A Nonracial Craniofacial Perspective on Human Variation: A(Ustralia) to Z(Uni)

AMERICAL JOURNAL OF PHYSICAL ANTHROPOLOGY 82:341-360 (1990)

A Nonracial Craniofacial Perspective on Human Variation: A(ustralia) to Z(uni)

C. LORING BRACE AND KEVIN D. HUNT Museum o Anthropology, University of Michi an, Ann Arbor, Michigan 48109 (C. .B.); Department of Anthropology, flarvard University, Cambridge,L Massachusetts, 02138 (K.D.H.) KEY WORDS Dental measurements, Race, C scores ABSTRACT Dental and craniofacial measurements were collected for 57 samples from Asia, the Pacific, the aboriginal western hemisphere, and Europe. The craniofacial dimensions include many that are not obviously under the control of specific selective forces. Similar configurations for these in different samples should yield indications of recency of common ancestry according to the logic expressed by Darwin and evident in the relationships indicated by nuclear DNA comparisons. Dental dimensions, however, vary according to the length of time that different intensities in selective forces have been in operation. The craniofacial measurements were transformed into C scores and used to generate Euclidean distance dendrograms. When all the material was used to generate a single dendrogram, the European and Amerindian samples sorted into two regionally identifiable clusters, and the Asian and Pacific material sorted into the three clusters identified in separate previous studies: a Mainland Asian cluster, a Jomon-Pacific cluster and an Australo-Melanesian cluster. Since these clusters are based on variation in traits that are basically nonadaptive in nature, no hierarchical ranking is possible. The clusters simply reflect degree of relationship. This technique holds forth the promise of producing a nonracial assessment of the relationships of all the peoples of the world, past and present.

There is a spectrum of variation in what is shared a common ancestor. There is almost confusingly labeled “anatomically modern” certainly some ethnocentrism inherent in Homo sapiens that is rarely taken into ac- viewing the s ectrum as running from Eu- count in ap raisals of human evolution in rope to Austrayia, but this quite literally does general an individual fossil specimens in extend from one geographical extreme of the articular. 8In tooth size alone, the difference earth to the other, and, dentally at least, the getween the average condition in Australia Australian aborigines can legitimately stand and that in Euro e (or China) would satisfy for a morphological extreme in contempo- the criterion usea to indicate specific differ- rary H. sapiens and Europeans come quite ence (Gingerich, 1974,1979,1980; Gingerich close to representing their antithesis (Brace, and Schoeninger, 19791, yet there is obvi- 1980; Brace et al., in press b). ously no reproductive barrier between Aus- The aboriginal inhabitants of Australia tralians and anybody else in the world. All have been a continuing source of fascina- living human beings are demonstrabl mem- tion for anthropologists and the general bers of the same species even thoug there public alike ever since they were first en- are average visible differences betweenE the countered by itinerant Europeans almost various geogra hically situated populations 400 years ago. With the addition of an evolu- of the world. #he spectrum of variation is tionary perspective developed by Charles there, however, and it should be possible to Darwin after his visit there a century and a deal with it in such a fashion that we can half ago, that fascination invested them with determine how much of it is due to differ- more than just the status of benighted sav- ences in the intensity of s ecific selective agery that had previously prevailed. To be forces and how much of it is Bue to the length of time since the groups being compared Received June 23,accepted July 3,1989.

@ 1990 WILEY-LISS, INC. 342 C.L. BRACE AND K.D. HUNT sure, much of that judgment continued una- 1944; Lack, 1961; Gould and Lewontin, bated, but to it was added an implication of 1979). The rise of the synthetic theory of the ‘‘ rimitive” in an evolutionary sense. evolution redirected scholarly enthusiasm Whet R er overt or uns oken, there has been a towards the study of characteristics with general assumption tR at Australian Aborig- measurable adaptive value, and the matura- ines represent the survival of an earlier tion of the science of genetics focused atten- stage in human evolution. Whatever their tion on traits with simple and discoverable age or sex, they have been repeatedly desig- modes of inheritance, whatever their adap- nated by the collective term “primitive man,” tive si ificance. Partially as a result of a kind of Pleistocene vestige that has been these cr evelopments, the relevance of skele- re arded as a veritable livin fossil. tal studies in general came under an increas- Fhis sort of appraisal has ta e same kind of ingly vocal attack (Boyd, 1950) that has con- t pological essentialism to it that underlies tinued up until the resent (Renfrew, 1987). tE e urge to identify cladistic status or “race” Initially, however, tR is simply served to shift (Brace, 1988). But if cladistic assessment the basis on which classification was estab- and racial designation are equally subjective lished from the mor hological henotypic enterprises, what can we do to make sense in level to the serologicaP genotypic Pevel. our consideration of those aforementioned Classification involved the identification Australian Aborigines or any other human and naming of a presumably fixed number of group during the course of an investigation “races,”whose individual members each con- of the spectrum of variation in H. sapiens? tained traits considered to be uniquely char- There are in fact two general approaches acteristic of the “race” in question. Still ear- that can be productive in very different ways lier, during the nineteenth century, there to help us understand the nature and rela- was a general consensus in physical anthro- tionships of any given population. One such polop that the various “races” had been approach is to assess the adaptive state of fixe entities for an immeasurably long pe- separate traits one by one in the perspective riod of time, possibly reflecting original cre- of the intensit and temporal duration of the ations-separate and unequal (Brace, 1982). individual seIy ective force that a plies to With the superficial addition of an evolution- each. The other is to assess the simiP arities to ary dimension to the expectations of physical and differences from other populations, both anthro ologists as the twentieth century near and far, by using traits and configura- rocee ed, the idea grew that each “race” tions that have little adaptive si ificance in {ad a differentB evolutionary trajectory, and and of themselves. No one stur y of limited that each had evolved to a greater or lesser scope can hope to do complete justice to both extent in comparison to each other (Coon, ap roaches. At best, one can as ire to pro- 1962). The assumption of differences in tim- vi cre an illustration of how and wR y each one ing and degree of that evolution allowed the can work. We attempt to do this in the sec- preservation of an assumed hierarchy no tions that follow. different in effect from the idea that differences in worth had existed ad initio, and it underlies the invidious comparisons that THEORETICAL BACKGROUND continue to be offered (Rushton, 1985,1987, The demise of racial classification 1988). A generation and more ago, physical an- Finally, the conce t of race itself was thropology considered racial classification to shown to be devoid of \ iological justification be one of its principal tasks, and the pre- (Livingstone, 1962; Brace, 1964a,b). Conse- ferred technique used in the pursuit of this uently the whole enterprise of racial classi- goal was the comparative use of nonadaptive Fication, once at the core of professional traits (Hooton, 1926, 1931). Subse uently physical anthropology, was de rived of its the enthusiasm for the use of nona?i aptive scientific credibility and basicaP ly ceased to dimensions of mor hology waned (Hooton, exist. The social scientist, cognizant of the 1946)) and, with tK e Fisherian insistence inequities and injustices perpetrated in the that selection was the sole mechanism con- name of “race,” could well cry “good rid- trollin evolution, it became the accepted dance.” view ta at all aspects of mor hology were But human beings are not identical clones, controlled by selection and notK ing could be and their differences have re ‘onal and tem- regarded as non-adaptive (Dobzhansky, poral aspects to them that sfl ould certainly CRANIOFACIAL PERSPECTIVE ON HUMAN VARIATION 343 be the legitimate concern for biological an- Smouse and Li, 1987; Cann, 1988; Excoffier thropologists. Some investigators have actu- and Langaney, 19891, but there are some ally chosen to deal with one or another trait stochastic problems and other pitfalls that whose manifestation can be associated with keep it from bein our best general model differences in the intensity or duration of (Weiss, 1987; Spuf ler, 1988, 1989). The ap- specific selective forces (Livingstone, 1958, praisal of nuclear DNA on the other hand, in press; Brace, 1967, 1977, 1979, 1980; and despite some of its own attendant prob- Brace et al., 1987, in press c). Others have lems, can function as our basic example. concentrated on particular single gene phe- Ideally this would be best a proached by nomena and investigated the evidence for codon sequencing. In practice,K owever, this genetic drift and the founder effect in island is a tedious and nearly interminable process. PO ulations or other remote genetic isolates Doubts have also been expressed by some (GPass et al., 1952; Friedlaender and Stein- concerning the effectiveness of what has berg, 1970; Neel, 1970; Morton and Lalouel, been demonstrated by crude “brute force” 1973). Somewhat more ambitious attempts DNA-DNA com arisons because of some have been made to use multiple loci to assess of the methodoP ogical problems involved within and between group similarities and (Marks, 1988; Marks et al., 1988). In addi- differences of a more extensive regional na- tion, it was thought that the different adap- ture (Smouse et al., 1982). tive significances of individual base Until recently, however, the whole matter have not been taken into account, an(pairs fur- of population relationships that used to be ther doubts were expressed because of the subsumed under the rubric of “racial classi- indeterminant nature of just what is being fication” has been left relatively untouched. measured when a stated level of concordance Underlying this de facto abdication of con- is reached (McKenna, 1987). This has been cern is the sometimes privately expressed geneticists who fear that any attempt to deal with this mat- the genome does ter would simply allow observers to establish hence is adaptively a hierarchical ranking of human groups with term “junk DNA all of the invidious implications once associ- (Ohno, 1970; Nei, 1983, 1987). Given this, ated with the racial classifications of the the differences observed should be mainly past. We ropose to show, however, that the related to time elapsed since divergence matter oP relationships can be treated in (Sibley and Ahlquist, 1983, 1987a). In any perfectly objective fashion without the impli- case, the arguments over why it should or cations inherent in the terms “race,” “type” should not work have been effectively muted or whatever, and in such a way that no group by the obvious fact that it does. Although can be judged either more or less fit, “higher” arguments continue about various as ects of or “lower,”or better or worse than any other. fine tuning, the overall pictures of lo e- In this endeavor, our efforts represent an netic relationshi it has produced Por i3ir s, extension of the pioneering work initiated at primates, and otR er living creatures (Sibley Mainz by Schwidetzky and those who have and Ahlquist, 1986, 1987a,b; O’Brien, 1987) continued in that tradition (Schwidetzky have convinced many observers that DNA- and Rosing, 1984; Sokal and Uytterschaut, DNA hybridization comparisons work. Fur- 1987; Sokal et al., 1987, 1988; Harding and thermore, they work precisely because so Sokal, 1988). much of the genome has no adaptive significance. The logic of why this should be so was articulated by Charles Darwin himself when Lessons from molecular biology he observed that “as a general rule,. . . the The accumulating record of ongoing and less any part of the organisation is concerned successful research in molecular biology may with special habits, the more important it have shown us a way out of the dilemma becomes for classification” (Darwin, 1859: inherent in “classification.” Certainly it is 414). And he continued with the converse, worth the effort to look at the matter and see noting that “adaptive characters, although if the underlying logic can be extended to of the utmost importance to the welfare of deal with morphology as well as with mole- the being, are almost valueless to the sys- cules. tematist’ (1859:427). The study of mitochondria1 DNA has pro- B “classification” he did not mean any duced promising and provocative results kin of ranking in the eyes of a putative (Wallace et al., 1985; Cann et al., 1987; “creator”B or position in a hypothetical chain 344 C.L. BRACE AND K.D. HUNT of being. Rather he was simply referring to ossible worlds (Gould and Lewontin, 1979). de ee of relationship between one individ- ioth the “hyperselectionists” and the pious uaY or group with another where that degree eighteenth-century objects of Voltaire’s sat- of relationship reflects evolutionary histo- ire represent the conviction that, although ry-the time elapsed since descent from a the workings of nature on the one hand and common ancestor. Darwin, of course, was God on the other may be beyond the grasp of s eaking of aspects of the henotype, and, human reason, the nature of the world as- aP though it is now obvious t Kat his logic also sumed by their various manifestations of a works at the genotypic level, it would be priori faith continues to be fully valid even if interesting to see if we could put it to the test not discernible or testable (Brace et al., in at the phenotypic level at which he conceived press c). of it in the first place.

