AMERICAN JOURNAL OF PHYSICAL ANTHROPOLOGY 84:399405 (1991)
Endocranial Capacity of Western Australian Aboriginal Crania: Comparisons and Association With Stature and Latitude L. FREEDMAN, W.F.C. BLUMER, AND M. LOFGREN Centre for Human Biology and Department of Anatomy and Human Biology, University of Western Australia, Nedlands W A 6009 (L.F., W.F.C.B.) and Department of Anthropology, Western Australian Museum, Perth W A 6000 (M.L.), Australia
KEY WORDS
Australian Aborigines, Allometry, Climate
ABSTRACT The endocranial capacities (ECCs) of 73 Western Australian Aboriginal crania were estimated. Using water-standardked mustard seed, ECCs were (in cm3) males-X 1,239, S.D. 92.3; females-X 1,118, S.D. 77.5. The male and female mean values were smaller than those previously published for Australia as a whole; sexual dimorphism (9.7%)was also slightly lower. Comparison of the Western Australian Aboriginal sample with a large Danish sample (Pakkenberg and Voigt, 1964; Holloway, 1980) permitted analysis of factors underlying sex and population differences in ECC. In both samples about 40% of the mean sex differences in ECC could be related to stature differences; for each sex almost 213 of the differences between the Western Australian and Danish means appear to be associated with differences in stature and latitude. Allometric adjustments are also involved. Most of the studies of the endocranial capacity (ECC) of the Australian Aborigines are based on small samples and the techniques used are often not outlined (e.g., Morant, 1927). In the present study, using a water-standardised, mustard seed technique to measure ECC, we report on a substantial sample of crania from Western Australia and compare our results with those of two earlier studies of Australian Aborigines (Morant, 1927; Wagner, 1937) and also with a large, well-anal sed Danish sample of brain weights (Pak enber and Voigt, 1964; Holloway, 1980). We use t e latter comparison to investigate the basis of sexual and population variations in ECC.
Most of the specimens had been sexed for three previous studies (Mar etts and Freedman, 1977; Freedman anc f Lofgren 1981; Milne et al., 1983) but all were re-assessed (independently by L.F. and M.L.) for the present investigation, using the seven best morphological cranial sexing features described by Larnach and Freedman (1964) and, when present, post-cranial bones (especially the pelvis-Davivongs, 1963). In the original description of the LarnachFreedman techni ue, using a coastal New South Wales samp e sexed by pelvic bones, it was concluded that 90-95% were correctly assessed. On a Murray Valley sam le, sexed by pelvic bones, Brown (1981)foun 92.5%of the female crania and 91.5% of the male MATERIALS AND METHODS crania could be correctly assessed by the The s ecimens measured for this study Larnach-Freedman technique. consiste of 73 Western Australian crania Methods (51male and 22 female) from the collection of It is extraordinaril difficult to get repeatthe De artment of Anthropology, Western Austra ian Museum, Perth. Almost all are able estimates of E C as the irregularity, the result of chance findings of burials and porosity, fissures, and foramina of crania none are believed to be of great antiquity. present difficult challenges (Todd, 1923; For 60 of the specimens specific localities Keen, 1951). The mustard seed, water-stanwere available (Fig. 1). For a few crania, some minor plasticine reconstruction was ReceivedMay 11,1989; accepted September 21,1990. necessary.
I R
B P
0 1991 WILEY-LISS, INC
7
8
8
400
L. F R E E D M ET AL.
WESTERN AUSTRALIA SCALE
I00 290 STATUTE MlLES
o Male
/?
