AMERICAN JOURNAL OF PHYSICAL ANTHROPOLOGY 62425-437 (1983) Climate and the Evolution of Brachycephalization KENNETH L. BEALS, COURTLAND L. SMITH, AND STEPHEN M. DODD Department of Anthropology, Oregon State Uniuersity, Coruallis, Oregon 97331 KEY WORDS Brachycephalization, Cranial index, Cephalic index, Climate, Ecotype ABSTRACT Significant associations of cranial shape for 82 ethnic groups and seven climate variables are described. Variation among current populations is partially attributed to cold adaptation throughout the Pleistocene. Application of data files tabulated by the authors is described for a number of problems. Temporal distribution of 115 specimens indicates a geometric trend (CI = 76.7 - 1.96 log time x lo3). Cranial indices are summarized within alternative taxonomic models and between climatic ecotypes. Evidence supports the hypothesis of cold adaptation among “Classic” Neandertals. Limitations of the thermodynamic model are discussed. It is probable that a decrease of the cranial index occurs from the Middle to Upper Paleolithic. During the Holocene, the index increases under all climatic conditions. Adaptive significance of head shape was first suggested in a classic study by Boas (1911), which indicated that generational change occurred with the cephalic index among the children of European immigrants. Weidenreich (1945) later described the trend of brachycephalization among recent hominids. Suggestions as to the cause have included effects of pedomorphism, population admixture, neurocranium balance, encephalization, stature increase, heterosis, nutritional deprivation, nomadic incursions, and abandonment of cradling (Bielicki and Welon, 1964; Hooton, 1946; Hulse, 1971). In 1955, Coon mentioned that brachycephaly would be advantageous in regions of cold climates. This hypothesis is based upon the effect of surface area:mass ratios that underlie the rules of Allen and Bergmann. Beak (1972) evaluated the climatic adaptation model with a worldwide sample of 339 populations. The climatic associations were highly significant. For example, ethnic groups exposed to winter frost have a mean cephalic index that is 4.3 units higher than those living within the tropics. The conclusions are independently supported by the data of Hiernaux (1968) for Africa and by Crognier (1981)for populations of Europe and the Mediterranean. 0 1983 ALAN R. LISS, INC. Such evidence suggests that a portion of the paleontological trend toward brachycephalization may be explained by the increased occupation of cold environments throughout the Pleistocene. The primary purpose of the current investigation is to evaluate the cold adaptation hypothesis empirically. “Adaptation” is used in the broad sense of any adjustment to ecological factors. MATERIALS AND METHODS For several years, the authors have been tabulating paleontological and distributional evidence relative to surface area:mass ratios. Our ultimate objective is to combine a computerized mapping program with rates of evolutionary change, distance analysis, and climatic correlation to project clinal distributions into the remote past. Development of such a research tool would then permit comparison of newly discovered hominids with trait distributions as expected at any point in time and space. The objective thus requires the statistical data pertaining to both paleontological and contemporary ecological associations. The present study is thus part of a more comprehensive project, and a portion of the Received January 19, 1983; accepted August 30,1983. 426 K.L. BEALS. C.L. SMITH, AND S.M. DODD evidence and discussion within the text originates from analyses of other related problem areas which are in various stages of manuscript completion and with results as yet unpublished. General methodology is first described, followed by a particular application to the brachycephalization problem. The concept is to use computer technology as a kind of “time machine” to construct outline maps, plot trait predictions by latitude and longitude, draw clinal maps for any point in time within any geographical area, compare the predictions to known reference points (the paleontological data already available), and calculate the probability of “correspondence” of a particular specimen to the empiric expectation (given multiple regression with time, sex, location, and climate). Such a tool would be analogous in some respects to x2 observatiodexpectation analysis within genetics-except that the deductive theory (i.e., Hardy-Weinberg Equilibrium, Law of Segregation) does not exist, and the system thus depends upon the empirical data bases. Paleontological data are derived from a reference file with 147 specimens from 75 sites (having a minimum antiquity of 10,000 years) and tabulated from 54 original and secondary sources. The file contains site, specimen number, cranial index, cranial capacity, date, sex, latitude and longitude, climatic association, taxonomic attributions, qualification notes, and references. Data on the cranial index are not available for