Bone Artefacts from the Middle Stone Age at Blombos Cave, Southern Cape, South Africa

Bone Artefacts from the Middle Stone Age at Blombos Cave, Southern Cape, South AfricaAuthor(s): Christopher Henshilwood and Judith SealySource: Current Anthropology, Vol. 38, No. 5 (December 1997), pp. 890-895Published by: The University of Chicago Press on behalf of Wenner-Gren Foundation for AnthropologicalResearchStable URL: http://www.jstor.org/stable/10.1086/204678 .

Accessed: 19/03/2013 13:47

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 ofcontent in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new formsof scholarship. For more information about JSTOR, please contact [email protected].

.

The University of Chicago Press and Wenner-Gren Foundation for Anthropological Research are collaboratingwith JSTOR to digitize, preserve and extend access to Current Anthropology.

http://www.jstor.org

This content downloaded from 129.174.21.5 on Tue, 19 Mar 2013 13:47:16 PMAll use subject to JSTOR Terms and Conditions

http://www.jstor.org/action/showPublisher?publisherCode=ucpress

http://www.jstor.org/action/showPublisher?publisherCode=wennergren

http://www.jstor.org/action/showPublisher?publisherCode=wennergren

http://www.jstor.org/stable/10.1086/204678?origin=JSTOR-pdf

http://www.jstor.org/page/info/about/policies/terms.jsp


Reports

An Interview withC. Loring Brace1

helke ferrieRR #2, Rivendell, 1997 Beechgrove Sideroad, Alton,Ont., Canada L0N 1A0. 16 i 97

Introduction. C. Loring Brace is Professor of BiologicalAnthropology in the Museum of Anthropology at theUniversity of Michigan in Ann Arbor. He descends froman old New England family members of which wereprominent in the history of American letters and whosearticulations of ideological principles influenced thebiological sciences. As a graduate student of W. W.Howells at Harvard University he began to question thethen accepted paleoanthropological dogma concerningthe separate species status of the Neanderthals. His re-search has emphasized the role of cultural behavior, es-pecially diet, in the interpretation of human evolution.Brace has probably the largest data base on the evolu-tion of hominid teeth, having examined teeth (and cra-nia) throughout the world for decades. He has alwaystaken a strong stance against taxonomic splitting andemphasized, instead, a dynamic, Darwinian, evolution-ary perspective. His research into the biology and his-tory of the concept of race, conducted partly with Ash- C. Loring Brace, 1992. (Photo Della Cook.)ley Montagu, has been very influential in anthropologyand recent cultural and political history. Many of his

ford. He married the granddaughter of Elder Brewsterstudents are now prominent paleoanthropologists. His(1566–1644), the preeminent Puritan in Plymouth fromscholarly work is graced by the unique and delightful1620 on. Any family that has roots that continue inaddition of humorous limericks on all aspects of thesitu from 17th-century New England is related to vir-practice of paleoanthropology. A small sample and onetually any other such family. The father of my great-new limerick, specially commissioned for this inter-grandfather, John Pierce Brace, was born in 1795, and heview (conducted in May 1996), are included here.was headmaster of Litchfield School. There one of hisstudents was Harriet Beecher, who later as Harriet Bee-HF: Your family claims to have come to the shores ofcher Stowe of Uncle Tom’s Cabin [1852] fame wroteNew England even before the Mayflower. Is that true?Old Town Folks [1869], whose hero, Mr. Rossiter, wasmodeled on John Pierce Brace. I mention this becauseCLB: My grandfather, Charles Loring Brace II, used tohis son, Charles Loring Brace (my great-grandfather),joke about that, but there may be something to it. Therenormally gets all the publicity in retrospect.were people in the early 17th century in small home-

steads on the East Coast who did not leave any records.HF: Was it your great-grandfather who brought DarwinHowever, I do know that the Brace family comes fromto America?southern England. My first recorded ancestor, Stephen

Brace, arrived in Cambridge, Massachusetts, and appar-CLB: Well, maybe not to science, but he helped bringently did not like it and went on in 1660 to Connecti-Darwin to the public, yes. His cousin was the wife ofcut, where he is regarded as one of the founders of Hart-Asa Gray.2 Between Christmas and New Year in 1859

2. Asa Gray, a botanist at Harvard, corresponded extensively with1. 1997 by The Wenner-Gren Foundation for Anthropological Re-search. All rights reserved 0011-3204/97/3805-0006$1.00. Darwin before and after the Origin was published. He actively

851



852 current anthropology

my great-grandfather saw Asa’s copy of Darwin’s Origin CLB: My father was a professor of English at BostonUniversity primarily, as a friend of mine put it, to sup-of Species [1859], which was then hot off the press. Dar-

win had been in correspondence with Gray for some port his habit, and that habit was writing novels. It wasmy mother who in the 1920s, prior to my birth, earnedtime. In fact, in Darwin’s effort to claim priority [over

Wallace concerning the selectionist theory of evolution] her M.A. in biology at Boston University and was a biol-ogy instructor until her marriage to my father. My par-one of the letters [Darwin 1857] to Gray was used as

part of the proof included in the package presented to ents were strikingly different, but each admired whatthe other had. My father was tall, elegant, gracious, ur-the Linnaean Society in 1858. Darwin had asked his

publisher, John Murray, to send a copy of the Origin to bane, educated, and an idealistic liberal of an old NewEngland family who took modest pride in the fact thatGray, and so my great-grandfather encountered it. He

was at once fascinated. At the time he was on his he had voted for the Socialist ticket every time thatNorman Thomas ran for president. My mother was away to Concord, Massachusetts, to present on New

Year’s Day to the Transcendentalists his vision of how Republican whose father had come from a Midwesternpig farm and whose mother was from a Vermont dairyto help the orphans of New York. You see, my great-

grandfather was also the founder of the Children’s Aid farm. She was short and dumpy, with a marvellous headof thick dark red hair; she exuded practicality, commonSociety. [Henry David] Thoreau and Bronson Alcott and

several others were present to hear him; [Ralph Waldo] sense, and competence and cooked like a good farmhousewife. She had the keenest sense of smell of anyoneEmerson was not present, but my great-grandfather cor-

responded with him regularly. They were all of the I have ever met. My father’s admiration of her worldspins through all his novels. She knew how to deal withsame intellectual network. Over dinner the ideas of

Darwin’s Origin were introduced to them, and that is any emergency. It came about naturally that she intro-duced me to Andrews’s Meet Your Ancestors [1943].how they all picked up on it. I did not know about any

of this until I was told by a Thoreau scholar a few years Now, that was absolutely fascinating! I thought to my-self, ‘‘You mean you can do that for a living? Oh, that’sago. This incident was not recorded in the life and let-

ters of my great-grandfather which my great-aunt had what I want to do!’’ In high school I did a project onEuropean skull types (curiously, for my physics class),published [Brace 1894]. As he had also been prominent

in fighting against slavery and had been a boyhood and I got an A for it. My mother also had a very strongaffection for animals which she instilled in me. Shefriend of Frederick Law Olmsted—whose Cotton King-

dom [1861] was one of the mainstays of the abolitionist tried to treat farm animals as practical things, but inpoint of fact she really loved them deeply.movement—several biographers tried to obtain the

original letters from my parents, but they were neverable to find them. Years later I thought, ‘‘I bet I know HF: That explains the passage in your review [Bracewhere they are! They must be at the Wistar Institute 1973] of a book by Paul Shepard where you poke fun atin Philadelphia, because my great-aunt was married to modern hunters as that ‘‘army of middle-aged masculin-Harry Donaldson there.’’ The institute published the ity that waddles off to the woods each fall to blaze awayAmerican Journal of Anatomy, the American Journal of at anything animate including each other.’’ I assumePhysical Anthropology, and so on. It was a premier sci- that hunting was not part of the family tradition.entific institution then.

CLB: You assume correctly! But target shooting is a dif-HF: And did you find those letters? ferent matter. I got my ‘‘marksman’’ when I was drafted

into the army and came within a whisker of getting myCLB: I mentioned my guess to a writer, Annette Fry, ‘‘sharpshooter.’’ I missed it by one point.who was planning to do a full-scale and updated biogra-phy of him. Later she called me back with the sad news

HF: You insist that you were ‘‘involuntarily’’ draftedthat although my guess had been right, the librarian ofinto the army in the Korean War.the Wistar had some years earlier felt that they were

taking up too much space, so he had culled them forCLB: My father was a pacifist, but I have always beenletters that were from people he considered ‘‘impor-primarily an antiauthoritarian. You know, ‘‘There is atant,’’ such as Emerson, Theodore Parker (the Bostonianright reason, there is a wrong reason, and there is anabolitionist preacher), John Stuart Mill, Sir Charlesarmy reason’’—that didn’t set well with me. I was withLyell, and the like, and then destroyed the rest. It’s athe SPP (Scientific and Professional Personnel), and thegreat loss.only person who did not have a degree was the companycommander; the rest of us had M.A.’s or Ph.D.’s. It wasHF: How did you become interested in the biologicalvery upsetting to be drafted out of graduate school andsciences?taken away from my studies with [Earnest] Hooton,who then died that same year. Howells took his posi-tion.

spread and supported Darwinism in America and focused on recon-ciling ‘‘natural theology’’ with Darwinian evolution. HF: What did you do during your stint with the army?



Volume 38, Number 5, December 1997 853

CLB: I was part of the gas-mask fitting program, which was asked ‘‘Rory, what would you do if you had to goto Harvard?’’ he would lie down and play dead. If askedturned out to be very useful later in my anthropological

work. Chemical warfare was a big worry because of ‘‘What would you do if you were to go to Yale?’’ the dogwould jump up and lick everybody’s face.)some intelligence obtained from Germany after World

War II. I applied the technique of factor analysis, whichI knew Howells had begun using in anthropology. HF: Where had you spent your undergraduate years?

CLB: At Williams College, where I had done geologyHF: Could you kindly explain briefly what factor analy-sis is? and biology—which proved very useful, but the purpose

of Williams College was to give some educational ve-neer to the young men who then went straight intoLCB: It is a multivariate analysis. You pull out a series

of factors from reduced dimensions. What you are try- Daddy’s business. At Harvard, however, I was intro-duced to the latest research, especially to what becameing to do is to find poles of significant dimensions

where all that cloud of facts will revolve around this or the so-called New Synthesis [neo-Darwinism]. Howellshad us all read G. G. Simpson, [Ronald] Fisher, [Theo-that axis of explanation, so a pile of otherwise disparate

facts can then be organized to make more sense. dosius] Dobzhanzky, and all the other leading people inthe field. [Hallam] Movius was also one of my teachersat Harvard. I wanted to do paleoanthropology and wasHF: How did you apply factor analysis to the fitting of

gas masks? very much aware that control of the archaeological rec-ord was absolutely essential if one wanted to figure outwhat people were doing in prehistory. I had studied theCLB: What the army wanted me to do was to make their

M1A1 gas-mask fit everybody. They fitted them onto Krapina material through the publications but wantedto work with the actual fossils in Yugoslavia—it is stillthousands of volunteers to get a random sample of the

population. Now, it occurred to me that you didn’t need one of the largest collections of Neanderthal fossils—and I also wanted field experience in archaeology. Now,thousands of people at the cost of hundreds of thou-

sands of dollars a year to achieve this end if, instead, it so happened that Movius was a close friend of my fu-ture father-in-law, William J. Crozier, and like an uncleyou had a sampling of the various dimensions of the hu-

man face. You could then select individuals that repre- to his daughter, my future wife. In fact, Movius indi-rectly introduced me to my wife. She had a well-to-dosent dimensions of extremes within that population

and use those each time you construct a new mask. If it relative who helped Movius raise the funds to purchasethe site of Abri Pataud, France (where, incidentally,fits them, it will fit everybody in the country. It worked

beautifully! But it was a matter of great chagrin to those K. C. Chang and I dug as fellow graduate students [formore detail see Ferrie 1995]). Although I was not an ar-who ran the project, because it eliminated a lot of jobs

and saved a lot of money, so they shelved my report and chaeology student, I was allowed to tag along and learn,which I appreciated greatly. Meanwhile, I had met mykept on fitting thousands of people every year with their

new gas masks. But in 1961 Minnesota Mining and wife in the basement of the Peabody Museum. It wasvery romantic indeed. My wife-to-be was in the base-Manufacturing discovered my report, obtained army

clearance, and bought it from me and wound up using ment of the Peabody Museum, washing pottery forPhil Phillips, a colleague of Hal Movius, and for Bobit to make surgical masks.Greengo, a graduate student of Phillips’s. I was measur-ing chimpanzee teeth in the lab on the fifth floor. WeHF: What did you do your Ph.D. on at Harvard?met by the Coke machine in the basement of the geol-ogy wing—the last nickel Coke to be had in CambridgeCLB: After my discharge from the army I didn’t go

straight back to Harvard but, at Howells’s suggestion, then. She had been taught the technique of drawingflint tools and fossils by Nancy Movius and subse-went to Jackson Laboratory at Bar Harbor in Maine to

analyze the data they had been collecting there since quently illustrated all my books. So I got a fellowshipfor the year 1959–60, and my wife and I went to Europethe mid-1940s on dogs. That project had been set up

originally by a student of Hooton’s to test the ties be- and dug at Abri Pataud, visited Combe Grenal, metFrancois Bordes, and spent some time at Oxford intween physical characteristics, physiology, behavior,

genetics, and so on. They had been raising generations [Niko] Tinbergen’s lab. The most important part of thistrip was, however, studying the Krapina material ofof dogs and running them through standard test pro-

grams. The following year I returned to Howells at Har- [K. D.] Gorjanovic’s collection in Zagreb.vard with my data to do a factor analysis on them. Thatwas then to become my thesis. It turned out that ties HF: Was that the basis for your subsequent reinterpreta-

tion of the Neanderthal problem [Brace 1964a]?between biology and behavior were a complete nega-tive. (I must add that family tradition dictated that ev-erybody went to Yale. Fortunately, my father had al- CLB: Yes, but I had already begun with that in a gradu-

ate course with Howells. I had read Weidenreich, butready broken tradition by first attending Amherst andthen getting his Ph.D. at Harvard. He had a sister who we were, of course, told to read him for his descriptions

rather than for what he said about evolution. But hishad an Irish setter called Rory. The story goes that if he




ideas about human evolution [Weidenreich 1940, 1943, Mechanics are never detected,In the popular view that’s projected,1947a, b] struck me as particularly interesting, and

I thought, ‘‘My goodness, why doesn’t anybody talk Since all that worksIs change by jerks,about this?’’ Now, I had already read piles of neo-

Darwinism and was quite familiar with evolutionary And Darwin is flatly rejected.theory, but when I read palaeoanthropology none of it

For that’s how most scholars behave,was there! Evolutionary biology was absolutelyAnd it’s easy enough to be bravemissing.

When objection at mostIs the groan of a ghostHF: You say that you are still trying to bring physical

As it turns in its Westminster grave.anthropologists ‘‘kicking and screaming into the 20thcentury’’ [Brace 1989].

HF: How did you come to develop the habit of com-CLB: Yes, exactly. Not much has changed in the past 30 menting on paleoanthropology with your doggerel?years, as I pointed out again recently in the AmericanAnthropologist [Brace 1995a]. CLB: My favorite uncle used to make up limericks for

us on picnics. In college I discovered the dirty limericks,HF: How did the Neanderthals get your attention? which I thought were absolutely delightful! In fact, I

could at one time recite those for two hours withoutCLB: In 1957, prior to my European trip, I had written stopping. I quite naturally began to write some of mya paper on the Krapina material for Howells’s graduate own and eventually wrote limericks also as commen-seminar on human evolution. What had prompted my tary on evolutionary issues. Some of my students re-interest was a comment by Le Gros Clark in The Fossil cently collected all my clean and X-rated poems of theEvidence for Human Evolution [1955:180]. He wrote: past 30 years and gave them to me in a bound edition‘‘On purely morphological grounds (and without refer- at a dinner following a session held in my honor at theence to paleontological evidence), there is no certain ar- AAPA [American Association of Physical Anthropolo-gument why H. neanderthalensis could not be ancestral gists] meetings last spring in Oakland, California.to H. sapiens. But, in this particular instance, the fossilrecord shows clearly that such was not the case.’’ How

HF: Your ‘‘perennial scold’’ takes a rather interestingcould he say that? Just because Sir Arthur Keith [1925d,turn in your observation [Brace 1981, 1985, 1988, 1989,vol 1:198–99] told us it didn’t happen? Incidentally,1991, 1993, 1995a, n.d. (1992)] that palaeoanthropologyKeith wrote that the Neanderthals were ‘‘replaced, withis in your view characterized by ‘‘medieval essen-the dawn of the Aurignacian period, by men of the sametialism,’’ especially when you go so far as to assert occa-type as now occupy Europe. . . . A more virile form ex-sionally that the forces of Wilberforce rather than thetinguished him. He was not an ancestor of ours, but aforces of Huxley have succeeded.distant cousin.’’ He maintained this stance right up to

the end of his life [Keith 1950:318–19]. Well, I went andCLB: Indeed. You see, once I began to investigate theread everything that had been written on the Neander-facts on the basis of which we supposedly ‘‘know’’ whatthal question. I even went and learned German at Ber-happened in human evolution, I encountered not onlylitz, as I was in a rip-roaring hurry and needed to trans-Weidenreich but also Hrdlicka, who was considered alate the monographs by Gorjanovic myself. Then, ofgood compiler but whose interpretations were consid-course, I discovered Gustav Schwalbe [1906] and readered beyond the pale, as it were. His idea of the ‘‘Nean-Marcellin Boule’s [1913, 1921, 1929, 1937] monographsderthal phase’’ of evolution [Hrdlicka 1927, 1930] madeand realized that Le Gros Clark’s statement owed noth-perfect sense to me. Incidentally, I met Dobzhansky ating to the evidence itself. It amounted to an effort toan AAA [American Anthropological Association] meet-deny that evolutionary continuity was even a possi-ing in Detroit in 1964 and asked him about the strikingbility. When I published my 1964 paper, I was naivesimilarity between his views on evolution and those ofenough to believe that all that was necessary was to ex-Weidenreich. Dobzhansky replied, ‘‘Yes, I used to haveplain to paleoanthropologists that their ideas about evo-lunch with him at the American Museum, and we usedlution came from a totally different source and wereto discuss these issues there. And yes, we had a gener-antithetical to Darwinism, and then they’d all turnally compatible outlook.’’around and attempt to incorporate modern evolutionary

theory. It never happened—which is why [Brace n.d.HF: Your crusade for the Neanderthals seems to have(1992)]had some measure of success, since they are now called

I repeat my perennial scold, Homo sapiens neanderthalensis.Our profession is shaped by the mold

Of the covert view CLB: But they are still considered a separate species!That the roots of the new Chris Stringer [Stringer and Gamble 1993] and Ian Tat-

Are not to be found in the old. tersall [1986] continue to be dedicated splitters, and sothe problem remains; as far as theory is concerned, by. . . . . . . . . . . . . . . .




and large, paleoanthropologists haven’t incorporated an looking at those Mousterian hearths, and they lookedlike a depiction of a nebula in an astronomy book!evolutionary perspective as yet. I discussed that in de-

tail in my review [Brace 1994b] of the book on speciesand speciation edited by Kimbel and Martin [1993]. Ul-timately a cladistic view amounts to typological essen- HF: What did this lead you to conclude?tialism, or mind-made categories, which have nothingto do with the adaptive aspects of evolution.

CLB: I’ll tell you, because it is absolutely fascinating. Inthe mid-1960s when I was at Santa Barbara (1961–67),HF: Why do you think this typological essentialism pre-

vails? my friend Ben Finney, who was then also on our faculty,had begun his experiments with Polynesian canoe man-ufacturing and navigation. Eventually, of course, a de-CLB: It’s originally the legacy of French intellectualism,

but today—with postmodernism—it has gotten to the cade later, he celebrated our bicentennial by setting sailfrom Honolulu on July 4, 1976, for Tahiti in such a ca-point where the very existence of reality is denied in

much of anthropology. This makes it possible to argue noe, the Hokule’a, using only stellar navigation tech-niques. He made it, and the saga of his trip was featuredabout mind-made categories, such as we have them in

cladistics, instead of living, adaptive reality. I recently in the October 1976 issue of National Geographic[Lewis 1976], and the Hokule’a graced the cover of Sci-satirized this attitude:ence on June 17, 1977 [Finney 1977]. Now, Ben wantedSome say an outer world exists to have a luau for the private donors who had agreed toWith strictures that define finance his experiment. A luau is a Polynesian party atwhich a whole pig is roasted in an earth oven. Well, co-The core of which my life consists,

But their world isn’t mine. incidentally, my wife had asked for an imu. Since sheis the curator of the family zoo, my first response hadFor if I were not here to see, been, ‘‘It will never fly!’’ But she set me straight: ‘‘No,Who else would give a damn? not the bird, silly, it’s a Polynesian earth oven!’’ So, Imade an imu for her in our backyard, and we naturallyI know that I am really me—

I think, therefore I am. volunteered to have the party for Ben Finney’s support-ers. We roasted two pigs, each 75 pounds, in this imu,And since I am, the world I know and my wife produced the works to go along: she hadIs the one that I can feel; haupia, which is coconut pudding, lomi-lomi salmon,and various other Polynesian specialties. My wife canMy being then just serves to show:

It’s the only one that’s real. cook virtually every style of cuisine in the world, andit is no accident that our second son became a profes-sional chef and has his own restaurant. The luau was aHF: You have addressed some of the problems of bio-

logical evolution, but what do you think about the great success, and come winter my wife suggested weresuscitate the imu and do our turkey in it. So, I wentfact that there appears to be continuity of tools and

other cultural material across these supposed species out to dig it out with a hoe and clean up the wreckageof the previous summer’s luau. And there, in the walllines?of my imu, in the trench of this 10 3 6-foot rock-linedearth oven, was a nebula pattern with fist-size fire-CLB: Well, exactly! You see, French intellectualism was

characterized by a lack of communication between the blackened rocks, swirls of ash, bits of burned bone. Aha!The wall of Francois Bordes’s trench at Combe Grenal!different specialists. Cultural and biological anthropol-

ogy belonged in two different realms. Specialists did A Mousterian hearth! I mean, one must ask oneself howelse a Mousterian hearth after 50,000 years could stillnot share ideas. Francois Bordes was an exception; he

looked at behavior and the continuity of stone tool ty- be up to 2 feet deep. Those weren’t the surface camp-fires of contemporary hunters and gatherers; the Nean-pology from the Mousterian into the Upper Paleolithic.

He was the one who pointed out [Bordes 1958] the con- derthals in Ice Age Europe had themselves some fine lu-aus in large earth ovens! These have, of course, beentinuities and criticized anthropologists by saying that

denying the evolution of modern humans from the Ne- found at many archaeological sites, and some of the Up-per Paleolithic examples are huge, up to 15 feet long.anderthals was rather like saying that the works of

Shakespeare were written by somebody else of the same Now, this information is neither flaking technique nortool typology nor raw material sourcing—it doesn’t fitname. I met him in 1959 at Combe Grenal and told him

how I felt about Neanderthal ancestry, and he fully sup- into the categories archaeologists use. If you point outthese earth ovens to them, they will say, ‘‘Yes, yes, ev-ported my views. Incidentally, it may be of interest here

to mention that whereas Movius dug a site from the top erybody knows that,’’ but it is never mentioned in a sin-gle publication. It’s like trying to find out how aborigi-down and exposed an entire living floor in good Ameri-

can archaeological fashion and then took the layers nal peoples actually live from a working ethnologist.They won’t talk about what’s for dinner; they’ll talkdown, Bordes dug a trench horizontally from the side

in, so you could see the entire spectrum. I remember about mental models and categories.




HF: This experience probably was a helpful clue for the CLB: The sheer chance that a trait will be perpetuatedor lost in small populations. If one has a limited amountdevelopment of your ideas on the importance of cook-

ing to the process of physical change in human evolu- of genetic material in a very small population, the prob-ability that everything is going to be transmitted to thetion.next generation is not very high. Something is likely tobe lost, or something that is present in a substantialCLB: Most definitely! In fact, in order to test my ideas

on the evolutionary dynamics that may have been in- number of the adults will then be present in all of theoffspring and thereby become fixed. It’s just a matter ofvolved in tooth size reduction I did a thorough study of

Australian ethnographies. I wanted to find out what the the luck of the draw and rarely of any adaptive value.The biological community is willing to admit now thatAborigines used their teeth for. And you know, starting

with the appearance of professional anthropology at the genetic drift does occur, and in human evolution wehave plenty of examples, especially in trivial differencesbeginning of this century, the ethnographies have been

getting worse and worse, and you find out less and less between human populations—in the shape of the eye,for instance, or the shape of the external ear, in nuancesabout what those people did. So you have to go back to

the unpolished, unsophisticated 19th-century explor- of the cheekbone, or the way the back and the bottomof the skull are put together. These features are underers, Major Mitchell [1839] and so on. Now, I needed to

know what the Walbiri were eating, so I read Mervyn genetic control, and they will differ at random from onepopulation to another and become fixed in some area byMeggitt’s [1962] accounts. (He had been on our faculty

here.) The information wasn’t in it, but then I found pa- sheer accident. Some of the differences between livinghuman populations which are erroneously called racialpers by him in obscure journals [Meggitt 1957, 1964] out

in the boonies, and there was the information I wanted are clearly the result of genetic drift. However, this doesnot apply to adaptive features. Now, I have looked espe-on what the Aborigines ate. I asked him about this at a

cocktail party once. ‘‘Mervyn, why do I have to go to cially at situations where the force of selection was re-duced or where it is absent. This is not something zool-such an obscure journal to find out what people in

Yuendumu make for dinner?’’ He looked at me with a ogists generally look for, because they try to identify theways in which the organism attempts to meet the forcestwinkle in his eye and replied, ‘‘I have to satisfy my col-

leagues! I don’t get any brownie points for putting that of selection.in my monographs!’’

HF: Why did you look for signs of the reduction or ab-sence of selective force?HF: You evidently weren’t too concerned with brownie

points yourself, since you developed an entire theory ofevolutionary biological change in which cooking figures CLB: Because it struck me as peculiarly human. This is

what I have called the ‘‘cultural ecological niche’’ [Braceprominently! I am referring to what you call the ‘‘proba-ble mutation effect’’ [Brace 1963, 1979, 1995a, b; Brace, and Montagu 1965:216; 1995b; chap. 9]. The same ap-

plies to domesticated animals where humans have re-Hunt, and Smith 1991].duced the selective pressure on them. In all those in-stances you get consequences that I would attribute toCLB: Standard evolutionary change in Darwinian the-

ory is driven by natural selection. Neo-Darwinian or- the effect of mutations alone in the absence of naturalselection. It was, of course, Darwin who was the firstthodoxy goes beyond what Darwin himself would have

stood for; it states that change can only occur as a result one to point out the possibility of self-domestication inpeople also. His survey of plants and animals under do-of natural selection. This has been satirized by Ste-

phen Jay Gould [Gould and Lewontin 1979] as ‘‘hyper- mestication [Darwin 1868] was an investigation of theconsequences of domestication for natural selection. Itselectionism,’’ but most biologists agree that natural

selection is the primary driving force for organic evolu- is absolutely extraordinary how Darwin always seemsto have thought of everything first!tion though not the only one—an important consider-

ation with regard to my proposed probable mutation ef-fect. Sewall Wright [Wright 1931] saw the possibility of HF: So, what is the ‘‘cultural ecological niche’’?chance alone’s producing a change through time. Unfor-tunately, Fisher’s intransigence was such that he at- CLB: It is a whole realm of learned behavior where that

learning is conditioned by vicarious experience of oth-tacked Wright consistently and made no attempt to un-derstand him. Wright had proposed several evolutionary ers—previous generations as well as contemporaries.

This gives its beneficiaries a tremendous advantagemechanisms in 1931, and one of these was genetic drift,which was satirized as the ‘‘Sewall Wright effect’’ over even the cleverest creature that has to learn every-

thing from scratch. But learning at a remove of either[Wright 1951].3 Wright defended this concept against allodds, and it is finally becoming accepted today. time or place can only be done symbolically, and that

means by the use of language. Obviously the pressuresto learn the rudiments of linguistic behavior had to beHF: Could you briefly define ‘‘genetic drift’’?tremendous, and they had to have been the commonheritage of virtually all living human populations. With3. For a full treatment of the work of Sewall Wright, see Provine

(1986). that as a general background, people could then learn




very specific survival strategies such as what is edible, tomcats. It is a picture of absolute futility: a 160-poundhuman being in pursuit of 5 pounds of cat! But it waswhat is not, what you do to make it edible, how you

avoid drought, or cold, or whatever. For as long as cul- my wife who solved the problem by appearing in frontof the door with a tin of cat food, tapping it loudly withture has been the key to human survival, people have

been discovering ways to circumvent the immediate ef- a can opener; the cat returned at once. (Parenthetically,I should add that this says something about the differ-fects of particular selective forces: particular cultural

elements have been devised to lower the effects of par- ent ways men and women use their brains!) Now, allyou have to do is decrease the hominid to a 50-poundticular aspects of selection. Clothing and shelter protect

us against the elements, tools and weapons compensate australopithecine and increase the house cat to a 200-pound Plio/Pleistocene cat and guess which is dinner!for physical weakness, and each time some learned de-

vice is used to provide that compensation, the selection Brain [1981] actually found a hominid skull with holesmatching the huge canine teeth of a leopard. Human lo-that had previously maintained the anatomical capabil-

ity is weakened. The probable mutation effect then comotion is generally not very efficient for escape orcatching prey; there needs to be something else for suc-takes over, and the original capability is reduced over

time. cess. Furthermore, hominids do not have large, pro-jecting canine teeth, so there needs to be something tocompensate for this lack. It was, of course, once againHF: So morphological appearance would change over

time also? Darwin [1871] who suggested that the reason for the re-duced canine teeth in humans is their reliance on hand-held tools. Meanwhile, Washburn [1959, 1960] had alsoCLB: Yes, because the long-term reliance for survival on

specific aspects of culture was paralleled by reductions pointed out that if you give a baboon a digging stick, itcan effectively double its food supply. The competitorsin bone and muscle and other raw physical capabilities,

and this accounts nicely for the emergence of the rela- early hominids had on the plains of Africa were wart-hogs and baboons, but an australopithecine with a dig-tively slender ‘‘modern’’ form from our Middle Pleisto-

cene and Neanderthal ancestors. Adaptation to the cul- ging stick could have reached tubers much farther downthan baboons and warthogs could reach with their fin-tural ecological niche is a common human heritage,

which is why children of all living human groups learn gers and tusks. Now, what I had done—and this was mymistake—was to attribute to the australopithecinestheir language at exactly the same stage of life and why

no language is easier or harder to learn than any other cultural characteristics in excess of the evidence pro-vided at that time. In fact, it was Elwyn Simons whoand all are equally effective in transmitting what others

have learned. gently pointed out to me that I had overreacted toKeith’s [1925a, b, c] suggestion that the australopithe-cines were just another bunch of apes by making themHF: Did this insight into the ‘‘peculiarly human’’ lead

you to postulate the single-species hypothesis? too human. Of course, Keith himself abandoned thisview in time, but Simons’s cautions were most timely.I had indeed jumped too far in the opposite directionCLB: Yes, But I made a mistake on this originally. The

controversy centered around KNM ER 3733 (a Homo [Brace, Mahler, and Rosen 1973]. At the end of thatAAA meeting it was Washburn who said that there waserectus) and KNM ER 406 (a robust australopithecine),

which are contemporary and very different indeed. The no reason there couldn’t have been more than one hom-inid who might have made a stab at inventing rudimen-stratigraphy convinced me that here we do have two dif-

ferent species, but there is more to this. I offered my tary cultural behavior. This criticism forced me to re-vamp my views, especially in the light of the supportmodel of evolution as a follow-up to the Neanderthal

problem. This was at the AAA meeting in Detroit in from the emerging fossil and stratigraphic evidence.1964, where I tried to present the australopithecines asa single species and interpreted the differences between HF: So, you took your own advice [Brace n.d. (1992)] to

heart! I am thinking of ‘‘Examine the strata/Containingthem as possibly sexual dimorphism. Clark Howell ridi-culed this as my inability to tell the boys from the girls! the data/And use the ensuing array!’’The point I made was that within the cultural ecologi-cal niche no more than one species could exist at any CLB: Yes! There is no reason you can’t have more than

one australopithecine using tools, and I believe there isone time. I realized that culture is the key to humansuccess and thought that australopithecines had to use evidence from certain kinds of scratches on their teeth

to suggest that indeed they did. The analysis of theirtools to survive, since they were obligatory bipeds andhad a locomotor system that was simply inadequate to fossil bones indicates that their likely source of food

was tubers [Ryan 1980, Ryan and Johanson 1989]. Oncesave them from anything that wanted to make lunch ofthem. Now, I must briefly digress, because there is an we get to Homo, however, we see striking changes, such

as the reduction of molar tooth size. In fact, evidenceautobiographical basis to this which I have, in fact, in-cluded in my textbook [Brace 1995b]. Picture a middle- for the cultural ecological niche comes also simply from

looking at how very different the trajectories of changeaged Homo sapiens attempting to catch a small femaleFelis domesticus that one spring morning has escaped are for different aspects of human evolution. The hu-

man ecological niche is not a monolith, as far as its ef-in response to the noises made by the neighborhood




fect on human capabilities is concerned. Of key impor- The full story is actually more complicated than sucha thumbnail sketch, but the basic picture is here. Thesetance is, of course, the need to learn from others so as

to avoid inventing the wheel over and over again. This experiments are available in the literature [Novick andWeiner 1957, Koch 1983, Baldwin, Kirkish, and Kochproduces an intellectual development which ultimately

must have resulted in language. The selective forces in 1994]. The problem is that paleoanthropologists are ask-ing different questions and therefore do not considerthe human ecological niche must be responsible for the

tripling of brain size. these research results in the context of human evolu-tion.

HF: How does your concept of the cultural ecologicalniche square with the popular evolutionary scenario HF: So, how do you see the probable mutation effectbased on the concept of the neotenous ape? playing itself out in hominid evolution after the austra-

lopithecines?CLB: It doesn’t. I have written about this several times[Brace 1995b:82–83], and not too long ago Shea [1989] CLB: The probable mutation effect applies only to traitsdid a very good job of blowing that whole concept out that are under selective-force control. A lot of the varia-of the water. Neoteny literally means ‘‘retention of the tion we see in the fossils of the Pleistocene is due tonew,’’ that is, of the supposedly advantageous juvenile variations that are not under selective-force control butcondition. However, the human brain is the develop- are far more likely the consequence of genetic drift. Inment of something really quite more. Humans are not order to get, say, Homo erectus out of an australopithe-neotenous in very many respects. For example, the cine two things are necessary—increase in brain size,enormous length of the leg is hardly neotenous. The which is under direct selective force, and reduction inseparate forces of selection acting on brains, on the one tooth size by 30% or more—and this is where I suggesthand, and those forces which influence the develop- that the relaxation of selection has allowed the accumu-ment of face size, on the other, have accomplished the lation of mutations which interfered with the develop-appearance of this proportion between face and brain mental track that would otherwise have produced anwhich is supposedly an indication of neoteny. But this australopithecine dentition. It isn’t so much a matter ofproportion in itself is of absolutely no evolutionary food processing as one of a change in diet. You find ameaning. It is the accidental consequence of different change from a largely vegetarian diet to an increase inselective regimes. The reduction of the face size in hu- animal protein which requires much less chewing. Puremans is, however, not the reduction of the previous face carnivores, like the big cats, don’t even have molars atsize; it is simply the failure to grow into that size and all.to those proportions.

HF: Your probable mutation effect suggests an evolu-HF: So, if you don’t use it you lose it? tionary dynamic with which to interpret the great varia-

tion in the Pleistocene hominid fossil record; is it not atCLB: No! If you don’t need it, you lose it. the same time a critique of the gay abandon with which

phylogenetic trees are created?HF: So how does this loss become heritable and fixed?

CLB: Indeed! As in the ‘‘Owed to Trees’’ [Brace 1981]:CLB: The traits in question are, of course, under genetic

As we survey the path we’ve trod,control. Tooth size, for example, is under genetic con-Of knowledge gained by labored plod;trol, but these primary gene products are also subjectNow each aspiring learned clod,to mutations. If you reduce the intensity of selection,Will try to see,variation is allowed to accumulate, and the obvious, av-How he can be a bit like Goderage variation will be in the direction of reduction.And make a tree.

HF: Is there experimental evidence to support yourHF: Is there a conscious connection between your effortview?to expose what you have called the ‘‘theatre of the ab-surd’’ [Brace 1991] in paleoanthropology and your tak-CLB: Yes, one example comes from single-cell organ-ing on the creationists?isms. The ability to synthesize a necessary substance

for metabolic continuity, for survival, becomes subjectto random genetic variations when that substance is CLB: Oh, debating [D. T.] Gish wasn’t something I

wanted to get into! But Gish has made it his businessprovided by the experimenter. The microorganismceases to synthesize the substance. If that situation is to go around the country and debate people everywhere.

He had called my friend Dick Alexander here at the mu-not allowed to last too long, a reversal can occur whenthe substance is removed once again, although most seum’s Department of Zoology, but he refused to get in-

volved. Now, I had once been suckered into debatingof the bacteria will die off and only a small number withthe right reverse mutation will once again synthesize it Henry Morris [director of the Institute for Creation Re-

search in San Diego] in Auckland, New Zealand, inwhen it is not provided by the surrounding medium.




