i Evolution of Human Growth

7/29/2019 i Evolution of Human Growth

1/14

ARTICLES

Evolution of Human GrowthSTEVEN R. LEIGH

Despite the importance of these ad-

aptations, our understanding of pre-cisely how and why human ontogenydiffers from ontogeny in other primatespecies remains qualitative and rather

speculative.16 Consequently, this studybuilds on previous analyses in fieldssuch as allometry, heterochrony, hu-man and primate growth studies, and

life-history theory710 to address severalissues regarding human growth. First itdevelops an interspecific allometric per-spective to determine how human

growth compares to trends establishedby other anthropoid primates. Becausehumans are comparatively large pri-mates, we may simply follow a growth

pathway that is consistent with expec-tations based on body size. Second, thisallometric foundation enables evalua-tion of two competing models thatseek to account for the pattern of

human growth prolongation. Pro-longation could represent extension

or retardation of all phases of growth.

Alternatively, the overall prolongationof human growth could represent ex-treme lengthening of a single growthphase. Obviously, a blend of these

possibilities, such as prolongation of afew growth periods coupled with ab-breviation of others, may exist. Third,this study reviews alternative life-his-

tory models that seek causal explana-tions for prolonged human growth. Allo-metric results are applied to qualitativeevaluations of these models with the ob-

jectives of clarifying the focus of thesemodels and directing theoretical atten-tion to new areas. For example, life-his-torymodels often emphasize certain vari-ables, such as age at maturation, while

looking past others, such as neonatalgrowth rates. Exploring human growthin light of alternative life-history modelshas the potential to refine and redefine

these models with the ultimate goal ofproviding an integrated explanation for

the evolution of human ontogeny.

THEORETICAL PERSPECTIVES

ON HUMAN ONTOGENY

Comparative and Historical

Questions

Climbing Schultzs ladder:

traditional perspectives on

human growth

The fact that human ontogeny

stands out among that of other mam-

mals has been recognized for centu-

ries, if not millenia.1113 However,

comparative analyses of human on-

togeny have only recently attracted at-

tention in biological anthropology.

Previous research generally conveyed

the impression that uniform exten-

sions in the duration of various

growth phases account for the long

period of human ontogeny. This viewstems mainly from Adolf Schultzs14

classic diagram of primate ontogenies

(Fig. 1). This model, regularly figured

in introductory biological anthropol-

ogy textbooks, is based on several on-

togenetic variables, including gesta-

tion length, dental eruption timing,

reproductive maturation, and total

life span. It predicts that the evolution

of primate ontogeny involves pro-

gressive and regular increments in

the duration of various growth stages,

reminiscent of a scala natura conceptof primate evolution. As noted by sub-sequent researchers,15,16 Schultzs

view implies that the evolution of

primate ontogeny may have in-

volved a regular, highly predict-

able, and perhaps orthogenetic pro-

cess that culminated in the evolution

of the human pattern. In effect,

Schultzs view reflects a predictable, if

rather uninspiring, mechanism for the

production of ontogenetic diversity

among primates.

Schultzs scheme, despite its sim-

plicity, promoted significant advancesin our understanding of human evolu-

tion. For example, Lovejoy17 inter-

preted Schultzs framework as a re-

flection of progressive prolongation

of life phases and gestation in pri-

mates (p. 342). According to Lovejoy,

Schultzs model provides powerful ev-

idence that a scala natura is an ade-quate metaphor describing develop-

mental variation within the primate

order. Lovejoys perspective implies

that prolongation of human growth

Steven R. Leigh, is an Associate Profes-sor of Anthropology at the University ofIllinois, Urbana, IL. His research interestsinclude evolutionary developmental biol-ogy, sexual dimorphism, and human evo-lution. E-mail: [email protected]

Key words: ontogeny; phylogeny; development;

growth spurts; childhood; brain development

Evolutionary Anthropology 10:223236 (2001)

The human pattern of growth and development (ontogeny) appears to differ

markedly from patterns of ontogeny in other primate species. Humans present

complex and sinuous growth curves for both body mass and stature. Many human

proportions change dramatically during ontogeny, as we reach sizes that are

among the largest of living primates. Perhaps most obviously, humans grow for a

long time, with the interval between birth and maturation exceeding that of all other

primate species. These ontogenetic traits are as distinctive as other key derived

human traits, such as a large brain and language. Ontogenetic adaptations are also

linked to human social organization, particularly by necessitating high levels of

parental investment during the first several years of life.

Evolutionary Anthropology 223


2/14

was central to the divergence of homi-nins, requiring models of homininevolution to include ontogenetic and

life-history components. Lovejoys in-terpretations have been controversial.Egregious errors can follow from ap-plying a scala natura to ontogeneticvariation within primates (Box 1).

Nevertheless, he placed life history

and ontogeny within the mainstreamof biological anthropology while pro-viding an impetus for critically evalu-

ating Schultzs model.

Allometric approaches

Research during the 1970s and

1980s, particularly interspecific allo-

metric studies, brought a quantitative

comparative perspective to studies ofprimate ontogeny.18,19 Unfortunately,few of these studies focused exclu-sively on human life histories. How-ever, these studies illustrated signifi-

cant variation in primate life-history

parameters that could not be accom-modated solely by body size; that is,they documented significant residual

variation.1820 This finding impliedless regularity in the evolution of pri-mate ontogeny than could be encom-passed by Schultzs model.

Building on this research, Gould21

pursued themes that emerged fromboth Schultzs observations and allo-

metric analyses to present the first

major attempt to evaluate ontogenies

as adaptations.21 Goulds theoreticalframework moved well beyond Schultzsmodel by emphasizing variables that

are important in the analysis of het-erochrony. These variables includesize and shape, as well as the timing ofevents in ontogeny. Gould promoted

his perspective by exploring the adap-tive bases of a matrix of retardationin human ontogeny. According toGould21:

A general, temporal retardation

of development has clearly char-

acterized human evolution. This

Figure 1. Adolf H. Schultzs rendition of primate life histories in comparative perspective. 14 This

figure has been used repeatedly to demonstrate the prolonged period of human ontogeny.

Gould pursued themesthat emerged from both

Schultzs observationsand allometric analysesto present the first majorattempt to evaluateontogenies asadaptations. Gouldstheoretical framework

moved well beyondSchultzs model byemphasizing variablesthat are important in theanalysis ofheterochrony.

224 Evolutionary Anthropology ARTICLES


3/14

retardation has established a

matrix within which all trends inthe evolution of human mor-phology must be assessed(p. 365).

Gould also attempted to explain onto-

genetic variation through the conceptof an r and K selection continuum.Despite the fundamental deficiencies

of r-K selection theory,10 Gouldsframework stimulated adaptive expla-nations of variation in age at maturityin primates20 and showed that life-his-tory theory could play a role in ex-

plaining differences in morphologicalontogeny.

Goulds views were consistent withwhat was known about comparative

growth and development during the1970s. However, we still lack a clearunderstanding of whether or not thismatrix of retardation adequately ac-counts for the human pattern. Conse-

quently, this analysis evaluates Gouldsidea along with alternative hypothesesthat seek to define and explain derivedchanges in human growth. Addressing

these issues involves consideration ofseveral contemporary theoretical per-spectives that explore complemen-tary, but often nonoverlapping as-

pects of this problem. Specifically,researchers have either focused ex-clusively on distinguishing how thepattern of human morphological on-

togeny compares to that of other pri-

mates or have applied general life-his-

tory models to humans, concentrating

mainly on maturation age. Thus, asecondary goal of the study is toachieve better integration betweenthese fields.