the total sample of 147. Those that do have such reports (118) are listed in Table 1. Head shape information among ethnic groups derives from 339 populations as originally separated by Beak (1972)into climatic stress zones. This sample has been increased to 362 by addition of group means from temperate regions. A composite of 82 populations (means of same ethnic group from different locations) were tabulated into a climatidanthropometric file with cephalic index and 11 additional anthropometrics. Climatic correlations for each can be calculated for latitude and longitude, climatic zone, solar radiation, total hours of sunshine, winter water vapor pressure, summer water vapor pressure, mean annual precipitation, coldest month minimum mean temperature, warmest month maximum mean temperature, and isothermic zone. (Solar radiation is total heat as expressed in annual kilogram calories per square centimeter of earth surface. Water vapor (millibars) indicates total amount of air moisture. Isothermic zones are taken from Schwidetzky’s [19521 arbitrary scale of isotherms-ranging from 1 [over 30°] to 6 [under 10“1 .) As another portion of the overall project, a mapping program has been devised to plot distributions. We have used it to construct an outline map and illustrate the geographic location of specimens included within the analysis (Fig. 1). It is presently limited to plotting individual points-isophenes must be added manually. The biological and climatic files can be merged with the much larger set of cultural traits computerized from the Human Relations Area Files-which includes as many as 1,170 ethnic groups and 150 variables. Such merger permits evaluation of biological, ecological, and cultural interrelationships. Specifically in regard to the evolution of head shape, there is a research design conflict between the two desirable goals of maximizing both sample size and reliability. The problem is that circumstances such as ontological age, reconstruction, postmortem deformation, uncertain dating, and measurement error cause decisions to be made with respect to inclusion or exclusion of particular cases. There is no completely objective solution, since the reliability is a matter of degree. Reports on different cranial indices from the same specimen are averaged. Dates are taken as the midpoints of estimates. A number of specimens have had recent dating revisions. These include material from Border Cave, Fish Hoek, Ofnet, Saldanha, Ehringsdorf, and Ngandong (Solo). Infants are not included within the calculations. Juveniles are included, since the index changes little from a time shortly following birth. Omo-1, Krapina-C, and Krapina-D are omitted from analysis for reasons of postmortem deformation or known measurement error. Reconstructions are not eliminated, since they should not produce a systematic bias. Evaluation of a variety of possible exclusiodinclusion sets (such as with or without Steinheim and Swanscombe) did not produce differences of sufficient magnitude to negate any general result. Notes within Table 1 are given to indicate briefly the interpretive limitations. It is not possible to correlate the paleontological cases with climatic variables as specific as those which can be associated with CI 67.5 73.0 62.8 72.3 67.0 67.4 75.5 62.4 78.8 74.2 68.8 67.9 68.8 72.0 72.3 71.4 72.6 75.5 66.8 76.0 74.6 78.3 72.0 72.6 74.0 78.0 68.6 67.4 68.3 78.9 83.7 85.5 78.9 66.1 67.3 76.8 78.4 75.5 76.5 76.0 73.6 71.3 76.5 80.0 78.4 Specimen Sterkfonteind Swartkrans-46 Sangiran-4 Koobi Fora 3733 Olduvai-5 Olduvai-9 Sangiran-10 Taung-1 Lantian-2 Sangiran-2 Sangiran-3 Sangiran-17 Trinil-2 Saldanha Choukoutien-3 Choukoutien-10 Choukoutien-12 Ngandong-1 Ngandong-6 Ngandong-7 Ngandong- 10 Ngandong-11 Ngandong- 12 Steinheim Ehringsdorf-H Swanscornbe Omo-1 Omo-2 Laetoli-18 Fontechevade Krapina-C Krapina-D Ganovch Kanjera-1 Kanjera-3 Gibraltar-1 Saccopastore-1 La Ferrassie-1 Le Moustier Monte Circeo Neandertal SPY- 1 SPY-2 Petralona Teshik-Tash 1200 1200 1641 1352 1552 1452 1525 1425 1220 1565 1435 1200 1350 1200 1450 1320 908 800 530 1067 855 440 780 813 900 1004 900 1225 915 1225 1030 1172 1251 1013 1135 1231 1090 1460 1450 1250 485 CC Date 2500 2100 1900 1700 1500 1300 830 800 775 710 710 7 10 650 500 300 300 300 250 250 250 250 250 250 225 220 175 130 130 120 110 85 85 70 70 70 60 60 52 52 52 52 52 52 50 50 F M F F M M M M M F F F M F F M F F F M F M M F M M F F F M M F M M S CZ2 TR TR TR TR TR TR TR TR TM TR TR TR TR TR TM TM TM TR TR TR TR TR TR TM TM TM TR TR TR GL GL GL GL TR TR GL GL GL GL GL GL GL GL GL GL TAX3 N-HE N N AA AR HE HE AR HE HE AA-AR HE HE HE HE HE HE-N HE HE HE N-HE N-HE N-HE N-HE N-HE N-HE N N N N-MM N-MM N N N N N N-MM N-MM N N N N N N N (continued on next page) 026S027E 026S028E 0 0 7 s 11E 004N037E 003S035E 003S035E 007S111E 026S028E 034N109E 007SlllE 007SlllE 007SlllE 007S112E 033S018E 040N115E 040N115E 040N115E 007S112E 007S112E 007S112E 007S112E 007S112E 007S112E 049N009E 051N011E 051NOOOE 005N036E 005N036E 004S034E 046NOOOE 046N016E 046N016E 049N020E 001S035E 