1973 when I was visiting professor there, so the cre- HF: You pursued the study of evolution largely throughthe study of teeth. How did that come about?ationists decided to call me. Saying no to Gish is what

he feeds on and interprets as fear of tackling his argu-ments. I just couldn’t stand for that. So I agreed to de- CLB: It was a stroke of good fortune. When I began grad-

uate school, I got into a chimpanzee-tooth-measuringbate him, and it took a whole year to go through all thatcreationist literature. My department wasn’t suppor- project with a dentist, Dr. E. Leon Schuman [Schuman

and Brace 1954], at the Peabody Museum at Harvard.tive; they considered it a waste of time, actually. Therewas virtually no publicity in the papers, the radio, etc., Some time later, after the gas-mask project, I was taking

an anatomy course at the Harvard Medical School withbut to my amazement there were more than 3,000 peo-ple who had been bused in by the creationists from G. E. Erikson. I was talking to a fellow student, Jack

Prost (who is now at the University of Illinois in Chi-churches all over the Midwest and from Canada. Some90% of the audience was sitting there clutching their cago), and I told him that I couldn’t really tell the differ-

ence between chimpanzee and human molars. He said,Bibles! Now, I knew from the recorded previous debatesthat Gish likes to come second in the debate and what ‘‘Nonsense!’’ So I extracted a lower second molar from

my chimp cadaver and slipped it into my pocket and thehis ploys would be at the end. Sure enough, he won theright to go first by the flip of the coin but chose to go next morning showed it to Jack and asked him what he

thought this was. He assumed it was human and said,second. At the end, for his rebuttal, he promptlybrought out his favorite cartoons which show a whale ‘‘Well, it’s lower right—hmm, first or second. I’ll guess

first.’’ I asked, ‘‘What kind of an individual?’’ His replybeing transformed from a disconcerted-looking domes-tic cow with an obviously milk-filled udder. I antici- was, ‘‘Probably adult, from the lack of wear, not middle-

aged.’’ Then, smelling the formaldehyde, he looked uppated this particular cartoon and had my rebuttal allwritten up and just read it out. and said, ‘‘No!’’ and I said, ‘‘Yes!’’ and I turned and

slipped it right back into the chimp socket from whichit had come and made a believer out of him. He stillHF: I am reminded of your frequent critique of the prac-uses this story in his classes.tice of matching parts of fossils from different parts of

the world.HF: Is there a chance of encountering such an identifi-cation problem in the fossil record?

CLB: Oh yes, yes! All those phylogenetic trees are badenough, but I get a bit frustrated when a new species is CLB: Absolutely! So when I began to look at Neander-named on the basis of a single tooth or a fragmentary thal tooth measurements and became involved with themandible without knowing more of the rest of the Krapina material I realized that we didn’t have informa-beast. And then combining a piece of a specimen from tion on the spectrum of tooth sizes in living humanEthiopia and putting it together with a piece from Tan- populations. Yet it was perfectly obvious that there waszania and giving them a new name—oh, dear, this gets a great difference between modern and Neanderthalreally awkward. With due apologies to Shakespeare, I teeth. There had been a substantial evolutionary changehave expressed my views on that as follows [Brace which needed documenting. There was information1995b:171]: here and there in bits and pieces, such as the fact that

contemporary Australian Aborigines had bigger teethDouble, double, toil and trouble,than modern Europeans. So I slowly began to collectMix the fragments from the rubble;this information by taking measurements myself andA bit of this, a piece of thatcreated a data base of tooth measurements of modernJaw that’s chinless, skull that’s flat,human populations as well as of their relatives in theOrphan molar as a token,past. In this way I built up a picture of modern humanScrap of femur, rib that’s broken,dental variation and ultimately also a picture of thatSome from here, some from there,variation in the past. The data reflect an evolutionaryAnother chunk from God knows where;sequence, pure and simple [Brace 1995b:195, fig. 13.3].Assembled in a dreamer’s bubble,Incidentally, there are now new absolute dates out onNamed a species: Endless trouble.the Krapina material placing it around 130,000 6 10,000years ago, so that material represents early Neander-

HF: What was the outcome of your debate with Gish? thals [Rink et al. 1995, Schwarcz et al. 1996]. Tooth sizehasn’t changed in Neanderthals since Homo erectus,except for the front teeth, which are very large in theCLB: My part of the debate focused on all the many de-

monstrable errors of fact in Gish’s book [1979]. I was latter. The mean didn’t change in half a million years.We also have new, firm dates for Zhoukoudian, for therather pleased that at the end a number of the clergy-

men in the audience came to speak to me because they lower, early material, which is now at 400,000 years orolder [New York Times, May 2, 1996; Shen 1992]. Untilwere profoundly troubled and asked why Gish would

publish such things. Well, I told them, that was their last year we all thought it was only half that old. Fromclose to 2 million to roughly 120,000 years ago humanproblem to sort out, not mine. I understand that Gish

hasn’t tried a major university since then. tooth size didn’t change. The size was an order of mag-




nitude smaller than during the previous australopithe- those things used for? To make clothing from hides.Ofer Bar-Yosef at Harvard told me some years ago thatcine stage and an order of magnitude larger than in mod-

ern humans. There is a variation, of course, especially those African Middle Stone Age sites could have beencalled Mousterian, but there is one thing that is not todue to sexual dimorphism—particularly among the Ne-

anderthals, where we can usually tell the males from be found there: the D side scraper! You only get thosewhere survival depended on clothing and shelter.the females with certainty. So there are these plateaus

in tooth size, from which I conclude that for long pe-riods of time they were using them for the same sorts HF: What do you think the classic handaxe was used

for?of activities. What we have since the Neanderthal pe-riod is a picture of reduction [Ruff 1994]. At Predmost,the earliest Upper Paleolithic population sample of CLB: Aah! Let me tell you a story: You probably have

seen that documentary that is shown year after year tomodern morphology we have had, tooth size is exactlybetween the Neanderthal and modern Europeans. The new undergraduate anthropology students. It’s called

The Hunters and was made by John Marshall and BobPredmost collection was spectacular. It was destroyedalong with the Mladec specimens when Mikulov Castle Gardner. I happen to have been present when the film

was being edited and put together at the Peabody Mu-was burned by the retreating Nazi forces and their sym-pathizers. However, the Czech anthropologist Matiegka seum at Harvard. There you see some Bushmen suppos-

edly hunting a giraffe—something they would never do[1934] wrote a superb monograph on this collection.in reality; if a giraffe feels threatened, its kick is hardenough to do in a lion. If you watch carefully, you canHF: You attribute these observations ultimately to be-

havioral changes? see the Bushmen shaking their heads in disbelief as theyare being instructed off-screen to hurl spears at this biganimal. The giraffe looks at them with utter disdain asCLB: Yes, and it was the Neanderthals who changed

their behavior. Once they started cooking, the pressure the spears hit its inch-and-a-half-thick skin just enoughto annoy it. But you hear the voice of the narrator in-to maintain tooth size was removed. Now we can see

how the probable mutation effect works! In glaciated tone ponderously, ‘‘And the poison is doing its work!’’In the next frame you see the giraffe tumble like a pileEurope survival was only possible if you had some

means of thawing whatever you had hunted. Anything of huge sticks—well, what had happened was that JohnMarshall, crouched in a bush nearby, had shot it in theyou didn’t eat right away was destined to freeze.head with a high-powered rifle! The whole giraffe huntwas a complete fake, but what follows is not fake! TheHF: Are our climatic data good enough to tell us just

how cold it really was? enthusiasm and glee with which the Bushmen now areseen butchering the giraffe is the real thing. To getthrough the joints they are actually using the best ofCLB: Oh, yes. We know that the animals contemporary

with the Neanderthals were woolly mammoth, woolly German steel knives and arrowheads provided from Es-sen and designed especially for the African Bushmanrhinos, and many others associated only with very cold

climates. We also have proof of plants like arctic willow market. Now, these knives were ovate and pointed, andthe hunters are seen chopping away exactly as Pleisto-and so on. It was definitely very, very cold. However, as

I tell my students, the situation was rather like living cene hunters must have done, if we can judge from theuse-wear signs on the ancient handaxes [Keeley 1980].in Detroit today instead of Florida because Detroit is

where the jobs are—or at least were, when the automo- The battering is around the pointed edge. What we needis some experimental archaeology to replicate this ac-bile industry was still healthy. The Ice Age tundra of

Europe was not like the arctic tundra of today. It was tivity with, for example, a dead elephant’s joints.cold all right, but the amount of sunlight was sufficientto produce an efflorescence during the summer that HF: What do you think of the Movius Line, with its as-

sertion that handaxes are not found in Asia?supported a very large quantity of animals, so there wasfood on the hoof for humans in vast quantities if theymade the effort to live there. Ice Age Europe was defi- CLB: Well, presumably you don’t have handaxes in

Asia! Part of this Movius Line may be indicative of anitely the happy hunting ground, but you had to havethe technology to survive there. If you look at the ear- time line, not a geographic line. The earlier stuff is not

bifacial; the later stuff is. Typological preconceptionslier parts of the Pleistocene, you find that people didmove north every time they had the opportunity, but get in the way also, but there is more to this problem

than even that! One of Movius’s doctoral students, Karleach time a new glaciation began they disappeared.They simply could not stand the cold until they had de- Heider, had investigated Paleolithic sites in Asia and

found some perfectly acceptable handaxes on the banksveloped the technology and control of fire. One of thethings that is ubiquitous in Neanderthal sites is the of the River Kwai, of all places. I saw them. He had

them laid out for all to see in Movius’s lab. If they hadclassic D scraper, a side scraper. Francois Bordes used tosay, ‘‘Ten scrapers, fifty scrapers, a hundred scrapers— been found in Europe, nobody would have hesitated to

call them handaxes, and even today anything like thisthis is Mousterian!’’ To him it was the epitome of thedull: one doggone side scraper after another. What were found in Asia is called bifaces at best. But Movius




would not tolerate this interpretation, and Heider was HF: How do you define a classic Neanderthal?forced to abandon his thesis topic, which resulted in hisleaving archaeology completely; he became an ethnolo- CLB: The definition I use requires a modern-size brain

and Middle Pleistocene levels of robusticity. That isgist. Don Tyler of the Department of Anthropology at[the University of] Idaho has told me that such things why I refer to the skull from Jinnuishan in China as Ne-

anderthal. It has the flat cheekbones and the nonele-are found on Java and he has seen them in situ. Some ofthe handaxes found in Asia are exactly like the classic vated nasal bones of the modern Chinese, but it has the

projecting occiput and browridges that you do not findEuropean handaxes.in modern people. It has, of course, whacking big teeth!The measurements of those haven’t been published yet,HF: Bifacial tools by whatever name seem to have been

universal, but the famous D side scraper was the Nean- but I have seen them and handled the specimen. It haslingual tubercles on the incisors and huge shovels un-derthal special! What does this technological specializa-

tion tell you about evolution? like anything modern Chinese have. It’s a full-scaleclassic Neanderthal dental morphology. The third mo-lars, however, are tiny, and in modern Chinese the thirdCLB: It tells me something about the Neanderthals

themselves and provides hints for a global pattern as molars are comparatively smaller than in any other liv-ing human population. I know this because some yearswell. Survival at that latitude would have implied a re-

duced significance of skin pigmentation, as that is of ago I got a sample of modern Chinese teeth from theDepartment of Orthodontics of Beijing University.importance only in hot climates to protect us from ul-

traviolet rays. So the people who have lived longest in They were casts taken from all the incoming dental stu-dents. Among these I had trouble getting any examples,those northern latitudes are the most depigmented in

the world today. That agrees with the fossil and mate- in male or female samples, of upper and lower third mo-lars, there were so few left!rial record, which goes back to the Mousterian. You find

this depigmentation in the same latitudes in the FarEast, of course, but their depigmentation has not pro- HF: What caused the disappearance of the third molar?

Has the face changed so much that there is no room forgressed as far because they obviously haven’t lived inthose latitudes as long. that many molars?

CLB: Oh, no! It is not a question of crowding. The toothHF: Do you see the fossil and archaeological evidenceas supporting regional continuity of evolution? dimensions themselves have been reduced. Crowding is

a consequence of lack of use, that’s all. Among the Eski-mos, prior to contact, you find glorious big dental ar-CLB: Yes, of course. When I was measuring my way

through Scandinavian teeth and skulls while creating a cades, but their children, who were brought up oncanned foods, wind up with crowding and malocclu-data base of modern humans, I noticed in the museum

in Copenhagen and then again in Oslo that the skulls sion. What we have here is not genetic reduction. Infact, this can happen even within the same generation.from the medieval church burials had those familiar pe-

culiarities: odd bun-shaped occiputs, little mastoid pro- We found that it was possible to tell when an AustralianAborigine family came out of the bush and joined a mis-cesses sloping in that didn’t even reach the bottom of

the skull, sloping foreheads. And I thought, there is only sion station. The older children would have great wide,short faces and perfect occlusions while the youngerone possible ancestor for a form like that [Brace 1995b:

196, fig. 13.4]! There they were, looking just like dilute ones had narrow faces with overbite and crowdedteeth—and this in the same family [Price 1939]! TheNeanderthals [Brace n.d. (1992)].older ones had used their teeth right from the start inthe bush on partially cooked gristle; the younger onesHF: And what about their teeth?were raised on oatmeal. The difference was the stimula-tion to the erupting teeth. Oddly, even the dental pro-CLB: The teeth, of course, were greatly reduced. As is

predictable from the probable mutation effect, I have fession assumes that the overbite is ‘‘natural.’’ Well, itisn’t. Some 30 years ago when I was in California, a den-found the most rapid tooth size change in the past

10,000 years, since the development of pottery. This tal student offered to bring me some dental casts fromthe Navajo reservation at Tuba City. The children therechange took place at a rate of 1% per 1,000 years. Of

course, when you are dealing with measurements of were happy to oblige for 50 cents. I was actually inter-ested in adult teeth until I saw what he had broughtthis nature, 2% is measurement error, so the 2% change

between 2,000 years ago and today is just noise. The back! He had everything from a complete deciduousdentition to the mixed dentition up to, say, 12 yearswear in teeth is enough to allow only for an accuracy of

two-tenths of a millimeter at the most. Lucia Yaroch old—because after that age the kids ceased to be inter-ested in 50 cents!—and all those teeth were coming to[1994] did a doctoral thesis for me using a sophisticated

thin-plate spline analysis, Fred Bookstein’s [1991] spe- meet edge-to-edge and there was no wear at all. You see,what happens is this: the first permanent teeth thatcial technique, and she was able to get a perfect, unbro-

ken chain from classic Neanderthals to the modern come in are the first upper and lower molars. They arethe most regularly shaped and have an intercuspal rela-Norwegians and Britons.




tionship. They serve to keep the upper and lower jaw in HF: Did you come across anything similar in this partof the world?an aligned relationship to each other while all the other

deciduous teeth are lost and the permanent teeth comein. As these come in, they are set in position by use. CLB: Of course, because, you see, the fork was not partNow, if people at an early age hold things in their of the culture of early colonial America either. Only onemouths with their front teeth, they will wind up with fork is known in all of Massachussetts during the sec-an edge-to-edge bite, not an overbite. Other people have ond half of the 1600s, and it was owned by Governornoticed this, too. The dentist who did the tooth-wear John Winthrop (1588–1649), the New England Puritananalysis in Bob Littlewood’s dissertation [1972; see also and first governor of Massachusetts Bay Colony in 1629.Boyd in Littlewood 1972] noted in eastern New Guineathat people there had no tooth wear but edge-to-edge HF: By the middle of the 18th century forks must havebite. I also remember seeing European road gangs at been more common. Thomas Jefferson had forks; I sawlunch: they would stick a sausage between their teeth them at his house, Monticello. But come to think of it,and with one hand cut off a piece close to the mouth— looking at profile portraits of him, that projecting chinwhat I would call the ‘‘stuff-and-cut’’ school of eti- suggests edge-to-edge bite.quette. And as we all know, until mothers tell them notto children put everything into their mouths. Teeth CLB: I wouldn’t be surprised! And remember, he wasovergrow if no opposition is met during growth. Now, derided for his Frenchified table, cuisine, and the for-if we look at the actual record we find that in Europe eign table manners. A few forks are associated with thethe overbite does not appear until the end of the 18th officers’ quarters in the British military compound atcentury, and it corresponds with the appearance of the Jamestown, but there are none in colonial New Englandfork. The fork was invented in Italy and began as an and not even following the colonization of the Midwest.upper-class phenomenon. Now, I was telling K. C. Large two-pronged forks were used for carving meat,Chang about all this as we were dining in Beijing in though. The four-tined dinner fork was invented by1980, and I said, ‘‘I am willing to bet that the overbite Gennaro Spadaccini, chamberlain to King Ferdinand IIbegins in China as soon as chopsticks become part of of Naples (1467–1496), and he was richly rewarded as athe table setting!’’ After all, in classic Chinese cuisine result [Ciaglia 1991]. The king apparently felt that itto this day everything is already cut up before it even wasn’t dignified to stick his hand into a pot full of noo-gets to the table. Chang replied in that mild, dignified dles and then tilt his head back and drop the whole pilemanner of his, ‘‘Of course we do not butcher at the into his mouth. The fork enabled the king and his din-table.’’ ing companions to twirl the noodles into a ball instead.

The use of the fork reduced the need for using the teethHF: Were you able to confirm your hunch about the in conjunction with the knife, and that had conse-chopsticks? quences for the teeth themselves. But [Brace 1997:204]

Now dentists still put up a fightCLB: Oh, yes, and Chang discussed it in great detail inAgainst the thought that it mighthis Food in Chinese Culture [1977]. The overbite is first

just be everyday usefound during the Sung Dynasty, roughly around a.d.that served to produce800. There are some bronze chopsticks in the Shang,

The norm of the edge-to-edge bite.but those may have been ceremonial and for royaltyonly. Even with the Sung period one can assume that . . . . . . . . . . . .chopsticks were used only by the aristocracy. Edge-to-

But the source of our dentists’ confusionedge bite would have remained the norm for the masses,Is just a Platonic illusion:and in fact it survives all the way to the end of World

Knives and forks are the causeWar II among the peasants.Of the shape of our jaws

In their post-industrial occlusion.HF: Did you happen to get any supportive evidence forthis theory from ancient Chinese skeletons?

The relations between the English, whose upper classat least had assimilated the use of the fork in the 18thCLB: Yes, indeed I did! In the Shanghai Museum of Nat-

ural History I saw the pickled remains of—as I tell my century, and colonial Americans were rather frostyfrom the time of the Revolution to at least the War ofstudents—a Sung Dynasty graduate student! This fel-

low was an aristocratic young man, an official, who 1812, when the British even burned the White House.Therefore, cultural relations were also frosty. Aspiringdied, as the label explained, around the time he would

have sat for the imperial examinations. Well, there he Americans sent their offspring to Scotland, to Edin-burgh, for an education rather than to London, Oxford,was, in a vat floating in a pickling fluid with his mouth

wide open and looking positively revolting. But there it or Cambridge. We know that the Scots did not useforks. Boswell, for example, tells us that Samuel John-was: the deep overbite of the modern Chinese! As a

member of the upper class he would have had their son ate with his fingers. So it took a little longer for thefork to appear on American tables. Eventually, forks dideating habits, which would have included chopsticks.




appear in America, of course, but we were all brought tions of English, Irish, and Germans made an almost ex-act match to the records of the workhouse. But for meup to use them rather like spoons—I remember my

mother saying, ‘‘Don’t stab your food!’’ You can spot the most gratifying of all was the 15 specimens withcomplete jaws and teeth. Ten of them (two-thirds) haveAmericans anywhere in the world even today because

of what I call the ‘‘American shuffle.’’ Emily Post called an edge-to-edge bite but really very little wear. I’ve gonethrough collection after collection of late-19th- or early-it ‘‘zig-zag’’ eating, but I didn’t find that out until much

later [Visser 1991]. Food is held down with a fork in the 20th-century hospitals and medical schools, and out ofthe thousands of individuals I have examined you canleft hand and cut with the knife in the right. Then

the knife is put down and the fork is shuffled over to find scarcely a single one with edge-to-edge bite.the right hand and turned over in spoon-fashion to bringa bit up to the mouth. Then the fork is shuffled back HF: Did your observation hold true for the Australianagain to the left hand and the knife is picked up once dental evidence as well?more by the right, and the whole awkward thing beginsall over again. By contrast, Europeans use them the way CLB: Splendid edge-to-edge bite there, of course, butthey were intended to be used: they hold the food down after full acculturation the overbite quickly becomeswith the fork, cut off a piece with the knife and use the the norm. One of the things that struck me about Aus-fork to pop that piece of speared food into their mouths. tralia was the north-south gradient of tooth size. ThisSo, up to the early 19th century Americans ought to nicely parallels the entrance of food-processing tech-have had edge-to-edge bites, while later on the overbite niques and genes, too, from the north. I did a detailedought to have begun. My friend James Deetz, formerly study of this issue [Brace 1980, Brace and Ryan 1980,the archaeologist of Plimouth Plantation, in Massachu- Brace and Vitzthum 1984, Brace, Shao, and Zhang 1984,setts, tried to provide me with the evidence to prove Brace, Hunter, and Tracer 1991] which is most suc-this prediction. (K. C. Chang, Jim Deetz, and I were all cinctly summarized as follows [Brace 1980:161]:graduate students at the same time.) Deetz had finally

Australian teeth show a grade,got permission from the Quakers to dig up a cemeteryFrom the north where reduction’s displayed,in Rhode Island and examine the material in his labora-

To those in the southtory. As fate would have it, the good Quakers who hadWhere the typical mouthbeen buried in their meeting-house plot had all lost

Has a Pleistocene dental arcade.their teeth long before death, and it was impossible totell what kind of bite they might have had earlier in life.

HP: What do you make of some of spectacular new fos-Years later, I hit the jackpot! I happened to share a taxisil finds, such as Atapuerca (Arsuaga et al. 1993, Bermu-to the airport with Joyce Siriani, who was chair atdez de Castro 1993)?SUNY [State University of New York], Buffalo, and like

me has had a long involvement with dental anthropol-CLB: Atapuerca looks like a modern Eastern Europeanogy. I was chatting with her about wear and occlusionraised to Neanderthal dimensions! I have always feltand mentioned my prediction that the edge-to-edge bitethat once we find a Neanderthal in Eastern Europe itshould have survived in America well into the 19th cen-will not look like the Western ones, that it will have atury but that I couldn’t find a sample to test this. Joycelong, high-bridged nose, a flat occiput, and big mastoidsaid, ‘‘I’ve got one in Buffalo!’’ It came from an un-processes. And what do we find in Atapuerca? Exactlymarked cemetery in Rochester, in Highland Park rightthat! And of course, to this day we find cranial charac-across the street from the dental school. As it turnedteristics among the Basques at the extreme western endout, there had been a workhouse, an insane asylum, andof Europe, for example, that echo Central and Easterna prison in the area that is now the park; they were allEuropean characteristics, and it is not impossible thatbuilt just after the completion of the Erie Canal in 1825,this goes way back in time. Atapuerca has all the dentaland the burial ground had clearly been used for thosemeasurements and general robusticity of a Neander-who had died in those institutions. It’s easy enough tothal, but it has the cranial proportions and emphases ofimagine the scenario: laborers on the canal, out of worka modern Eastern European, which is what you wouldwhen the thing was finished, went to the workhouse;expect.mindless make-work drove them into the insane asy-

lum. The more desperate ones ran afoul of the law andHF: Where does that leave us with regard to the elusivewound up in prison, and all the unclaimed dead got bur-origin of anatomically modern Homo sapiens?ied in the adjacent potter’s field. From the poorhouse

records we have a pretty good idea of where the peopleCLB: Well, what is ‘‘anatomically modern Homo sa-came from. All this was rediscovered by accident whenpiens’’?the recreational facilities in Highland Park were being

upgraded. The skeletons were housed pro-tem at SUNY,HF: I wish I knew!Buffalo, and provided me with a splendid sample from

just the time period I was looking for. So I measured upthe crania and ran my results against the worldwide CLB: The modern human form is, in my view, the re-

duced version of our immediate nonmodern ancestors.data sample I had compiled in Ann Arbor. The propor-




The process of converting that nonmodern ancestor— to get genetic material from people in rural and distantareas of northern Europe and found a diversity far largerwhat I am perfectly happy to call a Neanderthal wher-

ever it is found—into a ‘‘modern’’ form is something I than asserted earlier. The basic point is that certain por-tions of the DNA chain among Africans have reachedhave referred to as ‘‘evolution by entropy’’ [Brace

1996a]. From an evolutionary perspective there is no equilibrium in ways that are not found in the rest of theworld. In fact, Kidd is no longer nailing dates to thedifficulty with this.wall. It is clear that geneticists have no idea at all howlong it takes for such equilibrium to be achieved in aHF: I suppose you would see the Eve theory, with its

assumption that mtDNA holds the key to understand- population. So when assertions are made about x num-ber of years, it can’t be done with teeth, skin color, oring human evolution, as another version of typological

essentialism? any other variation found in modern human popula-tions, nor can it be done with the variation observed infossil populations. You can’t tell from an mtDNA mole-CLB: Yes, you might say that. The assumptions on

which the Eve theory is based are being steadily under- cule or a Y chromosome or the disequilibrium on the12th chromosome whether that individual has had thismined. Not only is mtDNA found to be contaminated

with paternal DNA after all [Hoeh, Blakely, and Brown feature for 10,000, 100,000, or any other number ofyears. All of this has no relationship whatsoever to1991, Gyllensten et al. 1991] but when recently a real

effort was made to find different types of Africans in Af- what we observe as modern human morphology.rica—rather than use American Blacks as they had donebefore—a picture of extreme genetic variety was found HF: So what does this say about our belief that human-

ity originated in Africa?which is unmatched anywhere in the world, and thisvariety includes not only mtDNA but also the Y-chro-mosomal and nuclear DNA, which show regional CLB: The question ‘‘Out of Africa?’’ is very real, but the

question ‘‘How did modern human morphology comevariation as well [Spurdle, Hammer, and Jenkins 1994,Hammer and Horai 1995, Whitfield, Sulston, and about?’’ is a totally different question. For example, if

we compare the skin pigmentation of Europeans withGoodfellow 1995, Paabo 1995, Ayala 1996]. At the mo-ment, the ‘‘coalescence’’ times, or indications of a di- that of East Asians we do have an indication that it

takes at least 200,000 years for this process of depig-vergence, are so different between what is projectedfrom the diversity visible in nuclear DNA, mitochon- mentation to occur—in Asia people have lived there

continuously for at least that long. So obviously I can’tdrial DNA, and DNA on the nonrecombining part of theY chromosome that it almost looks as though anything buy the idea that anatomically modern humans came

out of Africa 200,000 years ago. Another test is Austra-goes. Therefore, a time regression of any sort is now ex-tremely problematic. There also are some very exciting lia: we know they came to that continent at least 50,000

years ago, and we do find a skin-color gradient runningnew findings concerning the disequilibrium haplotypesof the CD4 gene locus on the 12th chromosome. Over from the north to the south, getting lighter the farther

from the equator you go. But because food-processingtime, the elements of which the haplotypes are com-posed should become distributed at random vis-a-vis technology got to the south later, those lighter types

have the biggest teeth in the world, basically the sameeach other, but recently a paper came out of Ken Kidd’s(Kidd et al. 1996) genetics lab at Yale showing that, out- size as in the Pleistocene. But depigmentation has not

progressed with them as far as it has for people of a com-side Africa, those elements largely remain linked in alimited set of haplotypes [Tishkoff et al. 1996]. parable distance from the equator to the north around

the Mediterranean or the southern parts of China. Fiftythousand years is not enough to produce advanced de-HF: One is almost tempted to ask whether the propo-

nents of the Eve theory know this! pigmentation. Whether it’s dental reduction or pigmentreduction, we need a lot more time.

CLB: Exactly! Recently I talked to Kidd about this andwas told that genetic equilibrium has been found in a HF: Are you opposed to the mtDNA research, then?wide variety of haplotype forms not seen outside of Af-rica. However, at the same time it turns out that the CLB: No! But the various approaches—the archaeologi-

cal, the evolutionary dynamics, the fossil stratigraphy,genetic picture outside of Africa is also far more compli-cated than had ever been imagined. Yet the Eve-theory and the genetics—all have to agree in the long run, just

as we needed the dates at Zhoukoudian to agree withproponents have not made the kind of effort needed toget truly representative samples. On April 10, 1995, the evolutionary dynamics required to produce a Nean-

derthal morphology such as we have at Jinniushan outSvante Paabo from the University of Munich gave a pre-sentation to the Department of Human Genetics here of a Homo erectus. Once the dates were shown to be

400,000 b.p. not 230,000 b.p. as we used to think, theat the [University of Michigan] Medical School. The ti-tle was ‘‘Molecular Genetic Approaches to the Study of picture made sense, because the original dates had re-

quired the evolution of a Neanderthal stage withinHuman History.’’ He provided evidence for the geneticdiversity in human populations as seen in mtDNA and 30,000 years or so, which hadn’t seemed possible. Now

we are all heaving a collective sigh of relief, as it were.aspects of nuclear DNA as well. Paabo made the effort




Clark Howell was quoted in the New York Times (May HF: What is the difference between ‘‘race’’ and ‘‘cline’’?2, 1996) as saying that this was ‘‘terrific,’’ and evenRightmire conceded that ‘‘gradual evolution seems CLB: A cline is a gradient, a geographic distribution, in

a single trait. So you can deal with variation in skinmore likely than overlap.’’ This sort of logical agree-ment is what is needed with regard to the evolution of color as a gradient. If you draw an artificial line of sepa-

ration, such as the concept of race does, between thisanatomically modern humans also.or that group, you lose the picture of the gradient. I hadbecome interested in this same issue through a paper byHF: All of these observations about the evolutionary dy-E. O. Wilson and W. Brown [Wilson and Brown 1953] onnamics involved in bones, teeth, skin, and cranial mor-the nonexistence of the subspecies. They demonstratedphology seem to help uncover notions about race whichthat gradients of variation have to be understood inAshley Montagu called ‘‘the dangerous myth’’ [Mon-terms of the individual trait’s distribution and that atagu 1942, 1997; see Brace 1997a, b].trait will go this way or the other independently of pop-ulation boundaries, so that, if you start giving differentCLB: That’s exactly right. When you study the evolu- names to that trait as it appears in different populations,tion of modern human variation you can only do that you have lost the whole phenomenon. Butterflies are anpiece by piece, not by linking it to a generality like example they gave; other examples are the many types‘‘race.’’ The metabolic pathway for skin color to change of seagulls around the world, or wolves and the ordinaryis very complicated and takes a long time, as we have leopard frog. The gulls and the frogs have huge geo-seen. And when we get to ‘‘intelligence,’’ any attempt graphic distributions, and when you get to, say, Floridato associate that with different cultures when agricul- the leopard frog is still a leopard frog but cannot mateture has only been around for 10,000 years or so is with the Canadian leopard frog. Same with seagulls.plainly absurd. All cultures have only a time depth of a There is actually no break along the line. Similarly,few thousand years at most. Add to that the fact that with humans—even though they can interbreed—theall human beings had the same lifeways from 10,000 observed differences are a question of gradient. In theyears ago all the way back to as long as the genus exists. book I am currently writing I make the point thatHence, more than 2 million years of the same selective Marco Polo had no concept of race, and neither did He-forces have operated on the whole genus, so how can rodotus or the world-traveling medieval Islamic geogra-one expect there to be different intellectual capacities pher Ibn Batutah; yet, they all had seen the whole spec-in different living populations? trum of human variation.

HF: In 1994 you were given the honor of presenting to HF: You collaborated on some major publications withAshley Montagu the Charles Darwin Award for Life- Montagu [Brace 1962b, 1964; Brace and Montagu 1965;time Achievement [Brace 1994a]. How did your collabo- Brace and Livingstone 1975].ration with Montagu come about?

CLB: Yes, in fact some of my ideas on race and humanevolution came into print only because of Montagu’sCLB: Now, Ashley Montagu was, as he remains, the

best-known biological anthropologist in the world. In collaboration. It was a volatile issue back then and al-most prevented me from getting tenure! When our bookfact, at the Charles Darwin Award presentation in Den-

ver, I described him as the world’s one and only free- went to Macmillan [Brace and Montagu 1965], they gotvarious opinions and the editor backed off. So, in latelance physical anthropologist [see Lieberman, Lyons,

and Lyons 1995]. He has made his living talking on ra- 1964 Montagu marched into the building, brushed bythe various secretaries, demanded to see the president,dio and television and lecturing to university audiences,

and he has produced an avalanche of books and articles who wasn’t there, and then strode into the vice presi-dent’s office and, pounding his fist on the table, de-on all kinds of areas where human biology intersects

with the social world. He’s a transplanted Englishman manded to know how our book could be questioned.His international reputation was such that the bookwho came to America well before World War II and

made his mark especially with his development of the was published without so much as a comma altered.theme that ‘‘race’’ is a myth of our own making and hasno foundation in biology. Our paths converged when we HF: It is an interesting historical coincidence that 1964

was the same year the United States disallowed any dis-were at Santa Barbara and looking to buy a house. Inconversation with the owner of one of the places we crimination against blacks with regard to their civil

right to vote. What was so offensive in your book, dolooked at, we were told that Montagu was moving toSanta Barbara. The idea that he might be moving there you know?was something that I thought might give a real boost tothe fledgling department I had only just joined myself. CLB: Indeed! You see, Americans have a very strong

concept of race, and it exerts a powerful influence. ItI wrote to Montagu and invited him to give a talk at ourAnthropology Club. He had become interested in this wasn’t so much the race issue that offended the ex-

perts Macmillan consulted—that, too, of course—asissue of human variation through Julian Huxley, whohad invented the concept of ‘‘cline’’ [Huxley 1938]. the view that the Neanderthals are asserted to be our




direct ancestors. That bothered the reviewers a lot. Cu- tropical Homo erectus to cold-climate Neanderthalsand Eskimos [Ruff 1993, Ruff et al. 1993, Ruff and Trin-riously, this offended some of our colleagues because

our evolutionary scenario was supposedly critical of liv- kaus 1996, Ruff and Walker 1993], but that’s not a rec-ord of cumulative change. What is missing and what,ing populations!therefore, is needed is a concern for the process and dy-namics that have produced the changes we can actuallyHF: How does this follow?see in the fossil record.

CLB: Paleoanthropological orthodoxy states that theNeanderthals were an evolutionary sideline that went HF: How does this lack of concern for dynamics andextinct and that they were nasty, brutish, and so on. process show?This goes back to Marcellin Boule. So if we see Nean-derthal traits in recent populations—well!

