Antiquity of Postreproductive Life: Are There Modern ...hawkes/Blurton Jones... · Antiquity of Postreproductive Life: Are There Modern Impacts on Hunter-Gatherer Postreproductive

Antiquity of Postreproductive Life: Are There Modern Impactson Hunter-Gatherer Postreproductive Life Spans?

NICHOLAS G. BLURTON JONES,1* KRISTEN HAWKES,2 AND JAMES F. O�CONNELL3

1Department of Anthropology, and Graduate School of Education, University of CaliforniaLos Angeles, Los Angeles California2Department of Anthropology, University of Utah, Salt Lake City, Utah3Department of Anthropology, and Archaeological Center, University of Utah, Salt Lake City,Utah

ABSTRACT Female postreproductive life is a striking feature of human life history and therehave been several recent attempts to account for its evolution. But archaeologists estimate that in thepast, few individuals lived many postreproductive years. Is postreproductive life a phenotypic out-come of modern conditions, needing no evolutionary account? This article assesses effects of themodern world on hunter-gatherer adult mortality, with special reference to the Hadza. Evidencesuggests that such effects are not sufficient to deny the existence of substantial life expectancy at theend of the childbearing career. Data from contemporary hunter-gatherers (Ache, !Kung, Hadza)match longevity extrapolated from regressions of lifespan on body and brain weight. Twenty or sovigorous years between the end of reproduction and the onset of significant senescence does requirean explanation. Am. J. Hum. Biol. 14:184–205, 2002. � 2002 Wiley-Liss, Inc.

Almost alone among animals, modernhumans live long after the age at whichfemales cease to reproduce. Because thisphenomenon seems at first sight to fly inthe face of theoretical predictions about se-nescence (Hamilton, 1966), it has long at-tracted the attention of anthropologists andbiologists who offer adaptationist accountsof its evolution (Williams, 1957; Kaplan,1997; Kaplan et al., 2000; Hawkes et al.,1997). But what if there is really nothing toexplain? What if postreproductive survivalis so recent that we cannot believe it re-sulted from natural selection?

There are many critiques of the compet-ing ideas about the evolution of post-repro-ductive life, for example, Packer et al.(1998) and Blurton Jones et al. (1999). Butin this article we attend to the most basiccriticism—that postreproductive life hascharacterized human life history for tooshort a time to require an evolutionary ex-planation. Some suggest it is a product ofmodern civilization, a novelty of the past200 years or so (Washburn, 1981), whereasothers suggest it is a product of agriculturalor industrial subsistence. Austad (1997)presents the contrast between modernhunter-gatherers (!Kung and Ache), andearlier populations and challenges us toresolve the contradiction between estimatesof postreproductive survivorship frompaleodemography and from modern hunter-

gatherers. Austad offers as representativeof prehistoric populations the Weiss (1973)life tables for the paleolithic and the well-known report of Libben by Lovejoy et al.(1977). Libben is convenient to discuss be-cause of Howell�s (1982) examination of itsdemographic implications. Libben repre-sents an agricultural population and someview early agricultural populations as hav-ing suffered more mortality than the popu-lations of foragers they replaced (Cohen,1989; but see Wood et al., 1992).

There is another family of argumentswhich suggest that postreproductive lifeneeds no adaptationist explanation. Theseaccount for postreproductive life as a by-product of other processes. In an early ex-ample, Weiss (1981) suggests that sincehuman lifespan falls within the rangepredicted for human body size, it requiresno further explanation.

The idea that postreproductive life is aproduct of the modern world deserves at-tention, not only because paleodemographysuggests very few survivors past the 40s, but

Contract grant sponsors: The National Science Foundation,the Swan Fund, B. Bancroft, the Leakey Foundation (by a grantto Frank Marlowe), the University of Utah, and the Universityof California Los Angeles.*Correspondence to: Nicholas G. Blurton Jones, anthropology,

UCLA, Los Angeles, CA 90095-1553Received 11 September 2000; Revision received 1 October

2001; Accepted 5 October 2001

AMERICAN JOURNAL OF HUMAN BIOLOGY 14:184–205 (2002); DOI 10.1002/ajhb.10038

ª 2002 Wiley-Liss, Inc.

also because basing a species characteristiconmodernhunter-gatherersmayneglect theinfluence of the modern world upon thesepeoples. Perhaps modern hunter-gatherershavemore access tomodernmedicine than isreadily apparent, perhaps trade with theirneighbors is strongly beneficial, and notdisadvantageous and exploitative, as somehave argued. Howell (1979) found lowermortality among !Kung born after 1950,when the herder population of the Dobe areaincreased, changing the economic opportu-nities available to !Kung. How general issuch an effect? Can such changes haveshaped mortality in the Ache or Hadza? Dothe answers change our picture of the dura-tion of postreproductive life?

Descriptions of characteristics of H. sap-iens (or earlier forms) have been based onthree main lines of evidence:

1) Characteristics of modern populations.Special attention is often given to contem-porary foragers on the assumption thattheir economy more closely displays thecosts, benefits, and trade-offs crucial to theperiod in which the species evolved itscharacteristics. In the case of demographyand life history, it is clear that agricultureand industrialization had effects, so specialattention should be given to hunters andgatherers. Here we add new data on Hadzamortality to two well-studied cases, the!Kung (Howell, 1979) and the Ache (Hilland Hurtado, 1996). Expectation of life ofwomen age 45 is between 19 and 22 years inthese populations (Table 1).

2) Quantitative comparative studies inwhich a character of interest is extrapo-lated from others less in question. Weissand others (Weiss, 1981; Sacher, 1959,

1975; McHenry, 1994) showed a correla-tion between body size and longevity andextrapolated early human longevity fromestimates of body size. Recently, Hammerand Foley (1996) assembled data fromanthropoid primates and used regressionsto predict lifespan of several hominidsfrom body and brain weights. After ex-amining the impact of modern conditionson contemporary hunter-gatherer mortali-ty, we compare the figures with the re-sults of Hammer and Foley (1996).

3) Archaeological data sometimes providedirect evidence about a trait. But in the caseof postreproductive life archaeological re-ports have offered a striking contrast withthe first two procedures. As yet, everypaleodemographic account suggests a trun-cation of human lifespan close to the age atwhich reproduction ceases. Archaeologistsand physical anthropologists actively in-vestigate several limitations of these re-constructions. The distribution of age atdeath, the measure from which paleode-mographers reconstruct demography, ismore sensitive to fertility than to mortality(Milner et al., 1989) and is highly sensitiveto rate of population change (Howell, 1979).Errors in age estimation, particularly of oldand active people, are quite severe, for ex-ample, comparison of documented age withage assessed from the skeletons in the Spi-talfields study (Molleson et al., 1993; Keyet al., 1994). Extreme differential loss ofskeletons of the young and old was shownwhen Walker et al. (1988; see also Walker,1995) compared documentary and archaeo-logical evidence from mission cemeteries inCalifornia. Analysis of 32 skeletons foundtwo individuals of estimated age >45 (about

TABLE 1. Published demographic values for !Kung, Hadza, and Ache

!Kung Hadza Ache

Population sizea 895/841/796->454 750/1000 550TFRb 4.7 6.2 8.2Rate of increasec 0.0026 0.013 0.025% over 60d 8.5 8.8 5.5Infant mortalitye 0.20/0.15 0.21 0.12–0.18Life expectancy at birthf 30.0 32.5 37.1Life expectancy for female age 45f 20.0 21.3 22.1aPopulation size. !Kung: Howell (1979) pp. 18, 30; Hadza: 750 in 1985 census, 1,000 ever recorded. Eastern Hadza only. Ache: Hill andHurtado (1996) Fig. 3.1 Northern Ache in 1970.bTFR: Howell (1979) table 6.1 and p 123. Hadza Dyson (1977). Ache, H&H p 253.cRate of increase: !Kung, Howell 1979:214 Table 11.1, pre-1950. Hadza Blurton Jones et al. 1992; Ache Hill and Hurtado (1996) p,468.dPercent over 60: !Kung 8.5 from Howell (1979), table 2.1. Hadza, Blurton Jones et al. table 1. Ache from p 141, table 4.4.eInfant mortality: Howell (1979), table 4.1 from reproductive history interviews / table 4.6 for 1964–73. Hadza: from reproductivehistory interviews. Model North 6 gives 20.5 Ache; H&H p 187 and table 6.1.fLife expectancy at birth and at 45: !Kung from Coale and Demeny (1983) Model West 5. Hadza from Coale and Demeny Model North6. Ache: H&H table 6.1.

POSTREPRODUCTIVE LIFE SPAN 185

5%). This is similar to many paleodemog-raphy results. But written records showedthat 539 individuals >45 had been buriedthere, 37% of the buried population. Walkeret al. (1988) point out that conditions at thesite did not favor good preservation ofskeletons. State of preservation may bebetter at other sites, but sites relevant tothe current topic are several orders ofmagnitude older. Claims of good preserva-tion should be supported by objective evi-dence. Walker (1995) suggests a method forquantifying state of preservation and showsthat if this is applied to the Californiaskeletal sample the distribution of ages atdeath conforms to a pattern more in linewith contemporary life tables and the writ-ten records. Paleodemographers and ar-chaeologists continue to work on improvingestimates of age at death, but apparentlyless on finding ways to assess the impor-tance of differential preservation of theyoung and the old. In contrast to thesetasks, this article attempts to evaluate waysin which the modern world could inflate thesurvivorship of modern hunter-gathereradults.

Contemporary hunter-gatherers live in aworld of farmers and herders who are inturn more or less effectively connected to aglobal economy and modern technology.Although it is easy to overemphasize theselinks, whose influence seems fleeting andtenuous when viewed from remote ruralcorners of developing countries, the situa-tion of modern hunter-gatherers might bedifferent in many ways from that of hunt-ers in a world of hunters. Examining twosub-Saharan populations (!Kung and Had-za, with special emphasis on new datafrom the Hadza) and one south Americanpopulation (Ache) we try to assess the ex-tent and effects of 1) errors in estimates ofindividual ages; 2) access to modern medi-cine; 3) access to modern hygiene and fa-cilities such as piped water, which aresometimes viewed as having greater effectsthan medicine; 4) law and order; 5) thepossibility that old people take shelter withfarmers, enabling them to live longer thanthey might among their more mobile kin;6) the possibility that foragers may obtainfood from farmers at critical times, andagain escape mortality in this way; 7) re-duction in predation as wild animals be-came more scarce; 8) outside ideologicalinfluences on leaving the sick or elderly

behind; and 9) infant nutrition and adultmortality (Barker, 1997). We then attempta crude ‘‘accounting,’’ developing a ‘‘worst-case’’ scenario for Hadza survivorship intoand during postreproductive life. We dis-cuss the difficulty of estimating the effectsof epidemics and warfare.

