Evolutionary Perspective on Human Growth Author(s): Barry Bogin Reviewed work(s): Source: Annual Review of Anthropology, Vol.

28 (1999), pp. 109-153 Published by: Annual Reviews Stable URL: http://www.jstor.org/stable/223390 . Accessed: 01/07/2012 20:32
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp

.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.

Annual Reviews is collaborating with JSTOR to digitize, preserve and extend access to Annual Review of Anthropology.

http://www.jstor.org

Annu. Rev. Anthropol.1999. 28:109-53 CopyrightD 1999 by Annual Reviews. All rights reserved

ON HUMAN EVOLUTIONARY PERSPECTIVE GROWTH

Barry Bogin
Behavioral Sciences, Universityof Michigan-Dearborn, Dearborn, Michigan 48128; e-mail. bbogin@umich.edu

Key Words:

evolution, life history, mammalian growth, primate growth, adolescent growth spurt

U Abstract To better understandboth the shared and special features of human growth, this article explores the evolution of growth patterns of mammals in general and primates in particular. Special attention is paid to several competing hypotheses concerning the adaptive value of the juvenile stage to the life history of the social mammals. One hypothesis claims that all social mammals have ajuvenile stage of life, but although most primate species are social, not all primates show a juvenile stage of life history. There is also controversy over whether the adolescent growth spurt is a uniquely human feature. On the basis of empirical observations and evolutionary considerations, I conclude that the human adolescent growth spurt in stature and skeletal maturation is species specific, not found in any other primate species. Finally, data and theory are used to advance a philosophy of human growth. An acceptable philosophy must acknowledge the mammalian and primate foundations for the human pattern of growth. But a robust philosophy of human growth must also account for the ecology to which the human species, indeed any species, is adapted. Accordingly, a philosophy of human growth must allow for the evolution of variations on common themes and of new stages of growth that may be unique to the human species.

CONTENTS Introduction ................................................... The Basic Curve of Growth ......... ............................. Stages of Mammalian Growth ........ ............................ Juvenile Mammals .................................................. Hypothesesfor the Evolution of the Juvenile GrowthStage ..... Primate Growth Patterns .........................................

............

110 110 111 118 118 122

0084-6570/99/1015-0109$ 12.00

109

110

BOGIN Some Historical Background ......................................... Of Brains and Bodies ............................................... 125 128

Is the HumanAdolescent Growth SpurtUnique? .......

A Surveyof WeightGrowthin Primates................................. Skeletal Growthin Primates..........................................

..............

129
130 132

Skeletal GrowthSpurtsin Primates.................................

TheHumanAdolescent GrowthSpurt ............ Human-RhesusDifference in Growth Velocity. ...................... ..........................

137
141 141

Patternsof Growthfor HumansVersus OtherPrimates ...... A Philosophy of HumanGrowth ..................................

..........

144 146

INTRODUCTION
Biologists and anthropologistshave proposed a numberof taxonomic schemes for classifying Homo sapiens. Lovejoy (1981) suggested five characteristicsof humansas defining features:bipedality,a largeneocortex,reducedanteriordentition with molardominance,materialculture,andunusualsexual and reproductive behavior. The developmentof each characteristiccan be seen in the ontogenetic unfolding of the humanpatternof growth and development.For instance, bipedality is made possible by differential growth of the legs and pelvis, including the bone, muscle, tendons, and ligaments, versus the arms and shoulder girdle (Simpson et al 1996). Our unusualreproductivebehaviorresults, in part, from a prolonged childhood, delayed maturation,and species-specific neuroendocrine physiology. The humanpatternof endocrinephysiology results in the menstrual cycle of women, continuoussexual receptivityof both sexes, developmentof our secondary sexual characteristics,e.g. hair patches on the groin and armpits,and menopause in women. These are only some of the features of H. sapiens and humangrowth. Although these characteristicsset us apartfrom all other species, they have their origins in evolutionaryhistory. Thus, we share many basic growth patterns with other species, but we differ in other ways because of special evolutionary developments. To better understandthe shared and special features of human growth, this article reviews the phylogeny of growth of mammals and primates-the groupthatincludes monkeys, apes, andhumans.Materialfor this article is taken from a broad survey of the relevant literature,including historically of data (Bogin 1988, importantwork, my research findings and interpretation 1991,1993, 1994a,b, 1995, 1996, 1997, 1999), andnew researchpublishedin the last year.

THE BASIC CURVE OF GROWTH

The growth and development of any organism may be viewed as movement throughspace or time (Thompson 1917, 1942). They move throughtime in two fromearlier ways: in days, months,andyears;and in developmentandmaturation

OFHUMAN GROWTH EVOLUTION

111

to later stages of life. Both ways can be representedin mathematicalform. Most nonhumanorganisms share the same basic curve of growth, an S-shaped, or sigmoid, curve (Figure 1A). Chickens, rats, and cattle follow this growth curve (Brody 1945, von Bertalanffy 1960, Laird 1967, Timiras 1972). The growth of parts of organismsand colonies of cells, such as bacteriaand tumors in animals, also conformto this sigmoid curve. Figure lB and C illustratesome othermathematical features of the general growth curve. The velocity, or rate of growth, is given in Figure 1B; only a single peak, or maximumrate, of growth occurs, indicating an initial accelerationfollowed by a period of decelerationin growthrate. In Figure 1C, the changes in accelerationare clearer:Point zero accelerationcorresponds to the inflection point in the velocity curve where the rate of growth stops increasing and begins to slow. Growthrate at any subsequentpoint on the curve decreases with time. In mathematicalterms,partsB and C of Figure 1 arerelatedto A as its first and second derivatives,respectively. That the curve of general growth is completely differentiablemeans, mathematically,thatthe growthprocess representedby this curve is smooth and continuous. That continuity means that we can predict changes in amounts and rates of growth duringthe course of development with precision. Such predictionsallow us to make quantitativeand qualitativeassessments of the growth of any individual organism and to make comparisons between differentindividuals,groupsof individuals,and even differentspecies of animals in terms of the mathematicalpropertiesof their growth curves.

OF MAMMALIAN STAGES GROWTH

All placental mammals share the basic skeletal and reproductive adaptations relatedto rapidand flexible locomotion and to efficient feeding of both the fetus, via a placenta, and the infant,via lactation.They also sharethe neurological systems, especially the brain, which allow for intense parentalinvestmentand high levels of learningby the offspring. In contrastto these universals of mammalian biology and behavior, mammals can also be grouped by pattems of postnatal growth. The velocity growth curve of the mouse (Figure 2) and of the Holstein cow (Figure 3) show the patternthat is typical for most mammals. The shapes of these velocity curves are fundamentallydifferent from human growth curves (Figures 4 and 5) in several ways. First, humans achieve their maximum rate of growth in length and weight during gestation, and postnatal growth rates decelerate during infancy. Humansachieve such rapid growth during gestationbecause they have both only one fetus per pregnancy(the most common case) and the most efficient placentation of all mammals. Mice and cows reachtheirmaximumrateof growthafterbirth,duringthe infancy stage. For mice this is probably because of competition among multiple fetuses for placental resources (litters are the norm). For both mice and cows, lactation provides improved nutrition over that provided via the placenta, which may promote a surgein growthrate:andforboth,growthratesbegin to decline soon afterweaning.

112

BOGIN

10 10 -Figure 8 7
-

/ / /

6 5 3 2 1

0
1 2 3 4 5 6 7 8 9 10 Time 10
9

growth 1 General curves: (A) weight versus time, (B) rate of growth (velocity) versus time, and (C) acceleration of growth rate versus time. The units of time (X axis) and growth (Y axis) are arbitrary. (From Medawar 1945.)

89 B

~
0

7 6

>5

4
3

2
0

1F
1 2 3 4 5 6 Time 7 8 9 10

10

8 .0 6 a) 4

-2 -4
-6

10

Time

EVOLUTIONOF HUMAN GROWTH

113

The second difference is thatfor mice and cows, sexual maturation occurs relatively soon afterweaning andjust afterthe maximal rate of growth.Within a few days of weaning, vaginal opening occurs in female mice and they can become pregnant. About two weeks after weaning, male mice produce sperm. Human growth and maturationcontrast sharply with that of mice and cows because humans delay puberty and sexual maturationfor many years after the end of infancy, i.e. afterweaning. Weaning is defined here as the cessation of feeding by lactation;for most mammals, the infant stage of life history ends with weaning. Life history stages are periods of time that an organismdevotes to growth,maintenance, reproduction,raising offspring to independence,and death. In the broad course of mammalian evolution, several different life history strategies have emerged. The humandelay in pubertyand sexual maturationis even longer if measured from the time of maximalrateof growthafterconception.Mice and cows achieve peak growthratevelocity afterbirth,but because humanpeak velocity occurs before birth, humans experience an average delay of more than a decade between gestation and puberty. A thirddifference is thatfor mice and cows, pubertyoccurs while growthrates are in decline but still near the maximal rate. Human puberty occurs when the postnatalrateof growthin height and weight are at theirlowest points since birth. The fourthdifference is that humanpubertyinitiates a series of changes in endocrine function that leads to the adolescent growth spurt.In contrast,growthrates of mice and cows continueto decline afterpuberty.The fifth, and last, difference

*
1.00

Males

-+-

Females

0.80
0)

0 10

20

30

40

50

60

70

80

90 100 110 120

Days postpartum
Figure 2 Velocity curvesforweightgrowthin the mouse.Weaning(W) takesplace between 10 days and20. Inbothsexes, puberty(*P), meaningvaginalopeningfor females or spermatocytes in testes of males, occursjust after weaning and maximal growthrate. [Fromdatareported by Tanner(1962).]

114

BOGIN

is that soon afterpuberty,mice are fertile and begin to reproduce,but for humans there is a delay of years between pubertyand the onset of the adultreproductive stage of life. Many of these differences between humanand nonhumanmammaliangrowth were recognized previously. In a classic analysis of mammaliangrowth, Brody thatmost mammalshave a patternof growth similar to that (1945) demonstrated of mice. Brody's analysis was confirmedby von Bertalanffy(1960) and Tanner (1962). Tannerpresents many growth curves for mice, rats, rabbits, sheep, and cattle, showing that size makes little difference in the pattern of growth. The growthcurve of Holstein cattle, animalswith a mean adultweight of 2640 kg and a shoulderheight of 135 cm (Figure 3), is virtually identicalto that of mice (Figure 2).

COI

EB
Fiue3Mnhy

5
an nhihta

0
h ihr

5202
foheeralvreis ofcos.

(DatafromBrody 1945.) Monthlyvalues for height as the meanvalue for manyanimonthlyfrombirthto mals (samplesize variesfrom67 to 239 at each age) measured 24 months.