CLB: Well, what one typically gets is declarations suchas the one by Phil Rightmire in his tome on The Evolu-HF: That’s really funny, especially in view of your out-tion of Homo erectus [1990:205–6] that issues such asspoken views on the question of intelligence and yourwhat that taxon might have evolved from or evolvedrecent vocal critique [1996b] of Herrnstein and Mur-into ‘‘are large questions’’ that ‘‘would be out of place’’ray’s Bell Curve [1994]!in his book! Instead, one gets page after page devotedto ‘‘specializations’’ such as details of the postglenoidCLB: That is another one of those slippy-slidy slopeseminence, the angular torus, the thickness of the infe-like the one the creationist scientists are on. Many ofrior border of the tympanic bone, and on and on, with-the IQ and adoption study data they published, espe-out a single word of what the function of those ‘‘special-cially the IQ data from Africa, came from studies origi-izations’’ might be. Then we have Chris Stringernally financed by the Pioneer Fund.4 One of those re-recycling Marcellin Boule’s old idea of ‘‘Neanderthalports [Owen 1992] was done for the express purpose ofthe Unspeakable,’’ as I recently put it [Brace 1995b:217].supporting apartheid in South Africa. The subjects wereThe dwellings constructed by those venerable fossilschildren whose English was so poor that the instruc-are allowed status as ‘‘nests’’ [Stringer and Gambletions were done in sign language of sorts; they chose to1993:207]. That puts them effectively more on the men-ignore the IQ results from English-speaking blacks intal level of beavers, or pigeons even, than of the den-Soweto. Herrnstein and Murray generalized from thosedwelling wolves which were their contemporaries. Andstudies for all of Africa as well as for American blacks!then Ian Tattersall has opined, in all seriousness, thatFor my current book I went through all of those datahuman evolution is now over because genetic isolationand found out how they were obtained, since I deal withis no longer possible [in Wade 1995]! He evidently takesthe issue of intelligence in depth. This book, by theit as axiomatic that isolation is essential for evolution-way, has the working title Race Is a Four-Letter Word!ary change. What bothers me is that the main activityof paleoanthropologists seems to be focused on this orHF: What, in your view, is currently needed in anthro-that piece of minutia to justify the creation of yet an-pological research?other esoteric name, to which my friend I. DoolittleWright5 has reacted as follows [Brace 1993:163]:CLB: By way of a preamble let me say that it has always

struck me as curious that the ones who are the most Consider the volatile splitter,reluctant to interpret the hominid fossils as a record of Excited and all of a twitter,the course taken in human evolution are the very stu- At the fun in the game,dents of those fossils—that is, the paleoanthropologists Of adding a namethemselves! And this extends all the way through to For each little variant critter.those biological anthropologists who deal with themeaning of biological variation in living populations as There is nothing new in my complaint. I said it all awell. There is absolutely no effort to try and tie the evi- generation ago in the pages of current anthropology.dence for change in particular aspects of morphology to The only difference is that then my queries were di-changes in the intensity of specific selective forces. I rected to the previous generation—the members ofdon’t mean to be completely negative. Chris Ruff, for Homo ‘‘oldguardensis,’’ as my teacher W. W. Howellsexample, has done a wonderful job of showing the tie archly called them [Howells 1964:26]. Now we arebetween limb proportions and climate all the way from seeing the same views without modification in the

younger generation. I have labeled this ‘‘the great leapbackwards’’ [Brace 1988, 1989]. If that demonstrates in-4. The Pioneer Fund was founded in 1937 by supporters of Hitler’stellectual stasis, stasis is what we have got. We havepolicies against ethnic minorities and the handicapped, and it pro-

vided money for introducing Nazi propaganda into the United data in mindless quantity. What we need is the applica-States [Kuhl 1994:49]. Initial funding was provided by the ‘‘eccen- tion of a modicum of processual thinking.tric Massachusetts textile heir’’ Wicliffe Draper [Miller 1995:172].Recently ‘‘the pro-Nazi Pioneer Fund’’ has been described as ‘‘asmall foundation that has funded the work of every major academicadvocate of racism, eugenics, and fascism since the late 1930’s’’ 5. I. Doolittle Wright is C. L. Brace’s alter ego and responsible for

the teasing limericks on paleoanthropological topics.[Rosenthal 1995:44].




brace, c. l. 1962. ‘‘Cultural factors in the evolution of the hu-HF: Your work seems to me to be characterized by en-man dentition,’’ in Culture and the evolution of man. Editedthusiasm and frustration in almost equal measure, andby M. F. A. Montagu, pp. 343–54. New York: Oxford Univer-

this what makes reading your books and papers such sity Press.fun and at times so bracing—if you’ll forgive the pun! I ———. 1963. Structural reduction in evolution. American Natu-

ralist 97:39–49.am wondering if you have recently heard what your———. 1964a. The fate of the ‘‘classic’’ Neanderthals: A consider-good friend I. Doolittle Wright has to say about the cur-

ation of hominid catastrophism. current anthropology 5:3–rent state of paleoanthropology. 43.———. 1964b. ‘‘A non-racial approach towards the understand-

ing of human diversity,’’ in The concept of race. Edited byCLB: Yes, indeed! I told him about this CA interview,M. F. A. Montagu, pp. 103–52. Glencoe: Free Press.and he favored me with this reply:

———. 1973. Back to the bow and arrow: Review of The tendercarnivore and the sacred game, by Paul Shepard (New York:As this age gives way to the next,Scribner, 1973). Natural History 82 (9):110–17.The fossils still count as our text; ———. 1977. ‘‘Occlusion to the anthropological eye,’’ in The biol-

But what passes for theory ogy of occlusal development. Edited by J. A. McNamara Jr.Is backward and bleary Centre for Human Growth and Development Craniofacial

Growth Series Monograph 7.And read by the partly perplexed.———. 1979. Krapina, ‘‘classic’’ Neanderthals, and the evolution

of the European face. Journal of Human Evolution 8:527–50.For most of them work to defend———. 1980. On Australian tooth-size clines and the death of aThe ancient catastrophist trend

stereotype. current anthropology 21:141–64.And try to protect ———. 1981. Tales of phylogenetic woods: The evolution and sig-When they ought to reject— nificance of phylogenetic trees. American Journal of Physical

A millennium now at an end. Anthropology 56:411–29.———. 1985. Punctuationalism, cladistics, and the legacy of me-

It’s a suitable time to renew dieval Neoplatonism. American Journal of Physical Anthropol-ogy 63:140–41.The selectionist stance of the few;

———. 1988. Punctuationalism, cladistics, and the legacy ofIt’s my hope that at lastmedieval Neoplatonism. Human Evolution 3:121–38.We will look at the past ———. 1989. Medieval thinking and the paradigms of paleo-

With the neo-Darwinian view. anthropology. American Anthropologist 91:442–46.———. 1991. On the eve of the problem, or Will paleoanthropol-

My anthropological theme ogy ever escape from the Middle Ages? Invited lecture pre-To promote an adaptive scheme sented at Southern Methodist University, Dallas, Tex., Febru-

ary 25.—And thusly equipped———. 1993. The creation of specific hominid names: Gloria inTo interpret our script—

excelsis Deo? or ego? or praxis? Human Evolution 8:151–66.Is not an impossible dream. ———. 1994a. Presentation of the Charles Darwin Award forLifetime Achievement to Ashley Montagu. American Journalof Physical Anthropology 95:456–57.References Cited ———. 1994b. Review of: Species, species concepts, and primateevolution, by W. H. Kimbel and L. B. Martin (New York: Ple-

andrews, r. c. 1948. Meet your ancestors. London: John num Press, 1993). American Scientist 8:484–86.Long. ———. 1995a. Biocultural interaction and the mechanism of mo-

ayala, f. j. 1996. Gene lineages and human evolution. Science saic evolution in the emergence of ‘‘modern’’ morphology.268:1141. American Anthropologist 97:1–11.

arsuaga, j.- l. , i. martinez, a. gracia, j.-m. car - ———. 1995b. 5th edition. The stages of human evolution. Engle-retero, and e. carbonell. 1993. Three new human wood Cliffs: Prentice-Hall.skulls from the Sima de los Huesos Middle Pleistocene site in ———. 1996a. ‘‘Modern human origins and the dynamics of re-Sierra de Atapuerca, Spain. Nature 362:534–37. gional continuity,’’ in Prehistoric Mongoloid dispersals. Edited

baldwin, w. w., m. a. k irkish, and a. l. koch. 1994. by T. Akazawa and E. J. Szathmary. New York: Oxford Univer-A change in a single gene of Salmonella typhimurium can dra- sity Press.matically change its buoyant density. Journal of Bacteriology ———. 1996b. Review of: The bell curve, by R. Herrnstein and176:5001–4. C. Murray (New York: Free Press, 1994). current anthropol-

bermudez de castro, j. m. 1993. The Atapuerca dental re- ogy 37:S.156–61.mains: New evidence (1987–1991 excavations) and interpreta- ———. 1997a. ‘‘Foreword,’’ in Man’s most dangerous myth: Thetions. Journal of Human Evolution 24:339–71. fallacy of race, revised edition, by A. Montagu. New York:

bookstein, f. l. 1991. Morphometric tools for landmark Macmillan.data: Geometry and biology. Cambridge: Cambridge Univer- ———. 1997b. Race is a four-letter word. MS.sity Press. ———. n.d. (1992). ‘‘Modern human origins: Narrow focus or

bordes, f. 1958. ‘‘Le passage du Paleolithique moyen au Paleo- broad spectrum?’’ in Conceptual issues in modern human ori-lithique superieur,’’ in Hundert Jahre Neanderthaler. Edited by gins. Edited by A. Clark and C. M. Willermet. New York: Al-G. H. R. von Koenigswald. Utrecht: Kemink en Zoon. dine de Gruyter. In press.

boule, m. 1913. L’homme fossile de La Chapelle-aux-Saints. brace, c. l., and f. b. l iv ingstone. 1975. ‘‘On creepingParis: Masson. Jensenism,’’ in Race and IQ. Edited by A. Montagu. New

———. 1921. Les hommes fossiles: Elements de paleontologie hu- York: Oxford University Press. (Reprinted from Race and intel-maine. Paris: Masson. ligence. Edited by C. L. Brace, G. R. Gamble, and J. T. Bond.

———. 1929. Le ‘‘Sinanthropus.’’ L’Anthropologie 39:455–60. Washington, D.C.: American Anthropological Association,———. 1937. Le Sinanthrope. L’Anthropologie 47:1–22. 1971.)boyd, r. c. 1972. ‘‘An odontometric and observational assess- brace, c. l., p. e. mahler, and r. b. rosen. 1973.

ment of dentition,’’ in Physical anthropology of the Eastern Tooth measurement and the rejection of the taxon ‘‘Homo ha-Highlands of New Guinea, by R. A. Littlewood, appendix 4. Se- bilis.’’ Yearbook of Physical Anthropology 16:50–68.attle: University of Washington Press.




brace, c. l., and a. montagu. 1965. Man’s evolution: An keeley, l. h. 1980. Experimental determination of stone tooluses. Chicago: University of Chicago Press.introduction to physical anthropology. New York: Macmillan.

brace, c. l., and a. s. ryan. 1980. Sexual dimorphism ke ith, a. 1925a. The fossil anthropoid ape from Taung. Nature115:234–35.and human tooth size differences. Journal of Human Evolution

9:417–35. ———. 1925b. The Taung skull. Nature 116:11.———. 1925c. The Taung skull. Nature 116:462–63.brace, c. l., x -q. shao, and z-b. zhang. 1984. ‘‘Prehis-

toric and modern tooth size in China,’’ in The origins of mod- ———. 1925d (1915). 2d edition. The antiquity of man. 2 vols.London: Williams and Norgate.ern humans: A world survey of the fossil evidence. Edited by

F. Spencer and F. Smith. New York: Alan Liss. ———. 1950. An autobiography. London: Watts.kidd, k. k., j. r. k idd, s.a. pakst is, c. m. ti shkoff,brace, c. l., s. l. smith, and k. d. hunt. 1991. ‘‘What

big teeth you had, Grandma! Human tooth size past and pres- c. m. cast igl ione, and g. s irugo. 1996. Use of link-age disequilibrium to infer population histories. American Jour-ent,’’ in Advances in dental anthropology. Edited by M. A. Kel-

ley and C. S. Larsen. New York: Wiley-Liss. nal of Physical Anthropology, suppl. 22:138.k imbel, w. h., and l. b. martin. Editor. 1993. Species,brace, c. l., p. tracer, and k. d. hunt. 1991. Human

craniofacial form as evidence for the peopling of the Pacific. species concepts, and primate evolution. New York: PlenumPress.Indo-Pacific Prehistory Association Bulletin 11:247–69.

brace, c. l., and v. v itzthum. 1984. Human tooth size koch, a. l. 1983. The protein burden of lac operon products.Journal of Molecular Evolution 19:455–62.at Mesolithic, Neolithic, and modern levels at Niah Cave, Sara-

wak: Comparisons with other Asian populations. Sarawak Mu- kuhl, s. 1994. The Nazi connection: Eugenics, American rac-ism, and German National Socialism. New York: Oxford Uni-seum Journal 33 (54):75–82.

brace, e. 1894. The life of Charles Loring Brace: Chiefly told versity Press.lewis, d. 1976. Hokule’a follows the stars to Tahiti. Nationalin his own letters, edited by his daughter. New York:

Scribner. Geographic 150:512–37.l ieberman, leonard, andrew lyons, and harrietbrain, c. k. 1981. The hunters or the hunted? An introduction

of African cave taphonomy. Chicago: University of Chicago lyons. 1995. An interview with Ashley Montagu. currentanthropology 36:835–44.Press.

chang, k. c. Editor. 1977. Food in Chinese culture: Anthropo- l ittlewood, r. a. 1972. Physical anthropology of the East-ern Highlands of New Guinea. Seattle: University of Washing-logical and historical perspectives. New Haven: Yale Univer-

sity Press. ton Press.matiegka, j. 1993. Homo predmostensis, fosilni, clovek zciaglia, p. 1991. Pasta. Smithsonian 22(4).

clark, w. e. le gros. 1955. The fossil evidence for human Predmosti na Morave. Prague: Nakladem Ceske Akademie Veda Umeni.evolution: An introduction to the study of paleoanthropology.

Chicago: University of Chicago Press. meggitt, m. 1957. Notes on the vegetable foods of the Walbiriof central Australia. Oceania 28:143–45.darwin, c. 1857. Abstract of a letter from C. Darwin Esq., to

Prof. Asa Gray, Boston, dated Down, Sept. 5, 1857. Journal of ———. 1962. Desert people: A study of the Walbiri Aboriginesof Central Australia. Sydney: Angus and Robertson.the Proceedings of the Linnaean Society, Zoology 3:50–53.

———. 1859. On the origin of species by means of natural selec- ———. 1964. ‘‘Aboriginal food-gatherers of tropical Australia.’’The ecology of man in the tropical environment: Internationaltion, or The preservation of the favoured races in the struggle

for life. London: John Murray. Union for the Conservation of Nature and Natural Resources,9th Technical Meeting, Proceedings and Papers (Publication———. 1868. The variation of animals and plants under domesti-

cation. 2 vols. London: John Murray. Series 4). Geneva: Morges.miller, a. 1995. ‘‘Professors of hate,’’ in The Bell Curve de-———. 1871. The descent of man and selection in relation to

sex. 2 vols. London: John Murray. bate. Edited by R. Jacoby and N. Glauberman. New York:Times Books.ferrie, h. 1995. An interview with K. C. Chang. current an-

thropology 36:307–25. mitchell, t. l. 1839. Three expeditions into the interior ofeastern Australia with descriptions of the recently explored re-f inney, b. r. 1977. Voyaging canoes and the settlement of

Polynesia. Science 196:1277–85. gion of Australia Felix, and the present colony of New SouthWales. 2 vols. London: T. and W. Boone.gish, d. t. 1979. 3d edition. Evolution: The fossils say no! El

Cajon, Calif.: Creation-Life Publishers. montagu, a. 1942. Man’s most dangerous myth: The fallacyof race. New York: Columbia University Press.gould, s. j., and r. c. lewontin. 1979. The spandrels

of San Marco and the Panglossian paradigm: A critique of the ———. 1997. Revised edition. Man’s most dangerous myth: Thefallacy of race. New York: Columbia University Press. Inadaptationist programme. Proceedings of the Royal Society of

London B 295:581–98. press.novick, a., and m. weiner. 1957. Enzyme induction as angyllensten, u., d. wharton, a. josefsson, and

a. c. wilson. 1991. Paternal inheritance of mitochondrial all-or-none phenomenon. Proceedings of the National Acad-emy of Sciences (U.S.A.) 43:553–66.DNA in mice. Nature 352:255–57.

hammer, m. f., and s. horai. 1995. Y chromosomal DNA olmsted, f. l. 1861. The cotton kingdom: A traveller’s obser-vations on cotton and slavery in the slave states. 2 vols. Newvariation and the peopling of Japan. American Journal of Hu-

man Genetics 56:951–62. York: Mason Brothers.owen, k. 1992. The suitability of Raven’s Standard Progressiveherrnste in, r., and c. murray. 1994. The bell curve.

New York: Free Press. Matrices for various groups in South Africa. Personality and In-dividual Differences 13:149–59.hoeh, w. r., k. h. blakely, and w. m. brown. 1991.

Heteroplasmy suggests limited biparental inheritance of Myti- paabo, s. 1995. Y chromosome and the origin of all of us(men). Science 268:1141–42.lus mitochondrial DNA. Science 251:1488–90.

howells, w. w. 1964. Comment on: The fate of the ‘‘classic’’ price, w. a. 1939. Nutrition and physical degeneration: Acomparison of primitive and modern diets and their effects.Neanderthals, by C. L. Brace. current anthropology 5:26–

27. New York: Paul B. Hoeber.prov ine, w. b. 1986. Sewall Wright and evolutionary biology.hrdli cka, a. 1927. The Neanderthal phase of man. Journal of

the Royal Anthropological Institute 57:249–69. Chicago: University of Chicago Press.rightmire, p. g. 1990. The evolution of Homo erectus: Com-———. 1930. The skeletal remains of early man. Smithsonian

Miscellaneous Collections 83. parative anatomical studies of an extinct human species. Cam-bridge: Cambridge University Press.huxley, a. 1937. Ends and means. New York: Harper.

huxley, j. 1938. Clines: An auxiliary taxonomic principle. Na- rink, w. j., h. p. schwarcz, f. h. smith, and j. radovi-cic. 1995. ESR dates for Krapina hominids. Nature 378:24.ture 142:219–20.




rosenthal, s. j. 1995. The Pioneer Fund: Financier of fascist ———. 1947a. Facts and speculations concerning the origin ofHomo sapiens. American Anthropologist 49:187–203.research. American Behavioral Scientist 39:444–61.

ruff, c. b. 1993. Climatic adaptation and hominid evolution: ———. 1947b. The trend of human evolution. Evolution 1:221–36.The thermoregulatory imperative. Evolutionary Anthropology

2:53–60. whitfield, l. s. , j. e. sulston, and p. goodfellow.1995. Sequence variation of the human Y chromosome. Nature———. 1994. Morphological adaptation to climate in modern and

fossil hominids. Yearbook of Physical Anthropology 37:65– 378:379–80.wilson, e. o. , and w. l. brown. 1953. The subspecies107.

ruff, c. b., and e. trinkaus. 1996. Body mass in Pleisto- concept and its taxonomic application. Systematic Zoology2(3):97–111.cene Homo. American Journal of Physical Anthropology,

suppl. 22:205–6. wright, s. 1931. Evolution in Mendelian populations. Genet-ics 16:97–159.ruff, c. b., e. trinkaus, a. walker, and c. s.

larsen. 1993. Postcranial robusticity in Homo. 1. Temporal ———. 1951. Fisher and Ford on the ‘‘Sewall Wright effect.’’American Scientist 39:452–58.trends and mechanical interpretation. American Journal of

Physical Anthropology 91:21–53. yaroch, l. a. 1994. Characterization of Neandertal cranialshape using the method of thin-plate splines. Ann Arbor: Uni-ruff, c. b., and a. walker. 1993. ‘‘Body size and body

shape,’’ in The Nariokotome Homo erectus skeleton. Edited by versity Microfilms.A. Walker and R. E. Leakey. Cambridge: Harvard UniversityPress.

ryan, a. s. 1980. Anterior dental microwear in hominid evolu-tion: Comparisons with human and nonhuman primates. Ann Sampling Strategies for theArbor: University Microfilms,

ryan, a. s., and d. c. johanson. 1989. Anterior dental Collection of Cultural Data: Anmicrowear in Australopithecus afarensis: Comparisons withhuman and nonhuman primates. Journal of Human Evolution Extension of Boas’s Answer to18:235–68.

schuman, e. l., and c. l. brace. 1954. Metric and mor- Galton’s Problem1phologic variation in the dentition of the Liberian chimpanzee:Comparisons with anthropoid and human dentitions. HumanBiology 26:239–67.

schwalbe, g. 1906. Studien zur Vorgeschichte des Menschen. w. penn handwerker and danielle f.Stuttgart: Schweizerbartsche Verlagsbuchhandlung (E. Nagele).wozniakschwarcz, h. p., w. j. r ink, f. h. smith, and j. rado-

vic ic. 1966. ESR dates for the Neanderthal site of Krapina, Department of Anthropology (U-176), University ofCroatia. American Journal of Physical Anthropology, suppl. Connecticut, Storrs, Conn. 06269-2176, U.S.A. 1 v 9722:211.

shea, b. t. 1989. Heterochrony in human evolution: The case Social and behavioral scientists conventionally judgeof neotony reconsidered. Yearbook of Physical Anthropology the reliability, validity, and generalizability of findings32:69–101.

by reference to criteria set by classical statistical theory.shen, g. 1992. Progress in the dating of Paleolithic sites inHowever, much social and behavioral research entailsChina. Journal of the Israel Prehistoric Society, suppl. 1:

108. the study of cultural phenomena and, thus, what an-spurdle, a. b., m. f. hammer, and t. jenkins. 1994. thropologists call ethnography. ‘‘Ethnography,’’ as we

The Y alu polymorphism in southern African populations and use the term, refers to the processes and the products ofits relationship to other Y-specific polymorphisms. Americanresearch that seeks to document what people believe,Journal of Human Genetics 54:319–30.

stowe, h. beecher. 1852. Uncle Tom’s cabin. New York: Li- feel, and do in a way that situates those phenomena atbrary of America. specific times in the history of their life experiences, in-

———. 1869. Old town folks. New York: Library of America. cluding pertinent global events and processes. Tostringer, c., and c. gamble. 1993. In search of the Nean-

achieve this goal, ethnography requires distinctivederthals. London: Thames and Hudson.methodological tools. First and foremost, one must in-tattersall, i. 1986. Species recognition in human paleontol-

ogy. Journal of Human Evolution 15:165–75. teract intensively and create personal relationshipst ishkoff, s. a., a. e. dietzsch, w. speed, a. j. paks- with the people one wants to understand. In contrast to

tis, j. r. k idd, k. cheung, b. bonne-tamir, p. research traditions in which ‘‘experts’’ study ‘‘sub-santachcara-benerecetti, m. moral, s. paabo,jects,’’ ethnography requires one to ask other people toe. watson, n. ri sch, t. jenk ins, and k. k. k idd.

1966. Global patterns of linkage disequilibrium at the CD4 lo- help dispell the ethnographer’s ignorance of them. Eth-cus and modern human origins. Science 271:1380–87. nography therefore rests on the collection of cultural

v isser, m. 1991. The rituals of dinner: The origins, evolution, data. ‘‘Culture,’’ as we use the term, refers to the sys-eccentricities, and meaning of table manners. New York: Pen-

tems of mental constructs people use to interpret them-guin.selves and the world around them (e.g., Barnett 1953).wade, n. 1995. Double helixes, chickens, and eggs. New York

Times Magazine, January 29, p. 20. Culture thus can be thought of as meaning and as con-washburn, s. l. 1959. Speculations on the interrelations of sisting of a series of labeled (e.g., ‘‘mother’’) definitions

the history of tools and biological evolution. Human Biology31:21–31.

———. 1960. Tools and human evolution. Scientific American1. 1997 by The Wenner-Gren Foundation for Anthropological Re-203 (3):63–75.search. All rights reserved 0011-3204/97/3805-0008$1.00. The re-weidenre ich, f. 1940. Some problems dealing with ancientsearch reported here was supported by a grant from the Universityman. American Anthropologist 42:375–83.of Connecticut Research Foundation, for which we are most grate-———. 1943. The skull of Sinanthropus pekinensis: A compara-ful. The opinions, findings, and conclusions expressed in this studytive study on a primitive hominid skull. Palaeontologia Sinica,are ours, however, and do not necessarily reflect the view of then.s. D, 10:1–484.funding organization. Special thanks go to Joseph G. Jorgensen.




of phenomena together with the myriad emotional and If there are substantive differences between life-expe-rience markers (names, labels) and cultural data (theintellectual associations individuals acquire and attach

to specific labels in the process of living. definitions and intellectual and emotional associationswe attach to names and labels), an unbiased (‘‘random’’)The culture theory that prevails in contemporary an-

thropology (e.g., Geertz 1973, 1994) comes to us from sample of life-experience data should yield accurate pa-rameter estimates whereas a convenience sample maythe seminal work of Cooley and Mead, and many of the

presumptions of what, in sociology, came to be called or may not (since convenience samples may, by chance,be unbiased samples). By contrast, if cultural data aresymbolic interactionism constitute fundamental ten-

ets. The process of living through which people con- socially constructed, unbiased and convenience sam-ples of people whose lives encompass equivalent experi-struct their symbolic world consists of interactions that

individuals experience over the course of their lives ential variability should yield identical findings. Differ-ences in cultural data (‘‘boundaries’’ between oneboth with others and with aspects of the nonsentient

world. We thus take for granted that people live in a cul- cultural consensus and another [after Keesing 1994])should reflect life-experience differences (e.g., racism,turally constructed world and that that world is so-

cially constructed. income constraints on opportunities) that influencehow individuals construct the meanings they use toconduct their lives.

the problemmethods

This conclusion poses an important challenge to re-search carried out by reference to the criteria of classical We tested these hypotheses with data on the construc-

tion of family relations collected from a population de-statistical theory. Violations of the assumption thatdata observations occur independently and that predic- fined as the list of 1,990 foster mothers maintained by

the Department of Children and Families (DCF), Statetion errors associated with any one case remain uncor-related with errors made for other cases (for example, of Connecticut, to which Danielle Wozniak had been

granted access for purposes of research on the social andcross-sectional heteroskedasticity or spatial or temporalautocorrelation) invalidate the usual statistical tests. self-construction of identity among foster mothers. We

selected two sets of informants independently. A sim-Questions like ‘‘How old are you?’’ or ‘‘What is yourrace?’’ elicit datum points—labels, names—that reflect ple random sample of 45 women was selected using Ex-

cel’s uniform random-number generator. We selected aelements of the unique life history that makes each ofus an independent being and therefore may yield data convenience sample by asking key informants to partic-

ipate and to ask people they knew who fostered childrenconsistent with the assumption of observation and errorindependence. Sophisticated diagnostic techniques now to participate. We collected data by means of mailed

questionnaires. The simple random sample yielded aallow us to evaluate whether or not our data violatethese assumptions, and equally sophisticated tech- 60% return rate (n 5 27). The convenience sample in-

cluded 42 key informants and their friends. We main-niques (for example, measurement unit transforms,weighted least squares, and differencing procedures) tained standard human-subjects protection procedures.

Life-experience data collected included (1) the num-give us means to avoid the analytical confusion such vi-olations create. ber of years a woman had been a foster mother (range

0–30 years, mean 5 5.9 years, s.d. 5 6.5), (2) how manyCultural data, however, come from questions like‘‘How do you know you are a mother, and how did you children she had fostered (range from 0 for women

newly certified as foster mothers to an estimated 250,learn it?’’ or ‘‘Who belongs to your family, and what dis-tinguishes family members from other people?’’ mean 5 23.0, s.d. 5 46.3), (3) how many of these chil-

dren she had adopted (range 0–8, mean 5 .8, s.d. 5 1.4),Whereas the answer any one person gives to a questionlike ‘‘How old are you?’’ has no necessary relationship (4) how many children she was currently fostering

(range 0–5, mean 5 1.8, s.d. 5 1.2), (5) how many ofto the answer any other person gives to the same ques-tion, culture theory posits that no one can respond to a these children were relatives (range 0–3, mean 5 .1, s.d.

5 .5), (6) how many persons currently lived in her homequestion like ‘‘How do you know you are a mother?’’without talking about the social interaction necessary (range 1–10, mean 5 5.1, s.d. 5 1.9), (7) how old she was

(range 28–78, mean 5 44.5, s.d. 5 9.2), (8) her raceto formulate the answers. Cultural data thus should re-flect the social (interactive) processes by which we con- (44.93% ‘‘white,’’ 47.8% ‘‘black,’’ 4.35% ‘‘Hispanic,’’

and 1.45% other), and (9) annual household income (ex-struct our knowledge of each other and the way thesesocial processes work. Cultural datum points should excluding DCF payments), from ‘‘under $15,000’’ (4.35%)

through ‘‘under $30,000’’ (24.65%) and ‘‘under $60,000’’ist only insofar as they reflect the lack of independencebetween sources. Cultural data thus appear to invali- (49.28%) to ‘‘$60,000 and over’’ (18.84%). These data

thus included variables for which the parameter wasdate the usual statistical techniques and make the clas-sical sampling criterion of case independence impossi- known (race, age) and variables that measured life-expe-

rience differences (e.g., racism, income constraints onble to attain. How, then, should we select cases, whatanalytical techniques are appropriate, and to whom can opportunities) that might influence the meanings that

individuals constructed. We constructed three addi-we legitimately generalize?




tional variables: (1) a measure of fostering experience, point scale from ‘‘not at all important’’ to ‘‘very impor-tant.’’as an interaction of age and number of years a woman

had been a foster mother, (2) the natural logarithm of SYSTAT (Wilkinson et al. 1997) and ANTHROPAC(Borgatti 1992) software generated the output reportedthe experience variable, and (3) a dummy variable mea-

sure of lower-class status, coded ‘‘1’’ for households here. We use simple point estimates and 95% confi-dence limits to test the hypothesis that the simple ran-with under $15,000 annual income and households

with under $30,000 annual income if household size ex- dom sample accurately estimates population parame-ters and the convenience sample does not. We useceeded 5 and ‘‘0’’ otherwise.

Cultural data collected included four sets of ques- consensus analysis (e.g., Romney, Weller, and Bat-chelder 1986, Weller and Romney 1988) to test explic-tions bearing on family relations, the significance of

which had emerged in texts (conversations and tran- itly for the presence or absence of cultural boundariesand, thus, to test the hypothesis that findings from thescribed interviews) bearing on the nature of families

collected over the course of two years of intensive par- simple random sample and the convenience sample donot differ. For each set of cultural data, consensus analy-ticipant-observation. The first set asked each woman to

evaluate the importance of nine criteria she might (or sis conducts a minimal residual factor analysis of thecases adjusted for random error, thus asking if the re-might not) use to identify family ‘‘members’’: that the

person (1) be loved, (2) be born into the family, (3) be sponses of each woman constitute just one measure-ment of an unobserved consensus about meaning. Thelegally adopted, (4) contribute work or income, (5) be

taken care of by her, (6) live with her for a long time, following analysis treats responses as qualitative as-signments of meaning rather than rankings and mea-(7) no longer live with her, (8) feel like a family member

to her, and (9) feel like a family member him or herself. sures similarities among individuals as simple match-ing coefficients. A large proportion of the varianceThese criteria thus encompass conventionally recog-

nized legal, biological, and instrumental criteria as well among individuals explained by a single factor the ei-genvalue of which is three or more times larger than theas self-defined affective criteria. We randomized the or-

der of presentation. Informants rated the importance of next largest (a dramatic scree fall) warrants (contingenton diagnostic analysis) the inference of cultural consen-each criterion on a four-point scale from ‘‘not at all im-

portant’’ to ‘‘very important.’’ sus. The correlation between individuals and the firstfactor evaluates the degree of variability in meaningsThe second set of questions that elicited cultural data

asked each woman to evaluate the extent to which she represented in the sample; their average provides ameans of measuring the reliability and validity of thethought of people identified by specific social labels as

belonging to her family. We used labels representing identified consensus. The consensus analysis procedureidentifies the existence of significant agreement differ-conventionally recognized legal and biological distinc-

tions and self-defined affective distinctions that had ences by the presence of more than a single major fac-tor. Initial analyses will test the null hypothesis thatemerged in earlier ethnographic research and random-

ized the order of presentation. These included (1) foster there are no differences in meaning between responsesfrom the simple random sample and the conveniencedaughter, (2) adopted son, (3) cousin, (4) daughter-in-

law, (5) stepdaughter, (6) son-in-law, (7) son, (8) stepson, sample. Cultural consensus among informants selectedrandomly and informants selected by convenience con-(9) best friend, (10) daughter, (11) uncle, (12) mother (13),

father, (14) foster son, (15) husband, (16) adopted daugh- stitute grounds for accepting the null hypothesis.This study uses multidimensional scaling andter, (17) daughter’s boyfriend, (18) aunt, (19) boyfriend,

(20) grandchild, and (21) close male friend. The third set PROFIT analysis to examine the data further. Multidi-mensional scaling works directly on matrices of simi-of cultural data asked each woman to evaluate the de-

gree to which these people (sans ‘‘uncle,’’ which we in- larity coefficients and transforms those coefficients intospatial distance: short distances imply great similarityadvertently left off the questionnaire) were permanent

family members. Informants rated each item on a four- and great distances imply great dissimilarity. The simi-larity matrices among informant cases analyzed by con-point scale from ‘‘not at all a member/a permanent

member’’ to ‘‘very much a member/very permanent.’’ sensus procedures may exhibit one of at least four dif-ferent patterns. A very high degree of consensus—suchThe fourth set of questions that elicited cultural data

asked each woman to assess the importance of nine cri- as would exist if informants worked with a single set of‘‘correct’’ answers—would appear as a tight centralizedteria she might (or might not) use to identify herself as

a ‘‘mother’’: that a child (1) be born to her, (2) be born cluster of cases with increasing degrees of random scat-ter as one moved away from the answers of the ‘‘ex-to a family member, (3) contribute work or income, (4)

be loved by her, (5) be taken care of by her, (6) live with perts.’’ A weaker consensus might appear as a more uni-formly distributed but still clearly clustered scatter ofher for a long time, (7) be adopted, and (8) be thought of

by her as her child, and (9) think of her as her or his cases. By contrast, the absence of consensus would ap-pear as a random scatter of cases. The existence of twomother. Like the first set of cultural data, these criteria

encompass conventionally recognized legal, biological, or more distinct consensuses—such as would exist, forexample, if ‘‘white’’ and ‘‘black’’ women thought aboutand instrumental ties and self-defined affective criteria,

and we randomized the order of presentation. Infor- fostering and the construction of family relations in dif-ferent ways—would appear as two or more foci in themants rated the importance of each criterion on a four-




table 1Parameter Estimate Accuracy of the Random and Convenience Samples

Random Sample Convenience SampleParameter Point Estimate Point Estimate

Parameter Value (95% CI) (95% CI)

Percentage of ‘‘white’’ foster mothers 58 67 31(42–84) (16–49)

Percentage of ‘‘black’’ foster mothers 37 26 62(9–48) (43–78)

Average age of foster mothers 47.2 47.0 43.0(42.9–51.1) (40.4–45.6)

multidimensional scaling map. However, it is possible family member by her, or feel like a family memberhim- or herself. It is ‘‘not at all important’’ that the per-that the mapped agreements among women reflect

more subtle differences. Thus, this study conducts a son be born into the family, be legally adopted, contrib-ute work or income, live with the woman a long time,PROperty-FITting (PROFIT) analysis (Kruskal and Wish

1978) of the mapped agreements among women to eval- or no longer live with the woman.A person with any of the following social labels isuate the extent to which variables such as race, income,

or the number of children adopted influence the manner ‘‘very much’’ a ‘‘family member’’: foster daughter,adopted son, cousin, daughter-in-law, stepdaughter,in which foster mothers construct family relations. The

probability values for the vector R2 are generated from son-in-law, son, stepson, best friend, daughter, uncle,mother, father, foster son, husband, adopted daughter,1,000 random permutations based on Fisher’s method

of randomization. We conducted the PROFIT analysis aunt, and grandchild. A person with any of the follow-ing labels, by contrast, is only ‘‘somewhat’’ a memberwith all 12 of the variables constructed with individual

data plus a variable that distinguished the simple ran- of the family: daughter’s boyfriend, boyfriend, and closemale friend. Everyone is a ‘‘very permanent’’ familydom sample from the convenience sample as a further

check on consensus analysis findings. This exploratory member except foster daughter, son’s girlfriend, boy-friend, and close male friend, who are only ‘‘somewhatPROFIT analysis of 13 variables run five times capital-

izes on chance findings of low probabilities; we cor- permanent’’ family members, and daughter’s boyfriend,who is only ‘‘somewhat’’ and ‘‘not very permanent’’ arected a conventional α of .05 with the Bonferroni pro-

cedure. family member.There exists a clear consensus that to be a ‘‘mother’’

it is ‘‘very important’’ to love a child, take care of thefindingschild, think of oneself as a mother, and be thought of asher or his mother by the child. The consensus about theTable 1 shows, as expected from classical statistical

theory, that the simple random sample accurately esti- affective meaning of being a ‘‘mother’’ breaks down en-tirely for the conventional biological, legal, and instru-mates known population parameters (albeit not very

precisely, owing to the small sample size) and that the mental criteria of ‘‘motherhood.’’ For these latter crite-ria, the ratio of the first to the next largest eigenvalueconvenience sample does not. All 95% confidence lim-

its for the simple random sample encompass the param- is only 1.3 and the average level of agreement amonginformants only .01. Whereas the multidimensionaleter; no 95% confidence limits for the convenience

sample encompass the parameter. scaling maps of consensus agreements exhibit clear,single foci, the map of the agreements over the impor-Table 2 shows, as expected from culture theory, that

the simple random sample and the convenience sample tance of biological, legal, and instrumental criteria of‘‘motherhood’’ exhibits random scatter (see fig. 1).yield identical findings. Initial factor eigenvalue ratios

are three or more times higher (and account for 65% to PROFIT analyses reveal that the patterns of agreementamong informants, whether or not they exhibit culturalnearly 90% of the variance in individual responses) for

cultural data sets bearing on the criteria for judging consensus, show no patterning by the number of yearsa woman has been a foster mother, how many childrensomeone as a ‘‘family member,’’ who belongs to the

‘‘family,’’ the permanency of ‘‘family membership,’’ she has fostered, how many of these children she hasadopted, how many children she currently fosters, howand the affective criteria for judging oneself a ‘‘mother,’’

thus identifying clear consensus about the meanings of many of these children are relatives, how many personscurrently live in the home, how old she is, her race, an-these issues. For someone to be considered a ‘‘family

member,’’ it is ‘‘very important’’ that the person be nual household income, fostering experience, the natu-ral logarithm of the experience variable, lower-class sta-loved, be taken care of by the woman, be felt to be a




table 2Consensus and PROFIT Analysis Findings

SignificantCorrelation Vectors

Ratio of 1st between Identified byEigenvalue to Informant PROFITNext Largest Average Responses and the Analysis(Percentage of Agreement Identified (Bonferroni-

Explained among Cultural Adjusted forCultural Data Set Variation) Informants Consensus α 5 .05)

Criteria for judging someone a family member 3.0 (64.7) 0.54 0.98 noneWho belongs to the family 11.6 (87.0) 0.82 0.99 nonePermanency of family membership 6.9 (78.8) 0.71 0.99 noneCriteria for judging self a mother (affective) 7.4 (83.5) 0.76 0.99 noneCriteria for judging self a mother (legal, 1.3 (42.9) 0.01 (no identifiable none

biological, instrumental) consensus)

Fig. 1. Multidimensional scaling maps of informant agreements about criteria for judging oneself a ‘‘mother’’.Left, affective criteria (stress 5 .0517), right, biological, legal, and instrumental citeria (stress 5 .1237).

tus, or the differences between the simple random people and aspects of the nonsentient world.2 Theytherefore validate the practice of sampling knowledge-sample and the convenience sample. Only 3 of 65

PROFIT runs yielded probabilities at or below .05, able informants—informants we carefully select (John-son 1990)—to learn about cultural domains with clearwhich we could expect merely by chance. No probabil-

ity fell at or below the Bonferroni adjusted probability rules or correct answers. These findings also validatethe practice of selecting informants because they areof (.05/65 5 .0008).available for cultural domains—such as those studied—for which there exist no set of clear rules and for whicheveryone is an ‘‘expert.’’implications

These findings support the theoretical distinction be-2. Our findings about how foster mothers construct family rela-tween ‘‘life-experience’’ data and ‘‘cultural’’ data and tions challenge American kinship models based on conventional

validate the hypothesis that individuals construct ‘‘cul- biological, legal, and instrumental criteria. We will deal separatelywith these issues of power and the construction of culture.tural’’ data through their social interaction with other




table 3Conceptualizing life stories in ways that coherentlyincorporate the implications of being born at a specific Cultural Data Reliability and Validity Estimates for

Different Sample Sizes and Levels of Agreementtime and place, the effects of maturation, and the expe-rience of specific historical events and processes at spe- Among Informantscific ages constitutes a major theoretical challenge.Studying them requires distinctive methodologies. In-

Number of Average Leveldeed, this problem arose the first time an anthropologist Informants of Agreement Reliability Validityused numerical methods—when Edward B. Tylor at-tempted to explain the origins and distribution of kin

9 0.5 0.9 0.95avoidances with a hodgepodge of psychological, func-18 0.5 0.95 0.97tional, and evolutionary hypotheses (1889). One of the 36 0.5 0.97 0.99

originators of the numerical tools that we use today, Sir 6 0.6 0.9 0.9512 0.6 0.95 0.97Francis Galton, asked Tylor how he knew that his cases24 0.6 0.97 0.99were independent, since the similarities he tried to ex-

4 0.7 0.9 0.95plain by reference to individual needs and social func-8 0.7 0.95 0.97

tions might merely reflect social interaction mediated 16 0.7 0.97 0.99either by a common history or geographical propin- 3 0.8 0.92 0.96

6 0.8 0.96 0.98quity. Tylor had no answer.18 0.8 0.99 0.99In the second use of numerical methods by an anthro-

3 0.9 0.96 0.98pologist, Boas (1894) addressed this problem with a clus- 6 0.9 0.98 0.99ter analysis of regional and historical social relation- 18 0.9 0.99 ,1.00ships bearing on the distribution of myths and storiesfound among tribes of the Northwest Coast, the adja-cent Plateau, and three widely scattered communities(interior Athapaskan, Ponca, and Micmac). In the pro- variability in experience that might explain cultural dif-cess, Boas adduced evidence for the regional and histori- ferences, if and when we identify them.cal influences to which Galton alluded in his question Over the past decade, research on ethnographic meth-about observation independence (spatial and temporal ods has produced tools that allow us to demonstrate theautocorrelation). Boas found that neighboring commu- evidentiary value of cultural data.3 Weller (1987) hasnities shared more than distant communities (e.g., shown that the Spearman-Brown prophecy formula canNorthwest Coast tribes were more similar to each other be applied to informants rather than items, allowing usthan to tribes on the Plateau, but Northwest Coast and to measure the reliability and validity of the culturalPlateau tribes shared more than either did with the inte- data we report. Indeed, table 3 shows that so long as therior Athapaskan, Ponca, and Micmac) and that people average level of agreement among informants is .5 orwho might share an ancestral community (who spoke higher, ethnographic conclusions based on informationlanguages from the same language family) shared more from small numbers of informants (3–36, depending onthan people who spoke languages from different lan- the level of agreement) exhibit exceptional reliabilityguage families. (.90–99) and validity (.95–,1). To assess internal valid-

Where Boas demonstrated case-dependence in cul- ity, consensus, multidimensional scaling, and PROFITtural data collected in and taken to characterize differ- analytical tools, among others, allow explicit tests forent communities, our study extends Boas’s findings to the existence of cultural consensus and the presencecultural data collected from and taken to characterize and basis of cultural boundaries and, thus, valid general-different individuals. The socially constructed nature of ization to populations defined by the life experiencescultural phenomena makes the classical sampling crite- studied.rion of independent case selection not only impossible Efficient research designs should differentiate care-to attain but also undesirable. Indeed, research on cul- fully between life-experience and cultural data. Life-ex-tural phenomena that ignores the historically and re- perience data consist of measurements of names or la-gionally specific social interaction through which we bels applied to categories of experience and therefore ofcreate cultural phenomena simply misses the point measurements that may be made independently of each(e.g., see Wolf 1984). This does not make sampling fluc- other. ‘‘Race,’’ ‘‘ethnicity,’’ and ‘‘gender’’ labels col-tuations go away, but it means that classical tests will lected from any one person have no necessary relation-not give us the information we need. As we reorient ship with the labels collected from any other person.theory and research in ways that transform ethnography Answers to questions about what a label refers to orinto ethnology by a focus on cultural variability be- about the intellectual and emotional associations indi-tween individuals rather than between reified and es- viduals attach to the construct (how the label was cre-sentialized groups (cf. Bidney 1944, Barnett 1953, Mur-dock 1971, Vayda 1994, Keesing 1994), this also means 3. Concern with measuring the reliability and validity of culturalthat we take seriously the question of who agrees with data explicitly, however, goes back to Harold Driver’s ‘‘The Reli-

ability of Culture Element Data’’ (1938).whom about what and to what degree and search for the




ated, how one learns which to attach to oneself and un- The Effects of Markets onder what circumstances, and the emotional and otherassociations evoked by the label) must reflect social in- Neotropical Deforestation: Ateraction. Understanding of the meanings people use to

Comparative Study of Fourexperience and respond to the world around themcomes from examination of the historically and region- Amerindian Societies1ally situated social interaction through which they cre-ate and change meaning.

