent in development andphysical growth, and allocation of en-ergy to lowering mortality. It also hasthree ecological parameters: factorsaffecting the pay-offs to investmentsin development, factors affecting mor-

tality rate, and the growth in produc-tivity after adulthood due to learning.The formal model is presented inBox 1.

The six main results of the mathe-matical model confirm the co-evolu-tionary selection pressures predictedby the theory. Ecological factors in-creasing the productivity of invest-ments in developmental embodiedcapital (in the context of the presenttheory, a skill-intensive foraging niche)increase both time spent as a juvenileand investments in mortality reduc-tion. Ecological factors that lowermortality rates increase both timespent in development and investmentin mortality reduction. Finally, thegreater the growth rate in productionduring the adult period, due to largebrains and “on-the-job” training, themore it pays to invest in developmentand mortality reduction. These resultsall show that investments in develop-ment and investments in mortality re-duction and longevity co-evolve.

EMPIRICAL TESTS OFTHE THEORY

Composition of the Diet

Figure 4 illustrates our proposalabout the differences between the di-ets of nonhuman primates and hu-mans. While the diets of nonhumanprimates vary considerably by speciesand by local ecology, the inverted tri-angle in Figure 4 represents thegreater importance of lower quality,easier-to-acquire foods in the diets ofmost nonhuman primates (excludinginsectivores). The upward-pointingtriangle in Figure 4 represents thegreater importance of large-package-size, nutrient-dense, difficult-to-ac-quire foods in human diets.

Table 3 presents data on the diets often foraging societies and four chim-panzee communities for which caloricproduction or time spent feeding weremonitored systematically. As far as weare aware, this is a complete sample ofthe available data. The diet is subdi-vided into vertebrates, roots, nuts andseeds, other plant parts (such asleaves, flowers, and pith) and inverte-brate resources. The diets of all mod-ern foragers differ considerably fromthat of chimpanzees. Measured in cal-ories, the major component of forager

Figure 4. The feeding ecology of humans and other primates.

164 Evolutionary Anthropology ARTICLE

Box 1. A Formal Model of Natural Selection on Age at First Reproduction andInvestments in Mortality Reduction

The model treats two phases of the life course, the juvenile andadult periods. The juvenile period begins after the high mortalityphase associated with infancy and weaning, is dedicated togrowth and development, and lasts a variable amount of time, t,upon which selection acts. During this time, all energy is investedin either embodied capital (growth and learning) affecting futureenergy production, P, or in reducing mortality rate, m. The twochoice variables during the juvenile period are its length, t, and theproportion of energy invested in mortality reduction, l (implyingthat {1 2 l} is the proportion allocated to growth and learning).The adult production of energy at the end of the juvenile period,Pa, is determined by t, l, and the combined ecological effects ofthe environment and the technology of production, captured bythe vector, «, which is assumed to increase energy production(i.e., ­Pa/]« . 0). We can think of this as composed of twofunctions, P(t, «), which captures the growth in production due tolearning, growth, and development, and C(l), which representsthe proportional loss in production due to investments in mortalityreduction during the juvenile period. Thus, we have Pa 5 p(t, l;«) 5 P(t, «)C(l).

The second period is reproductive. During this period, growth inbody size ceases and all excess energy is allocated to reproduc-tion. Production grows at some constant rate, g, due to the effectsof learning. Thus, production at some age, x, after adulthood, Px,is Pae

g(x2t).During both phases, the instantaneous mortality rate, m, re-

mains constant, at a level determined by the amount of energyproduction diverted to mortality reduction, l, and by ecologyfactors affecting mortality (such as density of predators and dis-eases), u, which is assumed to increase mortality rates (i.e., ]m/­u . 0). Thus, m 5 m(l, u). The net energy available for repro-duction at age x during the adult period, Pr,x, will be equal to totalenergy production times the proportion allocated to reproduction,Pr,x 5 (1 2 l)Pae

g(x2t).The basic logic of the model is that an organism will maximize

fitness by maximizing its lifetime energy allocated to reproduction.By increasing the length of the growth and development phase, t,the adult rate of energy capture increases, but the expected lengthof the reproductive period, R, decreases. This decrease resultsfrom the fact no energy is allocated to reproduction during thegrowth and development phase, though the organism is still ex-posed to mortality. Allocations to mortality reduction also haveopposing effects. Allocations to growth during the juvenile periodand to reproduction during the adult period are reduced by allo-cations to mortality reduction. Yet lowered mortality increases thelength of the reproductive period by increasing the probability ofreaching reproductive age and by increasing the expected timefrom reproductive age to death. Thus, R is determined by both tand l. If we consider only individuals who survive infancy, R isequal to expectation of adult reproductive years lived, given theprobability of dying at each age, x. Thus, the expected number ofadult years during which energy is allocated to reproduction is R 5r(t, m(l, u)) 5 *x5t

` (x 2 t)me2mxdx 5 1/me2mt, where 1/m is theexpected adult life span conditional on reaching age t and e 2 mt isthe probability of reaching age t, conditional on having survivedinfancy. The expected energy production during the adult period ise 2 mt(1 2 l)Pa *x5t

` e(g2m)(x2t)dx, which is equal to e 2 mt(1 2l)Pa(m 2 g)21 for m . g (which we assume to hold true, becauseexpected lifetime income otherwise would be infinite).

Selection is expected to optimize t and l so as to maximizelifetime energy allocated to reproduction. Thus, we have the fol-lowing maximization problem:

Maxt,l,m

W 5 e 2 m~l,u!t@1 2 l#Pa~t, l, «!~m~l, u ! 2 g! 2 1. (1)

Partially differentiating the fitness function, W, with respect to tand l, respectively, the following first-order conditions for anoptimum are obtained:

­Pa

­tPa

2 1 5 m (2)

and

2­m

­l~t 1 ~m 2 g! 2 1! 5 2

­P­l

Pa2 1 1 ~1 2 l! 2 1 (3)

Equation 2 for optimal t, taking m as given, replicates Char-nov’s1 result. Optimal t occurs at the age when the proportionalincrease in adult production due to a small increase in timespent in development (the left hand side of equation 2) is equalto the proportional loss in the probability of reaching adulthood(the right-hand side). The benefits and costs are measured interms of proportions because fitness is a product of the prob-ability of reaching adulthood and the production rate as anadult.

Equation 3 concerns optimal allocations to mortality reduction,given time spent in development. The left-hand side of the equa-tion is the benefit of a small increase in investment in mortalityreduction. It is the proportional increase in reaching adulthood,plus the proportional increase in the adult lifespan, adjusted forthe growth in income due to learning, g. The right-hand side is thecost, the proportional loss in production due to increased invest-ment in mortality reduction. This proportional cost is two-foldbecause allocations during the juvenile period reduce the growthrate and therefore reduce Pa (the first term), while allocations tomortality reduction during the adult period reduce the propor-tion of adult production allocated to reproduction (the secondterm).

Differentiating equations 2 and 3 with respect to the ecologicalparameters, «, u, and g, we derive our six main analytical results,confirming the co-evolutionary effects predicted by our theory. ­t/­«

and ]l/]« are both positive. This means that ecological factors in-creasing the productivity of investments in developmental embodiedcapital, as indexed by «, not only increase time spent in developmentbut also increase investments in mortality reduction. ­t/­u and ]l/]uare both negative, meaning that exogenous or extrinsic increases inmortality, as indexed by u, reduce both time spent in developmentand investments in mortality reduction. Conversely, ecological fac-tors that lower mortality rates increase both time spent in develop-ment and investment in mortality reduction). Finally, ­t/­g and ­l/­gare both positive, showing that the greater the growth rate in pro-duction during the adult period due to “on-the-job” training, as in-dexed by g, the more it pays to invest in development and mortalityreduction. (A formal proof of the six results was developed by ArthurRobson and is available from the authors). These results all show thatinvestments in development and investments in mortality reduction/longevity co-evolve.

ARTICLE Evolutionary Anthropology 165

TABL

E3.

DIE

TO

FM

OD

ERN

HU

NTE

R-G

ATH

ERER

SA

ND

CH

IMPA

NZE

ESa

Hun

ter-

Ga

the

rers

Chi

mp

anz

ee

s

Ong

eA

nba

rra

Arn

hem

Ac

heN

uka

kH

iwi

!Kun

g1

!Kun

g2

Gw

iH

ad

zaG

om

be

Kib

ale

Ma

hale

Tai

Sam

ple

b

kg/p

ers

on

da

yc12

5636

5427

636

4594

147

5686

692

8?

?Sa

mp

leb

kg/p

ers

on

da

yc

me

at

0.59

0.58

1.34

1.36

0.62

0.97

0.26

0.59

0.30

1.10

me

at

0.03

0.01

0.04

roo

ts0.

210.

070.

430.

000.

000.

440.

150.

40?

1.62

roo

tsse

ed

s,n

uts

0.00

0.00

0.00

0.00

0.00

0.00

0.21

0.00

0.00

see

ds,

nu

tsfr

uits

0.00

0.15

0.01

0.61

0.86

0.29

0.15

0.40

?0.

54fr

uits

oth

er

pla

nts

0.00

0.00

0.00

0.96

0.00

0.04

0.00

0.00

0.00

oth

er

pla

nts

inve

rte

bra

te0.

011.

010.

090.

110.

350.

020.

000.

000.

24in

vert

eb

rate

ca

lorie

s/p

ers

on

da

ym

ea

t98

082

218

2121

2676

413

5069

016

0241

719

40ro

ots

242

9345

60

026

815

060

0?12

14se

ed

s,n

uts

00

00

00

1365

00

fru

its0

4410

2274

782

150

600?

621

oth

er

pla

nt

00

325

50

360

00

inve

rte

bra

te20

127

6730

837

557

00

255

Tota

l12

4310

8523

5727

1218

8617

9323

5516

17?

4030

No

nfo

rag

ed

d0

1116

0tr

ac

e37

862

6tr

ac

etr

ac

e0

tra

ce

Die

tary

pe

rce

nta

ge

off

ora

ge

dfo

od

s

Fee

din

gtim

ep

erc

ent

ag

eG

om

be

Kib

ale

Ma

hale

Lop

e

me

at

7975

7778

4175

2968

?26

?48

me

at

1.5

0.9

2.5

2.1

roo

ts19

819

00

156

37?

30ro

ots

0.0

0.1

01.

4se

ed

s,n

uts

00

00

00

580

0se

ed

s,n

uts

5.1

00

7fr

uits

04

01

405

637

?15

fru

its60

.278

.557

.767

.6o

the

rp

lan

t0

00

90

20

00

oth

er

pla

nt

29.3

21.3

33.4

15.5

inve

rte

bra

te2

123

1120

30

06

inve

rte

bra

te3.

90

6.4

5.6

co

llec

ted

0.0

4.0

0.6

0.8

20?

4.6

4.9

?37

?15

co

llec

ted

94.2

99.1

91.1

92.3

ext

rac

ted

21.9

20.3

30.1

24.3

40?

21.6

63.4

?37

?36

ext

rac

ted

3.8

06.

45.

6h

un

ted

78.0

75.7

69.4

74.9

4073

.731

.768

.026

.048

hu

nte

d2

0.9

2.5

2.1

aFo

rin

form

atio

no

nm

eth

od

so

fc

alc

ula

tion

an

dd

ata

sou

rce

s,se

efo

otn

ote

s,Ta

ble

2.b

Pe

rso

nd

ays

sam

ple

d,

inc

lud

ing

all

me

nw

om

en

an

dc

hild

ren

as

eq

ua

lco

nsu

me

rs.

cTh

isis

the

we

igh

to

fth

ee

dib

lep

ort

ion

for

me

at

an

dfie

ldw

eig

ht

for

all

oth

er

reso

urc

es.

dIn

take

of

no

nfo

rag

ed

foo

ds

inc

alo

ries/

pe

rso

nd

ay

wh

en

me

asu

red

or

rep

ort

ed

.

166 Evolutionary Anthropology ARTICLE

diets is vertebrate meat. This rangesfrom about 30% to 80% of the diet inthe sampled societies, with most dietsconsisting of more than 50% verte-brate meat (equally weighted mean 560%). The emphasis on vertebratemeat would be even more clear if anyhigh-latitude foraging societies wereincluded in the sample. In contrast,chimpanzees spend only about 2% oftheir feeding time eating meat. Unfor-tunately, the diet of wild primates isnot usually expressed in calories, as isthat of human foragers. Field workersstudying nonhuman primates usetime spent feeding on specific foods asthe closest approximation of energyacquired and rarely either measure in-gestion rate or calculate calorie in-take. The absolute intake of meat perday also varies tremendously: Chim-panzee per capita meat intake is esti-mated at about 10 to 40 g per day,while human meat intake ranges fromabout 270 to 1,400 g per person perday. Although it is true that chimpan-zee males eat much more meat thando females and juveniles,20–22 we con-clude that, in general, members of for-aging societies eat more than tentimes as much meat as do chimpan-zees.

The next most important food cate-gory in our forager sample is roots,which make up an average of about15% of the energy in the diet and wereimportant in about half the societiesin our sample. In contrast, the chim-panzee diet is primarily composed ofripe fruit, which accounts for over60% of feeding time. Only two forag-ing societies ate large amounts of ripefruit, the Gwi San of the Kalaharidesert, who consume melons for wa-ter and nutrients during much of theyear, and the Nukak of Colombia, whoextensively exploit tropical palmfruits, which, however, are difficult toacquire. Other plant products are animportant secondary food for chim-panzees, making up about 25% of ob-served feeding time. This category isunimportant for the foragers in oursample.