Allometry and Life History of

Ontogeny

Redefining allometrichypotheses

Allometric analyses have great po-tential to increase our understanding

of how human ontogeny compares tothat of other primates. It is importantto emphasize that Schultzs scheme isinsufficient for this purpose for at

least three reasons. First, as noted, heused data from several developmentalsystems. We now know that each ofthese traits may evolve in response todiffering selective pressures.22 Conse-

quently, comparisons of ontogeny re-quire investigations that isolate spe-cific variables or use multivariatemeasures that capture variation in nu-

merous traits. Second, the variationSchultz observed could easily be ex-plained by simple size differencesamong species. In effect, the X-axis in

Schultzs diagram (Fig. 1) represents asize axis, with each species spendingthe same proportion or percentage ofits total growth period at each devel-

opmental phase. Third, allometric

analyses lead to quantitative estimates

of the degree to which human growth

differs from expectations. Such esti-

mates may be extremely important for

refining life-history models.

An allometric perspective reveals

two major possibilities to explain how

prolongation of human growth may

have occurred. The first is a general

extension model. In this model, pro-

longation of growth could be a resultof evolutionary extension in all growth

periods (for example, prenatal, neona-

tal, and infant). This model emerges

from both Schultzs model and

Goulds views on human temporal re-

tardation. For each growth period,

human values should either meet ex-

pectations based on size (reflecting

maintenance of proportions), or devi-

ate from expectations in a consistent

direction. The second model, differen-

tial extension, involves extension of

only one or a few specific periods. For

example, a long period of infancy canresult in delays in maturation, but this

may not be accompanied by increases

in the length of other developmental

periods. Obviously, the lengths of var-

ious phases could be altered in any

direction and still lead to an overall

delay in age at maturation. With dif-

ferential extension, certain periods of

human growth should show devia-

tions from expectations, and the pat-

tern of deviations should illustrate at

least one period of ontogeny that is

longer than expected. Deciding which

Box 1. Primate Life History and Race

Schultzs diagram is commonly

used as a representation of primate

life histories. Lovejoys deployment of

the figure in his 1981 article undoubt-

edly popularized the model.17 Unfor-tunately, Lovejoy also explicitly brought

the notion of a scala natura into this

framework, transcending and formal-

izing Schultzs general comments

about the progressive nature of pri-

mate life histories. Lovejoy obviously

mentioned the scala natura in a met-

aphorical sense to simplify variation

within the primate order. However,

this metaphor has been taken literally

in attempts to associate racial vari-

ation in IQ with developmental sched-

uling. Specifically, J.-P. Rushton73

used Schultzs diagram coupled with

explicit reference to Lovejoys scala

natura in constructing a racial rank-

ing system. Thus, Rushtons relative

ranking of races on diverse variablesis determined by differences among

races in developmental scheduling.

This simply amounts to extending

Schultzs figure along the X axis to

accommodate racial categories of

humans.73 More explicitly, Rushton

grounded his idea of an inverse re-

lation . . . between brain size and

speed of physical maturation (p. 214)

on Schultzs model.

Ranked differences stem from vari-

ation in maturation rates, which are

said to affect brain morphometrics

and IQ. Rushtons ideas are problem-

atic at a number of different lev-

els.74,75 However, Schultzs model

can no longer be used as an ade-quate description of primate life-his-

tory variation. Moreover, by relying on

a scala natura, Rushton forfeits any

insight into human variation that

could be gained from evolutionary

theory.76 Simply put, and as is clearly

reflected by Fleagles updated ver-

sion of this figure,77,78 a better under-

standing of primate ontogenetic vari-

ation fails to illustrate the progressive

climb so crucial to Rushtons narra-

tive.

ARTICLES Evolutionary Anthropology 225


4/14

of these two models applies to hu-

mans is critical to evaluating how welllife-history models explain humangrowth.

A complicating factor in choosingbetween these possibilities concerns

alternative mechanisms for differen-

tial extension. For example, an exten-sion of the total preadult period couldbe produced either by increases in

specific growth phases or through in-sertion of novel growth periods notpresent in ancestral forms. These pos-sibilities are very difficult to distin-

guish from one another. In each case,growth phases are defined based oninflections in growth rate curves.2325

Thus, Bogin23 has suggested that hu-

man statural growth is prolongedthrough insertion of novel growthperiods representing childhood andadolescence. His idea relies on evi-

dence that human statural growthspurts are unique to primates. In ad-dition, Bogin employs both behavioraland hormonal definitions to definethese different traits. Unfortunately,

testing these hypotheses for other pri-mates requires separate investiga-tions. On the other hand, mass growthspurts are very common among pri-

mate species, including humans15,26

and African apes,15,25 which enablescomparative analyses. While it may

not be possible to distinguish betweenthese hypotheses, this distinction maybe unnecessary for mass growth,given the qualitative similarities be-tween mass growth-rate curves for hu-

mans and other primates.25

Life-history approaches

Explanations of why human growthdiffers from the growth of other pri-mates must be sought in terms of life-history models. Causal explanationsrest on a strong foundation of demo-

graphic and life-history studies of hu-man populations that has developedin the last two decades.27,28 Integra-tion of allometric and demographic

approaches has considerable poten-tial to refine life-history models, butnot all models recognize precisely thesame classes of allometric variables

(for example, growth rates andgrowth phase durations). Thus, mostlife-history models concentrate on ex-plaining maturation age variation10

without attempting to subdivide the

time span prior to maturation. This

emphasis has repeatedly proven its

value, but understanding how human

ontogeny compares to that of other

primates should point life-history the-

ories to new research areas. For exam-

ple, either a protracted infancy or ad-

olescence may retard human age at

maturation, but these scenarios carryvery different implications for paren-

tal effort,29 functional differences

among age classes,30 mortality, and

social organization. Therefore, this

analysis briefly reviews alternative

life-history models and applies allo-

metric results to these models. Allo-

metric findings provide the opportu-

nity for qualitative assessment of

how well life-history models account

for ontogenetic differences betweenhumans and other primates. The

life-history models considered in-

clude a brain growth model, an adult

mortality risk model, a juvenile met-

abolic risk aversion model, and an

investment model.9,28 Although the

present analysis approaches these

questions qualitatively, it can be

noted that quantitative tests of life-

history models with primate inter-

specific data show both advantages

and disadvantages to several of these

models.31

Brains and Learning. Attempts tointerpret the long period of humangrowth by reference to either learningor brain ontogeny have a deep history.As early as the mid-1700s, Henry St.John (Lord Bolingbroke)32 proposed

that humans require long periods of

development because they have moreto learn and more to do than otheranimals (p. 383). This idea was refor-

mulated by numerous authors, thencast into Darwinian framework byJohn Fiske33 and Herbert Spencer.34

Subsequent authors amplified this no-

tion by relating growth prolongationto adaptive flexibility.21 As with manyhistorical ideas, claims about learn-ing, plasticity, and growth prolonga-

tion often rely on untested assump-tions. Recent formulations of thishypothesis suggest that human growthperiods are lengthened because we

must assimilate considerable informa-tion prior to adulthood.1,9,21,23,28,3140

Learning, behavioral-flexibility, andbrain-growth models of life historiesimply that the benefits of what is

learned during the preadult periodoffset the reproductive opportunitieslost by prolonged growth. These mod-els predict that the earliest periods of

growth have undergone the greatestprolongation because of ties betweensomatic and brain ontogeny. It can

also be suggested that brain growthand development necessitates slowrates of growth in body mass in orderto accommodate the metabolic costsof total brain size growth.