EVOLUTIONOF HUMAN GROWTH

115

80

Boys
-

E 70-

60 -

c50
-J/
-

40 30 10 _

T
0

320

s1
0

Birth t l l l I 1 10 20 30 40 10 20 30 40 50 Postmenstrual Postnatal

I I I

Age (weeks)
~~~~~~~Boys

>43

2
I
I AI

Birth
I

10 20 30 40 Postmenstrual

10 20 30 40 Postnatal

50

Age (weeks)
Figure 4 Distance and velocity curves for growth in body length during human prenatal and postnatal life. The figure is diagrammatic, based on several sources of data. (Dotted lines) The predicted curve of growth if no uterine restriction takes place. In fact, such restrictions do take place toward the end of pregnancy and may impede the flow of oxygen or nutrients to the fetus. Consequently growth rate slows but rebounds after birth and returns the infant to the size she or he would be without any restriction. (From Tanner 1990.)

116

BOGIN

200

180

ci 160 0 C 140 cu X 120 100 s 80 60

40 l

A
20 t 18 E 16 14 12 *o .010 0 8
~6I. ID 2

I,I.I,
4

6 8 10 12 14 16 18 20 22

l,,

I,

0
0

IN4

2 4

6 8 10 12 14 16 18 20 22

AGE, years

Figure 5 Idealizedmeandistancecurves(A) and velocity curvesof growth(B) in height for healthygirls (dashed lines) and boys (solid lines) showingthe postnatal stages of the patternof humangrowth.(B) Note the spurtsin growthrate at midchildhoodand adolescencefor both girls and boys. The stages of humanpostnatal as follows: I, infancy;C, childhood;J, juvenile; A, adolesgrowthare abbreviated (1984) the curvescome fromPrader cence;M, matureadult.[Dataused to construct andBock & Thissen(1980).]

EVOLUTION OFHUMAN GROWTH

117

These three researchersare in agreementthat the majorityof mammalsprogress from infancy to adulthoodseamlessly, withoutinterveningstages of development. In contrast, the path from conception to maturityis different for human beings. Brody states that his analysis "demonstrates the close similaritybetween the age curves of differentanimalspecies. The humanage curve, however, differs from the others in having a very long juvenile period, a long interval between weaning andpuberty(approximately3 to 13 years);this period is almost absentin laboratoryand farm animals. In these animals, weaning merges into adolescence without the interventionof the juvenile phase found in man"(1945:495). By today's standards,however, this terminology is out of date. For example, no distinction is made between puberty and adolescence, and both are equated with sexual maturation. We know now thatpubertyis an event of the neuroendocrine systems in the brain that control sexual maturation.In humans this event occurs at the end of the juvenile period and before the start of the adolescent growth spurtand life stage. We also know that only humans, and perhaps some apes, have an adolescent stage of growth and of life history.Nevertheless, Brody correctlynotes thatof the animalshe studied,only humansand chimpanzeeshave a juvenile growth period. Despite this importantdiscovery, Brody still tried to show that in essence all animals, parts of animals, colonies of cells, and even populationsof animalsfollow a common patternof growth."Thegeneralsimilarity between the curves of growth of individualsand of populationsis not surprising, since ultimately both are collections of individuals. Ourbodies are made of cells, and our bodies, in turn,are cells in a social body. Individualsare organisms and also units of a largerorganism,an epiorganism"(Brody 1945:495). Brody was searchingfor a type of grandunificationtheory of growth.His specific hypothesis was that all growing organisms and populations or organisms could be modeled with one curve of growth. He aligned the maximum rate of growth afterbirthfor many species, includingmice, cows, and humans,and then showed thatall have two phases of growthvelocity. The first is a self-accelerating phase, from birthto peak velocity, and the second is a self-inhibitingphase, from peak velocity to adult size. Humans, even with theirjuvenile growth period and adolescent growth spurt,would have to fit into this unification of growth curves. Accordingly, Brody equated the adolescent growth spurt of humans with the velocity curve of growth for animals such as mice, rats, and many domesticated farm animals (cows, pigs, sheep). In this hypothesis, all mammals followed the same curve of growth,but some mammalshad their growth spurtsoon afterbirth (e.g. mice and Holstein cows), whereas for others (e.g. humans), the spurtwas delayed until later in life. In many ways, Brody's methodology and speculative ideas are ingenious and stimulating.His analyses and findings apply to the laboratory and farmanimalsin his database,but they do not apply to humans.The main problemwith the unification hypothesis is that some animals, especially humans, have more than one phase of self-accelerationand self-inhibitionof growthrate.In fact, humanshave three such phases: The first occurs duringgestation, the second is the midgrowth spurt, and the third is the adolescent growth spurt. Brody knew only about the

118

BOGIN

adolescent growth spurtand incorrectlyequatedit with the solitary growth spurt of his laboratoryand farm animals. Brody's work is mentionedhere because it had a stronginfluence on the study of humangrowthfor many years. Even today, it is common to read aboutadolescent growth spurtsor an adolescent phase in animals, such in as rodents or farm animals,but it is not like humanadolescence. Although these ideas seem preposteroustoday, bearin mindthatduringthe firstthirdof the twentiethcentury,when Brody carriedout most of his originalresearch,the natureof the humanpostnatal growth curve, even the existence of the adolescent growth spurt,was still being debated(Bogin 1999:Ch. 1). The humanpostnatalstages of childhood andjuvenile growth,as well as the humanprenatalcurveof growth(Figures4 and 5), were unknownto most researchersat that time. Moreover,the neuroendocrinecontrol of growth and of pubertywere poorly understood.

JuvenileMammals
Brody's discovery of the humanjuvenile growthperiod is one of his lasting contributionsto the field. Since that time, juvenile growth stages have been discovered for several other mammalianspecies. Highly social mammals, such as the social carnivores(wolves, lions, hyenas), elephants,many cetaceans (porpoises, whales), and most primates, all evolved a new stage of development between infancy and adulthood-the juvenile stage (Bekoff & Byers 1985, Pereira& Fairbanks 1993). Juveniles may be defined as offspring who are weaned, i.e. no longer dependenton maternallactationfor feeding, but who are still prepubertal. Juveniles are largely responsible for their own care and feeding. They must find their own food, avoid predatorson their own, and compete with adults for food and space. Juveniles may even compete with their own mothers, who may be encumberedwith another pregnancy or a nursing infant. Clearly, the juvenile stage adds several new risks along the pathtowardreproductivematurity.In fact, the highest ratesof postneonatalmortality(i.e. deathsafterthe first monthfollowing births) for social mammals occur duringthe juvenile stage (Pereira& Fairbanks 1993).

Hypotheses for the Evolution of the JuvenileGrowthStage

The juvenile growth stage must have evolved because it added some benefits to the life of social mammals,but there is some debateas to its function.A stimulating review and analysis of the evolution of this stage is offered by Janson& van Schaik(1993). They proposetwo benefits. The firstis thatthejuvenile periodprovides the extendedperiod of braingrowthand learningtime necessary for reproductive success in various species of social mammals (the traditionallearning hypothesis explanation).Social carnivores,elephants,andprimatesmust all learn how to live within the social hierarchyof the group.They must also acquirecomplex feeding skills requiredfor huntinganimalprey, opening fruits or seeds with protective coverings, and discerningwhere and when food can be found. Reproductive skills must also be learned,including how to compete for mates and care for offspring. Johnson(1982) and Lancaster(1985) show thatthe selective bene-

EVOLUTION OFHUMAN GROWTH

119

fit of learning is that it permits adaptationto unpredictableecological changes. Includedare common problems faced by all extant and extinct huntingand gathering societies, such as seasonal variability in climate, plant growth, and animal migrations. The unique capacity for learning in humans is also brought to the forefront in rare crises, such as the 1943 drought in CentralAustralia. Birdsell (1979) relates thatduringthis time, an old Aborigine man, by the name of Paralji, led a band of people on a 600-km trek in search of water. After passing 25 dry waterholes,he finally led them to a fallbackwell. The old man hadnot visited that place for more than 50 years. The well was also dry, though, and so Paraljiwas forced to trek 350 km more on ancient trails, seeking to locate water holes by place names learnedfrom initiationrites and ceremonialsongs he had memorized as a juvenile. Even the way humansforage for food requiresextensive learninggained during the childhoodandjuvenile growthperiods. Much of the food humansutilize is hidden from view or is encased in protectivecoatings;tools areusually needed to extractand process these foods. The costs of tool manufacture,however, and the time and energy required to find and process raw materials are outweighed by the benefits. Tool-using human gatherersextract twice as many calories from savanna/woodlandenvironmentsas do non-tool-using primates.Otherfoods are poisonous before processing by washing, leaching, drying,or cooking. For example, acornsandhorse chestnuts,eatenby manyNorthAmericanIndians,andmanioc, a staple food for many extantculturesin the tropics,aretoxic if eaten raw and must be leached by boiling in water and driedbefore consumption.Furthermore, one must acquire knowledge of the location and methods of processing these foods, as well as the location of the necessaryraw materialsandtheirmanufacture into tools. One example is providedby the huntingand gatheringAruntaof Central Australia. They are compelled to live in self-sufficient nuclear families because in their habitatfood resources are widely dispersed (Service 1978). The children, once able, follow their mothersand fatherson the daily rounds of food collection and preparation. Elkin (1964) observed thatboys and girls as young as 5 years were takenby their fatherson huntingtripsand shown how to collect and prepareraw materialsfor use in spears, points, and other tools. Because it takes more than a decade to become proficient in the manufactureand use of these tools, early learning, slow growth, and delayed maturationduring the juvenile stage are all mutuallybeneficial. Learningseems like a greatidea, but the learninghypothesis does not take into considerationthe high mortalityassociatedwith thejuvenile stage. After all, evolution works by differentialmortalityandreproductivesuccess, and high levels of mortalityduringthe prereproductive juvenile stage will not lead to evolutionary success. The second benefit of juvenility, and a complement to the learning hypothesis, is the ecological risk aversion hypothesis. Janson & van Schaik (1993) developed this hypothesis to deal with the issue ofjuvenile mortalityrisk: When the risk of predationis high for the individualsof a species, naturalselection often favors the formationof social groups-there is relative safety in numbers. The formationof social groups comes at a cost, however, and that cost is an

120

BOGIN

increasein competitionfor food within the group-there aremore mouthsto feed. According to Janson& van Schaik, newly weaned individualsare most likely to be adversely affected by this competitionbecause they have less-developed foraging skills than adults do. The young must develop and grow to maturesize to reproduce,and there are two basic ways to get from infancy to adulthood.One is to develop quickly and minimize the duration of the nonreproductiveperiod between weaning and sexual maturation.This is the strategyfollowed by mice, bison, and most mammals. Because there is little time to learn, it places a premium on reproductive quantity rather than quality. The other strategy is to develop toward adulthood slowly, increase the time for learning, and produce higher-qualityadults. A corollary of each strategy for mammals of medium-tolarge body size is thatfast developmentrequiresfast growth,but slower development allows for slow growth.Janson& van Schaikhypothesizethatslow juvenile growth reduces the risk of death and simultaneouslyprovides the time for juveniles to learn. The risk of death during competition for food is an example. Juveniles must forage for theirown food, a skill thatmust be practiceduntil maturelevels of success are achieved. Muchof theirforagingis in competitionwith adults.This competition becomes clearest duringtimes of food scarcity, when juvenile primates die in greaternumbersthaninfantsor adults.But not all juveniles die, and thereis a relationship between growth and mortality. A small, slow-growing juvenile requiresless food than a larger,fast-growingindividualand may survive periods of food scarcity;thereis evidence for this from studiesof both primates(Janson& van Schaik 1993) and elephants(Lee & Moss 1995). Moreover,slow growthand smaller size allow juveniles to practice feeding skills with less risk of starvation duringall seasons. According to this risk-aversionhypothesis, the juvenile stage evolved when the rate of growth was slowed and the durationof growth prolonged. An example of thejuvenile patternof growth is given for elephants(Figure6). The distance curves are based on cross-sectional means of shoulder height, a common measure of skeletal growth for elephants [see Lee & Moss (1995) for The originalmeanswere plotted and details of the methodology of measurement]. fit with a distance-weighted,least-squaresregression.The curveproducedby this statisticaltechniquerepresentsthe maintrend,or averagegrowthpattern.The distance curve for both male andfemale elephantsshows the typicaljuvenile pattern: many decades of prolongedandrelatively slow growth.The incrementsin growth between successive datapoints were calculated,and as was done for the distance curves, the increments between means were plotted and fit with a distanceweighted least-squares regression. The shape of this fitted increment curve reveals the stages of postnatalgrowthmore clearly thanthe distance curves. The durationsof infancy,juvenile, andmatureadultstages areindicated. approximate with a greatdeal of cautionbecause The incrementcurves must be interpreted they are calculated from means ratherthan from individualdata. Incrementsare of growthvelocity. Overall,the presentedonly to provide a rough approximation fitted curve shows a fairly rapiddecline in rateof growthduringthe first decade of

EVOLUTIONOF HUMAN GROWTH

121

Elephat Shouhbnh1Jr

OV M0

"

320.
-280
240
'

>
( -1

l-----------:'
!.......
..............t.....
-

200

2Xtet-

-~~~~~~~~~~~*-1

' 0

F......

~12O.
A
40

.....I.t.
10 20 40 '30 AGE,yams s0 so 70fl.Fmale

?
1

Shoulder Elephant Height nm entcurveS

7a.

..

. r.-......... af......

k I... . fi ro
(0 ~0

~ ~~~~~~~~~~~~.........."~

2o

Ta

1.

1
a
50 60 70

-20 -L X W

P
10

. . ....,.
20 30 40

"'D Feale

years Age,
the patternof growthin social Figure 6 Growthcurves for elephantsillustrating adult(M) stagesof growth.(A) Diswith infant(I),juvenile(J), andmature mammals tance curves, (B) incrementcurves. Weaning(W) takes place at a mean age of 40 age 9 in females,with firstbirthat apmonths.Puberty(P) occursat approximately untilafter25 yearsof age. [Themean proximately10 years.Males do not reproduce supplied by Dr.PCLee, Department kindly heightwere of Biologivaluesof shoulder Cambridge, and publishedby Lee & Moss (1995).] of University cal Anthropology,