The data suggest that humans spe-cialize in rare but nutrient-dense re-source packages or patches (meat,roots, and nuts), whereas chimpan-zees specialize in ripe fruit and plantparts with low nutrient density. These

differences in the nutrient density offoods ingested are also reflected in hu-man and chimpanzee gut morphologyand food passage time. The chimpan-zees gut is specialized for rapid pro-cessing of large quantities and low-nutrient, bulky, fibrous meals.23

However, a stronger contrast is appar-ent when we consider how the re-sources are obtained. We have catego-rized all foods into three types.Collected foods are those that can be

obtained and eaten simply by gather-ing them from the environment. Ex-tracted foods are non mobile but areembedded in a protective contextfrom which they must be removed.Such foods may be underground, inhard shells or associated with toxins.Hunted foods include mobile re-sources that must also be extractedand processed before consumption.Collected resources include fruits,leaves, flowers, and other easily acces-

sible plant parts. Extracted resourcesinclude roots, nuts and seeds, mostinvertebrate products, and plant partsthat are difficult to extract, such aspalm fiber or growing shoots. Huntedresources include all vertebrates andsome mobile invertebrates.

Table 3 shows a breakdown of for-ager and chimpanzee foods accordingto our three acquisition categories.Chimpanzees obtain an average ofabout 95% of their diet from collectedfoods, whereas the foragers in oursample obtain an average of 8% oftheir food energy from collected re-sources. On the other hand, foragersobtain about 60% of their food energyfrom hunted resources and about 32%from extracted resources, whereaschimpanzees obtain about 2% of theirfood energy from hunted foods andabout 3% from extracted resources.While these categories may be some-what rough, it is clear that humansare much more dependent on re-sources that can be obtained only bycomplicated techniques. Thus, the di-etary data are consistent with our the-oretical model. Humans appear to bemore dependent on resources that re-quire skill and learning to acquire.

The Age Profile of Acquisitionfor Collected and ExtractedResources

The proposition that difficulty of ac-quisition predicts the age profile offood production can be tested in twoways, by looking at the daily amountof different resource types producedby individuals of different age and byobservational and experimental mea-sures of hourly rates of acquisition ofdifferent resource types by individualsof different ages. The daily data aredetermined by both time allocationand rates of return per unit of timespent on a resource, whereas thehourly data are based exclusively onrates of return. Both sources of datasupport the proposition that juvenilescannot easily obtain extracted andhunted foods.

Beginning with the daily data, Fig-ures 5 and 6 show, respectively, thedaily caloric contribution of variousfood types acquired by Ache and Hiwimales and females as a function ofage. The upper panels represent the

The absolute intake ofmeat per day alsovaries tremendously:Chimpanzee per capitameat intake is estimatedat about 10 to 40 g perday, while human meatintake ranges fromabout 270 to 1,400 g perperson per day.Although it is true thatchimpanzee males eatmuch more meat thando females andjuveniles, we concludethat, in general,members of foragingsocieties eat more thanten times as much meatas do chimpanzees.

ARTICLE Evolutionary Anthropology 167

calorically less important foods (thosefor which daily production is less than200 kcal/day); the lower panels repre-sent the more important foods (.200kcal/day). Among young Ache, bothmales and females acquire only fruits.Both sexes reach their peak rates ofdaily fruit acquisition in their mid- tolate teens. Extracting palm hearts re-quires strength and some skill (aboutthree minutes of chopping in the rightspot); daily palm heart acquisition isasymptotic for both sexes by the ageof 20 years. More skill and learningare required to extract honey or palmstarch (knowing how to open a “win-dow” to the resource and then extractit). The daily acquisition of these re-sources does not peak until individu-als are in their late 20s. Daily returnsfrom hunting do not peak until indi-viduals are about 35 years old.

Hiwi foragers show similar produc-tion patterns with age, except thatdaily fruit acquisition becomes as-

ymptotic at later ages. Hiwi fruit col-lection is more complicated becausemany trips entail walking through thenight to distant groves, followed by areturn trip of 10 to 20 km with a heavyload of fruit. Daily honey extractionrate also reaches its peak level whenthese foragers are in their early 30s(these are very small nests of nativebees). Female root-production ratesincrease four-fold from age 20 to age40, but men’s meat production doesnot peak until they are about age 35.

The fact that forager children, likechimpanzees, primarily acquire ripefruits is supported by additional data.Among the Ache, children acquire fivetimes as many calories per day duringthe fruit season as they do duringother seasons of the year.24 Amongthe Hadza, adolescent girls acquire1,650 calories per day during the wetseason, when fruits were available,and only 610 calories per day duringthe dry season, when fruits are not

available. If we weight the data for thewet and dry season equally, teenageHadza girls acquire 53% of their cal-ories from fruits, compared to 37%and 19%, respectively, for reproduc-tive-aged women and postreproduc-tive women (all calculated fromHawkes and coworkers.6) Hadza boys,like Ache and Hiwi boys, switch fromeasier tasks, such as fruit collection,shallow tuber extraction, and baobabprocessing to honey extraction andhunting in their mid- to late teens.13,14

Chimpanzee juveniles also focus onmore easily acquired resources thando adult chimpanzees. Juvenile chim-panzees practice difficult extractionactivities such as collecting termites,fishing for ants, or nut cracking lessthan adults do.20,26 Hunting is strictlyan adult or subadult activity.20,21,27

Hourly return rates provide furtherevidence that important human foodresources require long periods oflearning and skill development. Ob-

Figure 5 and 6. Age-sex specific daily energy acquisition is calculated as described for Figure 2 and Tables 2 and 3. Mean daily acquisitionfor each resource class was calculated by summing all the calories acquired for individuals of that age-sex category and that resourceclass, then dividing by all person days sampled in the relevant age category. Y values are plotted at the mean age for each class analyzed.

168 Evolutionary Anthropology ARTICLE

servations of Ache fruit collectionshow that foragers generally acquirethe maximum observed rate by aboutage 20 (Fig. 7). Some fruits (for exam-ple, pretylla) that are simply pickedfrom the ground are collected by chil-dren as young as one-and-a-half tothree years at 30% of the adult maxi-mum rate. For fruits such as vijulla,which must be picked off branches,children do not reach 50% of the adultmaximum rate until age 15 (Fig. 7). Asmentioned earlier, fruit collection bythe Hiwi (Fig. 8), unlike that done oth-ers, is labor-intensive and requirestravel to distant food sites. Both malesand females reach maximum returnrates by about age 25.

Hadza data also show competentfruit collection by children. TheKongoro berry collection rate ofyoung married girls is equal to theadult women’s rate.6 Baobab collect-ing and processing seems to reach50% of the adult rate by about age12.14 Baobabs are an interestingfood resource because they are bothcollected and extracted. While theycan easily be picked off the ground,much more food energy is obtainedwhen the pith is extracted with pound-

ing and water. Young children do notpractice these activities, whereas olderchildren do.14

In contrast to the hourly acquisitionrate of fruits, that of extracted re-sources often increases through earlyadulthood as foragers acquire neces-sary skills. Data on Hiwi women showthat their root-acquisition rates do notbecome asymptotic until the womenare about age 35 to 45 years (Fig. 8).The root-acquisition rate of 10-year-old girls is only 15% of the adult max-imum. For Hiwi males, the honey-ex-traction rates peak at about age 25.Again, the extraction rate of 10-year-olds is less than 10% of the adult max-imum. Experiments done with Achewomen and girls clearly show that theyoung adults are not capable of ex-tracting palm products at the rate ob-tained by older Ache women (Fig. 9).Girls take longer than women to cutpalms because they lack strength andbecause they cannot judge whether apalm will fall to the ground or getstuck in nearby tree branches. Girlstake longer than women to extract thegrowing shoot from the palm after itis on the ground because this task re-quires strength and knowing where to

cut across the palm leaf stalks. Girlstake longer to extract the starchy fiberfrom the trunk of a downed palm be-cause they do not know how to cutopen a window nor how most effi-ciently to pound the fiber away fromthe hard outer trunk wood. Whenthese components activities of palmextraction are combined, Ache womendo not reach peak return rates untiltheir early 20s (Fig. 9).

Supporting data are also availablefrom other groups. !Kung (Ju/’hoansi)children crack mongongo nuts at amuch slower rate than adults do.28

Bock29 has shown that nut-crackingrates among the neighboring Ham-bukushu do not peak until about age

35. Hadza women, however, appear toobtain maximum root-digging ratesby early adulthood,6 perhaps becausethey obtain a good deal of practicethroughout childhood13,14 and thusrequire only adult strength in order toproduce at adult rates.

Casual ethnographic observationsupports the generalization that fruitcollection is easily learned, extractionskills require more time to develop,and hunting is the most difficult for-aging behavior. Anthropologists work-ing with modern foragers often partic-ipate in fruit collection and canrapidly achieve aboriginal return rates.Some types of extraction can also be

Figure 7. All Ache data, including timed counts of fruits acquired per minute, were analyzedfor each age category. The mean plotted is the average from each independent moni-tored count for a given age class of acquirers. Fruits other than those that are simplycollected from the ground or low branches were not included in the analyses. Y values areplotted at the mean age for each class analyzed.

Hourly return ratesprovide further evidencethat important humanfood resources requirelong periods of learningand skill development.Observations of Achefruit collection show thatforagers generallyacquire the maximumobserved rate by aboutage 20.

ARTICLE Evolutionary Anthropology 169

mastered with practice. One of us(KH) chops down palms as fast asAche women do and can extract theheart at a slightly slower rate. How-ever, despite some practice, KH hasnot achieved the rate of palm-fiber ex-traction of Ache women. We know ofno ethnographers who successfullydig roots or crack nuts at the rate ofmembers of their study populations.

Indeed, the same patterns are seenamong captive animals released intothe wild. Animals that have grown toadulthood in captivity can be success-fully released into the wild if theirfeeding niche is simple (for example,if they are herbivores). However, thesuccess rate for carnivores and apeswith complex feeding strategies ismuch lower.

Hunting and the Role of Menin Human Reproduction

Male behavior plays a distinctiverole in human life histories in twoways. First, hunting, the primarysubsistence activity of adult men, isthe most learning-intensive foragingstrategy practiced by humans. Sec-ond, unlike most higher primates,men play a major role in the energet-ics of human reproduction.

Although a detailed quantitativeanalysis of the learning process in-volved in human hunting has not yetbeen conducted, it is clear that humanhunting differs qualitatively fromhunting by other animals. Unlikemost animals, which either sit andwait to ambush prey or use stealthand pursuit techniques, human hunt-ers use a wealth of information tomake context-specific decisions, bothduring the search phase of hunting

Figure 8. Because Hiwi foragers target specific resources when they leave their centralcamp, we recorded all time dedicated to foraging for different resource types over thesample period. Total energy production for each resource type and age-sex class wasdivided by the total number of hours reported to be out-of-camp foraging for that resourcetype to obtain the hourly return rate from foraging for that resource type. The time foragingincluded in-patch pursuit of resources as well as walking time to and from the patch. Yvalues are plotted at the mean age for each class analyzed.

We know of noethnographers whosuccessfully dig roots orcrack nuts at the rate ofmembers of their studypopulations. Indeed, thesame patterns are seenamong captive animalsreleased into the wild.Animals that havegrown to adulthood incaptivity can besuccessfully releasedinto the wild if theirfeeding niche is simple(for example, if they areherbivores). However,the success rate forcarnivores and apeswith complex feedingstrategies is much lower.

170 Evolutionary Anthropology ARTICLE

and then after prey is encountered.Specifically, information on ecology,seasonality, current weather, expectedanimal behavior, and fresh animalsigns are all integrated to form multi-variate mental models of encounterprobabilities that guide the searchand are continually updated as condi-tions change.30 Various alternativecourses of action are constantly com-pared and referenced to spatial andtemporal mental maps of resourceavailability.30 This information is col-lected, memorized, and processedover much larger spatial areas thanchimpanzees ever cover. For example,interviews with Ache men show thatfully adult men (those over the age of 35years) had hunted in nearly 12,000 km2

of tropical forest during their lives. Al-most all foragers surveyed use morethan 200 km2 in a single year, and manycover more than 1,000 km2 in a year(Table 4.1 in Kelly31). Male chimpan-zees, on the other hand, cover onlyabout 10 km2 in a lifetime.32,33

After potential prey are encoun-tered, humans employ a wide varietyof techniques to obtain them, usingastounding creativity. Here are just

some examples that Hill, Hurtado andKaplan have seen among the Ache,Hiwi, Machiguenga, and Yora: Arbo-real animals have been shot with ar-rows from the ground or a tree, drivenby climbing, shaken down frombranches, frightened into jumping tothe ground, brought down by felling atree with an axe, lured by imitatedcalls, lured by making captured in-fants emit distress calls, captured bythe spreading of sticky resin onbranches to trap them, and capturedby scaffolding constructed from treebranches and vines. Ground-dwellingprey are shot with arrows, driven toother hunters or capture devices, rundown upon encounter, slammed todeath against the ground, strangled atthe neck, or suffocated by stepping onthem while they are trapped in a tightspot. Burrowing prey are dug out,chopped out of tree trunks, stabbedthrough the ground with spears,frightened to the point at which theybolt from the burrow, smoked out,and captured by introducing a lassothrough a small hole. Aquatic prey areshot on the surface or below it, driveninto traps, poisoned, discovered on

muddy bottoms by systematicallypoking the bottom of a pond, andspeared underwater by randomthrusts in drying lakes. The widelyvaried kill techniques are tailored to awide variety of prey under a wide va-riety of conditions. Although allgroups probably specialize in themost abundant and vulnerable prey intheir area, the total array of speciestaken is impressive, and probably ismuch larger than that of most, if notall, other vertebrate predators. For ex-ample, from 1980 to 1996 our sampleof weighed prey taken by the Ache

included a minimum of 78 differentmammal species, at least 21 species ofreptiles and amphibians, probablymore than 150 species of birds (morethan we have been able to identify)and more than 14 species of fish.Moreover, human hunters tend to se-lect prey that is in prime conditionfrom the perspective of human nutri-tional needs rather than prey madevulnerable by youth, old age, or dis-ease, as do many carnivorous ani-mals.34,35

The skill-intensive nature of humanhunting and the long learning process

. . .it is clear that humanhunting differsqualitatively fromhunting by otheranimals. Unlike mostanimals, which either sitand wait to ambushprey or use stealth andpursuit techniques,human hunters use awealth of information tomake context-specificdecisions, both duringthe search phase ofhunting and then afterprey is encountered.