Female
- - - _ ~ ~
material listed above (Fig. 1). Descriptive ECC statistics were calculated, by sex, for each region and for the sample as a whole (Table 1, 2b). The regional samples by sex, and also the total male and female samples, were compared by “t” tests. Sexual dimorphism was also examined by the formula of Tobias (1975): (male mean - female mead male mean) x 100, and also by a histogram (Fig. 2). RESULTS
Table 1 lists the basic statistics for the water-standardised mustard seed ECCs of E the 73 Western Australian crania. The numbers in the regional samples are small, especially in the case of the females. No significant differences at the 5% level were found by “t”tests in either the maIe or the female regional sample comparisons, nor were there any clear trends in the data. Therefore, the samples were pooled for further analysis (Table 2b). 34 Sexual dimorphism was investigated for 1 -.. - -_ the total sample and, using the formula of 114 122 130 Tobias (19751, was 9.7%. Examined by a “t” Fig. 1. Distribution in north (N), central (C) and test, the male-female difference was highly south (S) coastal, and east (E) inland areas of Western significant (“t”= 5.60 with 71 d.f., P < .001). Australia of the male and female crania examined. The male range, 417 cm3, was greater than that of the females, 306 cm3 (Table 2b). Figure 2 is a histogram showing the degree of dardised rocedure which we devised is out- overlap between the samples; 18% of the lined in t e Appendix and gives repeatable females and 45% of the males are in nonresults. In this procedure, the volume deter- overlap regions. mined by mustard seed is reduced by 20 cm3 COMPARISONS AND ANALYSIS in order to make it comparable to waterIt is against a background of two widely estimation techniques. For many of the studies reported in the recognised findings that human ECCs must literature (e.g., Morant, 1927), there is un- be viewed: certainty as to whether water-standardisa1. Measurements of the endocranial cation was performed. Unless the estimation methods are fully described and standard- pacities of a wide variety of populations of ised, comparisons must be treated with cau- Homo sapiens supiens average about 1,420 tion. For example mustard seed estimations cm3 for males and 1,270 cm3 for females of the ECCs of 9 Western Australian crania, (Tobias, 1981). However, there is very conreported by Woodward (1901), were mea- siderable inter- and intra-population variasured in the present study and were found to tion. Using a very large sample, including average 61.2 cm3 larger than by our water- many different populations, Tobias (1971) gives a range for males of 900-2,000 cm3. standardised procedure. Individuals appear to be able to exhibit all of INITIAL DATA ANALYSIS the complex behavioural patterns considProcedures ered characteristic of anatomically modern Because Western Australia covers over 2.5 humans within these very wide limits. 2. It is widely acce ted that endocranial million km’, the material was initially subdivided into four samples: north, central shape, foramina and Essures make estimacoastal, south coastal, and east inland, as tions of ECC difficult and that different techwas done in the three previous studies of this niques may give values which are not compa-
v
f:
401
AUSTRALIAN ABORIGINAL ENDOCRANIAL CAPACITIES
TABLE 1. Endocranial capacities (cm3)of regional samples of Western Australian male and female crania’ Statistic Male N X SD Female N X SD
North
Central
South
East
?
9 1,251.3 88.9
16 1,261.1 93.5
10 1,210.0 87.8
6 1,238.3 80.5
10 1,222.4 110.2
4 1,098.0 91.5
5 1,100.0 114.2
9 1,132.3 71.6
1 1,097
3 1,141.7 18.5
-
‘Data on ECCs, cranial lengths, breadths, and heights on individual crania are available from the Department of Anthropology, Western Australian Museum, Perth.