The construction of nonracial clusters Although this will continue to remain a Nonadaptive features and controversial point, we suggest that there is population relationships atic approach that ma well produce There remains the problem of whether a de act0 solution. With the iernonstration there is such a thing as a “non-adaptivetrait” by human geneticists that “. . . on a per char- and, if indeed there is, how we could recog- acter basis, anthropometric traits discrimi- nize an number or even one of such. We can nate better than genetic traits” (Spielman argue t at old-fashioned physical anthropol- and Smouse, 1976:328), and the pragmatic ogy managedK to discover them in almost support of the theoreticians of numerical ludicrous abundance. For example, one need taxonomy (Sneath and Sokal, 19731, we have only scan the papers and monographs of two been encouraged to use morphometrics in an or three generations ago to encounter count- attempt to deal with human opulation rela- less enumerations of traits that have yet to tionships. Furthermore, rer) erring back to be shown to have any vestige of adaptive the example of the DNA-DNA hybridizers, if significance. What possible difference in se- we sim ly include a plethora of dimensions, lective value mi ht there be in havin a we migK t be able to produce the same kind of cranial contour ta at is ovoid as opposef to “brute force” results. We have inadvertently penta onal-or an other of the myriad of done something very much like this, and we name 2 shapes? Wi y on Earth would it be take this occasion to display the reliminary better to have high rounded orbits under results. This was inadvertent g ecause we some circumstances and low rectangular actually had tried to use a set of measure- ones under others? And when one turns to ments designed to discriminate between the assessment of the “soft parts,” the same opulations in features that presumably dif- kinds of questions remain. What can we pered because of differences in selective force learn about human survival capabilities intensit (Brace et al., 1989, in press a). from an assessment of differences in lip seam Althougi we did succeed in getting a picture development? And of what value were those of population relationshi s that served our extensive investigations of female breast purposes, we discovered tK at the picture be- form other than as measures of the ostensi- came even sharper when we simply added bly legitimized lechery of a previous gen- more measurements without any regard for eration of male investigators (Ploss, 1872; the possible adaptive si ificance of the un- Ploss and Bartels, 1899; Ploss et al., derlying dimension. In Pact, we have merely 1938; Mercier, 1873; Stratz, 1898, 1901; rediscovered what eighteenth century natu- Hoerschelmann, 1904; Lipiez, 1907; Wit- ralists such as Buffon and Blumenbach had kowski, 1907; Bloch, 1909; Rothe, 1912; realized, namely, that the best results will Hooton, 1918; Holl, 1920a,b;Martin, 1928a)? follow from the use of the “greatest possible Those who insist that selection must be the number of characters” (Mayr, 196573). In main controlling factor, even if we cannot see the same somewhat inadvertent vein, the how it works, have been com ared by Gould satisfactory nature of our results demon- and Lewontin to the figure of 5r. Pangloss in strates in em irical fashion the aptness of “Candide,” who maintained, despite the Sneath and Eokal’s observation that, al- mountin record of personal disasters, that though there may be redundancy in the use all was Hor the best in this the best of all of correlated variables, it is more than “coun- CRANIOFACIAL PERSPECTIVE ON HUMAN VARIATION 345 terbalanced” by the additional information whether the reduction recorded in a obtained (Sneath and Sokal, 1973:106). area actually took place in sztu or whet Per it Actually, what DNA-DNA hybridization was the result of immigration and dilution or studies and our own morphometric efforts re lacement by a population from some have produced are branching diagrams de- otK er area. Since we were attempting to test picting phenetic relationships using large the continuity or lack of continuity between quantities of available but unweighted char- prehistoric and modern groups, we were re- acters. To use Cain and Harrison’s term, stricted to the use of variables that could be these are phenograms (Cain and Harrison, assessed with equal ease on both. Of neces- 1960:3; Mayr, 1965) in the classic sense of sity, this meant that we had to seek our data numerical taxonomy (Sokal and Sneath, on skeletal material. And, since our initial 1963; Sokal, 1965; Sokal et al., 1965; Sneath focus was on odontometrics, it was only lo - and Sokal, 19731, or “numerical phenetics” cal that our test should focus on the use oftf e refers (Mayr, 1965, 1982:222).Al- craniofacial material that provided us with thougas tR ere is a perception that this ap- our dental samples. roachMar “has become a bit unfashionable To avoid the possible taint of sub’ectivity rately,” our efforts might just ualify as a associated with the assessment of ch aracter representation of the “comebacR ” expected state or other kinds of nonquantitative anal- by the author of that comment (Dawkins, ysis, we chose to restrict our efforts to the 1986:281). collection of craniofacial metric information. One choleric critic has characterized the We started with a selection of the measure- use of DNA-DNA hybridization to generate ments used by Howells in his study “Cranial such diagrams for the purpose of indicating Variation in Man” (Howells, 1973) and population relationshi s as “a dubious mix- added a few variables that we hoped would ture of phenetic and cP adistic methodology” im rove our ability to test similarities and (McKenna, 1987571, but at least it avoids di8 erences. In successive years, we added the manifest subjectivity of standard cladis- further measurements. until we now have a tic procedure (and see Schwartz, 1984, for a full two dozen. The complete roster appears particularly risible example of the latter). In in Table. .~~ 1.-. any case, we present a series of dendrograms Since we added items to our list after we for modern and recent prehistoric human had started, this means that the groups mea- groups that we suggest have several advantages. First, they agree well with linguistic, ethnographic, and historical accounts, and TABLE l. Cranrofacial measurements used the can be checked against the available in this study arcK aeological information. Second, this ap- 1 Nasal height (Martin No. 55)’ roach provides a means of dealing with 2 Nasal bone height (Martin No. 56[2]) R uman biolo ‘cal variation that can be used 3 Piriform aperture height (Martin No. 55[1]) to oup an%l compare human populations 4 Nasion prosthion length (Martin No. 48) 5 Nasion basion (Martin No. 5) witr out the attendant danger of creating 6 Basion prosthion (Martin No. 40) anything comparable to a racial typology 7 Superior nasal bone width (Martin No. 57[2]) with the possible attribution of relative a Simotic width worth that was inevitably associated with 9 Inferior nasal bone width (Martin No. 57[3]) 10 Nasal breadth (Martin No. 54) this activity in the past. In essence, it is a 11 Simotic subtensel nonracial and noninvidious means of com- 12 Inferior simotic subtense paring the general biological similarities and 13 FOW subtense at nasion differences of human populations. And, fi- 14 MOW subtense at rhinion:’ nally, it allows us to deal with both past and 15 Bizygomatic breadth (Martin No. 54) 16 Glabella opisthocranion (Martin No. 1) living populations, something that is beyond 17 Maximum cranial breadth (Martin No. 8) the reach of those who confine their analysis 18 Basion bregma (Martin No. 17) to the study of the genome itself. 19 Basion rhinion 20 Width at 13 (fmt-fmt) 21 Width at 14 DATA SELECTION AND TREATMENT 22 IOW subtense at nasion:’ Craniofacial variables 23 Width at 22 (fmo-fmo) 24 Minimum nasal tip elevation In the course of investigating the evidence ‘Martin numbers are from Martin (19ZXb). for dental reduction in the recent human /Howells (1973). past, we had to deal with the question of ‘Woo and Morant (1934). 346 C.L. BRACE AND K.D. HUNT sured at the beginning of our project were Data treatment: C score computation represented by fewer variables than the ones and use most recently studied. In some cases, it was To start with, we collected data from a possible to return to the collections and add series of samples using the measurements the variables not measured on the first visit. listed in Table 1. In Table 2, we show the It was in the course of this that we discovered names of the grou s we tested alon with the the curious fact that our ability to cluster number of indivi uals used in eac!E and the and distinguish was more dependent on the locations of the collectionsB in which they are sheer quantity of variables used than it was stored. on the care with which we selected them, a We thought that comparisons of the kind finding that rovides empirical support for we wished to make would be accomplished the intuitive6 based position taken by the best by concentratin on relative proportion promoters of numerical taxonomy (Sneath or "shape" alone an% reducing the effect of and Sokal, 1973). We also discovered that, sheer size on the results of our analysis. To beyond the first twenty or thirty individuals accomplish this, we converted our data to C per group, our reliability in constructing our scores according to the procedure described assessments was not changed even when we by Howells (1986). This adjusts for size by increased our numbers to one or two hun- comparing each measurement on an individ- dred cases for each of the variables. ual to the overall size of all measurements of