Figure 9. Data from experiments in which Ache females of different ages were asked tochop down a palm tree, extract its growing shoot (heart), and extract palm starch from thepalm for at least 1

2hour. Y values are plotted at the mean age for each class analyzed.

Age-specific hunting returns are based on measured kg of live weight of game acquired byAche males between 1980 and 1984 and measured out-of-camp foraging time for malesduring the same time period. Y values are plotted at the mean age for each class analyzed.Differences between age categories may be due to differences in encounter rates withgame or success in pursuits of game after encounter.

ARTICLE Evolutionary Anthropology 171

involved are demonstrated dramati-cally by data on hunting return ratesby age. Hunting return rates amongthe Hiwi do not peak until the menreach the age of 30 to 35 (Fig. 8). Theacquisition rates of 10-year-old and20-year-old boys reach, respectively,only 16% and 50% of the adult maxi-mum. In the 1980s, the hourly returnrate for Ache men peaked when theywere in their mid-30s (Fig. 9). The re-turn rate of 10-year-old boys is nowabout 1% of the adult maximum, andthat of 20-year-olds is still only 25% ofthe adult maximum.

It is not surprising that no ethnog-rapher has ever described being ableto hunt at a rate equivalent to that ofstudy subjects. Indeed, most who huntmake kills only after being led to thegame animal by a competent hunter,and then generally with a firearm (asis the case with recreational big-gamehunters). These patterns also mirrorthe effects of acculturation in mostgroups. Foragers who grow up in set-tled communities can often collectfruits and other plant resources atrates equivalent to those obtained byolder individuals who grew up in thebush (see, for example, Blurton Jonesand colleagues28 on the !Kung). Butmost young Ache men who havegrown up on reservation settlementscannot hunt using traditional technol-ogy nearly as successfully as oldermen who grew up in the forest (un-published data).

Chimpanzees too, appear to requiremany years to learn successful hunt-ing techniques. Older males are morelikely than younger adult males to am-bush prey during a group hunt andperform more complicated maneu-vers during the hunt.20 (But see Stan-ford.21)

Although the learning process islong, investments in hunting ability byhuman males allow them to be highlyproductive as adults. The comparativeanalysis of diets and productivityamong foragers shows that men play amajor role in the energetics of humanreproduction. For example, amongthe Ache the total expected net caloricproduction (food produced minusfood consumed) from age 18 to deathis 121,638,000 calories for males and2924,000 calories for females. Thecorresponding figures for the Hiwi are

111,151,000 for males and 23,096,000for females. Even among the Hadza,where women play a much greaterrole in subsistence production, malesprovide as least as much support forreproduction as females do, if notmore. While the estimates may re-quire revision when researchers com-plete their analysis of age-specific pro-duction among the Hadza, our initialapproximation, based on publisheddata, is that over the entire expectedlife course (including the probabilityof survival to each age), net produc-tion for Hadza males is 116,671,000calories, while that of females is only13,352,000 calories. Table 2 shows

that men provide more food energyper day than women do in all but oneor two of the ten foraging societies forwhich there is quantitative data. Menalso provide the vast majority of pro-tein in the diet. This is critical, sincehigher daily protein intake increasesweight gain,36,37 immune function re-sponse,38 and survival.39,40

The fact that men produce morefood than women do in most low-lat-itude foraging societies is not conven-tional wisdom. This is a result of theinfluence of Richard Lee’s41,42 pio-neering study of the !Kung (Ju/’ho-ansi), which showed that women pro-vide more food than men do. There

has been a tendency to generalizethose results to all foraging societies.Table 2 shows that those results arenot general; it is also shows the possi-bility that weaknesses in Lee’s studyhave been misleading for the !Kung aswell. Lee’s sample covered only 28days of one month of 1964, and ontwo of those days he took women outcollecting mongongo nuts in his vehi-cle, thus inflating the collected por-tion of the diet and female returnrates. During much of the year, mon-gongo nuts are not abundant, nor arethey as abundant at other !Kung studysites as at the Dobe site. Another studyof !Kung food production showsmuch higher hunting success thanLee reported. Yellen43 provided datashowing that !Kung men acquiredtwice as much meat per day whenthey were in bush camps than they didin the permanent dry-season water-hole settlement where Herero raisedcattle, which probably had a depres-sive effect on game densities. Yellen’ssample of person consumption dayswas larger than Lee’s and covered allmonths of the year. The daily meatconsumption in Yellen’s sample wasabout 1,600 calories per capita, whichwould represent 68% of all calories iftotal food consumption was the sameas Lee reported. In addition, both Leeand Draper (p. 262 in Lee42) foundthat !Kung men spent more hours perweek on food acquisition than did!Kung women.

The fact that humans were success-ful in colonizing high-latitude ecolo-gies where plant foods are not abun-dant and are available only for shortperiods also demonstrates the impor-tant role that men play in the energet-ics of reproduction. It is also interest-ing that many Neanderthal and earlyHomo sapiens are found at high lati-tudes where plant consumption wasminimal and, in some cases, whereeven fuel and residential constructionwere provided by animal products(see Hoffecker44 for a review of eastEuropean sites). Moreover, some ar-cheological sites beginning with Homoergaster (for example, Boxgrove, En-gland45) contain super-abundant ani-mal remains and evidence of spearsand other hunting tools, but no evi-dence of plant consumption. More re-cent low-latitude archeological as-

Although the learningprocess is long,investments in huntingability by human malesallow them to be highlyproductive as adults.The comparativeanalysis of diets andproductivity amongforagers shows that menplay a major role in theenergetics of humanreproduction.

172 Evolutionary Anthropology ARTICLE

semblages also often have extremelydense bone scatter, which suggestshigh meat consumption. At Kutikina,Tasmania, for example, bone frag-ments comprise nearly 20% of theweight per cubic meter of some arche-ological strata.46 On the other hand,complete assessment of plant pollenat some assemblages, among themTamar Hat on the north coast of Afri-ca,47 suggests no edible plant species,despite dense faunal scatters. Such ev-idence suggests that adult males haveoften been the main, and sometimesthe only food providers in foragingsocieties.

In another vein, the Hadza researchteam15,48,49 has argued that men donot play a major role in the energeticsof human reproduction because thevagaries of hunting luck render meatan unreliable and indefensible re-source. Two separate issues must bedistinguished in evaluating this argu-ment. The first concerns the sourcesof the caloric and nutritional subsidi-zation of human reproduction. Thesecond issue is whether or not the pro-ceeds from hunting preferentially sup-port the spouse and children of thehunter.

With respect to the first issue, thedata we present clearly demonstratethat hunted foods provide a substan-tial proportion of the energy and es-sential nutrients consumed by womenand children. The fact that humans insome places depend on hunting foralmost 100% of their energy needssuggests that this can be a reliablemeans of subsistence because theyhave developed cultural solutions tothe variability problem. The main so-lutions are food sharing and, to alesser extent, food storage. The secondissue is more complex and, as yet, un-resolved. Hawkes and associates49

claim that prey items are not con-trolled or owned because they are im-possible to defend. Thus, males do notprovision their families, but simplyprovide equally for everyone in a bandby hunting. There are, however, nodata to support this interpretation,other than the observation of sharingitself. In fact, there is considerable ev-idence that carcasses can be defendedwhen conditions do not favor foodsharing. The Ache of Paraguay, whosupplied the initial data for the inde-

fensibility view of sharing share gameresources widely when on long forag-ing treks, yet withhold even largegame, sharing mainly with preferredpartners when on reservation settle-ments.50 Thus, the same resources areshared differently in the forest and inthe reservation. The same is true inAfrica, where the large game itemstaken by the Hadza are treated as pri-vate property by other ethnic groupswho trade or sell them in the bushmeat market.

The indefensibility hypothesis alsoasserts that shares given up by suc-cessful hunters are never repaid in anyuseful currency (meat, other goods,or services) and that shares are givento everyone equally, regardless ofwhether they have done or will doanything to pay them back. However,recent research shows that contingentgiving is typical in all societies whereit has been examined. Ache and Hiwiforagers share more with those whoshare with them and, when they aresick or injured, receive help fromthose with whom they have shared

and in relation to how generously theyhave shared.50,51 Yanomamo garden-ers work more hours in the gardens ofnon-kin who work more hours in theirgardens.52 We believe that this typeof reciprocity, often in different cur-rencies, is the basis for all humaneconomies, divisions of labor, andspecialization, and that its critical de-velopment in the hominid line distin-guishes us from our ape relatives.

Taken together, currently availableevidence suggests that men generallyprovide a considerable portion of theenergy consumed by juveniles and re-productive-aged women. This doesnot mean that men contribute moreto society or reproduction than dowomen. Women process food (part ofproviding nutrients), care for vulnera-ble children, and do a variety of otherimportant tasks in all societies. It isthe partnership of men and womenthat allows long-term juvenile depen-dence and learning and high rates ofsurvival. Indeed, analyses show thatamong both the Ache and the Hiwi,individual women produce less food iftheir husband is a high producer.53

Divorce or paternal death leads tohigher child mortality among theAche,10,54 the Hiwi,55 and the !Kung,56

but not the Hadza.57

In our view, human pair bondingand male parental investment is theresult of complementarity betweenmales and females. The commitmentto caring for and carrying vulnerableyoung, common to primate females ingeneral, together with the long periodrequired to learn human huntingstrategies, renders hunting unprofit-able for women. The fact that humanmales can acquire very large packagesof nutrient-dense food means thatthey can make a great difference tofemale reproductive success. That isnot true of most other primates (theexception being callithricid males,who provide other necessary assis-tance). This difference creates is a ma-jor discontinuity between humansand apes, and results in a partnershipbetween men and women. That part-nership is ecologically variable, in thatthe roles of men, women, and childrenvary with the availability of food re-sources in the environment and therisks posed to children.28,55,58–60

When plants are abundant and

In our view, human pairbonding and maleparental investment isthe result ofcomplementaritybetween males andfemales. Thecommitment to caringfor and carryingvulnerable young,common to primatefemales in general,together with the longperiod required to learnhuman huntingstrategies, rendershunting unprofitable forwomen.

ARTICLE Evolutionary Anthropology 173

game is scarce (as in the Gwi environ-ment), men specialize in providing therarest nutrients in the environment,protein and lipid, because these nutri-ents are critical for growth and goodhealth,61,62 while women providemore food energy. When plant foodsare scarce, men’s hunting provides thebulk of energy and women concen-trate on the processing of food andother raw materials, and on childcare. There is much to be learnedabout the determinants of male andfemale roles in foraging societies, butthe primary activity of adult males ishunting to provide nutrients for oth-ers.

In sum, hunting is the human activ-ity that requires the longest periodbefore maximum return rates areachieved, but it also provides the high-est overall return rate once maximumskill levels are reached. As a result,hunting provides the greatest energy

component of human diets in manyforaging societies (Table 3) and is afundamental feature of the humanlife-history adaptation. Resolution ofthe debate about whether hunting isprimarily direct parental investment,as we contend, or mating effort, asproposed by Hawkes and cowork-ers,63 awaits further data and sophis-ticated tests. Nevertheless, the nutri-tional support of reproduction byhuman males is a fundamental fea-ture of our species.

Reduction in Juvenile andEarly Adult Mortality Rates

Human foragers have longer maxi-mum life spans than chimpanzees do,suggesting that they may have lowermortality rates over much of the lifespan.64–67 Both groups experienceminimum mortality rates in the latejuvenile and early adult period, a pat-

tern typical of many living organ-isms.68 However, the minimum mor-tality rate of foragers is about 1% peryear, whereas the minimum rate forchimpanzees is about 3.5% per year(Fig. 10). We propose that the charac-ter of food resources taken by humansultimately lowers the mortality rate ofjuveniles and young adults in threeways. First, the hunted and extractedfoods taken by humans generallycome in large packages. Large, valu-able packages favor food sharing,which reduces fluctuations in dailyfood intake, allows sick and injured

individuals to recover at higher rates,and facilitates provisioning of chil-dren. Second, the food types takenfavor larger foraging parties and resi-dential groups, which offer protec-tion, particularly for juveniles who as-sociate closely with adults. Third, thefood niche has led to the developmentof tools and an understanding of ani-mal behavior, which can be used ef-fectively to repel predators.