ETHNOGRAPHIC BACKGROUND

Hunters and gatherers

While our working category of ‘‘hunter-gatherer’’ is simply economic (people whoduring the period of observation acquirenearly all of their food by hunting wild ani-mals and foraging for wild plant foods) itneedsqualifying for !KungandHadza. !Kungincreasingly used resources fromHerero andTswanaherders during theperiod coveredbyHowell (1979), changing from apparentlyalmost no people using these resources toperhaps more than half the !Kung usingsome unmeasured amount of food acquiredfrom herders. Howell sets 1950 as a turningpoint in this process and presents data onmortality of children born before 1950 andthose born in and after 1950.

Hadza have been closely surrounded byneighbors and traded with them for at leastthe last 100 years. Although most reportsconcern receipt of cloth, pots, and iron,Hadza could also have obtained food if theyneeded it. Today individual Hadza fre-quently work for farmers guarding maizefields, or harvesting sweet potatoes, and allHadza sometimes beg off fanners and tradehoney with herders. Hadza use of farmfoods has varied greatly over the last 40years. Intermittent imposed settlementsincluded delivery of maize, irregularly andunreliably. Some Hadza have spent longperiods living in a settlement or village, butmost did not and few do today. Much of thetime almost no food was provided to settle-ments. Hadza trying to modernize by livingin larger, more stable communities still fedthemselves from the bush. Some Hadza livenear villagers (farmers) and others far fromthem. In 1985–1986, working in one of themost remote locations, Hawkes et al. (1997)and found that only 5% of food came fromfarmers (carried back from visits to villag-es). In 1995–1996, Marlowe (1997) workedin a variety of camps including some verynear villages and farmers. About 11% ofHadza food came from agricultural, non-

186 N.G. BLURTON JONES ET AL.

wild resources. Among Hadza who lived inMunguli, some farmed successfully untilthey all left in 1990. Thus, there has beengreat variation in Hadza reliance on bush orfarm food and there is some possibility thataccess to farm food has affected Hadzamortality. We examine data on this issuebelow.

Before peaceful contact and settlement in1971, Ache obtained all their food from wildresources, except for whatever they gotfrom the occasional and often fatal practiceof raiding farmers’ fields.

!Kung. The !Kung of northwestern Bot-swana and northeastern Namibia, and theirlifestyle of the1950s to 1970s, arewell knownto anthropologist readers. Howell�s (1979)demographic study centered on a study pop-ulation of some454 individuals out of aZhun/twa speaking population of some 800–900.Nearly 10% of the population were >60 yearsold, total fertility rate (TFR) was low, 4.7(Table 1). The pre-1950 population was bestrepresented by Coale and Demeny (1983)modelWest 5, but it was evident in the 1960sthat both fertility and mortality were un-dergoing rapid change as !Kung interactedmore closely with Herero and Tswana im-migrants.

Ache. The Ache of Paraguay are alsowell known to anthropological readers (Hilland Hurtado, 1996). Their presettlement,full-time forager lifestyle was much moremobile than the life of the other populationsdiscussed here. Small groups moved almostdaily through the forest, clearing a newsleeping area each evening. Couples andtheir children shared a fire, but people livedin much greater proximity at these over-night camps than in a Hadza or !Kungcamp. In some contrast to !Kung and Had-za, Ache men provided about 80% of thefood by hunting animals of a wide variety ofsizes and taxa. While !Kung are known tohave practiced occasional infanticide andHadza claim never to have heard of such apractice, Ache used to regularly kill one ormore children upon the death of their fa-ther, and sometimes in other circumstanc-es. Hill and Hurtado (1996) describeinformants’ interpretations of this practice.

Hill and Hurtado (1996) present datafrom ‘‘the forest period’’ before Ache werecontacted by Westerners or educated Para-guayans, and during which they lived al-

most entirely off hunted and gathered foods,the contact period of rampant infections anddeaths, and a settlement period duringwhich the population recovered and medicalcare became available. Life expectancy atbirth rose from 37.1 years for females in theforest period to 45.6 in the settlement. Lifeexpectancy at age 45 rose only from 22years in the forest period to 24 years in thesettlement. During the contact period,mortality rose mostly among the young andold, including old females. ‘‘Only a handfulof women over the age of fifty survived thecontact period’’ (Hill and Hurtado, 1996, p.189). These authors attributed the viru-lence of these epidemics to lack of previousexposure to the disease organisms. Thenorthern Ache had kept themselves isolatedfrom their neighbors. But a question of in-terest is whether many forager populationswere so isolated in a world of foragersamong foragers. It will be difficult to assessthe effect of disease reservoir and popula-tion density upon forager mortality. In theforest period 5.5% of people were >60 years,TFR was 8.2, and the population was in-creasing rapidly (Table 1).

Hadza. The eastern Hadza are a groupof 700–800 people who occupy an area of2,500 km2 in the Eastern Rift Valley, southand east of Lake Eyasi, in northern Tan-zania. The climate is warm and dry. Annualrainfall is in the 300–600 mm range, most ofit falling in the 6–7 month wet season (No-vember–May). Vegetation is primarilymixed savanna woodland; medium/largeanimals are locally abundant. Ethno-graphic data on the eastern Hadza is mostavailable in the publications of Woodburn(1964, 1968a, 1970, 1972, 1979, 1982, 1988).The western Hadza are about 250 peoplewho speak the same language but live westof Lake Eyasi. Contact is sufficiently low foreastern Hadza informants to often commentthat they do not know people in the west,although individuals and families do movebetween these areas. Eastern Hadza talk ofthemselves as living in three regions:Mangola-Balai, Siponga, and Tliika.

The language, Hadzane, has been studiedby Woodburn and others and recently bySands (1993, 1995). Linguists agree onlythat its connection to any other Africanlanguage family is very remote. Hadza musthave remained culturally distinct from theirneighbors (who currently represent the


Bantu, Nilotic, and Cushitic language fam-ilies) for many hundreds of years. At thebeginning of this century, it appears thatonly the Hadza occupied this country(Baumann, 1894; Obst, 1912). They appar-ently lived entirely by hunting and gather-ing. Local incursions by non-Hadza pastoraland agricultural groups are recorded asearly as the 1920s and have continued tothe present (McDowell, 1981; Dar es Sa-laam; Woodburn, 1988). Archaeological ev-idence suggests that farmers andpastoralists have been present intermit-tently for several centuries, hunter-gather-ers far longer (Mehlman, 1988). During thepast 50 years various segments of theHadza population have been subjected to aseries of government- and mission-spon-sored settlement schemes designed to en-courage them to abandon the foraging life infavor of farming. None of these schemes hasbeen successful and in every case most ofthe Hadza returned to the bush, usuallywithin a few months. In each instance someHadza avoided settlement and continued tolive as full-time hunter-gatherers.

During the 1990s hunting became pro-gressively more difficult as interference byoutside agencies continued and incursionsby people with other lifestyles increasedand the habitat was degraded. The arrivalof ‘‘ethno-tourism’’ in the late 1990s seemedlikely to herald the imminent submersion ofthe Hadza into a new lower rung of Tanza-nian rural society. But by the end of thecentury it appeared that this developmentmight instead have positive outcomes insome localities.

Many have argued that anthropologiststoo readily separate people into assumedlyisolated ‘‘tribes.’’ This could lead to demo-graphic study of populations that were notdemographic units. This study included asmembers of the Hadza population anyonewho claims Hadzane as his or her mothertongue. By and large this coincides withbeing regarded as Hadza by other Hadza,and by others who would regard themselvesas not Hadza, and with residing in a fairlycircumscribed area. It coincides less tightlywith many other traits, characteristics, orhabits: carrying a bow, knowing the foot-paths, knowing how to hunt and gather,attending epeme dances, using one set ofplace names rather than another, and so on.It coincides with a sharp difference inprobability of marriage—Hadza marry oth-

er Hadza much more often than they marrypeople who do not have Hadza as theirmother tongue. But some Hadza do marryothers (Stevens, 1977); Hadza may be moreculturally separable than genetically sepa-rable from their neighbors.

The field method allowed Hadza to definethemselves indirectly. The censuses andanthropometry was conducted from camp tocamp, finding each camp by asking peoplein the current camp where there is another.Informants were also asked the wherea-bouts of each person listed in the populationregister (‘‘where are they now’’ interviews).The register is a list of information about allthe people encountered in our research, orwho Lars Smith encountered in 1977, andothers before him. We also tried to identifyin this register everyone we saw in a cen-sus. These procedures give clues about thedifficulty of ‘‘becoming a Hadza’’ and ofceasing to ‘‘be a Hadza.’’ Only one adult waslocated, who our field assistant claimed wasnot born a Hadza but has ‘‘become one’’; hespeaks Hadza, his wife is Hadza, he lives asa Hadza and with Hadza, and has for a longtime. There is also one other man whoseems to be in the early steps of the sameprocess. These are two, contrasting withsome 200 adult men, who Hadza regard ashaving been born and raised Hadza. Thereare also a few people recorded earlier whomwe have not met but who Hadza know to bealive. The assistant and others stated ofsome of them have ‘‘gone to the Swahilis,’’and although in some sense apparent toHadza, they are still ‘‘really’’ Hadza, theyshould now be categorized differently fromthe majority.

HADZA DEMOGRAPHY

Blurton Jones et al. (1992) reported re-sults of a census conducted in 1985 by LarsSmith and Blurton Jones. Age structure,survival from an earlier census conductedas part of the International Biological Pro-gram (IBP) by Barnicot, Bennett, andWoodburn, a capture-recapture calculationof population increase, and other measureswere found to fit closely to Model North 6chosen by Dyson (1977) to fit the 1966–1967IBP data. The age structure was signifi-cantly younger than Model West 5 thatHowell (1979) reported as the best fit to the!Kung. Middle-aged and older Hadzaseemed to survive at a higher rate thanexpected from Model 6 North.


We acquired new data on Hadza mortal-ity from repeat censuses and reproductivehistory interviews conducted between 1990and 1999. Some births, deaths, and campcompositions were recorded during behavi-oral and ecological observations conductedduring 1985–1986, fall 1988, and winter1989. At each census, informants are askedfor news of everyone recorded in previouscensuses. This is our main source of infor-mation about deaths. Information on causeof death is sought at the same time but ishard to elicit, vague, and often uninter-

pretable (Table 2). More than one report of adeath is sought. Only rarely has a personreported as dead been seen or reported aliveat a later date. The method of estimatingindividual ages is discussed below.