Efficient sampling strategies for cultural data shouldricardo godoy, david wilkie, and jeffreytherefore focus on the selection of knowledgeable infor-franksmants whose lives encompass experiential variabilityDepartment of Anthropology, University of Florida,that may influence the studied meanings, track levelsGainesville, Fla. 32611, U.S.A. 22 iv 97of agreement, and expand sample sizes and sampling

criteria consistent with levels of agreement and identi-Ever since Gross et al.’s (1979) seminal article on thefied cultural boundaries.effects of acculturation on agricultural intensification,scholars have been debating the effects of markets on

References Cited indigenous people’s use of natural resources as part of alarger discourse on how economic development or in-barnett, h. g . 1953. Innovation. New York: McGraw-Hill.come affects the environment. Despite at least two de-b idney, david. 1944. The concept of culture and some cul-

tural fallacies. American Anthropologist 46:30–44. cades of debate, scholars still disagree on these effects.boas, franz. 1894. Indianische Sagen von der Nord-Paci- In this article we present a model, three hypotheses

fischen Kuste Amerikas. Berlin: Asher. arising from it, and empirical evidence from four Amer-borgatti, s. p. 1992. ANTHROPAC 4.0. Columbia, S.C.: Ana-

indian societies in the rain forest of Honduras (Ta-lytic Technologies.wahka) and Bolivia (Mojeno, Yuracare, Chimane) to ex-driver, harold e. 1938. Culture elements distribution. 8.

The reliability of culture element data. Anthropological Rec- amine the role of markets in the loss of old-growth rainords 1(4). forest. We explain the conditions under which markets

geertz, cl ifford. 1973. Interpretation of cultures. New hurt or help conservation and thus help to bridge oppos-York: Basic Books.ing views in the debate.———. 1994. ‘‘The uses of diversity,’’ in Assessing cultural an-

thropology. Edited by Robert Borofsky, pp. 454–65. New York: Our focus on Amerindians does not imply that theyMcGraw-Hill. are the main culprits in neotropical deforestation; the

johnson, jeffrey c. 1990. Selecting ethnographic infor- model and hypotheses apply to any group. Rather, theirmants. Newbury Park, Calif: Sage Publications.

behavior is of special interest because they are gaining akees ing, roger. 1994. ‘‘Theories of culture revisited,’’ in As-greater say in the use of their natural resources (Turnersessing cultural anthropology. Edited by Robert Borofsky, pp.

301–12. New York: McGraw-Hill. 1995). Although we stress the role of markets in the losskruskal, j. b., and m. wish. 1978. Multidimensional scal- of forest, we recognize that macroeconomic and sec-

ing. Newbury Park, Calif.: Sage Publications. toral policies also matter. Finally, we focus on the lossmurdock, george peter. 1971. Anthropology’s mythology.of old-growth rain forest because it generally contains(The Huxley Memorial Lecture.) Proceedings of the Royal An-

thropological Institute. more biological diversity than secondary-growth forestromney, a. k., s. c. weller, and w. h. batchelder. (Frumhoff 1995, Saldarriaga et al. 1985, Lawrence,

1986. Culture as consensus. American Anthropologist 88:313– Leighton, and Peart 1995, Finegan 1996, Silver, Brown,38.

and Lugo 1996, Denevan 1992).tylor, edward b. 1889. On a method of investigating the de-The debate about how markets or economic develop-velopment of institutions; applied to laws of marriage and de-

scent. Journal of the Royal Anthropological Institute of GreatBritain and Ireland 18:245–72.

vayda, andrew p. 1994. ‘‘Actions, variations, and change: 1. 1997 by The Wenner-Gren Foundation for Anthropological Re-search. All rights reserved 0011-3204/97/3805-0011$1.00. This pa-The emerging anti-essentialist view in anthropology,’’ in As-

sessing cultural anthropology. Edited by Robert Borofsky, pp. per was originally presented at a seminar in the Department of An-thropology at the University of Florida in Gainesville. We thank320–29. New York: McGraw-Hill.

weller, s. c. 1987. Shared knowledge, intracultural variation, the participants in the seminar for the opportunity to discuss ourideas with them. Financing for this study was provided by the Na-and knowledge aggregation. American Behavioral Scientist 31:

178–93. tional Science Foundation (SBR 9307588, SBR-9417570), the JointCommittee on Latin American Studies of the Social Science Re-weller, s. c., and a. k. romney. 1988. Systematic data

collection. Newbury Park, Calif.: Sage Publications. search Council and the American Council of Learned Societieswith funds provided by the Ford Foundation, the Harvard Institutewilkinson, leland, et al. 1997. SYSTAT: The system for

statistics. Chicago: SPSS, Inc. for International Development (Harvard University), the DavidRockefeller Center for Latin American Studies (Harvard Univer-wolf, eric. 1984. Europe and the people without history.

Berkeley: University of California Press. sity), USAID (through a grant to the Unidad de Analisis de PolıticasSociales), and BOLFOR. N. Brokaw, K. O’Neill, K. McSweeney, J.Overman, J. Demmer, A. Cubas, M. Alvarado, J. Romero, and V.Aliaga provided useful suggestions, logistical support, and field as-sistance. The views expressed here are those of the authors and notthose of the International Monetary Fund.




ment affect conservation has produced at least three po- same time. Since integration into the market for cropsand for labor has opposite effects on forest clearance,sitions. Anthropologists have often said that markets

hurt conservation (Gross et al. 1979, Painter and Dur- the relation between integration into both markets (ortotal income) and deforestation will be nonlinear.ham 1995). In their classic paper, Gross and his col-

leagues showed that circumscription and proximity to The four Amerindian economies studied vary in pop-ulation size (table 1) and in their history of contact. Themarkets caused declining yields, shorter fallow, and

horticultural intensification among Amerindians in the Mojenos have had centuries of contact with cattleranchers and with missionaries. The Tawahka, Yura-Brazilian Amazon (Gross et al. 1979). Development or-

ganizations often say that markets enhance conserva- care, and Chimane have had only more recent contactwith outsiders. All four groups sell crops. The Tawahkation provided that people enjoy secure property rights

in land (World Bank 1992, Hyde, Amacher, and Magrath sell only 7% of their rice harvest, whereas the Chimanesell nearly half (47%) of theirs. All four groups sell labor1995). Last, some researchers say that economic devel-

opment produces ambiguous effects on deforestation. to ranches, loggers, and nongovernmental organizationsin their territory; in Bolivia the heads of households sur-Cropper and Griffiths (1994) found that the relation be-

tween income and deforestation resembled a U-shaped veyed worked an average of 36 days/year outside theirvillage. Formal credit is limited among all groups. Aparabola in Asia and an inverted-U-shaped curve in

Latin America and in Africa. typical family in each of the four groups cleared abouthalf a hectare of old-growth rain forest each year.Three conclusions emerge from the debate. First,

most researchers have focused on documenting defores- The studies were done in 1995 and 1996 with the as-sistance of local researchers from each ethnic group. Intation rather than on presenting models or testable

hypotheses about the conditions under which markets each group we interviewed an average of 120 house-holds spread out over an average of 10 villages. We cap-may hurt or help conservation. Second, researchers

have not determined whether their findings hold across tured the extremes of integration into the market by in-terviewing households in the most remote and in thedifferent definitions of market integration. For instance,

Gross and his colleagues discussed only integration most modern villages; we also interviewed householdsin villages in the middle of an idealized autarky-to-mar-through the labor market. Last, anthropological case

studies often lack enough variance in integration into ket continuum. Sample size ranged from 3% (Mojeno)to 88% (Tawahka) of all households in the population.the market to reveal nonlinearities.

As economies modernize, households take part in the We returned to the field to discuss preliminary resultsof the analysis with the Chimane and with the Ta-market through the sale of crops, labor, or both. The in-

creasing sale of (chiefly) annual crops causes deforesta- wahka.During the interviews we collected information ontion to increase if technology is held constant. Growing

demand for crops from the outside world induces rural personal, household, and village attributes. We mea-sured the education and self-reported morbidity ofhouseholds to clear more and more old-growth rain for-

est. But economic development also opens up opportu- household heads, the size and composition of thehousehold, the length of residence in the village, andnities for nonfarm work inside and outside the village.

Increasing nonfarm work causes deforestation to de- household wealth (as represented by the ownership ofa radio or plastic buckets). Imputed farm income wascline. As people spend an increasing share of their time

working in nonfarm jobs, the opportunity cost of rural estimated by multiplying the total harvest of the mainannual crops by the village price. We estimated incomelabor rises, as does the cost of clearing old-growth rain

forest. earned from wage labor outside the village. Villagedummies were included to control for community fixedDuring economic development households generally

take part in the market for both crops and labor at the effects. We did not have instrumental variables to con-

table 1The People and the Surveys

Survey

% Rice Average Number NumberHarvest of Days Plot % Population of

Name Population Sold Worked/Year Sizea Sampled Households Year

Tawahka 900 7 n.a. 0.47 88 101 95Yuracare 3,339 17 46 0.27 10 62 95Chimane 5,124 47 26 0.66 14 209 96Mojeno 19,759 22 36 0.48 3 107 95

aHectares of old-growth forest cleared during the year of the interview.




table 2Effects of Crop and Labor Markets on Deforestation (Probit Model)

Bolivia Honduras

Mojeno Yuracare Chimane Tawahka

Variable Coefficient S.E. Coefficient S.E. Coefficient S.E. Coefficient S.E.

Education 1.09 3.5 2.19 6.5 24.24 1.8** 22.03 2.6Household size 24.92 2.4** 3.70 5.5 1.10 1.4 21.64 2.1Length of residence 20.13 0.6 1.92 1.1* 20.36 0.2 20.79a 0.5Illness 20.04 0.1 0.40 0.3 20.15 0.2 0.19 0.6Radio 210.54 12.7 257.20 25.7** 5.72 4.7 0.38 4.0Crop Yb 8.33 7.7 0.07 0.05 0.02 .005*** .003 .001***Wage Y 2.34 2.3 20.03 0.02 2.002 .002 214.73 15.8c

Pseudo-R2 18.6 55.1 32.50 26.89Observations 79 53 175 89

note: Dependent variable is categorical variable for old-growth rain forest cut; explanatory variables held constant at their meanvalue. Coefficients are probabilities of cutting forest for a unit increase in the explanatory variable. Village dummies included.*, **, and *** significant at #10%, 5%, and 1%.aBuckets instead of radio.bIn logarithms.cWage income 5 share of cash income earned in May 1995 from wage labor.

trol for endogeneity. A probit model was used to esti- and imputed farm income with total income, the sumof the two variables. As predicted, the results shown inmate the probabilities of cutting old-growth rain forest

when explanatory variables were held constant at their table 3 suggest that the relation between total income(or integration into the market of both crops and labor)mean value.

Table 2 contains the regression results. As predicted, and deforestation resembles an inverted-U-shaped curveamong the Chimane, Mojeno, and Yuracare.the estimated coefficient for imputed farm income

bears a positive relation to the loss of old-growth rain The signs of the estimated coefficients provide tenta-tive confirmation for the hypotheses that integrationforest. The estimated coefficient for wage earnings out-

side the village bears the hypothesized negative sign. into the market for annual crops worsens deforestationbut integration into the market for wage labor reducesIn table 3 we replace the variables for wage income

table 3Effects of Total Integration into Markets on Cutting of Old-Growth Forest (Probit Model)

Mojeno Yuracare Chimane

Variable Coefficient S.E. Coefficient S.E. Coefficient S.E.

Education 22.64 3.8 3.71 6.3 27.73 2.0***Household size 23.97 2.4 7.85 4.8* 1.37 1.5Length of residence 0.15 0.6 1.96 1.0** 20.27 0.2Illness 20.13 0.1 0.32 0.3 0.03 0.2Radio 26.89 13.4 265.97 21.1** 8.63 5.1*Income 0.01 0.01 0.05 0.05 .001 .004Income2 21.05e-8 2e-8 23.6e-7 2.6e-7 21.01e-9 3e-9Test (Probability)a 14.39 32.89 90.38Pseudo-R2 19.21 52.42 27.39Observations 88 53 174

note: Dependent variable and meaning of coefficients same as in table 2. Village dummies included. Tawahka excluded because oflack of data on wage income.*, **, and *** significant at #10%, 5%, and 1%.aJoint significance of income and income squared.




turner, terence. 1995. An indigenous people’s struggle fordeforestation. Integration into both markets producessocially equitable and ecologically sustainable production: Thean inverted-U-shaped parabola, though our model sug-Kayapo revolt against extractivism. Journal of Latin American

gests that the relation could also take other nonlinear Anthropology 1:98–21.forms. The study contains several lessons in methods world bank. 1992. World development report: Development

and the environment. Washington, D.C.: Oxford Universityand in theory.Press.On the methodological side, the study suggests that

researchers must measure integration into the marketfor both labor and crops if they wish to understand infull the effects of economic development on deforesta- Fertility and Mode oftion. Put differently, definitions matter; how one de-fines integration into the market influences the results. Subsistence: A PhylogeneticThe results also suggest that village-focused studieswill fail to detect that beyond a threshold of income or Analysis1

integration into the market households may cease todeforest because they will begin to switch out of agri-culture altogether into permanent nonfarm work out- daniel w. sellen and ruth mace

Department of Anthropology, University Collegeside the village. At the same time, standard rural house-hold surveys such as those done by the Living Standard London, Gower St., London WC1E 6BT, U.K. 3 iii 97Measurement Surveys of the World Bank will probablyfail to capture the initial transition out of autarky be- There has been much recent debate over the extent to

which the demography of human populations is depen-cause of the costs and the logistical difficulties of reach-ing households at the left-hand tail of integration. dent on the mode of subsistence. The resolution of this

debate will have important implications for anthropolo-On the theoretical side, we have shown through asimple model the conditions under which markets may gists working in several areas, including archaeologists

trying to understand the origins of agriculture and be-hurt or help conservation. Anthropologists and eco-nomics have often been at loggerheads about the effects havioural ecological anthropologists interested in envi-

ronmental effects on birth spacing and the reproductiveof markets on deforestation. The model and approachpresented here suggest that part of the disagreement strategies of individuals. Attempts to test the hypothe-

sis of a causal link between the practice of agriculturemay have arisen because the different disciplines havebeen looking at different parts of the same curve. and higher fertility have focused on the issue of whether

fertility rates differ between traditional agriculturalistsand non-agriculturalists in modern populations. Con-References Citedflicting conclusions have been published by workers us-

cropper, maureen, and charles griff iths. 1994. ing quite similar base samples but applying differingThe interaction of population growth and environmental qual- controls and exclusion criteria. For example, Campbellity. American Economic Review 84:250–54.

and Wood (1988) found great overlap in the ranges of fer-denevan, will iam h. 1992. The pristine myth: The land-tility rates for a large selected sample of natural-fertilityscape of the Americas in 1492. Annals of the Association of

American Geographers 82:369–85. populations and no significant differences in the meanf inegan, bryan. 1996. Pattern and process in Neotropical sec- fertility rates of populations grouped by principal mode

ondary rain forests: The first 100 years of succession. Trends in of subsistence. Hewlett (1991) reported higher fertilityEcology and Evolution 11:119–24.among a small sample of horticulturalists and pasto-frumhoff, peter c. 1995. Conserving wildlife in tropical for-

ests managed for timber. BioScience 45:456–64. ralists than among foragers, but the results were statis-gross, daniel, george eiten, nancy m. flowers, tically insignificant. Conversely, Bentley, Goldberg,

francisca m. le ioi, madel ine lattman ritter, and Jasienska (1993; see also Bentley, Jasienska, andand dennis w. werner. 1979. Ecology and acculturation

Goldberg 1993) excluded from Campbell and Wood’samong native peoples of central Brazil. Science 206:1043–50.sample a number of populations for which measures ofhyde, william f., gregory s. amacher, and wil -

liam magrath. 1995. Deforestation and forest land use: fertility were ambiguous or based on small sample size,Theory, empirical evidence, and policy implications. World populations with high rates of primary sterility, andBank Research Observer 11:223–48. populations whose subsistence practices were in transi-lawrence, d., m. le ighton, and d. peart. 1995. Avail-

tion at the time of demographic measurement andability and extraction of forest products in primary and man-aged forests around a Dyak village in West Kalimantan, Indone- found that the fertility of agriculturalists but not horti-sia. Conservation Biology 9:76–81. culturalists was significantly higher on average than

painter, michael, and will iam durham. Editors. that of non-agriculturalists. There is a closely related1995. The social causes of environmental degradation in Latin

debate over whether the process of settlement itself re-America. Ann Arbor: University of Michigan Press.saldarriaga, j. g., d. c. west, m. l. tharp, and c.

uhl. 1985. Long-term chronosequence of forest succession inthe Upper Rıo Negro of Colombia and Venezuela. Journal of 1. 1997 by The Wenner-Gren Foundation for Anthropological Re-

search. All rights reserved 0011-3204/97/3805-0012$1.00. WeEcology 76:938–58.s i lver, whendee l., sandra brown, and ariel e. thank Gillian Bentley for sharing her data and insights and four

anonymous referees for helpful comments. We are grateful to thelugo. 1996. Effects of changes in biodiversity on ecosystemfunction in tropical forests. Conservation Biology 10:17–24. Leverhulme Trust and the Royal Society for financial support.




sults in demographic shifts which may not necessarily published sources to develop a base sample of estimatesof total fertility rates (TFR) for natural-fertility popula-be associated with the adoption of agriculture (Lee

1972). Whereas Roth (1985) has suggested that fertility tions (no evidence of fertility regulation with increasedparity) representative of as many cultures as possible.of both foraging and pastoral groups increases with sed-

entism, Hewlett (1991) found no effect of settlement Selection criteria for inclusion in the final sample were(1) that systematically comparable quantitative esti-pattern among foragers, and Randall (1994) has argued

that demographic differences between sympatric no- mates of the importance of alternative traditional sub-sistence practices were available and (2) that the histori-mads and sedentary groups are minimal. Finally, there

is debate over whether the low fertility rates reported cal relationships within the sample could be estimatedusing published information on their genetic similari-for some foragers and horticulturalists result from high

rates of sterility and primary infertility due to sexually ties.We collated the sample of population-level demo-transmitted diseases rather than reproductive adapta-

tions to negative energy balance (Pennington 1992). graphic estimates of total fertility rates used in previousinvestigations of the relationship between fertility andA major problem with all these analyses is that none

adequately controls for Galton’s problem (Mace and Pa- subsistence in natural-fertility populations (Bentley,Goldberg, and Jasienska 1993, Campbell and Woodgel 1994, 1996). Historically related populations tend to

share sets of characteristics by chance alone rather than 1988) with the appropriate entries in two source compi-lations of data on human populations. First, cultural-because of any causal associations among those charac-

teristics and are therefore not statistically independent. level data on agriculture were obtained from the com-puterised version of the World Ethnographic SampleIt is possible that the reported cross-cultural associa-

tions between fertility and subsistence result from cul- made available to us by Patrick Gray (henceforth theWES). The WES is an expansion of Murdock’s originalture-historical tendencies rather than from a mechanis-

tic relationship between subsistence per se and the Ethnographic Atlas and contains recently revised cate-gorical codes for relative dependence on various typescomponents of fertility. Cultures practising a particular

form of subsistence are likely to have similar marriage, of subsistence for a large sample (n 5 1,267) of humancultures, both historical and ethnographic (see Murdockchild care, and other practices known to influence fer-

tility by descent from a common cultural ancestor. An- 1965, Murdock and White 1969, Murdock 1981). Toproduce the codes, estimates of total dependence, bothother problem is that these existing studies have

adopted a mutually exclusive categorical approach in subsistence and economic, on hunting, gathering, fish-ing, animal husbandry, and farming were made bydefining subsistence dependency. Useful information is

lost when comparing populations dependent on several teams of ethnographic readers using standardised cate-gorisation criteria developed by Murdock, White, Bur-forms of subsistence.

Here we present the results of an analysis of a se- ton, Gray, and co-workers. Second, population-leveldata on the distribution of human genetic polymor-lected sample of cultures designed to tease out such cul-

tural ‘‘phylogenetic’’ effects by comparing only the dif- phisms were obtained from the recent global analysis ofpatterns of genetic affinities by Cavalli-Sforza and asso-ferences in fertility and subsistence among cultural

sister populations (Felsenstein 1985, Pagel 1992). We ciates (Cavalli-Sforza, Menozzi, and Piazza 1994). Sincea large proportion of the fertility estimates used by pre-use quantitative measures of degree of dependence on

different subsistence types which take account of the vious investigators pertained to populations and cul-tures not included in one or both of these additionalfact that most populations depend to some degree on

several forms of subsistence. sources, they were dropped from our analysis. To boostsample sizes and bring some fresh data to bear, we im-plemented a retrospective snow-ball literature searchmaterials and methodsfor additional published estimates of natural fertilityusing the key-word ‘‘fertility’’ in the BIDS and MED-Our objective was to test the null hypothesis that no

culturally independent association between depen- LINE citation indices between the years 1989 and 1996.Any demographic data found on populations suspecteddence on agriculture and fertility exists. Populations for

which relevant data could be found were arranged on a by the original authors to fit the definition of ‘‘naturalfertility’’ were also collated with the other data sets. Wephylogeny (a branching tree diagram indicating histori-

cal relationships). The amount of change along the were able to collate the necessary data for a total of 69cultures, of which 46 had been included in previousbranches of the tree in the variables of interest was mea-

sured independently at each node (Felsenstein 1985). non-phylogenetic analyses by other investigators. Table1 summarizes the data from the selected sample.Regression analysis was performed using these ‘‘con-

trasts’’ rather than the raw values for populations at Treatment of data. For purely descriptive purposesthe cultures in table 1 are categorised with respect tobranch tips (Pagel 1992).

Our research strategy was to construct a representa- subsistence mode using the codes in the WES (variable42, with the exception that we recategorised the Hadzative sample of human populations for which compara-

ble quantitative measures of subsistence and fertility as gatherers rather than hunters). For the analysis, weused continuous estimates of dependence on the prod-were available and for which historical relationships

could be estimated in a consistent manner. We used ucts of various subsistence activities for a typical popu-




ta

bl

e1

Sam

ple

Sele

cted

for

An

alys

is

Dat

eE

thn

ogra

phic

Eth

no-

Per

cen

tT

ype

offo

rR

egio

nof

grap

hic

Pre

dom

inan

tD

epen

den

ceA

gric

ult

ure

Sett

le-

Subs

is-

Coh

ort

Per

iod

Oth

erD

ate

Ori

gin

WE

SA

tlas

Mod

eof

onan

dM

ain

men

tte

nce

Est

imat

eE

stim

ate

Est

imat

efo

rT

FRG

enet

ican

dC

ult

ure

aN

um

ber

Cod

eSu

bsis

ten

ceb

Agr

icu

ltu

reC

rop

Pat

tern

cD

ata

ofT

FRof

TFR

ofT

FRE

stim

ates

Sou

rce

Affi

liat

ion

Sub-

Sah

aran

Afr

ica

Bw

aka

1038

Ai2

3A

84E

xten

sive

,ro

ots

ortu

bers

P19

204.

200

——

1948

–57

Th

omas

(196

3)M

buti

Pyg

my

(Uba

ngi

an)

Mal

ink

e64

5A

g9A

59E

xten

sive

,ce

real

sP

1930

6.44

0d—

—19

51–

75B

ille

wic

zan

dM

cGre

gor

(198

1)W

est

Afr

ican

(Man

de)

Pla

teau

Ton

ga74

4A

c30

A58

Ext

ensi

ve,

cere

als

P19

406.

600

——

1946

–50

Col

son

(195

8)B

antu

Yao

304

Ac7

A58

Ext

ensi

ve,

cere

als

P19

205.

400

——

1946

–47

Mit

chel

l(1

949)

Ban

tu(M

acon

de)

Les

e10

45A

i30

A57

Ext

ensi

ve,

tree

fru

its

P19

202.

400

——

1985

Hew

lett

(199

1)N

ilo-

Sah

aran

Dog

on11

3A

g3A

57In

ten

sive

(no

irri

gati

on),

cere

als

P19

308.

400

——

1989

Cas

tle

(199

2)N

igor

Kor

dofa

nia

nSe

rer

967

Ag2

2A

48E

xten

sive

,ce

real

sP

1920

6.70

06.

700

—19

63–

65C

antr

elle

(196

9)W

est

Afr

ican

(Ser

er)

Bam

bara

12A

g1A

48In

ten

sive

,ce

real

sP

1920

7.56

08.

100

—19

81H

ill

(198

5)W

est

Afr

ican

(Man

de)

Kip

sigi

s64

8A

j9A

48In

ten

sive

(wit

hir

riga

tion

),P

1920

5.14

0—

—19

82–

83B

orge

rhof

fM

uld

er(1

988)

Nil

o-Sa

har

ance

real

sT

urk

ana

220

Aj5

A39

Ext

ensi

ve,

cere

als

NP

1920

7.23

0—

—19

84B

rain

ard

(199

1),

Les

lie

etal

.N

ilo-

Sah

aran

(198

8)D

atog

a10

78A

j28

P19

Ext

ensi

ve,

cere

als

NP

1940

6.90

0—

—19

87B

orge

rhof

fM

uld

er(1

992)

Nil

o-Sa

har

anH

erer

o2

Ab1

P0

No

agri

cult

ure

NP

1900

—2.

650

—19

09–

59P

enn

ingt

onan

dH

arpe

ndi

ng

Ban

tu(S

outh

wes

t)(1

993)

Mas

ai11

9A

j2P

0N

oag

ricu

ltu

reN

P19

00—

—8.

000

1970

?M

oln

os(1

972)

,ci

ted

inB

orge

r-N

ilo-

Sah

aran

hof

fM

uld

er(1

992)

Mbu

ti20

2A

a5H

0N

oag

ricu

ltu

reN

P19

305.

500

——

1960

Har

ako

(198

1)M

buti

Pyg

my

Boz

o31

2A

g7F

0In

ten

sive

(wit

hir

riga

tion

),N

P19

307.

900

8.34

0—

1987

–91

Mar

riot

t(1

993)

Nig

erK

ordo

fan

ian

cere

als

!Ku

ng

1A

a1G

0N

oag

ricu

ltu

reN

P19

504.

950e

——

1967

–69

How

ell

(197

9)Sa

nH

adza

726

Aa9

G0

No

agri

cult

ure

NP

1910

6.20

0—

—19

67–

85B

lurt

onJo

nes

etal

.(1

992)

San

Cir

cum

-M

edit

erra

nea

nH

un

gari

ans

1106

Ch

8A

66In

ten

sive

(wit

hpl

ows)

,ce

real

sP

1940

——

6.80

0f17

30–

60A

ndo

rka

(199

4)U

rali

c(H

un

gari

an)

Fren

chC

ana-

1133

Cf5

A66

Inte

nsi

ve(w

ith

plow

s),

cere

als

P19

30—

8.00

0—

1700

–30

Hen

ripi

n(1

954)

Cen

tral

Eu

rope

(Fre

nch

)di

ans

Tu

nis

ian

s90

7C

d21

A66

Inte

nsi

ve(w

ith

plow

s),

cere

als

P19

305.

920

9.15

0—

1840

–59

Gan

iage

(196

0)B

erbe

r5

Nor

thA

fric

an(T

un

isia

ns)

Ber

ber

893

Cd7

A47

Inte

nsi

ve(w

ith

plow

s),

cere

als

NP

1910

6.82

0—

—19

80s

Cro

gnie

ret

al.

(199

3)B

erbe

r5

Nor

thA

fric

an(B

erbe

r)A

rusi

841

Ca1

1P

38E

xten

sive

(wit

hpl

ows)

,ce

real

sN

P18

904.

500

——

1990

Lin

dtjø

rn,

Ate

mu

,an

dB

jor-

Afr

o-A

siat

ic(E

thio

pian

)ra

tin

(199

3)A

hag

gare

n88

0C

c9P

29In

ten

sive

(wit

hir

riga

tion

),ce

-N

P19

204.

300

——

1962

Gan

on(1

975)

Ber

ber

5N

orth

Afr

ican

real

s(T

uar

eg)

Zen

aga

891

Cc2

0P

28In

ten

sive

,ce

real

sN

P19

105.

420g

——

1977

Tra

ore

(198

5)N

ear

Eas

tern

(Bed

ouin

)If

ora

883

Cc1

2P

20In

ten

sive

(wit

hir

riga

tion

),tr

eeN

P19

445.

240g

——

1982

Hil

lan

dR

anda

ll(1

985)

Ber

ber

5N

orth

Afr

ican

fru

its

(Tu

areg

)Sh

uw

a87

4C

b16

P20

Ext

ensi

ve,

cere

als

NP

1920

6.38

07.

120

—19

90H

aile

mar

iam

(199

1)A

fro-

Asi

atic

(Eth

iopi

an)

Reg

eiba

t22

Cc1

P19

Inte

nsi

ve(w

ith

irri

gati

onan

dN

P19

505.

500

——

1950

Mu

hsa

m(1

950)

Nea

rE

aste

rn(B

edou

in)

plow

s),

tree

fru

its

Fulb

e68

1C

b8P

19E

xten

sive

,ce

real

sN

P19

206.

750

——

1981

,H

ill

(198

5);

van

den

Eer

em-

Wes

tA

fric

an(F

ula

ni)

1989

beem

t(1

985)

Wod

aabe

1082

Cb2

4P

9C

asu

al/i

nci

den

tal,

cere

als

NP

1950

5.80

0—

—19

81G

anon

(197

5)W

est

Afr

ican

(Fu

lan

i)A

mar

ar86

5C

a35

P9

Cas

ual

/in

cide

nta

l,ce

real

sN

P19

307.

500

——

1980

Lan

dan

dH

olte

r(1

996)

Eas

tA

fric

an(B

eja)

Saam

i12

9C

g4P

0N

oag

ricu

ltu

reN

P19

50—

—5.

000h

1791

–Fr

acca

ro(1

959)

Lap

p18

90R

wal

aB

e-13

2C

j2P

0N

oag

ricu

ltu

reN

P19

203.

600

4.90

0—

1965

Hen

in(1

968,

1969

)N

ear

Eas

tern

(Bed

ouin

)do

uin




Eas

tE

ura

sia

Pu

nja

bi,

Wes

t12

58E

a13

A74

Inte

nsi

ve(w

ith

plow

s),

cere

als

P19

50—

—7.

500h

1900

–19

14P

otte

ret

al.

(196

5)In

dian

(Pu

nja

bi)

Koy

a10

93E

g11

A66

Inte

nsi

ve,

vari

ous

—19

606.

545

——

1954

Ray

and

Rot

h(1

991)

Ela

mo-

Dra

vidi

an(K

oya)

Gu

jara

ti11

36E

f9A

66In

ten

sive

(wit

hpl

ows)

,ce

real

sP

1920

3.51

0—

—19

85C

asim

iran

dR

ao(1

995)

Indi

an(G

uju

rati

)Ir

ania

ns

1135

Ea9

A65

Inte

nsi

ve(w

ith

irri

gati

onan

dP

1950

——

8.20

019

50H

enry

(195

3)Ir

ania

npl

ows)

,ce

real

sSe

noi

1138

Ej1

4A

58E

xten

sive

,ce

real

sN

P19

605.

690

6.29

0—

1968

–69

Fix

(197

7)A

ust

ric

Bai

ga65

4E

g9A

58E

xten

sive

,ce

real

sN

P19

306.

700i

——

1990

Red

dyan

dM

odel

l(1

995)

Mu

nda

Mal

ays

366

Ej8

A57

Inte

nsi

ve(w

ith

irri

gati

onan

dP

1940

8.77

08.

200

—18

88–

1947

Smit

h(1

960)

Mal

aysi

anpl

ows)

,ce

real

sB

enga

li14

1E

f2A

57In

ten

sive

(wit

hir

riga

tion

and

P19

404.

200

——

1967

–69

Tal

uk

dar

(197

9)In

dian

(Wes

tB

enga

l)pl

ows)

,ce

real

sT

urk

men

234

Eb5

A56

Inte

nsi

ve(w

ith

irri

gati

onan

dP

1910

—7.

745

—19

73–

74Ir

ons

(198

0)T

urk

oman

(Alt

aic)

plow

s),

cere

als

Tib

etan

s23

9E

e4A

56In

ten

sive

(wit

hir

riga

tion

and

P19

20—

—5.

400j

1981

;19

76G

olds

tein

(197

6)T

ibet

anpl

ows)

,ce

real

sK

hal

ka

Mon

-13

4E

b3P

10C

asu

al/i

nci

den

tal

(wit

hpl

ows)

,N

P19

407.

530

——

1980

Ch

u(1

997)

Mon

gol

Tu

ngu

s(M

ongo

l)go

lsva

riou

sA

nda

man

ese

45E

h1

FG0

No

agri

cult

ure

NP

1870

2.60

0—

—19

80P

andy

a(1

987)

Indo

-Pac

ific

Ch

ench

u43

Eg1

G0

Cas

ual

/in

cide

nta

l,ce

real

sN

P19

405.

800

——

1970

Siba

judd

in(1

984)

Ela

mo-

Dra

vidi

an(C

hen

chu

)Se

man

g14

8E

j3G

0N

oag

ricu

ltu

reN

P19

205.

200

——

1978

Gom

es(1

988)

cite

din

Hew

lett

Au

stri

c(1

991)

Insu

lar

Pac

ific

On

ton

g26

5Ii

5A

F47

Hor

ticu

ltu

re,

tree

fru

its

NP

1920

——

6.30

0h19

20–

72B

ayli

ss-S

mit

h(1

975)

Indo

nes

ian

(Jav

a)Ja

van

ese

Gil

bert

ese

633

If14

AF

39H

orti

cult

ure

,tr

eefr

uit

sP

1890

6.30

0h—

—19

47–

72L

ambe

rt(1

975)

Mel

anes

ian

(Gil

bert

)Ja

van

ese

54Ib

2A

74In

ten

sive

(wit

hir

riga

tion

and

P19

506.

100

7.30

0—

1979

–83

Tak

asak

a(1

986)

Indo

nes

ian

(Jav

a)pl

ows)

,ce

real

sE

nga

334

Ie7

A66

Ext

ensi

ve,

root

sor

tube

rsP

1950

5.90

0—

—19

59–

64Si

nn

ett

and

Wh

ite

(197

3)N

ewG

uin

ean

Kap

auk

u57

Ie1

A58

Ext

ensi

ve,

root

sor

tube

rsP

1950

4.90

04.

700

—19

55P

ospi

sil

(196

3)N

ewG

uin

ean

(En

ga)

Du

sun

251

Ib5

A48

Ext

ensi

ve(w

ith

plow

s),

cere

als

P19

205.

100

——

1950

Gly

n-J

ones

(195

3)In

don

esia

n(B

orn

ean

)T

iwi

157

Id3

G0

No

agri

cult

ure

NP

1920

——

5.00

0h19

52–

61Jo

nes

(196

3)A

ust

rali

an(N

orth

ern

Ter

rito

-ri

es)

Nor

thA

mer

ica

Nav

aho

182

Nh

3A

39In

ten

sive

(wit

hir

riga

tion

),N

P19

30—

—6.

500h

1844

–19

44M

orga

n(1

973)

Na

Den

e(N

avaj

o)ce

real

sN

un

amiu

t46

1N

a12

H0

No

agri

cult

ure

NP

1950

—6.

900h

—19

35–

68B

info

rdan

dC

has

ko

(197

6)A

lask

aB

un

gi62

1N

e14

H0

No

agri

cult

ure

NP

1850

5.90

0h—

—18

94B

oas

(189

4)N

orth

Am

eric

an(O

jibw

a)Se

ri28

3N

i4F

0N

oag

ricu

ltu

reN

P19

006.

800

——

1960

Nee

lan

dW

eiss

(197

5)H

okan

Ku

tch

in46

9N

a20

F0

No

agri

cult

ure

NP

1880

5.40

0—

—18

58O

sgoo

d(1

936)

Na

Den

eG

reen

lan

ders

488

Na2

5F

0N

oag

ricu

ltu

reN

P19

10—

—3.

500h

1951

Mal

auri

e,T

abah

,an

dSu

tter

Esk

imo

(Gre

enla

nd)

(195

2)P

olar

Esk

imo

463

Na1

4F

0N

oag

ricu

ltu

reN

P18

804.