Large package size has been posi-tively associated with the probabilityof food transfers or the degree of foodtransfer in virtually every study in

Figure 10. Yearly mortality rates for the Ache are from Hill and Hurtado10 and for chimpan-zees from Hill and coworkers8 as described for Figure 1. Raw mortality data were smoothedusing a double running average. Ages 1–5, 3 pt; ages 5–10 5 pt; ages .10, 9 pt withtruncation at the end of the life table.

. . .hunting is the humanactivity that requires thelongest period beforemaximum return ratesare achieved, but it alsoprovides the highestoverall return rate oncemaximum skill levels arereached. As a result,hunting provides thegreatest energycomponent of humandiets in many foragingsocieties and is afundamental feature ofthe human life-historyadaptation.

174 Evolutionary Anthropology ARTICLE

which it has been examined.50,63,69–75

Among the Ache and Hiwi, the per-centage of foods not eaten by the ac-quirer’s nuclear family is directly re-lated to the mean package size ofdifferent food categories72 or packagesize within food categories.50 Dailyvariability in the acquisition of differ-ent food categories also positively cor-relates with the percent shared in bothsocieties. Importantly, meat comes inlarger packages and is more variablethan extracted resources, which, inturn, come in larger, more variablepackages than do fruits. Meat is trans-ferred most between non-kin, fol-lowed by extracted products, thenfruits.72 Chimpanzees also share meatmore than they do any other food re-source.76,77 Thus, we propose that ashumans moved into a hunting-extrac-tion feeding niche, levels of food shar-ing increased dramatically over thatseen among chimpanzees.

Among the Ache, individuals whoare sick or injured are frequently fedby other individuals, often ones whoare not kin. Those who share a greaterpercentage of their production areprovisioned by more individuals whenthey are sick or injured.51 Illness andinjury are common among those inforaging societies, as we might imag-ine, given their frequent exposure todangers, parasites, and pathogens,and the lack of modern medical care.A systematic health survey in one res-ervation Ache community in 1997showed that adults required care atthe community clinic on 6% of all per-son days, whereas children requiredcare on 3.5% of all person days. Sug-iyama and Chacon78 have calculatedthat Yora men of Peru were unable tohunt on about 10% of all person daysmonitored. Bailey79 reported that Efemen came to ask for medical treat-ment on 21% of all man days and of-ten had problems that precluded for-aging. Most relevant to our hypothesisis how often men, women, or childrenare sick or injured for many days in away that would preclude food acqui-sition. No data are yet available onthis topic. However, we have seen avariety of serious medical problemsincluding snakebites, injuries sus-tained in piranha and jaguar attacks,broken bones, large punctures, andanimal bites, as well as, arrow

wounds, massive infections, and occa-sional illness that have kept peoplefrom foraging for more than a week ata time. Indeed, we researchers haveall experienced medical problems thatlasted for a week or more and pre-vented us from working in the field. Ifpeople were not food-subsidized dur-ing such periods, mortality rateswould undoubtedly be higher.

Large, widely dispersed food patchesmay promote grouping among ani-mals.80,81 Human women and chil-dren generally spend the day in par-ties that contain many individuals ofthe foraging band, whereas chimpan-zee females and juveniles are oftendispersed into parties of one or twoindividuals.76,82 This appears to havetwo important consequences. First,

human males experience higher mor-tality from predators than do females.Eight of the nine Ache who died fromjaguar attacks in the twentieth cen-tury were men, as were 14 of the 18who died from snakebite.10 Second,human juveniles are rarely killed bypredators. In recent memory and his-torical mythology, no Ache child hasbeen killed by a jaguar. The only!Kung reportedly killed by a predatorin Howell’s83 demographic study wasan older man. Nevertheless, there isgood evidence that chimpanzee juve-niles are killed by predators20,76 andthat adult females are probably killedmore often than adult males.

Human tool use probably providesa good deal of protection from preda-tors. Fire appears to frighten many

predators and may provide a gooddeal of nocturnal protection. Huntingtools can inflict serious injury ordeath on predators, thus providingprotection for women and children,particularly when multiple males arepresent. The knowledge and analysisof animal behavior, fundamental tohuman hunting, can also be employedto fend off predators. Ache often knowwhen the group is being stalked by ajaguar by analyzing its footprints andmovement patterns. When they be-come aware of a predatory jaguar,they build brush walls around thecamp and take turns acting as senti-nels through the night.

These quantitative and qualitativedata about food sharing, tools, andknowledge provide suggestive evi-dence that the shift to large-package,high-quality foods indirectly acted tolower mortality rates. Research oncauses of death and risks of predationand illness among both humans andchimpanzees is necessary to deter-mine the factors responsible for themore than three-fold differences in ju-venile and adult mortality rates.

DISCUSSION

Primate Life-History Evolution

The life-history traits and largebrains of fully modern Homo sapiensmay be seen as the extreme manifes-tation of a process that defines theprimate order as a whole. Our theoryorganizes the major evolutionaryevents in the primate order and thespecific changes that occurred in thehominid line.

The early evolution of the primateorder (60 mya to 35 mya) was charac-terized by small increases in enceph-alization. Relatively little is knownabout early life-history evolution ex-cept that it appears that even the more“primitive” prosimian primates werelong-lived and delayed in reaching re-productive maturity as compared tomammals of similar body size. Austadand Fischer84,85 have related this evo-lutionary trend in the primates to thesafety provided by the arboreal habi-tat. They compare primates to birdsand bats, which are also slow in devel-oping and long-lived for their bodysizes. Thus, the first major grade shiftthat separated the primate order from

Large package size hasbeen positivelyassociated with theprobability of foodtransfers or the degreeof food transfer invirtually every study inwhich it has beenexamined.

ARTICLE Evolutionary Anthropology 175

other mammalian orders was achange to a lowered mortality rateand the subsequent evolution ofslower senescence rates.

The second major grade shift oc-curred with the evolution of the an-thropoids, the lineage containingmonkeys, apes, and humans, begin-ning about 35 mya. This was charac-terized by an increasing emphasis onplant foods as opposed to insects, andby more rapid increases in brain sizerelative to body size.80 The major de-fining characteristic of the evolutionof the anthropoids was reorganizationof the sensory system from one inwhich olfaction and hearing were rel-atively dominant to one completelydominated by binocular color vi-sion.80 This grade shift is almost cer-tainly tied to a dietary shift toward adiverse array of plant parts, particu-larly fruits and leaves.

The diet of the anthropoids hasbeen characterized as both “broadand selective.”86,87 The diet is broadprecisely because it is so selective. An-thropoid primates tend to select foodson the basis of the ripeness, fiber con-tent, nutrients, and toxicity of foodsconsumed early in the day.88–90 Thisselectivity requires the allocation ofincreased brain tissue to visual pro-cessing.91,92 It also requires that manydifferent species of plants and animalsbe included in the diet, increasing thedemands for memory and learn-ing.93,94

This grade shift is reflected in brainsize. Regressions of log brain size onlog body size show that the interceptis significantly lower for strepsirhineprimates (including most prosimians)than for the haplorhine primates (in-cluding all anthropoids and a few pro-simians).

The third major grade shift in pri-mates occurred with the evolution ofthe hominoid lineage, the branchleading to apes and humans. This shiftincluded further encephalization. Theintercept in a regression of log brainsize on log body size is significantlyhigher for apes than monkeys, andapes clearly perform better on mosttasks reflecting higher intelligence.93,95

Evidence on teeth and tooth wearamong early hominoids suggests adiet composed mainly of soft, ripefruits.96 The frugivorous emphasis of

the hominoid lineage is evident inlater hominoid species as well.96

Wrangham and colleagues97,98 alsoshow that the chimpanzee diet isbased on a much greater percentageof ripe fruits than is the diet of othersympatric primate frugivores. (This isprobably true of orangutans as well).The cognitive demands of a diet thatemphasizes ripe fruits are likely to bemuch greater than simple frugivory.For one thing, there are greater per-ceptual demands in detecting the stateof fruit from visual cues against abackground.91 For another, the lowerabundance of ripe fruits and the short

time in which they are available (ripe,but not yet eaten by competitors) islikely to impose greater demands withrespect to monitoring the environ-ment, remembering the state of indi-vidual trees, and predicting when thefruits will become ripe on the basis oftheir current state.

In addition to eating ripe fruits,chimpanzees and gorillas also usecomplex techniques for extractingfoods from protected substrates, suchas nut-cracking, termite and ant fish-ing, and removal of bark to get atpith.93 Gibson99,100 identified extrac-tive foraging in primates as an impor-

tant selective force in primate intelli-gence and presented evidence insupport of this idea. Furthermore,these behaviors vary between groupsas social traditions. In a comprehen-sive review of chimpanzee cultures us-ing 151 years of chimpanzee observa-tions from seven long-term studies, 39behavior patterns were found to becustomary or habitual in some com-munities but absent from otherswhere ecological explanations couldbe discounted.101 Of these, 19 werepatterns of extractive foraging. An ad-ditional 14 extractive foraging behav-iors were identified but failed toachieve habitual status in any onecommunity. Furthermore, chimpan-zee males are avid hunters. Boeschand Boesch20 present data indicatingthat older chimpanzee males are ca-pable of predicting escape patterns oftheir prey and of predicting how preywill respond to the behaviors of otherchimpanzees. This appears to requireeven greater levels of cognitive pro-cessing than are required for extrac-tive foraging. Although, as comparedto humans, chimpanzees engage inrelatively little extractive foraging andhunting, they do much more thanmonkeys. In this sense, their superiorintelligence and greater encephaliza-tion than occurs in monkeys illus-trates the same evolutionary forcesthat separate humans from apes. (SeeLancaster and associates102 for a re-view of chimpanzee behavior and cog-nition.)

There is some debate, however,about the relative importance of dietversus group living in the evolution ofprimate intelligence and brain size.According to one view, the increase inbrain size was largely driven by thecomplexities of the primate diet. Jeri-son103,104 suggested that brain tissueevolves in response to two kinds ofdemands: One depends on body size,based on monitoring and supportingan animal’s body tissue and particu-larly its surface area; the other is theability to assimilate, integrate, and re-member environmental information.He therefore predicted that differ-ences in brain size, after controllingfor body mass, would be associatedwith an animal’s ecological niche andits demands for information process-ing. Jerison103 hypothesized that the

Although, as comparedto humans,chimpanzees engage inrelatively little extractiveforaging and hunting,they do much morethan monkeys. In thissense, their superiorintelligence and greaterencephalization thanoccurs in monkeysillustrates the sameevolutionary forces thatseparate humans fromapes.

176 Evolutionary Anthropology ARTICLE

need to process information in a com-plex three-dimensional environmentwas the cause of the large brain ofprimates relative to the brains of othermammals.

Clutton-Brock and Harvey105 testeda version of this hypothesis with inter-genera comparisons within families ofprimates. They reasoned that frugi-vores need to assimilate and retainmore environmental information thanfolivores do because fruits are morescarcely distributed than leaves, re-quiring more specific locational mem-ory. They found that, after controllingfor body size, both dietary emphasis(leaves versus fruits and insects) andthe size of the home range predictedbrain size. Milton23,89,106 extendedtheir work, focusing on gut specializa-tion and brain size as alternativeroutes to energetic efficiency. Leaves,while abundant, tend to contain highamounts of fiber and often toxins aswell. The ability to extract nutrientsfrom leaves depends on the size of thegut and other specializations designedto facilitate fermentation for nutrientextraction. Fruits, on the other hand,are ephemeral resources, patchily dis-tributed but offering a higher densityof easily processed energy. Miltonshowed in paired interspecific com-parisons that gut size and brain sizewere inversely correlated, correspond-ing to the dietary emphasis on fruitsversus leaves.

Another view is that brain-size evo-lution was driven primarily by thecomplexities of social life in primategroups.107,108 Many species of pri-mates exhibit complex dominance hi-erarchies that are mediated by politi-cal alliances and relations amongrelatives in genetic lineages.109–112 Itis not clear why higher primates havesuch complex social relations, but itappears that group living is at leastpartially a response to predation thatsignificantly lowers predation. Amongspecies that tend to eat higher-qualityfoods that are easy to monopolize, so-cial relationships also mediate accessto foods within groups.113

The most recently published analy-ses with the largest samples show thatboth group living and diet are associ-ated with brain size and the size of theneocortex in primates.91,92 In a sepa-rate set of analyses, Allman and col-

leagues66 have shown that group liv-ing and diet are positively associatedwith maximum life span in primatesafter controlling for body size. Theyalso have shown a high positive corre-lation between brain size and lifespan, again controlling for body size.Smith114 showed that after controllingfor body size, brain size in primates ispositively associated with the age offirst molar eruption, an indicator ofthe age at which individuals begin toconsume adult diets, as well as withage of first reproduction and longev-ity. The relationship is strongest for

age of first molar eruption, probablybecause brain growth and postcranialmorphological development compete.115

Kaplan and colleagues116 have re-cently conducted multivariate analy-ses specifically designed to test thepresent theory. They found that whenbrain size is regressed on body size,age at first reproduction, maximumlife span, percent of fruit in the diet,range size, and group size in a multi-variate model, all but group size weresignificant predictors of brain size. Inaddition, the frugivory and range sizevariables, meant to capture the cogni-tive demands of the diet, also predictthe life-history variables (age of first

reproduction and maximum lifespan)after controlling for body weight.