001S035E 036N005W 042N013E 045N001E 045N001E 041N013E 051N005E 050N005E 050N005E 040N023E 038N067E Location Notes and sources4 (067) (W80) (B79) Djetis, Holloway Revision (B79) (D65) (075) ER 3733 (B79) “Zinjanthropus” (067) (H72-73)(B79) “Chellean Man” (B79) (H78) Holloway Revision (W80)(573) (075) Adult estimate of CC (H78) (067) (B79) (W80)(A731(H80) (P72) (075) Holloway revision (D65)(075) Juvenile, CC. for adult (K80) (D65) (075) W80) (573) (T81) (D65) (075) (K80) (075) Revised dating (B79)(067) (P72) 6 5 4 ) Adolescent CC +2% (W80)(D65) (075) (W80) (D65) (075) W80) (D65) (075) Holloway revision, Solo-1(K80)(P72) (B79) (075) Holloway revision, s o b 5 (K80)(P72) (B79) (075) Holloway revision, Solo-6 (K80) (P72) (B79) (075) Solo-9 (K80)(P72) (B74)(075) Holloway revision, Solo-10(K80)(P72) (B79) (075) Holloway revision, Solo-11(K80) (P72)(B79) (075) Distorted (P72) (B79)(H51) (W80B) Dating revised (W80)(B79) (H51) (071) Estimated CI (W80)(P72) (B79) (052) (D65) Kibish, uncertain date (K80)(R74) Kibish, uncertain date (K80)(R74) (D80) Specimen number uncertain (K80)(W80) (B79) CI auestioned (W80)(B79)(S80) CI questioned (W80)(B79) (”80) (071) Reconstructed, redated (L70) (W801 (P75) See above (L80) (W80)(P75) (B79) (H51) 6 8 0 ) (W80) (B79) (H51) (S80) (B79) (580) Sex doubtful (B79) (H51) (B79) (H51)(580) (B79) (H51) (S80) (B79) (H51) (B79) (H51) (W80) (B79) (W80B) Adolescent, CC +5% (H51)(W45) “P. Transuaalensis” (067)(P72) (P73) (B79) TABLE 1. Paleontological data of specimens’ Specimen Ingwavuma-1 Shanidar-1 La Quina-H5 La Quina-HI8 D. Irhound-1 D. Irhound-2 Subalyuk Tabun-1 Broken Hill-1 La Chapelle Florisbad-1 Quafzeh-6 Fish Hoek-1 Chatelperron Eyasi-1 G. des Enfants-4 G. des Enfants-5 G des Enfants-6 Skhiil-4 SkhGl-5 Skhiil-9 Mladec-5 Amud-l Predmost-3 Predmost-4 Combe Capelle Cro-Magnon Markina Gora Cap Blanc Staroselye Choukoutien-101 Choukoutien-102 Choukoutien-103 Barma Grande Barma grande Barma grande Brno-l Le Figuer Olduvai-1 Cape Flats Cheddar Gamble’s Cave-4 Gamble’s Cave-5 Laugerie 1230 1600 1500 1380 1300 1740 1580 1250 1440 1590 1715 1375 1580 1554 1520 1590 1568 1600 1271 1280 1600 1420 CC 1450 1600 1345 CI 70.5 76.2 67.6 77.0 73.2 75.1 78.2 77.0 65.9 75.0 75.0 73.7 75.0 85.5 74.3 76.3 68.6 69.3 71.8 74.5 68.1 73.1 72.1 71.3 70.2 65.7 73.8 71.5 76.3 73.1 70.2 69.3 71.3 71.6 76.3 72.2 69.0 74.7 66.0 69 .O 70.4 70.8 73.7 74.9 47 47 45 45 42 42 42 41 40 40 38 37 36 34 34 32 32 32 32 32 32 31 28 26 26 25 22 21 20 20 18 18 18 17 17 17 17 17 17 15 15 15 15 15 Date F M M F M F F F M M F M M M M M M M M M F M F M F M M M M S TR TM GL GL TM TM GL TM TR GL TR TM TR GL TR GL GL GL TM TM TM GL TM GL GL GL GL GL GL GL TM TM TM GL GL GL GL GL TR TR GL TR TR G1 CZ2 TAX3 MM N N N N-MM N-MM N N N-MM N MM-N N-MM MM MM N-MM MM MM MM N-MM N-MM N-MM MM N-MM MM MM MM MM MM MM MM MM MM MM MM MM MM MM MM MM MM MM MM MM MM (contin ued on tiLest page) 027S032E 037N044E 046NOOOE 046NOOOE 032N009W 032N009W 048N021E 033N035E 014S028E 045N002E 029S026E 033N035E 034S019E 044N004E 004S035E 044N008E 044N008E 044N008E 033N035E 033N035E 033N035E 049N017E 033N036E 049N017E 049N017E 045N003E 045N001E 051N039E 045N001E 046N034E 040N115E 040N115E 040N115E 044N008E 044N008E 044N008E 049N017E 045N004E 003S035E 035S018E 051N003W 001S036E 001S036E 045N001W Location Notes and sources4 Border cave, date revised (067) (P72) (P75) (W80)(K70)(B79) (075) Sex doubtful, adult (B79) (H51) Child (B79) (H51) (067) (W80) (P72) (B79) (067) W80) (P72) (B79) Child (7-9 years) (A76) (W80)(B79) (M74) (H51) “Rhodesian Man” (067) (P72) (B74) (B79) (H51)(580) (067) (P72) (B79) (P75) Jebel Kafzeh (W80) (V75) (067XP72 (P75) (P72) Sex doubtful (067) Grimaldi (F78) (NND) (F78) (NND) (F78) (NND) (B79)(P72) (H51)(M74) (075) (H51) (B79) (P72) (M74) (075) (M74)(P72) (H51)(B79) (075) Lautsch, (W80) (B79) (071) Record cranial volume, (W80)(B70)(075) (B79) (NND) (B79) (NND) (p72) (B79) (P72)(B79) (D65) Kostenki (A76) (071) Maedalenian. absolute date uncertain W45) (071) Chiid, CI estimated for adult (A76) Upper cave (W38)(W80)(B79) Upper cave (W80) (879) (W38)(075) Upper cave (W80) (B79)(W38)(075) Grimaldi, date uncertain (NND) Grimaldi, posthumous deformation (K80) (NND) Grimaldi, Mentone (K80) (NND) (W80)(P72) (B79) Child (B80) Low CI due to postmortum distortion (P74) (067) TABLE 1. Paleontological data of specimens (continued)’ b@ N tr 0 U tr m P F v, Lr m x em (Y, 74.9 75.1 74.6 70.0 73.8 72.0 72.6 73.4 80.5 77.7 88.9 86.2 77.0 83.3 78.7 75.7 72.7 76.8 78.9 70.5 73.7 78.2 65.8 67.0 72.4 68.6 77.4 72.5 Laugerie Liu Kwang-1 Obercassel Obercassel Springbok-1 Chancelade Keilor-1 Talgai-1 Ofnet “2.1” Ofnet “2.11” Ofnet “3.1” Ofnet “4.1” Ofnet “5.11” Ofnet “8.1” Ofnet “11.1” Ofnet “13.1” Ofnet “14.1” Ofnet “15.1” Ofnet “18.1” Ofnet “21.1” Ofnet “24.1” Ofnet “25.1” Cohuna Kow Swamp-1 Kow Swamp-5 Kow Swamp-14 Tze Yane-1 Wadjak-T 11 10 10 10 10 10 10 1260 1210 1550 Location 045N001W 024N109E 015N007E 051N007E 025S029E 045N001E 038S145E 027S150E 049NO 10E 049N010E 049N010E 049N010E 049N010E 049N010E 049N010E 049N010E 049N010E 049N010E 