Adult Mortality. A compelling al-ternative general explanation of pri-mate growth prolongation is an adultmortality model developed by Char-nov.7,41 This model predicts that ex-trinsic mortality determines the tim-ing of maturation. It relies on theobservation that the amount of energyavailable for growth and reproduction

(the production function) for pri-mates is lower than that of othermammals. The relatively low primateproduction function affects both

growth and reproduction, with opti-mal adult body size set by a balancebetween mortality and production.Selection for larger body size, given

the low primate production rate, in-creases the time necessary for growth,but the associated delay in maturationincreases mortality risk before matu-

ration. Large body size increases pro-

duction, but only if the increase in

. . . an extension of the

total preadult periodcould be producedeither by increases inspecific growth phasesor through insertion ofnovel growth periods notpresent in ancestralforms. These possibilitiesare very difficult to

distinguish from oneanother. In each case,growth phases aredefined based on

inflections in growth ratecurves.



5/14

size offsets increased mortality. De-

creased adult mortality is critical be-cause it extends the expected adult re-productive life span enough to offsetcosts of delayed maturation while pro-viding production benefits as a result

of larger adult size. Maturation delays

may not be related to selection directlyon juveniles. Long juvenile periods mayhave no special significance, and thus

may arise automatically as a result ofselection for larger size42 (p. 37).

A mortality risk model has difficultyaccommodating differences in growth

rates. Specifically, Charnov and Berri-gans7 version of this model assumesthat size differences among speciesare driven by differences in growth

duration rather than growth rates,with growth rates determined by theequation:

dw/dt A W.75 ,

where growth rate (dw/dt) is equal tothe production value (A) times massat maturation age (W) scaled to the0.75 power (or 0.7 power) (see Char-nov and Berrigan7 and Charnov,41 re-spectively). Charnov and Berrigan7 re-port an A value for primates at about0.4, which is substantially below thevalue of 1.0 for other mammals. How-ever, differences among species in

growth rates may pose problems forthe model simply because if ratesvary, then either A must vary or massmust be scaled by a different expo-nent. Increases in growth rates can

reduce the amount of time to reachadult size, thereby reducing time-related mortality risks during thegrowth period. On the other hand, thismodel cannot explain why growth

rates may decrease, because this the-oretically extends the period of pre-adult mortality in the absence of com-pensatory increases in growth rates

during other phases of ontogeny. In-sufficient knowledge about rate varia-tion does not, however, diminish thevalue of exploring correlations between

adult mortality and growth duration.

Metabolic Risk Aversion. Thethird model makes a very different setof predictions from the adult mortal-

ity model, but can be seen as an alter-native mortality risk model. Specifi-cally, Janson and van Schaiks43

metabolic risk aversion model sug-

gests that slow growth rates necessi-

tate a lengthy period of ontogeny. The

premise for this position is that low

growth rates (small changes in massper unit of time) can be selectivelyfavored. Janson and van Schaik sug-gest that primates generally encoun-ter conditions that increase metabolic

risk as a result of feeding competition

within groups, thus selectively favor-ing reduced growth rates. Feedingcompetition, which stems from pred-

ator pressure for group formation,may affect growing individuals moredrastically than adults. Young ani-mals, because of their small size,

should face relatively high risks ofmortality through predation. Feedingat or near the center of a group mini-mizes predation risks, but elevates

feeding competition and, therefore,the risk of mortality through starva-

tion. This, coupled with the fact thatjuveniles have less foraging efficiency

than adults do, means that juvenilesmay encounter selection against highgrowth rates, reducing metaboliccosts per unit of time. Maturational

delays essentially arise as byproductsof selection against high growth rates.On the other hand, increases ingrowth rates could be explained in

terms of reduced metabolic risks.Although Janson and van Schaik

emphasize differences between pri-mates and nonprimates, they also

apply their model to explanations ofdifferences among primates (for ex-ample, anthropoids versus nonan-thropoids and folivores versus nonfo-

livores). The relationship between

reduced growth rates and increased

growth duration may not be obviousin this model. However, this model

can accommodate prolongation ofgrowth if selection for optimal adultsize occurs late in ontogeny to balanceearly selection for reduced growth

rates. Thus, they suggest that selec-

tion for slow growth rates can forcedelayed maturation as a correlatedresponse43 (p. 72).

Future Investment. Kaplan andcolleagues9 have recently proposed anintegrative model to explain derivedattributes of human life histories.

They relate human growth prolonga-tion to slow rates of growth in bodymass and to the period of braingrowth. Their model couples reduc-

tions in mortality with increases inparental production to offset the costsof subsidizing offspring development.These authors recognize that preadulthumans are unable to support their

own development but argue that thisperiod of extremely low productivityis effectively an investment in futureproduction. Specifically, large bodies

that grow slowly, the acquisition ofdetailed knowledge, and the ability tolearn all pay off by higher productionin the adult period. This high payoff

enables very high production sur-pluses that can then be channeled todeveloping offspring. Kaplan and col-

leagues note that this life-history pat-tern requires the ability to offset threeimportant investment costs during theadult period, including low produc-tivity early in life, delayed reproduc-

tion, and a very expensive brain togrow and maintain (p. 161). Delays inmaturation enable the investmentnecessary to counterbalance thesecosts.

Ontogenetic data and attention togrowth phases are potentially very im-portant to this model. Distinguishingbetween general and differential ex-

tension hypotheses can lead to hy-potheses about which costs are likelyto be most important at variousphases in development. For example,

prolongation of all growth phases sug-gests an equal spread of investmentcosts throughout ontogeny. Earliestphase extensions imply that the costs

of brain growth and basic knowledgeacquisition are highest, while exten-sion of later phases suggests that themost costly knowledge pertains toadult behaviors, including hunting,

foraging, and social relations.

Janson and van Schaiksmetabolic risk aversionmodel suggests thatslow growth ratesnecessitate a lengthyperiod of ontogeny. The

premise for this positionis that low growth rates(small changes in massper unit of time) can beselectively favored.



6/14

Testing Human Life-History Mod-els. Evaluations of life-history modelsrequire data regarding ontogeny, de-mography, feeding competition, pre-dation, and socioecology. However,the performance of these models with

respect to the human case can be

qualitatively evaluated by assessinghow well they explain the observedpattern of human growth prolonga-

tion. First, a brain growth model pre-dicts protraction of the earliest phasesof human ontogeny. Exceptionallylong later growth phases, such as the

period encompassing the subadultgrowth spurt, may not be consistentwith this model because brain sizegrowth has been completed by this life

phase. Second, an adult mortalitymodel is consistent with prolongationof all growth periods at size-expectedrates of growth. Here, mortality re-ductions produced by growing at a

size-expected growth rate to a largesize, with attendant increases in pro-duction, can explain the human pat-tern. Differences in growth rates dur-

ing ontogeny are problematic for theadult mortality model.22 Third, a met-abolic risk-aversion model can ac-commodate differences in growth

rates. However, heavy subsidies forhuman juveniles9 complicate assess-ments of this model. Fourth, the in-

vestment model of Kaplan and col-leagues fits best with prolongation ofearly growth phases, given its empha-sis on learning, but can accommodatechanges in the duration of later

phases in any direction.