The fate of the 706 people censused in1985 was tabulated. Fifty-seven individualshad died by 1990. Of the survivors to 1990,a further 68 died by 1995. Year of death wasestimated for each individual who died by1995. If year of death was not known, it wasestimated as one-third of the time betweenthe last report of the person alive and thefirst year in which they were reported dead.Many of these estimated years of death canbe refined in further fieldwork. The numberentering each year of life was then calcu-lated and a life table was constructed (Table3, Fig. 1). Number entering each year wascomputed as in the following examples.Child A was 5 years old at the time of the1985 census and lived until 1988. He wasrecorded as entering the years 5, 6, 7, and 8,and dying by the end of year 8. Old lady Bwas calculated as 65 in 1985 and died in1992. She was recorded as entering years65, 66, 67, 68, 69, 70, 71, 72, and dying atthe end of year 72. Her sister C was

TABLE 2. Causes of death of Hadza 1985–1997

Hadza 1985–1997 Male Female

Respiratory(TB, pneumonia)

5 6

‘‘Oldness’’ 4 5Measles 8 4Other illness 15 10Homicide 3 1Childbirth 0 5Fall from tree 1 0In hospital or jail 2 1‘‘Poisoned’’ or bewitched 3 2Total given a cause* 41 34*Only 75 deaths out of a little more than 125 were ascribed a‘‘cause’’ by informants.

Fig. 1.


TABLE 3. Hadza lifetable 1985–1990

Age Enter Die qx lx probability

0 0 0 0.21 11 23 1 0.0435 0.792 54 3 0.0556 0.75563 77 5 0.0649 0.71374 92 4 0.0435 0.66735 105 7 0.0667 0.63836 117 1 0.0085 0.59587 133 1 0.0075 0.59078 149 2 0.0134 0.58629 163 1 0.0061 0.578410 177 1 0.0056 0.574811 193 5 0.0259 0.571612 191 3 0.0157 0.556813 176 3 0.0170 0.548014 168 0 0 0.538715 168 2 0.0119 0.538716 164 2 0.0122 0.532317 164 0 0 0.525818 161 2 0.0124 0.525819 156 2 0.0128 0.519220 151 2 0.0132 0.512621 145 1 0.0069 0.505822 140 3 0.0214 0.502323 133 1 0.0075 0.491524 132 0 0 0.487825 132 2 0.0151 0.487826 127 1 0.0079 0.480427 122 0 0 0.476728 115 0 0 0.476729 114 1 0.0088 0.476730 109 0 0 0.472531 109 0 0 0.472532 108 0 0 0.472533 106 0 0 0.472534 103 2 0.0194 0.472535 98 0 0 0.463336 94 1 0.0106 0.463337 91 2 0.0219 0.458438 88 1 0.0114 0.448339 90 1 0.0111 0.443240 85 1 0.0118 0.438341 85 3 0.0353 0.433142 79 0 0 0.417843 74 0 0 0.417844 72 0 0 0.417845 73 3 0.0411 0.417846 69 1 0.0145 0.400747 66 1 0.0152 0.394948 64 0 0 0.388949 64 0 0 0.388950 63 3 0.0476 0.388951 60 1 0.0167 0.370452 55 1 0.0182 0.364253 55 1 0.0182 0.357654 54 0 0 0.351155 52 3 0.0577 0.351156 47 2 0.0426 0.330857 46 2 0.0435 0.316758 46 0 0 0.302959 50 2 0.04 0.302960 46 2 0.0435 0.290861 41 3 0.0732 0.278262 38 1 0.0263 0.257863 39 0 0 0.251164 41 0 0 0.251165 41 1 0.0244 0.251166 42 1 0.0238 0.244967 40 2 0.05 0.2391


recorded as 55 in 1985 and survivedthroughout the study period. She was com-puted as entering every year of life from 55to 65 and dying in none of them. Becausethe censuses may undercount the smallestbabies, infant mortality is taken from thereproductive history interviews (proportionof babies born but dying before they couldwalk, age at walking appears to be close tothe 1 year, as found in many populations,including African populations).

Hadza survivorship is close to that forModel North 6 and a little higher thanModel 5 West (!Kung), and very close toprecontact Ache. The Hadza data confirmthe observation of high adult survivorshipand an expectation of life just over 20 yearsfor women 45 years of age. The data willsupport further analysis when a final cen-sus has been completed, the revised methodof estimating ages has been compared withthe original, and year of death has beenestablished more closely.

Now we consider a series of ways inwhich these data (and the data on Acheand !Kung) might conceivably underesti-mate adult mortality. Most have to dowith possible differences between the cir-cumstances of hunter-gatherers in a worldof hunters and hunter-gatherers as stud-ied in the modern world.

Are the age estimates bad?

At the time of study very few !Kung,Ache, or Hadza knew their ages or had hadthem recorded. Could errors in age esti-

mation exaggerate the survivorship of eld-erly contemporary hunter-gatherers andaccount for the difference between con-temporary and archaeological populations?

Howell (1979) estimated ages of !Kung byfirst ranking everyone in order from oldestto youngest. There were two known agepoints and a figure for infant and childmortality that allowed her to choose a modelfrom Coale and Demeny (1985). Howellthen matched the age ranking to the agestructure to obtain age estimates for everyindividual. Howell devised some checks onthe age estimates, as well as several checkson the fit to West 5. Particularly useful isthe age gap between mother and first, sec-ond, third children. The gap did not changewith mother�s age as it might have done ifolder women had their age grossly over- orunderestimated.

Hill and Hurtado (1996) estimated Achebirth dates by a combination of ranking,and informants matching contemporary in-dividuals to the developmental stage ofother individuals at the time of significantevents in the past. For example, A�s motherwas asked how large A was (‘‘which childhere was closest to A’’) when she bore hisnext younger sister. Many of the contem-porary reference children were of knownage. A�s mother might also be asked whoshe most resembled when she bore A and allhis siblings ‘‘which woman here is clos-est...?’’ Hill and Hurtado checked their es-timates in several ways.

Hadza ages were calculated by an ageranking which originally included some 44individuals of known age between 1 and

TABLE 3. (Continued)

Age Enter Die qx lx probability

68 38 2 0.0526 0.227169 36 0 0 0.215270 36 1 0.0278 0.215271 38 1 0.0263 0.209272 39 3 0.0769 0.203773 35 2 0.0571 0.188074 33 5 0.1515 0.177375 28 5 0.1786 0.150476 19 1 0.0526 0.123677 17 0 0 0.117178 15 4 0.2667 0.117179 11 1 0.0909 0.085880 8 1 0.125 0.078081 7 1 0.1429 0.068382 4 0 0 0.058583 3 1 0.3333 0.058584 1 0 0 0.0390


29 years. Ages of four older individualswere estimated by less direct andindependent methods. Rank order and agewere regressed and a polynomial equationused to calculate age of each individual inthe ranking. These estimates also passedthe test of the gap between mother andfirst child (Blurton Jones et al., 1992).

Subsequently, more age marks have beenestablished. Three cases are particularlyimportant for testing estimates of olderages. Dorothea Bleek�s notebooks of hervisit to the Hadza in 1930 (University ofCapetown Library, Manuscripts and Ar-chives Department) contain many partialgenealogies. Most of the people named inthem were remembered by informants inthe 1990s; they include their parents,grandparents, and other kin. Two individ-uals were recognized by informants as stillliving in 1991. Bleek describes both as‘‘boys,’’ and says the youngest was agedabout 8 and she reports his height as 116cm, nearer the height of a 10–12 year-oldHadza child. Supposing this to indicate thatboth were around 10 in 1930 (born 1920)then they were around 71 years in 1991.Their position in the age ranking had giventhem calculated ages of 66 and 74, mean69.5 in 1991, which agrees remarkably withthe new estimate.

Most of the men and women named byKohl-Larsen (1943, 1958) from his fieldworkat the other end of Hadza country between1934–1937 are also recognized by contem-porary Hadza and their descendants arewell known. Most of those named were deadby 1985. If they were adult (>20) in the mid-1930s, thus born around 1915, then theyoungest of these adults would have been70 in 1985. Four men and four women werestill alive in 1985, all ranked among those>60 years. Kohl-Larsen gave no indicationof their ages. But he named a boy (‘‘DerKnabe...’’) who informants claimed theycould recognize and who had been known toLars Smith in the 1970s. Though dead by1985, ‘‘the boy’’ had an older brother alive

(who had been estimated to be 2–5 yearsolder). Suppose the boy was 5 (born, say1930), and his brother 7–10 years old (born1925–28), then the brother was 57–60 in1985. His age calculated from his rank was56 in 1985. This match is also extremelyencouraging. The quest for more calendarmarks and historical information continues.

While the estimates of age are reasonablyreliable for older people, three ways inwhich inaccuracy could affect the conclu-sions about postreproductive life, need dis-cussion.

Assumptions about maximum age. Thefigure obtained for life expectancy at 45years (e45) is affected by the assumptionabout maximum age, particularly the de-cision about which will be the highest agetabulated. For example, Wood (1987) pub-lished a life table for the Gainj with thefinal age group as 60–65. If we assumethat all !Kung in Howell�s Table 4.6 haddied by 60, instead of between 85 and 90,the estimate of e45 for !Kung in the 1960sdeclines from 27.31 to 16.57. Since esti-mates of the age of the older people areprobably the least reliable in any samplewe should take account of the effect ofvariation in estimated oldest age. Theproportion surviving to the highest agegroup tabulated, or to any lower age, is notaffected. We can replace e45 with, for ex-ample (25e45), the number of years livedbetween 45 and 70 by someone whoreaches 45. Table 4 shows such life ex-pectancy figures from age 45 and from age20 (used below) for the three populationsand for the eventual Hadza ‘‘worst case.’’In these populations the high expectationof life at age 45 is not dependent on ahandful of very old individuals.

Difference between model life tables andAche and Hadza. Among the Ache, and lessso among the Hadza, more people age 50–65years appear to survive than in the nearest

TABLE 4. Expectation of life at age 45 with different assumptions about age of oldest individuals

e45 to Tmax e45 to 70 e20 to Tmax e20 to 70

!Kung west 5* 20.0 17.5 34.0 32.4Ache forest period 22.1 19.6 39.8 37.8Hadza 1985–1995 23.0 20.0 41.4 38.1Hadza ‘‘worst case’’ 19.3 17.1 33.7 32.2*From Coale and Demeny (1983) Model West 5. Expectancy to 70 calculated from (Tx-T70)/lx.54.


Coale and Demeny (1983) models. If theHadza 50–70-year-olds were in fact 40–65-year-olds, the fit to the model might be close.But estimates of the age of these olderwomen gave a similar gap between thewoman and her oldest child as observedamong younger women (Blurton Jones et al.,1992). If the age estimates were lowered forthe older women this would no longer holdtrue. The effect on e45 would be quite small,as can be seen by comparing the observedHadza e45 of 23.0 years, with that fromModel 6 North 21.3.