122

BOGIN life and then a slower, prolonged,decline to age 25 years for females and age 50 for males. The difference seen in the distance curve in shoulderheight between the sexes is explainedby the additional25 years of growth for males. The increment estimates fluctuatewidely, and some are even negative. This is due to the from different cross-sectionalnatureof the data,i.e. to the adjacentmeasurements individual elephants, and to the unequal spacing between increments for each individual.The fluctuationsalso may be due to differentecological conditionsfor growthduringthe life of these individuals.It is possible thatnegative increments representthe differencein size between individualsin good ecological conditions and those who grew up under harsh conditions. Notice that the negative increments, and most of the fluctuations,are confined to the first 20 years of life. That time includes the infancy andjuvenile stages. The fluctuationsin these increment data, despite their limitations, do add some support, as an explanation for the juvenile stage of growth, to the ecological risk-aversion hypothesis. In other and a stratwords, the environmentin which these elephantslive is unpredictable, egy of slow growthto adulthoodmay well provide the best chance for avoidance of starvationduringbad years. The dominancehypothesis is anotherpossible explanationfor the existence of a juvenile stage for social mammals. Researchusing wild and captive primates, including elephants and social camivores (wolves, lions, hyenas), shows that high-ranking individuals in the social hierarchy can suppress and inhibit the reproductivematurationof low-rankingindividuals(Pereira& Fairbanks1993). The inhibitionmay be due to the stress of social intimidationacting directlyon the endocrinesystem, or it may be secondaryto inadequatenutritionbecause of competition for food. Juveniles are almost always low-rankingmembers of primate social systems. In the past, individualswith slow growthand delayed postinfancy reproductivematurationmay have survived to adulthoodmore often than individuals with rapid growth and maturation,and in this context, the juvenile stage may have evolved. Whateverthe roots of the juvenile growthstage, leaming, risk aversion, or social inhibition, one scholar points out that in a broadperspective, "...juvenilelife has two main functions:to get to the adultstage withoutdying and to become the best possible adult" (Alexander 1990). Adding a juvenile stage must have served these functions well for it to have evolved independentlyin so many social mammals.

PRIMATE GROWTHPATTERNS
Most primatesarehighly social mammalsand, as would be expected, have ajuvenile stage of growth and development. The primates, however, are a highly diverse groupthat includes the prosimians(such as lemurs, lorises, and tarsiers), New Worldmonkeys (rangingfrommarmosetsandtamarinsto cacajosandhowlers), Old World monkeys (such as baboons, rhesus, and colobus), apes (gibbons and siamangs, orangutans, gorillas, common chimpanzee, and bonobo), and humans.In all, there are more than 50 generaand more than200 species of living primates(Napier & Napier 1967). Primateslive in environmentsthatrange from

EVOLUTIONOF HUMAN GROWTH

123

0.3 -

Common Marmoset Callithrixjacchus 5

Red Baboon Papio papio . ,3

Colobus Monkey Colobus guereza ,

4-J

5~~~~~~~~~

2~~~~~~~~ e~~~
0 5 10

Ct~~
15

(
0 5 10

t
15

1
0

\
5 10 15

Age (years)

Age(years)

Age (years)

Male

--

Female ----

Figure 7 Three primate species showing different patterns of postnatal growth in weight. Abbreviations: I, infancy; J, juvenile; A, adult. Marmosets are weaned at 63 days and can breed at 1 year of age (Harvey et al 1987). Marmoset growth rate shows no postnatal growth spurt and no change in growth rate typical of a juvenile growth stage. Baboons are weaned by 18 months and begin puberty at about 3.5 years for females and 4.5 years for males (Harvey et al 1987). Baboons have a juvenile growth stage for both males and females, but only males have clear weight growth spurt. Colobus monkeys are weaned at about 13 months and females have their first birth at about 4.6 years of age. Colobus monkeys show a postnatal spurt for both sexes, but it is not clear if there is a juvenile growth stage. ( Velocity curves) Fit to cross-sectional data collected from captive animals using the statistical estimates based on lowess regression. See the text for further details. (From Leigh 1994b.) semiarid scrub, to savanna, to woodland, to rain forest, and from sea level to more than 2000 meters above sea level. Adult sizes range from that of male gorillas, 160 kg, to that of mouse lemurs (Microcebus murinus), 0.08 kg. Maximum life span ranges from 8.8 years for dwarf lemurs (Cheirogaleous major) to approximately 120 years for humans (Harvey et al 1987). Given this variation in local ecologies, world distribution, general environments, adult body size, and life span, it is reasonable to expect that primates have a variety of patterns of growth and development, and indeed they do. Three patterns of primate postnatal growth in weight (Figure 7) are based on the work of Leigh (1992, 1994a,b). There are differences in growth patterns between the three species (marmoset, baboon, colobus) and between males and females within species. Furthermore, none of these species follows a pattern of weight growth similar to that of the human species. The marmoset seems to follow the type of growth curve Brody described for most mammals, i.e. an acceleration phase followed immediately by a deceleration phase. The switch between phases seems to occur just before the time of birth, and the infancy stage merges seamlessly into the reproductively mature stage. Based on the velocity curves of weight growth, there is no evidence of a juvenile stage. The lack of a juvenile

124

BOGIN growth stage may relateto marmosetsocial behavior.Marmosets,and the closely related tamarins,live in a type of family group in which females form a social bond with one or a few males. These polyandrousfamilies establish feeding and breeding territoriesand exclude other marmosets. The young, usually born in pairs or triplets, are raised cooperatively by the mothers and her male consorts (young of the same littermay be sired by differentmales). After weaning, which takes place as early as 6 months for some species and by 12 months for others, the young are protected and provisioned with food by both adult males and the mother (Goldizen 1987) until they reach sexual maturity, approximately 15-24 months. Following the risk aversionhypothesis of Janson& van Schaik, a juvenile stage of growth would not be needed in this type of territorialfamily Leigh social group [see Garber& Leigh (1997) for an alternativeinterpretation]. (1996) analyzedgrowthdatafor five other species of marmosetsor tamarins,and all these follow a patternof growth essentially the same as that for the common marmoset. The baboon velocity curves show evidence for infant, juvenile, and adult stages of growth.Baboon infants are weaned within approximately18 months of birth, but growth rates remain fairly stable until approximately4 years of ageclearly a juvenile patternof growth. Baboons reach a peak in growth rate before birth,just as humansdo, but by the time of birth,baboons are alreadygrowing at the slower, steady rate that characterizeslate infancy for humans.Both male and female baboons follow this patternduringinfancy, but duringthe juvenile phase males show a pronouncedaccelerationin weight growth-a spurt.Baboons live in large social groups composed of males and females of all ages, and therefore their growth pattern conforms to the prediction that they will have a juvenile growth stage. Why males and not females have a juvenile growth spurtin weight is not known. It is known that the spurt is due mostly to an increase in muscle mass, which in males is used in competitionfor scarceresourcessuch as food and mates. Colobus monkeys show yet a thirdpatternof postnatalgrowth.Weight growth afterbirthhas a clear accelerationin both males and females . Whatis not clear is whetherthese accelerationscan be classified as juvenile growthspurts.It is possible that colobus monkeys grow more like marmosetsthan like baboons. Like the marmoset,the colobus may also have only a single peak rate of growth but one that is reached soon after birth. The colobus, then, would be following a growth patternlike thatfor mice and cows (Figures2 and 3). Colobus social behaviorand diet do not provide a clear case for the prediction of a juvenile stage. Colobus monkeys live in troopsof 3-15 animals,with a trooptypically composed of a single adult male, three to four adult females, and their offspring. These troops are highly social, there is little evidence for a female dominancehierarchyor female aggression,and infantsseem to be groomedand caredfor by troopmembersother than the mother. This shared infant care, as well as the intense allo-grooming between adult females, is believed to maintain a highly cohesive social group (Struhsaker& Leyland 1987). Colobus monkeys eat mostly young leaves from the hackberrytree (Celtis durandii) and two other tree species, which together

EVOLUTIONOF HUMAN GROWTH

125

comprise 69% of their total diet. Colobus, and relatedleaf-eating monkeys of the subfamily Colobinae, digest this large quantityof leaves in a specialized stomach composed of three or four subcompartments. Their digestive system is similarto the ruminantdigestive system of cows. Given the wide distributionand availability of theirfood, and given theircohesive andnoncompetitivesocial organization, it is possible thatyoung colobus monkeys have a low risk for mortality.If so, then the risk aversionhypothesis would predictthe absence of a juvenile growthstage, or the presence of a very brief stage. There is some empiricalsupportin favor of this prediction.In an analysis of 42 primatespecies, Leigh (1994a) founda significant correlationbetween diet andrate of growth.Folivores, such as colobus monkeys, have faster rates of growth than do species of non-folivore monkeys and apes, such as rhesus monkeys and gibbons. So, perhapsit is not so far-fetchedto statethatcolobus have both cow-like stomachsandcow-like patternsof growth.

Some Historical Background

A more detailed discussion of primatejuvenile growth, especially of postnatal growth spurts, is presented below. Before that review, some of the history and findings of research into primate growth are discussed so as to provide a better appreciation of what primates share with other mammals and what is special about primate growth. Comparativestudies of primate growth began with the work of Schultz (1924). Frominception,his studies were aimed at ".. .the relation of the growth of primates to man's evolution... (1924:163). Perhaps Schultz's most lastingcontributions were summarized in his 1960 illustration of the approximate ages of some life periods of the primates(Figure 8). Despite the grandeur and vision of Schultz's research on primate growth and primate evolution, few primate species were actually studied in detail by 1960. In fact, until the 1980s, details of skeletal, dental, and somatic growth were known from only three species, rhesus monkeys (Macaca mulatta), chimpanzees (Pan troglodytes), and humans. LairdcontinuedSchultz's interestin evolution of the humangrowthcurve. In a 1967 paper, she reviewed studies of the growth of the three well-studied species: rhesus monkeys, chimpanzees,and humans(Laird 1967). She took a mathematical approachto the study of growth. By fitting mathematical functions to the growth data, Lairdhoped to reveal more precisely the stages in the evolutionary development of the humangrowth curve. The curves were fitted to the monthly weight data points by the method of least squares regression. Essentially, this method minimizes the sum of the squareddeviations of each datapoint from the fitted curve, i.e. it produces an average line between the datapoints. Laird found that monthly weight increases in rhesus monkeys and chimpanzees followed two separategrowth curves. For male rhesus, the first curve fit the datafrombirthto 22 months,andthe second from 23 monthsonward.A change in growth rate that occurredbetween months 22 and 23 necessitated the use of different mathematicalcurves to model growth in the two periods. Sexual development in male rhesus takes place duringthe second growthphase, aftermonth 40. The deviationsof weight growthabove the fitted curvebetween months48 and 54

126

BOGIN

70-

60

= FrenatalFeriod =_____ Feriod= InFantile = JuvenilePeriod= MM\' Adult Period= FemaleReproductive Feriod=

111111

Gibbon Chimp. Man LemurMacaq?ue 50_

40-

30-_

WEK

24

334

20~~~~~~~1 -TU IILI____

IK51 I

2.

3 .___0

34

38XX~~~ The\\ roianthun.

sam _liesae,btte have theL

inces

in legt

fro

Figure 8 Schultz's diagram of the proportional increase in the length of life stages across the scala naturae of living primates. Schultz used eruption of the permanent teeth to mark the boundary between life periods. Note that Schultz did not recognize

the childhoodor adolescentstages for modernhumans.Indeed,all primatespecies to human.The have the same life stages,but they increasein lengthfromprosimian rather thantheoretiestimatesfortotallengthof life arebasedon averageexpectations cal maximums. (FromSchultz1969.)

EVOLUTIONOF HUMAN GROWTH

127

correspondedto the time of reproductivematurationand the beginning of adult levels of gonadal hormone secretions. The curves of growth for female rhesus were similar to those of male, except that the time of onset of sexual maturation was earlier,occurringat about42 months. For chimpanzees,the first curve fit the early phase of growth, from birth to 6 years. This early phase is essentially the infancy period, which lasts until 5 years of age in chimpanzees(Teleki et al 1976, Goodall 1983, Nishida et al 1990). The second curve was fit to what Lairdcalled the adolescent phase. Male and female chimpanzees followed the same curve of growthfrombirthto 6 years. Duringthe adolescentphase males grew in weight at a faster rate than did females, and this requiredseparatemathematicalfunctions for each sex. Because of the differentrates of weight growth, sexual dimorphism in weight became well marked.The sexual dimorphismreachedits greatestlevel at the age when male chimpanzeesbegin to sire offspring. Laird,in her analysis, confirmedthe work of Tanner(1962), who had shown thatthe sexual dimorphism in the weight of adult chimpanzeeswas largely due to a weight growth spurtfor the males (a more recent example is shown in Figure 11). Laird (1967) found that the velocity curve of human growth (Figure SB) requiredthree mathematicalfunctions to model its course. This conclusion was confirmed independently(Bock & Thissen 1976, Bogin 1980, Karlberg 1987). The need for the thirdfunctionis one aspect of humangrowththatmakes it different fromthatof otherprimates.Lairddescribedthe similaritiesof growthbetween rhesusmonkeys and chimpanzeesandthe distinctpatternof humangrowthas follows: "[T]he curvilinear growth by which the body weight of an organism approachesits mature value and during which sexual maturationcharacteristically occurs, startsat birth in sub-primatemammals and birds, but is deferredin monkeys ... the preliminary growth occupying about 1/3 and the adolescent growth about 2/3 of the time requiredto reach fully maturesize. In the chimpanzee, adolescentgrowthis deferredto the last 1/2 of the total period... In the human a furtherdelay has occurredso thatadolescentgrowthwith its concomitantdevelopment of sexual maturityoccupies only the last 1/3 of a prolonged growthperiod. The delay in the human can be interpretedas being due to the insertion, between birthand adolescence, of two growthphases, ratherthanthe single phase identifiable in the monkey and the chimpanzee...."(1967:351-52). It seems thatLairdconsideredall threeprimatesto be alike in having ajuvenile phase between infancy and what she refers to as adolescence. The second growth phase for humansbetween infancy and pubertyis childhood (Figure SB). I define childhood as a life history stage in the human life cycle that occurs between the end of infancy and the startof the juvenile growthperiod (from approximately3 to 7 years of age). Childrenareweaned frombreastfeeding (or bottle feeding) but must be provided especially preparedfoods and require intensive care by older individuals.In contrast,juvenile primatesprovide for their own feeding and care. Human childhood is characterizedby relatively rapidneurological development and slow physical growth and development. Based on the mathematicalanalysis by Laird,I conclude that in terms of anatomy, physiology, and behavior only humanshave a childhood life history stage.