TABLE 2. Average and range of N and location of collections for the samples used in phenograms (footnotes give locations at which skeletal collections are housed) N N Population Range Average Population Range Average Admiralty Islands' 2-3 3 Japan, Chiba2 73-80 79 Ain~~,~ 78-114 85 Japan, KofunIfi 9-25 19 Andaman Islands4 22-26 25 Japan, TohokuI7 83-109 102 Australia, North5 18-19 19 Jauan. Tokvo3 116-119 119 Australia, South5 19-22 22 Japan, Kyushu16J8 29-32 31 Baining, New Britain' 8-13 12 JGmon, Early/Middle2,'y 5-19 14 Burma4 36-38 38 Jamon, Late2,Iy 9-37 27 California, South6 27 27 Korea,' 17 17 China, East Coast7 196-241 224 Maori' 27-41 39 China, Hong Kong' 47-52 50 Marq~esas','~ 33-40 37 China, Neolithic9 21-33 27 Michigan Indians, North"' 7-15 12 China, North3 4 1-44 42 Michigan Indians, South2o 33-53 45 China, Siehuan'O 83-92 89 New Caledonia12 20 20 China, Yunnan9 80-106 97 New Hebrides, Vanuatal2 13-14 14 Dayak, Borneo3 10-13 13 New Ireland' 4-6 6 Denmark" 14-22 16 Northwest Coasth 48-50 50 Denmark, Neolithic'' 26-32 30 Norway"' 44-46 46 Easter Island' 15-17 16 Perufi 18 18 England4 33-42 41 Philippines, Visayas2" 29-41 36 Fiji12 7-8 8 Plains Indiansfi 3 3 FranceI2 38-40 40 Ponape' 5-9 8 France, Neolithic'" 3-6 5 Ralum, New Britain' 62-71 70 Germany' 24-41 36 Samurai' 25-32 30 GuamI4 55-79 69 Siberia' 5-7 7 Gulf of Papua, New Guinea4 14-15 15 Southwest Indians" 7 7 HawaiiI4 76-91 87 Thailand"' 72-87 82 Irian Jaya, W. New Guinea4 6 6 Trobriand Islands4 10-11 11 Italy'5 42-50 46 VietnamI2 7-10 9 Yavoi"j 28-37 35 'American Museum of Natural History, New York. ' {Institutde Palebntoiogie Humaine, Paris. 'Sapporo Medical College. 14B.P.Hishup Museum, Honolulu. .'University Museum, University of Tokyo. "Peahody Museum, Harvard. 4i)uckworth Laboratory, Cambridge University. l61l)epartment of Anatomy, Kyushu University Medical School, '1)epartment of Anatomy, University of Edinburgh Medical Fukuoka. School. 171)epartment of Anatomy, Tohoku University Medical School, hLuwie Museum, University of California Berkeley. Sendai. 'Institute of Anthropology, Fudan University, Shanghai. 'XDepartmentof Anatomy, Nagasaki University Medical Schuoi. "Prince Philip Dental Hospital, University of Hong Kong. "'Laboratory of Physical Anthrupology, Kyoto University. gIVPP, Heijing. L"Museum of Anthropology, University of Michigan. "'Chengdu College of Traditional Chinese Medicine. LIAnatomicalInstitute, University of Oslo. "F'anum Institute, Copenhagen. Z'Siriraj Hospital, Bangkok. 1LMusBe de l'Homme, Paris. CRANIOFACIAL PERSPECTIVE ON HUMAN VARIATION 347 the cranium. C scores are similar to ratios in lating missing variables, but we decided that they both are measures of relative size. against this. In our analysis we use up to 24 The advanta e of a C score over a simple measurements, but, instead of calculatin ratio is that ta e C score reflects the relative an average Z score only for individuals wit a size of a given feature in comparison to the all 24 variables, we calculated average Z average relative size of all the other traits scores if 15 or more variables were present used, whereas a ratio can reflect only rela- on any one individual. These average Z tive size in comparison with a single refer- scores were then used to compute the C ent. C scores are calculated as follows. First, scores for each of the variables present in individual unweighted measurements are each individual as in equation 3. A mean C converted into sex-specific standardized score was then calculated for each variable scores (Z scores), where each Z score repre- by sex and by population, and these sex- sents the number of standard deviation specific C scores were used to generate the units by which the value in question departs male-female midsex mean C scores (CiJ)that from the grand mean of that variable for all we then used as the basis of our group- the samples used in a given analysis. by-group com arisons. This can be represented as: More formaP ly,