600

——

1950

Mal

auri

e(1

956)

Esk

imo

(Can

ada

Eas

tA

rcti

c)N

etsi

lik

714

Na4

3F

0N

oag

ricu

ltu

reN

P19

206.

400

——

1923

Shri

rean

dSt

eige

r(1

974)

Esk

imo

(Can

ada

Nor

thW

est

Esk

imo

Ter

rito

ry)

Cop

per

169

Na3

F0

No

agri

cult

ure

NP

1920

——

4.50

019

10–

20Je

nn

ess

(192

2)E

skim

o(C

anad

aN

orth

Wes

tE

skim

oT

erri

tory

)So

uth

Am

eric

aY

anom

amo

1264

Sd9

A84

Ext

ensi

ve,

tree

fru

its

NP

1965

8.20

07.

900

—19

58;

1987

Ear

lyan

dP

eter

s(1

990)

Yan

omam

oM

aya

345

Sa6

A66

Ext

ensi

ve,

cere

als

P15

209.

200

8.20

0—

1964

–66

Ear

ly(1

970)

May

aA

ymar

a19

3Sf

2A

57C

asu

al/i

nci

den

tal

(wit

hpl

ows)

,P

1940

——

7.00

0h19

65–

66C

ruz-

Cok

eet

al.

(196

6)So

uth

Am

eric

an(A

ymar

a)ro

ots

ortu

bers

Bla

ckC

arib

388

Sa7

A58

Hor

ticu

ltu

re,

root

sor

tube

rsP

1940

5.80

0—

—19

46–

47Fi

rsch

ein

(198

4)So

uth

Am

eric

an(C

arib

)G

oaji

ro39

1Sb

6P

10C

asu

al/i

nci

den

tal,

cere

als

NP

1940

5.10

0—

—19

85–

88H

urt

ado

and

Hil

l(1

987)

Sou

thA

mer

ican

(Ara

wak

an)

Aw

eik

oma

199

Sj3

H0

No

agri

cult

ure

NP

1910

——

2.20

019

58Sa

lzan

oan

dde

Oli

veir

a(1

970)

Sou

thA

mer

ican

(Cai

nga

ng)

Shav

ante

1184

Sj11

G20

Ext

ensi

ve,

cere

als

NP

1910

5.70

0—

—19

58Sa

lzan

o,N

eel,

and

May

bury

-So

uth

Am

eric

an(X

avan

te)

Lew

is(1

967)

aA

sde

fin

edin

Mu

rdoc

k(1

967)

.bA

,ag

ricu

ltu

re;

P,p

asto

rali

sm;

H,

hu

nti

ng;

F,fi

shin

g;G

,ga

ther

ing.

cP

,pe

rman

ent;

NP

,n

on-p

erm

anen

t.dO

nly

figu

res

for

vill

age

ofM

andu

arar

eu

sed,

sin

ceth

eot

her

vill

age

incl

ude

dh

adac

cess

tom

oder

nh

ealt

hca

re.

eT

he

mea

nva

lue

for

the

tota

lfe

rtil

ity

rate

sfo

rtw

ose

para

teco

hor

ts(4

.7an

d5.

2)is

use

d.f T

otal

mar

ital

fert

ilit

yra

te;

mar

ital

rate

sre

port

edto

beve

ryh

igh

.gSe

vera

let

hn

icsu

b-gr

oups

pool

ed.

hE

xclu

ded

byB

entl

ey,

Gol

dber

g,an

dJa

sien

ska

(199

3).

i Mea

nof

con

san

guin

eou

san

dsa

ngi

neo

us

mar

ital

fert

ilit

yra

tes.

j Com

plet

edfa

mil

ysi

ze.




lation within each culture. ‘‘Dependence’’ here implies For some cultures, such as the Fulani and the Tuareg,fertility estimates were available from several differentacquisition of food and other goods for direct consump-

tion and for exchange, so that, for example, foragers populations. Unweighted mean values were calculatedfor use in this analysis. For other cultures, such as thewho trade hunted foods or agricultural foods will be

coded as dependent primarily on hunting rather than Hadza and the Herero, multiple fertility estimates wereavailable from single populations censused at differentagriculture. We took the midpoints of the coded ranges

for percent dependence on hunting, gathering, fishing, times. In these cases, the estimate from a time periodmost nearly matching that to which the WES subsis-herding, and agriculture (variables 1–5 in the WES re-

spectively; the coded ranges are 0–5%, 6–15%, 16– tence data referred was used. In a few cases, total fertil-ity rate estimates from different subpopulations (e.g.,25%, 26–35%, 36–45%, 46–55%, 56–65%, 66–75%,

76–85%, 86–100%) and divided by the summed total castes and marriage types) were pooled to give meanpopulation estimates (see notes to table 1).dependence across all subsistence types to arrive at a

scaled estimate of percent dependence on each type of In their reanalysis of Campbell and Wood’s originalfertility data, Bentley, Goldberg, and Jasienska (1993)subsistence for each culture.

The cultures were categorised for analysis by the type excluded some populations known or suspected to beundergoing subsistence changes or to have high rates ofof agriculture, if any, practised (intensive, making use

of fertilisation, crop rotation, or other techniques to primary and/or infectious sterility at the time of mea-surement. We chose to include these data. We reasonedshorten or eliminate the fallow period, versus exten-

sive, involving shifting, swidden, or long fallow [WES that since rates of sterility were not known for many ofthe populations retained in that analysis, they mightvariable 28] and whether use was made of irrigation or

plows (either aboriginal or long-established [WES vari- still be high even with high total fertility rates and thusBentley et al.’s exclusions might have introduced someable 39]). Cultures were also coded as non-permanently

or permanently settled (based on a recoding of WES vari- sort of bias. We also reasoned that infertility might bea correlate of subsistence in the same way as has beenable 35; cultures described in the WES as ‘‘nomadic or

fully migratory,’’ ‘‘semi-nomadic,’’ ‘‘semi-sedentary,’’ postulated for other components of fertility and there-fore in terms of the broader issue there was little to beor having ‘‘compact but impermanent settlements’’

were all coded as non-permanently settled for this anal- gained from factoring out such associations.A phylogeny representing one likely interpretation ofysis).

The methods used to estimate fertility varied in the the genetic history of populations in the sample wasdrawn based on the global genetic phylogeny con-original sources, and therefore a ‘‘best available esti-

mate’’ of fertility was obtained for each culture as fol- structed by Cavalli-Sforza, Menozzi, and Piazza (1994)using information on the distribution of human geneticlows: Cohort estimates of the fertility of postreproduc-

tive women were available for 52 of the 69 populations polymorphisms. We used only those cultures for whichboth fertility and subsistence data could be collated.selected and were used in preference to other types of

fertility estimate on the ground that they captured the Figures 1 and 2 show the phylogeny on which our analy-sis was based. No attempt was made to calculate branchfull range of effects on fertility operating within a popu-

lation, including maternal mortality. Period estimates lengths from average genetic distances between popula-tions. Figure 1 shows the distribution of best availableof fertility calculated from age-specific data assuming

stable age-structure and fertility rates were available for estimates of fertility in relation to estimates of culturalhistory. Figure 2 shows how this compares with the cul-12 of these 53 populations but were not used directly.

Since not all women will survive into higher age- tural phylogenetic distribution of dependence on agri-culture.groups, the period estimate is an upwardly biased esti-

mate of population total fertility. However, the cohort Statistical approach. The final sample of cultures hasa fairly wide geographic distribution, and a wide rangeand the period estimate are highly positively associated

(R2 5 0.977; p , 0.001) in these 12 cultures according of subsistence modes is represented. Table 2 summa-rizes the pattern of differences by subsistence mode into the regression equation TFRc 5 0.9299 (TFRp).

Therefore, in the case of 4 cultures for which only pe- the best published estimates of total fertility rates ob-served in this particular sample. Increased fertility is ap-riod estimates were available (Turkmen, Nunamiut,

Herero, and French Canadians), the total fertility rate parently positively associated with dependence on ani-mal husbandry and intensification of agriculture, butwas indirectly estimated from this equation, boosting

the number of cases for which some estimate of cohort direct comparison would be inappropriate because ofGalton’s problem (Mace and Pagel 1994, 1996). It isfertility was available to 56. For the remaining 13 popu-

lations selected, the original authors either did not de- clear from figures 1 and 2 that although both levels offertility and dependence on agriculture vary quitescribe precisely how the total fertility estimate was ob-

tained for their populations or used a ‘‘weak’’ measure widely among historically related cultures, there is atendency for high levels of fertility to cluster in theof total fertility such as completed family size. Since

identical patterns were found after excluding the data same branches of the phylogeny as high dependence onagriculture and also pastoralism. In an attempt to re-on these 13 populations and after disaggregating the pe-

riod and cohort data, we report only the results of tests duce the possible influence of the non-independence ofcultures on the association between subsistence andperformed on the whole data set.




Fig. 1. Genetic phylogeny of 69 cultures for which Fig. 2. Genetic phylogeny of 69 cultures for whichdemographic and subsistence data were collated,demographic and subsistence data were collated,

showing the best estimate of total fertility levels showing traditional percent dependence onagriculture for each culture as reported inmeasured in one or more populations from each

culture. White, 2–4.99 live births per woman; grey, ethnographic references. White, agriculturetraditionally absent or providing less than 15% of5–6.99 live births per woman; black, 7 or more live

births per woman. Cross-hatching indicates subsistence; grey, 16–55% dependence on agriculture;black, 56–100% dependence on agriculture. Cross-uncertainty about the likely fertility levels of

ancestral cultures. However, a rough estimate of the hatching indicates uncertainty about the likelyagricultural dependency of ancestral cultures.likely fertility levels of some ancestral populations is

indicated on some branches. These estimates are Dependence on agriculture was treated as acontinuous variable in the phylogenetic analysis.based on parsimony and the assumption that

unresolved polytomies represent uncertainties ingenetic affiliation. Total fertility was treated as acontinuous variable in the phylogenetic analysis. ance in these phylogenetically unbiased sets of con-

trasts rather than variance in the raw values of thevariables (Felsenstein 1985). Multiple regressionthrough the origin was used to calculate partial F valuesfertility, the size and direction of numerical differences

in both dependence on agriculture and total fertility of these variances. The independent contrasts in thetest variables at nodes on the genetic-cultural phylog-rate between pairs of historically related cultures

(‘‘nodes’’ in the phylogeny)—so-called phylogenetically, eny were calculated by the method of Pagel (1992),which is implemented in a computer program widelyhistorically, or evolutionarily ‘‘independent contrasts’’

(see Felsenstein 1985, Pagel 1992)—were calculated. employed by ecologists in comparative studies (CAIC[Purvis and Rambaut 1995]). All statistical tests wereThe significance of the associations between subsis-

tence and fertility was assessed by examining the vari- performed using SPSS/PC (SPSS Inc. 1993).




table 2Estimates of Total Fertility Rates by Subsistence Mode

Mean BestPercent Dependence on Various Forms of Subsistenceb Estimate of

TotalPredominant Mode Fertilityof Subsistencea n Fishing Gathering Hunting Pastoralism Agriculture Rates

Fishing 8 53 13 26 0 5 5.2Gathering 5 11 53 30 0 0 5.3Hunting 4 11 31 54 0 0 5.2Pastoralism 16 4 5 10 66 15 5.6Incipient agriculturec 3 32 14 14 5 35 6.1Extensive agriculture 17 13 6 7 14 60 6.5Intensive agriculture 16 7 6 4 25 59 6.6Foragers 18 30 29 34 0 0 5.3Non-foragers 51 9 6 7 33 45 6.1Non-agricultural 34 17 18 22 33 10 5.4Agriculturald 35 12 6 6 19 59 6.3

aType of subsistence estimated to provide the greatest contribution to both food and economic needs according to variables 1–5 inthe WES. Our coding agreed with variable 42 of the WES except in the case of the Hadza, who were originally coded as ‘‘hunters’’but are reported to derive approximately 40% of subsistence from hunting and honey-collecting and 60% from gathering both in theearlier codes in the WES and by recent observers (K. Hawkes, personal communication, 1990).bAdjusted to sum to 100% within each culture; here row totals by subsistence type may sum to more than 100% because ofrounding.cThis group consists of the Gilbertese, the Ontong Javanese, who are equally dependent (about 40% respectively) on fishing and hor-ticulture based on fruiting trees, and the Shavante.dExcludes ‘‘incipient agriculturalists.’’

results creased dependence on livestock substituted for depen-dence on all other forms of subsistence.

When phylogenetically independent contrasts in per-Table 3 shows the bivariate associations between phy-logenetically independent contrasts in percent depen- cent dependence on the major types of subsistence (for-

aging, animal husbandry, and agriculture) were com-dence on all types of subsistence (fishing, gathering,hunting, animal husbandry, and agriculture) and in fer- pared with independent contrasts in fertility using

multiple regression through the origin, the only vari-tility. Differences in dependence on hunting and gather-ing were found to be positively correlated, indicating able showing an independent association with fertility

was dependence on agriculture (table 4). In stepwise de-that these two modes of subsistence tend to be practisedtogether, while differences in dependence on hunting letion procedures beginning with the model containing

all subsistence practices, dependence on agricultureand agriculture were negatively correlated, indicatingthat these two activities substituted for each other was always retained as the only significant predictor of

fertility rates (table 5). Using a cultural phylogeny basedacross cultures. Negative correlations showed that in-

table 3Correlations between Contrasts in Subsistence Dependence in a SelectedSample of Human Populations

Hunting Gathering Fishing Pastoralism

Agriculture 2.5542 2.4744 2.0759 2.1105p , .001 p , .001 p 5 .544 p 5 .377

Hunting — .3159 2.1896 2.3042p 5 .010 p 5 .127 p 5 .013

Gathering — — 2.1275 2.4452p 5 .303 p , .001

Fishing — — — 2.4533p , .001




table 4Contrasts in Total Fertility Rate and Percent Dependence on Various Forms ofSubsistence in a Selected Sample of Human Populations

Variables in β CoefficientRegression Modela (Standardised) T p

% dependence on agriculture 10.3749 3.059 .0032% dependence on animal husbandry 20.1219 2.975 .3334% dependence on extracted resourcesb 20.1201 0.980 .3308

note: An analysis of variance produced a model-adjusted R2 of .100 (F2, 64 5 4.68, p 5 .0127).aScaled to sum to 100% within cultures; results were almost identical when unscaled valueswere used.bValues shown are ‘‘model in’’ values; this variable was excluded from the analysis of variancemodel because tolerance criteria were exceeded.

on genetic similarity (figs. 1 and 2), a 10% increase in of the phylogeny (genetic or linguistic) or of time lagsbetween estimates of subsistence and fertility.dependence on agriculture between sister cultures was

associated, on average, with a fertility increase of ap- The sample was divided into permanently settledgroups and those which had some degree of mobility. Aproximately 0.2 live births per woman in this sample

(fig. 3). The relationship appeared stronger in the sub- clear association between fertility and agriculture wasfound among the non-permanently settled groups (tablesample for which the demographic data were collected

within 20 years of the ethnographic observations on 5). Among the permanently settled (sedentary) ones,most of which are dependent to a high degree on agri-which the subsistence codes were based. Fertility was

estimated to increase by an average of approximately culture, variance in fertility was not associated with de-pendence on agriculture, foraging, or herding. Cultures0.4 births per woman with each 10% increase in ag-

ricultural dependence between sister cultures. When a more highly dependent on agriculture than on any othermode of subsistence, categorised either as ‘‘non-forag-language phylogeny based on Ruhlen’s (1991) classifi-

cation of world languages was used to structure the in- ers’’ (excludes foragers and fishers but includes pasto-ralists) or as ‘‘agriculturalists’’ (excludes pastoralistsdependent comparisons, the estimated relationship was

almost identical to that derived using a genetic phylog- and foragers [see table 1]), were also examined sepa-rately. Again, variance in fertility was not associatedeny. Thus, the finding of a positive association between

fertility and agricultural dependence is not an artefact with dependence on agriculture among these groups.

table 5Relationship between Contrasts in Total Fertility Rate and Percent Dependence on Agriculture in a SelectedSample of Human Populations

Number ofContrasts Significant

Compared/ Predictor of RegressionExclusion Criteria Number of Difference Coefficient AdjustedApplied to Sample Cultures Used in TFRa (6 S.E.) d.f. F p R2

Genetic phylogeny (figs 1 and 2, afterCavalli-Sforza et al. 1994)

No exclusions 66/69 Agriculture 10.0212 6 0.0074 65, 1 8.407 0.0051 .101Cultures with , 20-year discrepancy 15/16 Agriculture 10.0430 6 0.0126 14, 1 11.560 0.0043 .413

between demographic and subsis-tence data

Non-permanently settled only 39/41 Agriculture 10.0310 6 0.0103 38, 1 8.995 0.0048 .191Permanently settled only 26/30 None — — — n.s. .136‘‘Non-foragers’’ only 49/51 None — — — n.s. .061‘‘Agriculturalists’’ only 34/35 None — — — n.s. .176

Linguistic phylogeny (Ruhlen 1991)No exclusions 28/46 Agriculture 1.0263 6 0.0115 27, 1 4.771 0.0295 .133

aFrom stepwise multiple regression.




Fig. 3. Fertility and agricultural dependence in a selected sample of human populations. Differences betweensister cultures were calculated as the difference in the values of each variable at each node in the phylogenyin figures 1 and 2.

These results suggest that a threshold of agricultural de- less agriculturally dependent populations. It has beensuggested that the fertility of recent foragers such as thependence exists above which increases in dependence

on agriculture along a phylogenetic branch are not sig- !Kung may have been limited by undernutrition (Bent-ley 1985, Howell 1979, Lee 1979), high workloads (Blur-nificantly associated with further increases in fertility.

Finally, after controlling for phylogeny, fertility was ton Jones 1986, Lee 1979), and prolonged breastfeeding(Konner and Worthman 1980). Such arguments havefound to increase with agriculture among polygynous

cultures but not among monogamous ones. been extended to some horticulturalists with low fertil-ity (Bailey et al. 1992, Wood et al. 1985). By inference,the adoption of agriculture over alternative modes ofdiscussionsubsistence can be postulated to influence fertility posi-tively in at least three ways: by increasing the amount,In our sample, differences in dependence on agriculture

were the strongest predictors of differences in fertility quality, or predictability of the food supply to reproduc-tive-aged women, by reducing women’s workloads, andbetween closely related cultures. This finding corrobo-

rates previous reports that increased dependence on by increasing the availability of weaning foods foryoung children.agriculture is associated with increased fertility across

human populations. In contrast to previous studies, it It is likely that any differences in energy balance bysubsistence mode will have complex effects on fecund-is unlikely to be due to the general tendency for clusters

of historically related cultures to display similarities, ability, since the mechanisms regulating fecundabilityappear to be distinct from the mechanisms regulatingbecause our method does not assume that all cultures

are independent. Rather, it ensures that the test vari- maintenance of pregnancy (Leslie et al. 1996, Leslie,Campbell, and Little 1993). Evidence from modern pop-ables are compared only within closely related pairs of

cultures, so that only the variance in the data that is ulations suggests that the low maternal nutritional sta-tus and high workloads observed in many agriculturaldue to changes along the branches of the phylogeny is

utilised in hypothesis testing. populations (Ferro-Luzzi 1990), particularly on a sea-sonal basis associated with agricultural cycles (JenikeOur results reveal demographic concomitants of the

persistence of high dependence on agriculture in recent 1995), will tend to reduce fecundity (Ellison 1994). Ar-chaeological evidence, although perhaps not directlypopulations. Despite the observations that many of the

‘‘non-foraging’’ cultures practised hunting, gathering, relevant to modern agriculturalists, also suggests a de-cline in nutritional and general health status and anherding, and fishing to some extent and that all of these

activities substituted for farming, fertility rate differ- increase in physiological stress, presumably due tochanges in activity patterns and increases in workloadsences among historically related populations are larger

when differences in agricultural dependence exist than (Larsen 1995). Both sets of observations might lead usto expect lower, not higher, fertility rates in more agri-when other subsistence activities differ. It is not clear

how limitations on mean fertility might operate among culturally dependent populations. Differences in mater-




nal time allocation and child care may therefore be im- tinguished from hypotheses of origin. Our findings can-not be used to evaluate any of these hypotheses becauseportant. Reductions in the frequency or duration of

breastfeeding may reduce energy demands on mothers they derive from synchronic rather than diachronicdata.and reduce birth intervals. There is very little informa-

tion available for testing the assumption that agricul-turalists introduce weaning foods earlier than otherpopulations either because appropriate foods are more References Citedavailable or because mothers are forced to spend longer

andorka, r. 1994. The historical demography of a proper Hun-periods away from weanlings because of the nature of garian village: Atany in the 18th and 19th centuries. Journal ofagricultural labour. If either were true, this would po- Family History 19:311–31.

armelagos, g., a. goodman, and k. jacobs. 1991.tentially lead to higher fertility and possibly poorerThe origins of agriculture: Population growth during a periodgrowth and higher child mortality (Sellen and Maceof declining health. Population and Environment 13:9–n.d.). 22.

Clearly, our results justify further investigations as to bailey, r. c., m. r. jenike, p. t. ell ison, g. r. bent-ley, a. m. harrigan, and n. r. peacock. 1992. Theprecisely what mechanisms underlie the trend towardsecology of birth seasonality among agriculturalists in centralincreasing fertility with increasing agricultural activityAfrica. Journal of Biosocial Science 24:393–412.and under what conditions this relationship breaks

bayl iss -smith, t. 1975. ‘‘Ontong-Java: Depopulation and re-down. The fertility differences between natural-fertility population,’’ in Pacific atoll populations. Edited by V. Carroll,populations appear to be mediated primarily through pp. 417–84. Honolulu: University of Hawaii Press.

bentley, g. 1985. Hunter-gatherer energetics and fertility: Abehavioural rather than physiological mechanismsreassessment of the !Kung San. Human Ecology 13:79–109.(Wood 1990), although controversy continues over the

bentley, g., a. goldberg, and g. jas ienska. 1993.wider demographic importance of lowered maternal nu- The fertility of agricultural and non-agricultural traditional so-tritional status and workloads (Ellison 1991, 1994) rela- cieties. Population Studies 47:269–81.

bentley, g., g. jas ienska, and a. goldberg. 1993. Istive to other factors. Most of the substantial interpopu-the fertility of agriculturalists higher than that of nonagricul-lation variation in total fertility rates (Campbell andturalists? current anthropology 35:778–85.Wood 1988) has been attributed to differences in age-

billewicz, w., and i. mc gregor. 1981. The demographyspecific patterns of marriage (Bongaarts 1983, Trussell of two West African (Gambian) villages, 1951–75. Journal of1979) and in the intensity and duration of breastfeeding Biosocial Science 13:219–40.

b inford, l., and w. chasko jr. 1976. ‘‘Nunamiut demo-(Gray 1994, Wood 1990). These are largely cultural re-graphic history: A provocative case,’’ in Demographic anthro-sponses with important underlying physiological ef-pology: Quantitative approaches. Edited by E. Zubrow, pp. 63–fects. Evidence suggests that the more strongly physio- 143. Albuquerque: University of New Mexico Press.

logical mechanisms (i.e., those less amenable to blurton jones, n. 1986. Bushman birth spacing: A test foroptimal interbirth intervals. Ethology and Sociobiology 7:91–cultural modification, such as changes in the capacity105.to conceive and rates of foetal loss) appear to be impor-

blurton jones, n., l. smith, j. o ’connell, k.tant only in determining fertility differences among in-hawkes, and c. kamuzora. 1992. Demography of the

dividuals within traditional populations. Hadza, an increasing and high-density population of savannaAlthough the results suggest that high fertility rates foragers. American Journal of Physical Anthropology 89:159–

81.may contribute to the persistence of agriculture, theyboas, f. 1894. The half-blood Indian: An anthropometric study.do not throw light on the demographic concomitants of

Popular Science Monthly 45:761–70.the origin of agriculture. The archaeological record sug- bongaarts, j. 1983. ‘‘The proximate determinants of naturalgests that human population increases 10,000 years ago marital fertility,’’ in Determinants of fertility in developing

countries, vol. 1. Edited by R. Bulatao and R. Lee, pp. 103–38.were linked to the emergence of food production (Has-New York: Academic Press.san 1973, Armelagos, Goodman, and Jacobs 1991, Co-

borgerhoff mulder, m. 1988. ‘‘Reproductive success inhen 1989). This has led some to argue that the inventionthree Kipsigis cohorts,’’ in Reproductive success. Edited by

and intensification of agriculture were responses to T. H. Clutton-Brock, pp. 419–35. Chicago: University of Chi-high fertility and population increase during human cago Press.

———. 1992. Demography of pastoralists: Preliminary data onevolution (the increase-of-carrying-capacity hypothesisthe Datoga of Tanzania. Human Ecology 20:383–405.[Boserup 1965, Cohen 1977, and contributions in Cohen

boserup, e. 1965. The conditions of agricultural growth. Chi-and Armelagos 1984]). Others interpret the shift to agri- cago: Aldine.culture as fostering higher fertility, at least until new brainard, j. 1991. Health and development in a rural Ken-

yan community. New York. Peter Lang.production limits are reached, because of increasedbuikstra, j., l. konigsberg, and j. bullington.availability of dietary energy or appropriate weaning

1986. Fertility and the development of agriculture in the prehis-foods (the nutritional-release hypothesis [Buikstra, Kon-toric Midwest. American Antiquity 51:528–46.

igsberg, and Bullington 1986, Howell 1979, Stephens campbell, k., and j. wood. 1988. ‘‘Fertility in traditional1975, Sussman 1972]). It is also possible that population societies,’’ in Natural human fertility: Social and biological

mechanisms. Edited by P. Diggory, S. Teper, and M. Potts, pp.increase was not dependent on changes in fertility but39–69. London: Macmillan.rather occurred with settlement because of a change in

cantrelle, p. 1969. Etude demographique dans la region duage-specific patterns of survival (Handwerker 1983, Sine-Saloum (Senegal): Etat civil et observation demograph-Pennington 1996, Sattenspiel and Harpending 1983). ique 1963–1965. Paris: Office de la Recherche Scientifique et

Technique Outre-Mer.Unfortunately, hypotheses of maintenance must be dis-




casimir, m., and a. rao. 1995. Prestige, possessions, and gray, s. 1994. Comparison of effects of breastfeeding practiceson birth spacing in three societies: Nomadic Turkana, Gainj,progeny: Cultural goals and reproductive success among the

Bakkarwal. Human Nature 6:241–72. and Quechua. Journal of Biosocial Science 26:69–90.hailemariam, a. 1991. Fertility levels and trends in Arsi andcastle, s. 1992. Intra-household variation in illness manage-

ment and child care in rural Mali. Ph.D. diss., University of Shoa regions of central Ethiopia. Journal of Biosocial Science23:387–400.London, London, U.K.

cavall i - sforza, l., p. menozzi, and a. piazza. 1994. handwerker, p. 1983. The first demographic transition: Ananalysis of subsistence choices and reproductive consequences.The history and geography of human genes. Princeton:

Princeton University Press. American Anthropologist 85:5–27.harako, r. 1981. ‘‘The cultural ecology of hunting behaviorchu, p. 1996. Fertility decline in Mongolia: Who, when, where,

how, and why? Paper presented at the 95th Annual Meeting of among Mbuti Pygmies in the Ituri Forest, Zaire,’’ in Omnivo-rous primates. Edited by R. Harding and G. Teleki, pp. 499–the American Anthropological Association, San Francisco,

Calif. 555. New York: Columbia University Press.hassan, f. 1973. On mechanisms of population growth duringcohen, m. 1977. The food crisis in prehistory. New Haven:

Yale University Press. the Neolithic. current anthropology 14:535–40.henin, r. 1968. Fertility differentials in the Sudan (with refer-———. 1989. Health and the rise of civilization. New Haven:

Yale University Press. ence to the nomadic and settled populations). Population Stud-ies 22:147–64.———. 1989. ‘‘Paleopathology and the interpretation of eco-

nomic change in prehistory,’’ in Archaeological thought in ———. 1969. The patterns and causes of fertility differentials inthe Sudan. Population Studies 23:171–89.America. Edited by C. Lamberg-Karlovsky, pp. 117–32. Cam-

bridge: Cambridge University Press. henrip in, j. 1954. La fecondite des menages canadiens au de-but du XVIIIe siecle. Population 9:61–84.cohen, m., and g. armelagos. Editors. 1984. Paleopa-

thology at the origins of agriculture. New York: Academic henry, l. 1953. Aspects demographiques d’une region rurale del’Iran. Population 8:590–92.Press.

colson, e. 1958. Marriage and family among the Plateau hewlett, b. 1991. Demography and childcare in preindustrialsocieties. Journal of Anthropological Research 47:1–37.Tonga of Northern Rhodesia. Manchester: Manchester Univer-

sity Press. hill, a. 1985. ‘‘The recent demographic surveys in Mali andtheir main findings,’’ in Population, health, and nutrition incrognier, e., c. bernis, s. elizondo, and c. varea.

1993. The pattern of reproductive life in a Berber population of the Sahel. Edited by A. G. Hill, pp. 41–64. London: Routledgeand Kegan Paul.Morocco. Social Biology 40:191–99.

cruz-coke, r. , a. cri stoffanini, m. asp illaga, and hill , a., and s. randall. 1985. ‘‘Issues in the study of thedemography of Sahelian pastoralists and agro-pastoralists,’’ inf. b iancani. 1966. Evolutionary forces in human popula-

tions in an environmental gradient in Africa, Chile. Human Bi- Population, health, and nutrition in the Sahel. Edited by A. G.Hill, pp. 3–14. London: Routledge and Kegan Paul.ology 38:421–38.

early, j. 1970. Demographic profile of a Maya community. Mil- howell, n. 1979. Demography of the Dobe !Kung. New York:Academic Press.bank Memorial Fund Quarterly 43:167–78.

early, j., and j. peters. 1990. The population dynamics of hurtado, a. m., and k. r. hill. 1989. Seasonality in a for-aging society: Variation in diet, work effort, fertility, and sex-the Mucajai Yanomamo. New York: Academic Press.

ell ison, p. 1991. ‘‘Reproductive ecology and human fertility,’’ ual division of labor among the Hiwi of Venezuela. Journal ofAnthropological Research 34:293–346.in Applications of biological anthropology to human affairs.

Edited by C. G. N. Mascie-Taylor and G. W. Lasker, pp. 14–54. irons, w. 1980. ‘‘Is Yomut social behavior adaptive?’’ in Socio-biology: Beyond nature/nurture? Edited by G. Barlow and J. Sil-Cambridge: Cambridge University Press.

———. 1994. Advances in human reproductive ecology. Annual verberg, pp. 417–63. Boulder: Westview Press.j en ike, m. 1995. Variation in body fat and muscle mass: Re-Review of Anthropology 23:255–75.

felsenstein, j. 1985. Phylogenies and the comparative sponses to seasonal hunger among tropical horticulturalists,Zaire. Ecology of Food and Nutrition 34:227–49.method. American Naturalist 125:1–15.

ferro-luzzi, a. 1990. Seasonal energy stress in marginally j enness, d. 1992. The life of the Copper Eskimos. Report ofthe Canadian Arctic Expedition, 1913–1918, vol. 12.nourished rural women: Interpretation and integrated conclu-

sions of a multicentre study in three developing countries. Eu- jones, f. 1963. A demographic survey of the Aboriginal popula-tion of the Northern Territory, with special reference to Bath-ropean Journal of Clinical Nutrition 44:41–46.

f irschein, i. 1984. ‘‘Demographic patterns of the Garifuna urst Island Mission. Australian Institute of Aboriginal Studies.Occasional Papers in Aboriginal Studies 7.(Black Carib) of Belize,’’ in Black Caribs: A case study in bio-

cultural adaptation. Edited by M. Crawford, pp. 67–94. New konner, m., and c. worthman. 1980. Nursing frequency,gonadal function, and birth spacing among !Kung hunter-gather-York: Plenum Press.

f ix, a. 1977. The demography of the Semai Senoi. Ann Arbor: ers. Science 207:788–91.lambert, b. 1975. ‘‘Makin and the outside world,’’ in PacificMuseum of Anthropology.

fraccaro, m. 1959. Fertility differential in two Lappish popu- atoll populations. Edited by V. Carroll, pp. 212–85. Honolulu:University of Hawaii Press.lations. American Journal of Human Genetics 11:92.

ganiage, j. 1960. La population europeenne de Tunis au mi- lang, h., and u. holter. 1996. Arab camel nomads in theNorth West Sudan: The Northern Mahria from a census pointlieu du XIXeme siecle: Etude demographique. Universite de

Tunis Publications de la Faculte des Lettres, 4eme serie (His- of view. Anthropos 91:19–32.larsen, c. 1995. Biological changes in human populationstoire) 2.

ganon, m. 1975. ‘‘The nomads of Niger,’’ in Population with agriculture. Annual Reviews of Anthropology 24:185–213.growth and socioeconomic change in West Africa. Edited by J.

Caldwell, N. Addo, S. Gaisie, A. Igun, and P. Olusanya, pp. lee, r. 1972. ‘‘Population growth and the beginnings of seden-tary life among the !Kung Bushmen,’’ in Population growth:694–700. New York: Columbia University Press.

glyn-jones, m. 1953. The Dusun of the Penampang Plains in Anthropological implications. Edited by B. Spooner. Cam-bridge: MIT Press.North Borneo. London: Foreign Office.

goldste in, m. 1976. Fraternal polyandry and fertility in Ne- ———. 1979. The !Kung San. Cambridge, England: CambridgeUniversity Press.pal. Human Ecology 4:223.

gomes, a. 1988. Demographic implications of villagisation leslie, p., k. campbell, m. l ittle, and c. kigondu.1996. Evaluation of reproductive function in Turkana womenamong the Semang of Malaysia. Paper presented at the 5th In-

ternational Conference on Hunting and Gathering Societies, with enzyme immunoassays of urinary hormones in the field.Human Biology 68:95–117.Darwin, Australia.




leslie, p. w., k. l. campbell, and m. a. little. 1993. ductive behavior in a tribal population (the Baiga) in Madhya-Pradesh, India. Annals of Human Biology 22:235–46.Pregnancy loss in nomadic and settled women in Turkana,

Kenya: A prospective study. Human Biology 65:237–54. roth, e. 1985. A note on the demographic concomitants of sed-entism. American Anthropologist 87:380–82.l indtjørn, b., t. alemu, and b. bjorvatin. 1993. Pop-

ulation growth, fertility, mortality, and migration in drought- ruhlen, m. 1991. 2d edition. A guide to the world’s languages.Vol. 1. Classification, London: Edward Arnold.prone areas in Ethiopia. Transactions of the Royal Society of

Tropical Medicine and Hygiene 87:24–28. salzano, f., and r. de ol ive ira. 1970. Genetic aspectsof the demography of Brazilian Terena Indians. Social Biologymace, r., and m. pagel. 1994. The comparative method in

anthropology: A phylogenetic approach. current anthropol- 17:217–23.salzano, f., j. neel, and d. maybury-lewis. 1967. Fur-ogy 35:540–64.

———. 1996. ‘‘Tips, branches, and nodes: Seeking bio-cultural ad- ther studies on the Xavante Indians. American Journal of Hu-man Genetics 19:463–89.aptation by comparative study,’’ in Human nature: A critical

reader. Edited by L. Betzig. New York: Oxford University sattenspiel, l., and h. harpending. 1983. Stable popu-lations and skeletal age. American Antiquity 48:489–98.Press.

malaurie, j. 1956. The last kings of Thule: A year among the schrire, c., and w. steiger. 1974. A matter of life anddeath: An investigation into the practice of infanticide in thePolar Eskimos of Greenland. London: Allen and Unwin.

malaurie, j., l. tabah, and j. sutter. 1952. L’isolat es- Arctic. Man 9:161–84.sellen, d. w., and r. mace, n.d. A phylogenetic analysisquimau de Thule (Groenland). Population 7:675–92.

marriott, h. 1993. Determinants of natural fertility differen- of the relationship between sub-adult mortality and mode ofsubsistence. MS.tials: A comparative survey of the rural populations of the in-

ner Niger Delta of Mali. Ph.D. diss., University of London, Lon- s ibajuddin, s. 1984. Reproduction and consanguinity amongChenchus of Andrha Pradesh. Man in India 64:181–92.don, U.K.

mitchell, j. 1949. An estimate of fertility in some Yao ham- s innett, p., and h. white. 1973. Epidemiological studiesin a highland population of New Guinea: Environment, cul-lets in Liwonde district of southern Nyasaland. Africa 29:293–

308. ture, and health status. Human Ecology 1:245–77.smith, t. 1960. The Cocos-Keeling Islands: A demographic lab-molnos, a. 1972. Cultural source materials for population

planning in East Africa. Vol. 1. Nairobi: East African Publish- oratory. Population Studies 14:94–130.spss, inc. 1993. SPSS proprietary software, release 6.0. Chi-ing House.

morgan, k. 1973. ‘‘Historical demography of a Navajo commu- cago.stephens, c. 1975. On mechanisms of population growth. cur-nity,’’ in Methods and theories of anthropological genetics. Ed-

ited by M. Crawford and P. Workman, pp. 263–314. Albu- rent anthropology 16:228–29.sussman, r. 1972. Child transport, family size, and increase inqurque: University of New Mexico Press.

muhsam, h. 1950. Fertility and reproduction of the Beduin. human population during the Neolithic. current anthropol-ogy 13:258–59.Population Studies 4:354–63.

murdock, g. 1965. Ethnographic atlas. Pittsburgh: University takasaka, k. 1986. Post-partum amenorrhea, waiting time toconception, and prevalence of pregnancy in a Sundanese ag-of Pittsburgh Press.