TABLE 2. Australian Aboriginal male and female endocranial cauacities lcm”i (a) Major previous studies Morant (1927) Male State
N
Northern Australia/ Territory Queensland New South Wales Victoria South Australia Western Australia Tasmania Unknown Total
18
-
X
N
1,224.2
19 38 18 39 20 33 30 215
Wagner (1937) Female
1,287.7 1,287.7 1,311.3 1.319.6 1,255.5 1,264.3 1,310.9 1.286.8
1
1
Male
-
Female
-
X
X
N
10
1,142.5
10
1,261
27
1,141.4
19 12
39
1.149.9
15 22 7
1,277 1,320 1,388 1.278 1,231
25 9 110
1,153.8 1,149.4 1.148.0
98
1.294
N
\j
--
X
11
1.103
11
1.103
(b) Present study State
N
x
Western Australia
51
1.239.1
Male SD 92.3
rable (Todd, 1923; Keen, 1951). The repeat measurements, a t different times and with different observers using our standard procedure, lead us to believe that our method is acceptably reproducible (Appendix A). As noted, our water-standardised technique includes a subtraction of 20 cm3from the mustard seed estimate. Comparisons with data in the literature are, however, hazardous due to varying techniques of estimation. Having become fully cognizant with the difficulties involved, we agree with Wagner (1937)that differences of less than 50 em3between population figures from authors using different or less well standardised methods, cannot be considered significant. Australian comparisons There are scattered records of Australian Aboriginal ECCs which have been estimated
~
Range
N
X
1.467-1.050
22
1.118.5
Female SD 77.5
Range 1.248-942
by a variety of techniques for more than 100 years. In 1927 Morant collected together the early published data. The 18 studies he summarised were made between 1865 and 1918 and included the capacities of 215 males and 110 females (Table %a). He noted that: “. . . all capacities were accepted, although the different methods used to determine them might well have led to substantially different results . . .” (p. 421). The next, and apparently only other, large study was by Wagner (1937). He assessed 98 male and 11 female ECCs from mainland Australia and quotes Morant’s data for the Tasmanians (Table 2a). The mainland sample he measured included previously undescribed material but apparently also some crania included by Morant (1927). Com arisons between the regional samples o Morant (1927) and Wagner (1937) show mean differences as great as 76.7 cm3 and the difference for their Western Austra-
P
402
L. FREEDMAN ET AL 14 12
4
8
Fig. 2. Histogram of male and female Western Australian Aboriginal endocranial capacities (cm3).
lian male samples is 24.5 cm3(Table 2a). The because the extra volume of the menin es, male mean values for the present study of about 50 cm3,is balanced by the amount t at the Western Australian males (Table 2b) is a skull shrinks, 50-60 cm3, when it is dried similar to that of Wagner but 16.4 cm3 less (Todd, 1923). than that of Morant. Compared with the On mean values, the Danish male and Australia-wide totals, the present Western female sample ECCs (above)are each considAustralian male value is about 50 cm3 less eater than those for Western Australia ( able 2bbmales 167.5 cm3, females than those of both Wagner and Morant. The erably female Western Australian mean ECC is 153.6 cm3. The degree of sexual dimorphism is simi29.5 cm3 less than that of Morant’s Australia-widefigure and sexual dimorphism in the lar in the Western Australian and Danish present sample (9.7%)is also lower than that samples, 9.7% and 9.6%, respectively. In a ofMorant’s sample (10.8%). major study of brain evolution, Tobias (1975) has found the mean ECC sexual dimorphism of 67 worldwide populations to be 10.57%, Sex and population variation in ECC with a range from 5.2% for Chinese to 18.4% In order to investigate sexual dimorphism for Singhalese. So sexual dimorphism in and population differences in ECC, the both the Australian and Danish Sam les is Western Australian data were compared just slightly lower than the mean va ue for with those of the large Danish Sam le of 502 Tobias’world range. Brain size and stature: Pakkenberg and males and 165 females (Pakken erg and , 1964), recently fully re-analysed by Voigt (19641, in a linear regression analysis vOi? Holoway (1980). This was the only large of Danish brain weight, body weight, and sample we could locate with which compari- stature, showed that “. . . brain weight desons could be made. As these data_were for pends significantly on height. . .” (p. 303). brain weight in grams (males-X 1,457.2, Re-analysing the Danish data, Holloway S.D. 119.8; females-X 1,317.9, S.D. 109.81, (1980) using partial correlation statistics, to make the data comparable, the weights reported a significant (P < .001) zero-order were converted to volumes using the average correlation (0.35) for males between brain specific gravity of brain tissue, 1.036 (Vo- weight and stature but the female value neida, 1966). Thus the Danish male brain (0.07) was not significant ( P = 0.19); the volume (cm3)is estimated as X 1,406.6, S.D. combined male-female correlation (0.47)was 115.6 and that of the females as X 1,272.1, si ificant (P < .001). %en examining the relationships beS.D. 106.0.These brainvolumes may be used directly for comparison with ECCs of the tween stature and ECC, the question of aldried Western Australian crania. This is lometry needs to be considered. In order to