Regardless of the importance ofgroup size in determining brain sizein monkeys, the grade shift in brainsize between monkeys and apes is al-most surely related to diet and not togroup size.95 Apes show dietary spe-cialization, but do not live in particu-larly large groups. In fact, gibbons,orangutans, and to a lesser extent, go-rillas, live in relatively small groupswhile chimpanzees and bonobos formgroups of the same size as do baboonsand vervets.

It seems likely that the cognitiveability associated with foraging in acomplex three-dimensional environ-ment is an important pre-adaptationfor social intelligence and complex so-cial relations. Of particular interesthere are the specializations inherentin the primate visual system, includ-ing binocular vision, high visual acu-ity, and associated increases in thesize of the lateral geniculate nucle-us.64,65,91 Barton91 has shown that thesize of the parvocellular system in-volved in the processing of visual stim-uli increases with increasing group sizeamong nonhuman primates. Thus, itmay be that selection favoring socialintelligence increases once the neces-sary cognitive pre-adaptations exist.This may be one reason why primatesexhibit such high levels of social com-plexity relative to other group-livingmammals, such as many herbivores.Adept social manipulation that leadsto a higher position in a dominancehierarchy may be favored by naturalselection, for dominance has corre-lated positively with measures of fit-ness in a multitude of primate stud-ies.109–113,117 Yet this would probablybe true for other social mammals aswell. Many of the same abilities tostore and analyze information may beemployed to solve dietary and socialproblems.94 Once social intelligenceand social complexity evolve, the cog-nitive adaptations may be maintainedeven when dietary complexity is sec-ondarily reduced in some species, asis the case with the derived simplicityof the diet of colobine monkeys.105

Apes show remarkably sophisticatedsocial intelligence94,112,118,119 andstand out in comparison to monkeys.Although this cannot be a result of

It seems likely that thecognitive abilityassociated with foragingin a complex three-dimensionalenvironment is animportant pre-adaptation for socialintelligence andcomplex socialrelations. Of particularinterest here are thespecializations inherentin the primate visualsystem . . .

ARTICLE Evolutionary Anthropology 177

greater group size per se, it is againpossible that selection on social appli-cations of intelligence increased whenthe cognitive adaptations to feedingacted as a pre-adaptation.

It is useful to think of primate evo-lution as both a branching and a di-rectional process. It is branching inthe sense that ecological variation, in-traniche competition, and segregationgenerate variable selection pressuresthat lead to the evolution of multiplespecies with different traits. Some areselected to rely on more easily ac-quired foods and travel less, to investmore in gut physiology and less inbrains, and to mature more rapidlyand live shorter lives. Others are se-lected to rely on more complex feed-ing strategies and exhibit their respec-tive, correlated life histories. It is adirectional process in the sense thatcognitive and life-history adaptationsthat evolved previously act as pre-ad-aptations for further selection in thesame direction in response to nichecompetition and ecological change.Homo sapiens is an extreme in one ofthose directions.

The Hominid Line

The fourth major grade shift in pri-mate evolution occurred with diver-gence of the hominid line, particularlythe evolution of genus Homo. Thebrain and life span of modern humansare clearly “outliers” compared tothose of other mammals, and even ascompared to the relatively large-brained, slow-living primates. Theevolution of these extreme adapta-tions in the hominid line is built on ahominoid base that already showed asignificant tendency toward largebrains, long lives, and exploitation ofhigh-quality foods. Available evidenceon Australopithecines suggests thatbipedalism preceded changes in brainsize and life history.114,115

Although the record is still incom-plete, it appears that brain enlarge-ment and life-history shifts co-oc-curred. Early Homo ergaster showsboth significant brain expansion andan elongated developmental period,120

but much less so than modernhumans. Neanderthals display bothbrain sizes and dental developmentthat are in the same range as those ofmodern humans. Bipedalism can be

thought of as a pre-adaptation thatevolved to facilitate terrestrial loco-motion, allowing for long day rangesthrough energetic efficiency.121 Achemen walk an average of 15 km perday, about three times the distancecovered by chimpanzee males,32,33

even though both inhabit tropical for-ests. Bipedality also had the second-ary effect of freeing the hands for spe-cialization in manipulative activities.Environmental conditions in the earlyPleistocene interacted with bipedalityto favor an increased emphasis on ex-tractive foraging, hunting, learning,prolonged development, and longlives for at least one line of the homi-nid family. It may be that this uniquesuite of conditions is responsible forthe extreme differences between hu-mans and other primates in intelli-gence, development, longevity, and re-source flows across generations. The

complexities of a highly variable cli-mate also would have favored a trendtoward cognitive solutions that led tonew foraging opportunities. Homin-ids with such solutions were probablybetter able to withstand rapid climateand habitat change than those whowere more inflexible.122

The shift toward a high-quality dietbased on learned foraging techniquesalso implies another co-evolutionaryprocess. Just as the juvenile periodand the life span co-evolved, so too didbrain and gut sizes co-evolve with life-history traits and with each oth-er.23,123 Many other physiologicaltraits are probably related to thisevolved complex. For example, it islikely that hidden estrus among hu-mans is related to the male energysubsidization of adult female repro-duction,62 and that the postreproduc-

tive life span of both sexes is related tohigh food production by older individ-uals. Almost certainly, the reason thathuman females have shorter inter-birth intervals than apes do124 (andprobably different physiological sensi-tivity to nursing stimulation) is becausereproductive-aged human women areable to decrease rather than increasetheir food production during lactationdue to subsidies by other age and sexclasses.102 This commitment to foodsharing is evident in human physiol-ogy and behavior. Women obtainmore of the extra energy required forpregnancy and lactation by increasesin energy intake125 and reductions inenergy expenditure12,126 rather thanby an increase in fat mobilization ormetabolic economy.125 Reductions inbasal metabolic rate during earlypregnancy127 are not sufficient tomake up for the increased energy re-quirements. It is likely that this re-sponse evolved in the context of pro-visioning.

Figure 11 illustrates our historicalhypothesis regarding the hominid dietand its impact on both pay-offs tolearning and development and onmortality rates. The figure begins withtwo important exogenous changes.The first is a change in the distribu-tion of foods with the emergence ofAfrican savannas in the Pleistocene,which increased the abundance ofhigh-quality but protected plant foods(nuts and tubers, in particular) andanimal foods. The second change isthe pre-adaptation of bipedality,emerging in the Australopithecines asa locomotor adaptation. This adapta-tion frees the hands for tool use,which allows more efficient extractionand hunting and frees the hands forcarrying large packages of food suit-able for sharing.128,129 It also results inefficient terrestrial locomotion andhigher daily mobility that would in-crease the daily encounter rate withrare but energetically rich resources.These changes led to an increased em-phasis on large, high-quality, but dif-ficult-to-acquire foods. Our hypothe-sis is that this feeding niche hadmultiple effects: It increased the pre-mium on learning and intelligence,delaying growth and maturation; in-creased nutritional status and de-creased mortality rates through food

Although the record isstill incomplete, itappears that brainenlargement and life-history shiftsco-occurred.

178 Evolutionary Anthropology ARTICLE

sharing (predicted by the provisioningof young sick or injured individuals);and released selection against largergroup size, which lowers predationmortality.

The Grandmother Hypothesis

The present theory shares some fea-tures with a model of menopause andhuman life history recently proposedby Hawkes and colleagues,130 often re-ferred to as the grandmother hypoth-esis. This hypothesis proposes thathumans have a long life span relativeto that of the other primates becauseof the assistance that older postrepro-ductive women contribute to descen-dant kin through the provision of dif-ficult-to-acquire plant foods. Women,therefore, are selected to invest inmaintaining their bodies longer thanchimpanzee females do. Thus, boththeories focus on the exceptionallylong human life span. Rather than re-garding the cessation of reproductionin the fifth decade of life as the criticaladaptation (a feature shared withapes), both theories attempt to ex-plain the extension of the human lifespan beyond menopause.24,130 Thisimplies that some fertility earlier inlife is given up in order to prolong theadult life span.24 The theories differ inimportant ways, however.

First, the grandmother theory fo-cuses on the productivity of olderwomen, but not on investments in

learning and development. Thepresent theory specifically links highproductivity later in life to invest-ments in a large brain and to theunique features of growth in humanchildren. The grandmother hypothe-sis is silent about the expansion of thecostly human brain. It also offers noexplanation of why human childrengrow so slowly and take so long tomature, except for the fact that adultmortality rates are low. In fact,Hawkes and colleagues130 suggestthat learning is a secondary effect of along juvenile period (determined bythe long lifespan) rather than thecause of the long juvenile period. Sec-ond, the grandmother theory fails toaccount for why men live to about thesame age as women. If the benefits ofliving longer derive, for women, fromprovisioning descendant kin and, formen from direct reproduction, thereis no reason why their life spansshould be so similar. Third, the grand-mother model fails to capture the im-portant role that human males play insupporting women’s reproduction. Italso fails to explain the age profile ofproduction among males and whymen take so long to reach their pro-ductive peaks.

The evidentiary basis of the grand-mother hypothesis is also very weak.There is no direct evidence that post-reproductive women are the majorfood providers in any society. WhileHadza data do suggest that older

Hadza women produce more foodthan younger women do, postrepro-ductive women produce less thanHadza men (Fig. 2). Furthermore, it isunlikely that the Hadza pattern ofhigh food production by postrepro-ductive women is common amongother foragers. Ache and Hiwi pos-treproductive women do not acquireeven half the daily food energy ac-quired by adult men. !Kung data alsosuggest that older women producevery little, even leading the Hadza re-search team to ask “Why don’t elderly!Kung women work harder to feedtheir grandchildren?” (Blurton Jones

and coworkers,13 p. 388). And cer-tainly none of the high-meat-consum-ing societies of Table 4 nor any of thesocieties in human history dwelling athigh latitudes could have been mostlydependent on the food production ofpostreproductive women. While weagree that the age profile of produc-tion is shifted toward older agesamong women as well as men, andthat postreproductive women providemany important services in foragingsocieties, among them child care, toolmaking, food processing, and campmaintenance, there is no evidence tosupport the hypothesis that they have

Figure 11. Ecology and co-evolutionary process in the Hominid line. The present theoryspecifically links highproductivity later in lifeto investments in a largebrain and to the uniquefeatures of growth inhuman children. Thegrandmother hypothesisis silent about theexpansion of the costlyhuman brain. It alsooffers no explanation ofwhy human childrengrow so slowly and takeso long to mature. . .

ARTICLE Evolutionary Anthropology 179

been the “breadwinners” in most soci-eties during human history.

The present theory organizes all ofthese facts. Both males and femalesexploit high-quality, difficult-to-ac-quire foods (females extracting plantfoods and males hunting animalfoods), sacrificing early productivityfor later productivity, with a life-his-tory composed of an extended juvenileperiod in which growth is slow, alarge brain is programmed, and a highinvestment is made in mortality re-duction and maintenance to reap therewards of those investments.

PREDICTIONS OF THE THEORYTO BE TESTED IN FUTURE

RESEARCH

Comparative Biology

The co-evolutionary selection pres-sures posited by our theory and sup-ported by the quantitative theoreticalmodel should be generally applicable.Holding pay-offs to time spent as ajuvenile constant, increased survivalrates during the juvenile period shouldselect for delayed reproduction, aspredicted by the models of Kozlowskiand Weigert5 and Charnov.1,9 Holdingconstant extrinsic mortality hazardsand the pay-offs to investments inmortality reduction, increased pay-offs to time spent in development as ajuvenile should select for increased al-locations to survival and lower mor-tality rates. The latter is a novel pre-diction of our theory. The testing ofour predictions will require carefulcomparative research within taxo-nomic groups at different levels ofanalysis (for example, among birds,mammals, and reptiles, and amongorders and genera within those highertaxonomic levels). It will be necessaryto distinguish pay-off functions andmortality hazards from the actual lev-els achieved, given the observed allo-cations. For example, it will be neces-sary to distinguish the risk of dyingfrom different causes, such as preda-tion and infectious disease, from ob-served mortality rates due to thosecauses, because the observed mortal-ity rate will be affected by allocationsto predator avoidance and feeding,and by allocations to immune func-tion.

A principal innovation of the pro-posed theory is the incorporation of

co-evolutionary selection among life-history traits without requiring anytrait to be treated as extrinsic. Thisis more realistic than previous ap-proaches, but makes for increasinglydemanding empirical analysis. Thereare many possible pay-offs to timespent in development. The most gen-eral benefit of increased time spent indevelopment is increased body sizeand its effects on energetic turnover,survival, and mating success. The co-evolutionary selection measures pos-ited here should apply to those effects,but of primary interest in our theoryare the benefits associated with learn-

ing and information storage. Varia-

tion within birds, carnivores, pinni-peds, and primates should beparticularly fertile ground for testingpredictions generated by the theorywith respect to learning, brain size,and life-history traits. The generalprediction is that within those taxo-nomic groups, species that rely onmore learning-intensive, complexfeeding strategies will have longer de-velopmental periods, longer lifespans, and larger brains relative tobody size. With respect to brain size,the correlations of brain size with life-history characteristics and feedingstrategies should be reflected in thoseparts of the brain associated withlearning and information storage, and

not with raw perceptual processing,such as sonar and visual acuity (un-less, of course, those functions aremore developed in organisms thatlearn and store information). Sugges-tive evidence in support of the theoryis available for birds131 and pri-mates,116 but rigorous testing of thosepredictions awaits further researchand analysis.