———. 1981. Atlas of world cultures. Pittsburgh: University of ricultural community. Human Biology 58:933–44.talukdar, s. 1979. Genetical demography of two Dule BagdiPittburgh Press.

murdock, g., and d. white. 1969. Standard cross-cultural demes. Man in India 59:43–70.thomas, j. 1963. Les Ngabaka de la Lobaye: Le depeuplementsample. Ethnology 8:329–69.

neel, j., and k. weiss. 1975. The genetic structure of a rural chez une population forestiere de la Republique Centra-fricaine. Paris: Mouton.tribal population, the Yanomamo Indians. American Anthropol-

ogist 42:25–52. traore, b. 1985. ‘‘Demographic characteristics and trendsamongst the nomads of Mauritania,’’ in Population, health,osgood, c. 1936. The distribution of the northern Atha-

paskan Indians. Yale University Publications in Anthropol- and nutrition in the Sahel. Edited by A. G. Hill, pp. 105–15.London: Routledge and Kegan Paul.ogy 7.

pagel, m. 1992. A method for the analysis of comparative data. trussell, j. 1979. ‘‘Natural fertility: Measurement and use infertility models,’’ in Natural fertility. Edited by H. Leridon andJournal of Theoretical Biology 156:431–42.

pandya, v. 1987. Above the forest: A study of the physical, be- J. Menken, pp. 29–64. Liege: Ordina Editions.van den eerembeemt, m.-l. 1985. ‘‘A demographic profilehavioral, and cognitive development of Pygmy children. Ph.D.

diss., University of Chicago, Chicago, Ill. of the Fulani of central Mali with special emphasis on infantand child mortality,’’ in Population, health, and nutrition inpennington, r. 1992. Did food increase fertility? Evalua-

tion of !Kung and Herero history. Human Biology 64:497– the Sahel. Edited by A. G. Hill, pp. 79–104. London: Routledgeand Kegan Paul.521.

———. 1996. Causes of early human population growth. Ameri- wood, j. 1990. Fertility in anthropological populations. AnnualReviews of Anthropology 19:211–42.can Journal of Physical Anthropology 99:259–74.

———. 1993. The structure of an African pastoralist community. wood, j., d. lai, p. johnson, k. campbell, and i.maslar. 1985. Lactation and birth spacing in highland NewOxford: Clarendon Press.

pospi s il, l. 1963. Kapauku Papuan economy. New Haven: Guinea. Journal of Biosocial Science suppl. 9:159–73.Yale University Press.

potter, r., m. new, j. wyon, and j. gordon. 1965. Afertility differential in eleven Punjab villages. Milbank Memo-rial Fund Quarterly 43:185–201.

purv is, a., and a. rambaut. 1995. Comparative analysisby independent contrasts (CAIC): An Apple Macintosh applica-tion for analysing comparative data. Computer Applied Biosci-ences 11:247–51.

randall, s. 1994. ‘‘Pastoralists demographically different fromsedentary agriculturalists?’’ in Environment and populationchange. Edited by B. Zaba and J. Clarke, pp. 325–47. Liege: Or-dina Editions/IUSSP.

ray, a., and e. roth. 1991. Indian tribal fertility patternsfrom Orissa. Man in India 71:235–39.

reddy, p., and b. modell. 1995. Consanguinity and repro-




(Brooks et al. 1995, Yellen et al. 1995; S. Ambrose, per-Bone Artefacts from the Middle sonal communication; R. Yates, personal communica-tion), whether MSA people in Africa can be regarded asStone Age at Blombos Cave,behaviourally modern. Some researchers suggest that

Southern Cape, South Africa1modern behavioural traits are apparent in Africa possi-bly as far back as 100,000 years ago (H. J. Deacon 1995,Yellen et al 1995). Others argue that MSA people, com-

christopher henshilwood and judith pared with their fully modern LSA descendants, weresealy not behaviourally modern because, among other things,Department of Archaeology, University of Cape they seldom worked organic materials, lacked symbolicTown, Private Bag, Rondebosch 7700, South Africa. behaviour, and did not fish (see Thackeray 1992, Klein13 v 97 1995). Here we examine new evidence from Blombos

Cave indicating that by at least 40,000 years ago MSAThe Middle Stone Age in Africa (approximately equiva- people at this site were making bone artefacts, scoringlent to the Middle Palaeolithic) is currently the focal bone with possible symbolic connotations, and fishing.point of global research investigating the origins of In Later Stone Age sites in Africa and Upper Palaeo-‘‘modern’’ humans. Two main issues are in the spot- lithic assemblages in Eurasia, formal, standardised toolslight: whether anatomically modern humans evolved in in a variety of materials and works of art are regularAfrica between 150,000 and 50,000 years ago and components of artefact assemblages. They are muchwhether fully modern human behaviour appeared con- rarer, if, indeed, they are present at all, in Middle Stonetemporaneously with the modern human form or first Age or Middle Palaeolithic contexts (Klein 1995, Mel-developed in Africa during the latter part of the Middle lars 1993, Thackeray 1992). The most notable exceptionStone Age (MSA), around 50,000–40,000 years ago (cf. to this pattern is the barbed and unbarbed bone points,Klein 1994, 1995), perhaps then spreading to Eurasia as well as a bone ‘‘dagger,’’ reported from a riverside siteand elsewhere (Mellars 1993, Klein 1995). in the Semliki Valley in Zaire and dated by association

The model for early modern behaviour is derived to 82,000–174,000 years ago (Brooks et al. 1995, Yellenfrom a suite of characteristics associated with Later et al. 1995). These are anomalous, since larger, well-ex-Stone Age (LSA) or Upper Palaeolithic people (see Klein cavated assemblages from sites of a similar age have not1994, 1995) and is recognisable in the archaeological yielded formal bone tools. Moreover, the contexts inrecord, both in Africa and Eurasia, from around 40,000– which these artefacts appear are clearly complex, since30,000 years ago. By comparison, Middle Stone Age or there are varying ESR, TL, and U-series dates for theMiddle Palaeolithic (MP) people prior to around sites. In southern Africa, ground warthog-tusk ‘‘dag-40,000–45,000 years ago are often deemed ‘‘non-mod- gers’’ were recovered from Border Cave, some thoughtern’’ in their behaviour (see Thackeray 1992, Mellars to be as much as 80,000 years old (Beaumont, de Vil-1993). Critical questions are when fully modern behav- liers, and Vogel 1978), but the dating of these by associa-iour is first recognisable, what elements of this behav- tion with the strata in which they were recovered is ten-ioural ‘‘package’’ are evident prior to ca. 40,000 years uous (Klein 1989, Sillen and Morris 1996, Thackerayago, and, in the light of recently recovered evidence 1992). A similar case can be made for the several bone

artefacts, including a bone point, found at Kabwe (Clarket al. 1947). A bone point recovered from the Howiesons1. 1997 by The Wenner-Gren Foundation for Anthropological Re-Poort levels of Klasies River Mouth (Singer and Wymersearch. All rights reserved 0011-3204/97/3805-0013$1.00. We

thank Cedric Poggenpoel for his help in excavating the site and an- 1982) may be a more recent intrusion into older layersalysing the fish. Stanley Ambrose, Graham Avery, Geoff Bailey, (Klein 1995, Thackeray 1992). Two notched bones haveDesmond Clark, Richard Klein, Paul Mellars, Duncan Miller, been found at Klasies River Mouth and two at ApolloRoyden Yates, and others have discussed the finds with us, contrib-

11 in Namibia (Wendt 1976), though these are not for-uted valuable insights, and in several cases drawn our attention tomal tools. In summary, the Semliki material apart,relevant articles in the literature. Francesco d’Errico kindly exam-

ined the incised bone and suggested that the incisions are likely ‘‘Formal bone artefacts are rare or absent in MSA sites,to have been made by the same stone tool. Tjeerd van Andel gave and where they occur they could easily represent unde-informative advice and drew figure 3. Richard Klein kindly ar- tected intrusions from LSA levels’’ (Klein 1989:377).ranged for radiocarbon dates to be run at the Lawrence LivermoreLaboratory and organised funding. Peter Nilssen examined thebone point from unit TOB and suggested that it might be a seal the blombos findsfibula. Carbon and nitrogen analyses were performed by G. Be-nincasa of the Department of Chemistry at the University of Cape Preliminary excavations at Blombos Cave in 1993 andTown. Trevor Sewell advised on methods of examining and record-

1997 (figs. 1 and 2) have yielded Later Stone Age depos-ing the artefacts. Analytical work was supported by the Foundationits dating to within the last 2,000 years. Below these isfor Research Development and the University of Cape Town.

Grants to CH from the Anglo-American Chairman’s Fund, the a sterile layer, and deeper still is a Middle Stone AgeUniversity of Cape Town, the Cambridge Overseas Trust, the Cen- industry, the Still Bay. The cave is today some 100 me-tre for Science Development, the Louis Leakey Foundation, the tres from the coast and 35 metres above sea level; theSmuts Memorial Fund, the Overseas Research Students Award

MSA occupation probably dates to a period of similarlyScheme, and the Swan Fund, Oxford, supported the excavations.We thank them all. high sea level, since shellfish and fish in the MSA levels




table 1Radiocarbon Determinations on MSA Shelland Charcoal

No. Material Level Date b.p. (S.D.)

CAMS 15874 charcoal 11 39,200 (1,680)CAMS 15875 shell 9 35,670 (1,600)CAMS 15876 shell 9 35,640 (1,440)CAMS 15877 shell 11 32,600 (1,310)Pta-6557 shell 9 38,500 (1,760)

note: Calibration of the radiocarbon time scale at this period byU-series dating shows that 14C dates are about 3,000 years ‘‘tooyoung’’ by 30,000 b.p. (Bard et al. 1990). Increased atmospheric14C at about 40,000 b.p., as argued by Sternberg and Damon(1992), would additionally make dates of that period systemati-cally ‘‘too young.’’ See also Laj, Majaud, and DuPlessy (1996).

Fig. 1. Location of Blombos Cave. adjacent to Blombos Cave, van Andel (personal commu-nication) has estimated that the site-to-shore distancewas approximately 1 km at ca. 50,000 years ago andeven less at ca. 80,000 and 100,000 years ago (fig. 3). Atare almost as abundant a component of the deposits as

in the Later Stone Age layers. Five radiocarbon determi- the Last Glacial Maximum, ca. 17,000 years ago, the seawas more than 160 km from Blombos Cave, and for thenations from two laboratories on MSA shell and char-

coal have yielded results which the laboratories have in- balance of the Last Glacial period it was between 1 and5 km distant. Since the Still Bay industry is a late ex-dicated are best regarded as infinite dates (table 1). If the

Blombos MSA occupation is indeed older than 40,000 pression of the MSA, we expect that the 50,000–60,000years ago date is the most likely.years, the most likely dates are those when the sea was

closest to the cave, although it is possible that the ma- The MSA levels at Blombos have yielded remainswhich are unusual in several respects. Stone artefactsrine assemblage was transported to the site at a period

when the sea was more distant. Oxygen isotope (OI) ra- are dominated by bifacial points: 52 of a total of 98 for-mal (retouched) tools are bifacial points or parts thereof,tios on marine shell indicate that warmer interstadials

during the Last Glacial occurred at around 50,000– and an additional 10 artefacts are unifacial points. Theyare made mainly by pressure flaking fine-grained sil-60,000 years ago (OI stage 3), 80,000 years ago (OI stage

5a), and 100,000 years ago (OI stage 5c) (Shackleton crete (fig. 4). Similar points are known from other SouthAfrican MSA sites (J. Deacon 1995, Thackeray 1992,1982, van Andel 1989). From the offshore topography

Fig. 2. Stratigraphy of Middle Stone Age layers at Blombos Cave (west section). Positions of bone artefacts insquares are approximate.




Fig. 5. TOB bone point; the flat end is broken.Length 61 mm.

(Klein 1989:321) or subsistence choices on the part ofFig. 3. Distance from Blombos Cave to shore duringthe inhabitants (Deacon 1989).the past 120,000 years.

During excavations in 1993 and 1997 more than 20items of worked bone were recovered from the Still Baylevels at Blombos Cave, including ‘‘awls,’’ a possibleVolman 1984), though nowhere else are they dominant.peg, two bone points, and an incised bone. Most speci-Associated faunal remains include a number of speciesmens are shaped at least partly by utilisation as imple-of shellfish, known from other MSA sites, and large fishments for piercing, gouging, or drilling. A detailed anal-include the species Cymatoceps nasutus, Aries feliceps,ysis of the whole worked-bone industry is pending.and Liza richardsonii. Large fish are notably absentHere we focus on two specimens which were intention-from MSA levels at all coastal southern African sites,ally shaped by grinding and polishing to symmetricalincluding Klasies River Mouth main site, approxi-bone points very similar to Later Stone Age examplesmately 330 km east north-east of Blombos Cave. Herefrom a number of southern African sites. These arethe bulk of the MSA deposits are probably older thanstandardised (‘‘formal’’) tools.Blombos, dating to 120,000–80,000 years ago (Deacon

The bone points (figs. 5–7) were recovered from unitand Geleijnse 1988). Small fish may have been broughtTOB, square E3, and unit A1, square E4, respectively.into the Klasies site in the stomachs of seals or by seaThe TOB bone point is snapped at one end; the re-birds and therefore do not represent human food re-maining fragment is 61 mm long with a maximum di-mains. The absence of fish at Klasies has variously beenameter of 7 mm. It was smoothed to a point by grinding,considered to reflect premodern human behaviourand striations are clearly visible, especially near the tip.The tip is darker in colour than other parts of the arte-fact and may have been fire-hardened. Examination ofthe cross section shows that the outer bone is relativelyporous while the core is trabecular bone. The porouscharacter is similar to that of seal bone, and the point

Fig. 6. TOB bone point: close-up of area withindentations probably derived from binding to a haft.Fig. 4. Bifacially flaked Still Bay stone points, the

largest 83 mm long. (This margin is not visible in figure 5.)




made by dragging the point of a stone tool repeatedlyover the bone surface (d’Errico, personal communica-tion), perhaps with the aim of producing a decorativeimage.

Although the bone points and the incised fragmentwere found in Middle Stone Age levels, there was somedoubt whether they really belonged there. A possibleexplanation was that they had slipped down from over-lying Later Stone Age levels. The point from unit TOBweighs only 1.3 g, and accelerator mass spectrometerdating would destroy most of it. Instead, we chose tomeasure the carbon and nitrogen contents of the pointsand of the incised bone for comparison with carbon and

Fig. 7. A1 bone point, broken at distal end. Length nitrogen measurements on non-artefactual bone from55 mm. both LSA and MSA levels in the site.

carbon and nitrogen analysismay have been made from a seal fibula. On the sharpestmargin, near the snapped end, is a series of slight inden- In addition to the bone points and incised bone, carbontations visible under oblique lighting. The regular spac- and nitrogen levels were determined in six non-arte-ing of these indentations and their position on the shaft factual bones from the MSA and five from the LSA.indicate that they may have been caused by tight bind- Specimens were surface-cleaned, approximately 5 mg ofing which attached the point to a haft or to some other bone drilled out using a diamond-tipped drill bit, andcomponent of a composite tool. When fresh, the bone the powder collected for analysis. C and N assays werewould have been fairly soft, allowing binding materials performed using a Fisons Instruments elemental ana-to leave a permanent imprint. lyser, model 1108. Each specimen was analysed in du-

The A1 point is also snapped at its distal end; it is 55 plicate. Results are given in table 2, and mean valuesmm long with a maximum diameter of 7 mm. Made on for each specimen are plotted in figure 9.dense, compact bone, it is ground to a point and highly As expected, both carbon and nitrogen were muchpolished. Patches of what appear to be ochre are visible more abundant in Later than in Middle Stone Age bonesin the polish. A longitudinal groove runs from the butt as a result of the more complete preservation of the or-end almost to the point. Under magnification, minute ganic component of the bone in more recent specimens.scratches on the bone surface are suggestive of use wear. Gradual, time-dependent breakdown of bone proteinThere is no evidence of burning or charring on this bone and subsequent leaching of the degraded material occurpoint. in most depositional contexts.

The incised bone (fig. 8) was recovered from unit Results for the incised bone and the bone points areSAN, square E3. It is a flat fragment of bone, perhaps closest to the values for MSA bones: the incised bonefrom a scapula. It has a series of subparallel grooves, and the bone point from A1 cluster tightly with non-probably all made by the same stone tool. It is possible artefactual MSA bones. The TOB bone point hasthat these are simply cut marks, produced by repeated slightly higher C and N contents. We have suggestedattempts to remove meat from the bone. If this were the above that the darker colour of the tip may indicate thatcase, however, we might expect the long edge of a blade- it was fire-hardened. Burning destroys bone protein:type tool to have been used; here the incisions were heating bone to 200 to 400° C tends to convert carbon

in protein to inorganic carbon (‘‘char’’); at temperaturesabove 400° C organics are burned away entirely (Sillenand Hoering 1993). Could the TOB point actually derivefrom LSA levels but have low C and N contents as aresult of heating-induced loss of protein?

We think not. First, the very tip of the TOB bonepoint is slightly darkened; the remainder of the speci-men is not. Bone powder for analysis was drilled fromnear the snapped surface, at the opposite end from theputatively charred area. Second, the carbon and nitro-gen contents are inconsistent with such a hypothesis.C/N ratios for LSA bones in table 2 lie between 3.1 and3.7. Values for modern bone collagen (the major boneprotein) range from 2.9 to 3.6 (DeNiro 1985); in addi-tion, there is a small amount of carbon in bone apatite.The LSA bones from Blombos are essentially modern inoverall composition. Laboratory experiments by SillenFig. 8. Incised bone. Length 28 mm.




table 2Results of Carbon and Nitrogen Analyses

Result 1 Result 2 Mean

Layer Unit Square Period Description C N C N C N C/N Ratio

3 MC3 F2 LSA Bone 11.58 3.11 11.77 3.27 11.68 3.19 3.74 MC3 E3 LSA Bone 13.10 3.99 12.37 3.68 12.74 3.84 3.34 MC3 E3 LSA Bone 15.27 4.83 15.25 4.84 15.26 4.84 3.25 MC4 E4 LSA Bone 13.44 3.84 13.17 3.81 13.31 3.83 3.55 MC4 E4 LSA Bone 15.34 4.96 15.29 4.95 15.32 4.96 3.17 BR1 E3 MSA Bone 2.91 0.19 2.62 0.18 2.77 0.19 14.97 BR1 E4 MSA Bone 3.63 0.22 3.57 0.23 3.60 0.23 16.08 TOB E3 MSA Bone 2.65 0.21 2.82 0.23 2.74 0.22 12.49 PIP E4 MSA Bone 3.08 0.15 3.16 0.15 3.12 0.15 20.8

10 SAN E4 MSA Bone 2.90 0.13 2.86 0.12 2.88 0.13 23.010 SAN E4 MSA Bone 2.93 0.10 3.00 0.10 2.97 0.10 29.710 SAN E3 MSA Incised bone 2.99 0.11 3.04 0.12 3.02 0.12 26.210 TOB E3 MSA Bone point 4.13 0.45 4.04 0.49 4.09 0.47 8.711 A1 E4 MSA Bone point 3.28 0.04 3.33 0.04 3.31 0.04 82.6

and Hoering (1993) have shown that charred modern bones. Its surface was worked (smoothed) in order tomake it more resistant to damage; this may also havebone retains significant quantities of nitrogen. C/N ra-

tios of chars were never higher than 6.0 in experimental rendered it slightly less susceptible to chemical degra-dation.bones heated to various temperatures up to 600°C. The

C/N ratio of the TOB bone point is 8.7—in other words, The results for the incised bone and bone points aresubstantially separated from those for LSA bones. Theseit contains very little nitrogen indeed. This makes it un-

like charred modern bones but very like the non-arte- artefacts are part of the MSA assemblage.factual MSA bones. Its C and N contents are, however,slightly higher than those of non-artefactual MSA

significance of results

The two bone points are the first standardised bone ar-tefacts in the world that have been demonstratedthrough direct analysis to be associated with assem-blages older than about 40,000 years. Taking into ac-count the rarity of bone tools from MSA sites, the re-covery of two bone points in addition to more than 20other items of worked bone from the limited excava-tions at Blombos Cave clearly indicates that both for-mal and ad hoc bone tools did form a part of the toolkit of some MSA people.

The demonstration that the Blombos bone points areindeed Middle Stone Age prompts reconsideration ofother infrequent finds of bone tools with MSA assem-blages. Certainly the recent recovery of Middle Pleisto-cene 400,000-year-old wooden spears at Schoningen(Thieme 1997) and the bone tools from the Semliki Val-ley (Brooks et al. 1995, Yellen et al. 1995) suggest thatorganic raw materials were being shaped for use at anearlier stage than previously recognised. The bone pointfrom Klasies River Mouth may well be reliably associ-ated with the Howiesons Poort levels in which it wasFig. 9. Results of carbon and nitrogen analyses of

LSA and MSA bone, compared with bone points and found; the excavators remarked that ‘‘the bone isstained and in identical condition to other bones in theincised bone. For LSA bones, mean carbon content 6

one s.d. is 13.66 6 1.60%, mean nitrogen content 4.13 same layer, being quite unlike the yellow, greasy bonestypical of the LSA middens’’ (Singer and Wymer 1982:6 0.75%. For non-artefactual MSA bones, mean

carbon content is 3.01 6 0.32%, mean nitrogen 115). There are, nevertheless, only a handful of bonetools from MSA contexts; possibly bone tools did notcontent 0.17 6 0.05%.




deacon, h. j. 1989. ‘‘Late Pleistocene palaeoecology and ar-play a significant role in most MSA artefact kits, al-chaeology in the southern Cape, South Africa,’’ in The humanthough this seems not to have been the case during therevolution: Behavioural and biological perspectives on the ori-

period of the as yet undated Still Bay industry. The ab- gins of modern humans. Edited by P. Mellars and C. Stringer,sence of worked bone at other sites containing Still pp. 547–64. Edinburgh: Edinburgh University Press.

———. 1995. Two Late Pleistocene-Holocene archaeological de-Bay–type lithics may be due to poor preservation; cur-positories from the southern Cape, South Africa. South Africanrently no other known Still Bay–type sites contain or-Archaeological Bulletin 50:121–31.ganic materials. deacon, h. j., and v. b. gelei jnse. 1988. The stratigra-

The bone points and the finely made stone points phy and sedimentology of the main site sequence, KlasiesRiver, South Africa. South African Archaeological Bulletin 43:from the Middle Stone Age levels at Blombos Cave5–14.comprise a sophisticated technology. We have direct ev-

deacon, j. 1995. An unsolved mystery at the Howieson’sidence that the bone point was hafted, probably to makePoort name site. South African Archaeological Bulletin 50:

a spear or harpoon. The stone points were most likely 110–20.similarly lashed onto handles or shafts. Other MSA de niro, m. j. 1985. Postmortem preservation and alteration of

in vivo bone collagen isotope ratios in relation to palaeodietarysites have also yielded evidence which can be taken toreconstruction. Nature 317:806–9.indicate hafting, for example, a blade with traces of

klein, r. g. 1989. ‘‘Biological and behavioural perspectives onmastic from the uppermost MSA at Apollo 11 (Wendt modern human origins in southern Africa,’’ in The human rev-1976). olution: Behavioural and biological perspectives on the origins

of modern humans. Edited by P. Mellars and C. Stringer, pp.At Blombos, the bone and stone technology and food529–46. Edinburgh: Edinburgh University Press.remains differ from earlier MSA assemblages and open

———. 1994. ‘‘Southern Africa before the Iron Age,’’ in Integra-a new window to our understanding of the developmenttive paths to the past: Paleoanthropological advances in hon-

of modern human behaviour. The co-occurrence of for- our of F. Clark Howell. Edited by R. S. Corruccini and R. L. Ci-mal bone tools and the remains of large fish, almost cer- ochon, pp. 471–519. Englewood Cliffs: Prentice-Hall.

———. 1995. Anatomy, behavior, and modern human origins.tainly caught for food, may signal significant develop-Journal of World Prehistory 9:167–98.ment of technology and subsistence practices in late

laj, c., a. mazaud, and j-c. du plessy. 1996. Geomag-MSA times. Pressure-flaking of fine-grained raw materi- netic intensity and 14C abundance in the atmosphere and oceanals to produce consistently shaped bifacial points may during the past 50 kyr. Geophysical Research Letters 23:2045–

48.be a further step towards modern human behaviour, butmellars, p. 1993. ‘‘Archaeology and the population-dispersalas the chronological positions of pressure-flaked points

hypothesis of modern human origins in Europe,’’ in The originin the MSA Middle Palaeolithic are unclear, it is uncer-of modern humans and the impact of chronometric dating. Ed-

tain if these tools can be interpreted in this way. Parel- ited by M. J. Aitken, C. B. Stringer, and P. A. Mellars, pp. 196–lel incisions on the bone fragment may be decorative or 216. Princeton: Princeton University Press.

shackleton, n. j. 1982. ‘‘Stratigraphy and chronology of thesimply cut marks, but until further evidence of this na-KRM deposits: Oxygen isotope evidence,’’ in The Middle Stoneture is recovered we must reserve judgement on itsAge at Klasies River Mouth in South Africa. Edited by R.maker’s intentions. Singer and J. Wymer, pp. 194–99. Chicago: University of Chi-

There is a dearth of well-preserved, carefully exca- cago Press.s illen, a., and a. morris. 1996. Diagenesis of bone fromvated, and well-curated MSA assemblages of this age

Border Cave: Implications for the age of the Border Cave homi-from southern Africa or elsewhere in Africa. Further ex-nids. Journal of Human Evolution 31:499–506.cavations at Blombos and other African sites should

s inger, r., and j. wymer. Editors. 1982. The Middlegive us a better picture of the late Middle Stone Age and Stone Age at Klasies River Mouth in South Africa. Chicago:lay a stronger foundation for assessing the questions University of Chicago Press.

sternberg, r. s., and p. e. damon. 1992. Implications ofwhether modern human behaviour dates to the Middledipole moment secular variation from 50,000–10,000 years forStone Age and whether its origins are indeed in Africa.the radiocarbon record. Radiocarbon 34:189–98.

thackeray, a. i. 1992. The Middle Stone Age south of theLimpopo River. Journal of World Prehistory 6:385–440.References Cited thieme, h. 1997. Lower Palaeolithic hunting spears from Ger-many. Nature 385:807–10.bard, e., b. hamelin, r. g. fa irbanks, and a. z ind-

van andel, t. h. 1989. Late Pleistocene sea levels and the hu-ler. 1990. Calibration of the 14C timescale over the pastman exploitation of the shore and shelf of southern South Af-30,000 years using mass spectrometric U-Th ages from Barba-rica. Journal of Field Archaeology 16:133–54.dos corals. Nature 345:405–10.

volman, t. p. 1984. ‘‘Early prehistory of southern Africa,’’ inbeaumont, p. b., h. de villiers, and j. c. vogel.Southern African prehistory and palaeoenvironments. Edited1978. Modern man in sub-Saharan Africa prior to 49,000 yearsby R. G. Klein, pp. 169–220. Rotterdam: Balkema.b.p.: A review and evaluation with particular reference to Bor-

wendt, w. e. 1976. ‘‘Art mobilier’’ from the Apollo 11 cave,der Cave. South African Journal of Science 74:409–19.South West Africa: Africa’s oldest dated works of art. South Af-brooks, a. s., d. m. helgren, j. s. cramer, a.rican Archaeological Bulletin 31:5–11.franklin, w. hornyak, j. m. keating, r. g. klein,

yellen, j. e., a. s. brooks, e. cornel issen, m. j.w. j. r ink, h. schwarcz, j. n. leith smith, k.mehlman, and k. stewart. 1995. A Middle Stone Agestewart, n. e. todd, j. verniers, and j. e. yellen.worked bone industry from Katanda, Upper Semliki Valley,1995. Dating and context of three Middle Stone Age sites withZaire. Science 268:553–56.bone points in the Upper Semliki Valley, Zaire. Science 268:

548–53.clark, j. d., k. p. oakley, l. h. wells, and j. a. c.

mc clelland. 1947. New studies on Rhodesian man. Jour-nal of the Royal Anthropological Institute 77:7–32.




month-old female is capable of becoming pregnant; ges-Reconsidering the tation takes only 63–74 days (Bolton and Calvin 1981:275; Gade 1967:214), and the number of newborns var-Archaeological Rarity of Guineaies from three to four or even more. Significantly, im-

Pig Bones in the Central Andes1mediately after delivery females are again receptive andmay become pregnant. Finally, and most important, be-cause adult male guinea pigs (kututus) are territorial

lidio m. valdez and j. ernesto valdez (Bolton 1979:278; Brothwell 1983:117), they are fre-Department of Archaeology, University of Calgary, quently killed in large numbers to prevent fighting and2500 University Dr., N.W., Calgary, AB, Canada T2N unwanted deaths. It is very common to keep only one1N4/Departamento de Ciencias Sociales, Universidad adult male with a large number of females (Gade 1967:de Huamanga, Ayacucho, Peru. 18 iv 97 214) while the rest are killed and eaten. In short, ac-

cording to ethnographic studies, guinea pig raising in-In his Peru Before the Incas, E. P. Lanning suggested volves continual killing of large numbers of animals.that guinea pigs (Cavia porcellus) might have been one One therefore wonders why their bones are seldomof the most important food animals in the ancient cen- found in archaeological sites.tral Andes (1967:18): ‘‘If we had any way of estimating To explore this question, a contemporary Peruvianthe number of guinea pigs eaten in ancient times, we central highland kitchen was chosen for excavation.3

might find that they ranked with seafood as the most This kitchen, located in the Ayacucho Valley at aboutimportant sources of protein in the ancient diet, well 2,500 m above sea level, was first built in the 1930s.ahead of the camelids and the Andean deer.’’ Lanning Guinea pigs (locally called cuyes) were introducedwas convinced that these small rodents, often kept in shortly thereafter, and the structure was used for raisingthe kitchen and usually fed table scraps, were seriously them until the early 1980s, when it was abandoned. Ac-underrepresented in the archaeological record and thus cording to the structure’s owner, who has recently es-the quantity of their bones uncovered during excavation tablished a small new kitchen a few meters from the oldwas not a true reflection of what might have been one, the average number of guinea pigs raised at a timeeaten in the past. Since Lanning’s observation, excava- in the old kitchen was 20–25 adult females and a ku-tions have been carried out at many central Andean tutu.4 Then as now, guinea pigs were raised for house-archaeological sites, and they have yielded guinea pig hold consumption only, although some might be givenbones only occasionally (e.g., Wing 1972; 1975:79; 1980; to a recently married young couple as a present.Pozorski 1976:136; 1979:175; Shimada 1982; Hastorf An excavation unit of 1 3 2 m was established at the1993:180; Burger 1992:267–68; Valdez 1988; Miller and southeast corner of the abandoned kitchen to look forBurger 1995; Pozzi-Escot and Cardosa 1986). Compared guinea pig remains. The excavations5 revealed the pres-with the quantity of bones of the South American cam- ence of two levels, one from the period when theelids, the quantity of guinea pig bones is insignificant. kitchen was in use and an overlying one from the period

The low frequency of guinea pig bone remains in ar- since it was abandoned. Most of the deposits belongedchaeological sites has led many to emphasize the role to the latter level. The first occupation was basicallyof the South American camelids as the most important represented by the kitchen floor. A hard, compact, andsource of meat in the ancient central Andes.2 Indeed, a very well-preserved floor covered the entire excavationbrief review of zooarchaeological reports for the central unit, which was basically clean. Indeed, no material re-Andes shows not only that these are entirely concerned mains were found in this level except dark burned spotswith studies of camelid bone remains (e.g., Wing 1972; in the floor itself. The owner had anticipated this result,Shimada 1982, 1985; Miller and Burger 1995) but also saying that ‘‘the kitchen was kept clean.’’ The secondthat they consider llamas (Lama glama) and alpacas level was composed entirely of postabandonment de-(Lama pacos), followed by some deer, the principal posits.6 What is interesting is that although the resi-sources of meat. This view, however, is entirely based dents of this household, like neighboring rural residentson the predominance of camelid bones and the relative of this valley, consumed not only guinea pigs but alsoabsence of guinea pig bones.

Fortunately for zooarchaeologists, many Andean 3. We thank our aunt Delfina Cardenas Palomino for allowing uscommunities still raise guinea pigs (Andrews 1975; Bol- to dig in her old kitchen and for sharing a delicious guinea pig dish

with us. We thank Isaac Cardenas for helping us in the excavations.ton 1979; Morales 1994, 1995), perhaps in the same way4. The new kitchen contains 12 adult female guinea pigs and a ku-as in ancient times. It is well known that guinea pigtutu; their number has dropped because during the conflictive pe-populations increase dramatically with very little care riod of the 1980s this residence was partially unoccupied.

(Gade 1967:219; Bolton 1979:231). In fact, a three- 5. A 4-mm sieve was used to screen the dirt removed.6. Among these were plastic remains, a shellfish, a piece of metal,a piece of rope, a small bottle and a bottlecap, two ball-point pens,seven pottery fragments, a comb, a pair of running shoes and an-1. 1997 by The Wenner-Gren Foundation for Anthropological Re-

search. All rights reserved 0011-3204/97/3805-0010$1.00. other shoe, a corncob, 23 bone fragments (most of them Europeansheep and some cow), and only 1 guinea pig bone (a juvenile right2. As at present, llamas were most likely valued as cargo animals

and alpacas as wool providers, not basically as meat animals. tibia).




corn, the excavations failed to reveal any considerable mained one of the important meat sources in the cen-tral Andes and beyond them (Morales 1995). It shouldremains of either.

The owner explained why there were no guinea pig be clear from this that not all of what was eaten in thepast is represented in the archaeological record.remains in the abandoned kitchen. Guinea pigs eat all

the time, she said, and consequently the kitchen floorwas often covered with guinea pig feces, which required

References Citedsweeping at least once a day. Guinea pig manure wasregarded as one of the best soil fertilizers, and therefore andrews, d. h. 1975. On the ethnozoology of the guinea pig.the sweepings were taken to the cornfields. In other Nawpa Pacha 10–12:129–34.

bolton, ralph. 1979. Guinea pigs, protein, and ritual. Ethnol-words, most of the evidence of guinea pig raising andogy 18:229–52.consumption was daily removed as the kitchen floor

bolton, ralph, and l. calvin. 1981. ‘‘El cuy en la cul-was swept. At the same time, guinea pig bones were dis- tura peruana contemporanea,’’ in Runakinap kawsayninkupaqcarded not in the kitchen but outside. Therefore it was rurasqankunaqa. Edited by H. Lechtman and A. M. Soldi, pp.

261–326. Mexico, D.F.: Universidad Nacional Autonoma dea mistake to look for guinea pig bones in the placeMexico.where these rodents were raised. Indeed, a meal made

brothwell, don. 1983. ‘‘Why on earth the guinea-pig?’’ inof guinea pigs was served us for lunch,7 and all the bonesSite, environment, and economy. Edited by B. Proudfoot, pp.

were kept until everyone had finished and then placed 115–19. British Archaeological Reports International Series 173.on the patio for the three dogs to share. In a very few burger, richard l. 1992. ‘‘The sacred center of Chavın de

Huantar,’’ in The ancient Americas: Art from sacred land-minutes all the greasy, fragile bones had been entirelyscapes. Edited by R. F. Townsend, pp. 265–77. Chicago: Arteaten by the dogs; inspection of the patio did not locateInstitute of Chicago.a single piece of guinea pig bone. Again, the informants gade, daniel. 1967. The guinea pig in Andean folk culture.

had anticipated that this would be the case, because Geographical Review 57:213–24.hastorf, christ ine a. 1993. Agriculture and the onset of‘‘dogs like guinea pig bones’’ (and see Valdez 1995).

political inequality before the Inka. Cambridge: CambridgeA 1 3 2-m excavation conducted in an abandonedUniversity Press.kitchen where guinea pigs had been raised for nearly 50

lanning, edward p. 1967. Peru before the Incas. Englewoodyears failed to report any substantive evidence of such Cliffs: Prentice-Hall.activity. In the first place, the daily sweeping of the miller, george r., and richard l. burger. 1995. Our

father the cayman and our dinner the llama: Animal utiliza-kitchen floor had removed most of the evidence oftion at Chavın de Huantar, Peru. American Antiquity 60:421–the presence of these animals in the kitchen. Secondly,58.the fact that guinea pig bones were discarded outside morales, edmundo. 1994. The guinea pig in the Andean

had made it less likely that any bone would be found economy: From household animal to market commodity. LatinAmerican Research Review 29:129–42.inside. Finally, the direct intervention of dogs had made

———. 1995. The guinea pig: Healing, food, and ritual in the An-it unlikely that any such bones would have survived.des. Tucson: University of Arizona Press.When a test excavation such as this one fails to pro-

pozorski, sheila g. 1976. Prehistoric subsistence patternsvide any conclusive evidence that guinea pigs were and site economics in the Moche Valley, Peru. Ph.D. diss., Uni-raised and eaten in a particular place, it is likely that versity of Texas, Austin, Tex.