Can the 76 Hadza women over 45 be ex-plained away? It seems impossible thatthere are errors gross enough to remove the76 Hadza women who are over 45 years ofage. Two arguments can be given.

i) Many women calculated to be over 45years have living daughters and grandchil-dren and a few have great-grandchildren.Ages of young women are quite well estab-lished and provide an age at first childbirtharound 18–19 years. There are also anumber of women born before a well-re-membered 1964 earthquake, who are thusnow in their mid-30s. Many of these haveliving mothers, who are thus likely to bewomen over 55 years unless age at firstbirth has recently increased markedly.

ii) If a population is decreasing it willhave an older age pyramid than if it wasgrowing. What if the old women are relicsof a rapidly decreasing population? Even ina population with an age pyramid as shal-low as Libben (Howell, 1982), not dissimilarto many archaeological populations, forageror nonforager, a few might survive beyond45 years. But very few survive, 2% in Ho-well�s calculation of the age structure im-plied by the Libben life table. If the 76 olderHadza women were the relics of such apopulation, it would have had to number8,000 and exist sometime in the lifetime ofthese women. There are many previousaccounts of the Hadza and none indicate apopulation any greater than 800 or so.There is good evidence that the Hadzapopulation has been increasing for the pastseveral decades at a rate that fits well withthe few previous accounts (Blurton Joneset al., 1992).

It must be concluded that neither popu-lation decline, nor errors of age estimationare sufficient to do away with the large ob-served population of older people, or to

greatly affect estimates of survivorship ofthose who reach age 45.

Modern medicine

Did researchers intervene enough to affectthe mortality statistics? Researchers canhave had no effect on Ache survival in theforest. The data were gathered in retro-spective interviews. No researchers werewith the northern Ache before the contactand settlement period (Hill and Hurtado,1996). Howell (1979), discussing the lowmortality of the !Kung between 1964 and1973, considers migration of the less heal-thy to health care on the periphery of !Kungcountry and the possibility that the re-searchers had an effect by giving healthcare and emergency transport.

In fieldwork with the Hadza we respondedto requests for helpwith sickness or injury intwo ways: transporting individuals to ruralhealth clinic or hospital, or giving medicineswhen available. We justified this by the ex-treme isolation of our field camps (by journeytime if not miles) from professional help andthe extremity of life in a bush camp. Itseemed unethical to withhold help fromHadza that we gave, or would have given, toourselves and each other. Did we transportenough sick people to affect the mortalitystatistics? Table 5 shows that the cases werenot so obviously concentrated on accidentalinjuries as Howell indicates for the !Kung(Howell, 1979). Two facts stand out in Table5. We mainly transported men, as also re-ported by Howell (1979). We only transport-ed twowomen, one of thempostreproductive.Most of the people that we transported died;our help was ineffective.

Did we treat enough people againstgenuinely threatening conditions success-fully enough to affect the statistics? Wetabulated the interventions of one of us.Most treatment was given to children. Earinfections were fairly common, as werecuts and scrapes, and we saw one seriousburn. Only one treated child died (3 yearslater). But we suspect that the condition ofonly about two was sufficiently serious tomake it possible that they might have diedwithout our intervention. Older womenasked for medicine for sore eyes and theoccasional cough or abdominal pain. Wheneye infection was obvious and we weregoing to be in camp long enough, we of-fered treatment.


How much access do Hadza (and others)have to modern medicine and do they makeuse of it? Throughout the study period,government clinics have been available nomore than a day and one-half walk from anyHadza camp. Hadza seem extremely reluc-tant to go to these clinics. They are con-vinced that there will be no medicines andthat if there were the attendant would saythere were not, refusing to treat pennilessHadza. Our experience is that there is sometruth, but only some, in these suspicions.When we have accompanied Hadza to theseclinics they have been fairly well receivedand treated. The clinic attendants appear tobe very competent but supply lines are longand very unreliable in the countryside.

Bennett et al. (1973) reported a very dif-ferent picture from the 1960s. During the1966 and 1967 IBP surveys, shortly afterthe establishment of the Yaeda and Mung-uli settlements (now long deserted by mostHadza), 242 out of 491 Hadza interviewedhad at some time sought Western medicalaid, 75% of the people in settlements butonly 40% of eastern Hadza not in a settle-ment, and only 18% of western Hadza. Thisgreater access to medicine may explainwhy, even though informants describe anunusual number of illnesses and deaths atthe settlements, there was no trace in the1985 age structure.

Towards the end of the recent period ofdemographic data collection there havebeen two improvements in access to modernmedicine. Between 1994 and 1998 CUSO(‘‘a Canadian organization for global socialjustice’’) field representatives have mademore frequent visits to one previously iso-

lated bush location and become more in-clined to transport Hadza for medical care.Since 1995 a mission-supported doctor hasbeen stationed in Mangola at the north endof Hadza country and he tries to be acces-sible to Hadza. In other respects medicalcare remains remote and forbidding to theHadza. In towns such as Arusha, a widevariety of modern medicines (antibiotics,chloroquine, and variants) are available inshops without prescription, and increas-ingly in smaller towns. But Hadza still haveno access to such sources.

Reservoir of disease. Modern hunter-gatherers try to isolate themselves fromtheir farmer neighbors. This might makeepidemics less frequent but may make themmore severe. Contact die-offs confirm thepotential severity. In a world of huntersamong hunters isolation may have been lessrigorous, with the result that diseasesspread more easily but perhaps with lessdire consequences.

The modern world differs in two otherways. The surrounding farmer populationsare much more dense than previous foragerneighbors would have been, perhaps im-plying a larger reservoir of disease. On theother hand, these dense modern farmerpopulations have the advantage of immu-nization programs and other public healthefforts. While such efforts have led to theglobal extinction of smallpox, sleepingsickness returned to the Eyasi area after anabsence. It may now be disappearing again.As a result of clearance by herders, theTsetse fly vectors have become noticeablyscarce. Hadza themselves have received few

TABLE 5. Hadza transported by researchers in search of medical attention

ID Sex Age Year Request by Illness Outcome

SE M 65 1985 Family Sleeping sickness Died inhospital

SM M 58 1986 Family Not diagnosed DiedKS M 23 1986 Community Psychosis? Died in

hospitalGM M 50 1986 Self Injured leg Survived

>10 yrsSG F 2 1986 Parents Skin lesions. Scabies AliveNS F 7 1990 Observer and assistant,

weight lossTB Alive

HW M 45 1992 S asked return to campafter failed treatment.

TB? Died

OP M 43 1995 Observer and assistant Clinic gave antibiotics AliveNI F 34 1997 Family asked for medicine. Too ill Acute pneumonia AliveMG M 62 1997 Family TB recurrence DiedMM M 65 1996 Fall from tree Injuries DiedNM F 61 1998 Head wound from drunk son-in-law Alive


successful immunizations (Bennett et al.,1973). We organized a measles immuniza-tion expedition from the Mbulu governmenthospital in 1992 and some may have beenreached by the 1997 polio vaccination drive.

It is difficult to assess the relative im-portance of these counteracting factors. TheHadza may have avoided the 1917 flu pan-demic, as it receives no mention in Bag-shawe (1924–1925), Bleek (1931a, b), orKohl-Larsen (1943, 1958), despite Bag-shawe reporting Hadza as giving refuge toIsanzu during the 1918–1920 famine. ButHadza did not totally escape the polio epi-demic of the mid-1960s. Bennett et al.(1973) imply that few Hadza were affectedby a smallpox epidemic in Shinyanga Re-gion, adjacent to the western Hadza. ManyHadza children succumbed to measles dur-ing a brief settlement attempt in 1986.Hadza claim that many died of diseaseduring the 1964–1965 settlements. No caseof AIDS is presently known among Hadza.The first case in the study area was re-ported in Mangola (a large village in thenorthern part of Hadza country) amongnon-Hadza in 1991 and the disease isabundant elsewhere in northern Tanzania.Meningococcal meningitis has had severalsmall epidemics in northern Tanzania inrecent years, including one near the north-ern end of Hadza country in which themission doctor rapidly interceded, but wehave heard of no Hadza deaths suggestiveof this disease.

Before modern medicine existed? Mod-ernmedicine andpublic health interventions(‘‘sanitary engineering’’ and immunizationprograms) are effective and avoiding foodshortfall aids survival. In trying to assesstheir effect upon the three modern hunter-gatherer populations, another comparisonmay help. Figure 1 shows that Hadza adultand elder mortality was very similar to thatof the English between 1640 and 1809. TheEnglish of that period had no access tomodern, science-based medicine. It had notbeen invented. There seems no reason tosuppose that English traditional folk medi-cine was any better than that of the Hadza,Ache, or !Kung. Thus modern hunter-gath-erers show a level of mortality close to that offarming populations with no access to mod-ern medicine. Similar data (Laslett, 1995)are available for France, Sweden, and otherEuropean countries. Absence of modern

medicine did not lower European postre-productive survivorship below that observedamong modern hunter-gatherers.

Sanitary engineering

Much of the 19–20th century improve-ment in life expectancy has been attributedto piped water, sewage disposal, and otherimprovements in hygiene. Does this con-tribute to survivorship of Hadza, !Kung orAche in the forest period? We know only oneHadza who has access to piped water (heworks in the house of a European farmer),and very few individuals among theirfarmer or herder neighbors have access topiped water. At three Hadza settlements inthe 1960s an effort was made to providepiped water. All of these have been defunctsince at least 1985. Relics can be seen andin one place (Endamagha) the source is usedby farmers and herders. At some settlementattempts, pit latrines were dug. We have nodata on their use or longevity. In Hadzacamps people defecate a little outside camp,mostly during the hours of darkness. Eachcamp is surrounded by a band of drying fe-ces some 20–40 yards from the houses.Camps are moved relatively frequently andlarge populations of flies do not accumulate.Water is obtained from a variety of naturalsources. As described by Woodburn (1968a),many of these are also used by wild animalsand Datoga livestock. Likewise, Ache in theforest period, and !Kung described by Ho-well (1979), had no access to piped water orsewage disposal. All three populationsachieve their survival figures without theassistance of modern sanitary engineering.

Law and order; murder/homicide

Hadza are not noted in the literature foran unusual incidence of homicide. Theirrelationship with their herder neighbors,the Datoga, has been described as uneasy.It is said that sometimes a Datoga youthwill attempt to enhance his status by killinga Hadza instead of a lion or a cattle thief. Itis said that Hadza have made the occasionalretaliatory killing. Murders of Hadza byHadza are known but appear rare. Wood-burn (1979) has described in some detail theease with which Hadza men, all lethallyarmed with poison arrows, could ambush,kill, and remain undetected. He suggeststhat this is an important ‘‘equalizer’’ andrestraint in Hadza social life.