(4) where i = number of the measurement (e.g., 1 . . .24), j = number of the individual, Xij = valNe of measurement “i” for individ- = ual ‘3,” Xi overall sex specific average where i = number of the measurement (e.g., ui = value for measurement “i,” and overall 1 . . .24),j = number of the individual, J = a sex specific standard deviation for measure- opulation, e.g., Japan, Jf = females only ment ‘5.” i-om population J, J, = males only from From Z scores, C scores are calculated as population J, and, therefore, CijJfrepresents the difference between the Z score of a single the C score for the ith measurement, for the measurement for a given individual and the jth individual among the females of the Jth mean Z score of that individual for all the PO ulation. measurements used in the analysis. 8nce we had generated a mean C score for The mean Z score of an individual is calcu- each variable by population (CiJ), we ana- lated as follows: N lyzed the data using a cluster algorithm available in the Michigan Interactive Data 1 Zij Analysis System (Fox and Guire, 1976). This algorithm produces branching diagrams in which populations are arranged according to their Euclidean distance from one another. where Zj = the average Z score for all the Groups are displayed as twigs on a tree variables for individual “j” and N = the num- where the distance between each branching ber of variables used (e.g., 24 if all are re re- point is roughly pro ortional to Euclidean sented). Given this, the C score for varia f le i distance (Sokal and geath, 1963; Sokal and for individual j is: Camin, 1965; Sokal, 1966; Sokal and Rohlf, - 1969; Sneath and Sokal, 1973; Fox and c..1J = z..1J - z.J’ (3) Guire, 1976). We did a series of trial runs to test whether As Howells described this procedure, it our use of individuals with different num- has the requirement that all variables be bers of variables present for the pur ose of present on each of the specimens used. This generating mean C scores had any e ect on frequently presents a problem in the analy- the nature of the clusters that wereK: pro- sis of rehistoric populations since the num- duced. In the case of the treatment of rela- ber oP complete specimens is often small to tively large groups, i.e., those with Ns of over nonexistent. 25 for each sex, the major clusters were Others have occasionally used various exactly the same. In the cases with smaller multiple regression procedures for interpo- Ns, our procedure tended to produce more 348 C.L. BRACE AND K.D. HUNT satisfying results than was the case when ular aspects of population history and has on1 individuals with complete data were been dealt with elsewhere in greater detail incr uded. For exam le it eliminated the cu- (Brace et al., 1989). It is a curious irony that riosity of seeing an erindian group or two the facial features associated with high so- appear in the Europeank or the Australo- cial status in Japan and glorified in Japa- Melanesian clusters. As will be seen in what nese art evidently represent the genetic con- follows, our results are sufficiently consis- tribution that the despised Ainu have made tent for us to be confident in the validity of to the traditional Samurai class. our procedure. Ja an, however, for all its long and inde- pen CQent history, shares many cultural ele- PRACTICAL APPLICATION ments with the neighboring Asian mainland, Japan and it has been a matter of long-term obser- Our first attem t to use this technique for vation that the Japanese characteristicall a population anaf ysis problem was focused display facial features that ally them wit{ on Japan. We used C scores as the input to their mainland Asian neighbors. At the same generate a Euclidean distance dendrogram time, during work on the prehistoric Jdmon to see whether the modern Japanese clus- skeletal collections, we were struck by the tered with their J6mon predecessors or with similarities between the configuration of other previous and modern Asian their features and those visible in some of the First we restricted our attention to t PPS.e data crania from Oceania. The obvious course of from the Japanese archi elago itself. action was to test this by including Japanese, Figure 1 shows that the mo ern Japanese Oceanic, and mainland Asian crania in a fall into the same cluster asL9 the Yayoi rice single analysis. The results of this can be agriculturalists who entered Japan in 300 BC seen in Fi re 2. and also with the Kofun tomb builders, who As can re seen, the Japanese along with were the immediate predecessors of the the prehistoric Yayoi rice agriculturalists emergence of the Yamato State in the sixth are included in the same cluster with the century. The Jomon, associated with the de- mainland Asian groups tested. The Main- velopment of potte in Ja an more than land Asian cluster further is broadly sepa- 10,000 ears ago, fa 1 into t e cluster with rated into a northern and a southern compo- the mo ern Ainu. ThisTR confirms the conclu- nent. The northern component includes the sions previouslyB reached by the appraisals of northern Chinese Neolithic, modern coastal craniofacial form (Koganei, 1903, 1927; Chinese, Koreans, and Japanese. The Ainu Brace and Nagai, 1982) and the nonmetric and the prehistoric Jomon, however, fall into characteristics of the dentition (Turner, a cluster with Micronesians and Polyne- 1976,1986). sians. This has been called the J6mon-Pa- The J6mon-Ainu association of a utative cific cluster (Brace et al., 1989, in press a, b). group of Samurai, killed during t ge over- throw of the Kamakura Shogunate in 1333, is an interesting demonstration of how this Oceania and Australia analytical approach can cast light on partic- The process by which the settlement of Oceania was accomplished has been an ac- tive focus of interest by archaeologists and Samurai culture historians for some time (Green, Late J6mon 1973,1986; Bellwood, 1975,1979; Jennings, EarlyIMiddle Jdmon 1979), and some very contradictory models have been proposed. These range from the Yayoi vision of a figurative Austronesian “fast Kofun train” from the Asian mainland, or the pos- Tdkyo tulation of a Lapita “homeland in the Bis- Tohoku Chiba marcks of northwestern Melanesia (Allen, YvGshG 1984; Kirch et al., 1989), to the suggestion that Polynesians emerged by transformation out of eastern Melanesians somewhere in Fig. 1. Euclidean distance dendrogram comparing the rehistoric and modern populations of Japan. All the neighborhood of New Caledonia (Terrell, denfrograms are based on C scores; see text for explana- 1986). This would appear to be an ideal tion. situation in which to test the effectiveness of CRANIOFACIAL PERSPECTIVE ON HUMAN VARIATION 349

Samurai Polynesia Micronesia JBrnon

North China 1 Japan 1 Chinese Neolithic Yayoi Korea Hong Kong East Coast China Sichuan Y unnan Vietnam Thailand

Fig. 2. Euclidean distance dendrogram comparing a series of populations from the Asian mainland, Japan, and the islands of Micronesia and Polynesia. the approach we are advocating. We have Ainu and the prehistoric Jdmon of Japan, presented a reliminary treatment of this there is virtually no likelihood that any matter elsewK ere (Brace et al., in press a), members of what has been termed the but we have subsequently added some fur- Jomon-Pacificcluster could have emerged by ther populations to fill in areas previously transformation from anything in either unrepresented. western or eastern Melanesia. And with the In dealing with this matter, we have had to representatives of the latter from New Brit- organize things to consider the oft-reiterated ain and New Ireland to New Caledonia and observation that the Melanesian inhabitants the New Hebrides solidly associated in the of Oceania are closely related to Tasma- same cluster with both northern and south- nians, Australians, and the inhabitants of ern Australian aborigines as well as groups New Guinea (Sarasin, 1924; Wagner, 1937; from the Gulf of Papua in New Guinea, there Howells, 1973). Figure 3 shows what hap- is every reason to believe that the members pens when we compare the various eoples of of the Australo-Melanesian cluster have a the Pacific with those who rofuced the common inheritance that goes well back into Mainland Asian cluster. The !omon-Pacific the Pleistocene. Archaeological evidence for cluster retains its identit , and all the human occupation of the Swan River region oups found in Melanesia, IJew Guinea, and of southwestern Australia (Pearce and Bar- Tustralia form a separate coherent cluster, betti, 1981) and the Huon Peninsula of which we have termed the Australo-Melane- northern New Guinea goes back more than sian cluster. It is a matter of added interest 40,000 years (Groube et al., 1986) and is that the available specimens representing supported by prepottery sites of more or less the Andaman Islands also fall into this clus- comparable antiquity in the Bismarck Archi- ter, althou h the tie is not close. Still, the pelago of western Melanesia. Australo-2 elanesian group with which they Altogether, this sustains the suspicion form their closest association is the western- previously offered, largely on the basis of most for which we have data from the island tooth size alone, that an area extending from ofNew Guinea itself. This, with other subjec- the Andaman Islands in the west through tive hints previously noted, gives some sup- Borneo and the Philippines to the large is- port to the expectation that groups that lands of Melanesia and includin New would have been reco ized as members of Guinea and Australia was populated By peo- this cluster once may 8"ave been distributed ple of common ancestry and similar appear- west of New Guinea throughout- what is now ance durin the latter part of the Pleistocene Indonesia. (Brace an8 Hinton, 1981). Figure 4 shows With Easter Island, Hawaii. and the Maori the putative extent of this related series of New Zealand in the far Pacific fitting of populations and suggests the routes comfortably in the same cluster with the by which their spread into the regions be- 350 C.L. BRACE AND K.D. HUNT

Andaman Islands West New Guinea Trobriand Islands Gulf of Fapua New Hebrides New Britain North Australia South Australia FIJI New Caledonia

Jdmon Japan Ainu Samurai Hawaii Guam Maori Marquesas FonaDe

Shanghai China Korea Neolithic China Yayoi Tokyo Japan Siberia Vietnam Thai Burma Borneo Philippine

Fig. 3. Euclidean distance dendro am comparing samples from the Asian Mainland, Micronesia, Polynesia, Melanesia, and fuustralia. yond Wallace’s line was presumably accom- further evidence (Suzuki, 1981; Suzuki and plished. Hanihara, 1982; Turner, 1983, 1989, in Then, to get a population of a ver differ- press), then there is at least a tentative ent appearance out into the small isr ands of reason to offer Japan and the Ryukyu Archi- the far Pacific between 4,000 and 2,000 years pelago as the source from which subsequent ago, some version of Bellwood’s ex ress train ex ansion produced the distribution of the model has to be invoked. There Kave been reP ated peoples who form the Jomon-Pacific archaeological attempts to identify a south- cluster. This possibility is illustrated by the east Asian source (Solheim, 1964, 1972, routes of movement depicted in Figure 5. 1976; Spoehr, 1973), and Turner has used data from dental morphology to suggest that Other parts of the world what had been Sundaland at the time of The success we have had using this clus- lowered sea levels in the Late Pleistocene ter-producing method to suggest population was the most likely area (Turner, 1986, relationships in Asia and the Pacific has led 1987). us to extend our efforts to include represen- However, the oldest human skeletal mate- tatives from other major regions of the world. rial solidly aligned with the members of the Figure 6 shows what happens when a series Jomon-Pacific cluster are the Early Jomon of European and Amerindian groups is specimens themselves from approximately added to what has already been examined. It 6,000 years ago in Japan. On purely archae- is no surprise to discover that the two addi- ological evidence, the Jomon goes back much tional series make distinct, separate clus- farther than that (Ikawa-Smith, 1980; Pear- ters. In the Amerindian cluster, one of the son, 1986), and, if the 18,000-year-old Mi- constituent groups, the Southwest, appears natogawa skull from Okinawa can be used as to be less closely related to the others and, in CRANIOFACIAL PERSPECTIVE ON HUMAN VARIATION 351

Fig. 4. The stippling represents the suggested area of occupation late in the Pleistocene of a large-toothed series of people of relatively similar appearance. The arrows show the presumed routes of initial Late Pleistocene movement into the Phili pines and beyond Wallace’sLine into the large islands of Melanesia and Australia. (Adapteffrom Brace and Hinton, 1981, with permission of the publisher.) a sense, almost intermediate between the producing pro am, but we included it any- Amerindian cluster and the European clus- way since we i?- ad nothing else to represent ter. As it happens, we were dubious about that end of the Australia-to-Zuni spectrum. this even before we ran the cluster- Our suspicions arose from the possibility 352 C.L. BRACE AND K.D.HUNT