———. 1979. Prehispanic diet and subsistence of the Moche Val-the rarity of their bones in archaeological sites does notley, Peru. World Archaeology 11:163–84.necessarily represent their limited use in the past. In-

pozzi -escot, denise, and carmen rosa cardosa.stead, it appears that guinea pig bones do not show up 1986. El consumo de camelidos en Ayacucho desde el For-in the archaeological record because just after discard mativo a Wari. Ayacucho: Universidad de Huamanga/Lima: In-

stituto Arqueologico de Estudios Andinos.they are subjected to taphonomic processes that tend toshimada, izumi. 1994. Pampa Grande and the Mochica cul-eliminate them. Therefore, the relative abundance of

ture. Austin: University of Texas Press.camelid bone remains, for instance, compared with theshimada, melody. 1982. ‘‘Zooarchaeology of Huacaloma: Be-

remains of guinea pig bones should not be seen as con- havioral and cultural implications,’’ in Excavations at Huaca-clusive evidence of the greater importance of the former loma in the Cajamarca Valley, Peru, 1979. Edited by K. Ter-

ada and Y. Onuki, pp. 303–36. Tokyo: University of Tokyospecies. As noted by Lanning (1967), it is difficult to findPress.‘‘a way of estimating the number of guinea pigs eaten’’

———. 1985. ‘‘Continuities and changes in pattern of faunal re-in the past or of demonstrating that these small domes- source utilization: Formative through Cajamarca periods,’’ inticates were raised in ancient times. Guinea pigs are, as The Formative period in the Cajamarca Basin, Peru: Excava-

tions at Huacaloma and Layzon, 1982. Edited by K. TeradaIzumi Shimada (1994:186) argues, ‘‘greatly underrepre-and Y. Onuki, pp. 289–310. Tokyo: University of Tokyo Press.sented in the archaeological record.’’

valdez, lidio m. 1988. Los camelidos en la subsistenciaNevertheless, the simple fact that guinea pigs wereNasca: El caso de Kawachi. Boletın de Lima 57:31–35.

domesticated is indicative of their food value in ancient ———. 1995. Camelids or cuyes? Paper presented at the 14th an-times as in many contemporary Peruvian highland nual Northeast Conference on Andean Archaeology and Ethno-

history, October 20–22, Providence, R.I.communities. Most important, despite the introductionwing, elizabeth s. 1972. ‘‘Utilization of animal resources inof Old World animal species, the guinea pig has re-

the Peruvian Andes,’’ in Andes 4: Excavations at Kotosh, Peru,1963 and 1964. Edited by I. Seiichi and K. Terada, pp. 327–51.Tokyo: University of Tokyo Press.7. Two adult female guinea pigs were killed to prepare the lunch,

which was shared by four individuals. ———. 1975. Informe preliminar de los restos de fauna de la




cueva de Pachamachay, Junın, Peru. Revista del Museo Nacio- indeed West Polynesian—prehistory. In particular, thenal 41:79–80. late ceramics were perplexing in that they appeared to

———. 1980. ‘‘Faunal remains,’’ in Guitarrero Cave: Early mandate some 1,000 years after pottery was thought to havein the Andes. Edited by T. F. Lynch, pp. 149–72. New York:been abandoned in West Polynesia (Clark 1993, 1996).Academic Press.In light of this contradiction between the ’Aoa data andthe conventional model, we report here the results of atest of the late-ceramics hypothesis through the use ofhydration-rim measurements from a sample of volca-Late Ceramics in Samoa:nic-glass artifacts collected from the site.A Test Using Hydration-Rim

Measurements1the ’aoa site

In 1986, as part of an archaeological research programon the eastern end of Tutuila Island, Clark and Herdrichjeffrey t. clark, peter sheppard, and(1993) conducted an intensive survey of the ’Aoa Valleymartin jonesand surrounding ridges. ’Aoa is a small amphitheaterDepartment of Sociology and Anthropology, Northvalley with an eastern lobe that is slightly higher in ele-Dakota State University, P.O. Box 5075, Fargo, N.D.vation than most of the valley floor (fig. 2). The modern58104, U.S.A. ([email protected])/village of ’Aoa occupies most of the eastern lobe. SixDepartment of Anthropology, University ofstreams cut down the surrounding ridge slopes andAuckland, Private Bag 92019, Auckland, Newacross the valley floor. A band of calcareous sand frontsZealand ([email protected];most of the valley, and inland of that are silty clays [email protected]). 22 iv 97stoney silty clays that are slightly to mildly acidic (pH5.6–6.0). There is also a small estuary and area of man-Archaeological research on Tutuila Island, in the cen-grove where three of the valley streams converge. Thetral Pacific, in the late 1980s and early 1990s producedwater table is normally 1.0–1.5 m below the surface.some quite startling results. Tutuila is one of nine ma-

Survey of the valley revealed archaeological materialsjor islands in the Samoan Archipelago of western Poly-(artifacts, shell, and traces of house floors) over most ofnesia (fig. 1). In the ’Aoa Valley, on the north coast ofthe lower (coastal) half of the valley. Because these ma-eastern Tutuila, Clark and others discovered and par-terials occurred in a continuous pattern, one site num-tially excavated what was at the time the only ceramicber, AS-21-5, was assigned to the entire area, with 16residential site known for the island. (Numerous otherlocalities distinguished on the basis of geomorphic fea-ceramic sites had, however, been reported from ’Upolutures or density of surface artifacts. Most of the arti-and Manono Islands to the west.) The ’Aoa site pro-facts, including all of the pottery, have come from theduced an unusual abundance of volcanic glass and ba-eastern lobe.saltic rock artifacts as well as a large collection of pot-

Locality 2 is in the western part of the eastern lobe,tery sherds. A set of radiocarbon determinationsimmediately inland of the primary settlement zone ofsuggested that site occupation began perhaps as much’Aoa Village. Areas adjacent to this locality were in-as 3,000 years ago and, quite surprisingly, that potteryfilled a few decades ago to reclaim swampy ground. Atuse continued until the 15th to 17th centuries a.d. andabout the same time, Puna Stream was diverted from aperhaps later. Furthermore, none of the ceramics col-course through the village to its current channel behindlected from the site, either from excavation or from theand around the west side of the village. The streamsurface, displayed the Lapita dentate stamping thatchannel now cuts along the edge of Locality 2. The landwould be expected from a site at the early end of thatsurface gently slopes upward (3.5–4.2 m a.s.l.) to thetime range (see Green 1979). These data challenged thenorth and east until reaching the base of a steeply risingconventional culture-historical model of Samoan—andridge.

The site was first tested in 1986, when Clark and Her-1. 1997 by The Wenner-Gren Foundation for Anthropological Re- drich (1988) discovered pottery sherds in the Punasearch. All rights reserved 0011-3204/97/3805-0014$1.00. Funding Stream bed and a bank cross section that revealed a cul-for the fieldwork discussed here was provided by the National Sci-

tural deposit including at least two cultural layers andence Foundation, grant no. 9111566. The hydration study wasmade possible through the support of the University of Auckland several features. Excavations (fig. 3) in three 1-m2 unitsDepartment of Anthropology and the development of an OHD fa- (designated XU-1, -2, -3) were terminated when seepagecility under FORST grant UOA315. Additional support for the from the water table flooded the units. Two years later,fieldwork was provided by contracts and grants from the American

Clark (1989) returned briefly to the site to dig a smallSamoa Government, Office of Historic Preservation, and North Da-(1.0 by 0.5 m) unit (XU-4) into the stream bank to ob-kota State University. Specific thanks for their assistance in vari-

ous ways are due Douglas Sutton, Michael Michlovic, David tain charcoal for dating. Three charcoal samples (Beta-Herdrich, and Aaron Bergstrom. We are grateful to the many people 28210, -28211, -28212) yielded results suggesting occu-who over the years have assisted in the various aspects of the re- pation between a.d. 1450 and 1650 (calibrated) (Clarksearch on Samoan prehistory, and, as always, we extend special

1989). In 1991, excavations at Locality 2 were expandedthanks to the people of Samoa and to the matai and residents of’Aoa Village in particular. with four additional units (XU-5–8) covering 10 m2




Fig. 1. The main islands of the Samoan chain.

(Clark and Michlovic 1996) and, with the use of a pump, across the site, and not all layers were revealed in everyunit because of either unit depth or layer distribution.reaching a final depth in two units (XU-7 and -8) of over

2 m. A summary of site stratigraphy and associated artifactdistribution is presented in table 1. The A horizonsThe deeper excavations of 1991 led to the identifica-

tion of eight soil layers. The layers varied in thickness were identified in the field on the basis of color and tex-

Fig. 2. ’Aoa Bay and Valley, with the excavation area at Locality 2 on the eastern side of the valley.




Fig. 3. Site AS-21-5, Locality 2, with the locations of excavation units (XU-1–8).

ture, and those assessments were later confirmed in lab A detailed examination of the stream bank on thenorthern edge of the site revealed eight features in pro-analysis primarily by organic content. Clark and Mich-

lovic (1996) have interpreted the A horizon layers as file. An additional ten features were investigated duringexcavations. All 18 features are in layers II through Vrepresenting times of landform stability with minimal

deposition. The sediments of the other layers, however, and are ovens/fireplaces, postmolds, and boulder align-ments.were derived in large part from episodic deposition from

the mountain slopes. The basaltic rock artifacts (mostly hawaiite) are 12

table 1’Aoa Natural and Cultural Stratigraphy

%Organic Basalt Tools Volcanic

Layer Texture Color Carbon Horizon and Flakes Glass Pottery

I gravelly coarse sandy loam 5YR3/2 0.71 A,C 40 — —II loam 5YR2/2 1.04 2Ab 330 1 3

III gravelly loam 5YR3/2 0.34 2C 90 — 2IV loam 7.5YR3/2 0.32 2C — — —V clay loam 5YR3/2 0.84 3Ab/3B 2,431 39 126

VI silt loam 5YR2/2 0.98 4Ab 214 26 46VII clay loam 5YR3/2 0.43 4B 465 183 568

VIII loam 5YT3/3.5 0.15 4C 34 27 130




table 2Radiocarbon Age Determinations from ’Aoa, Site AS-21-5

Calibrated Calibrated RangeSample Depth 14C 13C Range b.p. at 2 b.c.–a.d. at 2No. Unit Layer (cm) Context Years Correction Sigma Sigma

B-48910 SB II — 540 6 70 510 6 70 653–476 a.d. 1297–1474B-28210 4 II 113–123 Fea. 1 330 6 40 492–293 a.d. 1443–1657B-48048 5 V 84–94 520 6 60 470 6 60 643–476 a.d. 1307–1474B-48047 SB V 140–156 430 6 80 400 6 80 553–299 a.d. 1397–1651B-28211 4 V 150–169 350 6 50 507–293 a.d. 1443–1657B-28212 4 V 160–165 Fea. 9 170 6 40 293–0 a.d. 1657–1955B-48911 8 VII 128–148 2,450 6 110 2,460 6 110 2,764–2,195 b.c. 814–245B-48049 7 VII 170 2,940 6 140 2,890 6 140 3,455–2,759 b.c. 1505–809

note: Calibrations are based on Stuiver and Reimer (1993).

adzes and adze fragments, 12 preforms, 15 flake tools Clark and Michlovic (1996; Clark 1993) conclude thatthe radiocarbon data support the interpretation of two(mostly scrapers/graters), 3,559 waste flakes, 1 grinding

stone, and 1 possible hammerstone. Volcanic glass (geo- broad components at the site. The lower component isrepresented by layers VII and VIII and is 2,000–3,000chemically not true obsidian) was represented by 276

items. Of these, 152 pieces are chips and flakes, 98 are years old. The upper component consists of layers II–Vand dates to sometime between 550 and 300 b.p. (cali-cores or core fragments, and 24 are flakes that display,

microscopically or macroscopically, edge damage sug- brated). The lack of datable material from layer VI, to-gether with the absence of features and the significantlygestive of tool use. While this tool count is low, almost

certainly many more pieces were utilized but evidence lower artifact content relative to layer V, created someuncertainty as to the actual relationship of layer VI toof wear does not show up under a light microscope (see

Fredericksen and Sewell 1991). Technological and geo- the upper component. As a paleosol, it seemed laterthan and distinct from the lower component, but thechemical characteristics of the volcanic-glass assem-

blage are discussed elsewhere (Clark and Wright 1995). question remained whether it was effectively part ofthe upper component or separated in time from both theThe ceramic sample from site excavations consists of

878 sherds, many of which are small (90% at , 2.5 cm lower and upper components. In any case, the dates forthe upper component are astonishingly recent for a ce-in diameter) and eroded and have crumbly surfaces. The

occurrence of what appears to be goethite (FeO[OH]), ramic site.which forms as a weathering product, together with thegeneral condition of the sherds led Clark and Michlovic implications(1996) to suggest postdepositional alterations in the pot-tery. Hundreds more sherds have been collected from Clark and Michlovic (1996) have considered alternative

explanations for the late-dated ceramics at ’Aoa. Down-the surface of the valley east of Locality 2.Data on the eight dated charcoal samples from the slope superpositioning, either of the entire upper com-

ponent or of just the artifacts, seems to them unlikelysite are presented in table 2 (Clark 1993). These datesand the site stratigraphy suggest two cultural compo- because of the condition of the total artifact assemblage

(i.e., little evidence from lithic edges of transport), thenents widely separated in time. The upper componentproduced six dates, four of which (one from layer II and presence of numerous features in the upper component,

and the stratigraphic differences in proportions of arti-three from layer V) cluster very tightly in the range of550 to 300 b.p. (calibrated, two sigma). One of the layer fact types between the components. That lower-compo-

nent artifacts have been brought to a higher level byV samples (B-28212) appears too recent and one layer IIsample (B-48048) may be too old. In short, layers II and later disturbances (gardening or pits) appears unlikely

because of the intact paleosol of layer VI and the depthV are radiometrically indistinguishable, though almostcertainly there is an actual time difference, as is sug- of the lower component. Consequently, they conclude

that the best explanation for the late ceramics is the ex-gested by the slightly different artifact contents—layerII has notably fewer sherds and volcanic-glass artifacts. istence at ’Aoa of a late ceramic-bearing component.

The suggested presence of a ceramic industry dated toAlso, in the eastern portion of the site—but only in thatarea—layers II and V are separated by layers III and IV, the 15th–17th centuries stands in marked contrast to

the conventional model for Samoan culture history aswhich represent short-term alluvial depositionalevents. Two additional dates came from different levels it has been articulated by Green (1974a, b) and adopted

by others (e.g., Bellwood 1978, Davidson 1979). In thatwithin layer VII and indicate considerably older depos-its: 2,764–2,195 b.p. and 3,455–2,759 b.p. (calibrated, model, based on sites from ’Upolu Island, Green put the

abandonment of pottery in Samoa between the 3d andtwo sigma).




7th centuries a.d., probably by the 5th century (Green1974b:248–53). Volcanic glass co-occurs with pottery,and it, too, fell out of use at about the same time. How-ever, Clark (1994, 1996) has pointed out that a reevalua-tion of previously reported sites in Samoa (Green andDavidson 1969, 1974; Jennings and Holmer 1980; Jen-nings et al. 1976) suggests that some of the other siteswhere late ceramics were previously attributed to dis-turbances may in fact represent legitimate cases of latepottery use. That view, along with the proposed latecomponent at ’Aoa, is certainly contentious and in needof further scrutiny.

With such scrutiny in mind, Sheppard proposed thata set of volcanic-glass artifacts from the two compo-nents at ’Aoa be subjected to hydration-rim, or rind,measurements at the recently established facility at theUniversity of Auckland. Such measurements would, heargued, provide a test of an alternative hypothesis tothat of Clark and Michlovic—that the sherds and vol-canic glass are in a late-dated context because of dis-turbance. Since there is no chronometric calibrationfor hydration measurements of Samoan glass, the testwould be based on relative thicknesses of the hydrationrims. The disturbance hypothesis would be supported ifthere were no appreciable difference in the rim mea-surements on the upper- and lower-component speci-mens. Such evidence would also refute the hypothesisof a late ceramic component at ’Aoa.

methods and results

A group of specimens was selected by Clark and sent intwo batches to Sheppard and Jones. Of these, 14 werefound suitable for analysis. The artifacts were selected Fig. 4. Typical volcanic-glass hydration bandto provide thorough coverage of the cultural deposit. observed on measured samples (artifact #53, width ofVolcanic glass was not recovered from layers I, III, and field 43 microns, flake interior on the right).IV, and pieces large enough for analysis were very lim-ited in layers II and VIII. The samples were sent to Shep-pard and Jones without provenience data so as not to of a uniform distribution. Thus, there is no overlap be-

tween the two groups, and they can be interpreted asinfluence their analysis.The volcanic-glass hydration-rim measurements (fig. the result of two distinct events. Table 3 displays the

provenience information for each specimen along with4) were made using the video imaging technique de-scribed by Ambrose (1993). This approach allows more the measurement. The data are arranged by measure-

ment and by stratigraphic depth. Because of the varia-accurate rim measurements than possible with stan-dard optical microscopy. Rim thickness measurements tion in layer distribution and thickness at the site, the

depth measurements are meaningful only within units.were taken from only one face; the ventral face wasselected where identifiable, and otherwise the least The different thickness groupings will have arisen as

the result of three factors: glass chemistry, ground tem-weathered face was chosen. This approach ensured thatthe section cut would be made normal to the face asso- perature, and time. To make the assumption that the

different groupings are the result of activities separatedciated with the cultural context. This avoided any pos-sible error due to rim ‘‘inflation’’ introduced by nonnor- in time, we must argue that glass chemistry and tem-

perature are not factors that systematically vary amongmal cuts or reading of a rim not associated with thearchaeological context. The measurements were made the samples.

We can be confident that there is no systematic varia-using pixel intensity histograms integrated over a pass25 pixels wide. Theoretically, this gives a reading accu- tion in glass chemistry between the different layers. A

recent geochemical study by Wright has shown thatrate to 60.075 µm.As illustrated in table 3, samples fall into two groups. volcanic-glass artifacts from ’Aoa that cover the range

of macroscopic variability are quite similar in elemen-Group 1 samples range from 0.6 to 0.75 6 0.075 µm,while group 2 samples range from 0.9 to 0.975 6 0.075 tal composition (Clark and Wright 1995). ‘‘In fact, the

variation between flakes is approximately the same asµm. The error ranges should be interpreted as the limits




table 3Hydration-Rim Measurements from Samples Arranged byStratigraphy and Depth

Sample Unit No. DepthNo. Measurement (XU-) Layer (cm)

16 0.6 2 II 60–7056 0.67 7 VI 95–10538 0.675 5 V 84–94

111 0.675 7 VII 124–134161 0.75 8 V 46–56166 0.75 8 V 54–6447 0.75 7 V 70–7553 0.75 7 VI 85–95

214 0.9 8 VII 128–138137 0.9 7 VII 144–154209 0.975 8 VII 128–138233 0.975 8 VII 138–148252 0.975 8 VII 148–158267 0.925 8 VIII 158–168

the variation within individual flakes, suggesting that discussionall flakes found at ’Aoa came from the same, slightly

Several observations can be made on the basis of theseheterogeneous, glass source,’’ and that source is almostdata. First and foremost, the disturbance hypothesis iscertainly on Tutuila (Clark and Wright 1995:253). Werefuted. The measurements fall into two groups thatassume, therefore, that there is little variability in hy-correspond to the two proposed components. Just howdration rate due to differences in material compositionmuch time actually separates the two components can-and source.not be ascertained via the rim measurements. If the rateSimilarly, there will be no significant difference be-of hydration is, as expected, nonlinear, the age differ-tween the samples in terms of soil temperature. Whileence will be greater than suggested by the numbersit is not possible to give an accurate estimate of the ab-taken at face value. Moreover, if the ages of the twosolute temperature history of the artifacts, we can con-components are not as widely separated in time as indi-fidently assume that there is little relative differencecated by the radiocarbon determinations, it is mostamong those temperature histories. A recent study di-likely that the age of the lower component, which isrected at establishing the factors that govern variationless well dated, is off, possibly because of an old-woodin archaeological soil temperatures (Jones, Sheppard,effect.and Sutton n.d.) indicates that only depth could gener-

The upper component holds together on a ratherate systematic temperature variations among the sam-broad level. The single measurement from layer II sug-ples in this study. As the bulk of the samples comesgests that this layer may indeed sort out in time fromfrom two contiguous excavation units, the surface en-layer V, though this is far from conclusive. The sug-ergy budgets for the samples will be uniform and thegested time difference is supported by notable differ-only variation will be due to depth of burial. Any differ-ences in artifact content between layers II and V. It isence would tend to produce larger rims in the shallowerconceivable, also, that after the depositional events ofupper component, as this will experience a higher aver-layers III and IV, ceramics and volcanic glass were noage temperature regime than the lower componentlonger used at ’Aoa and the very few examples of these(Jones, Sheppard, and Sutton n.d.). The only sensibleartifacts in layer II are indeed the result of disturbances.temperature variation scenario in this case is that theSample VG-III is clearly aberrant and represents a pieceupper component will have experienced a higher aver-that is out of context, because of in-site disturbance orage storage temperature than the lower componentfield contamination. The position of layer VI is clarified(though as both components are below the first damp-by these results; it is, as suspected, part of the uppering moment of soils the differences will be negligible).component.As this would have the effect of producing larger (mar-

ginally) rims in the upper component, we can reject thepossibility that the groupings observed are the result of conclusiondifferences in the storage temperature of the artifacts.

Since there is no systematic variation among the This study illustrates the potential value of hydration-rim measurements for relative age assessments. Thesamples in terms of chemistry or storage temperature,

we can safely conclude that the different groupings rep- test reported here provides support for the hypothesisthat two cultural components existed at ’Aoa. The hy-resent events separated in time.




h. l. smith. 1976. Excavations on ’Upolu, Western Samoa.dration-rim data do not provide a calendrical date forPacific Anthropological Records 25.those components and therefore do not directly confirm

jones, m., p. sheppard, and d. sutton. n.d. Soil temper-the hypothesis of late ceramics. However, by rejecting ature and obsidian hydration dating: A clarification of the vari-the hypothesis that the artifacts of the upper compo- ables affecting accuracy. Journal of Archaeological Science. In

press.nent reached that stratigraphic position through super-stuiver, m., and p. j. reimer. 1993. Extended 14C data-positioning or disturbance, we are left with the conclu-

base and revised CALIB Radiocarbon Calibration Program. Ra-sion that the association of artifacts and radiocarbon diocarbon 35:215–30.dates in the upper component is valid, and that associa-tion indicates that pottery was used at ’Aoa until the15th–17th centuries.

Stressed to the Max?References Cited Physiological Perturbation inambrose, w. r. 1993. ‘‘Obsidian hydration dating,’’ in Archae- the Krapina Neandertals1

ometry: Current Australasian research. Edited by B. L. Frank-hauser and J. R. Bird, pp. 79–84. Research School of PacificStudies, The Australian National University, Occasional Pa-pers in Prehistory 22. dale l. hutchinson, clark spencer larsen,

bellwood, p. s. 1979. Man’s conquest of the Pacific: The pre- and inui choihistory of Southeast Asia and Oceania. New York: Oxford Uni-

Department of Anthropology, East Carolinaversity Press.University, Greenville, N.C. 27858 (Hutchinson)/clark, j.t. 1989. The Eastern Tutuila Archaeological Project:

1988 final report. MS, Historic Preservation Office, American Research Laboratories of Anthropology andSamoa Government, Pago Pago. Department of Anthropology, University of North

———. 1993. Radiocarbon dates from American Samoa. Radiocar- Carolina, Chapel Hill, N.C. 27599-3120 (Larsen)/bon 35:323–30.Department of Anthropology, California Polytechnic———. 1994. The Samoan ceramic sequence reconsidered. Paper

presented at a symposium at the 15th Congress of the Indo- Institute, San Luis Obispo, Calif. (Choi), U.S.A.Pacific Prehistory Association, Chiang Mai, Thailand, January 3 iii 975–12.

———. 1996. ‘‘Samoan prehistory in review,’’ in Oceanic cultureReconstructions of Neandertal living conditions gener-history: Essays in honour of Roger Green. Edited by J. David-ally emphasize their difficulty. According to Jelinekson, G. Irwin, F. Leach, A. Pawley, and D. Brown. New Zea-

land Journal of Archaeology Special Publication. (1994:77), for example,clark, j. t., and d. j. herdrich. 1993. Prehistoric settle-

One thing that the biological evidence indicates isment system in eastern Tutuila, American Samoa. Journal ofthe Polynesian Society 102:147–85. that life was not easy for the Neandertals. They

clark, j. t., and m. g. michlovic. 1996. An early settle- show a higher incidence of physical trauma thanment in the Polynesian homeland: Excavations at ’Aoa Valley, later people, and an extreme level of what appearsTutuila Island, American Samoa. Journal of Field Archaeology

to be periodic nutritional stress. Even the most aged23:151–67.clark, j. t., and e. wright. 1995. Volcanic glass in individuals appear not to have survived much be-

Samoa: A technological and geochemical study. Journal of the yond 40 years, and the great majority seem to havePolynesian Society 104:230–66. died at least a decade earlier. So, the successful adap-davidson, j. m. 1979. ‘‘Samoa and Tonga,’’ in The prehistory

tation of the Neandertals appears to have included aof Polynesia. Edited by J. D. Jennings, pp. 82–109. Cambridge:shorter and more difficult life than that of later Up-Harvard University Press.

fredericksen, c. f. k., and b. sewell. 1991. The reliabil- per Paleolithic hunters, but not so short and diffi-ity of flaked tool function studies in New Zealand archaeology. cult as to prevent them from surviving as a biologi-Archaeology in Oceania 26:123–26. cal population.green, r. c. 1974a. ‘‘Excavations of the prehistoric occupa-tions of SU-Sa-3,’’ in Archaeology in Western Samoa, vol. 2. Among the many Hobbesian depictions of NeandertalsEdited by R. C. Green and J. M. Davidson, pp. 108–54. Auck-

have been the protein-starved, cannibalistic hominid,land Institute and Museum Bulletin 7.———. 1974b. ‘‘A review of portable artifacts from Western

Samoa,’’ in Archaeology in Western Samoa, vol. 2. Edited byR. C. Green and J. M. Davidson, pp. 245–75. Auckland Insti- 1. 1997 by The Wenner-Gren Foundation for Anthropological Re-

search. All rights reserved 0011-3204/97/3805-0013$1.00. Wetute and Museum Bulletin 7.———. 1979. ‘‘Lapita,’’ in The prehistory of Polynesia. Edited by thank the Louis B. Leakey Foundation, Northern Illinois Univer-

sity, Purdue University, and the Midwest Universities ConsortiumJ. D. Jennings, pp. 27–60. Cambridge: Harvard UniversityPress. for International Activities for providing the financial support that

made possible the collection of data and production of casts of thegreen, r. c., and j. m. davidson. Editors. 1969. Archae-ology in Western Samoa. Vol. 2. Auckland Institute and Mu- Krapina teeth in Croatia. Jakov Radovcic and the Croatian Natural

History Museum provided facilities for study and access to theseum Bulletin 7.———. 1974. Archaeology in Western Samoa. Vol. 2. Auckland Krapina collections. Laboratory space and equipment for analysis

were provided by the Departments of Anthropology at East Caro-Institute and Museum Bulletin 7.j ennings, j. d., and r. n. holmer. Editors. 1980. Archae- lina University and the University of Illinois and the Department

of Sociology and Anthropology at Purdue University. We thankological excavations in Western Samoa. Pacific Anthropologi-cal Records 32. Lorraine V. Aragon, Simon Hillson, Marsha Dean Ogilvie, Erik

Trinkaus, and Linda Wolfe for useful comments on the manuscript.j ennings, j. d., r. n. holmer, j. c. janetski, and




the warm, caring hominid in the face of overwhelming Broadly speaking, both hypoplasias and hypocalcifi-cations reflect stress episodes occurring during thedifficulty, and the crippled, degenerative hominid (Con-

stable 1973). The general impoverishment of Neander- growth and development of tooth crowns in childhood(Goodman and Rose 1990, Yaeger 1990). Hypoplasiastals is often tied to the glacial climate that dominated

Europe during the major portion of their existence in are serial pits or linear grooves resulting from amelo-blast death and the failure to form enamel matrix. Hy-the later Pleistocene (Poirier 1993). The glacial land-

scape of Europe was continuous, however, in neither pocalcifications are opacities or bands of discolorationthat contrast with normal enamel. Hypocalcificationgeographic nor temporal distributions, and therefore the

environmental stressors that purportedly formed the results from a disruption of the mineralization of theenamel matrix, although recent research by Sucklingbasis of Neandertal living conditions were not uniform

in space or time. The Hobbesian portrayal of Neandertal (1986) suggests that it can be associated with enamelmatrix formation. Hypoplasias are often associatedlife is also related to the ongoing debate over modern

human origins; it is easier to accept the demise of Euro- with hypocalcifications, suggesting metabolic disrup-tion of both the matrix secretion and maturation (Hill-pean late-Pleistocene populations on the assumption

that they were struggling unsuccessfully to survive. son 1996). In their study of the Krapina series, Molnarand Molnar (1985) and Ogilvie, Curran, and TrinkausAmong the clues that can provide an indication of the

success or failure of Neandertals to adapt to their envi- (1989) included isolated pits as a type of enamel de-fect. Little attention has been devoted to isolated pitsronmental challenges are skeletal and dental indicators

of physiological stress. In some Neandertal samples, although Goodman and coworkers (1992) suggest thatthey result from chronic physiological disruption. Forsuch as Krapina or Shanidar, trauma is prevalent (Berger

and Trinkaus 1995, Trinkaus 1983) and degenerative the purposes of this analysis, the category ‘‘surface de-fects’’ includes pits, pit patches, linear pits, and linearjoint disease has been observed (Trinkaus 1983, 1985b).

Yet, there is little skeletal evidence for infectious dis- hypoplasias.Larsen conducted a visual inspection of 260 teethease (Ogilvie, Curran, and Trinkaus 1989). Floral and

faunal remains could provide additional information from Krapina at the Croatian Natural History Museumin Zagreb in 1990, recording hypoplasias, hypocalcifi-about nutrition, but preservation, taphonomic change,

and the sharing of niche space with other predators and cations, defect pits, and depressions (table 1). Negativecast impressions were made of all teeth and later castcarnivores has made interpretations of Neandertal di-

etary preferences difficult (Stiner 1994). in the United States. Data were collected from both theoriginal specimens and the positive cast impressions,In this paper we report on levels of physiological

stress in the Neandertals from the Krapina rock shelter but our discussion here reports on observations of theoriginal specimens only. No estimates for the chrono-in Croatia, the largest sample of Neandertals recovered

from a single site (Radovcic 1985, Radovcic et al. 1988). logical age at which growth disruptions occurred weremade because of uncertainty regarding rates of Nean-Excavated under the direction of Gorjanovic-Kram-

berger between 1899 and 1905, the rock shelter at Kra- dertal growth and development and their potential dif-ferences from recent modern humans (e.g., Dean,pina yielded the first large sample of archaic human re-

mains and provided the first test of fluorine dating (Wol- Stringer, and Bromage 1986, Stringer, Dean, and Martin1990; but see Smith 1991), although recent analysis ofpoff 1979). Although they are highly fragmentary, the

abundance of the Krapina remains makes them enor- Krapina perikymata counts by Mann, Monge, andLampl (1991) indicates that the time of formation maymously valuable for drawing inferences about the

health, lifestyle, and behavioral ecology of these extinct be essentially modern.For purposes of analysis, teeth were grouped ac-hominids.

Given the predominance of dental remains at Krapina cording to the Krapina Dental Person (KDP) designa-tions utilizing the Krapina catalog (Radovcic et al.and other fossil hominid sites, enamel defects resulting

from metabolic disturbance are heavily utilized as indi- 1988). Each individual (KDP) consists of associated setsof teeth grouped together on the basis of matching inter-cators of health by various researchers (Bermudez de

Castro and Perez 1995, Molnar and Molnar 1985, Ogil- proximal facets or unusual features (Wolpoff 1979). Allteeth were examined, but quantifying defects on teethvie, Curran, and Trinkaus 1989, Robinson 1956, Skin-

ner 1996, White 1978). Ogilvie, Curran, and Trinkaus from both sides of the dentition would duplicate dataregarding a systemic growth disruption. Consequently,(1989) and Molnar and Molnar (1985) found that most

individuals from Krapina exhibited enamel defects and only teeth from the left side of dental individuals wereincluded in our tabulations; if available, teeth from thesuggested that the Krapina people were under severe

stress. The findings we report here elaborate on these right side were substituted in the case of missing teethfrom the same dentition. The final study sample sizeimportant previous studies by incorporating our study

of hypoplasias and adding data on another type of consisted of 176 teeth representing 35 individuals. Inaddition, 28 unassociated teeth were examined, bring-enamel defect, hypocalcification. We also compare the

frequency of enamel defects in the Krapina population ing the total sample to 204 teeth. The sample distribu-tion of teeth is presented in table 1, and a summary ofwith frequencies from several Holocene populations to

provide additional context for assessing stress in this tooth types is presented in table 2.The percentage of dental individuals affected by met-group of Neandertals.




table 1Distribution of Hypoplasias and Hypocalcifications by KrapinaDental Individual

Tooth N NKDPa Numberb Tooth Typec Calcd Hype

1 18 mandibular lt i1 — —1 13 mandibular lt i2 — —1 24 maxillary lt c — —1 17 maxillary lt i1 — —1 15 maxillary lt i2 — —1 XA.3 MAXILLARY LT M1 — —1 XA.1 maxillary lt m1 — —1 XA.2 maxillary lt m2 — —1 116 MAXILLARY LT PM3 — 11 117 MAXILLARY LT PM4 — —1 98 MAXILLARY RT M2 — —2 191 MAXILLARY LT C — 12 XB.7 maxillary lt c — —2 194 MAXILLARY LT I1 — 12 XB.5 maxillary lt i1 — —2 XB.2 maxillary lt m1 — —2 192 MAXILLARY LT M2 — 12 XB.3 maxillary lt m2 — —2 196 MAXILLARY RT I2 1 —2 XB.1 maxillary rt i2 — —2 XB.4 MAXILLARY RT M1 — —2 110 MAXILLARY RT PM3 2 —3 119 MANDIBULAR LT C 1 43 107 MANDIBULAR LT M2 — 13 102 MAXILLARY LT C 1 33 XC.2 MAXILLARY LT M1 — —3 XC.3 MAXILLARY LT M2 — —3 XC.1 maxillary lt m2 — —3 131 MAXILLARY RT I2 1 13 53 MAXILLARY RT PM3 1 —4 ND.2 MANDIBULAR LT C 1 —4 ND.1 MANDIBULAR LT I2 1 —4 ND.5 MANDIBULAR LT M1 — —4 ND.6 MANDIBULAR LT M2 — —4 ND.3 MANDIBULAR LT PM3 — 24 ND.4 MANDIBULAR LT PM4 1 —4 155 MAXILLARY LT I1 1 —4 160 MAXILLARY LT I2 — 14 XD.3 MAXILLARY LT M1 — —4 XD.4 MAXILLARY LT M2 — —4 XD.1 MAXILLARY LT PM3 — —4 XD.2 MAXILLARY LT PM4 1 —4 141 MAXILLARY RT C — 24 163 MAXILLARY RT M3 — —5 XE.5 MAXILLARY LT C — 15 XE.3 MAXILLARY LT I1 — 15 XE.4 MAXILLARY LT I2 — 15 XE.6 MAXILLARY LT PM3 — 15 XE.7 MAXILLARY LT PM4 — 16 NH.3 MANDIBULAR LT C — —6 NH.1 MANDIBULAR LT I1 — —6 NH.2 MANDIBULAR LT I2 — —6 NH.5 MANDIBULAR LT M1 — —6 NH.6 MANDIBULAR LT M2 — —6 NH.7 MANDIBULAR LT M3 — —6 NH.3.1 MANDIBULAR LT PM3 — —6 NH.4 MANDIBULAR LT PM4 1 —6 142 MAXILLARY LT C — 16 133 MAXILLARY LT I1 — —6 125 MAXILLARY LT I2 — 16 164 MAXILLARY LT M1 — —6 173 MAXILLARY LT M3 — 16 45 MAXILLARY LT PM3 — 1




table 1(Continued)


6 49 MAXILLARY LT PM4 — —7 NA.1 mandibular lt c — —7 NA.2 mandibular lt m1 — —8 120 MANDIBULAR LT C — —8 NB.1 mandibular lt c — —8 NB.5 MANDIBULAR LT M1 — —8 NB.3 MANDIBULAR LT PM3 — —8 NB.4 MANDIBULAR LT PM4 — —8 NB.2 MANDIBULAR RT I2 — —8 103 MAXILLARY LT C 1 19 73 MANDIBULAR LT I1 — —9 NC.1 MANDIBULAR LT I2 — —9 67 mandibular lt m1 — —9 66 mandibular lt m2 — —9 111 MANDIBULAR LT PM3 — 19 NC.3 MANDIBULAR RT M1 — —9 NC.4 MANDIBULAR RT M2 — —10 NE.4 MANDIBULAR LT C 1 310 NE.2 MANDIBULAR LT I1 — 110 NE.3 MANDIBULAR LT I2 — 110 NE.7 MANDIBULAR LT M1 — —10 NE.8 MANDIBULAR LT M2 — —10 106 MANDIBULAR LT M3 — —10 NE.5 MANDIBULAR LT PM3 2 —10 NE.6 MANDIBULAR LT PM4 — —11 NF.1 MANDIBULAR LT PM4 — 111 34 MANDIBULAR RT PM3 — 112 NG.1 MANDIBULAR RT I1 — —12 NG.2 MANDIBULAR RT M2 — —12 NG.3 MANDIBULAR RT M3 — —13 NJ.3 MANDIBULAR LT C 1 —13 NJ.1 MANDIBULAR LT I1 — —13 NJ.2 MANDIBULAR LT I2 — —13 NJ.5 MANDIBULAR LT M1 — —13 NJ.6 MANDIBULAR LT M2 — —13 NJ.4 MANDIBULAR LT PM4 — —13 NJ.7 MANDIBULAR RT M3 — —14 4 MANDIBULAR LT M3 — 115 RA1 MANDIBULAR LT M3 — —16 RA4 MANDIBULAR LT M3 — 117 147 MAXILLARY LT C — 217 158 MAXILLARY LT I1 — 117 156 MAXILLARY LT I2 1 218 145 MANDIBULAR RT C 1 —18 84 MANDIBULAR RT M1 — —18 2 MANDIBULAR RT M2 2 —18 28 MANDIBULAR RT PM3 1 —18 31 MANDIBULAR RT PM4 3 —18 36 MAXILLARY RT C 2 118 129 MAXILLARY RT I1 1 118 127 MAXILLARY RT I2 1 —18 38 MAXILLARY RT PM3 1 —18 47 MAXILLARY RT PM4 — 119 139 MAXILLARY LT C 1 219 54 MAXILLARY LT PM3 — 119 57 MAXILLARY RT I2 — 219 169 MAXILLARY RT M2 — 119 170 MAXILLARY RT M3 — 119 46 MAXILLARY RT PM4 1 —20 82 MANDIBULAR LT M1 — —20 3 MANDIBULAR LT M2 — —20 5 MANDIBULAR LT M3 — —20 32 MANDIBULAR LT PM4 1 —21 93 MAXILLARY LT I1 1 1




table 1(Continued)


21 186 maxillary lt m2 — —21 181 maxillary rt m1 — 122 16 mandibular rt i2 — —22 68 mandibular rt m2 — —22 94 MAXILLARY LT I1 1 —22 100 MAXILLARY LT M1 — —22 183 maxillary lt m1 — 222 185 maxillary lt m2 — —22 12 maxillary rt i2 — —23 138 MANDIBULAR LT C 2 —23 72 MANDIBULAR LT I1 — —23 69 MANDIBULAR LT I2 — 223 85 MANDIBULAR LT M3 — 123 27 MANDIBULAR LT PM3 — —23 79 MANDIBULAR RT M1 1 —23 1 MANDIBULAR RT M2 — 123 30 MANDIBULAR RT PM4 — 223 136 MAXILLARY LT M1 — —23 135 MAXILLARY LT M2 1 —23 52 MAXILLARY LT PM3 — 123 41 MAXILLARY LT PM4 1 —24 176 MAXILLARY LT M2 — 124 179 MAXILLARY LT M3 — 125 172 MAXILLARY RT M2 — —25 162 MAXILLARY RT M3 — 126 59 MANDIBULAR RT C — 126 60 MANDIBULAR RT I2 — —27 81 MANDIBULAR LT M1 — —27 6 MANDIBULAR LT M2 2 —27 64 mandibular lt m2 — —27 33 MANDIBULAR LT PM3 2 —27 50 MANDIBULAR LT PM4 — 228 63 mandibular lt m2 — —29 123 MAXILLARY LT I1 — 229 122 MAXILLARY LT I2 2 —30 76 MAXILLARY RT C — 130 128 MAXILLARY RT I2 1 131 121 MANDIBULAR LT C — 131 90 MANDIBULAR LT I2 — 131 114 MANDIBULAR LT PM3 — 131 113 MANDIBULAR LT PM4 — 132 166 MAXILLARY RT M2 — 232 58 MAXILLARY RT M3 — —33 177 MAXILLARY RT M2 — —33 178 MAXILLARY RT M3 — —34 29 MANDIBULAR RT PM3 — 234 35 MANDIBULAR RT PM4 — 135 126 MAXILLARY RT I1 — —35 130 MAXILLARY RT I2 — —

86 MANDIBULAR LT M2 — 165 mandibular lt m2 — —

9 MANDIBULAR LT M3 — —118 MANDIBULAR LT PM4 — —

75 MANDIBULAR RT C 1 180 MANDIBULAR RT M1 — —

105 MANDIBULAR RT M1 — —104 MANDIBULAR RT M2 — —

8 MANDIBULAR RT M3 — 126 MANDIBULAR RT PM4 1 —37 MAXILLARY LT C — 1

144 MAXILLARY LT C 4 —146 MAXILLARY LT C 3 —101 MAXILLARY LT M1 — —

108.2 MAXILLARY LT M1 — —




table 1(Continued)


97 MAXILLARY LT M3 — —180 MAXILLARY LT M3 — 1

43 MAXILLARY LT PM3 — —40 MAXILLARY LT PM4 1 —

132 MAXILLARY RT I1 2 —148 MAXILLARY RT I2 2 —

109.2 MAXILLARY RT M1 — —99 MAXILLARY RT M3 — 1

109 MAXILLARY RT M3 — —39 MAXILLARY RT PM3 1 151 MAXILLARY RT PM3 — 142 MAXILLARY RT PM4 — 244 MAXILLARY RT PM4 1 —

aKrapina Dental Person.bTooth number or jaw of numbered dental remains (X 5 maxilla; N 5 mandible; e.g., XA.1 5tooth 1 of Maxilla A).cLT 5 left; RT 5 right; I 5 incisor; C 5 canine; PM 5 premolar; M 5 molar. Upper-case lettersindicate permanent teeth, lower-case letters indicate deciduous teeth.dNumber of hypocalcifications.eNumber of hypoplasias.

abolic disruption was determined on the basis of at least A comparison of the mean frequencies of the twotypes of enamel defects reveals that increased frequen-one dental defect on at least one tooth. Most of the 35

individuals (N 5 30; 86%) display at least one enamel cies of hypoplasias are associated with increased fre-quencies of hypocalcifications (fig. 2). The mandibulardefect on one or more teeth (table 1). Of the 30 affected

individuals, 7% (N 5 2) have only hypocalcification, and maxillary canines and the mandibular second molarhave the greatest mean frequency of defects per tooth.43% (N 5 13) have only surface defects (hypoplasias and

pits), and 50% (N 5 15) have a combination of surface A higher frequency of defects in anterior teeth has beenobserved by us in the study of modern (Holocene) hu-defects and hypocalcification. The prevalence of surface

defects and hypocalcification is shown in figure 1 as a mans (Hutchinson and Larsen 1988, 1990; Larsen andHutchinson 1992). The unusually high frequency ofpercentage of teeth affected for each tooth type. In gen-

eral, more maxillary than mandibular teeth are affected; enamel defects in posterior teeth, especially the thirdmolar, appears to be unique to Neandertals (see alsothe tooth type most consistently displaying surface de-

fects is the permanent maxillary canine (85%). Ogilvie, Curran, and Trinkaus 1989).The strong association between hypocalcification and

surface defects suggests that the two types of defects totable 2some degree represent parallel metabolic stress experi-Summary of Krapina Dental Elements Examined byences. In fact, the addition of hypocalcification to theTooth Typedata base resulted in few changes to the research re-sults, at least with regard to percentage of individuals

Tooth Type Maxillary Mandibular affected by enamel defects. The factors that determinewhether a stress event is manifested by matrix deposi-tion defects (hypoplasia) or mineralization defects (hy-Permanent

CENTRAL INCISOR 11 5 pocalcification) depend largely on the timing of stressLATERAL INCISOR 12 9 events in relation to the timing of enamel developmentCANINE 13 11 and maturation (Yaeger 1990). Although hypocalcifica-FIRST PREMOLAR 12 11

tion can be a direct result of either surface or subsurfaceSECOND PREMOLAR 10 14FIRST MOLAR 10 13 defects, it can also result from metabolic disruption fol-SECOND MOLAR 10 13 lowing matrix secretion or during matrix maturationTHIRD MOLAR 11 11 (Suga 1992).Deciduous

We further compared the pattern of surface defects oncentral incisor 2 1the anterior teeth (maxillary and mandibular incisorslateral incisor 3 2

canine 2 2 and canines) from Krapina with that found in a prehis-first molar 4 2 toric foraging population from the Stillwater Marshsecond molar 5 5 region of the American Great Basin (Hutchinson and




Fig. 1. Percentage of Krapina teeth affected by enamel hypoplasia (black bars) and enamel hypocalcification(stippled bars).