Our records show four Hadza killed, threeby Datoga (two men, one woman) a mankilled by other Hadza in 10 years. Thesedeaths come from a sample of 1,000 peoplefollowed in our ‘‘where are they now’’ inter-views and known to have been alive at sometime between 1985 and 1995. This is fourdeaths in 10,000 person-years (40/100,000).Lee (1979) gives a rate for the !Kung of 29.3per 100,000 person-years. (Kim Hill, per-sonal communication, suggests this figure isan underestimate because it includes theyears after police began to intervene andbecause the people who lived in Namibiawere included, while it is not clear that allhomicides that occurred in Namibia wouldhave been recorded). All these killings wereof !Kung by !Kung. Thus, compared to!Kung, Hadza lose more people to homicidebut themselves kill fewer.

Hill and Hurtado (1996) describe homi-cide as ‘‘by far the most common cause ofdeath to forest-dwelling Ache’’ (p. 163),mostly at the hands of non-Ache raidingparties. Hill (personal communication) cal-culated homicides per person-year. Achehomicide rate is much higher than that ofHadza or !Kung. The difference appearspartly due to the high rate of killings bynon-Ache, and the practice of killing chil-dren under a variety of special circum-stances. Nonetheless, adult Ache kill otheradult Ache at a higher rate than Hadza or!Kung kill each other.

Although Lee (1979; see also Cashdan andDraper, 1992 ) reports police intervention aslong ago as 1952, !Kung recognize they are‘‘remote from any immediate authority fig-ure to settle disputes.’’ We have witnessedtwo large-scale disputes among Hadza, eachreaching borderlines of deadly force. Eachresulted in large delegations of people goingfrom camp to camp debating the issue andapparently influencing a peaceful settle-ment. We have also seen communal discus-sion with a young woman and her suitorsabout the danger of her failure to chose be-tween them.

Does knowledge that homicide is illegaland could bring trouble from the outsidehave no effect? Several observations sug-gest that knowledge of state law and thepolice may have had little effect among theHadza until very recently.

1) Hadza recounting their versions of anambush and murder (apparently the sameincident as reported by Woodburn, 1979)

show no indication that they believed ex-ternal authority should have been involved.

2) Informants reporting the three killingsof Hadza by Datoga made no mention ofhaving sought outside intervention or pun-ishment.

3) When asked whether any one had re-ported these cases to the police or other lo-cal authority, informants seemed surprisedat the idea they could have done this.

4) Hadza seemed unsurprised when,sometime in the late 1980s or early 1990s,one of their own responded to villagers�pleas (because the police were 2 days away,so could not be called in time) to bring hisarrows to deal with some young non-Hadzawho had turned violent at the travelingmarket.

5) In 1995 one of the best-educated andmost softly spoken Hadza was interviewedin the district capital by a Tanzanian econ-omist from a government commission.Aiming for a ‘‘household budget’’ of theHadza, the economist latched instantly ontothe rationality of hunting—why hunt toprovide a public good? ‘‘You get more wom-en.’’ And since most are married, the nextquestion: ‘‘What if you catch another manwith your wife?’’ ‘‘Arrows’’ was the unhesi-tating reply. Earlier authors have reportedthat Hadza regard a wronged husband asjustified in killing his rival. The conversa-tion implies that this idea is still accepted.

6) The police are 2 days walk away fromthe closest Hadza camps, and unlikely tohave a functioning vehicle or want to chasemiscreants through the bush.

Among the Hadza, understanding, beliefin, and access to the rule of law has grownonly slowly. But they have long had evi-dence of the power that outside authoritiescan wield. Bleek�s 1930 field notebooks(University of Capetown Library, Manu-scripts and Archives Department) show herinformants had been impressed by thepower of the Germans but it is not clear thatthey understood this as anything but ran-domly applied. But in 1934–1939, some re-quested Kohl-Larsen�s intervention in apotentially deadly dispute (Kohl-Larsen,1958). In 1964 many Hadza witnessed first-hand the power of the new postcolonialgovernment when some 100 or more wererounded up at gunpoint and taken to theYaeda settlement scheme. In 1985 Hadzareported the rapid and severe punishmentvisited upon Datoga after some bird trap-


pers had been killed in a remote location.But even by 1990, Hadza understanding ofthe legal system was not sufficient for themto recognize that a local official had no legalright to be sending groups of young meninto the bush to round people up and beatthem if they refused to join yet anothersettlement scheme. By 1995 Hadza wereready to have the ward secretary call policein response to a murder.

Our current impression is that externallyimposed law and order had little influenceupon Hadza homicide rates during most ofour study period. The similarities betweenHadza and !Kung methods of dealing withdisputes and violence suggest a long-lastingcondition of a low but recordable rate ofhomicide.

Note that among !Kung only three of thekillings were of a woman, and only oneamong the Hadza. Ache, on the other hand,reported killing older women. Yet Hill andHurtado (1996) cite very low mortalityamong women 45 to 65 years. We concludethat in the past mortality due to homicidewithin the ethnic group is unlikely to haveaccounted for much larger numbers ofdeaths to postreproductive women thanwere observed in the contemporary foragerpopulations.

Do the elderly or infirm take refuge in villages?

Two possibilities are envisaged:1) Elderly, sick, and infirm go to villages

and thereby survive longer. Then, if theyare included in our censuses, high eldersurvivals would be recorded.

2) Elderly sick, infirm go to villages, dienonetheless, but get left out of censusesbecause the villages were not visited. Thus,only the healthiest, toughest old peoplecould be included in the census, and thismight give an inflated estimate of eldersurvival (a possibility discussed for the!Kung by Howell, 1979).

Neither can have affected mortality esti-mates for the Ache in the forest period.Relations with villagers were too hostile forsuch an arrangement: 15 adult females age15–59 were shot by Paraguayans and onewas captured. Two females over 60 wereshot by Paraguayans.

Do older Hadza take refuge near farmers?We rated all camps in the 1977 census byLars Smith and in our 1985 and 1995 cen-suses for proximity to farmers. The ratingscale of 0–7 was intended to reflect the ef-

fort required to be in a position to beg forfarm foods: 0 being easiest, 7 hardest. Thescale primarily reflects distance in walkingtime from a camp to the nearest farmers orfields. All camps named in the 1977, 1985,1995 censuses were rated with respect tothe distribution of farmers in those years.Farmers have expanded their range, andthus many locations have gained lowerscores as time went on. Each individualreceived a score on these ratings, deter-mined by the camp he or she was found in,for each of the censuses (1977, 1985, 1995).We calculated changes from 1977 to 1985and from 1985 to 1995 and correlatedthese with age. There was no significantrelationship for 1977–1985 (Pearson�s r =)0.026, n = 403, P = 0.601), and apparentlyno great tendency for people in general to becamped nearer farmers in 1985 than 1977.Between 1985 and 1995 there was a differ-ent picture. More people were camped nearfarmers in 1995 than in 1985 (probably dueto the encroachment by farmers) and therewas a significant correlation between ageand increased proximity to farmers. But itwas young Hadza who were camped nearerto farmers, not older Hadza (Pearson�s r =0.259, n = 367, P < 0.0001). There is nosignificant tendency for older Hadza tomove near to farmers. Taking shelter nearfarmers thus seems unlikely to account forHadza elder survivorship.

Farmer or herder neighbors protect modernhunter-gatherers from food shortage?

Both !Kung and Hadza have been ob-served regularly trading with, working for,and begging from farmers. In a world ofhunters among hunters, opportunities forlong-distance exchange may have beenpresent but the availability of food to begmay have been much reduced. But evalu-ating the difference between the modernand ancient circumstance is not easy.Farmers have occupied the locations withthe highest rainfall or best groundwatersupply, thus perhaps areas of greatestbiomass productivity (see Foley, 1982).Farmers or herders have degraded the en-vironment in parts of !Kung and Hadzacountry. We may be observing foragermortality in degraded versions of the leastproductive habitats. The question can beposed as: ‘‘Do farmers give more food thanthey, ‘take� by removing and degrading wild


habitat?’’ The 11% of the Hadza diet ob-tained from farm foods (1995–1996) couldbe weighed against the measurable per-centage of land lost to farmers and the lesseasily measured loss of game and other re-sources. The balance is likely to tip in thedirection of greater hardship for Hadza inrecent years. In contrast, Balee (1992)claims that long-departed, undocumentedfarmers had enriched the Ache habitat.

Anthropological literature comments onthe reliability of foraging and chanciness offarming, citing striking instances of farmersseeking refuge among foragers to escapefamine due to crop failure. Lee (1979) givesan instance among the !Kung. Bagshawe(1924; p. 126) reports an instance in Hadzacountry ‘‘during the 1918–1920 faminehundreds of natives belonging to othertribes took refuge in their [Hadza] countryand lived on game which the Hadza assistedthem to kill.’’ ‘‘There are amongst the tribea few foreign girls, mostly Anisanzu, whomarried Kangeju [Hadza] men during thefamine.’’ Bagshawe (1924) and Obst�s (1912)Hadza informants, present-day Hadza, andDatoga informants of Borgerhoff-Mulder,Sellen and Sieff (personal communication)in the 1990s report Datoga herders takingto the forager lifestyle after losing theircattle to the Maasai in the 19th century.They soon returned to herding but theirclan are still remembered as the people whoonce lived among the Hadza as hunters.Opportunities for foragers to benefit fromfarmers are also quite evident. Might theseopportunities have had significant affectson mortality or lifespan? How great couldthese effects be?

Among Ache in the forest period, the ef-fect was surely minimal, or even negative.Stealing from farmers� fields was a verydangerous occupation, many Ache wereshot by Paraguayans, and others were cap-tured and enslaved. Trying to benefit fromfarm produce may thus have increasedmortality! Howell (1979) dates the arrival inthe Dobe area of a substantial number ofherders to the 1950s and reports the lowermortality of !Kung born after 1950. In the1960s and 1970s, many !Kung spent time atcattle posts, working for and begging offherd owners.

Hadza are so mobile and opportunisticthat it might be hard to directly assess aneffect of farm food. Marlowe�s recent ob-servation that even in 1995, in a sample of

camps that included many close to farmers,the proportion of food of agricultural originwas only 11%, suggests the effect may besmall (Marlowe, 1997). We tested for aneffect of proximity to farmers upon Hadzamortality, using the ratings of camps byproximity to farmers. The proximity ratingwas entered as an independent variable topredict survival or death across the 5years, 1985–1990, after controlling for age(age+age2). There was no significant effectof proximity to farmers (P = 0.9350). Thisresult is to be expected if Hadza are free tomove to where food is most available tothem. But with so few deaths this cannotbe a very powerful test.