Fig. 5. The suggested route of s read of the Jomon-Pacific movement from its hypothetical area of origin out into the small isEnds of the Pacific. (Adapted from Brace et al., 1989, with permission of the publisher.) that the form of the skulls from the American generates Mahalanobis D2 figures, which Southwest had been altered by artificial de- express a numerical form of the Euclidean formation. A great deal of the available ma- distance between any given group and each terial was rejected out of hand because the of the others (Sokal, 1965). It is our hope that deformation was obvious. Although we tried we may eventually be able to use such fig- to select specimens that did not appear to ures as measures of se aration time when have been deformed, nonetheless we still we have been able to col ect enough informa- had the uneasy feeling that there were virtu- tion on linea e change throuP h time and on ally no completely undeformed individuals. reliably date5 separations of anown groups. Obviously we need to add many more repre- At the moment, however, this is a long sentatives of each region treated before we way from being achieved, which is why we can be more than tentative about the nature have not included a table of the already- and extent of each such grouping. ponderous matrix of D2figures. What does come as something of a surprise, however, is the relatively complete TOOTH SIZE AS A MODEL FOR ADAPTIVE separation of the Amerindian cluster from TRAIT CONSIDERATIONS the Mainland Asian and the Jomon-Pacific Because of the quantity of information clusters. Evidently the 12,000 or 13,000 year available and the various other reasons men- time span since the presumed initial move- tioned in extensive previous discussions ment into the New World (Irving, 1985; (Brace, 1967, 1978, 1979, 1980; Brace and Fagan, 1987; Haynes, 1988) was sufficient Mahler, 1971; Brace and Hinton, 1981; for the roduction of the distinctions ob- Brace and Nagai, 1982; Brace et al., 1984, served. 5he computer program that pro- 1987; Brace and Vitzthum, 1984), we have duces dendrograms from our C scores also concentrated our consideration of dimen- CRANIOFACIAL PERSPECTIVE ON HUMAN VARIATION 353

Group Ainu

Samurai I Jdmon I Hawaii Guam Maori Marquesas Ponape Easter Island

Vietnam Thailand Burma Japan China Neolithic Yayoi Korea China

lrian Jaya ~ I Trobriands Gulf of Papua I New Ireland New Britain N Australia 7 S Australia New Caledonia 1

Southwest Mid-West Plains I Northwest Coast S. California Peru

Italy Germany England Norway Denmark France Da n is h Neolithic French Neolithic

Fig. 6. Euclidean distance dendrogram showing the nature of the clusters when a series of representative Amerindian and Euro ean groups are added. The clusters shown in Figure 3 retain their identit , and Europe and t\e whole ofthe aboriginal western hemisphere appear as comparably identigable clusters. sions under obvious selective force control on and postcranial remains by individual, if the single trait of tooth size. We are aware, of indeed the postcranium is represented at all. course, that there are many parameters in- On the other hand, the teeth on which our volved in tooth size, as can be seen when measurements have been made are usually mesial-distal and buccal-lingual dimensions still in or at least associated with the skulls are treated separately or when anterior and of their original owners. Since cranial mea- postcanine or maxillary and mandibular surements can give us a fair ap roximation teeth are compared (Brace et al., 1987). For of brain size (Pearson, 19261, an I! since brain purposes of simple demonstration, however, size is allometrically related to body size there is considerable heuristic value in treat- within a given species (Lande, 1979; Martin, ing tooth size as though it were a simple 1983; Martin and Harvey, 1985; Riska and adaptive trait. Atchley, 19851, we have used the cranial and However, tooth size, like any other metric dental dimensions of the collections studied feature, shares a measure of common vari- to investigate the relations between tooth ance with body size as a whole. It would be size and body size. Our measure of tooth size, desirable, then, to treat it in conjunction TS, is the cumulative sum of the cross-sec- with some kind of body size index. Stature is tional areas (MD x EL) of all the maxillary notorious1 oor as an indicator of body and mandibular tooth classes (Brace, 1980). mass, an8 urthermore, in many skeletal To calculate brain size, we treated the collections, thereP is no way to sort cranial height, width, and length measurements of 354 C.L. BRACE AND K.D. HUNT the brain case as though they were the major Flower himself observed over 100 years ago axes of an ellipsoid and used the standard when he treated groups from those areas formula to calculate its volume. Obviously (Flower, 1885). we are aware that this will produce an over- When we break down our com arisons to estimate of the actual size of the brain for single dimensions and plot tootYl size and several reasons (and note the attempt to brain size separately on the same graph, as correct for this made by Pearson in 19261, in Figure 8, it is obvious that they vary in but it does ’ve us a crude figure with which quite unrelated fashion for the same set of to work, anf a case can be made that this will populations plotted here in the same order as rise and fall in close approximation to the in Figure 7. The brain size-tooth size corre- variation of true brain size itself. lation across these groups in fact is negative, Over a century ago, the English anatomist being -0.384. This provides a graphic dem- W.H. Flower used an even cruder index to onstration of the point that tooth size and determine whether the dentition in a given body size have become decoupled during the population was relatively large, medium, or recent course of human evolution (Brace et small--“microdont, mesodont, and mega- al., 1987).A full treatment of the changes in dont” to use his terms (Flower, 1885). To do human tooth size that have followed the this, he measured the distance from the alterations in the relevant selective forces is mesial surface of the maxillary first premo- being resented elsewhere (Brace et al., in lar to the distal surface of the third molar, a press cI; . length that can be ap roximated by sum- ming the mesial-distal fimensions of the five CONCLUSIONS teeth involved. He then divided this figure A full understanding of the course of evo- by the nasion-basion length. lution requires the study of both phyloge- Flower’s index, of course, is simply a pro- netic relationships and the dynamics by portion between two linear measures. Since which individual traits change through time, our own appraisal of tooth size involves the Darwin realized this 130 years ago with a calculation of areas and our approximation clarity that has been appreciated by too few. ofbrain size is avolumetric measure, we took Now that we have the advantages of both the square root of the first and the cube root data and techniques that were unavailable of the second in order to allow a direct com- in the past and can put his suggestions to the parison of Flower’s index with our own. De- test, his genius seems ever more impressive. spite the relatively crude approximations But the effort to apply these to deal with used by Flower, his measures show surpris- roblems in human phylogeny and evolution ingly high correlations with ours. For exam- [as suffered from their association with the ple, the correlation between our calculation invidious enterprise of racial classification of cranial volume and basion-nasion length in the physical anthropology of yesteryear. is 0.813, and that between his length mea- This is clearly expressed by Colin Renfrew in sure of the maxillary postcanine teeth and his recent book “Archaeology and Lan- our cross-sectional area of the entire denti- gua e,” in which he says, “craniometry, the tion is 0.926. Even though we believe that stuf y and measurement of human skulls, our own approach should yield a more reli- has in recent years enjoyed about as much able assessment, it is clear that Flower’s prestige in scientific circles as phrenology” index does indeed produce an easily calcu- (Renfrew, 1987:4). He continues, “there are lated and very useful assessment of relative at present few conclusions that can be relied tooth size. upon” (p. 5) and concludes, “it would in my Figure 7 compares the assessment of a view be wrong to place much weight upon series of Asian and Oceanic populations by conclusions drawn from physical anthropol- Flower’s method and by our own, where the ogy until the methodology is better devel- left-to-right position of the bars on the graph oped (p. 93). is determined by increasing magnitude of We would argue, however, that the prob- the TSkranial volume index. As can be seen, lem is less with the methodolo than with the populations that are placed in the the vision, or, rather, the lack oB vision, of its Jbmon-Pacific cluster in our previous cranio- users. As we have been able to show above, facial analysis fall largely at the microdont the use of nonadaptive traits to survey some end of the spectrum, the Australo-Melane- re resentative human populations has en- sians all prove to be megadont, and the abedP us to produce an assessment of the Mainland Asians are in between, just as strength of their genetic relationships one to CRANIOFACIAL PERSPECTIVE ON HUMAN VARIATION 355

40 Legend rZa Flower ITS/Volume 35

Fig. 7. The tooth sizekranial volume index for a series of Jomon-Pacific, Mainland Asian, and Australo-Melanesian PO ulations compared to Flower's index for the same groups. The increase in magnitude from peft to right is keyed to the TSBS index. Based on the data in Tables 3 and 4. another. Then, when we turned to the treat- 1985;by the University of Michigan Museum ment of a trait under the changing influence of Anthropology Field Research Fund in of selection for different known eriods of 1954, 1985, 1986, and 1988; by the L.S.B. time in different areas of the worl$ we have Leakey Foundation in 1986; by Diana Bla- been able to make interpretive sense out of a ban Holt in 1987; and by the National Sci- specific instance of evolutionary change. It is ence Foundation, grant BNS-8616298, in our hope that other investi ators will follow 1987. Essential assistance has also been pro- this lead, dealing with stilf further human vided by G. Brace of Nanterre, France; by populations and other traits in the manner M.L. Brace and Roger C. and Pe gy L. Brace that we recommend. By these means, a co- of Ann Arbor, Michigan; by J.a . Deetz of herent picture of human populations rela- Berkeley, California; by Professors Ben R. tionships and trait modifications-in sum, Finney and Wilhelm G. Solheim, 11, Univer- human evolution-has a very real chance of sity of Hawaii; by Mark A. Gordon and emerging in the near future. Chacma, Inc., New York, New York; by Pro- fessor William w. Howells, Harvard Univer- sity; by Professors Raymond C. Kelly, Frank ACKNOWLEDGMENTS B. Livingstone, and Henry T. Wright, Uni- The research on which the project is based versity of Michi an; by Dr. Pamela D. Raspe, was accomplished in art with sup ort pro- University of 8ambridge; by Dr. B. Holly vided by the Horace !I.Rackham 8chool of Smith, Museum of Anthropology, University Graduate Studies at the University of Mich- of Michigan; and by Professor Robert R. igan in 1973, 1977, 1980 and 1983; by the Sokal, State University ofNew York at Stony Committee on Scholarly Communication Brook. Important financial support was also with the People's Republic of China, Na- provided by the late E.B. Hoa land and by tional Academy of Sciences, in 1980 and the late Professor and Mrs. G.#. Brace. 356 C.L. BRACE AND K.D. HUNT

2000 2000

1800 1800 “E 0 “E .-c E 1600 1600 .g i a, -3 .-N 2 m - f .-0 1400 1400 o c 0 e cf

1200 1200

Legend tZTooth Size ICranial Volume 1000 1000

Fig. 8. Absolute tooth size and cranial volume figures for the same groups in the same order as that seen in Figure 7. Based on the data in Tables 3 and 4.