128

BOGIN

100
?3

80 60
-

Brain

cD

0
C

4 20 0 40 80 120 160 200 240 280 320

Age (days)
Figure 9 Growthof differenttypes of tissue in rats.(FromTimiras1972.) The insertion of childhood into the human life cycle not only requires a third mathematicalfunction to model growth, but, more important,changes the biological and social ecology of the humanspecies. These biosocial changes are discussed in detail in otherpublications(Bogin 1996, 1997, 1999).

Of Brainsand Bodies
Another key difference between nonprimateand primate growth involves the relative rates of growthof the body, the brain,and the reproductivesystem. Most mammals, including rats (Figure 9) and the social mammals, show an advancement of brain growth relative to body growth. In rats, reproductivematuration occursbefore the brainor the body achieve final adultsize (Donaldson 1895). Primates delay body growth and reproductivedevelopmentbut do not delay brain growth. These relationshipsin humansare illustratedin Figure 10. The weight of the humanbrainreaches 80%of adultsize by age 4 and almost 100%of adultsize by age 7. Yet body growth continues until age 18 and beyond. Even more to the point, braingrowthis finished before reproductivematurityeven reaches 10%of the adult value. This patternof relative growth of the brain,body, and reproductive system is also found in rhesusmonkeys and chimpanzees(Laird1967). Other organs of primatebodies (e.g. heart, lungs, liver) follow the body growth curve. Primatebrain, then, is most unusual in its patternof acceleratedgrowth in comparison to other organs and to the body as a whole. In an obvious way, this pattern of growth relates the fact that primates are learning creaturespar excellence. There are other consequences of the primate

OFHUMAN EVOLUTION GROWTH

129

100

~3

80 60
605-0
40 -

##/#

Dentition
/ 100
'

:Body
~
-

0~~~~~~

Age (years)
curves fordifferent Forthebrain, thecurve is for Figure10 Growth bodytissues. totalweight et al 1993); fordentition, thecurve is themedian score (Cabana maturity forgirlsbased teeth onthesevenleftmandibular M1,M2)using 1,PM2, (I1,12,C,PM thereference dataof Demirjian forthebody,thecurve in (I1986); represents growth ortotalbodyweight; stature andforreproductive, thecurve theweight of represents

thegonads and reproductive (Scammon primary 1930). organs thatarenot as obvious. of growth It is worth thatrapid brain pattern noting andprogressively deferred sexualmaturation growth, bodygrowth, delayed enhance thequality andquantity of reproductive inprimates. greatly efficiency Notonlyis a givenoffspring withgreat endowed of learned behavior, flexibility butalsoas anolder thisindividual canhelpitsmother prereproductive juvenile, carefor new infants. In humans, is further provide reproductive efficiency enhanced with our childhood and adolescent ofgrowth. These periods advantages in selection resulted at thegenetic levelof primate to develop andto biology maintain these of growth. patterns IS THE HUMAN ADOLESCENT GROWTHSPURTUNIQUE? Oneof thehallmarks of human is theadolescent thepostgrowth growth spurt, in rateof growth acceleration in height andweight. individual juvenile Many andmost theadolescent organs 1966, bodydimensions experience (Shock spurt Cameron et al 1982,Satake et al 1993,Dasgupta & Das 1997,Greil1997). theadolescent is detectable inboth Moreover, spurt boysandgirls,andin every human so farexamined. Theadolescent population spurt SB)is a regular (Figure andnormal of human feature andsevere malnutriillness, growth. Onlychronic
tion, or physiological stress can obliterate the growth spurt. For instance,

Indian atthehighaltitudes Andes ofthePeruvian Quechua boysandgirlsliving havea lateand defined adolescent Thereason forthisis that poorly growth spurt.

130

BOGIN

these childrensuffer from the combinedstressof hypoxia (insufficientdeliveryof oxygen to the tissues of the body), energy malnutrition,heavy workloads, and cold temperatures.Instead of a clear spurt,the Quechuaexperience a prolonged adolescent growth period, lasting until age 22 and beyond (Frisancho 1977). There is some controversy as to whether the adolescent growth spurt is a uniquely humanfeature.On the basis of empiricalobservationsand evolutionary considerations,I conclude that the humanadolescent growthspurtin statureis a species-specific characteristic; thatis, a skeletal growthspurtof the humantype is not found in any otherprimatespecies (Bogin 1993, 1994a, 1999; Bogin & Smith 1996). This conclusion stands in contrast to previous research. As discussed above, Brody (1945) believed that the humanadolescent growth spurt,both for height and weight, was homologous with the peak velocity of growth shownjust afterbirthby most laboratoryand farmanimals.von Bertalanffy(1960) and Laird (1967) also believed this to be the case. Forthese researchers, the special features of primategrowth were the prolongationof thejuvenile period, found in all haplorhine primates, and the evolution of the childhood period, found only in humans.The additionof thejuvenile, and then the childhood,growthphases simply delayed the ubiquitouspostnatalgrowthspurtof all animalsto the adolescent phase of humans.As shown above, this interpretation is incorrect. PerhapsTanner(1962) was the first to reject the grandunificationhypothesis proposed by Brody. A pediatricianand a human biologist, Tannerwas able to understand thatprimatesfollow a patternof growththatis differentin many ways from thatof othermammals.Accordingto his analysis,one such differenceis that Old World monkeys, apes, and humans show two growth spurts:the universal spurtsharedby all mammalsthatoccursjust beforeor afterbirth,anda spurtat the time of puberty.Although Tanner'sanalysis sufficiently clarifies this point, confirmationand furtherclarificationof this unusualpatternhas since been provided (Laird 1967, Timiras& Valcana 1972). Tanneranalyzedgrowthcurves for male and female rhesusmonkeys and chimpanzees.Theirweight velocity curves show thatgrowthratespeak at orjust before birthand thendecline afterbirth.The period of decline lasts for more than a year for rhesusmonkeys and for more than 3 years for chimpanzees.Both species then show an accelerationin weight velocity that has all of the characteristicsof the humanadolescentspurtin weight, including the earlieronset of the spurtin females andthe greaterintensityof the spurtin males. Tannerpresents velocity curves for crown-rumplength for rhesus monkeys and concludes, again, thatboth males and females have an adolescent spurt. This skeletal spurt,however, is much smallerthanthe spurtin weight andof much shorterduration.The rhesus weight spurtspans almost 3 years in males, but the skeletal spurtlasts at most 6 months. Based on otherskeletal measurements, Tanner concludes thata rhesus growthspurtis "...largelyone of shoulderbreadthand, above all, of muscle mass" (1962:235).

A Surveyof Weight Growth in Primates

Based on the evidence available today, there is no question that some monkeys and apes have pubertalgrowth spurts in weight. To date, Leigh (1996) has pro-

EVOLUTION OFHUMAN GROWTH

131

vided the most comprehensive study directed at the question of the primate growth spurtsin weight. Some of Leigh's findings arepresented(Figure7). Leigh collected chronological age and body weight data for 2395 captive primates, housed at both zoos and primatelaboratories.There are 35 nonhumanspecies in his sample, including representativesof New World monkeys, Old World monkeys, and apes. He also included a data set for healthy humans(English boys and girls measured in the 1970s and 1980s). Although there are many longitudinal records for individual animals in the database,the data are analyzed in a crosssectional fashion. Leigh explains that cross-sectional analysis "...leads to artificially depressed estimates of growth spurt magnitudes...[and] should lead to a conservative diagnosis of the presence of growth spurts"(1996:457). He also used the nonparametricmathematicaltechnique called lowess regression to fit smoothed curves to the growth data for each primate species. There are many advantagesand disadvantagesto the use of mathematicalcurve-fitting.Interested readers should consult books on the theory and application of regression for details. Leigh's decision to use the mathematicalsmoothingtechniqueof lowess regression is useful, because it helps to reduce the unimportant variability(often called noise) in time-series datasets and helps to reveal the important main trends thatoccur over time. Lowess curves were fit to the distancedata,andthen the first derivative of the distance curve was estimatedto produce a velocity curve. This follows exactly the procedureused to constructFigure 1. A summary of Leigh's results for the 35 nonhuman species he analyzed is given in Table 1. Only one species of New Worldmonkey, Cebusapella, shows a spurt, and only for males. All studies of Old World monkeys show a spurt for males, and for many species they also show a spurtfor females. Among ape species, gibbons (Hylobates spp.) show no spurt,orangutans(Pongo) show a possible male spurt,common chimpanzees (Pan troglodytes) shows a male spurtbut no female spurt, and both bonobos (Pan paniscus) and gorillas (Gorilla) have male and female weight spurts.The variationin presence or absence of a postnatal weight spurt among primate species, and between sexes within a species, is in keeping with the diversity of primate biology and ecology. Leigh's data show clearly that a postnatalweight spurtis not a primatecharacteristic; rather,it is a variable trait found in one, or both, sexes of some species. The velocity curves for the apes are illustratedin Figure 11. The rateof weight growthof the gibbon generallydeceleratesfrombirthto sexual maturityin a manner similar to marmosets (Figure 7). Gibbons are also socially somewhat like marmosets in that both species live in family groups. But unlike the marmosets, gibbons have a juvenile stage of growth. Gibbon infants are weaned at approximately 3 years of age, andthe age when females first give birthis approximately9 years, which means that gibbons spend about 6 years as juveniles. The gibbon data show that postnatal growth spurts, especially spurts around the time of puberty,andthejuvenile growthstage arenot causally associatedwith each other. Great apes, including orangutans,gorillas, bonobos, and chimpanzees, all have infant, juvenile, and adult stages of growth but show varying patterns of stage durationandvaryingpatternsof growthspurts.Weaningtakes place by 4 years of

132

BOGIN TABLE 1 Summaryresultsof Leigh's (1996) analysis of weight growthin primatea Species Ceboidea (New World monkeys) Cebuellapygmaea Callithrixjacchus Cailirnico goeldi Saguinusfuscicollis Saguinus geoffroyi Saguinus imperator Saguinus oedipus Leontopithecusrosalia rosalia Saimiri sciureus Cebus apella Callicebus moloch Aotus trivirgatus Sample size (Male/Female) Growth spurt (by sex)

36/51 48/71 42/27 18/19 9/10 11/14 46/18 26/31 32/28 26/28 30/23 25/23

None None None None None None None None None Male None None

Cercopithecoidea(Old World monkeys) 30/30 Cercopithecusaethiops 27/37 Cercopithecusmitis Cercopithecusneglectus Erythrocebuspatas Cercocebus atys Macaca arctoides 29/23 41/52 38/71 52/58

Both Male Male Both Male Male (continuedon nextpage)

age for gorillas, by 5 years for chimpanzeesand bonobos, and by approximately age 6 for orangutans.Adult females first give birthat age 10 for gorillas, age 12 for orangutans,and between 13 to 15 years for chimpanzees and bonobos. Each species, then, varies in amountsof time spent in the juvenile stage. Even though gorillas have the shortest infancy and juvenile stages, they achieve the largest body weights of all the apes. This is because, at any age, gorillas are always growing faster than any of the other great ape species (Figure 11).

SkeletalGrowthin Primates
In contrastto Leigh's comprehensivesurvey of primateweight growth, there are only a few studies of skeletal growth. Coelho (1985) measuredgains in crownrump length (CRL) and weight in a mixed-longitudinalsample of 250 male and 452 female olive baboons(Papio cynocephalus anubis). The animalswere partof a laboratorycolony living undernature-likeconditions in terms of physical envi-

EVOLUTIONOF HUMAN GROWTH

133

TABLE 1 (continued) Summaryresults of Leigh's (1996) analysis of weight growth in primatesa Species Macacafascicularis Macacafuscata Macaca mulatta Macaca nemestrina Macaca silenus Papio hamadryas Mandrillussphinx Colobus guereza Presbyns entellus Presbutis obscura Hominoidea (apes and humans) Hylobates lar Hylobates syndactylus Pongo pygmaeus Gorilla gorilla Pan paniscus Pan troglodytes Homo sapiens Sample size (Male/Female) 13/13 64/71 52/58 39/64 39/41 33/53 49/59 46/49 29/24 19/17 Growth spurt (by sex) Both Both Both Both Male Male Both Both Both Male

25/25 19/21 42/42 77/64 13/23 22/23 LiteratureData

None None Male' Both Both Male Both

aSpecies evaluated, sample sizes, and presence or absence of a postnatal growth spurtby sex. bPresence of spurtis uncertain.

ronmentand social group composition. All animals were healthy and well nourished, and none showed signs of clinical obesity. The date of birth for all individualswas known and the animals were measuredonce a year. The mixedlongitudinal design of the study provided data on growth between birth and 8 years of age, which for this species is the total span of the growingyears. Velocity curves for CRL arepresentedin Figure 12 (weight velocity is essentially identical to the data in Figure 7). The crown-rumpvelocity data are presentedin a crosssectional fashion, which, again, would tend to reduce the apparentsize of any growth spurt.Even so, the crown-rumpvelocity curves stand in sharpcontrastto the weight velocity curves that show a large spurtfor males, for neithermale nor female baboons show a spurtin CRL. Baboons do show threephases of declining growth rate in CRL: an infancy phase until 2 years, a juvenile phase from 2 to 6 years, and a postpubertal phase until 8 years. Similarpatternsof decreasingvelocity in skeletal growth from birthto maturitywere found in studies of three other monkey species, Macaca nemestrina (Orlosky 1982), Papio cynocephalus (Siri-

134

BOGIN

Hylobateslar
2.0
26e0

H.syndactylus
16

Pongo pygmaeus

co

0.)
1.0

126