INVESTIGATING HUMAN

GROWTH PROLONGATION

These models attempt to explain ex-tremely important aspects of humanevolutionary biology. However, theyare difficult to test because of limita-

tions on comparative growth data.Several factors account for this prob-lem. Comparative ontogenetic datafrom known-age individuals of spe-

cies other than macaques and chim-panzees44 are extremely rare. Even theavailable data are inadequate formany comparisons, having been col-

lected from small captive samplesmany decades ago, then analyzedwithout the benefits of new scatterplotsmoothing techniques.45,46 Growth

studies often failed to include suffi-

cient numbers of adult animals for

gauging later growth. Certainly, new

data are emerging, particularly from

noncaptive populations,4649 and such

data will be critical to future analyses.

Unfortunately, chronological develop-

mental data for traits other than body

mass are virtually nonexistent. The only

currently viable approach to compar-

ing primate growth patterns involves

studying growth in body mass based

on data derived from captive sources.

These data have certain limitations,50,51

but these should not obscure broad-

scale interspecific trends.

Interspecific Comparisons of

Growth TrajectoriesThis study uses human growth data

representing Western populations,

which can be assumed to represent

circumstances that do not include nu-

tritional stress. Information regarding

early aspects of growth are derived

from Tanner, Whitehouse, and Taka-

hashis British growth standards.52

Because the attributes of growth

spurts are important in comparisons

across taxa, detailed data for human

growth spurts reported by Bucklers26

study of children in Leeds, UK are

used. Body mass and chronological

age data are available for nearly fiftynonhuman primate species, all fromcaptive sources.50 For the purposes ofthis study, only those species with lategrowth spurts will be quantitatively

evaluated, reducing the sample to

twenty-one species with male spurtsand only twelve with female spurts.

This protocol has been employed

previously6,25 in comparisons be-tween noncaptive and captive adultmasses.53 The analytical protocol in-volves several steps. First, mixed-lon-

gitudinal mass data collected fromcaptive individuals are plotted againstage. Second, nonparametric lowess re-gressions45,46 are calculated for males

and females. These regressions are de-sirable because they make no assump-tions about functional form and canbe applied across species. However,

lowess regressions have a tendency tooverfit some ranges of the data. Con-sequently, in a third step, data wereresmoothed using nonparametric splineregressions.54 Splines fit the data using

a broader or global criterion, resultingin smoother curves that reflect less sen-sitivity to localized differences in dataabundance. Fourth, values predicted by

spline regression are used to calculatevelocity or rate curves. These pseudo-velocity curves55 portray changes in

rates of growth (kg/yr).6,25 A pseudo-velocity curve approximates the first de-rivative of a parametric curve.

The next step is to define variousgrowth periods based on clearly de-

fined inflection points in rate curves(Fig. 2). Subdividing body massgrowth rate curves enables the analy-sis of several different variables. These

are broadly divided into variables thatdescribe either the time course or therate of ontogeny. Additional compari-sons are based on size-for-age,which represents size (kg) at various

points during ontogeny. All variablesare assessed allometrically (that is,relative to size) in two ways. First,timing and velocity variables can be

gauged relative to size-for-age. Sec-ond, these variables can be regressedon adult mass. Literature sources areused for investigating human gesta-

tion period and neonatal mass.56,57

Time components can be extractedfrom rate curves by defining segmentsalong the X (age or time) axis. Thisstudy examines only three time seg-

ments, defined by the age at growth

Comparative

ontogenetic data fromknown-age individualsof species other thanmacaques andchimpanzees areextremely rare. Even theavailable data are

inadequate for manycomparisons, havingbeen collected fromsmall captive samplesmany decades ago,then analyzed withoutthe benefits of newscatterplot smoothingtechniques.



7/14

spurt initiation, the age at peak veloc-ity, and the age at return to take-offvelocity (Fig. 2). In the interest of clar-ity, time-course results are presentedschematically for only a few selected

species. These graphs show an ex-pected time course of ontogeny basedon adult mass as well as an observedtime course of ontogeny (schematicsare expressed in natural logarithms to

facilitate direct comparisons with pre-viously presented bivariate allometricplots6). These graphs show howgrowth timing differs from allometric

expectations, enabling discriminationbetween the general and differentialextension models. Human data pointswere excluded in the calculation of

expectations for all variables.The second category of variable that

can be extracted from rate curves arevelocities (Fig. 2), or variables along

the Yaxis. Velocities are especially rel-evant to tests of metabolic risk aver-sion hypotheses because Janson andvan Schaiks43 model predicts covaria-

tion between growth rates and meta-

bolic risks. The growth rate early in

ontogeny (early growth point) is mea-sured at 50% of age at take-off veloc-ity. Finally, analyzing the size of eachspecies at various points in ontogeny(size-for-age) provides additional infor-

mation. Specifically, sizes both at theearly growth point and peak velocity areinvestigated. Size-for-age values repre-sent species averages at given ages ob-tained directly from lowess regressions.

All of these variables (timing estimates,rates, and size-for-age) have been ana-lyzed in greater detail by previous stud-ies.6,25 Here, the focus is on the both the

time course of human ontogeny and oncontrasts between early and later on-togeny. Finally, it can be noted that theeffects of phylogeny on these variables

have been discussed elsewhere.6

HUMAN GROWTH IN

COMPARATIVE PERSPECTIVE

Timing of Human Ontogeny

Gestation

The time course of human gestation

is consistent with interspecific and

size-based expectations. Martin and

MacLarnon56 reported that humans

match the typical pattern for preco-

cial mammals generally, and for pri-

mates in particular (p. 73). Their

analysis shows that the protracted hu-

man prenatal period reflected by

Schultzs diagram may be nothing

more than what is expected for a pri-

mate of our size. More importantly,

comparative data for large-bodied

apes suggest that humans do not ex-

hibit obviously extreme departures in

the attributes of the prenatal period.

These results indicate that the gesta-

tion period does not markedly pro-

long the total duration of human on-

togeny.

Postnatal growth

Schematic diagrams representing

the time course of growth point to

considerable diversity across primates

(Figs. 3 and 4). The expected durationof the first growth period (the pre-

spurt period) is independent of size

across this sample,6,25 but individual

species differ substantially. Humans

of both sexes demonstrate exception-

ally long pre-spurt periods (Fig. 4), al-

though protracted early growth peri-

ods clearly are not unique among

primates. Humans also compress late

growth periods relative to expecta-

tion. Human females show a more ab-

breviated second growth phase than

do males, but males contract the third

. . . comparative data forlarge-bodied apessuggest that humans donot exhibit obviouslyextreme departures inthe attributes of theprenatal period. Theseresults indicate that the

gestation period doesnot markedly prolongthe total duration ofhuman ontogeny.

Figure 2. Attributes of growth velocity (pseudo-velocity) curves measured in the present

study. The X axis represents age (in years), and the Y axis calibrates velocity (kg/yr) for male

red baboons. Lines connecting the velocity curve to the Y axis measure velocities analyzed.

Dashed vertical lines connecting the velocity curve to the X axis measure the age at

specific periods of growth.



8/14

phase as compared to females. Thegeneral human pattern is consistent

with that seen in other large-bodiedapes. Human growth prolongationseems to be mainly a product ofchanges in the earliest period of

growth; later phases are actually

briefer than expected.