Predation

Hill and Hurtado (1996) begin their ac-count of Ache demography with an inform-ant�s chilling story of the loss of a friend to ajaguar. Predation has loomed large in theevolutionary anthropologist’s imagination,from Bowlby�s (1969) account of the originsof the child�s attachment to its caregiver, tocontemporary discussions of the effective-ness of scavenging. Yet data on modernhunter-gatherers show few losses to preda-tors. Ache greatly fear jaguars, and nineadults were reported killed by them butthese contributed only .0235% of deaths.Snakebites contributed only .0548%. Hadzarespect lions, and !Kung fear them, anddeaths or serious injuries inflicted by lionshave been reported. Yet in neither popula-tion is there a single death in the demo-graphic data attributed to a large predator.Hadza and !Kung also regard buffalo asdangerous, and Hadza greatly fear ele-phants. Rhinoceros were abundant, andmuch feared by Hadza for their unpredict-ability during fieldwork in 1959–1961 byWoodburn (1964) but disappeared by thetime Lars Smith was in the field in the mid-1970s. We have heard of no deaths of Hadzaattributed to Rhinoceros. Bennett et al.(1973) report a number of injuries inflictedby wild animals, or sustained while runningaway from them. Such accidents continue.Falls from trees were a more common causeof injury and still are.

Is this low predation rate because modernforagers (but only men, not women) arewell-armed (with bows and poison arrows,none had guns during the study periods), orbecause predators have become scarce? Li-


ons cannot be described as scarce in eithersub-Saharan habitat during the main peri-ods of data collection. !Kung researchersoften saw lion tracks near their camps, of-ten heard lions and hyenas at night. Be-tween 1984 and 1992 in Hadza bush campsit was very unusual not to hear lions atnight, and extraordinary not to hear hye-nas. Numbers of these predators are likelyto vary with numbers of prey, which is inturn closely linked to rainfall (Coe et al.,1976). In 1995 lions were noticeably scarcein Hadza country and they remain less of-ten heard than before that time.

Perhaps lions took more foragers whenthere were no herders conducting a ven-detta on lions. Blurton Jones and Konner(1976) linked the reportedly more cautiousscavenging practices of !Kung as contrastedwith Hadza to a difference in the behaviorof lions in the two regions, which mightresult from the different practices of Hereropastoralists in Ngamiland (see also Howell,1979), and Maasai and Datoga in northernTanzania. Different abundance of preymight also account for the difference in lionbehavior.

GETTING LEFT BEHIND

Walking was the only means of transportavailable to Hadza, !Kung, or Ache duringthe study periods. In each population, thereare accounts of old and infirm people beingleft behind to die while camp mates moved.Such observations are well ingrained in thepaleontological imagination and may wellexemplify the final exit of many of the veryold. But did mobility cause many deathsamong the 45–65-year-olds? Would the re-coverably sick 50-year-old who falls behindbe left to die, or would they live to catch upanother day? Has the penalty of mobilityameliorated in the world of hunters amongfarmers? Have outside ethics been imposed?Informant accounts seem to imply ambiva-lence, people were not easily abandoned.This might imply that it would be easy topersuade people to cease this practice, oralternatively that the ‘‘new’’ ethic wouldreadily be acknowledged but the toughchoices imposed by the forager lifestylemight nonetheless prevail.

Among the Hadza there are some cases oflong survival and travel by the very infirm.A man who could barely walk (G.M. inTable 5) asked to be driven to a village. We

obliged but years later found him manymiles away, and he survived about another10 years unable to walk without the aid of apole. We also visited an old woman badlyafflicted with leprosy, barely able to see orwalk, and apparently confined to her houseat Mongo wa Mono, site of a recent settle-ment effort. A couple of years later we foundher 55 km away at her son�s house. She hadwalked there and stayed until her death ayear or so later. The case of desertion dis-cussed by Woodburn in Lee and DeVore(1968) begins with the parents carrying theparalyzed boy from camp to camp for someyears. Likewise, from at least 1985 to 1995a blind youth�s parents led him and fed himinto their own late 70s. There are also re-ports of blind adults surviving for long pe-riods among the !Kung. Hill and Hurtado(1996) report 11 deaths under theircategory ‘‘left behind’’ but also describe thecase of an abandoned man who recoveredand rejoined his band.

The significance of mobility for adultmortality must have always been quitevariable. When !Kung move from one dryseason waterhole to another they may coversome 20–30 miles in a day, across waterlessand nearly shadeless country. Someone whoembarked on such a journey and failed tocomplete it would have to spend a cold anddangerous night in the bush, with or with-out a helper. Among the Ache, with appar-ently daily trekking, anyone unable to keepup, or recruit help, must have been at con-siderable risk. Journeys accomplished byEskimos and other northern foragers mightbe even more challenging. However, exceptamong the Ache, little is known about theoption of not embarking on the journey. Didit always mean staying alone in a totallyexhausted habitat? Many camp movementsinvolve camp fission. The infirm elder mightbe left with a less than optimal faction butnonetheless not alone.

If the goal is to estimate an average riskof a prehistoric African forager dying bybeing left behind, then Foley�s (1982) argu-ment is relevant. He points out that in Af-rica, contemporary foragers inhabit areas oflowest biomass productivity. Savanna areasof high productivity are now occupied byherders or farmers. This might imply thatthe !Kung situation is extreme and theHadza situation is a more typical repre-sentation of prehistoric mobility. Hadzacountry is well-supplied with small water


sources. Although camp moves are fre-quent, the distances are often quite small.Woodburn (1968a) has described Hadzacountry as relatively rich in food resources.We think there may always have been onlyfew fatal desertions among the Hadza,mainly because of the short distancestraveled on most moves. Perhaps the‘‘paleontological imagination’’ misunder-stood the variability, extent, and nature ofsavanna hunter-gatherer mobility.

Changes in infant nutrition affectadult survivorship

Barker (1997) has shown that impairedfetal growth predicts higher adult mortali-ty, suggesting that the fetus adjusts to im-paired nutrition in ways that exert costslater in life. Impaired fetal growth leads togreater risk of several diseases of adult life,particularly coronary heart disease and di-abetes. These ‘‘diseases of civilization’’ havelittle relevance to the adult mortality ofcontemporary hunter-gatherers (Barnicotet al., 1972, Truswell and Hanson, 1976).But if poor fetal nutrition were also to im-pair the immune system, forager mortalitycould be strongly affected by changes inmaternal nutrition. There is no evidence forsignificant secular changes in maternalnutrition. Adult Hadza heights and weightsdo not differ between 1966–1967 (Hiernauxand Hartono 1980) and today. Both resem-ble the figures given by Obst (1912). It isalso not clear whether the effects describedby Barker are due to shortfall in mother�snutritional intake, or due to problems withpregnancy that might be related to stress,or genetic variation in response to stress, orparent–offspring conflict (Haig, 1993).

Accounting: the Hadza ‘‘worst case’’

The investigations reported above give abetter feel for possible variation in Hadzamortality. Here we attempt a crude ‘‘ac-counting’’ to show how Hadza survivorshipwould be affected by increases in the effectsexamined above. Researcher interventionprobably saved one boy, no men, two girls,and two women in 10 years. Let us adddeaths of three children and two adultwomen. Let us double the number of homi-cides, despite our having argued that po-licing had not held down the homicide rate.Thus, we add another four deaths of adults.We will count nothing for older people

moving nearer farmers, they do not do it.We found no significant effect of proximityto farmers and we will ignore this. But wewill assume that the opportunity to seek outmedical help from government clinics savesone death per 5 years in each adult 5-yearage class (12 classes, 12 deaths added foreach 5 years for each sex, 48 added deaths).We add these 57 ‘‘deaths’’ to the life table.The deaths of the three children were addedat the age they were when we intervened.The deaths of adults were added to any agefor which zero deaths had been recorded.The deaths represent a 46% increase in thenumber of deaths used to calculate Qx andIx. But the effect on the Ix curve is quitesmall, moving the Hadza down from nearthe Ache to near the !Kung, leaving a lifeexpectancy at 45 of 19.3 years. We concludethat these features, or possible features, ofthe modern world have only a slight effecton the picture of adult survivorship ob-tained from contemporary hunter-gathererpopulations and are not sufficient to sug-gest that postreproductive life is a productof the modern world.

How do the data from hunter-gatherersmatch the extrapolations from brain weight

and body weight?

Comparative studies have extrapolatedhuman life spans that are nearly as long asare observed in living populations. In themost recent comparative study, using 13anthropoid subfamilies Hammer and Foley(1996) extrapolate H. sapiens maximumlifespan as between 66 and 78 years. Life-span in the contemporary populations, de-fined as the age of the oldest individuals, isvery vulnerable to sampling error and thedifficulty of allotting ages to older individ-uals. The comparative method may smoothout some of this variation, being based onlarge numbers of species. A better way tocompare the two sets of observations is tocompare average adult lifespan, the expec-tation of life at onset of reproduction. Theextrapolated maximum lifespans, convertedby Charnov�s (1993) formula, correspond toaverage adult life spans for H. sapiens of26.2–30.9 years. If reproduction begins at20 years, Hammer and Foley�s, extrapola-tions can be compared with modern hunter-gatherer life tables and with Coale andDemeny (1983) life tables for other modernpopulations. Table 6 shows that Hammerand Foley�s (1996) results are not far out of


line with observed life tables of modernhunter-gatherers, and are far from the re-ports of archaeological populations. Theextrapolated average adult lifespan is closeto that for the !Kung and the Hadza ‘‘worstcase,’’ and greater than that estimated forthe Gainj horticulturalists in New Guineaby Wood (1987).

DISCUSSION

Ways in which mortality of adult con-temporary hunter-gatherers might under-estimate adult mortality of prehistorichunter-gatherers were considered. ‘‘Moderninfluence’’ can be shown in the comparisonof !Kung before and after 1950. Several po-tential modern influences were suggested,tested, and mostly discarded among theHadza, and are barely imaginable for theAche in the forest period. Such underesti-mation was not sufficient to explain awaythe 20-year life expectation of hunter-gath-erers who reach 45 years of age. Compara-tive studies such as Hammer and Foley(1996) predict a human lifespan of 66–78years from regressions of lifespan on brainand body weight. Converted to averageadult lifespan, these results overlap withthe figures from contemporary populations.These two sources support the conclusionthat the evolved average adult lifespan forHomo sapiens includes a lengthy postre-productive period. Postreproductive life isnot a product of civilization and moderncircumstances, nor of agricultural or in-dustrial economies.

Of course, confirming the existence of apostreproductive lifespan does not permit achoice between alternative explanations ofthis phenomenon. For example, we havegiven no attention to ‘‘by-product’’ accounts.