TABLE 3. Total tooth size (TS),cranial volume (CV),summ.ed mesial-distal dimensions for maxillary Pl-M3, nasion-basion figures, Flower’s index, and the TS/CVinder for a series of Asian and Pacific groups Pl-M3 Flower’s Cranial Tooth TS’ L/cranial Population Basion-nasion Length index volume size volume‘ ~ Jomon 102.90 40.60 39.5 1,838.1 1,134.3 27.5 Ainu 104.15 41.68 40.0 1,805.4 1,132.3 27.6 Maori 102.80 42.22 41.1 1,800.6 1,223.0 28.8 Easter Island 107.40 44.42 41.4 1,786.8 1,278.6 29.5 Ponape 103.00 43.13 41.9 1,848.2 1,218.1 28.4 Samurai 101.14 42.62 42.1 1,751.7 1,196.5 28.7 Yayoi 99.14 43.09 43.5 1,740.3 1,231.4 29.2 Guam 102.43 44.65 43.6 1,828.2 1,310.9 29.6 Japan 99.24 43.34 43.7 1,708.6 1,225.8 29.4 Korea 97.31 42.88 44.1 1,707.3 1,229.3 29.4 China 97.00 42.87 44.2 1,708.3 1,197.3 29.0 Thailand 95.81 43.22 45.1 1,673.7 1,218.3 29.4 Borneo 96.89 43.72 45.1 1,645.4 1,195.9 29.3 Fiji 99.52 45.57 45.8 1,706.1 1,341.9 30.6 N. Australia 99.13 45.64 46.0 1,560.8 1,359.7 31.8 S. Australia 100.98 46.67 46.2 1,694.2 1,474.8 32.2 New Hebrides 97.64 45.12 46.2 1,632.4 1,337.6 31.1 New Guinea 97.80 45.31 46.3 1,587.4 1,356.0 31.5 New Britain 98.39 45.77 46.5 1,587.3 1,355.2 31.5 Andamans 92.23 43.20 46.8 1.506.5 1.198.5 30.2 New Ireland 96.13 45.45 47.3 11620.4 1:266.0 30.3

For access to and help in the various collec- Panum Institute, University of Copenhagen; tions under their care, we are grateful to Dr. to Professor James J.F. Deetz, University of Verner Alexandersen and Dr. Pia Bennike, California, Berkeley; Professor Yukio Dodo, CRANIOFACIAL PERSPECTIVE ON HUMAN VARIATION 357

TABLE 4. Average and range of N for the TS/Cranial Bellwood P (1975) The prehistory of Oceania. Curr. Volume index Anthropol. 16:9-28. Bellwood P (1979) Man’s Conquest of the Pacific: The TSN Cranial - Prehistory of Southeast Asia and Oceania. New York Population volume N Range Average Oxford University Press. ~ ~~ Bloch A (1909) Presentation de portraits des jeunes J6mon 23 17-73 49 negresses pour faire voir la form articuliere de Ainu 106 91-196 137 l’areole de la mamelle. Bull. Mem. loc. Anthropol. Ponape 7 8-15 11 Paris (V serie) 10:141-142. “Samurai” 28 29-50 42 Boyd WC (1950) Genetics and the Races of Man: An Maori 40 1-26 9 Introduction to Modern Physical Anthropology. Bos- China 227 25-234 108 ton: Little, Brown and Co. Borneo 13 3-1 1 6 (1964a) Yayoi 36 25-64 48 Brace CL The concept of race. Curr. Anthropol. Japan 331 153-305 264 5:313-320. Korea 17 13-44 22 Brace CL (1964b) A nonracial approach towards the Thailand 82 50-76 63 understanding of human diversity. In AMontagu (ed): Easter Island 15 1-21 10 The Concept of Race. New York: The Free Press of Guam 68 38-89 70 Glencoe, pp. 103-152. Andaman Islands 27 1-19 7 Brace CL (1967) Environment, tooth form and size in the New Ireland 6 3-21 8 Pleistocene. J. Dent. Res., 46[Suppll:809-816. Fiji 7 2-5 3 Brace CL (1977) Occlusion to the anthropological eye. In New Hebrides 14 3-14 11 McNamara, JA (ed): The Biology of Occlusal Develop- New Britain 71 13-67 42 ment. Craniofacial Growth Series Monograph No. 7 New Guinea 15 2-1 1 6 Ann Arbor, Michigan: Center for Human Growth and N. Australia 18 143-150 149 Development, pp. 179-209. S. Australia 21 164-338 209 Brace CL (1978) Tooth reduction in the Orient. Asian Perspect. 19:203-219. Brace CL (1979) Krapina “classic”Neanderthals,and the evolution of the European face. J. Hum. Evol. 8.527- Sa poro Medical College; Dr. James P. Gar- 550. licE and Corinne Duhig, Cambridge Univer- Brace CL (1980) Australian tooth-size clines and the sity; Professor Kazuro Hanihara, University death of a stereotype. Curr. Anthropol. 21:141-164. Brace CL (1982) The roots of the race concept in Ameri- of Tokyo; Dr. Lean-Louis Heim, Musee de can physical anthropology. In F Spencer (ed): A His- YHomme, Paris; Dr. Per Holck, University of tory of American Physical Anthropology 1930-1980. Oslo; Professor Jiro Ikeda, K oto University; New York: Academic Press, pp. 11-29. Professor Toshihiro Ishii, $ohoku Univer- Brace CL (1988) Punctuationism, cladistics and the leg- acy of medieval Neoplatonism. J. Hum. Evol. 3:121- sity; Professor Nina Jablonski, University of 138. Hong Kong; Professor Li Yongyi, Chengdu Brace CL, Brace ML, Dodo Y, Hunt KD, Leonard WR, Li College of Traditional Chinese Medicine; Y, Shao X-q, Sangvichien S,and Zhang Z-b Microne- Professor Henry de Lumley, Institut de Pa- sians, Asians, Thais and relations: A craniofacial and leontologie Humaine, Paris; Professor Masa- odontometric perspective. J. Micrones. SOC.(in press a). fumi Nagai, Kyushu University; Professor Brace CL, Brace ML, and Leonard WL (1989) Reflections Yoshiatsu Naito, Nagasaki University; Pro- on the face of Japan: A multivariate craniofacial and fessor Michael Pietrusewsky, University of odontometric perspective. Am J. Phys. Anthropol. Hawaii; Professor David Pilbeam, Harvard 78:93-113. Brace CL, and Hinton RJ (1981) Oceanic tooth-size University; Professor George J. Romanes, variation as a reflection of biological and cultural University of Edinbur h Medical School; mixing. Curr. Anthropol. 22:549-569. Professor Sood Sangvica ien, Siriraj Hospi- Brace CL, Li Y, Smith SL, Hunt KD, and Zhang Z-b A tal, Bankok; Professor Shao Xiangqing, craniofacial comparison of circum-Pacific eo les In Fudan University, Shanghai; Dr. Yoshihiko DR Croes (ed): Circum-Pacific History &nkrence, Seattle, August 9,1989. Pullman Washington, Wash- Sinoto, B.P. Bishop Museum, Honolulu; Dr. ington State University Press (in press b). Christ0 her B. Stringer, British Museum, Brace CL, and Mahler PE (1971) Post-Pleistocene Natura History, London; Dr. Ian Tattersall, changes in the human dentition. Am. J. Phys. Anthro- AmericanP Museum of Natural History, New pol. 34:191-204. York; and Professors Wu Jukang and Wu Brace CL, and Nagai M (1982) Japanese tooth size, past and present. Am. J. Phys. Anthropol. 59r399-411. Xinzhi and Zhang Zhenbiao, Institute ofVer- Brace CL, Rosenberg KR, and Hunt KD (1987) Gradual tebrate Paleontology and Paleoanthropol- change in human tooth size in the late Pleistocene and ogy, Academia Sinica, Beijing. post Pleistocene. Evolution 41 :705-720. Brace CL, Shao X-q, and Zhang Z-b (1984) Prehistoric LITERATURE CITED and modern tooth size in China. In FH Smith and F (eds): The Origins of Modern Humans: A Allen J (1984) In search of the Lapita homeland. J. orld Survey of Fossil Evidence. New York: Alan R. Pacific History 19(4):186-201. Liss, Inc., pp. 485-516. 358 C.L. BRACE AND K.D. HUNT