Hominid Evolution

Following this line of reasoning, thefirst major increases in brain size inthe hominid line should be accompa-nied by extensions of the juvenileperiod, increased longevity, and in-creased complexity of learned forag-ing strategies. Smith’s analyses oftooth eruption among early Homo er-gaster provide suggestive evidence ofextension of the juvenile period be-yond that of chimpanzees and Austra-lopithecines, but to a lesser extentthan that of fully modern humans. Sofar, the record is silent with respect tolongevity and mortality rates. Ourability to test the prediction that lon-gevity increases with brain size duringhominid evolution may await new de-velopments such as advances in theextraction of DNA from fossil remainsand in understanding the genetics ofaging. Nevertheless, this extension ofthe expected life span with increasedemphasis on learning in developmentand brain size is a firm prediction ofour theory.

Although the early evolution of thegenus Homo could have been accom-panied by increased complexity of for-aging strategies with respect to eitherhunting or gathering, we stronglysuspect that future research willdemonstrate increases in both theimportance of hunted foods andcomplex extractive technologies forgathering.23,121,128,129,132,133 Recently,O’Connell and associates49 have re-jected hunting as being important inearly hominid evolution on twogrounds. First, some archeological as-semblages that include large accumu-lations of animal bone, thought to il-lustrate the hunting life-style, havenow been reinterpreted as possibly re-sulting from natural processes. Sec-ond, recent data show that chimpan-zees hunt considerably more than waspreviously thought. Therefore, hunt-ing cannot be the cause of the changes

A principal innovation ofthe proposed theory isthe incorporation of co-evolutionary selectionamong life-history traitswithout requiring anytrait to be treated asextrinsic. This is morerealistic than previousapproaches, but makesfor increasinglydemanding empiricalanalysis.

180 Evolutionary Anthropology ARTICLE

in hominid life histories and socialsystems.

With respect to the first point, it ispremature to reject the “hunting hy-pothesis” simply because the causes ofbone accumulation at hominid sitesare not well understood and are opento various interpretations. It has neverbeen demonstrated that early homin-ids did not rely heavily on huntedfoods. It is also the case that too muchemphasis has been placed on huntinglarge game (Jones presents suggestiveevidence that bone accumulations atOlduvai reflect reliance on small gamehunting134). Chimpanzees hunt smallgame, as do many modern huntergatherers. For example, the Ache relyheavily on small game: About 50% ofthe animal food in the Ache diet issmall game acquired using hand-hunting techniques (no projectileweapons). In fact, it may be that thelearning demands of a diet based onsmall game may be greatest because itcan require killing many different spe-cies at regular intervals. Thus, muchencounter-specific and species-spe-cific knowledge and creativity mayalso be required.

With respect to the second point,Tables 1 and 2 show that humans andchimpanzees exploit very differentfood niches. While chimpanzees dohunt, the most successful chimpanzeehunters obtain less than 10% of thedaily per capita energy intake frommeat reported for any human forag-ing group. To claim that humans andchimpanzees both must have thesame life history and social system be-cause they both hunt is equivalent toasserting that chimpanzees must havethe same social system as black andwhite colobus monkeys because theyboth eat some ripe fruit. We predictthat early hominids will show a majorincrease in the consumption ofhunted foods. Analyses of vitaminsand minerals in hominid bone re-mains may provide new avenues fortesting that hypothesis, in addition toarcheological evidence.135

It is clear that the process of homi-nization occurred over a long time.Our theory is silent about the dates ofimportant events and evolutionarychanges in the hominid line. Whetherthe process will look ratchet-like, inthat small increases in longevity willprecede small increases in brain size,which, in turn, lead to further in-

creases in longevity, is also an openquestion. Whether the process is grad-ual or punctuated, our theory predictsthat changes in brain size, the dietaryimportance of meat and other diffi-cult-to-acquire foods, gut size, and lifespan will be seen to co-evolve in thearcheological and paleontologicalrecord. It is important to recognizethat many of those changes may bequantitative as well as qualitative. Forexample, although chimpanzees hunt

monkeys, they only pursue them in avery small proportion of their encoun-ters with them, presumably becausemost of those encounters do not occurunder conditions likely to result in a

successful kill. Ache foragers onlyrarely ignore an encounter with mon-keys, even when they hear them at agreat distance, because most such en-counters do result in successful kills,presumably because of the Ache’s skilland effective weaponry. The quantita-tive changes in brain size and life-history traits during hominid evolu-tion should be accompanied byquantitative changes in diet and skill-intensive foraging techniques.

Modern Humans VersusChimpanzees

There are at least two major ave-nues of future research with extant

hunter-gatherers and chimpanzeesfor testing the predictions of our the-ory. First, we have provided only sug-gestive evidence that food sharing andprovisioning lower both juvenile andadult mortality rates in humans andthat this is a major cause of the differ-ence between human and chimpanzeeage-specific survival probabilities. Weneed to know a great deal more aboutthe frequency of illness in both speciesand the relationship between morbid-ity and mortality. What is the relation-ship between illness and food intakerates in both species? What is the re-lationship between reduced food in-take and mortality? Our theory pre-dicts that weight loss accompanyingillness will be more frequent and se-vere, and that illness will more fre-quently result in mortality amongchimpanzees than humans. Our the-ory also predicts that chimpanzee ju-veniles will spend more time than hu-man juveniles in contexts that exposethem to predation, and that this dif-ference will be due to both group sizeand time spent in the food quest. Ourtheory also predicts absolutely greaterfrequency of predation on chimpan-zees than on humans.

A second avenue of research is onthe components of foraging success. Aprincipal difference between our the-ory and the grandmother hypothesisis that we propose that the human lifecourse and large brain are the resultsof a long learning process that is nec-essary for successful foraging, andthat both males and females engage inthis process. Recently, on the basis ofsome experiments among Hadza juve-niles, Blurton Jones and Marlowe136

have suggested that human foraging isnot very difficult to learn. We proposethat hunting, as practiced by humans,but not necessarily by other preda-tors, is exceedingly difficult to learnand requires many years of experi-ence. Our observations of hunters insix different groups suggests to us thatit is not marksmanship, but the knowl-edge of prey behavior and remotesigns of that behavior such as tracksand vocalizations that are the mostdifficult features of human hunting.30

This impression is testable. Weshould find that marksmanship is ac-quired relatively early in the learningprocess and that naive individualssuch as anthropology graduate stu-dents could be trained to be effective

To claim that humansand chimpanzees bothmust have the same lifehistory and socialsystem because theyboth hunt is equivalentto asserting thatchimpanzees must havethe same social systemas black and whitecolobus monkeysbecause they both eatsome ripe fruit.

ARTICLE Evolutionary Anthropology 181

marksmen with relatively little prac-tice. In contrast, we should find thatknowledge of animal behavior andsigns of behavior are learned onlygradually over many years. We shouldalso find that encounter rates withprey are much more strongly age-de-pendent than are kill rates upon en-counter. It should also be the case thatkills that rely on marksmanship areless age-dependent than kills that relyon finesse and skill, such as luring an-imals out of holes. In addition, itshould be much more difficult to trainnaive individuals to find prey than toshoot projectile weapons accurately.Similarly, we should find that age ef-fects on chimpanzee hunting shouldbe a result of differences in knowledgeof prey behavior. Boesch and Boesch20

provide suggestive evidence of sucheffects, but more research is necessaryfor a definitive test. An analogous setof predictions also could be testedwith extractive gathering techniquesand the specific knowledge requiredfor each (Bock29,137 presents a seriesof experiments with grain processingas a model of how such research couldbe conducted.

CONCLUSIONS

The human adaptation is broad andflexible in one sense, but narrow andspecialized in another sense. It isbroad in the sense that, as hunter-gatherers, humans have existed suc-cessfully in virtually all of the world’smajor habitats. This has entailed eat-ing a very wide variety of foods, bothplant and animal, both within andamong environments. It also has en-tailed a great deal of flexibility in thecontributions of individuals of differ-ent ages and sex. The relative contri-butions of men and women to foodproduction appear to vary from groupto group. Even the contributions ofchildren and teens to food productionvary predictably with the abundanceof easy-to-acquire foods.

Our adaptation is narrow and spe-cialized in that it is based on a dietcomposed of large, nutrient-dense,difficult-to-acquire packages and a lifehistory with a long, slow develop-ment, a large commitment to learningand intelligence, and an age profile ofproduction shifted toward older indi-viduals. We do not expect to find anyhuman population that subsists on

leaves or other low-quality foods. In-deed, we expect humans to remain atthe very top of the food hierarchy inevery environment they live in (for ex-ample, humans often exterminate allother top predators in their habitat).Humans ingest foods that are alreadyhigh in quality and do not requiremuch digestive work or detoxification.And if a food contains toxins, they aregenerally removed prior to ingestionby processing techniques. This dietarycommitment is reflected in the ex-tremely reduced size of the human

hindgut.23 Humans use their great in-telligence to extract and hunt thosefoods. In order to achieve this diet,humans also engage in extensive foodsharing both within and among ageand sex classes of individuals. Finally,the effect of the commitment to foodsharing is evident in the reproductivephysiology of human women. Provi-sioning permits human women, incontrast to other female primates, toreduce rather than increase their rateof energy production during their re-productive years, when they have bothinfant and juvenile nutritional depen-

dents and a greatly reduced spacingbetween births.

The model and the data we havepresented suggest that the human lifecourse is based on a complex set ofinterconnected, time-dependent pro-cesses and the co-evolution of physi-ology, psychology, and behavior.There appears to be a tight linkageamong the ordering of major psycho-logical milestones (language learning,understanding and mastering thephysical, biological, and social envi-ronment); the timing of brain growth;growth rates during childhood andadolescence; developmental changesin survivorship; behavioral, psycho-logical, and physiological changeswith the transition to adulthood; pro-files of risk with age; and rates of se-nescence and aging. It is very likelythat a species-typical life courseevolved in response to the demands ofa hunting and gathering lifestyle thatwas broad and flexible enough to al-low successful exploitation of theworld’s environments, but specializedtoward the acquisition of learnedskills and knowledge to obtain veryhigh rates of productivity later in life.

ACKNOWLEDGEMENTS

We thank K.G. Anderson, TheodoreBergstrom, John Bock, Robert Boyd,Pat Draper, Peter Ellison, MichaelGurven, Sarah Hrdy, Monique Borg-erhoff Mulder, David Lam, and PeterRicherson for helpful discussions ofthese issues as well as Anne Pusey andJanette Wallis for sharing data onchimpanzee development and birth-spacing. We especially thank NicholasBlurton Jones, Eric Charnov, KristenHawkes, and James O’Connell whoseseminal work in this area both chal-lenged us and helped stimulate ourthinking, and who took time to cri-tique various versions of this manu-script. Arthur Robson deserves specialrecognition for performing the math-ematical analysis to generate the re-sults outlined in Box 1 and writing theMathematical Appendix, which isavailable from the authors.

Field research on the Ache, Hiwi,and Machiguenga has been supportedby grants to Kim Hill and A. M. Hur-tado by the National Science Founda-tion (BNS-8613215, BNS-538228,BNS-8309834, BNS-8121209, BNS-9617692), the National Institutes of

Finally, the effect of thecommitment to foodsharing is evident in thereproductive physiologyof human women.Provisioning permitshuman women, incontrast to other femaleprimates, to reducerather than increasetheir rate of energyproduction during theirreproductive years,when they have bothinfant and juvenilenutritional dependentsand a greatly reducedspacing between births.

182 Evolutionary Anthropology ARTICLE

Health (RO1HD16221-01A2), Funda-cion Gran Mariscal de Ayacucho, Ca-racas, Venezuela, and two grants fromthe L.S.B. Leakey Foundation. Fur-ther support has gone to H. Kaplanfor field research on the Machiguengaand Yora by the National ScienceFoundation (BNS-8718886) and theL.S.B. Leakey Foundation, and for lifehistory theory and life course evolu-tion by the National Institute of Aging(1R01AG15906-01).