For instance, if hunter-gatherer adult mor-tality remained constant throughout adultlife (as apparently observed in some birds,see Stearns 1992), one could regard postre-productive individuals as the chance resid-uum of a constant rate of attrition of theadult population. Whether this situationtruly applies to birds is disputed, and itwould pose some interesting problems forlife history evolution. Ache, Hadza, and!Kung show very similar adult mortality(Table 4). Each population also shows theusual Gompertz-like increase in probabilityof death in the late 60s and 70s. (Thesepopulations are too small to show the small‘‘non-Gompertz’’ tail that currently inter-ests both demographers and biologists, e.g.,Partridge and Mangel, 2000.) The picture iscompatible with Hamilton�s (1966) pictureof evolution of senescence, but only if weassume that individuals age 45–65 havereproductive value.

Two other possible lines of argumentwhich seem likely to support our conclusionwere omitted:

1) Demonstrating the comparative vigorof older adults. Whatever the mortalitylevels, hunter-gatherers appear to maintainphysical ability at a similar level to, orbetter than observed in other active modernhumans (Finch, 1990; see also Barnicot,1972; Truswell and Hanson, 1976). For ex-ample, postreproductive Hadza womenwork longer hours and as effectively asyounger women (Hawkes et al, 1997) andHadza men aged 60 are about as good attarget archery as men in their 30s (unpub-lished data collected by Marlowe and Blur-ton Jones). Kaplan (1994) plotted foodproduction against age for Piro and Mach-iguenga horticulturalists and showed pro-ductivity continuing at a high level into the

TABLE 6. Maximum lifespan (Tmax), average adult lifespan (assuming first birth at 20, e20),for modern hunter-gatherers

Tmax

Averageadult lifespan(e20) from

0.4 Tmax–0.l

Observedaverage adultlifespan (e20)

Nearestmodel

!Kung 88 35.1 33.9 West 5 f.Ache (forest) 77 30.7 39.8Hadza 84 33.5 41.4 North 6 f.Hadza ‘‘worst case’’ 33.7H. sapiens upper estimate 77.6 30.9 West 3 f. North 1 f.H. sapiens lower estimate 65.9 26.2 <West 1Gainj 26.8 �West 1Libben 16.2

Hammer and Foley (1996) estimates for H. sapiens from body size and brain size, compared with Gainj (Wood, 1987), and Libben(Lovejoy et al., 1977).


60s. More systematic data on physical ca-pacities of older foragers are needed in or-der to make a quantitative comparison ofsenescence.

2) Examining data from other ThirdWorld and historical populations. Manypopulations may show adult survivorshipsimilar to contemporary hunter-gatherers(child mortality may vary much more),suggesting a life history that is almostuniversal, found everywhere except underindustrialization, ad lib food, and advancedinfrastructure and health services. Howev-er, published after the tabulations for thisarticle were completed, Early and Headland(1998) may provide a contrary examplefrom the Agta. Agta economy, with exten-sive trading of meat for rice andlong-standing if limited horticulture, fallsoutside our economic definition of hunter-gatherer. But the existence of such ahigh-mortality population anywhere, andespecially with a postreproductive surviv-orship somewhat worse than the !Kung andthe Hadza ‘‘worst case,’’ is important. Fromthe figures available, an Agta 40-year-oldcan expect only about 16–17 more years.Because of the low maximum age appar-ently attributed, this figure should be com-pared to the years of expected life to age 70in Table 4. Headland and Headland (1997)report the high mortality of the Agta asevidence of the progressive loss of theirlivelihood by a well-documented series ofdeprivations by outside forces. But withoutbetter understanding of the reported ratesof death among 30-, 40-, and 50-year-olds,which exceed the highest mortality modelsin Coale and Demeny (1983), and among the60-year-olds where mortality is lower thanin the models, we cannot claim to haveclosed all argument about the extent ofhuman postreproductive life.

Archaeological accounts do not agree withcomparative extrapolations, or with thehunter-gatherer data; few describe individ-uals living beyond their 40s. The agreementbetween the comparative extrapolationsand the modern data supports Howell�s(1976) proposition that ‘‘the uniformitarianassumption’’ (that the overall pattern ofage-specific fertility and mortality is similarfor any human population) be employed byarchaeological demography. To discard theuniformitarian assumption and support theview that few lived past 45 would requirethe following: a) to discard the comparative

findings, and therefore require us to showwhy human life span should be so muchshorter than predicted by patterns of cor-relation found among other taxa; b) to findsome significant modern influence; and c) toshow why the Ache, to whom none of themodern influences proposed and examinedin this article can have applied, had adultsurvivorship so close to that of the otherpopulations.

Hill and Hurtado (1996) discuss the un-reality of excluding warfare from demo-graphic models. It is excluded because it isoften very extreme and very sporadic, andthus cannot be encompassed by stablepopulation models. But many societies haveexisted in which warfare was frequent, ifsmall-scaled. Among Ache, risk of homicideseems to have been constant, and high. Wehave seen no accounts of mortality due towarfare among !Kung. Data on the Hadzacover no period of warfare. Colonial andpostcolonial governments may have re-duced traditional indigenous warfare. Eventoday Hadza talk of a time when Hadzawere killed in some numbers by Maasai.Early written accounts (Obst, 1912; Bag-shawe, 1924–1925; Jaeger, 191l) suggestthey are referring to a period sometime inthe mid-to-late 19th century when Maasaiexpanded into Ngorongoro crater and otherareas around Hadza country. Indeed,Blurton Jones et al. (1992) argued thatHadza population was still increasing froma low produced by these attacks. Had Ger-man colonization and Rinderpest notweakened the Maasai it is doubtful thatHadza would have survived into the 20thcentury.

The sparse historical information aboutepidemics and the Hadza was discussedearlier. Modern foragers have tended toisolate themselves from their densely pop-ulated neighbors. The isolation might haveincreased the intensity of epidemics butperhaps reduced their frequency. The densepopulation and connections to the largerworld might have increased the frequencyof epidemics among their neighbors. It ispossible that in a world of hunters amonghunters, diseases flowed more steadily andepidemics were less frequent and less se-vere, or that populations were simply toosmall to support certain diseases. As withthe history of warfare, feeling runs so in-tensely in historical research on Africa andSouth America that we may be excused for


reserving judgment until less passionatelyoverinterpreted data become available.

Discussing the very low adult mortality ofthe Ache, Hill and Hurtado (1996) suggest-ed that a ‘‘sawtooth’’ pattern of rapid in-crease and sudden crash may not have beenunusual in human population history. Theirargument goes beyond the Ache illustrationof such a pattern. Hill and Hurtado pointout that rapid increase seems built in tohuman life history parameters. While Mal-thus suggested that human fertility mustoutrun available land, and Darwin modifiedthis view to point out that populations mustgrow rapidly unless curbed by mortality(which then exerts selection), Hill andHurtado (1996, p. 471) imply that perhapshuman mortality could rarely have kept upwith fertility: ‘‘foragers not affected bySTDs do show completed family sizes ofaround six live births. Most also show sur-vivorship to adulthood around 50%. Thisleads to a very high population growth thatwould be impossible to sustain. ...then ju-venile mortality must have been around67% during most of human history. Suchhigh mortality has never been observed inany traditional population... An alternativeview on human life history is that rapidgrowth has been a common characteristic ofour species.’’ Hill and Hurtado further pointout that if human mortality levels wereother than observed, we would be a verydifferent animal. Higher adult or preadultmortality would lead to an earlier age atmaturity, thus an earlier age at cessation ofgrowth: we would be a smaller animal thanwe are.

Hill and Hurtado�s suggestion needs to beapproached cautiously. Fertility varies aswell as child mortality, and the !Kung caseshows that an approximation to zero popu-lation growth is possible in a hard envi-ronment. But if a sawtooth history doescharacterize the human species, its impli-cations should be considered. Life historytheory based on stationary populations maybe inadequate [although there are someindications that selection due to rapid in-crease and selection at density dependencehave similar effects on life history(Charlesworth, summarized in Stearns,1992)]. Life history effects of repeated pop-ulation crashes have not been exploredtheoretically, although population bottle-necks are expected to leave genetic traces(Harpending et al., 1993) and empirical

study long ago recorded changes in genefrequency following a population crash. Akey question is whether genotypes adaptedto survive the rare disaster can prevail overthose adapted to make the most of thelonger periods during which rapid multi-plication is possible.

The Hadza data support Hill and Hurt-ado�s suggestion. Perhaps zero populationgrowth cannot be regarded as the primitivecondition of mankind. Would our attitudesto, and interventions with, populationgrowth be different if we thought of rapidincrease as the natural state of affairs, in-stead of as an annoying aberration of peoplein developing countries? As evolutionaryanthropologists we should also continue towonder about how such a life history cameabout.

ACKNOWLEDGMENTS

We thank the Tanzania Commission onScience and Technology for permission toconduct research in Tanzania. We thankseveral hundred individual Hadza for theirpatience and good spirits and our field as-sistants Gudo Mahiya and the late SokoloMpanda for their expertise and collegiality.We thankLarsC. Smith for introducing us tothe Hadza and allowing us to use data fromhis 1977 census of the eastern Hadza, DavidBygott and Jeannette Hanby for providing ahome away from home and vital logistic fa-cilities, Professors C.L. Kamuzora and Au-dax Mabulla of the University of Dar esSalaam, and numerous citizens of theMbuluand Karatu districts for help and friendship.The University of Capetown Library gave usaccess to Dorothea Bleek�s Hadza field note-books and photographs.

LITERATURE CITED

Austad S. 1997. Postreproductive survival. In: WachterKW, Finch CE, eds. Between Zeus and the salmon:the biodemography of longevity. Washington, DC:National Academy Press. p 161–174.

Bagshawe FJ. 1924–1925. The peoples of the HappyValley (East Africa), Part 2. The Kangeju. J Afr Soc24:117–130.

Balee W. 1992. People of the fallow: a historical ecologyof foraging in lowland South America. In: RedfordKH, Padoch C, eds. Conservation of neotropical for-ests: working from traditional resource use. NewYork: Columbia University Press. p 35–57.

Barker DJ. 1997. The long-term outcome of retardedfetal growth. Clin Obstet Gynecol 40:853–863.

Barnicot NA, Bennett FJ, Woodburn JC, PilkingtonTRE, Antonis A. 1972. Blood pressure and serumcholesterol in the Hadza of Tanzania. Hum Biol44:87–116.


Baumann O. 1894. Durch Massailand zur Nilquelle.Berlin: Reimer.

Bennett FJ, Barnicot NA, Woodburn JC, Pereira MS,Henderson BE. 1973. Studies on viral, bacterial, rick-ettsial, and treponemal diseases of the Hadza of Tan-zania, and a note on injuries. Hum Biol 45:243–247.

Bleek DF. 1931a. The Hadzapi or Watindiga of Tan-ganyika territory. Africa 4:273–286.