~~~~~~~~~~~~~~~~~~~~~~~~10

0.6

*2X

0.0 0 10 16

-100 0 4 a 12 16

0 0 6 10 16 20 26

20

Gorilla gorilla 16

Pan paniscus
*5

P. troglodytes

0)

teeis

4vdnefrtepeec

fhmnlk

dlsetgot

prsi

Fig

o ay of t es Veletocit

species of

Another peys.

1996ni

tudinal study of growth in weight and CRL with baboons suggests a pubertal spurt in skeletal growth. The authors of this study report ";...the presence of a pubertal growth spurt in body weight, and possibly CRL, in male but not in female baboons" (Crawford et al 1997). The male CRL spurt amounts to about a 2-cm increase in growth rate from prepubertal values. But the analysis of the spurt is based on a varying number of individual baboons, with as few as two and as manly as eight individuals measured on different occasions. It is difficult, if not impossible, to derive an accurate and reliable estimate of growth velocity from data of this type. Cross-sectional studies of both weight and skeletal growth (e.g. limb lengths, trunk length, etc) in Japanese macaques (Macaca fuscata) and chimpanzees (Pan troglodytes) have been published by Hamada (1994) and Hamada et al (1996). The macaque group consisted of both laboratory colonies and seminatural colonies reared in monkey parks in Japan. The data are cross sectional, representing 2886 animals (1288 males, 1598 females) between the age of birth to 15 years. The chimpanzee data, which are also cross sectional, came from animals reared at a primate park owned by a Japanese corporation. There were a total of 172 chim-

EVOLUTION OFHUMAN GROWTH

-

135

Males

-s-

Females

is

-2
o
1

Age (years)
Figure 12 Velocity curves for crown-rump length (CRL) in the baboon. (From Coelho 1985.) panzees (81 male, 91 female) between the ages of birthand 15 years. Hamadaand colleagues analyzed the data by fitting spline curves to the median values of growth for each age class, e.g. 3 months, 6 months, 12 months, 2 years, etc. The first derivativeof these distancecurves was used to producea velocity curve. The spline curve method is similar to the lowess regressionmethod used by Leigh in thatboth aredesigned to fit a smooth, continuouscurveto a set of datapoints. One difference between the methods is that spline curves are forced to pass through each datapoint, whereas lowess curves may pass between data points. Hamadafound that the postnatalgrowth of the Japanesemacaque,in termns of both distance and velocity, had, in general, three distinct phases: ". . .[1] from birth to two years when the growth velocity decreases rapidly; [2] from two to three and a half (in females) or four and a half (in males) years of age when the velocity stays constantor growthaccelerates;[3] fromthose ages to eight years of age when velocity decreases again to zero)". This overall patternis a textbook case of an animal with infant,juvenile, and adultgrowth stages. Note that during the juvenile stage the growth rate stays constantor accelerates(Hamada'sphase 2). Weight growth for Hamada'smale and female monkeys provides an example of bothjuvenile patterns.Male monkeys show a clear spurtin weight, but female monkeys show no spurt(similar to the baboons in Figure 7). The analysis of the macaques also includes the calculationof distance and velocity curves of growth for 15 different body segments or cranial dimensions, such as trunklength, tail length, upper arm and forearm length, thigh length, and head breadth.Some of these dimensions show what might be a growthspurt,for example in handand leg length, but other dimensions, such as forearmlength and head breadth,do not. Generally,male monkeys show spurtsmore often in these 15 dimensionsthan do female monkeys.

136

BOGIN

Chimpanzeedataanalyzedby Hamadaet al (1996) includedmeasurementsof body weight andtrunklength(measuredfromthe cranialend of the sternumto the cranialend of the pubic symphysis). Velocity curves for these two measurements show that the chimpanzees have three stages of growth: infancy, juvenile, and adult. During the juvenile stage male chimpanzees experience a clear spurt in weight but not in trunk length. Female chimpanzees do not show a postnatal growth spurtin eitherweight or length. In anotherstudy, Hamadaet al (1998) analyzed the patternof bone maturation for a group of 65 male and female chimpanzeesfrom the same primatepark colony used in the somatic growth study. The authors employed the TannerWhitehouse2 method(Tanneret al 1975) to assess the stages of bone maturation. Fromthese bone stages, Hamadaet al computedRUS scores for the 13 epiphyses of the radius (R), ulna (U), and short bones (S) (metacarpalsand the first, third, and fifth phalanges). The RUS scores are calculated as points, with zero points meaning thatno epiphyses are presentand 1000 points indicatingfull maturityof 2 andRUS methodsarewidely used to the 13 epiphyses. The Tanner-Whitehouse in humanboys and girls. The findings of Hamadaet al for assess bone maturation chimpanzees are shown in Figure 13 as incrementsin RUS points between successive measurements.The chimpanzeedataarecomparedwith similardatafrom humanboys and girls. The humanRUS point incrementsare taken from the 50th percentile curve publishedby Tanneret al (1975); in otherwords, from the sample used to develop the RUS scoring system. Chimpanzeesand humanshave distinctly differentpattems of skeletal development.Boys and girls show the typical humancurve of growth,with a deceleration in bone maturationduring infancy, a steady rate of bone maturationduring the childhood and juvenile stages, and an adolescent spurt in bone maturation. Note thatthe humanadolescent spurthas but a single peak. Both male and female chimpanzees also show a decelerationduringtheir infancy stage but, in contrast to humans, show an immediate accelerationin the rate of bone maturation,then anotherdeceleration,and then a secondaryacceleration.Hamadaet al call these last two phase of bone maturationthe midadolescenttroughand the postadolescent peak. I would substitutethe word juvenile for adolescent, but whatever one they have no parallelin the human calls these changes in rate of bone maturation, condition. The studiesby Leigh andby Hamadaand colleagues arevaluablebecause they include large samples of nonhuman primates and many species of primates. Taken together, these studies show that some primate species have postnatal growth spurts in weight, linear dimensions, and bone maturationthat occur aroundthe time of sexual maturation.About half of the species studied show a weight spurtfor males only, andwhen a female weight spurtis presentit is always smallerthanthe male spurt.These accelerationsin weight growtharethe resultof an increase in size of several tissues, particularlyin muscle mass and body fat. Behavioralstudies of free-ranginganimals show that adultmales use this muscle mass in reproductivecompetition with each other. Female primates may also choose to mate with largermales, who may be healthier and socially more suc-

EVOLUTIONOF HUMAN GROWTH

137

cessful than smallermales. The absence or reduced size of a female growth spurt allows females to maintain smaller adult body size comparedwith males. This may be advantageous during pregnancy and lactation because smaller bodies requireless energy and other nutrientsand thereforewould permitthe motherto divertmore of the food she eats to her developing offspring.Withinthis context of reproductivestrategies,one may begin to understand how evolutionarypressures and naturalselection have shapedthe type and intensityof growthspurtsthatnonhuman and human primates experience. Human boys and girls experience a growth spurtin weight and in height duringadolescence, and the male spurtsare more intense than the female spurts. Discussions of the types of evolutionary pressuresthat may have shapedhumangrowth at adolescence can be found elsewhere (Bogin 1993, 1994a, 1999). Here, attentionis focused on a more detailed considerationof skeletal growth spurts,for it is in the growth of the skeleton that we see the clearest differencebetween the patternsof growthfor humanand nonhumanprimates.

SKELETAL GROWTHSPURTSIN PRIMATES

A serious limitationof the otherwise excellent studiesby Leigh andby Hamadaet al is that the growth velocity curves were computed from cross-sectional data, which tends to minimize the size of any actual growth spurt.But cross-sectional dataanalysis may also producea growthspurtwhen none is really present.Differences in the care, feeding, or health of animals from which informationis gath16 . . . . . . .

..... .............. ....... ....... .....I ...... E o1 2 ........ a . ......................... male male H m .......................... P an ..... omo........... / H . 12- ............. .
c 10 o
.........

female. .

female , / .

.......................i...............

......

......

... ...........

.....
.. .. ......

_2

.... ............................ ........ . . . .. . . . . . .

m -2

50

100

150

200

Age in months scores for chimpanzees Figure 13 Rate of developmentin RUS bone maturation and humans.Rateof bone developmentis presentedas increments in the numberof RUS (radius, ulna,andshortbones)pointsbetweensuccessivemeasurements. [ChimpanzeedatafromHamadaet al (1998). HumandatafromTanneret al (1975).]

138

BOGIN

ered for adjacent data points may result in discordantpatterns of growth that produce spuriousgrowth spurts.Only a pure longitudinalanalysis can faithfully representthe growthof a primate,or groupof primates,over time. One of the first longitudinal analyses of nonhuman primate growth was published by Gavan (1953), who analyzed the skeletal growth of nine male and seven female chimbeginning in panzees. The datawere collected at the Yerkes PrimateLaboratories 1939. The chimpanzees were measuredonce a year from birth to age 12. Influmethodsand devices (i.e. enced by Brody's speculations,andusing mathematical addingmachines) availableat thattime, Gavanconcludedthat chimpanzeeshave a human-like skeletal growth spurt at the time of puberty. In 1971, however, Gavan reanalyzed his chimpanzee data based on newer mathematicalmethods and the use of a computer and found that no spurt in linear growth could be detected."Afterall, a smoothly deceleratingcurve gave the best fit to most of my datawith a very small residualvariance"(Gavan 1982:3). Watts & Gavan (1982) reanalyzedthe same chimpanzeedataand new growth data for the rhesus monkey. The rhesus data were collected by Gavan between 1960 and 1967 from a laboratorycolony. Longitudinalassessments of growth were made at least once a year frombirthto 5 years of age. They foundthatsimple plots of height or weight for age did not reveal a growthspurtin eitherthe rhesus monkey or the chimpanzee.This standsin contrastto the humancase where simple graphicalmethodsof analysis reveal the adolescentgrowthspurtin most individuals. To searchfor nonhumanprimategrowth spurts,the longitudinaldatafor each chimpanzeeor rhesus was fit with an exponentialregressionformula."This model depicts growthas a process thatis graduallyand constantlydeceleratingas size increases. Its use, therefore,assumes that an adolescent spurtdoes not exist" (Watts & Gavan 1982:56). This is the same model Gavan (1971) used when he concludedthatthe chimpanzeehad no growthspurtin trunklength or in otherlinear dimensions. Watts & Gavan (1982) found that deviations in skeletal growth from the regressionmodel followed a consistentpatternin all 16 chimpanzeesand in all the rhesus monkeys. An example is depicted in Figure 14 for growth of the thigh of one chimpanzee.Deviations arepositive, above the curve, in early infancy, negative in later infancy and just prior to puberty, and positive again following puberty. The authorsnoted that the differences between the observed and predicted values at each age are small, less than a centimeter,and often only a few millimeters. It was impossible for them to depict these differences graphicallyas to exaggeratethe differencesfor drawingthis figtwo distinctcurves. "Therefore, were by a factorof three..."(Figure 14; Watts the actual deviations multiplied ure & Gavan 1982:58). Watts& Gavanemphasizedtheirconsistentfindingof growth deviations from the predicted curve as evidence for the presence of adolescent growth spurtsin nonhumanprimates.No statisticaltest of the mathematicalsignificance of the deviations was made. Watts does make the following qualitative assessment:"...the magnitudeof the change is very small" (1985b:56). Tanner et al (1990) published a longitudinal analysis of skeletal growth in female rhesusmonkeys (Macaca mulatta)and found evidence for "clearpubertal