REFERENCES

1 Charnov EL. 1993. Life history invariants:some explanations of symmetry in evolutionaryecology. Oxford: Oxford University Press.2 Gadgil M, Bossert WH. 1970. Life historicalconsequences of natural selection. Am Nat 104:1–24.3 Janson CH, Van Schaik CP. 1993. Ecologicalrisk aversion in juvenile primates: slow andsteady wins the race. In: Pereira M, Fairbanks L,editors. Juvenile primates: life history, develop-ment and behavior. New York: Oxford UniversityPress. p 57–76.4 Kozlowski J, Wiegert RG. 1986. Optimal allo-cation to growth and reproduction. TheoretPopul 29:16–37.5 Kozlowski J, Weigert RG. 1987. Optimal ageand size at maturity in the annuals and perenni-als with determinante growth. Evol Ecol 1:231–244.6 Hawkes K, O’Connell JF, Blurton Jones N.1989. Hardworking Hadza grandmothers. In:Standen V, Foley RA, editors. Comparative socio-ecology of humans and other mammals. London:Basil Blackwell. p 341–366.7 Hawkes K, O’Connell F, Blurton Jones N. 1997.Hadza women’s time allocation, offspring provi-sioning, and the evolution of long postmeno-pausal life spans. Curr Anthropol 38:551–577.8 Hill K, Boesch C, Pusey A, Goodall J, WilliamsJ, Wrangham R. 2000. Chimpanzee mortality inthe wild. University of New Mexico.9 Charnov E. 1993. Why do female primates havesuch long lifespans and so few babies? Evol An-thropol 1:191–194.10 Hill K, Hurtado AM. 1996. Ache life history:the ecology and demography of a foraging peo-ple. Hawthorne, NY: Aldine de Gruyter.11 Hill K, Hawkes K, Hurtado A, Kaplan H.1984. Seasonal variance in the diet of Ache hunt-er-gatherers in eastern Paraguay. Hum Ecol 12:145–180.12 Hurtado AM, Hill K. 1990. Seasonality in aforaging society: Variation in diet, work effort,fertility, and the sexual division of labor amongthe Hiwi of Venezuela. J Anthropol Res 46:293–345.13 Blurton Jones N, Hawkes K, O’Connell J.1989. Modeling and measuring the costs of chil-dren in two foraging societies. In: Standen V,Foley RA, editors. Comparative socioecology ofhumans and other mammals. London: BasilBlackwell. p 367–390.14 Blurton Jones NG, Hawkes K, O’Connell J.1997. Why do Hadza children forage? In: SegalNL, Weisfeld GE, Weisfield CC, editors. Unitingpsychology and biology: Integrative perspectiveson human development. New York: AmericanPsychological Association. p 297–331.15 Hawkes K, O’Connell JF, Blurton Jones NG.1991. Hunting income patterns among theHadza: big game, common goods, foraging goals

and the evolution of the human diet. Philos TransR Soc London (B) 334:243–251.16 Blurton Jones N, Smith L, O’Connell J,Hawkes K, Samuzora CL. 1992. Demography ofthe Hadza, an increasing and high density popu-lation of savanna foragers. Am J Phys Anthropol89:159–181.17 Kaplan HK. 1994. Evolutionary and wealthflows theories of fertility: empirical tests and newmodels. Popul Dev Rev 20:753–791.18 Becker GS. 1975. Human capital. New York:Columbia University Press.19 Mincer J. 1974. Schooling, experience, andearnings. Chicago: National Bureau of EconomicResearch.20 Boesch C, Boesch H. 1999. The chimpanzeesof the Tai Forest: behavioural ecology and evolu-tion. Oxford: Oxford University Press.21 Stanford CG. 1999. The hunting apes: meateating and the origins of human behavior.Princeton: Princeton University Press.22 Wrangham RW, Van E, Riss ZB. 1990. Ratesof predation on mammals by Gombe chimpan-zees, 1972–1975. Primates 3:157–170.23 Milton K. 1999. A hypothesis to explain therole of meat-eating in human evolution. Evol An-thropol 8:11–21.24 Kaplan HK. 1997. The evolution of the humanlife course. In: Wachter K, Finch CE, editors.Between Zeus and Salmon: the biodemographyof aging. Washington, D.C.: National Academy ofSciences. p 175–211.25 Hiraiwa-Hasegawa M. 1990. The role of foodsharing between mother and infant in the ontog-eny of feeding behavior. In: Nishida T, editor.The chimpanzees of the Mahale Mountains: sex-ual and life history strategies. Tokyo: Tokyo Uni-versity Press. p 267–276.26 Silk JB. 1979. Feeding, foraging, and food-sharing behavior in immature chimpanzees. Fo-lia Primatol 31:12–42.27 Teleki G. 1973. The predatory behavior ofwild chimpanzees. Lewisburg, PA: Bucknell Uni-versity Press.28 Blurton Jones NG, Hawkes K, Draper P. 1994.Foraging returns of !Kung adults and children:why didn’t !Kung children forage? J AnthropolRes 50:217–248.29 Bock JA. 1995. The determinants of variationin children’s activities in a Southern Africancommunity. Ph.D. dissertation, University ofNew Mexico.30 Leibenberg L. 1990. The art of tracking: theorigin of science. Cape Town: David Phillip.31 Kelly R. 1995. The foraging spectrum: diver-sity in hunter-gatherer lifeways. Washington,D.C.: Smithsonian Institution Press.32 Wrangham W. 1975. The behavioral ecologyof chimpanzees in Gombe National Park, Tanza-nia. Ph.D. dissertation, Cambridge University.33 Wrangham RW, Smuts B. 1980. Sex differ-ences in behavioral ecology of chimpanzees inGombe National Park, Tanzania. J Reprod Fertil(Suppl) 28:13–31.34 Alvard M. 1995. Intraspecific prey choice byAmazonian hunters. Curr Anthropol 36:789–818.35 Stiner M. 1991. An interspecific perspectiveon the emergence of the modern human preda-tory niche. In: Stiner M, editor. Human predatorsand prey mortality. Boulder: Westview Press. p149–185.36 Martorell R, Lechtig HA, Yarbrough C, Del-gado H, Klein RE. 1976. Protein-calorie supple-mentation and postnatal growth: a review offindings from developing countries. Arch Lati-noam Nutr J 26:115–128.37 Mora JO, Herrera MG, Suescun J, DenavarroL, Wagner M. 1981. The effects of nutritional

supplementation on physical growth of childrenat risk of malnutrition. Am J Clin Nutr 34:1885–1892.38 Coop R, Holmes P. 1996. Nutrition and para-site interaction. Int J Parasitol 26:951–962.39 Baertl JM, Morales G, Verastegui G, GrahamGG. 1970. Supplementation for entire communi-ties: growth and mortality of infants and chil-dren. Am J Clin Nutr 23:707–715.40 Kielmann AA, Taylor CE, Parker RL. 1978.The Narangwal nutrition study: a summary re-view. Am J Clin Nutr 31:2040–2052.41 Lee RB, De Vore I. 1968. editors. Man thehunter. Aldine, Chicago: 1968.42 Lee RB. 1979. The !Kung San: men, women,and work in a foraging society. Cambridge: Cam-bridge University Press.43 Yellen J. 1977. Archaeological approaches tothe present: models for reconstructing the past.New York: Academic Press.44 Hoffecker JE. 1999. Neanderthals and mod-ern humans in Eastern Europe. Evol Anthropol7:129–141.45 Pitts M, Roberts M. 1997. Fairweather Eden:life in Britain half a million years ago as revealedby the excavations at Boxgrove. London: Cen-tury.46 Jones R. 1990. From Kakadu to Kutikina: thesouthern continent at 18,000 years ago. In: Gam-ble C, Soffer O, editors. The world at 18,000 BP.London: Unwin Hyman.47 Close A, Wendorf F. 1990. North Africa at18,000 BP. In: Gamble C, Soffer O, editors. Theworld at 18,000 BP. London: Unwin Hyman.48 Hawkes K. 1991. Showing off: tests of an hy-pothesis about men’s foraging goals. Ethol Socio-biol 12:29–54.49 O’Connell JF, Hawkes K, Blurton Jones NG.1999. Grandmothering and the evolution ofHomo erectus. J Hum Evol 36:461–485.50 Gurven M, Allen-Arave W, Hill K, HurtadoAM. 2000. It’s a wonderful life: signaling gener-osity among the Ache of Paraguay. Evolution andHuman Behavior, in press.51 Gurven M, Hill K, Kaplan H, Hurtado M,Lyles R. 1999. Food transfers among Hiwi forag-ers of Venezuela: tests of reciprocity.52 Hames R. 1987. Relatedness and garden laborexchange among the Ye’Kwana: a preliminaryanalysis. Ethol Sociobiol 8:259–284.53 Hurtado AM, Hill K, Kaplan H, Hurtado I.1992. Tradeoffs between female food acquisitionand childcare among Hiwi and Ache foragers.Hum Nat 3:185–216.54 Hill K, Hurtado AM. 1991. The evolution ofreproductive senescence and menopause in hu-man females. Hum Nat 2:315–350.55 Hurtado AM, Hill K. 1992. Paternal effects onchild survivorship among Ache and Hiwi hunter-gatherers: implications for modeling pair-bondstability. In: Hewlett B, editor. Father-child rela-tions: cultural and biosocial contexts. Haw-thorne, NY: Aldine de Gruyter. p 31–56.56 Pennington R, Harpending H. 1988. Fitnessand fertility among the Kalahari !Kung. Am JPhys Anthropol 77:303–319.57 Blurton Jones NG, Hawkes K, O’Connell JF.1996. The global process, the local ecology: howshould we explain differences between the Hadzaand the !Kung? In: Kent S, editor. Cultural diver-sity in twentieth century foragers. Cambridge:Cambridge University Press.58 Kaplan H, Dove H. 1987. Infant developmentamong the Ache of Eastern Paraguay. Dev Psy-chol 23:190–198.59 Kaplan HK. 1996. A theory of fertility andparental investment in traditional and modernhuman societies. Yearbook Phys Anthropol 39:91–136.

ARTICLE Evolutionary Anthropology 183

60 Hawkes K, O’Connell JF, Blurton Jones HG.1995. Hadza children’s foraging: juvenile depen-dency, social arrangements and mobility amonghunter-gatherers. Curr Anthropol 36:688–700.

61 Hill K. 1988. Macronutrient modifications ofoptimal foraging theory: An approach using in-difference curves applied to some modern forag-ers. Hum Ecol 16:157–197.

62 Hurtado AM, Hill K, Kaplan H, Lancaster J.1999. The origins of the sexual division of labor.In preparation.

63 Hawkes K, O’Connell JF, Blurton Jones NG.2000. Hadza hunting and the evolution of nu-clear families, submitted.

64 Allman J, McGuiness E. 1983. Visual cortex inprimates. In: Comparative primate biology. NewYork: Alan R. Liss. p 279–326.

65 Allman J. 1987. Primates, evolution of thebrain. In: Gregory RL, editor. The Oxford com-panion to the mind. Oxford: Oxford UniversityPress. p 663–669.

66 Allman J, McLaughlin T, Hakeem A. 1993.Brain weight and life-span in primate species.Proc Natl Acad Sci 90:118–122.

67 Hakeem A, Sandoval GR, Jones M, Allman J.1996. Brain and life span in primates. In: BirrenJE, Schaie KW, editors. Handbook of the psy-chology of aging. San Diego: Academic Press. p78–104.

68 Roff DA. 1992. The evolution of life histories.London: Chapman and Hall.

69 Altman JC. 1987. Hunter-gatherers today: anaboriginal economy of North Australia. Can-berra: Australian Institute of Aboriginal Studies.

70 Bahuchet S. 1990. Food sharing among thepygmies of Central Africa. Afr Stud Monogr 11:27–53.

71 Endicott K. 1988. Property, power and con-flict among the Batek of Malaysia. In: Ingold T,Riches D, Woodburn J, editors. Hunters andgatherers: property, power and ideology. NewYork: St. Martin’s Press. p 110–128.

72 Kaplan H, Hill K. 1985. Food-sharing amongAche foragers: Tests of explanatory hypotheses.Curr Anthropol 26:223–245.

73 Lee R. 1972. The !Kung bushmen of Bot-swana. In: Bicchieri MG, editor. Hunters andgatherers today. New York: Holt, Rinehart andWinston. p 326–368.

74 Marshall L. 1976. Sharing, talking and giving:relief of social tensions among the !Kung. In: LeeR, DeVore I, editors. Kalahari hunter-gatherers.Cambridge: Harvard University Press. p350–371.

75 Winterhalder B. 1996. Social foraging and thebehavioral ecology of intragroup resource trans-fers. Evol Anthropol 5:46–57.

76 Goodall J. 1986. The chimpanzees of theGombe: patterns of behavior. Cambridge: Cam-bridge University Press.

77 McGrew WC. 1996. Dominance status, foodsharing, and reproductive success in chimpan-zees. In: Weisner P, Schiefenhovel W, editors.Food and the status quest. Providence, RI:Berghahn Books. p 39–46.

78 Sugiyama L, Chacon R. 2000. Effects of ill-ness and injury on foraging among the Yora andShiwar: pathology risk as adaptive problem. In:Cronk L, Irons W, Chagnon N, editors. Humanbehavior and adaptation: an anthropological per-spective. New York: Aldine de Gruyter. In press.

79 Bailey RC. 1991. The behavioral ecology ofEfe pygmy men in the Ituri Forst, Zaire. AnnArbor: University of Michigan. Museum of An-thropology.