Bleek DF. 1931b. Traces of former Bushman occupationin Tanganyika Territory. S Afr J Sci 28:423–429.

Blurton Jones NG, Konner MJ. 1976. !Kung knowledgeof animal behavior. In: Lee RB, DeVore I, eds. Kala-hari hunter-gatherers. Cambridge, MA. HarvardUniversity Press. p 325–348.

Blurton Jones NG, Smith LC, O�Connell JF, Hawkes K,Kamuzora CL. 1992. Demography of the Hadza, anincreasing and high density population of savannaforagers. Am J Phys Anthropol, 89:159–181.

Blurton Jones NG, Hawkes K, O’Connell JF. 1999.Some current ideas about the evolution of the humanlife history. In: Lee PC, ed. Comparative primate so-cioecology. Cambridge: Cambridge University Press.p 140–166.

Bowlby J. 1969. Attachment and loss, vol 1. Attach-ment. London: Hogarth Press.

Cashdan E, Draper P. 1988. Technological change andchild behavior among the !Kung. Ethnology 27:339–365.

Charnov EL. 1993. Life history invariants. Oxford:Oxford University Press.

Coale A, Demeny P. 1983. Regional model life tablesand stable populations. New York: Academic Press.

Coe MJ, Gumming DH, Phillipson J. 1976. Biomass andproduction of large African herbivores in relation torainfall and primary production. Oecologia (Berlin)22:341–354.

Cohen M. 1989. Health and the rise of civilization. NewHaven: Yale University Press.

Dyson T. 1977. The demography of the Hadza in his-torical perspective. African Historical Demography,Centre for African Studies, University of Edinburgh,Scotland.

Early JD, Headland TN. 1998. Population dynamics ofa Philippine rain forest people. Gainsville, FL: Uni-versity Press of Florida.

Finch CE. 1990. Longevity, senescence, and the ge-nome. Chicago: University of Chicago Press.

Foley RA. 1982. A reconsideration of the role of preda-tion on large mammals in tropical hunter-gathereradaptation. Man (N.S.) 17:383–402.

Haig D. 1993. Genetic conflicts in human pregnancy.Q Rev Biol 68:495–532.

Hamilton WD. 1966. The moulding of senescence bynatural selection. J Theor Biol 12:12–45.

Hammer MLA, Foley RA. 1996. Longevity and lifehistory in hominid evolution. Hum Evol 11:61–66.

Harpending HC, Sherry ST, Rogers AR, Stoneking M.1993. Genetic structure of ancient human popula-tions. Curr Anthropol 34:483–496.

Hawkes K, O�Connell JF, and Blurton Jones NG. 1997.Hadza women�s time allocation, offspring provision-ing, and the evolution of long post-menopausal life-spans. Curr Anthropol 38:551–577.

Hawkes K, O�Connell JF, Blurton Jones NG, Alvarez H,Charnov EL. 1998. Grandmothering, menopause, andthe evolution of human life histories. Proc Natl AcadSci USA 95:1336–1339.

Headland TN, Headland JD. 1997. Limitation of humanrights, land exclusion, and tribal extinction: the AgtaNegritos of the Philippines. Hum Org 56:79–90.

Hiernaux J, Hartono DB. 1980. Physical measurementsof the adult Hadza of Tanzania. Ann Hum Biol 7:339–346.

Hill K, Hurtado AM. 1996. Ache life history: the ecologyand demography of a foraging people. New York:Aldine de Gruyter.

Howell N. 1976. Toward a uniformitarian theory ofhuman paleo-demography. J Hum Evol 5:25–40.

Howell N. 1979. Demography of the Dobe area !Kung.New York: Academic Press.

Howell N. 1982. Village composition implied by apaleodemographic life table: the Libben site. Am JPhys Anthropol 59:263–269.

Jaeger F. 1911. Das Hochland der Riesenkrater und dieumliegenden Hochlander Deutshc-Ostafrikas. Berlin:Mittler.

Kaplan H. 1994. Evolutionary, and wealth flows theo-ries of fertility. Empirical tests and new models. PopDev Rev 20:753–791.

Kaplan H. 1997. The evolution of the human life course.In: Wachter KW, Finch CE, eds. Between Zeus andthe salmon: the biodemography of longevity. Wash-ington, DC: National Academy of Sciences Press. p175–211.

Kaplan H, Hill K, Lancaster J, Hurtado AM. 2000. Atheory of human life history evolution: diet, intelli-gence, and longevity. Evol Anthropol 9:156–185.

Key CA, Aiello L, Mollesen T. 1994. Cranial sutureclosure and its implications for age estimation. Intl JOsteoarchaeol 4:193–207.

Kohl-Larsen L. 1943. Auf den Spuren des Vormens-chen. Stuttgart: Strecker und Schroder.

Kohl-Larsen L. 1958. Wildbeuter in Ost-Afrika: DieTindiga, ein Jager- und Sammlervolk. Berlin: Die-trich Reimer.

Laslett P. 1995. Necessary knowledge: age and aging inthe societies of the past. In: Laslett P, ed. Aging in thepast: demography, society, and old age. Berkeley, CA:University of California Press. p 3–77.

Lee RB. 1979. The !Kung San. Cambridge: CambridgeUniversity Press.

Lovejoy CO,Meindl RS, Pryzbeck TR, Barton TS, HeipleKG, Kotting D. 1977. Paleodemography of the Libbensite, Ottawa County, Ohio. Science 198:291–293.

Marlowe FW. 1997. Paternal care and mating effortamong hadza hunter-gatherers. Doctoral disserta-tion, University of California, Los Angeles.

McDowell W. 1981. Report to the Rift Valley Project.Dar es Salaam: Ministry of Information and Culture.

McHenry H. 1994. Behavioral ecological implications ofearly hominid body size. J Hum Evol 27:77–88.

Mehlman M. 1988. Later quaternary archaeologicalsequences in northern Tanzania. Doctoral disserta-tion, University of Illinois, Champaign-Urbana.

Milner GR, Humpf DA, Harpending HC. 1989. Patternmatching of age-at-death distributions in palaeo-demogrpahic analysis. Am JPhys Anthropol 80:49–58.

Molleson T, Cox M, Waldron AH, Whittaker DK. 1993.The anthropology. The Spitalfields Project, vol. 2.London: Council for British Archaeology.

Obst E. 1912. Von Mkalama ins Land der Wakindiga.Mitteilungen der Geographischen Geselleschaft inHamburg 26:3–45.

Packer C, Tatar M, Collins A. 1998. Reproductive ces-sation in female mammals. Nature 392:807–811.

Partridge L, Mangel M. 2000. Messages from mortality:the evolution of death rates in the old. Trends EcolEvol 14:38–440.

Sacher G. 1959. Relation of lifespan to brain size andbody weight in mammals. Ciba Foundation Colloquia5:115–141.


Sacher GA. 1975. Maturation and longevity in relationto cranial capacity in hominid evolution. In: Tuttle R,ed. Primate functional morphology and evolution. TheHague: Mouton. p 417–441.

Sands B. 1995. Evaluating claims of distant linguisticrelationships: the case of Khoisan. Doctoral disserta-tion, University of California, Los Angeles.

Sands B, Maddieson I, Ladefoged P. 1993. The phoneticstructures of Hadza. UCLA Working Papers in Pho-netics 84:67–88.

Stearns SC. 1992. The evolution of life histories. Ox-ford: Oxford University Press.

Stevens A, Morissette J, Woodburn J, Bennett FJ. 1977.The inbreeding coefficients of the Hadza. Ann HumBiol 4:219–223.

Truswell AS, Hansen JDL. 1976. Medical researchamong the !Kung. In: Lee RB, DeVore I, eds. Kalaharihunter-gatherers. Cambridge, MA: Harvard Univer-sity Press. p 166–194

Walker PL. 1995. Problems of preservation and sexismin sexing: some lessons from historical collections forpalaeodemographers. In: Saunders SR, Herring A,eds. Grave reflections, protraying the past throughcemetery studies. Toronto: Canadian Scholars� Press.p 31–47.

Walker, PL, Johnson JR, Lambert PM. 1988. Age andsex biases in the preservation of human skeletal re-mains. Am J Phys Anthropol 76:183–188.

Washburn S. 1981. Longevity in primates. In:McGaugh JMJ, ed. Aging, biology and behavior. NewYork: Academic Press. p 11–29.

Weiss KM. 1973. Demographic models for anthro-pology. (Memoirs of the Society for AmericanArcheology, no 27). American Antiquity 38: No 2. PartII. p 1–186.

Weiss KM. 1981. Evolutionary perspectives on humanaging. In: Amos PT, Harrell S, eds. Other ways ofgrowing old. Stanford, CA: Stanford University Press.p 25–58.

Williams GC. 1957. Pleiotropy, natural selection, andthe evolution of senescence. Evolution 11:398–411.

Wood JW. 1987. The genetic demography of the Gainj ofPapua New Guinea. 2. Determinants of effectivepopulation size. Am Natural 129:165–187.

Wood JW, Milner GW, Harpending HC, Weiss KM.1992. The osteological paradox: problems of inferringprehistoric health from skeletal samples. Curr An-thropol 33:343–370.

Woodburn JC. 1964. The social organisation of theHadza of north Tanganyika Doctoral dissertation,University of Cambridge.

Woodburn JC. 1968a. An introduction to Hadza ecology.In: Lee RB, DeVore I, eds. Man the hunter. Chicago:Aldine. p 49–55.

Woodburn JC. 1968b. Stability and flexibility in Hadzaresidential groupings. In: Lee RB, DeVore I, eds. Manthe hunter. Chicago: Aldine. p 103–110.

Woodburn JC. 1970. Hunters and gatherers. The ma-terial culture of the nomadic Hadza. London: BritishMuseum.

Woodburn JC. 1972. Ecology, nomadic movement andthe composition of the local group among hunters andgatherers: an East African example and its implica-tions. In: Ucko PJ, Tringham R, Dimbleby GW, eds.Man, settlement and urbanism. London: Duckworth.p 193–206.

Woodburn JC. 1979. Minima politics: the political or-ganization of the Hadza of north Tanzania. In: ShackWA, Cohen PS, eds. Politics and leadership: a com-parative perspective. Oxford: Clarendon Press. p 244–246.

Woodburn JC. 1982. Egalitarian societies. Man (NS)17:431–451.

Woodburn JC. 1988. African hunter-gatherer socialorganisation: is it best understood as a product ofencapsulation? In: Ingold T, Riches D, Woodburn J,eds. Hunters and gatherers, vol. 1. Oxford: Berg. p31–64.


Documents

Antiquity of Postreproductive Life: Are There Modern ...hawkes/Blurton Jones... · Antiquity of Postreproductive Life: Are There Modern Impacts on Hunter-Gatherer Postreproductive