OFHUMAN EVOLUTION GROWTH

139

growthspurts"(1990: 101) for tibia length, CRL, andweight. Theirdataconsist of measurementstaken every threemonths on nine monkeys housed indoorsand six monkeys housed outdoors at the Yerkes Regional Primate Center in Atlanta, Georgia. The original figures from the article are reproducedhere (Figure 15) because they are essential to the analysis by Tanner et al. The velocity curves shown in Figure 15 are mean values that were calculatedby combining the data from all nine indoor-housedmonkeys. This was done by aligning their curves on their age at maximum velocity to produce peak length velocity curves. This method is identical to the procedureinvented by Boas in 1930 to study human adolescent curves and is "...an exact copy of Shuttleworth's (1937) famous analysis of human growth and produces essentially a mean-constant velocity curve"(Tanneret al 1990:102). According to Tanneret al, these figures of rhesusmonkey growth". . .establish the occurrenceof a pubertalgrowth spurtbeyond any reasonabledoubt, at least for this species" (1990:101). The mean curve for tibia length certainly seems to supportthis, as there is a single peak velocity of about 2.25 cm/3 month (on an annualbasis this peak velocity equals about 0.6 cm/year). The mean curves for CRL and weight are less easily interpreted.Both CRL and weight show two or three spurts,albeit the one thatprecedes menarcheis the largest. Tannerand colleagues offer only these figures as evidence for the primacyof the premenarchial

300

Observed

E
E

,24~~~~~~Predicted

21200

1/
/

100
, , , , , , , , l I I L I

6 8 10 Age (years)

12

Figure 14 Observedandpredictedgrowthcurvesfor thigh lengthin a male chimpanzee.(FromWatts& Gavan1982.)

140

BOGIN
0

E
E
0)

-6

.3

1IS

15

*6

12

15

Sm*z
E
CZ)
0.6

0)0

peak velocity
Figure 15 Mean-constant velocity curves for growth of female rhesus monkeys. Tibia length, crown-rump length (CRL), and weight. (From Tanner et al 1990.)

EVOLUTION OFHUMAN GROWTH

141

spurt.They do not carryout any statisticalanalysis to demonstratethatthis spurt representsa significant increasein growthratecomparedwith any otherperiod of growth. Nor do they analyze the durationof the spurt,which is importantif we want to understandthe contributionof this spurtto the total patternof growth for these monkeys. Another complication arises because the patternof growth and growthspurtsseen for the nine monkeys housed indoorsdiffers significantlyfrom that of the six outdoor-housedmonkeys. Both the outdoor- and indoor-housed groups show a tibia-length spurtthat occursjust before menarche.The outdoorhoused females show either no spurtfor CRL, or a small postmenarchialspurt, and two spurtsfor weight (Wilson et al 1988). Theirfirst weight spurtoccurs before menarche and the second well after menarche,unlike the situation for the indoor-housedmonkeys. Tanneret al provide illustrationsof the velocity curves for the tibia and CRL for "...three individual indoor-housedmonkeys...selected at random from the total group" (1990:103). That illustrationis reproducedhere as Figure 16. The data show much variability,with skeletal growth rates oscillating between 0.00 cm/3 months and more than 2.5 cm/3 months for adjacentmeasurements.Maximum individualpeak velocities occur both before and after menarche(CRL for monkey no. 1 and 4, respectively), and smallerspurtsoccur both before and after first ovulation (CRL for monkeys no. 4 and 7; note also the variationin spurtsfor tibia length for all three monkeys).

The Human Adolescent GrowthSpurt

Longitudinaldatafor primateskeletal growtharerare,and it is even less common to see these dataanalyzedin a properlongitudinalfashion. The 1990 work of Tanner et al is valuable because it presents and analyzes such longitudinaldata. The quality of their data is unquestionable,and their analysis meticulous. Nevertheless, humanadolescents do not follow the patternsof growth seen in rhesusmonkeys. Figure 17 is a mean-constantcurve of humanvelocity growth in height and weight produced from the longitudinal growth records of 12 Guatemalangirls. The data are derived from a study of children,juveniles, and adolescents at a private school in GuatemalaCity (Bogin et al 1992). The girls are from families of high socioeconomic status,and all were healthyandwell nourishedat the times of measurement-once per year. The same methods were used to constructFigure 17 as were used by Tanneret al (1990) to constructthe mean-constantvelocity curves for the nine rhesus monkey females of Figure 16. One indicatorof human pubertyis the decreasingto increasingchange of growth velocity that marksthe transitionfrom thejuvenile stage to adolescent stage. Thatpoint is labeled on Figure 16. The mean age at peak height velocity and peak weight velocity for these 12 girls is 12.5 ? 1.07 years, and the standarderrorof the mean is 0.31 year.

Human-Rhesus Difference in GrowthVelocity

Thereare at least five differencesbetween humansandrhesusmonkeys in the pattern of velocity growth: (a) Humanshave one relatively large and unambiguous growth spurtfor both height and weight duringadolescence, (b) the magnitudeof

<

E~~~~~~~~~~~o X

>

>

>n~~~~~~~~~~~~~~~~~~~~~~l

t
_

>
-

n~~~~~~~~~~~~~~~~~~~~~~~ ?
i~~~~~~~

<,, z

<t.'0

<

,c~~~~~~~~~~~~~~~CS X

Z'

<

it

>

<

S;~~~~~~~~~~~

j to

.4.'

.'

<

c;>~~~~~~~~~~~~~~C1 e

@ ; < L
t0 -

<
FO
(owglluo)

t~~~~~~~~~~~~~~~~~~~~~~~ E

Q> t =

E~~~

lZ

-t'-

, > > _ O >O

g~~~~~~~~~~~~~

^uo)

3~~~~~~~~o
=ce
v

!3?ta,&

Blq^L M)9U8tI

^!3?l@^ tlf}uaq

dwnl-u*solD

GROWTH EVOLUTION OFHUMAN

Meanconstant heightvelocity for12 Guatemalan girs 9
8 .. ...
. ..

143

.
-5-

.... . ... ... . . .... .. ...,. -4 -3

....

... . ... . . ... ... . ... . . . . . . . . . . . .

4 ......-6

1 PHV+ ......X

+2

+:+

5+

Years beforeand afterPHV

C 3 ./.........
0

weight velocity Mean-constant ... ... . ... . ... . ... . . ....... ........ for girls \i.... .. 12 Guatemalan - ....... 9 -----,...----_-r- ... . . . . . . -----. . . . . . . . . . .. .. .. . . . . . ... 2.

... .. .

. . . . ..... .. . . ...

8l
-6 -5 -4 -3 -2 -1 PHV +1 +2 +3 +4 +S +6

Years before and after PW\t

Figure 7 17 ~~~Mean-constanthegtvlcyan
*

weight velocity.......... cre

2Gae

.... ..... velocities... .... pae with..... that..... nonhuman.. of...... the..... primates, ( ....)...human growth do.not osiltfrmna-aiu veoiytnermnmmvlctbeweada measurements cent ...d)... the.....human..... adolescent... spurt..lasts.............. for.. years.an the..... rhesus.....

pare wit tha of he nn

Years befriaes(n) after PW whVeoite

curveesnor1 Gateawea-ightu velocity 17o Mean-cosanthight velocity and Fsigurte

humanadolescent spurt ithraeosklalagroth cetmaueet,()the

absluel lare com-u ioyas

144

BOGIN

spurtslast for only a few months, which is a relatively shortamountof time even when consideringthatthe life span of a monkey is only one thirdthat of a human being, and (e) thatthe humanadolescentgrowthspurtis found in all humanpopulations, including those housed indoors and those housed outdoors (that is, in to the ruralhuntingand gathering societies rangingfrom the urban/industrialized groups). Exceptions to each of these six points are reflected in the rhesus data. Because the Guatemalangirls were measuredonly once per year, it was not possible, with that data set, to show that humansmaintaina fairly smooth trajectory of adolescent growth even over shortertime periods. At monthly intervals there is a good deal of variationin statureincrementsfor boys and girls (Bogin 1977, 1978; Togo & Togo 1982). Even so, duringthe durationof the adolescent spurtin stature,humangrowthincrementsnever approachzero, as they do for the most of this human rhesusmonkey. At 6-monthintervalsbetween measurements, a fair comparison would be A interval humans for variabilitydisappears. 6-month with the rhesus monkey 3-month interval, as humans take more than twice the time to grow up. One famous example of humangrowthmeasuredevery 6 months was provided by the growth data gatheredby the Count Philibert Gueneau du Montbeillardof France.He measuredthe statureof his son every 6 months from the boy's birth in 1759 to his 18th birthday. George-Louie Leclerc de Buffon includedthe measurements,and his commentaryon them, in a supplementto his Histoire Naturelle published in 1777. These data are usually considered as the first longitudinalstudy of humangrowth, and because of Buffon's commentary, the most famous. Togo (1995) analyzed this Frenchboy's growth based on the full data set gathered at 6-month intervals between measurementsand showed that duringhis adolescent growth spurtin height, the boy increased smoothly to peak height velocity and then decreasedsmoothly for the next year. Togo did not find any of the oscillations that typify rhesus monkey growth after puberty.

PATTERNS OF GROWTHFORHUMANSVERSUSOTHER PRIMATES

There is no doubt that some nonhumanprimates have increases in the rate of skeletal growth for some body dimensions at the time of puberty.Considerable evidence shows that these small increases in skeletal growth representreal and in growth around importantbiological processes. Most likely, the perturbations the time of pubertyare due to the onset of adultlevels of gonadalhormonesecretions. It has been well established that testosterone in male primates and estrogens, such as estrodiol, in female primatescan increasegrowthrates (Martinet al 1977, Bercu et al 1983, Prader 1984). But there is more involved: Roughly equivalent amountsof hormoneproductionin both chimpanzees and in humans result in strikinglydifferentrates of growth. In male chimpanzees,the concentration of testosteronein blood serumpriorto puberty(from 1 to 6 years of age) averages 13 ng per dl (Martinet al 1977). For human males, the prepubertalserum testosterone concentration(from ages 1 to 12 years) averages 9 ng/dl (Winter 1978). Thepeak velocity in long bone growthof the eight male chimpanzeesstud-

EVOLUTION OFHUMAN GROWTH

145

ied by Watts & Gavan (1982) occurredat a mean age of 10.96 years, with a standarddeviation of 1.31 years. At this age, serum testosteroneaverages about 400 ng/dl (Martin et al 1977). Peak height velocity in human boys from western Europe occurs at a mean age of 14.06 years, with a standarddeviation of 0.92 years (Marshall& Tanner 1986) when serum testosterone levels average about 340 ng/dl (Winter 1978). Based on these data, the serumtestosteroneconcentration of chimpanzees increases about 31-fold from the prepubertalto pubertal state. In humanmales, serumtestosteroneconcentrationincreases about 38-fold, or 1.23 times the increase for chimpanzees. Clearly, both chimpanzee males and humanboys have large increases in testosterone productionafter puberty,but the effects on skeletal growth are not so similar.Accordingto Watts& Gavan(1982), chimpanzeeshave a relatively small increase in the velocity of growthof individuallong bones duringpuberty"...usually less thana centimeter"(1982:58). Similarfindings are reportedby Copeland et al (1985), who analyzed growth in weight and CRL in 86 male and female chimpanzees. These researchersfound that in males therewas a markedincrease in serum testosteronebetween the ages of 6 and 8 years. This was followed by a slight increase in the rate of weight gain but no detectablespurtin CRL. Females also showed a rise in serumtestosteronebetween the ages of 6 and 8 years andrise in estradiolafter 10 years of age. In contrastto males, female chimpanzeeshad a much smaller increase in the rate of weight gain in relationto testosteronelevels. Females had no detectablespurtin CRL at any age. Hamadaet al (1996) confirm these findings when they reportno increase in the rate of skeletal growth at the time of pubertyfor their samples of male or female chimpanzees. In sharpcontrastto the chimpanzeeresearchare the findings for humanboys and girls that show relatively large and easily detectable growth spurts for both height and weight duringadolescence. Because chimpanzeesare not bipedal, it is not possible to take a measurementthatis exactly equivalentto humanheight. It is thereforemore appropriate to comparechimpanzeeandhumangrowthin termsof body segments. Cameron et al (1982) performeda longitudinal analysis of the growth of individual limb segments in British boys. They found that the peak value in velocity duringthe adolescent growth spurtrangedbetween 1.34 cm per year for the forearmand 2.44 cm per year for the tibia. Satake et al (1993) report that duringhumanadolescence, peak velocity for sitting height (roughly equivalent to CRL) equals 7.5 cm per year for Japaneseboys and 6.2 cm per year for Japanese girls. Based on these findings, one may propose that there are differences in the effect of testosteroneand estrogens on the skeletal growth of chimpanzees and humanbeings. The growthresponse of the humanskeleton to rising testosterone levels is greaterthan the response of the chimpanzee skeleton. The change in the serumhormonelevels for males of the two species differs by a factor of about 1.23, but the change in the velocity of growth of various body segments is two to four times greaterfor the human arm or leg, and more than 10 times greaterfor human sitting height. The velocity of growth in weight of male chimpanzees (and, presumably,of male baboons and some female primates) is better correlated with endocrine