7-l
F
P
E
403
AUSTRALIAN ABORIGINAL ENDOCRANIAL CAPACITIES
maintain a balance between biomechanical and hysiological requirements with body size c an es, “. . . a different growth rate in different irections for different structures” (Gans, 1974,p. 144)is usually required. If all linear dimensions of a population increased by a similar percentage, a 5% increase in stature and in cranial length, breadth, and height would result in an ECC increase of 15.8%. Using Abbie’s(1976)Australia-wide mean stature data (males 168.4 cm, females 157.2 cm), the male excess of stature over that of the females is 7.1%.If each of the Western Australian cranial dimensions (length, breadth, and height) were greater by a similar percentage (7.1%),the male ECC would be greater by28.2%.The differences in ECCs actually found was 10.6%. When the three basic cranial measurements of the Western Australian Aborigines (Margetts and Freedman, 1977) are examined, it is found that, on mean values, the male excesses over those of the female are length 5.8%(male 186.5 cm, female 176.3 cm), breadth 2.5% (male 131.4 cm, female 128.2 cm), and hei ht 5.3%(male 132.5 cm, female 125.8 cm). hus all three ercentage differences are less than those or stature (7.1%)and the cranial breadth excess is only about one-third. However, in considering these figures in relation to ECC, bone thickness and cranial contours need to be considered, and it becomes clear that breadth, with only relativelythin bone on each side, is most closely related to ECC. The marked development of the Australian Aboriginal browridges, especiallyin the male, greatly adds to the external length measurement. As the height dimension is made from basion, because of the contour shape of the cranial base, the cranial height component of internal volume is exaggerated. Calculations show that a 3.4%increase in each of the three linear cranial measurements would be enough to produce the 10.6% difference in ECC which was found. The allometricreduction in the percentage difference in each of these three dimensions has resulted in a 40 cm3smaller ECC than would have been expected if each dimension was greater by the same ercentage. It would seem that there has een an appropriate adaptive, evolutiona adjustment for the new balance required etween ECC and the greater body size in males. In a comparison of the Western Australian
% %
8
g
t 73
Aboriginal data with Holloway’s (1980) data on Danes, in each case the difference between males and females in mean ECC is about 125 cm3 (Western Australians 120.6 cm3, Danes 134.5 an3)and, in stature (using Abbie’s 1964 data for the former), it is about 10 cm (11.2 cm and 10.0 cm, respectively). Pakkenberg and Voigt (1964) roduced a common (for Danish males and emales) regression slope of 4.56 of brain weight on stature. If, as an ap roximation, this regression, which takes a lometry into account, is applied to those two PO ulations, the 10 cm stature difference coul account for almost 50 cm3(40%)of the ECC difference between the Western Australian and Danish samples. Further, for each sex, the mean ECCs of the Danes were found to be about 160 cm3 lar er than those ofthe Western Australians (Ta le 2b) but, ain using the AustraliaY Iie, on average, Danish wide data of Ab males and females are each also about 5 cm. taller than the Australian Aborigines (Australian Aboriginal males 168.4 cm, females 157.2 cm; Danish males 173.2 cm, females 163.0 cm). If the Pakkenberg-Voigt common regression slope (4.56) is applied to these data, the Western Australian 5 cm lower mean height in both males and females could account for 23 cm3of the difference in ECC. Bruin size and latitude: In addition, latitude has been shown by Beals et al. (19841,to have a significant correlation with ECC; for example, larger ECCs are more often found in very cold regions of the world. Their broad assessment for 82 ethnic groups is that ECC increases by 2.5 cm3 per degree of latitude away from the e uator. The mean latitude for the Western ustralian sample can be taken as 25”s (15”535”S),while for Denmark it is about 56“N. This 31° latitude se aration would lead to an expected 77 cm3 E&C! difference, which is about half of that observed between the male and between the female Western Australian and Danish samples. To ether with the ECC differences ascribed a ove to stature, two-thirds of the mean ECC differences between each sex of the two population samples could be accounted for. The influence of latitude may be due to climate, i.e., to its thermodynamic effects. In the past, for mammals including the Primates, the allometric exponent for bod wei ht-brain weight was generally consii ere to be about two-thirds (e.g., Gould,