Size During Ontogeny and

Rates of Growth

Variables that describe size duringgrowth and the velocity of body massgrowth illustrate moderate to strong

positive correlations as size increasesacross primates (Figs. 57). Duringearly ontogeny, humans generally de-viate considerably from interspecific

expectations, but these deviations are

often minimal during the growth

spurt. Thus, humans tend to fall

within the range of primate variation

as ontogeny progresses. In addition,

common chimpanzees (Pan troglo-dytes) typically deviate in the same di-

rection that humans do.

Comparisons of mass at various

points in ontogeny (size for age) inrelation to adult mass (Fig. 5AC) il-

lustrate these trends clearly. Correla-

tions between neonatal and adult

mass are strong (Fig. 5A). Further, hu-

mans deviate from interspecific ex-

pectations to a degree not seen in

other primates.57 Size at the early

growth point shows a lower correla-

tion with adult mass than does neona-

tal mass, and humans still deviate

considerably from expectations (Fig.

5B). However, during the peak of the

growth spurt, humans are nearly the

size expected for primates of our size,

and fall within the range of residual

variation.

Growth rates at different points in

ontogeny are strongly correlated with

adult size (Fig. 6AC). Human growth

rates are extremely low during early

phases, but align with interspecific ex-pectations later in ontogeny. Human

growth rates do, however, fall within

the range of residual variation even

during the early phases of growth

(particularly males). Peak velocities

for human females and males are

highly consistent with interspecific

trends.

When growth rates are plotted

against size at each growth point,

lower correlations are apparent (Fig.

7AC). Here, the pattern of deviationsis closely consistent with the pattern of

deviations seen when velocities are re-gressed against adult mass (Fig. 6). Aswith previous regressions, humans fall

well within the range of residual varia-tion as ontogeny proceeds (Fig. 7C).

HUMAN GROWTH EVOLUTION

Differential Extension of

Human Growth Phases

These analyses point to similaritiesand distinctions between humans andother primates. The time course of hu-man growth is unexceptional both be-

fore birth and after initiation of the

subadult growth spurt. In contrast,

During early ontogeny,humans generallydeviate considerablyfrom interspecificexpectations, but thesedeviations are oftenminimal during thegrowth spurt. Thus,

humans tend to fallwithin the range ofprimate variation asontogeny progresses.

Figure 3. Schematic representation of growth-period durations for selected monkey spe-

cies. The first filled area represents the period before initiation of the subadult growth spurt

(birth to take-off velocity). The second, unfilled area represents the portion of growth from

take-off to peak velocity, while the cross-hatched region denotes the period from peakvelocity to return to take-off velocity. The upper bar shows observed values; the lower bar

for each species is based on reduced-major axis predicted values at values for species

average mass. Regressions used to calculate these expected values have been presented

previously.6 Each inch represents one natural logarithm unit.



9/14

the attributes of human growth arequite distinctive during early postna-

tal phases. These results reject thegeneral extension model of humanontogeny, suggesting that Schultzs

schematic model is no longer tenableas a description of ontogenetic varia-tion among primates, at least for bodymass. Moreover, the answer to thequestion of how different humans are

from other primates depends on whatpoint in ontogeny is being investigated:early periods of postnatal growth de-viate substantially from interspecificexpectations, but later periods are

quite consistent with expectations.As previous authors1,15,38,58,59 have

recognized, spurts in the growth ofbody mass occur in many primate

species,25 but in humans are unusualin some respects. We initiate abbrevi-ated growth spurts quite late, but theother properties of the growth spurt

are comparable to those of other pri-mates. In effect, we displace an other-wise standard but condensed pri-mate subadult growth spurt to a fairly

late age. Other things being equal,later human growth periods shouldreduce the overall growth period be-cause human subadult growth spurts

are actually briefer than might be ex-

pected. The degree to which humans

displace this standard subadult pri-mate growth spurt to later ages is

quite exceptional, making increases inthe early period of human ontogenyall the more remarkable. Many hu-

man traits are extraordinary duringthe period of growth framed by birthand the initiation of the growth spurt.These unusual attributes include largeneonatal size, slow growth rates, and

deferral of the subadult growth spurt.Large size during this period is prob-ably a product of high neonatalweight. Because the duration of hu-man gestation follows a size-based ex-

pectation, it is clear that human ges-tation reflects a strategy described asenergy expensive rather than timeexpensive.60 However, just the oppo-

site strategy seems to be pursued be-tween birth and the subadult spurt,with low growth rates distributed overa long period. An energy-expensive

approach reflects heavy maternal in-vestment in gestation, while a switchto a time-expensive strategy probablyspreads the metabolic burdens of in-

fants over time.Reconstructing an ancestral condi-

tion based on our close relatives(chimpanzees, pygmy chimpanzees,

and gorillas) reinforces findings from

anthropoid-wide analyses.60 Humans,

gorillas, and pygmy chimpanzees all

show both male and female subadult

growth spurts, whereas female subadult

spurts appear to be absent in chim-

panzees.25,60 If gorillas reflect an an-

cestral condition for humans, then

prolongation of an early human

growth phase is quite exceptional be-

cause the time span before the

subadult growth spurt in gorillas is

very short. However, gorilla ontogeny

is closely tied to folivorous diets.61,62

Specifically, gorillas, like other foli-

vores, show elevated growth rates and

enhanced growth spurts. Thus, gorilla

growth patterns are derived as an ad-

aptation to folivory, manifested as aderived extension of the early growth-

spurt period. On the other hand, male

chimpanzees and both sexes of pygmy

chimpanzees have relatively long

growth periods before the subadult

growth spurt. Therefore, small differ-

ences separate the chimpanzee and

human patterns, and the deviations of

humans and chimpanzees are consis-

tently in the same direction. Humans

and chimpanzees each prolong the

earliest phase but abbreviate spurt

phases (Figs. 3, 5).

. . . later human growth

periods should reducethe overall growthperiod because humansubadult growth spurtsare actually briefer thanmight be expected. Thedegree to which

humans displace thisstandard subadultprimate growth spurt tolater ages is quiteexceptional, makingincreases in the earlyperiod of humanontogeny all the moreremarkable.

Figure 4. Schematic representation of growth periods for selected ape species. Female

common chimpanzees appear to lack a clearly defined subadult growth spurt and are thus

not used in this comparison.



10/14

At a more general level, compari-

sons reflect heterogeneity in the waysin which ontogenies can be assem-bled. For example, humans growvery rapidly over a run-of-the-mill ges-

tation period, but reverse this patternafter birth and before the growthspurt, growing slowly over a long pe-riod. Then, at the spurt, we grow

much as primates of our body sizeshould. These observations, coupledwith variation across primates, fur-ther suggest that the attributes of

these phases are highly evolvable.63 In

other words, neonatal mass, early

growth rates, and the duration of

early growth periods seem to be quiteresponsive to selection. This point isperhaps underappreciated becauseSchultzs model conveys such a strong

sense of orthogenetic regularity. Incontrast, the high level of variationobserved in the present study suggeststhat primate growth periods can re-

spond quite readily, apparently inde-pendently of body size, to selection fora delay in the initiation of subadultgrowth spurts. In addition, the vari-

able presence of subadult growth

spurts, plus their wide taxonomic dis-

tribution, suggests that later periods

of body mass ontogeny are also quiteresponsive to evolutionary pressures.