146

BOGIN

changes at puberty(Figures7, 11). This observationsuggests thatthereare differences in the growth response of skeletal and nonskeletal tissue, such as muscle and adipose tissue. Thereare also differences in the effect of hormoneson tissue growthbetween males and females within the same species. Female chimpanzees experience a much smallerweight spurtthanmales, and female baboons show no weight spurtat all. It is importantto contrastthis to the human case; most girls experience significant growth spurtsin both height and weight. Thus, the growth effects of pubertalendocrinechanges differ between individualsof the same sex but of different species (e.g. male chimpanzees and humanboys), and between the sexes within the same species (e.g. male and female chimpanzees). Three major differences between humanand nonhumanprimategrowth may be highlightedat this point. These are (a) the residualgrowthpotentialof the nonhuman versus the humanprimateat adolescence, (b) the sensitivity of different body tissues to growth-promotingstimuli, and (c) sex differences in the expression of growthspurtsat adolescence. Monkeys, apes, andhumansall experiencea prolongationin the time for growthand a delay in the onset of sexual maturation. As noted by Brody (1945) more than 50 years ago, the delay is relatively and absolutely greaterin humansthan in nonhumanprimates,and in additionto this, humans also have a markedly increased potential over apes and monkeys for growth in height and weight during adolescence. This growth potential is more likely to be regulatedby the sensitivity of neuroendocrinereceptorsand postreceptors (i.e. biological tissues) to growth stimuli than by the rate or amount of productionof the stimuli (hormones)themselves. The lack of linear associations between testosterone concentrationsand growth velocities in skeletal and nonskeletal tissue of chimpanzeesand humansshow this. The differences in cellular sensitivity to growth stimuli between nonhumanand humanprimategrowth are probablycontrolledat the genetic level. Although it is possible that the different patternsof growthin the variousprimatespecies are the resultof the evolution of new structural genes (genes that code for specific proteins), it is more likely that the variationin growthcontrollies in the regulatorygenes thatinitiate,terminate, and control the duration of each of the distinct periods of growth (Britten & Davidson 1969, King & Wilson 1975, Bogin 1999:Ch. 7).

A PHILOSOPHY OF HUMAN GROWTH

Nonhumanprimatemodels are used in studies of human growth because of the similarities in anatomy and physiology between the species. When Schultz startedhis studies, he explicitly assumed that there was an evolutionarycontinuum between the living primate species. His illustrationof primate life history (Figure 8) shows this assumed continuumvery nicely. There is an evolutionary connection relating all primate species, but living monkeys, apes, and humans have separateevolutionaryhistories.A separationbetween Cercopithecoids(Old World monkeys) and hominoids (apes) occurredsome 20 million years ago and the hominoid-hominid (ape-human) split occurred approximately 5-6 million years ago. Thereis no evolutionaryreasonto expect thatthe patternsof growthfor

EVOLUTION OFHUMAN GROWTH

147

these threedivergentand ecologically distinct species should be identicalor even similar. As Gavan (1971:54) observed, chimpanzee postnatal growth begins "...withan initially high ratewhich deceleratessmoothly as size increases,but it is well known thathumangrowthis characterized by a growthspurt....Some change must have occurred in human growth since we and the chimpanzee have had a common ancestor." The notion of an evolutionarycontinuumis one legacy of the philosophy of a GreatChainof Being (Lovejoy 1936). This is a popularculturalconstructin westem society thathas historicalroots going back at least to ancientGreekwriters.In biology, the GreatChaintakes the form of the scala naturae. In its originalusage it erroneously implies that all living creatures,from the amoebae to the human, form a living evolutionarysequence from the simplest to the most complex creature.We now understand thathumansarenot the culminationor the goal of evolutionary history. We are just one of more than 2 million species of animal alive today, each the end productof its own history and with its own unique place in nature.Yet, the scala naturaeis sometimes misappliedto the connectionbetween humanand nonhumanprimates. Some observerstend to see monkeys and apes more as models for humanbiology and behavior than as creaturesin their own right. This point was cogently arguedby Scott (1967:72): "Subhuman primatesarenot small humanbeings with fur coats and (sometimes) tails. Ratherthey are a group which has diversified in many ways, so that they are as different from each other [and humans] as are bears, dogs and raccoons in the order Carnivora."Although Scott referredspecifically to psychological attributesof species within the orders Carnivora and Primates, his cautionaryremarksapply equally to morphology,physiology, and, in the presentcontext, patternsof growth.As mentionedabove, within the Carnivora, certainsocial species (dogs, wolves, lions) experiencea prolongedperiodof relatively slow growth between infancy and adulthood that correspondsto the juvenile growth phase. Nonsocial species of carnivoresmaturefrom infancy to adulthoodwithout a juvenile stage of growth. There is no reason to expect, a priori, that the primates, as an order of mammals, would be any more uniform in growth patterns than the Carnivora. Newell-Morris & Fahrenbach (1985) reviewed the use of nonhumanprimatesas models for human development and growth and concluded that "...thereare problemswith the extrapolationfrom the nonhumanprimatemodel to the humancondition because of intergenericdifferences in size, growth and developmentrates, and timing. Although investigators justify direct extrapolationof their findings on the basis of the close genetic relationships of all primates,this assumptionin many cases may be little more than absolute faith in the evolutionaryargumentfrom which it stems" (1985:35). Watts (1990) points out thatthe patternsof growth of the New World monkey Cebus, Old World rhesus monkeys, chimpanzees, and humans, are all derived from an ancestor or ancestors. Each of the living species is likely to be derived from its ancestor in ways that are independentand unequaldue to "differentecological and adaptivecircumstances" (1990:99). Wattsarguesthatrhesusmonkeys make poor models for humangrowth because of their advanced state of skeletal

148

BOGIN

development at birth and early onset of puberty and menarche compared with humans. The rhesus is also a seasonal breederin nature;ovulation takes place in the fall or wintermonthsand birthoccurs in the springor summer.Female rhesus monkeys raised outdoorsmay reach menarchein any season but have their first ovulation only in the fall (Wilson et al 1988). When housed in artificial conditions, such as for the indoor-housedmonkeys studiedby Tanneret al (1990), sexual maturation for female rhesus is very much alteredbecause they can reachfirst ovulation in any season. Several studies show that when animals, including primates, are housed indoors and exposed to 12 h of light followed by 12 h of darkprocedure),many of theirnormalendocrinesystem functionsare ness (a standard and disturbed,or even obliterated(Bogin 1977). Because both sexual maturation body growth are influenced by many of the same hormones, it seems unreasonable to expect nonhumanprimatesrearedindoors to serve as models for human growth and development. One colleague, Leigh, offered the following comment after readinga draftof this review: "Thepresence of a spurtin tibia or in otheranatomicalunits certainly does not, as your review indicates, imply homology with the human statural spurt."In biology, the term homology is used to describe anatomicalstructures, physiological processes, or behaviorsthat are found in differentspecies because of a common evolutionary origin. The five fingers and toes and the intimate mother-infantrelationshipcharacteristicof almost all primatespecies are due to homology. Leigh's point is thatgrowthspurtsfor variouspartsof the body of different species of primates are not necessarily homologous. He continues: "For example, my recent studies of baboons demonstrateunambiguousgrowth spurts in the snout, probablycomparablein magnitudeto tibial spurtsobservedby TanThe presence of these spurts, however, does not imply ner et al in macaques... homology with humangrowth spurtsin stature.CRL and sitting height would be the closest homologues, but even then, locomotor differences among these primates would greatlycomplicateassessmentsof homology. WhatI derive fromthe ideas you are presentinghere, and in the last section, is that it might be best to think of growth spurts as modular and highly evolvable features of ontogeny. Natural selection (or sexual selection) can effectively 'put' spurts where (anatomically) and when they are needed to increase fitness. I think that the field has labored a little too long under Brody's desire to find a universal pattern.What may be happeninghere is more akin to a universalprocess. Perhapswe could propose thatthe universalprocess is modularityand evolvability of growthspurts(or decelerations),and that this process need not producea uniformpatternbetween clade"(S Leigh, personalcommunication). any two species or within a particular An acceptablephilosophy of humangrowth must acknowledge the mammalian andprimatefoundationsfor the humanpatternof growth.But thatphilosophy must account for the ecology to which the human species-indeed any species-is adapted. A recent evolutionary perspective on the ecology of human growthand developmentis available(Bogin 1999:Ch.4). Groundedin evolutionary ecology, a robustphilosophy of humangrowth ought to allow for the evolution of variationson common themes. One variationmay be the evolution of new

EVOLUTION OFHUMAN GROWTH

149

stages of life history, such as humanchildhood(Bogin 1997). Anothervariationis new patternsof growth in body segments, such as the pubertalgrowth spurtsin a baboon snout or a rhesus monkey tibia. A final variationmay be the coevolution of both new stages and new patternsof growth, such as human adolescence and the humanadolescent growth spurt. ACKNOWLEDGMENTS I thankDr. Steven Leigh of the Departmentof Anthropology,University of Illinois, for his commentson an earlierversion of this article.I also thankDr. Yuruzu Hamadaof the PrimateResearchInstitute,Kyoto University, for kindlyproviding the data to create Figure 13 of this article. Both Dr. Leigh and Dr. Hamadaconduct empiricalresearchon patternsof primategrowth. Their findings and interpretationshave greatlyenrichedmy own perspectiveson the evolution of primate growth.

Visit the Annual Reviews home page at www.AnnualReviews.org.

LITERATURE CITED Alexander RD. 1990. How Did Humans Evolve? Reflections on the Uniquely UniqueSpecies. Spec. Publ. No. 1. Ann Arbor: Univ. Mich. Mus. Zool. Bekoff M, Byers JA. 1985. The development of behaviorfrom evolutionaryand ecological perspectives in mammals and birds. Evol. Biol. 19:215-86 Bercu BB, Lee BC, Spiliotis BE, Pineda JL, Denman DW, et al. 1983. Male sexual development in the monkey. I. Crosssectional analysis of pulsatilehypothalmicpituitary-testicular function. J. Clin. Endocrinol. Metab. 56:1214-26 Birdsell JB. 1979. Ecological influences on AustralianAboriginal social organization. In Primate Ecology and Human Origins, ed. IS Bemstein, EO Smith, pp. 117-51. New York: Garland Boas F. 1930. Observations on the growth of children.Science 72:44-48 Bock RD, Thissen D. 1980. Statistical problems of fitting individualgrowth curves. In Human Physical Growth and Maturation, Methodologies and Factors, ed. FE Johnston, AF Roche, C Susanne, pp. 265-90. New York: Plenum Bock RD, Thissen DM. 1976. Fitting multicomponent models for growth in stature. Proc. 9th Int. Biometric Conf 1:431-42 Bogin BA. 1977. Periodic rhythmin the rates of growth in height and weight of children and its relation to season of the year. PhD Diss., Temple Univ. Ann Arbor,MI: Univ.
Microfilms

Bogin BA. 1978. Seasonal patternin the rate of growth in height of children living in Guatemala. Am. J. Phys. Anthropol. 49: 205-10 Bogin B. 1980. Catastrophetheory model for the regulation of human growth. Hum. Biol. 52:215-27 Bogin B. 1988. Patterns of Human Growth. Cambridge,UK: CambridgeUniv. Press Bogin B. 1991. The evolution of humanchildhood. BioScience 40:16-25 Bogin B. 1993. Why must I be a teenager at all? New Sci. 137:34-38