P
P
%
1
%
f
B
404
L. FREEDMAN ET AL
1975) and a possible relationship with surface area was proposed. More recently, the exponent has been revised, using a larger number of mammalian species, t o threequarters and an association with basal metabolic rate suggested (Harvey and Bennett, 1983).Thus the Western Australian Aboriginal-Danish comparison, although being based on stature, would seem to suggest that stature and climate could well be associated with ECC through factors such as surface area and metabolic rate. SUMMARY
In this paper we report on the endocranial capacities (ECCs) of a series of male and female Western Australian Aboriginal crania. Sexual dimorphism is about 9.7%. We found no significant differences between local samples, and the mean male and female values for the whole sample appear to be slightly lower than those for the rest of Australia. A large, well-analysed sample of Danish brains (Pakkenberg and Voigt, 1964; Holloway, 1980)was used in combination with the Western Australian Aboriginal data to analyse sex and population differences in ECC. Differences in mean stature and in latitude (the latter reflecting the thermodynamic effects of climate), plus allometric adjustments, appear to account for 40% of the sex differences and two-thirds of the population differences. ACKNOWLEDGMENTS
We thank Dr. Matt Cartmill and three anonymous readers for some most useful comments on a previous version of this paper. Ms. Rita Bonjour patiently typed a number of drafts of this paper and Mr. Martin Thompson drew the diagrams. LITERATURE CITED Abbie AA (1976) Studies in Physical Anthropology, Vol. 11. Canberra: Australian Institute of Aboriginal Studies, p. 77. Beals KL, Smith CL, and Dodd SM (1984) Brain size, cranial morphology, climate and time machines. Curr. Anthropol. 25:301-330. Brown P (1981) Sex determination of Aboriginal crania from the Murray River Valley: A re-assessment of the Larnach and Freedman technique. Archaeol. Oceania 16:53-63. Davivongs V (1963) The pelvic girdle of the Australian Aborigine. Sex differences and sex determination. Am. J. Phys. Anthropol. 21:443-455. Freedman L, and Lofgren M (1981) Odontometrics of Western Australian Aborigines. Archaeol. Oceania 16:87-93.
Gans C (1974)Biomechanics. An Approach to Vertebrate Biology. Ann Arbor: University of Michigan Press, p. 144. Gould SJ (1975) Allometry in Primates, with emphasis on scaling and the evolution of the brain. In F Szalay (ed.): Approaches to Primate Biology. Contributions to Primatology Basel: Karger, Vol. 5, pp. 244-292. Harvey PH, and Bennett PM (1983) Brain size, energetics, ecology and life history patterns. Nature 306:314, 315. Holloway RL (1980) Within-species brain-body weight variability: A reexamination of the Danish data and other primate species. Am. J. Phys. Anthropol. 53: 109-121. Keen JA (1951) Standardization of the technique of cranial capacity determination. S. Afr. J. Clin. Sci. 2:170-182. Larnach S, and Freedman L (1964) Sex determination of Aboriginal crania from coastal New South Wales, Australia. Records Aust. Mus. 26:295-310. Margetts BM, Freedman L (1977) Morphometrics of Western Australian Aboriginal Skulls. Records West. MUS.6.63-105. Milne N. Schmitt LH. and Freedman L (1983) Discrete trait variation in ' Western Australian Aboriginal Skulls. J. Hum. Evol. 12:157-168. Morant GM (1927)A studv of the Australian and Tasmanian skulls based on "previously studied measurements. Biometrika 19:417-440. Pakkenberg H, and Voigt J (1964) Brain weight of the Danes. Acta Anat. (Basel) 5:297-307. Tobias PV (1971) The distribution of cranial ca acity values among living hominoids. Proc. 3rd int. 8ongr. Primat., Zurich 1970, vol. 1, pp. 18-35. Tobias PV (1975) Brain evolution in the Hominoidea. In RH Tuttle (ed.): Primate Functional Morphology and Evolution. The Hague: Mouton Publishers, pp. 353392. Tobias PV (1981) Evolution of the Human Brain, Intellect and Spirit. South Australia: University of Adelaide, p. 19. Todd TW (1923) Cranial capacity and linear dimensions in White andNegro. Am. J. Phys. Anthropol. 6:97-194. Voneida TJ (1966) Central nervous system. In JAAnson (ed.): Morris' Human Anatomy, 12th Ed. New York: McGraw Hill Book Company, p. 928. Wagner K (1937) The Craniology of the Oceanic Races. Norske Videnskaps-akademi i Oslo. I Mat.-Naturv. Klasse No. 2. Woodward BH (1901) Catalogue of skulls of Natives of Western Australia in the Museum. Perth, W.A.: Western Australian Museum. APPENDIX A