Research currently in progress onOld World monkey corroborates these

ideas. For example, skeletal growthspurts are evident in the faces, but not

necessarily in the postcranial lineardimensions, of baboons. Thus, skele-tal growth spurts appear to be mod-

ular and highly evolvable features ofontogeny: natural and sexual selec-

tion5 can effectively modify the ana-tomical and chronological place-

ment of skeletal growth spurts. Theseprocesses mean that there are also dif-ferent ways of assembling ontoge-

nies. Thus, we should expect multifac-torial causes of variation at different

stages of ontogeny and within differ-

ent anatomical units or systems. This

variation may ultimately be explained

by reference to social and ecological

selective factors, with potential links

to life-history processes.

Implications for Life-History

Models

Support for the hypothesis that ex-

tension of early growth best accounts

for human growth prolongation has

important implications for life-history

models. Comparative analyses of hu-

man ontogeny help establish the fea-

sibility of various models and may

point to additional hypotheses that

can be derived from these models.

. . . the high level ofvariation observed inthe present studysuggests that primategrowth periods can

respond quite readily,apparentlyindependently of bodysize, to selection for adelay in the initiation ofsubadult growth spurts.

Figure 5. A: Regression plot of neonatal mass against adult mass for male (left) and female

(right) anthropoid primate species. Humans are represented by male and female symbols;

African ape species are designated by letters (G gorilla, P pygmy chimpanzee, T

common chimpanzee). Ellipses represent 60% confidence intervals on the major axis re-

gression for each bivariate relation. All subsequent figures follow these conventions. B:

Regression of mass at the early growth point (1/2 age at take-off velocity) plotted against

adult mass. C: Regression of mass at peak velocity against adult mass.



11/14

Brains and learning

Extension of the early growth pe-riod is clearly compatible with a tra-

ditional brain growth or learning life-history model. Based on analyses ofindustrialized populations,64 growthin the size of the human brain occurs

only during the first seven to eightyears of life. Human body massgrowth rates, also from industrializedpopulations, increase steadily after

about two years of age until roughlytwo years after the cessation of braingrowth.52 Thus, the relationship be-tween brain growth rates and body

mass growth rates is not simply an

inverse relation, as might be expected

based on straightforward metabolic

competition between brain and

body development.65,66 Consequently,

the impact of brain growth on the to-tal duration of growth is uncertain, at

least at a morphological level. On the

other hand, processes at the cellular

level may bear a relation to the total

duration of ontogeny. Specifically,

studies at the cellular level show two

distinct phases of mammalian brain

ontogeny, termed the experience-ex-

pectant and experience-dependent

periods.67 During the experience-

expectant phase, primary synaptic

connections are overproduced, then

pruned in response to predictable en-

vironmental stimuli that are normally

present during critical developmental

periods. The subsequent experience-

dependent period responds to unpre-

dictable stimuli, leading to the storage

of information unique to the individ-

ual. This information is encoded by

forming new synapses or new connec-tions in response to new knowledge.68

These periods are crucial to normal

brain development, but they take

time. Thus, the protracted period of

human growth may be related to evo-

lutionary increases in the duration of

one or both of these brain develop-

ment periods. Furthermore, asyn-

chrony in the time courses of these

periods is evident when humans are

compared to other primates. For ex-

ample, many regions in the brain of

rhesus macaques (Macaca mulatta)

seem to pass through these stages at

about the same age.69 On the other

hand, various regions of the human

brain seem to go through experience-

expectant and experience-dependent

periods at different ages, with the de-velopment of some systems even oc-

curring during adolescence.70

These provocative findings suggest

a need for future interspecific com-

parative investigations that evaluate

potential correlations between synap-

togenesis and growth prolongation.

The present study suggests that this is

plausible in the case of humans be-

cause the early growth period is so

protracted. Although a brain-growth-

learning model is feasible, we need

much more comparative data on both

. . . the impact of braingrowth on the totalduration of growth isuncertain, at least at amorphological level. Onthe other hand,processes at the cellular

level may bear arelation to the totalduration of ontogeny.

Figure 6. A: Regression plot of velocity at the early growth point (early growth velocity)

against adult mass. B: Regression of take-off velocity against adult mass. C: Regression of

peak velocity against adult mass.



12/14

size growth of the brain and the devel-opment of synapses in the brain be-

fore we can test this model further.

Adult mortality models

Localizing human growth prolon-gation to the earliest periods of ontog-

eny may have important implicationsfor adult mortality models. Additionalevaluation of mortality models requiresmuch more demographic data.31,71

However, an adult mortality modelcould suggest that long intervals ofearly growth require exceptionallylow adult mortality. On the other

hand, extension of later periods may

not have the same requirements, in

part because offspring can contribute

to their own subsistence. It may be

possible to prolong later growth peri-

ods, and thus the total growth period,without exceptional reductions in

adult mortality.

The present analysis reveals areas of

concern for adult mortality models. It

is clear that primate growth rates vary

considerably, and that growth rates

scale to a variety of allometric coeffi-

cients,6 possibly violating assump-

tions of Charnovs adult mortality

model. Growth rate variability could

suggest variable production costs, ne-

cessitating changes in the assumption

of a growth constant. A constant

growth rate means that size increase

is a product of differing growth perioddurations, but this is not compatiblewith empirical evidence showinggrowth rate variation in primates.45

Variation in adult size can be reached

through many pathways involving

timing differences and rate differ-ences. Adult mortality models may re-quire revision in order to accommo-

date these findings.

Metabolic risk aversion

Models of metabolic risk aversionprovide compelling alternatives toboth brain growth and adult mortality

models.43 As with demographic hy-potheses, the data needed to test met-abolic models are currently insuffi-

cient.29,60,63 However, residual growth

rate variation across primates may

signal consistency with at least some

aspects of a metabolic risk-aversion

model. This variation could reflect dif-

fering profiles of metabolic risk, withperiods of high growth rates indicat-

ing low risks and periods of low

growth rates reflecting high risk.

Considerations of human growth

prolongation suggest that metabolic

risk-aversion models are compatible

with other life-history models. This is

particularly clear for the human case,

in which low early growth rates could

serve as a metabolic risk-aversion

strategy that complements processes

involved in learning or brain develop-

ment. Extended early growth at a slow

A constant growth ratemeans that size increaseis a product of differinggrowth period durations,but this is notcompatible withempirical evidenceshowing growth rate

variation in primates.Variation in adult sizecan be reached throughmany pathways . . .

Figure 7. A: Regression of early growth velocity against mass at the early growth point (EGP).

B: Regression of take-off velocity against mass at take-off velocity. C: Regression of peak

velocity against mass at peak velocity.



13/14

rate may offset the risks of developing

metabolically expensive structuressuch as the human brain (although, asnoted, the relations between the brainand body growth are far from under-stood). For example, although adults

devote about 25% of resting metabolic

energy to brain maintenance,72 in-fants devote roughly 85% to the sametask.65,66 Thus, the metabolic risks hu-

mans encounter may stem from thecosts of a large brain. Evaluating met-abolic risk models in light of informa-tion about the pattern of human

growth prolongation suggests that on-togenetic allometric analyses of brain-body relations during the early periodof growth can aid in further testing

this model.