Brace CL, Smith SL, and Hunt KD What big teeth you tion and Transformation. Cambridge: Cambridge Uni- had Grandma! Human tooth size, past and present. In versity Press, pp. 51-54. CS Larsen (ed): Advances in Dental Anthropology. Groube L, Cha ell J, Muke J, and Price D (1986) A New York: Alan R. Liss, Inc., in press c. 40,000 year o8human occu ation site at Huon Penin- Brace CL, and Vitzthum VJ (1984) Human tooth size at sula, Papua New Guinea. dture 324:453-455. Mesolithic, Neolithic and modern levels at Niah Cave, Harding RM, and Sokal RR (1988) Classification of the Sarawak: Comparison with other Asian populations. European langua e families by genetic distances. Sarawak Museum J. 33(54):75-82. Proc. Natl. Acad. &i. USA 85:937&9372. Cain AJ, and Harrison GA (1960) Phyletic weighting. Haynes CV Jr (1988) Geofacts and fancy. Nat Hist. Proc. Zool. SOC.London 135:l-31. 97(2):4-12. Cann RL (1988) DNA and human origins. Annual Re- Heim J-L (1976) Les Hommes Fossiles de La Ferrassie. view ofAnthropology 17:127-143. Tome I. Le Gisement. Les Squelettes Adultes (Crane et Cann RL, Stoneking M, and Wilson AC (1987) Mitochon- Squelette du Tronc). Paris: Masson. Archives de drial DNA and human evolution. Nature 325:31-35. I'Institut de Paleontologie Humaine, Memoire 35. Coon CS (1962) The Origin ofRaces. New York: Alfred A. Hoerschelmann E (1904) Uber die form der Mamma bei Knopf. der Estin, mit Bemerkungen uber die Mammaform bei Darwin CR (1859) On the Origin of Species by Means of einigen anderen Volkerschaften Z. Morphol. Anthro- Natural Selection, or the Preservation of the Favoured pol. 7:22-62. Races in the Struggle for Life. London: John Murray. Holl H (1920a) Uber ein merkwiirdige Verhalten der Dawkins R (1986) The Blind Watchmaker. New York: Briiste bei einer Buschmannfrau. Mitteilungen An- WW Norton. thropol. Gesellschaft Wien 50:7-10. Dobzhansky T (1944) On species and races of living and Holl H (1920b) Uber den Abstand der Brustwarzen fossil man. Am. J. Phys. Anthropol. 2:251-265. voneinander und die Sinusbreite beim weiblichen Ge- Excoffier L, and Langaney A (1989) Origin and differen- schlecht. Mitteilungen Anthropol. Gesellschaft Wien tiation of mitochondria1 DNA. Am. J. Hum. Genet. 5O:ll-22. 44:73-85. Hooton EA (1918) Some early drawings of Hottentot Fagan BM (1987) The Great Journey: The Peo ling of women. Harvard Afr. Stud. (Varia Africana 11) 2:8% Ancient America. New York: Thames and HuAon. 99. Flower WH (1885)On the size of the teeth as a character Hooton EA (1926) Methods of racial analysis. Science of race. J. Anthropol. Inst. Great Britain Ireland 63: 75-81. 14:183-186. Hooton EA (1931) Up From the Ape. New York: Mac- Fox DJ, and Guire KE (1976) Documentation for MIDAS millan Co. (3rd Ed). Ann Arbor: The Statistical Research Labora- Hooton EA (1946) Up From the Ape (2nd Ed). New York: tory, University of Michigan. Macmillan Co. Frayer DW (1978) Evolution of the Dentition in Upper Howells WW (1973) Cranial Variation in Man: A Study Paleolithic and Mesolithic Europe. Lawrence, Kansas: by Multivariate Analysis of Patterns of Difference University of Kansas Publications in Anthropology, among Recent Human Populations. Papers of the Pea- No. 10. body Museum of Archaeology and Ethnology, Vol. 67. Friedlaender JS, and Steinberg AG (1970) Anthropolog- Cambridge: Harvard University Press. ' ical significance of the gamma globulin (Gm and Inv) Howells WW (1986) Physical anthro ology of the prehis- antigens in Bougainville Island, Melanesia. Nature toric Japanese. In RJ Pearson, E)L Barnes, and KL 2359-61. Hutterer (eds): Windows on the Japanese Past: Stud- Gingerieh PD (1974) Size variability ofthe teeth in living ies in Archaeology and Prehistory. Ann Arbor: Center mammals and the dia osis of closely related sympat- for Japanese Studies, pp. 85-99. ric fossil species. J. Pgontol. 48:895-903. Ikawa-Smith F (1980) Current issues in Japanese Ar- Gingerich PD (1979) The stratophenetic approach to chaeology. Am. Sci. 68:134-145. phylogeny reconstruction in vertebrate paleonotology. IrvingWN(1985)Context andchronologyofearlymanin n J Cracraft and N Eldredge (eds): Phylogenetic the Americas. Annu. Rev. Anthropol. 14:529-555. Analysis and Paleontology. New York: Columbia Uni- Jennings JD (ed) (1979) The Prehistory of Polynesia. versity Press, pp. 41-77. Cambridge: Harvard University Press. Gingerich PD (1980) Evolutionary atterns in early Kirch PV, Swindler DR, and Turner CG I1 (1989) Human Cenozoic mammals. Annu. Rev. Earth Planet Sci. skeletal and dental remains from Lapita sites (1600- 8:407-424. 500 BC) in the Mussau Islands, Melanesia. Am. J. Gingerich PD, and Schoeninger MJ (1979) Patterns of Phys. Anthropol. 79t63-76. tooth size variability in the dentition of primates. Am. Koganei Y (1903) Uber die Urbewohner von Japan. J. Phys. Anthropol. 51:457-466. Globus 84(7):101-106. Glass B, Sacks MS, Jahn EF, and Hess C (1952) Genetic drift in a religious isolate: An analysis of the causes of Koganei Y (1927) Zur Frage der Abstammung der Aino variation in blood groups and other gene frequencies in und ihre Venvandtschaft mit anderen Volkern. An- a small population. Am. Nat. 86:145-160. throp. Anzeiger 4(3):201-207. Gould SJ, and Lewontin RC (1979) The s andrels of San Lack D (1961) Darwin's Finches. New York: Harper Marco and the Panglossian aradigm 1critique of the Torchbooks. adaptationist programme. {roc. R. Soc. London [Biol.] Lande R (1979) Quantitative genetic analysis of multi- 205:581-598. variate evolution, applied to brain:body size allometry. Green R. (1973) Lapita pottery and the origins of Polyne- Evolution 33:402-416. sian culture. Aust. Nat. Hist. June pp. 332-337. Lipiez M (1907) Uber ein Schema zur Bestimmung der Green RC (1986) Some basic components ofthe ancestral Brustform. Korrespondenz-Blatt Deutsch. Gesell- Polynesian settlement system: buildin blocks for schaft Anthropol. Ethnol. Urgeschichte 38(8): more complex Polynesian societies. In P6 Kirch (ed): 175-180. Island Societies: Archaeological Approaches to Evolu- Livingstone FB (1958) Anthropological implications of CRANIOFACIAL PERSPECTIVE ON HUMAN VARIATION 359