In the present study we used mustard seeds to assess ECC. The crania were preared by blocking all of the externally visible oramina and fissures through which seeds might be lost when in the inverted position. The seeds were kept under stable conditions for the duration of the study. In filling the endocranium, the inverted cranium with its anterior end tilted slightly upward was placed on a stand. Five hundred cubic centimeters of mustard seed were poured into a plastic funnel held with a
P
405
AUSTRALIAN ABORIGINAL ENDOCRANIAL CAPACITIES
finger blocking its outlet. The spout of the funnel was placed just inside the foramen magnum, and the seeds allowed to drop into the cranium under their own weight. The cranium and its stand were then sharp1 jerked left, right, forward, and backward: each movement being done twice. The funnel was refilled with seed and the procedure was repeated. Using a beaker, the cranial cavity was then filled to the internal lip of the foramen magnum, and the jerking was repeated, this time more gently. Any space that appeared after this procedure was refilled. Because of the contour of the bone just inside the foramen magnum, a heap of seeds commonly accumulated below the foramen and prevented flow sideways. This heap was eased out with a finger just before a final filling all the way up to the external lip of the foramen magnum. For volume assessment, the mustard seeds were poured from the cranium through the same funnel into a 2 litre measuring cylinder. They were gently agitated and tamped to a level meniscus before reading. Reliability of the technique was tested by inter- and intra-observer error experiments. A typical result of an intra-observer error test for 10 successjve estimations (cm3) of one cranium was X 1,320, S.D. 7.82, range
1,310-1,330. For the study, all assessments were made by the same observer (W.F.C.B.) and each cranium was measured three times. If a range greater than 15 cm3 was found, testing was continued until that degree of consistency was obtained for three successive tests. The technique was standardised using water on a specially prepared cranium of which the calotte was initially removable. All relevant foramina, fissures, and sutures were internally and externally sealed with plasticine. The whole internal surface was then coated with rubber latex (to ensure waterproofing), followed by a coat of pol urethane (to prevent the mustard see s adhering). The calotte was then re-attached using “Bedacryl”.ECC was estimated using mustard seed and then water, five times each. A typical result ofsne such test was (in cm2 mustard seed -X 1,304, S.D. 9.6; water-X 1,284, S.D. 2.2. This, and subsequent tests, showed that our mustard seed technique, compared to water, overestimated ECC by about 20 cm3. Variations in seed compacting in either the cranium, or in the measuring cylinder, can significantly vary the volume. Hence all our data are given as water-standardised values.
B
Endocranial Capacity of Western Australian Aboriginal Crania: Comparisons and Association With Stature and Latitude L. FREEDMAN, W.F.C. BLUMER, AND M. LOFGREN Centre for Human Biology and Department of Anatomy and Human Biology, University of Western Australia, Nedlands W A 6009 (L.F., W.F.C.B.) and Department of Anthropology, Western Australian Museum, Perth W A 6000 (M.L.), Australia
KEY WORDS
Australian Aborigines, Allometry, Climate
ABSTRACT The endocranial capacities (ECCs) of 73 Western Australian Aboriginal crania were estimated. Using water-standardked mustard seed, ECCs were (in cm3) males-X 1,239, S.D. 92.3; females-X 1,118, S.D. 77.5. The male and female mean values were smaller than those previously published for Australia as a whole; sexual dimorphism (9.7%)was also slightly lower. Comparison of the Western Australian Aboriginal sample with a large Danish sample (Pakkenberg and Voigt, 1964; Holloway, 1980) permitted analysis of factors underlying sex and population differences in ECC. In both samples about 40% of the mean sex differences in ECC could be related to stature differences; for each sex almost 213 of the differences between the Western Australian and Danish means appear to be associated with differences in stature and latitude. Allometric adjustments are also involved. Most of the studies of the endocranial capacity (ECC) of the Australian Aborigines are based on small samples and the techniques used are often not outlined (e.g., Morant, 1927). In the present study, using a water-standardised, mustard seed technique to measure ECC, we report on a substantial sample of crania from Western Australia and compare our results with those of two earlier studies of Australian Aborigines (Morant, 1927; Wagner, 1937) and also with a large, well-anal sed Danish sample of brain weights (Pak enber and Voigt, 1964; Holloway, 1980). We use t e latter comparison to investigate the basis of sexual and population variations in ECC.