049N010E 049N010E 049N010E 049N010E 036S144E 036S144E 036S144E 036S144E 030N105E 008S112E S F F F F M F F F F M F M M M F ~~ F M M M M F ;: y 11 11 11 11 11 11 11 11 11 11 15 15 15 15 15 14 13 12 11 M M M GL GL GL GL GL GL GL GL GL GL G1 GL GL TR TR TR TR TM TR G1 TM GL GL TR GL TR TR GL CZ2 MM MM MM MM MM MM MM MM MM MM MM MM MM MM MM MM MM MM MM MM MM MM MM MM MM MM MM MM TAX3 Notes and sources4 (N79) (NND) (N79) (NND) (N79) (NND) (N79)(NND) (N79)(NND) (P72) (075) (T81) (075) (T81) (075) (T81) (075) (P72) (W58) (075) (P72) (B79) (W45B) (D65) (075) L. Basse (F78) (NND) Liukiang (075) (P72) (F78) (NND) (B79) (F78) (NND) (B79) May be later burial (067) (P72) (P72) (W45B) (075) Adolescent, CC +5%(W45B) Ofnet redated from Mesolithic-Newell-79 (N79) (NND) ID numbers from Neumann catalogue (N79) (NND) (N79) (NND) (N79)(NND) (N79) (NND) TABLE 1. Paleontological data of specimens (continued)’ Date 1500 420 1540 1530 1593 1370 1480 1500 CC ‘The 118 specimens are abstracted from the 147 listed in HOMDAT, which include the cranial index. The specimens are in chronological order. Columns include specimen and site, cranial index (CI), cranial capacity (CC), estimate of date ( X 1,000), presumed sex (S), location, climatic zone (CZ), taxonomic code (TAX), and notes and principle sources. A complete file may be obtained by writing. The authors are grateful for attention to errors or omissions. *Climatic codes: TR, tropical; TM, temperate; GL, glacial. 3Taxonomic codes: AA, A. africanus; AR, A. robustus; HH, H habilis; HE, H. erectus; N, Neandertal, Neandertaloid, archaic H. sapiens; MM, early modern, H. saoiens. 4Reference codes: A73, Aigner and Laughlin, 1973; A76, Alexeyev, 1976; B35, Von Bonin, 1935; B50, Briggs, 1950; B70, Brain, 1970; B79, Brace, Nelson, Korn and Brace, 1979; B80, Billy, 1980; D62, Dart, 1962; D65, Day, 1965; D80, Day, Leakey, and Magori, 1980; F78, Frayer, 1978; H51, Howell, 1951; H72, Holloway, 1972; H73. Hollowav. 1973: H78. Hollowav. 1978: H80. Hollowav. 1980: H81. Hollowav. 1981: H80. Howells. 1980: J66. Jacob. 1966: 573. Jacob. 1973: K70. Kelso. 1970: K80, Kenned;,’ 1980; L70, Leakey, 7970; L72, Leakey, MLngai and Walker, 19?2; L73, Leakey, 1973; L74; Leakey, 1974; L75, Lestrel, 1975; ‘M62, McKern and Kozlik, 1962; M74, Mann and Trinkaus, 1974; NND, Neumann, no date; “79, Newell, 1979; 052, Oakley, 1952; 067, Oakley and Campbell, 1967; 071, Oakley, Campbell and Molleson, 1971; 075, Oakley, Campbell and Molleson, 1975; P72, Phenice and Sauer, 1972; P73, Parenti, 1973; P74, Protsch, 1974; P75, Protsch, 1975; R74, Rightmire, 1974; S54, Singer, 1954; S77, Sigmon, 1977; S80, Smith, 1980; T71, Tobias, 1971; T81, Thorne and Wolpoff, 1981; V49, Vallois, 1949; V75, Vallois and Vandermeersch, 1975; W39, Weidenreich, 1939; W45, Weidenreich, 1945; W45B, Weidenreich, 1945B; W58, Woo, 1958; W71, Wolpoff, 1971; W80, Wolpoff, 1980; W80B, Wolpoff, 1980B. CI Specimen m m 30W Fig. 1. Illustration of computerized mapping (HOMPLOT). Desired scale and region are specified by input of SWNE coordinates. Customized maps and data analysis may be drawn from existing files (HOMDAT = 147 specimens, 8 variables; CLIMDAT = 82 populations, 11 anthropometric and 10 climatic variables; CRANDAT = 122 populations, cranial capacity data; CULMAP = 1,150 cultures, 150 variables. Inquiries should be directed to the 50s authors. The map plots locations of cranial index reports for Pleistocene speciments. Data points may be labeled or numbered. “Choukou” illustrates label character length limitation. Plotting accuracy is limited to nearest degree (note Cape Flats). Any datum associated with points can be plotted (i.e., “35S, 18E”). Isophenes must currently be drawn manually. Options include color differentiation, paper gloss, and grid overlay. U 0 U U r F 431 CLIMATE AND BRACHYCEPHALIZATION living groups. We have classified the specimens as reflecting a general adaptation among their immediate ancestors toward tropical, temperate, or glacial conditions. “Tropical” is considered as the normal absence of winter frost, and includes both dry heat and wet heat. “Temperate” refers to regions having winter frost, but a moderate mean temperature. “Glacial” includes both wet cold and dry cold. The world mean cephalic index for 362 populations is 78.5 units. This value, however, overestimates the actual amount of phylogenetic change, since it is well known that the cephalic index averages 1 to 2 units higher than the cranial index. While it is possible that variation with frontoparietal musculature may be a factor, Krantz (1980) argues that the observed higher values for the cephalic index are the result of geometric effect-removal of approximately equal surface tissue produces a disproportionate length/ breadth ratio with a n oval object. To more accurately compare the paleontological record to contemporary populations, we have reduced the world mean for the cephalic index by 1.5 units, which gives a n average of 77 units for the cranial index. It is desirable to organize the data by taxonomic category as well as time. The