Larsen 1995). Available archaeological and biocultural ception of the maxillary canine, a striking similarity offrequency and pattern of defects (fig. 3).evidence suggests that the Great Basin and Krapina pop-

ulations share a number of behavioral attributes; both To extend the comparison further, we added two skel-etal series from the prehistoric North American Atlan-were highly mobile and likely experienced seasonally

based resource shortages. A comparison of the Krapina tic coast, an earlier preagricultural, foraging populationand a later agriculture-dependent population (fig. 4).and the Stillwater series in terms of the percentage of

teeth exhibiting one or more defects shows, with the ex- Prevalence of periosteal reactions and porotic hyper-

Fig. 2. Mean frequency of enamel hypoplasias (black bars) and enamel hypocalcifications (stippled bars) inKrapina teeth.




Fig. 3. Percentage of teeth affected at Krapina (black bars) and Stillwater Marsh (diagonal-striped bars); dataon the latter from Hutchinson and Larsen (1995).

ostosis, skeletal indicators of inflammation, and iron- fects is high, within the range of highly stressed Missis-sippian agricultural populations in North America (fig.deficiency anemia indicate that these Georgia coastal

populations were relatively stressed, especially during 5). Judging from enamel defects, most Krapina individu-als experienced one or more episodes of metabolic dis-later prehistory (Hutchinson and Larsen 1990, Larsen

1982, Larsen and Harn 1994, Larsen et al. 1992). The ruption. The number of individuals affected by surfacedefects and hypocalcifications (86%) is in agreementpercentage of teeth affected by surface defects in the

Georgia series is far greater than that for either the with the 72.2% reported by Molnar and Molnar (1985)and the 80% reported by Ogilvie, Curran, and TrinkausKrapina or the Stillwater series.

Viewed in isolation from those of other populations, (1989). Although these figures seem high, we believethat they are only a rough indicator of the magnitudethe dental defects of the Krapina people are difficult to

interpret. Ogilvie, Curran, and Trinkaus (1989) noted and pattern of metabolic disruption. When comparedwith Holocene populations for percentage of affectedthat the percentage of individuals exhibiting enamel de-

Fig. 4. Percentage of teeth affected at Krapina (black bars), Stillwater Marsh (diagonal-striped bars), and theGeorgia Coast, preagricultural (fine-stippled bars) and agricultural (coarse-stippled bars); data from Hutchinsonand Larsen (1995, 1990).




Fig. 5. Percentage of individuals affected by enamel hypoplasia from foraging and horticultural populationsrepresenting the Mousterian (black bars), Archaic (stippled bars), prehistoric foraging (dotted bars), LateWoodland (diagonal-striped bars), and Mississippian (white bars). Data sources: for the Perigord, Brennan(1991); for the Ohio Valley, Cassidy (1984); for the preagricultural Georgia Coast, Hutchinson and Larsen(1990); for Stillwater Marsh, Hutchinson and Larsen (1995); for Oman, Mack and Coppa (1992); for DicksonMound, Goodman, Armelagos, and Rose (1980); for Irene Mound, Hutchinson and Larsen (1990); forMoundville, Powell (1988).

teeth by tooth type, the Krapina sample does not appear people lived was environmentally unstable, involvingwarm and cold periods of relatively short duration, itto stand out. If the metabolic disruptions were sys-

temic, as would be predicted by a high-stress model, was not the extreme environment of a glacial maxi-mum. Electron spin resonance (ESR) dates (mean ofthen many of the anterior teeth should have enamel de-

fects because the enamel forms at roughly the same 130,000 6 10,000 years) obtained from tooth enameland dentin demonstrate that the entire Krapina se-time. Intertooth variability has been demonstrated in

humans (Condon and Rose 1992, Goodman and Rose quence was deposited within a short time period duringthe last interglacial (stage 5e), prior to the Wurm glacia-1990), but it would not explain the dramatic differences

between the Krapina and Georgia coastal samples. tion (Rink et al. 1995). Economically important re-sources available for human use—both plants and ani-We are in agreement with Ogilvie, Curran, and Trin-

kaus (1989) that it is unlikely that infectious disease mals—were undoubtedly greatly altered by rapidlychanging climates and oscillating landscapes. In this re-contributed appreciably to Krapina physiological stress.

Indicators of infection or related factors leading to in- gard, periodic reduction in resources may well have re-sulted in stresses (see also Rigaud and Simek 1987,fection are minimal in the Krapina skeletal remains

(Larsen, personal observation), although the fragmen- 1989) that would have profoundly affected nutritionalhealth during the late Pleistocene. Furthermore, Stinertary nature of the postcrania might prohibit a full docu-

mentation of skeletal pathology in the sample. Another (1991) suggests that at least some Neandertal popula-tions were using parts of animal carcasses with veryconsideration is that relatively thin enamel in Krapina

Neandertals may indicate stress (see Molnar et al. limited nutritional value. Some Neandertal remainsdisplay bone modification patterns indicating the possi-1993). With regard to enamel defects, however, these

stress levels are within the range of variation seen in a bility of nonritual, dietary cannibalism, perhaps madenecessary by severe food shortages (e.g., Arsuaga et al.limited sample of modern humans. Some individuals do

display unusually severe hypoplasia (for example, indi- 1993, Defleur et al. 1993, White 1988).These findings speak to a range of emerging studiesvidual 23 has multiple large, linear hypoplasias com-

bined with large pits), but from a population perspective that deal with the quality of life and adaptive ‘‘effi-ciency’’ in Neandertal populations. For example, bio-this sample is not unusual in regard to prevalence of

enamel defects. mechanical analysis of modern humans and Neander-tals reveals an unusually pronounced degree of right-Although the time period during which the Krapina




constable, george. 1973. The Neandertals. New York:side dominance in the latter (Churchill 1994; Trinkaus,Time-Life Books.Churchill, and Ruff 1994, Trinkaus, Ruff, and Churchill

dean, christopher m., chr istopher b. stringer,n.d.). In other words, Neandertals have much more ro- and timothy g. bromage. 1986. Age at death of thebust right humeri than left humeri compared with mod- Neanderthal child from Devil’s Tower, Gibraltar, and the im-

plications for studies of general growth and development inern humans. Furthermore, controlling for body size andNeanderthals. American Journal of Physical Anthropologyproportions, Neandertals have much more pronounced70:301–9.upper limb robusticity than modern humans (Churchill defleur, a., o. dutour, h. valladas, and bernard

1994). This suggests that Neandertals used their arms vandermeersch. 1993. Cannibals among the Neander-thals? Nature 362:214.more extensively and intensively than modern humans,

goodman, alan h., and jerome c. rose. 1990. Assess-especially in activities involving the right side. The rea-ment of systemic physiological perturbations from dentalsons for this difference are unclear, but Trinkaus, Ruff,enamel hypoplasias and associated histological structures. Year-

and Churchill (n.d.) suggest that Neandertals may have book of Physical Anthropology 33:59–110.been less efficient in collection and processing of foods goodman, alan h., george j. armelagos, and je -

rome rose. 1980. Enamel hypoplasias as indicators of stressthan modern humans; perhaps they simply pursuedin three prehistoric populations from Illinois. Human Biologyfoods that required more work to capture and consume.52:512–28.If so, then it seems possible that Neandertals may have goodman, alan h., debra l. martin, christopher

been more nutritionally stressed than modern humans. p. klein, m. scott peele, n jer i a. cruse, leah r.mc ewen, mmna saeed, and binta m. robinson.However, our study suggests that at least with regard to1992. ‘‘Cluster bands, Wilson bands, and pit patches: Histologi-the Krapina population nutritional stress was not espe-cal and enamel surface indicators of stress in the Black Mesacially elevated compared with that affecting severalAnasazi population,’’ in Recent contributions to the study of

other world populations. enamel development defects. Edited by Alan H. Goodman andNeandertals and contemporary populations form the Luigi L. Capasso. pp. 115–27. Journal of Paleopathology Mono-

graphic Publications 2context for understanding the origins and evolution ofhillson, s imon w. 1996. Dental anthropology. Cambridge:modern humans in the late Pleistocene. Much of that

Cambridge University Press.understanding comes from close examination of factors hutchinson, dale l., and clark spencer larsen.that potentially influenced their ability to adapt to envi- 1988. Determination of stress episode duration from linear

enamel hypoplasias: A case study from St. Catherines Island,ronmental challenges. Further comparisons with Holo-Georgia. Human Biology 60:93–110.cene populations may aid us in providing a behavioral

———. 1990. ‘‘Stress and lifeway change: The evidence fromcontext to these closely related cousins. Comparisonenamel hypoplasias,’’ in The archaeology of Mission Santa Cat-

of the enamel-defect evidence from Krapina with that alina de Guale, vol. 2, Biocultural interpretations of a popula-for Holocene populations suggests that the late Pleisto- tion in transition. Edited by Clark Spencer Larsen, pp. 50–65.

Anthropological Papers of the American Museum of Naturalcene Krapina hominids may have experienced stressesHistory 68.roughly analogous to those affecting at least some mod-

———. 1995. ‘‘Physiological stress in the prehistoric Stillwaterern foragers. Marsh: Evidence of enamel defects,’’ in Bioarchaeology of theStillwater Marsh: Prehistoric human adaptation in the West-ern Great Basin. Edited by Clark Spencer Larsen and Robert L.References Cited Kelly, pp. 81–95. Anthropological Papers of the American Mu-seum of Natural History 77.

arsuaga, juan luis, ignacio mart inez, ana jelinek, arthur j. 1994. ‘‘Hominids, energy, environment,gracia, jos e miguel carretero, and eudald and behavior in the Late Pleistocene,’’ in Origins of anatomi-carbonell. 1993. Three new human skulls from the Sima cally modern humans. Edited by Matthew H. Nitecki andde los Huesos Middle Pleistocene site in Sierra de Atapuerca, Doris V. Nitecki, pp. 67–92. New York: Plenum Press.Spain. Nature 362:534–37. larsen, clark spencer. 1982. The anthropology of St. Cath-

berger, thomas, and erik trinkaus. 1995. Patterns of erines Island. 3. Prehistoric human biological adaptation. An-trauma among the Neandertals. Journal of Archaeological Sci- thropological Papers of the American Museum of Natural His-ence 22:841–52. tory 57, pt. 3.

bermudez de castro, jos e, and pilar jul ia p erez. larsen, clark spencer, and dawn e. harn. 1994.1995. Enamel hypoplasias in the Middle Pleistocene hominids ‘‘Health in transition: Disease and nutrition in the Georgiafrom Atapuerca (Spain). American Journal of Physical Anthro- Bight,’’ in Paleonutrition: The diet and health of prehistoricpology 96:301–14. Americans. Edited by Kristin D. Sobolik, pp. 222–34. Carbon-

brennan, mary. 1991. Health and disease in the Middle and dale: Southern Illinois University Press.Upper Paleolithic of Southwestern France: A bioarchaeological larsen, clark spencer, and dale l. hutchinson.study. Ph.D. diss., New York University, New York, N.Y. 1992. ‘‘Dental evidence for physiological disruption: Biocul-

cassidy, claire m. 1984. ‘‘Skeletal evidence for prehistoric tural interpretations from the Eastern Spanish Borderlands,subsistence adaptation in the Central Ohio River Valley,’’ in U.S.A.,’’ in Recent contributions to the study of enamel devel-Paleopathology at the origins of agriculture. Edited by Mark opmental defects. Edited by Alan H. Goodman and Luigi L. Ca-Nathan Cohen and George J. Armelagos, pp. 307–45. New passo, pp. 151–69. Journal of Paleopathology Monographic Pub-York: Academic Press. lications 2.

churchill, steven e. 1994. Human upper body evolution in larsen, clark spencer, chr istopher b. ruff, mar-the Eurasian Later Pleistocene. Ph.D. diss., University of New garet j. schoeninger, and dale l. hutchinson.Mexico, Albuquerque, N.M. 1992. ‘‘Population decline and extinction in La Florida,’’ in

condon, keith, and jerome c. rose. 1992. ‘‘Intertooth Disease and demography in the Americas. Edited by John W.and intratooth variability in the occurrence of developmental Verano and Douglas H. Ubelaker, pp. 25–39. Washington,enamel defects,’’ in Recent contributions to the study of D.C.: Smithsonian Institution Press.enamel developmental defects. Edited by Alan H. Goodman mack, m. e., and a. coppa. 1992. ‘‘Frequency and chrono-and Luigi L. Capasso, pp. 61–77. Journal of Paleopathology logical distribution of enamel hypoplasias from the Ra’s al-Monographic Publications 2.




Hamra-5 (RH5) skeletal collection (Oman),’’ in Recent contri- tr inkaus, erik, steven e. churchill, and christo-pher b. ruff. 1994. Postcranial robusticity in Homo. 2. Hu-butions to the study of enamel developmental defects. Edited

by Alan H. Goodman and Luigi L. Capasso, pp.131–41. Journal meral bilateral asymmetry and bone plasticity. American Jour-nal of Physical Anthropology 93:1–34.of Paleopathology Monographic Publications 2.

mann, alan e., j. m. monge, and m. lampl. 1991. In- tr inkaus, erik, christopher b. ruff, and steven e.churchill. n.d. ‘‘Upper-limb versus lower-limb loading pat-vestigation into the relationship between perikymata counts

and crown formation times. American Journal of Physical An- terns among Near Eastern Middle Paleolithic hominids,’’ inNeandertals and modern humans in West Asia. Edited by T.thropology 86:175–88.

molnar, stephen, c. hildebolt, i. m. molnar, j. Akazawa, K. Aoki, and O. Bar-Yosef. Princeton: Princeton Uni-versity Press.radovc i c, and m. gravier. 1993. Hominid enamel

thickness. 1. The Krapina Neandertals. American Journal of white, t im. 1978. Early hominid enamel hypoplasia. Ameri-can Journal of Physical Anthropology 49:79–84.Physical Anthropology 92:131–38.

molnar, stephen, and i. m. molnar. 1985. The inci- ———. 1988. Anasazi cannibalism at Mancos Canyon. CollegiumAntropologicum 12:346.dence of enamel hypoplasia among the Krapina Neandertals.

American Anthropologist 87:536–49. wolpoff, milford h. 1979. The Krapina dental remains.American Journal of Physical Anthropology 50:67–114.ogilvie, marsha d., b. k. curran, and erik trin-

kaus. 1989. Incidence and patterning of dental enamel hypo- yaeger, j. a. 1990. ‘‘Enamel,’’ in Orban’s Oral Histology andEmbryology, 11th edition. Edited by S. N. Bhaskar, pp. 49–105.plasia among the Neandertals. American Journal of Physical

Anthropology 79:25–41. St. Louis: C. V. Mosby.poirier, frank e. 1993. 3d edition. Understanding human

evolution. Englewood Cliffs: Prentice-Hall.powell, mary lucas. 1988. Status and health in prehistory:

A case study of the Moundville chiefdom. Washington, D.C.:Smithsonian Institution Press.

radov ci c, jakov. 1985. ‘‘Neanderthals and their contempo- Did the First Hominidsraries,’’ in Ancestors: The hard evidence. Edited by Eric Del-son, pp. 310–18. New York: Alan R. Liss. Build Nests?1

radov ci c, jakov, fred h. smith, erik trinkaus,and milford h. wolpoff. 1988. The Krapina hominids:An illustrated catalog of skeletal collection. Zagreb: Mladost

jordi sabater pi , joaquim j. vea , andand Croatian Natural History Museum.r igaud, j -p., and jan f. s imek. 1987. ‘‘Arms too short to jordi serrallonga

box with God: Problems and prospects for Palaeolithic archaeol- Departament de Psiquiatria i Psicobiologia,ogy in Dordogne, Southwest France,’’ in The Pleistocene Old Universitat de Barcelona, Passeig de la Vall d’HebronWorld: Regional perspectives. Edited by Olga Soffer, pp. 47–61.

171, 08035 Barcelona (Sabater Pi and Vea)/New York: Plenum Press.Department de Prehistoria i Historia Antiga,———. 1989. Waves from the shore: Western Europe from 40,000

to 19,000 b.p. Paper presented at the annual meeting of the Universitat de Barcelona, Baldiri Reixach s/n, 08028American Anthropological Association, Washington, D.C. Barcelona (Serrallonga). Spain. 8 iv 97

rink, w. j., h. p. schwarcz, f. h. smith, and j.radovc i c. 1995. ESR ages for Krapina hominids. Nature

We believe that primatology can offer valid information378:24.robinson, j. t. 1956. The dentition of the Australo- regarding the behaviour of fossil hominids (Groves and

pithecinae. Transvaal Museum Memoir 9. Sabater Pi 1985, Wrangham 1987, Stanford and Allenskinner, mark. 1996. Developmental stress in immature hom- 1991, McGrew 1992). In our view, nesting is one fossil

inines from Late Pleistocene Eurasia: Evidence from enamel hy-hominid behaviour on which the behaviour of modernpeplasia. Journal of Archaeological Science 23:833–52.apes may shed light.smith, b. h. 1991. Dental development and the evolution of

life history in Hominidae. American Journal of Physical An- Gorillas, chimpanzees, pygmy chimpanzees, andthropology 86:157–74. orangutans build nests every night in which to sleep

stiner, m. c. 1991. The faunal remains from Grotta Guattari: and also to engage in social behaviours such as groom-A taphonomic perspective. current anthropology 32:103–17.ing, play, and sexual relations. They also build less elab-———. 1994. Honor among thieves: A zooarchaeological study

of Neanderthal ecology. Princeton: Princeton University Press. orate nests in the middle of the day to rest in during thestringer, c., m. c. dean, and r. d. martin. 1990. ‘‘A hottest hours. Gorillas usually build their nests on the

comparative study of cranial and dental development within a ground, and chimpanzees, pygmy chimpanzees, andrecent British sample and among Neandertals,’’ in Primate life

orangutans build theirs in trees. These nests—simplehistory and evolution. Edited by C. J. DeRousseau, pp. 115–52.constructions made of branches, stalks, leaves, and lia-New York: Wiley-Liss.

suckling, grace. 1986. Sheep—and research into develop- nas—have been studied in detail (Baldwin et al. 1981,mental defects of dental enamel. New Zealand Dental Journal Groves and Sabater Pi 1985, Sabater Pi and Vea 1990,82:68–71. Fruth and Hohmann 1996). All great-ape species sleepsuga, s. 1992. ‘‘Hypoplasia and hypomineralization of tooth

in the recumbent position because of muscular atonyenamel,’’ in Recent contributions to the study of enamel devel-opmental defects. Edited by Alan H. Goodman and Luigi L. Ca- due to the fact that their REM sleep is longer than thatpasso, pp. 269–92. Journal of Paleopathology Monographic Pub- of other primates and their sleeping pattern includeslications 2. stage 4 of slow-wave sleep, a feature not found in other

trinkaus, er ik. 1983. The Shanidar Neandertals. New York:nonhuman primates. According to Bert et al. (1967) thisAcademic Press.stage of NREM sleep slows down waking.———. 1985a. Cannibalism and burial at Krapina. Journal of Hu-

man Evolution 14:203–16.———. 1985b. Pathology and posture of the La Chapelle-aux-

Saints Neandertal. American Journal of Physical Anthropology1. 1997 by The Wenner-Gren Foundation for Anthropological Re-67:19–41.search. All rights reserved 011-3204/97/3805-0006$1.00.




Since the three living species of African apes and hu- tioned, are nocturnal—in open spaces may have obligedhominids to sleep in nests in forested areas.mans are derived from a common ancestor, ‘‘the proba-

bility of a given behavior occurring in the ancestor can 4. Fire is absolutely essential to ward off possible at-tacks by carnivores. Even today, African hunter-gather-be judged from its distribution in the four descendants’’

(Wrangham 1987:52). Considering this phylogenetic ers keep a fire going throughout the night, even thoughthey sleep in huts (Silberbauer 1981). The earliest ar-proximity between African apes and australopithecines

and the first forms of the genus Homo, we suppose that chaeological evidence of controlled fire in Africa is de-scribed by Clark (1969): he dates it to 180,000 b.p., atthese species shared nest behaviour. Some researchers

have suggested that the australopithecines may have Kalambo Falls (Zambia). Claims to have traced fire back1.5 million years are controversial (Bellomo 1994). Cur-slept in trees (Sept 1992a, Tuttle 1981, Susman, Stern,

and Jungers 1984), and Desmond Clark (personal com- rently accepted archaeological data associate controlledfire with final Homo erectus and archaic H. sapiensmunication) believes that even Homo habilis did so.

However, none of these specialists mention the possi- (James 1996). Australopithecines, H. habilis, and Par-anthropus had no knowledge of fire, so as they movedbility that these primates built nests in trees.

We hold that the following factors justify our hypoth- through a mosaic of biotopes with an abundance offorest areas and a high presence of predators, it is highlyesis that the first hominids built and used nests in trees:

1. Traditionally, the beginning of the process of homi- likely that they built nests in trees in which tosleep.nization was associated with the gradual disappearance

of forest and its replacement by grassland. Nonetheless, 5. Initially it was thought that the locomotor activityof the first hominids was characterized exclusively byrecent findings suggest that throughout the Miocene

and at the beginning of the Pliocene the ecological envi- terrestrial bipedalism (Johanson and Edey 1981, Latimer1991). However, the persistence of arboreal anatomicalronment of eastern Africa was a mosaic of distinct bio-

topes—thick forest, degraded forest, gallery forest, and characteristics in Australopithecus (their prehensile ex-tremities) seems to indicate their arboreality (Bergedifferent types of grasslands with varying presence of

trees (Hill 1987, Andrews 1989, Kappelman 1991, Cer- 1992; Clarke and Tobias 1995; Coppens 1991; Deloison1991, 1992; McHenry 1991; Rose 1991; Shreeve 1969;ling 1992, Kingston, Marino, and Hill 1994, Wolde Ga-

briel et al. 1994). In spite of the climatic change in east- Senut 1989a, b; Susman and Stern 1991; Susman, Sternand Jungers 1984; Tardieu 1983). Without fire, withern Africa throughout that period, not all types of forest

were replaced by grassland. Therefore the abundance of practically no nocturnal and crepuscular vision and atype of sleep that requires a recumbent position, andtrees in those habitats may have favoured nocturnal

nesting among hominids. given the danger represented by the large carnivores,Australopithecus and even H. habilis—which ac-2. Diurnal primates, including humans, have ex-

tremely deficient nocturnal and crepuscular vision. cording to Susman and Stern (1991) preserved the typi-cal archaic arboreal characteristics of apes—shouldThey are thus at a considerable disadvantage in compar-

ison with carnivores, whose efficient visual systems have nested in the trees as do modern apes.Nesting is an adaptive behaviour, in all likelihood ex-allow them to operate in poor light. There are a number

of reports of attacks by predators, especially leopards tremely archaic, and persists in modern apes (gorillas,chimpanzees, pygmy chimpanzees, and orangutans) andand lions, on gorillas (Schaller 1963, Fay et al. 1995) and

chimpanzees (Rahm 1967, Boesch 1991, Tsukahara was probably the only behaviour open to African homi-nids of the Pliocene and lower Pleistocene because of1993). Apes seek safety at night in relatively inaccessi-

ble refuges, individual nests which they build in the the factors we have described: predation, ecology, typeof sleep, lack of fire, and arboreal morphological adapta-branches of trees.

3. Savannah ecosystems are inhabited by a large num- tions. Nesting unites members of social units duringsleep and the long periods before and after sleep. As theber of mammals, 98–99% of which are herbivores and

1–2% carnivores. Only 18–19% of herbivore species in- process of hominization progressed, it may well havehad an active socializing function, culminating in homehabit forest areas, while the other 80% exploit open

spaces rich in Gramineae (Schaller 1972, Potts 1988, bases (Isaac 1980) or fix points (Groves and Sabater Pi1985). According to Sept (1992b), the initial distributionDomınguez-Rodrigo 1994). It is clear, then, that in

dense forest trophic pressure is low: the only carnivo- of fix points was similar to that of nesting areas of mod-ern chimpanzees. In this sense, the campsites of humanrous mammal in this area is the leopard (Panthera

parda), a solitary, territorial animal whose trophic pres- nomads and, by extension, the fix points (home bases)of humans have the same basic plan as the nests of apes,sure is low in any case (Domıguez-Rodrigo 1996, Sa-

bater Pi and Vea 1994). The only highly active, dynamic and ‘‘the social organization that lies at the base of the‘nest’ and campsite organisation is derivable from thepredator in dense forest is the eagle Stephanoetes coro-

natus (a monkey-eating eagle), which may attack young common denominator of the apes’ societies’’ (Grovesand Sabater Pi 1985:22).chimpanzees or even young gorillas (Sabater Pi, per-

sonal communication); therefore the building of diurnal In our view, nesting behaviour in Australopithecusand the first forms of hominids would account for thenests in trees is also adaptive. The greatest biomass and

consequently the majority of carnivores are found in persistence of arboreal characteristics in hominids andexplain the continuity of these species in environmentsopen spaces. Here trophic pressure is high. The greater

density of predators—many of which, as we have men- which were not readily habitable by higher primates.




latimer, b. 1991. ‘‘Locomotor adaptations in AustralopithecusReferences Citedafarensis: The issue of arboreality,’’ in Origine(s) de la bipediechez les Hominides. Edited by Y. Coppens and B. Senut, pp.andrews, p. j. 1989. Palaeoecology of Laetoli. Journal of Hu-

man Evolution 18:173–81. 169–76. Paris: CNRS.mc grew, w. c. 1992. Chimpanzee material culture: Implica-baldwin, p. j., j. sabater pi, w. c. mc grew, and

c. e. g. tut in. 1981. Comparisons of nests made by differ- tions for human evolution. Cambridge: Cambridge UniversityPress.ent populations of chimpanzees. Primates 22:474–86.

bellomo, r. v. 1994. Methods of determining early hominid mc henry, h. m. 1991. ‘‘First steps? Analyses of the postcra-nium of early hominids,’’ in Origine(s) de la bipedie chez lesbehavioral activities associated with the controlled use of fire

at FxJj 20 Main, Koobi Fora, Kenya. Journal of Human Evolu- Hominides. Edited by Y. Coppens and B. Senut, pp. 133–41.Paris: CNRS.tion 27:173–95.

berge, c. 1992. ‘‘Analyse morphometrique et fonctionnelle du potts, r. 1988. Early hominid activities at Olduvai. NewYork: Aldine de Gruyter.pelvis des Australopitheques (Australopithecus afarensis, A. af-

ricanus): Interpretations locomotrices et obstetricales,’’ in Cinq rahm, r. 1967. ‘‘Observations during chimpanzee captures inthe Congo,’’ Neue Ergebnisse der Primatologie. Edited by D.millions d’annees: L’aventure humaine. Edited by M. Tous-

saint, pp. 49–62. Liege: ERAUL. Starck, R. Schneider, and H. J. Kuhn. Stuttgart: Gustav FischerVerlag.bert, j., h. ayats, a. martino, and h. collomb.

1967. Le sommeil nocturne chez le babouin Papio papio: Obser- rose, m. d. 1991. ‘‘The process of bipedalization in hominids,’’in Origine(s) de la bipedie chez les Hominides. Edited by Y.vations en milieu naturel et donnees electrophysiologiques. Fo-

lia Primatologica 6:28–43. Coppens and B. Senut, pp. 37–48. Paris: CNRS.sabater pi, j., and j. j. ve a. 1990. Nest-building and popu-boesch, c. 1991. The effects of leopard predation on grouping

patterns in forest chimpanzees. Behavior 117:222–42. lation estimates of the bonobo from the Lokofe-Lilungu-Ikomaloki region of Zaire. Primate Conservation 11:43–48.cerling, t. e. 1992. Development of grasslands and savan-

nahs in East Africa during the Neogene. Palaeogeography/ ———. 1994. Comparative inventory of foods consumed by thewild pygmy chimpanzee (Pan paniscus; Mammalia) in thePalaeoclimatology/Palaeoecology. 97:241–47.

clark, j. d. 1969. Kalambo Falls prehistoric site. Vol. 1. The Lilungu-Lokofe region of the Republic of Zaıre. Journal of Afri-can Zoology 108:381–96.geology, palaeoecology, and detailed stratigraphy of the exca-

vations. Cambridge: Cambridge University Press. schaller, g. b. 1963. The mountain gorilla. Chicago: Univer-sity of Chicago Press.clarke, r. j., and p. v. tob ias. 1995. Sterkfontein Mem-

ber 2 foot bones of the oldest South African hominid. Science ———. 1972. The Serengeti lion. Chicago: University of ChicagoPress.269:521–24.

coppens, y. 1991. ‘‘L’evolution des hominides, de leur locomo- senut, b. 1989a. Le coude des primates hominoıdes: Anato-mie, fonction, taxonomie, evolution. Paris: Editions du CNRS.tion et de leurs environnements,’’ in Origine(s) de la bipedie

chez les Hominides. Edited by Y. Coppens and B. Senut, pp. ———. 1989b. ‘‘La locomozione dei preominidi,’’ in L’evoluzionedegli ominidi: Preominidi e Australopiteci. Edited by G. Gia-295–301. Paris: CNRS.

deloison, y. 1991. ‘‘Les Australopitheques marchaient-ils cobini, pp. 43–52. Milano: Jaca-Book.sept, j. 1992a. Archaeological evidence and ecological perspec-comme nous?’’ in Origine(s) de la bipedie chez les Hominides.

Edited by Y. Coppens and B. Senut, pp. 177–86. Paris: CNRS. tives for reconstructing early hominid subsistence behavior. Ar-chaeological Method and Theory 4:1–56.———. 1992. ‘‘Sur les traces de pas de Laetoli en Tanzanie,’’ in

Cinq millions d’annees: L’aventure humaine. Edited by M. ———. 1992b. Was there no place like home? curent antropol-ogy 33:187–206.Toussaint, pp. 63–72. Liege: ERAUL.

dominguez -rodrigo, m. 1994. El origen del comporta- shreeve, j. 1996. New skeleton gives path from trees toground an odd turn. Science 272:654–64.miento humano. Madrid: Tipo.

———. 1996. En el principio de la humanidad. Madrid: Sıntesis. s ilberbauer, g. 1981. Hunters and habitat in the Central Ka-lahari Desert. Cambridge: Cambridge University Press.fay, j. m., r. carrol, j. c. k. peterhans, and d.

harri s. 1995. Leopard attack on and consumption of gorillas stanford, c. b., and j. s. allen. 1991. On strategic story-telling: Current models of human behavioral evolution. cur-in the Central African Republic. Journal of Human Evolution

29:93–99. rent anthropology 32:58–61.susman, r. l., and j. t. stern. 1991. ‘‘Locomotor behav-fruth, b., and g. hohmann. 1996. ‘‘Nest building behav-

ior in the great apes: The great leap forward?’’ in Great ape so- ior of early hominids: Epistemology and fossil evidence,’’ in Or-igine(s) de la bipedie chez les Hominides. Edited by Y. Cop-cieties. Edited by W. C. McGrew, L. F. Marchant, and T. Nis-

hida, pp. 225–39. Cambridge: Cambridge University Press. pens and B. Senut, pp. 121–31. Paris: CNRS.susman, r. l., j. t. stern, and w. l. jungers. 1984.groves, c. p., and j. sabater pi. 1985. From ape’s nest

to human fix-point. Man 20:22–47. Arboreality and bipedality in the Hadar hominids. Folia Prima-tologica 43:113–56.h ill, a. 1987. Causes of perceived faunal change in the late

Neogene of East Africa. Journal of Human Evolution 16:583– tardieu, c. 1983. L’articulation du genou: Analyse morpho-fonctionnelle chez les Primates, application aux hominides fos-96.

i saac, g. 1980. ‘‘Casting the net wide,’’ in Current argument siles. Paris: CNRS.tsukahara, t. 1993. Lions eat chimpanzees: The first evi-on early man. Edited by L. K. Konigsson. Oxford: Oxford Uni-

versity Press. dence of predation by lions on wild chimpanzees. AmericanJournal of Primatology 29:1–11.james, s. r. 1996. ‘‘Early hominid use of fire: Recent ap-

proaches and methods for evaluation of the evidence,’’ in The tuttle, r. h. 1981. Evolution of hominid bipedalism and pre-hensile capabilities. Philosophical Transaction of the Royal So-Lower and Middle Paleolithic (Colloquium IX of XIII Interna-

tional Congress of Prehistoric and Protohistoric Sciences). Ed- ciety, London B 292:89–94.wolde gabriel, g., t. d. white, g. suwa, p. renne,ited by O. Bar-Yosef, L. L. Cavalli-Sforza, R. J. March, and M.

Piperno, pp. 65–75. Forli: Abaco. j. de he inzel in, w. k. hart, and g. helken. 1994.Ecological and temporal placements of early Pliocene hominidsjohanson, d., and m. edey. 1981. Lucy: The beginnings of

humankind. New York: Warner Books. at Aramis, Ethiopia. Nature 371:330–33.wrangham, r. w. 1987. ‘‘The significance of African apes forkappelman, j. 1991. The paleoenvironment of Kenyapithecus

at Fort Ternan. Journal of Human Evolution 20:95–129. reconstructing human social evolution,’’ in The evolution ofhuman behavior: Primate models. Edited by W. G. Kinzey, pp.k ingston, j. d., b. d. marino, and a. hill. 1994. Iso-

tropic evidence for Neogene hominid paleoenvironments in the 51–71. Albany: State University of New York Press.Kenya Rift Valley. Science 264:955–59.



Documents

Bone Artefacts from the Middle Stone Age at Blombos Cave, Southern Cape, South Africa