sickle cell gene distribution in West Africa. Am. An- predict patterns of brain-size evolution. Science thropol. 60533-562. 229:668-671. Livingstone FB (1962) On the non-existence of human Rothe F (1912) Untersuchungen iiber die Bruste von races. Curr. Anthropol. 3:279. tausend norddeutschen Frauen und Madchen. Luck- Livingstone FB Who gave whom hemoglobin S: The use enwalde: FCE Richten. of restriction site haplotype variation for the interpre- Rushton JP (1985) Differential K theory: The sociobiol- tation of the B-globin gene. Am. J. Hum. Biol. (in ogy of individual and group differences. Person. Indi- press). vid. Differ. 6:441452. McKenna MC (1987) Molecular and morphological anal- Rushton JP (1987) Toward a theory of human multiple of high-level mammalian interrelationships. In C birthing: Sociobiology and r/K reproductive strategies. atterson (ed): Molecules and Morphology. New York: Acta Genet. Med. Gemellol. 36.289-296. Cambridge University Press, pp. 55-93. Rushton JP (1988) Do r/K reproductive strategies apply Marks J (1988) DNA hybridization: holy grail or golden to human differences? SOC.Biol. 35:337-340. calf? Am. J. Phys. Anthropol. 75:245-246. Sarasin F (1924) Sur les relations des Neo-Caledoniens Marks J, Schmid CW, and Sarich VM (1988) DNA hy- avec le groupe de 1’Hornoneanderthalensis. Anthropol- bridization as a guide to phylogeny relations of the ogie 34:193-227. hominoidea. J. Hum. Evol. 17:769-786. Schwidetzk I, and Rosing FW (1984) Vergleichend- Martin R (1928a) Lehrbuch der Anthropologie in Sys- statistiscie Untersuchungen zur Anthropologie der tematischer Darstellung, Vol. 1,Somatologie (2nd Ed). Neuzeit (nach 1500).Homo 35:149. Jena: Gustav Fischer. Schwartz JH (1984) The evolutionary relationships of Martin R (1928b) Lehrbuch der Anthropologie in Sys- man and orang-utans. Nature 308.501605. tematischer Darstellung, Vol. 2, Kraniologie, Osteolo- Sibley CG, and Ah1 uist JE (1983) Phylogen and classi- gie (2nd Ed). Jena: Gustav Fischer. fication of birds %ased on the data of DdA-DNA hy- Martin RD (1983) Human Brain Evolution in an Ecolog- bridization. In RF Johnston (ed):Current Ornithology, ical Context. New York: American Museum of Natural Vol. 1. New York: Plenum Press, pp. 245-252. History. Sibley CG, and Ahlquist JE (1986) Reconstructing bird Martin RD, and Harvey PH (1985) Brain size allometry: phylogeny by comparing DNA’s. Sci. Am. 254(2): ontogeny and phylogeny. In WL Jungers (ed): Size and 82-92. Scaling in Primate Biology. New York: Plenum, Sibley CG, and Ahlquist JE (1987a) Avian phylogeny DD. 21-32. reconstructed from comparisons of the genetic mate- Ma& E (1965) Numerical phenetics and taxonomic the- rial, DNA. In C Patterson (ed): Molecules and Mor- OW.Syst. ZOO^. 14(2):73-97, phology. Cambridge: Cambridge University Press, MerEierCFX (1873) Eloge du sein des femmes (4th Ed). PP. 95-121. Paris: A. Barraud. Sibley CG, and Ahlquis MortonNE, andLalouelJM(1973)Bioassayofkinshipin evidence of hominoid Micronesia. Am. J. Phys. Anthrop. 38:709-719. panded data set. J. M Nee1 JV (1970) Lessons from a “primitive” people. Sci- Smouse PE, and Li W-H (1987) Likelihood analysis of ence 170:815-822. mitochondria1 restriction-cleavage patterns for the Nei M (1983) Genetic polymorphism and the role of human-chimpanzee-gorilla trichotomy. Evolution 41: mutation in evolution. In M Nei and RK Koehn (eds): 1162-1 176. Evolution of Genes and Proteins. Sunderland, Massa- Smouse PE, Spielman RS, and Park MH (1982)Multiple- chusetts: Sinauer Associates, Inc., pp. 165-190. locus allocation ofindividuals to groups as a function of Nei M (1987) Molecular Evolutionary Genetics. New the genetic variation within and differences among York: Columbia University Press. human populations. Am. Nat. 119:445463. OBrien SJ (1987) The ancestry of the giant panda. Sci. Sneath PHA, and Sokal RR(1973) Numerical Taxonomy. Am. 257(5):102-107. The Principles and Practice of Numerical Classifica- Ohno S (1970) Evolution by Gene Duplication. New tion. San Francisco: WH Freeman and Co. York: Springer-Verlag. Sokal RR (1965) Statistical methods in systematics. Biol. Pearce RH, and Barbetti M (1981) A 38,000-year-old Rev. 40:337-391. archaeological site at Upper Swan, Western, Austra- Sokal RR (1966) Numerical taxonomy. Sci. Am. lia. Archaeol. Oceania 16(3):173-178. 215(6):10&116. Pearson K (1926) On the reconstruction of cranial capac- Sokal RR, and Camin JH (1965) The two taxonomies: ity from external measurements. Man 27:46-50. areas of agreement and of conflict. Syst. Zool. 14:17& Pearson RJ (1986) Introduction to The Jomon, in Part 195. 111: Prehistory, culture history and societ In RJ Sokal RR, Camin JH, Rohlf FJ, and Sneath PHA (1965) Pearson. GL Barnes, and KL Hutterer (eds): kindows Numerical taxonomy: Some points of view. Syst. Zool. on the Japanese Past: Studies in Archaeology and 14:237-243. Prehistory. Ann Arbor: Center for Japanese Studies, Sokal RR, Oden NL, and Thomson BA (1988) Genetic pp. 219-221. changes across language boundaries in Europe. Am. J. Ploss H (1872) Die ethnographischen Merkmale der Phys. Anthropol. 76:337-361. Frauenbrust. Arch. Anthropol. 5:215-220. Sokal RR, and Rohlf FJ (1969) Biometry: The Principles Ploss H. and Bartels M (1899) Das Weib in der Natur- and Practice of Statistics in Biological Research. San und Volkerkunde: Anthropologische Studien (6th Ed). Francisco: WH Freeman and Co. Leipzig: Grieben’s Verlag. Sokal RR, and SneathHA(1963)Principles ofNumerical Ploss H. Bartels M. Bartels P. and Dinewall EJ (1938) Taxonomy. San Francisco: WH Freeman and Co. Woman: An Historical G aecolo ‘carand Anthropo- Sokal RR, and Uytterschaut H (1987) Cranial variation logical Compendium. St. Euis: CfMosby. in European populations: A spatial autocorrelation Renfrew C (1987) Archaeology and Language: The Puz- study at three time periods. Am. J. Phys. Anthropol. zle of Indo-European Origins. London: Jonathan Cape. 74:21-38. Riska B, and Atchley WR (1985) Genetics of growth Sokal RR, Uytterschaut H, Rosing F, and Schwidetzky I 360 C.L. BRACE AND K.D. HUNT

(1987) A classification of Euro ean skulls from three Terrell J (1986) Prehistory in the Pacific Islands: A time periods. Am. J. Phys. AntRropol. 74:l-20. Study ofvariation in Language, Customs, and Human SolheimWG (1964)Pottery and the Malayo-Polynesians, Biology. New York: Cambridge University Press. Curr. Anthropol. 5:360,376-384,40C-403. Trinkaus E (1983) The Shanidar Neandertals. New Solheim WG (1972) Prehistoric pottery of Southeast York: Academic Press. Asia. In N Barnard (ed): Early Chinese Art and its Turner CG I1 (1976)Dental evidence on the origins ofthe Possible Influence in the Pacific Basin. New York: Ainu and Japanese. Science 193.911-913. Intercultural Arts Press, pp. 508-532. Turner CG I1 (1983) The origin of the Ainu. Review of H Solheim WG (1976) Reflection on the new data of South- Suzuki and K Hanihara (eds): Minatogawa Man: The east Asian prehistory: Austronesian origin and conse- Upper Pleistocene Man from the Island of Okinawa. quence. Asian Perspect. 18(2):146-160. Quart. Rev. Archaeol. 4(2):11-12. Spielman RS, and Smouse PE (1976) Multivariate clas- Turner CG I1 (1986) Dentochronological se aration esti- sification of human oouulations I. Allocation of mates for Pacific rim populations. Icience 232: Yanomama Indians to biflages. Am. J. Hum. Genet. 1140-1142. 28:317-331. Turner CG I1 (1987) Late Pleistocene and Holocene Spoehr A (1973) Zamboanga and Sulu: An Archeological population history of East Asia based on dental varia- Approach to Ethnic Diversity. Ethnology Monographs tion. Am. J. Phys. Anthropol. 73:305-321. Number One. Pittsbur h Department of Anthropol- Turner CG II (1989) Teeth and prehistory in Asia. Sci. ogy, University of Pittsfurgh. Am. 206(2):88-96. Spuhler JN (1988) Evolution of mitochondrial DNA in Turner CG I1 Origin and affinity of the peo le of Guam: A monkeys, apes, and humans. Yrbk. Phys. Anthropol. dental anthropological assessment. dcronesica (in 31: 15-48. press). Spuhler JN (1989) Evolution of mitochondrial DNA in Verworn M, Bonnet R, and Steinmann G (1919) Der human and other organisms. Am. J. Hum. Biol. diluviale Menschenfund von Obercassel bei Bonn. 1(5):509-528. Wiesbaden: JF Bergmann. Stratz CH (1898) Die Schonheit des weiblichen Korpers. Wa er K (1937) The craniology of the Oceanic races, Stuttgart: Ferdinand Enke. Sgifter Utgitt Av Det Norske Videnskaps-Akademi I Stratz CH (1901) Die Rassenschonheit. Stuttgart: Ferdi- Oslo, I. Matematisk-Naturvidenskapelig Klasse 2: nand Enke. 1-193. Suzuki H (1970)The skull ofthe Amud man. In H Suzuki Wallace DC, Garrison K, and Knowler WC (1985) Dra- and F Takai (eds): The Amud Man and His Cave Site. matic founder effects in Amerindian mitochondrial Tokyo: University of Tokyo Press, pp. 123-206. DNAs. Am. J. Phys. Anthropol. 68:149-155. Suzuki H (1981) Racial history of the Japanese. In I Weiss ML (1987) Nucleic acid evidence bearing on hom- Schwidetzky (ed): Asien I: Japan, Indonesien, Ozean- inoid relationships. Yrbk. Phys. Anthropol. 30:41-73. ien. Munchen: R. Oldenbourg Verlag, pp. 7-69. Witkowski GJ (1907) Les seins a l’eglise. Paris: A. Mal- Suzuki H, and Hanihara K (eds) (1982) The Minatogawa oine. Man: The Upper Pleistocene Man from the Island of Woo-TL, and Morant GM (1934) A biometric study of the Okinawa. The University Museum, The University of “flatness” of the facial skeleton in man. Biometrika Tokyo, Bulletin No. 19. 2fi.196-260.