ubiquitous problem is a lack of consensus for a number of specimens-largely those that are morphologically intermediate. It is a problem that affects all of the statistical interpretations, since each possible arrangement produces a difference in the sample size, descriptive statistics, probabilities of significance, and rates of change. As a practical matter, we have evaluated the data with two taxonomic models. The first is “gradualistic” in the sense that the attributions correlate almost exactly with time. The second is based upon the most common alternative attribution of disputed specimens. For example, the material from Solo (Ngandong 1, 6, 7, 10, 11, 12) is included as Neandertal in the former model and as Homo erectus in the latter. Details are given in Table 1. Within the text H. erectus, Neandertal, and early modern H. sapiens refer to the gradualistic list. RESULTS Previous work (Beak, 1972) has already demonstrated that head shape has a significant relationship to zones of climatic stress. Weather is a multivariate phenomenon, how- ever, and it may be that some other climatic element could be more appropriately applied to the fossil record. Correlations were therefore calculated between the cephalic index and the nine variables previously described. Results are summarized in Table 2. Climatic zone has the virtue of being the simplest variable to classify throughout the past. It may be noted that it also has the highest correlation with the cephalic index, explaining approximately 21% of the variance itself. In Figure 3, contemporary groups are compared to the fossil record, with a humidity factor (wetnessldryness) distinguished for the former. Differences between wetldry heat and wetldry cold are significant. However, these zones were originally based upon annual precipitation, and such moisture by itself seems to have no effect. The apparent anomaly was investigated by correlating climatic zone with vapor pressure-arbitrarily assigning 15 for dry heat through dry cold, respectively. As cold increases, vapor pressure decreases (r = -0.41 in winter and -0.56 in summer). Annual precipitation and vapor pressure have different meteorological causes. However, by simultaneously classifying temperature and precipitation, climatic zone incidentally correlates with vapor pressure as well. It also may be noted that Schwidetsky’s (1952) scale of isotherms (temperature only) predicts the index virtually as well as climatic zone. It has a correlation of -0.45 compared to -0.39 for coldest month temperature and -0.14 for warmest month temperature. Again, however, the %uccess” of the scale is enhanced by its intercorrelation with TABLE 2. Correlations and probabilities of significant difference between climatic variables and mean cephalic index among ethnic groups‘ Climatic zone Isothermic zone Winter temperature Solar radiation Winter water vapor Hours of sunshine Summer water vapor Summer temperature Annual precipitation r Fr P 0.46 -0.45 -0.39 -0.39 -0.29 -0.22 -0.22 -0.14 0.09 0.09 0.09 0.09 0.10 0.11 0.11 0.001 0.001 0.001 0.001 0.005 0.023 0.021 0.104 0.03 0.16 0.399 0.11 ‘N = 82 composite means (subgroups, i.e., Australians and Eskimo, in different locations averaged). Vapor pressure is in millibars (IO’dynes cm’). Solar radiation is in kcal cm’ earth surface. Variables are elaborated in text. K.L. BEALS, C.L. SMITH, AND S.M. DODD 432 . .-m .-m ? x . a u0 . r a s c W c 0 D 0 d CI 0 Ln 0. 0 0 0 0 0 In 0 . 0 0 0 0 0 e 0 e 0 0 0 0 0 0 Ln 0 *a 0 0 c L .n Ln 0 0 0 In 433 CLIMATE AND BRACHYCEPHALIZATION vapor pressure (r = -0.68 in winter and -0.73 in summer). Our interpretation of cephalic index among living groups is that winter conditions are more important than those of the summer. Temperature is more important than humidity. Temperature and vapor pressure have interactive effects so that general climatic classifications correlate more highly with the index than do the individual variables. Turning to the Pleistocene, the mean hominid cranial index has increased 9 units. Figure 2 plots the distribution through time. The trend is general-not simply confined to the recent past. It is, however, geometric. A log transform of individual dates yields a n intercept of 76.6 at 10,000 B.P., compared to 77 as the observed current mean (CI = 76.7 - 1.96 log time x lo3). The log of absolute date for the 115 samples gives a correlation of -0.28 back through time, with a significance of 0.001. A slow rate of increase is observed while hominids occupied only tropical environments, but it accelerated dramatically with a n adaptation to the rigors of winter. The appearance of hominids in cold is relatively recent. The earliest specimen (for which we have a n estimate of the index) from a nontropical location is Lantian, with a n approximate date of 775,000 B.P. Even this should be considered as a marginal case of mixed broadleaved forest-grassland ecology, and semitropical to temperate climate. If we take 500,000 years of time for adjustment to any substantial amount of winter stress, this is