150

BOGIN population of toque macaques (Macaca sinica). J. Hum. Evol. 23:51-77 Coelho AM Jr. 1985. Baboon dimorphism: growth in weight, length and adiposity from birth to 8 years of age. In Nonhuman Primate Modelsfor HumanGrowth,ed. ES Watts, pp. 125-59. New York: Liss Copeland KC, Eichberg JW, Parker CR Jr, Bartke A. 1985. Puberty in the chimpanzee: somatomedinC and its relationshipto somatic growth and steroid hormone conJ. Clin.Endocrinol.Metab.60: centrations. 1154-60 Crawford BA, Harewood WJ, Handelsman DJ. 1997. Growth and hormone characteristics of pubertaldevelopmentin the hamadryasbaboon.J. Med Primatol. 26:153-63 Dasgupta P, Das SR. 1997. A cross-sectional growthstudyof trunkandlimb segments of the Bengali boys of Calcutta. Ann. Hum. Biol. 24:363-69 DemirjianA. 1986. Dentition. See Falkner & Tanner 1986, pp. 269-98 Donaldson HH. 1895. The Growth of the Brain. A Study of the Nervous System in Relation to Education. London: Scott Elkin AP. 1964. The Australian Aboriginies. New York: Doubleday/Anchor Falkner F, Tanner JM, eds. 1986. Human Growth. Vol. 2, Postnatal Growth. New York: Plenum. 2nd ed. Frisancho AR. 1977. Human growth and development among high-altitude populations. In TheBiology of High AltitudePeoples, ed. P Baker, pp. 117-71. Cambridge, UK: CambridgeUniv. Press Garber PA, Leigh SR. 1997. Ontogenetic variation in small-bodied New World primates: implications for patterns of reproduction and infant care. Folia Primatol. 68:1-22 Gavan JA. 1953. Growth and development of the chimpanzee, a longitudinal and comparativestudy. Hum. Biol. 25:93-143 Gavan JA. 1971. Longitudinal postnatal growth in the chimpanzee. In The Chimpanzee, ed. G Bourne, 4:46-102. Basel, Switz: Karger

Bogin B. 1994a. Adolescence in evolutionary perspective. Acta Paediatr. 406(Suppl.): 29-35 Bogin B. 1994b. The evolution of learning. In The InternationalEncyclopedia of Education, ed. T Husen, TN Postlethwaite, pp. 2681-85. Oxford, UK: Pergamon.2nd ed. Bogin B. 1995. Growth and development:recent evolutionaryand bioculturalresearch. In Biological Anthropology. The State of the Science, ed. N Boaz, LD Wolfe, pp. 49-70. Bend, OR: Int. Inst. Hum. Evol. Res. Bogin B. 1996. Childhoodin evolutionaryand bioculturalperspective.In Long TermConsequence of Early Environment,ed. CJK Henry, SJ Ulijaszek, pp. 7-22. Cambridge, UK: CambridgeUniv. Press Bogin B. 1997. Evolutionary hypotheses for humanchildhood. Yearb.Phys. Anthropol. 40:63-89 Bogin B. 1999. Patterns of Human Growth. Cambridge, UK: Cambridge Univ. Press. 2nd ed. Bogin B, SmithBH. 1996. Evolutionof the human life cycle. Am.J. Hum.Biol. 8:703-16 Bogin B, Wall M, MacVean RB. 1992. Longitudinal analysis of adolescent growth of ladino and Mayan school children in Guatemala: effects of environment and sex. Am. J Phys. Anthropol. 89:447-57 Britten RJ, Davidson EH. 1969. Gene regulation for higher cells: a theory. Science 165: 349-57 Brody S. 1945. Bioenergetics and Growth. New York: Reinhold CabanaT, Jolicoeur P, Michaud J. 1993. Prenatalandpostnatalgrowthandallometryof stature, head circumference, and brain weight in Quebec children. Am. J. Hum. Biol. 5:93-99 Cameron N, Tanner JM, Whitehouse RH. 1982. A longitudinalanalysis of the growth of limb segments in adolescence. Ann.
Hum. Biol. 9:211-20

Cheverud JM, Wilson P, Dittus WPJ. 1992. Primate population studies at Polannaruwa. III. Somatometricgrowth in a natual

EVOLUTIONOF HUMAN GROWTH Gavan JA. 1982. Adolescent growth in nonhuman primates: an introduction. Hum. Biol. 4:1-5 Goldizen AW. 1987. Tamarins and marmosets: communal care of offspring. In Primate Societies, ed. BB Smuts, DL Cheney, RM Seyfarth, RW Wrangham,TT Struhsaker, pp. 34-43. Chicago, IL: Univ. Chicago Press Goodall J. 1983. Population dynamics during a 15-yearperiod in one communityof freeliving chimpanzees in the Gombe National Park, Tanzania.Z. Tierpsychol.61:1-60 Greil H. 1997. Sex, body type and timing in bodily development-trend statements based on a cross-sectional anthropometric study. In Growth and Development in a Changing World,ed. DF Roberts,P Rudan, T Skaric-Juric,pp. 59-88. Zagreb: Croatian Anthropol. Soc. Hamada Y. 1994. Standard growth patterns and variations in growth patterns of Japanese monkeys (Macaca fuscata) based on an analysis by the spline function method. Anthropol.Sci. 102(Suppl.):57-76 HamadaY, Udono T, TeramotoM, Hayasaka I. 1998. Development of the hand andwrist bones in chimpanzees. Primates 39:15769 HamadaY, Udono T, TeramotoM, Sugawara T. 1996. The growth pattern of chimpanzees: somatic growth and reproductive maturation in Pan troglodytes. Primates 37:279-95 Harvey PH, Martin RD, Clutton-Brock TH. 1987. Life histories in comparative perspective. In Primate Societies, ed. BB Smuts, DL Cheney, RM Seyfarth, RW Wrangham, TT Struhsaker, pp. 181-96. Chicago, IL: Univ. Chicago Press Janson CH, van Schaik CP. 1993. Ecological risk aversioninjuvenile primates:slow and steady wins the race. In Juvenile Primates: Life History, Development, and Behavior, ed. ME Pereira, LA Fairbanks,pp. 57-74. New York: Oxford Univ. Press Johnson TD. 1982. Selective costs and bene-

151

fits in the evolution of learning.Adv. Study Anim. Behav. 12:65-106 KarlbergJ. 1987. On the modelling of human growth. Stat. Med. 6:185-92 King MC, Wilson AC. 1975. Evolution at two levels: molecular similarities and differences between humans and chimpanzees. Science 188:107-16 Laird AK. 1967. Evolution of the human growth curve. Growth31:345-55 LancasterJB. 1985. Evolutionaryperspectives on sex differences in the higher primates. In Gender and the Life Course, ed. AS Rossi, pp. 3-28. New York: Aldine Leclerc de Buffon G-L. 1777. Histoire Naturelle. Paris:ImprimerieRoyale Lee PC, Moss CJ. 1995. Statural growth in known-age African elephants (Loxodonta africana). J. Zool. Soc. London 236:29-41 Leigh SR. 1992. Patterns of variation in the ontogeny of primate body size dimorphism. J. Hum. Evol. 23:27-50 Leigh SR. 1994a. Ontogenetic correlates of diet in anthropoidprimates. Am. J. Phys. Anthropol.94:499-522 Leigh SR. 1994b. Growing up to be a primate. Evol. Anthropol.3:106-8 Leigh SR. 1996. Evolution of human growth Am.J. Phys.Anthropol.101:455-74 spurts. Lovejoy AO. 1936. The Great Chain ofBeing. Cambridge,MA: HarvardUniv. Press Lovejoy AO. 1981. The origin of man. Science 211:341-50 MarshallWA, TannerJM. 1986. Puberty.See Falkner& Tanner 1986, pp. 171-203 Martin DE, Swenson RB, Colins DC. 1977. Correlation of serum testosterone levels with age in male chimpanzees.Steroids 29: 471-81 MedawarPB. 1945. Size, shape andage. In Essays on Growthand Form, ed. WE LeGros Clark, PB Medawar, pp. 157-87. Oxford: Clarendon Napier JR, Napier PH. 1967. A Handbook of Living Primates. London:Academic Newell-Morris L, FahrenbachCF. 1985. Practical and evolutionary considerations for use of the non-human primate model in

152

BOGIN the Libben population and the HamannTodd chimpanzeesample.Am.J. Phys. Anthropol. 99:66-78 SirianniJE, VanNess AL, SwindlerDR. 1982. Growth of the mandible in adolescent pigtailed macaques (Macaca nemestrina). Hum. Biol. 54:31-44 Smuts BB, Cheney DL, SeyfarthRM, Wrangham RW, StruhsakerTT, eds. 1987. Primate Societies. Chicago, IL:Univ. Chicago Press StruhsakerTT, Leyland L. 1987. Colobines: infanticide by adult males. See Smuts et al 1987, pp. 83-97 Tanner JM. 1962. Growth and Adolescence. Oxford:Blackwell Sci. 2nd ed. TannerJM. 1990. Fetus Into Man. Cambridge: HarvardUniv. Press. 2nd ed. Tanner JM, Whitehouse RH, Marshall WA, Goldstein H. 1975. Assessment of Skeletal and Prediction of Adult Height (TW2 Method). London:Academic Tanner JM, Wilson ME, Rudman CG. 1990. Pubertalgrowth spurtin the female rhesus monkey: relationto menarcheand skeletal Am. J Hum. Biol. 2:101-6 maturation. Teleki GE, Hunt E, Pfifferling JH. 1976. Demographic observations (1963-1973) on the chimpanzees of the Gombe National Park, Tanzania. J. Hum. Evol. 5: 559-98 Thompson D'Arcy W. 1917. On Growthand Form. Cambridge,UK: CambridgeUniv. Press Thompson D'Arcy W. 1942. On Growthand Form. Cambridge,UK: CambridgeUniv. Press. Rev. ed. Timiras PS. 1972. Developmental Physiology and Aging. New York: Macmillan Timiras PS, Valcana T. 1972. Body growth. See Timiras 1972, pp. 273-302 Togo M. 1995. Time-series analysis in human growth studies. In Essays on Auxology, ed. R Hauspie, G Lindgren, F Falkner, pp. 100-5. Welwyn GardenCity, UK: Castlemead Togo M, Togo T. 1982. Time-series analysis

pre-natal research. See Watts 1985a, pp.

9-40

Nishida T, TakasakiH, TakahataY. 1990. Demography and reproductive profiles. In The Chimpanzees of the Mahale Mountains: Sexual and Life History Strategies, ed. T Nishida, pp. 63-97. Tokyo: Univ. Tokyo Press Orlosky FJ. 1982. Adolescent midfacial growth in Macaca nemestrina and Papio cynocephalus. Hum. Biol. 54:23-29 PereiraME, FairbanksLA, eds. 1993. Juvenile Primates: Life History, Development, and Behavior. New York: Oxford Univ. Press Prader A. 1984. Biomedical and endocrinological aspects of normal growth and development. In Human Growthand Development,ed. J Borms, R Hauspie, A Sand, C Susanne, M Hebbelinck, pp. 1-22. New York: Plenum Satake T, KirutkaF, Ozaki T. 1993. Ages at peak velocity and peak velocities for seven body dimensions in Japanese children. Ann. Hum. Biol. 20:67-70 Scammon RE. 1930. The measurementof the body in childhood. In TheMeasurementof Man, ed. JA Harris, CM Jackson, DG Paterson, RE Scammon, pp. 173-215. Minneapolis:Univ. Minn. Press Schultz AH. 1924. Growthstudies on primates bearingupon man's evolution.Am.J. Phys. Anthropol.7:149-64 Schultz AH. 1969. TheLife of Primates. New York: Universe Scott JP. 1967. Comparativepsychology and ethnology. Annu.Rev. Psychol. 18:65-86 Service ER. 1978. The Aruntaof Australia.In Profiles in Ethnology, pp. 13-34. New York: Harper& Row Shock NW. 1966. Physiological growth. In Human Development, ed. F Falkner, pp. 150-77. Philadelphia:Saunders and FK. 1937. Sexual maturation Shuttleworth the physical growthof girls age six to nineteen. Monogr. Soc. Res. Child Dev. 2(5). 253 pp. Simpson SW, Russell KF, Lovejoy CO. 1996. Comparisonof diaphysealgrowthbetween

EVOLUTIONOF HUMAN GROWTH of statureand body weight in five siblings. Ann. Hum. Biol. 9:425-40 von BertalanffyL. 1960. Principlesandtheory of growth. In FundamentalAspects of Normal and Malignant Growth,ed. WN Nowinski, pp. 137-259. Amsterdam:Elsevier Watts ES, ed. 1985a. Non-human Primate Models for Human Growth and Development. New York: Liss Watts ES. 1985b. Adolescent growth and development of monkeys, apes and humans. See Watts 1985a, pp. 41-65 Watts ES. 1990. Evolutionary trends in primate growth and development. In Primate Life History andEvolution, ed. CJDeRous-

153

seau, pp. 89-104. New York: Wiley-Liss Watts ES, Gavan JA. 1982. Postnatal growth of nonhumanprimates:the problemof adolescent spurt.Hum. Biol. 54:53-70 Wilson ME, GordonTF, RudmanCG, Tanner JM. 1988. Effects of naturalversus artificial environmenton the tempo of maturation in female rhesus monkeys. Endocrinology 123:2653-61 Winter JSD. 1978. Prepubertaland pubertal endocrinology. In Human Growth,Vol. 2, Postnatal Growth, ed. F Falkner JM Tanner, pp. 183-213. New York: Plenum. 1st ed.