Investment modelThis model accords well with the

observed pattern of human growthprolongation. As anticipated by themodel, slow growth for a long periodis observed, with the most obvious dif-

ferences between human- and pri-mate-based expectations occurringduring the phase when production isprobably lowest. Moreover, a large

portion of this stage corresponds tothe period of brain growth, whichmay further increase the gap between

production and consumption. Thus,this model seems to have predictivepower for slow growth rates. The per-spective of an investment model alsosuggests that early growth prolonga-

tion ensures a long period in whichthe cost of provisioning offspring ishigh. This raises particularly interest-ing problems for first offspring be-cause they may necessitate rapid at-

tainment of parenting efficiency. Thisrequirement could help explain theabbreviation of later stages of humangrowth and rapid increases in produc-

tion during these periods. An abbrevi-ated human growth-spurt period mayreflect rapid acquisition of adult skillsduring later growth in order to con-

tend with the first offspring.An investment model explains the

pattern of human growth more ade-quately than does other models in part

because it accommodates key ele-ments of other life-history models.For example, an investment model isconsistent with a brain and learning

perspective, but in a manner that does

not require strict adherence to meta-

bolic demands of brain development.

Instead, brain development is tied to

the acquisition of skill and knowledge.

In terms of this model, finding that

the earliest phases of human growth

are the most prolonged implies that

basic knowledge acquisition is more

costly in terms of production than isknowledge acquired during later phases

of growth. As with a mortality model,

a long period of child dependence

would require especially low levels of

adult mortality and high adult pro-

ductivity during the earliest years of

reproduction. Finally, an investment

model is consistent with elements of

metabolic risk-aversion models. While

slow growth reflects low production,

this model suggests that low growth

rates should also reduce the costs of

subsidizing a developing individual.

High metabolic costs of rapid growth

clearly raise the costs of subsidizingoffspring growth, but slow growthrates may help offset these costs.

CONCLUSIONS

Understanding the evolution of hu-man growth and development re-quires comparative analyses of otherprimates and tests of competing life-

history models. Comparisons betweenhumans and other anthropoids showthat Schultzs influential model of pri-mate ontogeny can no longer be ap-

plied to human growth, development,and life-history evolution. Similarly,Goulds general retardation modelmay not fit the human case. Exten-

sions of early growth periods most

clearly distinguish human ontogeny

from that of other primates. The laterstages of human ontogeny are abbre-

viated relative to interspecific expecta-tions, while other attributes of latergrowth periods are compatible withthese expectations. Humans and

chimpanzees illustrate comparable

deviations from interspecific expecta-tions.

These results support some current

life-history models while pointing toareas for new research within life-his-tory theory. Brain growth models maybe tenable, given that the majority of

derived changes in human ontogenyappear to have occurred during theearliest postnatal periods. However, itis apparent that future analyses of

brain growth models must incorpo-rate ontogenetic data as well as inves-tigate processes at the cellular level.Adult mortality models may providefurther insight into the human prob-

lem, particularly if they can shed lighton the correlations between extendedearly development and adult mortal-ity. Comparisons of growth rates

across primates may prompt modifi-cations to adult mortality models. Ju-venile metabolic risk models also haveconsiderable potential to account for

patterns of growth across primates.Finally, an investment model, with aspecial focus on investment during

early growth, is compatible with theapparently derived pattern of humanbody mass ontogeny.

ACKNOWLEDGMENTS

Data for this study were providedthrough the generosity of many re-search institutions and zoologicalparks. Data from the Yerkes PrimateCenter were supported by NIH grant

RR-00165. I thank John Fleagle forconsiderable patience during prepara-

tion of the manuscript. ProfessorsKim Hill, Bill Jungers, and Brian

Richmond offered helpful advice onearlier versions of the paper, as didseveral anonymous referees.

REFERENCES

1 Bogin B. 1988. Patterns of human growth. NewYork: Cambridge University Press.

2 Shea BT. 1990. Heterochrony in primates. In:McKinney MJ, editor: Heterochrony in evolu-tion. New York: Plenum Publishing. p 237266.

3 Shea BT. 1992. Developmental perspective onsize change and allometry in evolution. Evol An-

thropol 4:125134.

An investment modelexplains the pattern ofhuman growth moreadequately than does

other models in partbecause itaccommodates keyelements of other life-history models.



14/14

4 Smith BH.1992. Life history and the evolution ofhuman maturation. Evol Anthropol 4:134 142.

5 Leigh SR. 1995. Socioecology and the ontogenyof sexual dimorphism in anthropoid primates.Am J Phys Anthropol 97:339 356.

6 Leigh SR, Park PB. 1998. Evolution of humangrowth prolongation. Am J Phys Anthropol 107:331350.

7 Charnov EL, Berrigan D. 1993. Why do female

primates have such long lifespans and so fewbabies? Or life in the slow lane. Evol Anthropol1:191194.

8 Stearns SC. 1992. The evolution of life histo-ries. Oxford: Oxford University Press.

9 Kaplan H, Hill K, Lancaster J, Hurtado AM.2000. A theory of human life history evolution:diet, intelligence, and longevity. Evol Anthropol9:156 183.

10 Hill K, Kaplan H. 1999. Life history traits inhumans: theory and empirical studies. Ann RevAnthropol 28:397 430.

11 Lovejoy AO. 1959. Herder: progressionismwithout transformism. In: Glass BH, Temkin O,Strauss WL, editors. Forerunners of Darwin:17451859. Baltimore: Johns Hopkins UniversityPress. p 207222.

12 Locke J. 1690. Second treatise on govern-ment. Cambridge: Cambridge University Press.

13 Fiske J. 1909. The meaning of infancy. Bos-ton: Houghton-Mifflin.

14 Schultz AH. 1969. The life of primates. NewYork: Universe Press.

15 Gavan JA. 1982. Adolescent growth in nonhu-man primates: an introduction. Hum Biol 54:15.

16 Gavan JA, Swindler DR. 1966. Growth ratesand phylogeny in primates. Am J Phys Anthropol24:181190.

17 Lovejoy CO. 1981. The origin of man. Science211:341350.

18 Harvey PH, Martin RD, Clutton-Brock TH.1987. Life histories in comparative perspective.In: Smuts BB, Cheney DL, Seyfarth RM, Wrang-

ham RW, Struhsaker TT, editors. Primate societ-ies. Chicago: University of Chicago Press. p 181196.

19 Harvey PH, Clutton-Brock TH. 1985. Life his-tory variation in primates. Evolution 39:559 581.

20 Ross C. 1988. The intrinsic rate of naturalincrease and reproductive effort in primates. JZool (Lond) 214:199 219.

21 Gould SJ. 1977. Ontogeny and phylogeny.Cambridge: Harvard University Press.

22 Periera ME, Leigh SR. 2001. Modes of pri-mate development. In: Kappeler P, Periera ME,editors. Primate socioecology and life history.Chicago: University of Chicago Press. In press.

23 Bogin B. 1997. Evolutionary hypotheses forhuman childhood. Yearbook Phys Anthropol 40:6390.

24 Bogin B. 1999. Evolutionary perspective onhuman growth. Ann Rev Anthropol 28:109 153.

25 Leigh SR. 1996. Evolution of human growthspurts. Am J Phys Anthropol 101:455 474.

26 Buckler J. 1990. A longitudinal study of ado-lescent growth. London: Springer-Verlag.

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

28 Hill K, Hurtado AM. Ache life history. Haw-thorne, NY: Aldine de Gruyter.

29 Garber PA, Leigh SR. 1997. Ontogenetic vari-ation in small-bodied New World primates: im-plications for patterns of reproduction and infantcare. Folia Primatol 68:122.