still less than one eighth of hominid existence. If the hypothesis is correct, however, we should expect this relatively short period of time to have a disproportionate effect upon the total amount of change. It has been noted already that the rate of change was indeed extremely slow prior to the occupation of cold environments. Table 3 provides a summary by taxon, independently of time. A comparison of the means with the percentage of samples from temperate and cold environments within the hominid file is shown below. - Taxon A ustralopithecusl H. habilis H. erectus NeandertaV Neandertaloid Early modern H. sapiens x CI Temperate1 Cold 67.5 _+ 2.2 0% 71.2 k 1.0 73.8 k 0.6 35% 70% 73.9 5 0.6 73% The percentages are not meant to be taken as a measurement of the actual geographical TABLE 3. Summary of cranial indices among hominids from climatic zones Gradualist model N X r H. erectus Tropical Temperate Glacial Total Neandertal Tropical Temperate Glacial Total Early modern H. sapiens Tropical Temperate Glacial Total 9 5 70.0 73.5 14 71.2 13 12 16 41 14 5 37 56 - Alternate model N X r 3.9 15 5 1 21 71.5 73.5 80.0 72.4 70.9 73.9 76.2 73.8 4.4 2.6 3.2 4.1 1 5 15 21 68.3 75.6 75.9 75.5 71.4 72.7 75.0 73.9 2.9 3.5 5.i 4.7 - 3.9 3.0 _ 20 12 37 69 4.3 3.0 4.3 2.2 3.1 3.2 70.5 72.6 75.0 73.3 3.2 2.7 5.1 4.7 distribution, since they reflect more intensive field work in certain areas, i.e., Europe. The mean of the index is, however, partially determined by the particular frequency of samples that happen to come from different ecological zones. Among contemporary populations, the large amount of heterographic information on the cephalic index derives from the historical circumstance of the trait being regarded as nonadaptive and hence indicative of racial affinity. In actuality, the trait has become one of the most extensively documented examples of climatic adaptation known. It has no taxonomic (racial) utility beyond the secondary correlation of particular groups (such as Africans and Europeans) being inhabitants of particular climatic regions. A similar circumstance is apparent from the fossil record. Variation between tropical and glacial ecotypes exceeds variation between taxa. For example, the difference between tropical and glacial Neandertals is 5.1 units, whereas the difference between all Neandertals and early modern H. sapiens is only 1.2 units. The hypothesis in question predicts that means within taxa will systematically increase in accordance with decreasing temperature. Including modern populations, there are seven possible trials, and the means are shown below. H. erectus Neandertal Early modern H. sapiens Contemporary ethnic groups Tropical Temperate Glacial 70.0 70.9 71.4 73.3 73.9 72.7 76.2 75.0 75.5 78.9 80.0 - 434 00 K.L. BEALS, C.L. SMITH, AND S.M. DODD I LOWER PLEl STOCENE I H. ERECTUS - - I NEANDERTAL I I EARLY MODERN ETHNOGRAPHIC PRESENT Fig. 3. Means of cranial index for hominids from tropical, temperate, and glacial climates. Ethnographic data is subdivided into humid and dry thermal stress zones (N = 362 group means). General exceptions to the theory do not exist, and the probability of such a n occurrence by accident alone is 0.008 (0.ij7). In Figure 3, the relative difference between taxa is illustrated for the gradualist model. The individual taxonomic means are, of course, dependent upon the specific specimens which are included. The alternative model lists the data (see Table 3) by attribution of disputed forms to a n alternative classification. A primary difference is the number of specimens included as Neandertals (N = 21 rather than 41). It is apparent, however, that the overall pattern of climatic adaptation remains the same; there are again no exceptions. The results cannot be explained as the arbitrary effect of a particular phylogenetic model. It is true that variations in time between ecotypes within taxa influence the difference between means. For example, the cases classified as temperate H. erectus are collectively much younger than those from the tropics. It is also apparent from Figure 3 that other factors in addition to climate are required to explain the evolutionary change. There exists during the Pleistocene a slight increase of cranial index even among tropical ecotypes. It is obvious that a theory that accounts for a n increase as being due to adaptation to winter cold cannot explain such an observation when winter cold is not present. We may nonetheless evaluate the effect of the consequence of climate in addition to whatever the nonclimatic factors are through time. This may be done by comparison of the regression of specimens unexposed to winter frost to each separate observation among those whose immediate ancestors were exposed. Such a procedure also permits the ambiguity with taxonomy to be eliminated since it is simply a test of time and climate only. A sign test further reduces any effect on the CLIMATE AND BRACHYCEPHALIZATION parametric summaries that result from measurement error-existing to some unknown extent in regard to fragmentary