80 Fleagle JG. 1999. Primate adaptation and evo-lution. New York: Academic Press.

81 Wrangham RW. 1979. On the evolution of apesocial systems. Soc Sci Information 18:335–368.

82 Hiraiwa-Hasegawa M. 1990. A note on theontogeny of feeding. In: Nishida T, editor. Thechimpanzees of the Mahale Mountains: sexualand life history strategies. Tokyo: Tokyo Univer-sity Press. p 277–283.83 Howell N. 1979. Demography of the Dobe!Kung. New York: Academic Press.84 Austad S, Fisher KE. 1991. Mammalian ag-ing, metabolism, and ecology: evidence from thebats and marsupials. J Gerontol B:47–53.85 Austad S, Fischer KE. 1992. Primate longev-ity: its place in the mammalian scheme. Am JPrimatol 28:251–261.86 Hladik CM. 1988. Seasonal variations in foodsupply for wild primates. In: de Garine I, Harri-son GA, editors. Coping with uncertainty in foodsupply. Oxford: Clarendon Press. p 26–32.87 Milton K. 1993. Diet and primate evolution.Sci Am 269:70–77.88 Altmann SA. 1998. Foraging for survival:yearling baboons in Africa. Chicago: Universityof Chicago Press.89 Milton K. 1988. Foraging behaviour and theevolution of primate intelligence. In: Byrne RW,Whiten A, editors. Machiavellian intelligence.Oxford: Clarendon Press. p 285–305.90 Terborgh J. 1983. Five New World primates: astudy in comparative ecology. Princeton: Prince-ton University Press.91 Barton RA. 1998. Visual specialization andbrain evolution in primates. Proc R Soc LondonB 265:1933–1937.92 Barton RA. 1999. The evolutionary ecology ofthe primate brain. In: Lee PC, editor. Primatesocioecology. Cambridge: Cambridge UniversityPress.93 Byrne R. 1995. The thinking ape: evolutionaryorigins of intelligence. Oxford: Oxford UniversityPress.94 Menzel CR. 1997. Primates’ knowledge oftheir natural habitat. In: Whiten A, Byrne R, ed-itors. Machiavellian intelligence II. Cambridge:Cambridge University Press. p 207–239.95 Byrne R. 1997. The technical intelligence hy-pothesis: an additional evolutionary stimulus tointelligence? In: Whiten A, Byrne R, editors. Ma-chiavellian Intelligence II. Cambridge: Cam-bridge University Press. p 289–311.96 Andrews P, Martin L. 1992. Hominoid dietaryevolution. In: Whiten A, Widdowson EM, editors.Foraging strategies and natural diet of monkeys,apes and humans. Oxford: Clarendon Press. p39–50.97 Wrangham RW, Conklin-Brittain, Hunt KD.1998. Dietary response of chimpanzees and cer-copithecines to seasonal variation in fruit abun-dance. I. antifeedants. Int J Primatol 19:949–970.98 Conklin-Brittain NL, Wrangham RW, HuntKD. 1998. Dietary response of chimpanzees andcercopithecines to seasonal variation in fruitabundance. II. micronutrients. Int J Primatol 26:951–962.99 Gibson KR. 1986. Cognition, brain size andthe extraction of embedded food resources. In:Else JG, Lee PC, editors. Primate ontogeny, cog-nition, and social behavior. Cambridge: Cam-bridge University Press. p 93–105.100 Gibson KR. 1990. New perspectives on in-stincts and intelligence: brain size and the emer-gence of hierarchial mental constructional skills.In: Parker S, Gibson K, editors. “Language” andintelligence in monkeys and apes: ComparativeDevelopmental Perspectives. Cambridge: Cam-bridge University Press.101 Whiten A, Goodall J, McGrew WC, NichidaT, Reynolds V, Sugiyama Y, Tutin CEG, Wrang-ham R, Boesch C. 1999. Cultures in chimpan-zees. Nature 399:682–685.102 Lancaster J, Kaplan H, Hill K, Hurtado AM.1999. The evolution of life history, intelligence

and diet among chimpanzees and human forag-ers. In: Tonneau F, Thompson NS, editors. Evo-lution, culture and behavior. New York: PlenumPress.103 Jerison H. 1973. Evolution of the brain andintelligence. New York: Academic Press.104 Jerison HJ. 1976. Paleoneurology and theevolution of mind. Sci Am 234:90–101.105 Clutton-Brock TH, Harvey PH. 1980. Pri-mates, brains and ecology. J Zool, London 109:309–323.106 Milton K. 1981. Distribution patterns oftropical plant foods as an evolutionary stimulusto primate mental development. Am Anthropol83:534–548.107 Barton RA, Dunbar RIM. 1997. Evolution ofthe social brain. In: Whiten A, Byrne RW, editors.Machiavellian intelligence II. Cambridge: Cam-bridge University Press. p 240–263.108 Dunbar RIM. 1996. The social brain hypoth-esis. Evol Anthropol 6:178–190.109 Harcourt AH. 1988. Alliances in contests andsocial intelligence. In: Byrne R, Whiten A, edi-tors. Machiavellian intelligence. Oxford: Claren-don Press. p 132–152.110 Harcourt AH. 1988. Cooperation as a com-petitive strategy in primates and birds. In: Ito Y,Brown JL, editors. Animal societies: theories andfacts. Tokyo: Japan Scientific Societies Press. p147–157.111 Walters JR, Seyfarth RM. 1987. Conflict andcooperation. In: Smuts BB, Cheney DL, SeyfarthRM, Wrangham RW, Struhsaker TT, editors. Pri-mate societies. Chicago: University of ChicagoPress. p 306–318.112 de Waal F. 1996. Conflict as negotiation. In:McGrew WC, Marchant LF, Nishida T, editors.Great ape societies. Cambridge: Cambridge Uni-versity Press. p 159–172.113 Harcourt AH. 1987. Dominance and fertilityamong female primates. J Zool 213:471–487.114 Smith BH. 1991. Dental development andthe evolution of life history in Hominidae. Am JPhys Anthropol 86:157–174.115 Foley RA, Lee PC. 1992. Ecology and ener-getics of encephalization in human evolution. In:Whiten A, Widdowson EM, editors. Foragingstrategies and natural diet of monkeys, apes, andhumans. Oxford: Oxford University Press. p 63–72.116 Kaplan H, Gangestad S, Muller TC, Lan-caster JB. 2000. The evolution of primate lifehistories and intelligence. Submitted.117 Cowlishaw G, Dunbar RIM. 1991. Domi-nance rank and mating success in male primates.Anim Behav 41:1045–1056.118 de Waal F. 1992. Intentional deception inprimates. Evol Anthropol 1:86–92.119 de Waal F. 1994. Overview: culture and cog-nition. In: Wrangham RW, McGrew WC, DeWaal F, Helte P, editors. Chimpanzee cultures.Cambridge: Harvard University Press. p 263–266.120 Smith BH. 1993. The physiological age ofKNM-WT 15000. In: Walker A, Leakey R, editors.The Nariokotome Homo erectus skeleton. Cam-bridge: Harvard University Press. p 196–220.121 Leonard WR, Robertson ML. 1997. Compar-ative primate energetics and hominid evolution.Am J Phys Anthropol 102:265–281.122 Potts R. 1998. Variability selection in Hom-inid evolution. Evol Anthropol 7:81–96.123 Aiello L, Wheeler P. 1995. The expensive-tissue hypothesis: the brain and the digestive sys-tem in human and primate evolution. Curr An-thropol 36:199–221.124 Galdikas BMF, Wood JW. 1990. Birth spac-ing patterns in humans and apes. Am J PhysAnthrop 83:185–192.125 Piers LS, Diggavi SN, Thangam S, Van Raaij

184 Evolutionary Anthropology ARTICLE

JM, Shetty PS, Hautvast JG. 1995. Changes inenergy expenditure, anthropometry, and energyintake during the course of pregnancy and lacta-tion in well-nourished Indian women. Am J ClinNutr 61:501–513.126 Lawrence M, Whitehead RG. 1988. Physicalactivity and total energy expenditure of child-bearing Gambian village women. Eur J Clin Nutr42:145–160.127 Poppitt SD, Prentice AM, Jequier E, SchutzY, Whitehead RG. 1993. Evidence of energy spar-ing in Gambian women during pregnancy: a lon-gitudinal study using whole-body calorimetry.Am J Clin Nutr 57:353–364.128 Lancaster JB. 1978. Carrying and Sharing inhuman evolution. Human Nature Magazine1:82–89.129 Lancaster JB. 1997. The evolutionary historyof human parental investment in relation to pop-ulation growth and social stratification. In: Go-waty PA, editor. Feminism and evolutionary bi-ology. New York: Chapman and Hall. p 466–489.130 Hawkes K, O’Connell JF, Blurton Jones NG,Alvarez H, Charnov EL. 1998. Grandmothering,menopause, and the evolution of human life his-tories. Proc Natl Acad Sci 95:131 Bennett PM, Harvey PH. 1985. Relativebrain size and ecology in birds. J Zool London(A) 207:151–169.132 Isaac GL. 1978. Food sharing in human evo-lution: archaeological evidence from the Plio-Pleistocene of East Africa. J Anthropol Res 34:311–325.133 Walker A, Shipman P. 1989. The cost of be-coming a predator. J Hum Evolution 18:373–392.134 Jones K. 1984. Hunting and scavenging byearly hominids: a study in archeological methodand theory. University of Utah.

135 Sillen A, Hall G, Armstrong R. 1995. Stron-tium calcium ratios (Sr/Ca) and strontium isoto-pic ratios (87Sr/86Sr) of Australopithcus robustusand Homo sp. from Swartkrans. Hum Evol 28:277–285.136 Blurton Jones NG, Marlowe FW. 1999. Theforager Olympics: does it take 20 years to be-comes a competent hunter-gatherer? Salt LakeCity: Human Behavior and Evolution Society.137 Bock J. 2000. The socioecology of children’sactivities: a new model of children’s work andplay. Am Anthropol. Submitted.138 Hill K, Kaplan H. 1999. Life history traits inhumans: theory and empirical studies. Ann RevAnthropol 28:397–430.139 Sugiyama Y. 1989. Population dynamics ofchimpanzees at Bossou, Guinea. In: Heltne P,Marquandt L, editors. Understanding chimpan-zees. Cambridge: Harvard University Press. p134–145.140 Pusey A. 1990. Behavioral changes at adoles-cence in chimpanzees. Behaviour 115:203–246.141 Tutin CEG. 1994. Reproductive success sto-ry: variability among chimpanzees and compar-isons with gorillas. In: Wrangham RW, McGrewWC, DeWaal F, Helte P, editors. Chimpanzee cul-tures. Cambridge: Harvard University Press. p181–193.142 Wallis J. 1997. A survey of reproductive pa-rameters in the free-ranging chimpanzees ofGombe National Park. J Reprod Fertil 109:297–307.143 Nishida T, Takasaki H, Takahata Y, editors.1991. The chimpanzees of the Mahale Moun-tains: sexual and life history strategies. Tokyo:University of Tokyo Press.144 Bose S. 1964. Economy of the Onge of LittleAndaman. Man in India 44:298–310.

145 Meehan B. 1982. Shell bed to shell midden.Canberra: Australian Institute of AboriginalStudies.146 Silberbauer G. 1981. Hunter and habitat inthe central Kalahari Desert. Cambridge: Cam-bridge University Press.147 Tanaka J. 1980. The San, hunter-gatherersof the Kalahari: a study in ecological anthropol-ogy. Tokyo: Tokyo University Press.148 Hurtado AM, Hill K. 1986. Early dry seasonsubsistence ecology of the Cuiva (Hiwi) foragersof Venezuela. Hum Ecol 15:163–187.149 Hill K. 1983. Adult male subsistence strate-gies among Ache hunter-gatherers of EasternParaguay. Ph.D. dissertation, University of Utah.150 McArthur M. 1960. Food consumption anddietary levels of groups of aborigines living onnaturally occurring foods. In: Mountford CP, ed-itor. Records of the American-Australian scien-tific expedition to Arnhem Land. Melbourne:Melbourne University Press. p 90–135.151 Politis G. 1996. Nukak. Columbia: InstitutoAmazonico de Investigaciones Cientificas-SINCHI.152 Wrangham RW, Chapman CA, Clark-ArcadiAP, Isabirye-Basuta G. 1996. Social ecology toKanyawara chimpanzees: implications for un-derstanding the costs of great ape groups. In:McGrew W, Marchant L, Nishida T, editors.Great ape societies. Cambridge: Cambridge Uni-versity Press. p 45–57.153 Uehara S. 1997. Predation on mammals bythe chimpanzee (Pan troglodytes): a review. Pri-mates 38:193–214.154 Stanford CB. 1998. Chimpanzee and redcolobus: the ecology of predator and prey. Cam-bridge: Harvard University Press.

© 2000 Wiley-Liss, Inc.

MEETINGS OF INTEREST

November 15-19, 2000American Anthropological Association99thAnnual MeetingSan Francisco, California

Contact:AAA Meetings Department4350 N Fairfax Dr., Suite 640Arlington, VA 22203-1620Tel: 703 528 1902 ext. 2E-mail: jmeier@aaanet.org

January 7-12, 2001XVIIIth Congress of the InternationalPrimatological SocietyAdelaide, Australia

Theme is comparisons and parallel-isms between primates and marsupi-als

Contact:Conventions WorldwidePO Box 44Rundle MallSA 5000, AustraliaTel: 161 8 8363 0068Fax: 161 8 8363 0354

E-mail: satconv@camtech.net.auFor more information on accomoda-tions, air travel, registration, call forpapers, and symposium registrationvisit the website www.primates.on-.net

April 18-22, 2001Society for American Archaeology(SAA) Annual MeetingNew Orleans, Louisiana

Contact:Society for American Archaeology900 Second St., NE #12Washington, D.C. 20002-3557Web site: http://www.saa.orgVisit the SAA website for details re-garding symposia requests and pre-sentation submissions.

July 21-26, 20016th International Congress of VertebrateMorphologyUniversity of Jena, Germany

Contact:ICVM-6

Institute of Systematic Zoology andEvolutionary BiologyFriedrich-Schiller-University of JenaErbertstaß 1D-07743 JENAGermanyTel:149 3641 949155Fax: 1 49 3651 949152E-mail: icvm6@pan.zoo.uni-jena.deFor more information visit the web-site: http://www.sgiloco.zoo.uni-jena/icvm-6.html

March 20-24, 2002Society for American Archaelogy (SAA)Annual MeetingDenver, Colorado

Contact:Society for American Archaeology900 Second St., NE #12Washington, D.C. 20002-3557Web site: http://www.saa.orgVisit the SAA website for details re-garding symposia requests and pre-sentation submissions.

MEETINGS OF INTEREST Evolutionary Anthropology 185