30 Mayr E. 1963. Animal species and evolution.Cambridge: Harvard University Press.

31 Ross C, Jones KE. 1999. Socioecology and the

evolution of primate reproductive rates. In: LeePC, editor. Comparative primate socioecology.Cambridge: Cambridge University Press. p 73110.

32 Bolingbroke H St. J. Lord Viscount. 1754.Fragments or minutes of essays. Hildesheim:Greg Olms Verlagsbuchhandlung (1968 facsimileedition).

33 Fiske J. 1885. The destiny of man, viewed in

the light of his origin. Boston: Houghton, Mifflin.34 Spencer H. 1888. The principles of biology,Vol. I. New York: D. Appleton.

35 Dobzhansky T. 1962. Mankind evolving. NewHaven: Yale University Press.

36 Wasburn SL, Hamburg D. 1965. Implicationsof primate research. In: DeVore I, editor. Primatebehavior: field studies of monkeys and apes. NewYork: Holt, Rinehart, and Winston. p 607 622.

37 Poirier FE. 1977. Introduction. In: Chevalier-Skolnikoff S, Poirier FE, editors. Primate bioso-cial development. New York: Garland Publish-ing. p 1 42.

38 Tanner JM. 1978. Foetus into man. Cam-bridge: Harvard University Press.

39 Smith BH, Tompkins RL. 1995. Toward a lifehistory of the Hominidae. Ann Rev Anthropol

24:257279.40 Sarnat HB, NetskyMG. 1981. Evolution of thenervous system. New York: Oxford UniversityPress.

41 Charnov EL. 1991. Life history invariants.New York: Oxford University Press.

42 Pagel MD, Harvey PH. 1993. Evolution of thejuvenile period in mammals. In: Periera ME,Fairbanks LA, editors. Juvenile primates: Lifehistory, development, and behavior. New York:Oxford University Press. p 28 37.

43 Janson CH, van Schaik CP. 1993. Ecologicalrisk aversion in juvenile primates: slow andsteady wins the race. In: Pereira ME, FairbanksLA, editors. Juvenile primates. New York: OxfordUniversity Press. p 5774.

44 Watts ES, Gavan JA. 1982. Postnatal growthof nonhuman primates: the problem of the ado-lescent spurt. Hum Biol 54:5370.

45 Efron B, Tibshirani R. 1991. Statistical dataanalysis in the computer age. Science 253:390 395.

46 Moses LE, Gale LC, Altmann J. 1992. Meth-ods for analysis of unbalanced, longitudinal,growth data. Am J Primatol 28:49 59.

47 Pusey A. 1990. Behavioral changes at adoles-cence in chimpanzees. Behaviour 115:203246.

48 Strum SC. 1991. Weight and age in wild olivebaboons. Am J Primatol 25:219 237.

49 Cheverud JM, Wilson P, Dittus WPJ. 1992.Primate populations at Polannaruwa. III. So-matometric growth in a natural population oftoque macaques. J Hum Evol 23:5177.

50 Leigh SR. 1992. Ontogeny and body size di-morphism in anthropoid primates. Ph.D. disser-

tation, Northwestern University, Evanston, Il.

51 Leigh SR. 1992. Patterns of variation in theontogeny of primate body size dimorphism. JHum Evol 23:2750.

52 Tanner JM, Whitehouse RH, Takahashi M.1966. Standards from birth to maturity forheight, weight, height velocity, and weight veloc-ity: British children, 1965, Part II. Arch Dis Child41:613 635.

53 Leigh SR. 1994. The relations between captiveand noncaptive weights in anthopoid primates.Zoo Biol 13:21 44.

54 Schluter D. 1988. Estimating the form of nat-ural selection on a quantitative trait. Evolution42:849 861.

55 Coelho AM. 1985. Baboon dimorphism:growth in weight, length and adiposity from

birth to 8 years of age. In: Watts ES, editor.Nonhuman models for human growth and devel-opment. New York: Alan R. Liss. p 125 159.

56 Martin RD, MacLarnon AM. 1990. Reproduc-tive patterns in primates and other mammals.The dichotomy between altricial and precocialoffspring. In: DeRoussea CJ, editor. Primate lifehistory and evolution. New York: Wiley-Liss. p47 88.

57 Smith RJ, Leigh SR. 1998. Sexual dimor-phism in primate neonatal body mass. J HumEvol 34:173201.

58 Bogin B. 1994. Adolescence in evolutionaryperspective. Acta Paediatr 40(suppl):29 35.

59 Watts ES. 1982. Adolescent growth and devel-opment of monkeys, apes, and humans. In: WattsES, editor. Nonhuman primate models for hu-man growth and development. New York: AlanR. Liss. p 41 65.

60 Lee PC. 1999. Comparative ecology of postna-tal growth and weaning among haplorhine pri-mates. In: Lee PC, editor. Comparative primatesocioecology. Cambridge: Cambridge UniversityPress. p 111139.

61 Leigh SR. 1994. Ontogenetic correlates of dietin anthropoid primates. Am J Phys Anthropol94:499 522.

62 Leigh SR, Shea BT. 1996. The ontogeny ofbody size variation in African apes. Am J PhysAnthropol 99:43 65.

63 Houle D. 1992. Comparing evolvability andvariability of quantitative traits. Genetics 130:195204.

64 Jolicoeur P, Baron G, Cabana T. 1988. Cross-sectional growth and decline of human statureand brainweight in 19th century Germany.Growth, Dev, Aging 52:201206.

65 Leonard WR, Roberston ML. 1992. Nutri-tional requirements and human evolution: abioenergetics model. Am J Hum Biol 4:179 195.

66 Leonard WR, Robertson ML. 1994. Evolu-tionary perspectives on human nutrition: the in-fluence of brain and body size on diet and me-tabolism. Am J Hum Biol 6:77 88.

67 Greenough WT, Black JE, Wallace CS. 1987.Experience and brain development. Child Dev58:539 559.

68 Kleim JA, Vij K, Ballard DH, Greenough WT.1997. Learning-dependent synaptic modifica-tions in the cerebellar cortex of the adult ratpersist for at least four weeks. J Neurosci 17:717 721.

69 Rakic P, Bourgeois J-P, Eckenhoff MF,Zecevic N, Goldman-Rakic PS. 1986. Concurrentoverproduction of synapses in diverse regions ofthe primate cerebral cortex. Science 232:232235.

70 Huttenlocher PR, Dabholkar AS. 1997. Re-gional differences in synaptogenesis in humancerebral cortex. J Comp Neurol 387:167178.

71 Gage TB. 1998. The comparative demographyof primates. Ann Rev Anthropol 27:197221.

72 Lehninger AL. 1982. Principles of biochemis-try. New York: Worth Publishers.

73 Rushton J-P. 1995. Race, evolution, and be-havior. New Brunswick, NJ: Transaction Pub-lishers.

74 Armelagos G. 1995. Race, reason, and ratio-nale. Evol Anthropol 4:103109.

75 Harpending HH. 1995. Human biological di-versity. Evol Anthropol 4:99 103.

76 Lovejoy AO. 1936. The great chain of being.Cambridge: Harvard University Press.

77 Fleagle JG. 1988. Primate adaptation and evo-lution. New York: Academic Press.

78 Fleagle JG. 1999. Primate adaptation and evo-lution, 2nd ed. New York: Academic Press.

2001 Wiley-Liss, Inc.DOI 10.1002/evan.20002


Documents

i Evolution of Human Growth