or reconstructed remains. The cranial index regression for tropical forms only is -0.70128 log age (in thousands B.P.) f 0.76, with a n intercept of 71.99 at 10,000 B.P. If we take the null hypothesis that the unknown factors associated with time can explain the observations within temperate and cold regions, then we expect a n equal number of signs above and below the regression. Our observations yield 61 cases above the tropical regression, but only 12 below, with a probability less than 0.0001. By comparing each trial to the prediction, it is then possible to adjust for collective temporal difference between ecotypes within taxa. Results are summarized below, in which the values represent the mean additive magnitude of the index when compared to tropical contemporaries. 435 likewise has a known mechanism of selection that will result in differential reproductionnamely death by exposure. Such deaths include any life crisis in which heat radiation is a contributory factor, such as increased physiological stress in cases of traumatic injury. Variations in the cephalic index of modern populations are at least partially a consequence of adaptation to cold during the Pleistocene. Ecotypic differences within taxa exist both dependently and independently of time, and with time independently of taxon. The historically important assumption that head shape was a nonadaptive trait (and hence indicative of racial affinity) is no longer tenable. Furthermore, there is no reason to believe the trait serves any better for taxonomic diagnosis. To illustrate, the difference between Neandertals and H. erectus is 2.6 units (gradualist model) and 3.1 units (alternate model), while the ecotypic means vary as much as 7.6 and 8.5 units. Complicating any taxonomic utility is the Temuerate Glacial substantial “natural” variation. For example, the difference between Skhiil 5 and H. erectus 4.40 Neandertals 4.17 6.94 S k h d 9 is 6.4-equivalent to the average Early modern 1.80 4.70 change between the Lower Pleistocene and H. sapiens Upper Paleolithic-with a time span in excess of 2 million years. At Ofnet, specimens The relative differences between climatic vary by 8.4 units within centimeters in space, regions among early modern H. supiens is and identical time and climate. approximately the same as between current The implication is that if head shape is ethnic groups. Glacial Neandertals, living included among the traits used for taxonomic under arduous winter conditions, show a n assessment, some adjustment for the effect of even higher relative magnitude. climate must be made. It is further suspected Significance of difference between means that any trait that plays a role in surface was calculated by t test, comparing tropical area:mass ratios will require a similar apspecimens to those from both temperate and proach. glacial conditions. Data for H. erectus do folCranial capacity, for instance, is almost low the expected direction; yet, the probabil- universally used for taxonomic comparison ities are not significant (p = 10% for the and assessment. The distribution, however, gradualist model and = 20% for the alter- mirrors the adaptive pattern of the cranial nate). While it is probable that differentia- index. The modern ethnic group data show a tion by climatic ecotype began a t such a n sex-combined increase from 1293 to 1300 to early date, the small sample size precludes a 1362 to 1385 to 1425 cm3 from dry heat definite answer. through dry cold zones, respectively. These Results are significant for Neandertal spec- are ecotypic rather than taxonomic associaimens and early modern H. supiens. The tions, and are of substantial magnitude. Cligradualist model yields probability levels less mate affects size as well as shape; if one than 0.005, the alternate, less than 0.001. compares values from one taxon to another, the ecotypic and taxonomic variation is conDISCUSSION founded. A given mean may, for example, be A partial explanation for the relation be- significantly greater for no reason other than tween climate and the architecture of the containing a higher proportion of specimens human cranium is found in the physical prin- from colder thermodynamic environments. ciples of heat radiation and by the distribuA number of specimens represent populational evidence through time and space. It tions exposed to periglacial conditions. Mul- 436 K.L. BEALS, C.L. SMITH. AND S.M. DODD tivariate analysis among current groups living under comparable stress has been performed, although results are as yet unpublished. Briefly, extreme cold is typified by moderate stature, moderate nasal index, round cranium, large cranial volume, small brain size relative to weight, large brain size relative to stature, and lateral body build. These are examples of traits that should also be expected to show ecotypic variation evolving through the fossil record and also those that are especially indicative of the extent of cold adaptation among fossil forms. Such adaptation has been most widely discussed in reference to Neandertals (e.g., Howell, 1957). Head form data support the independent affirmative conclusions of Trinkaus (1981) with regard to postcranial remains. Glacial Neandertals fit the roundheaded model that is expected under such conditions. It is true that their average cranial index is less than the means of populations of the present living under extreme cold (Fig.3). However, if we make the comparison with specimens that were contemporary, the difference is striking; Glacial and Tropical Neandertals vary more than the average difference between A ustralopithecus and early modern H. sapiens. The climatic model succeeds in two major ways: It explains the general variation tween groups, and it explains why a trend toward brachycephalization occurred throughout the Pleistocene. As is often the case, however, the apparent failure of a theory is more interesting than its success. There are two apparent and perplexing failures. As may be seen in Figure 3, the index seems to decrease between Temperate/Glacia1 Neandertals and early modern H. supiens. Secondly, the index increases even among forms and populations unexposed to winter frost. The first case bears upon the longstanding debate on the “Neandertal problem” (i.e., Howells, 19761, particularly with respect to extinction, migration, or gene flow coincident with the origin of the Upper Paleolithic. We plotted the available information on the assumption that the observed decrease was not significant. Using all cases from Europe and southwest Asia between 15,000 and 75,000 B.P., the overall slope is 0.04971, with a correlation through time of 0.21 and a probability of 0.052. The level of significance is so close to the arbitrary 5% level that we cannot reject the statistical probability of a decrease. There are, nonetheless, multiple interpretations of the evidence. Some of these include a possible effect from the postglacial warming trend, insulation from extreme cold by improved technology, sampling inadequacy, and gene flow (or replacement event) from the tropics. Further work with multivariate analysis may resolve the problem. The second apparent failure concerns the fact that the index increases even when adaptation to cold cannot be a plausible explanation. The extremely slow rate of increase within the tropics for 2 million years could conceivably be some combination of sample size, dating, or measurement error; yet we must accept the rapid increase within the same type of heat stress for the recent past as genuine-as it is based upon thousands of individual observations and dozens of ethnic groups. The basic problem is that the index increases not only under conditions of cold (as expected), it also increases under conditions of severe heat (when unexpected). Our interpretation is that a climatic explanation cannot be the sole adaptive factor. It may not even be the most important. Whatever such additional processes may be, circumstantial evidence indicates that their effect would have to be exponential-with small significance until about 15,000 years ago. Such circumstantial speculation immediately suggests a connection with isolation breakdown. Convincing evidence requires, however, the demonstration that isolation breakdown not only changes the distribution of the variance, but also the mean of the trait over time. ACKNOWLEDGMENTS The authors are indebted to David Frayer and Gerry Brush for assistance in the preparation of the files. Evaluation of data was supported by unsponsored research funds from the Oregon State University Computer Center. Dave Fuhrer and Bob McNaughton aided in the creation of the mapping program. LITERATURE CITED Beals, KL (1972) Head form and climatic stress. Am. J. Phys. Anthropol. 37r85-92. Bielicki, T, and Welon, Z (1964)The operation of natural selection on head form in an East European population. Homo I5r22-30. Boas, F (1911) Changes in Bodily Form of Descendants of Immigrants. Washington: U.S. Government Printing Office. Coon, CS (1955) Some problems of human variability and natural selection in climate and culture. Am. Nat. 89:257-279. CLIMATE AND BRACHYCEPHALIZATION Crognier, E (1981) Climate and anthropometric variations in Europe and the Mediterranean area. Ann. Hum. Biol. 8:99-107. Hiernaux, J (1968) La Diversit6 humaine en Afrique Sub-Saharienne. Bruxelles: Universitk Libre de Bruxelles. Hooton, EA (1946) Up From the Ape. New York: Macmillian. Howell, FC (1957)Evolutionary significance of variation and varieties of Neanderthal man. Quart. Rev. Biol. 37:330-347. Howells, WW (1976)Explaining modern man: Evolution- 437 ist versus migrationist. J. Hum. Evol. 5:477-495. Hulse, FS (1971) The Human Species. 2nd Ed. New York: Random House. Krantz, GS (1980)Climatic races and descent groups. N. Quincy, Massachusetts: Christopher. Schwidetzky, I (1952) Selektions Theorie und Rassenbildung beim Menschen. Experientia 8:85-98. Trinkaus, E (1981) Neanderthal limb proportions and cold adaptation. Aspects Hum. Evol. 21:187-224. Weidenreich, F (1945)The brachycephalization of recent mankind. SW. J. Anthropol. 1:l-54.