Most of the specimens had been sexed for three previous studies (Mar etts and Freedman, 1977; Freedman anc f Lofgren 1981; Milne et al., 1983) but all were re-assessed (independently by L.F. and M.L.) for the present investigation, using the seven best morphological cranial sexing features described by Larnach and Freedman (1964) and, when present, post-cranial bones (especially the pelvis-Davivongs, 1963). In the original description of the LarnachFreedman techni ue, using a coastal New South Wales samp e sexed by pelvic bones, it was concluded that 90-95% were correctly assessed. On a Murray Valley sam le, sexed by pelvic bones, Brown (1981)foun 92.5%of the female crania and 91.5% of the male MATERIALS AND METHODS crania could be correctly assessed by the The s ecimens measured for this study Larnach-Freedman technique. consiste of 73 Western Australian crania Methods (51male and 22 female) from the collection of It is extraordinaril difficult to get repeatthe De artment of Anthropology, Western Austra ian Museum, Perth. Almost all are able estimates of E C as the irregularity, the result of chance findings of burials and porosity, fissures, and foramina of crania none are believed to be of great antiquity. present difficult challenges (Todd, 1923; For 60 of the specimens specific localities Keen, 1951). The mustard seed, water-stanwere available (Fig. 1). For a few crania, some minor plasticine reconstruction was ReceivedMay 11,1989; accepted September 21,1990. necessary.
I R
B P
0 1991 WILEY-LISS, INC
7
8
8
400
L. F R E E D M ET AL.
WESTERN AUSTRALIA SCALE
I00 290 STATUTE MlLES
o Male
/?
Female
- - - _ ~ ~
material listed above (Fig. 1). Descriptive ECC statistics were calculated, by sex, for each region and for the sample as a whole (Table 1, 2b). The regional samples by sex, and also the total male and female samples, were compared by “t” tests. Sexual dimorphism was also examined by the formula of Tobias (1975): (male mean - female mead male mean) x 100, and also by a histogram (Fig. 2). RESULTS
Table 1 lists the basic statistics for the water-standardised mustard seed ECCs of E the 73 Western Australian crania. The numbers in the regional samples are small, especially in the case of the females. No significant differences at the 5% level were found by “t”tests in either the maIe or the female regional sample comparisons, nor were there any clear trends in the data. Therefore, the samples were pooled for further analysis (Table 2b). 34 Sexual dimorphism was investigated for 1 -.. - -_ the total sample and, using the formula of 114 122 130 Tobias (19751, was 9.7%. Examined by a “t” Fig. 1. Distribution in north (N), central (C) and test, the male-female difference was highly south (S) coastal, and east (E) inland areas of Western significant (“t”= 5.60 with 71 d.f., P < .001). Australia of the male and female crania examined. The male range, 417 cm3, was greater than that of the females, 306 cm3 (Table 2b). Figure 2 is a histogram showing the degree of dardised rocedure which we devised is out- overlap between the samples; 18% of the lined in t e Appendix and gives repeatable females and 45% of the males are in nonresults. In this procedure, the volume deter- overlap regions. mined by mustard seed is reduced by 20 cm3 COMPARISONS AND ANALYSIS in order to make it comparable to waterIt is against a background of two widely estimation techniques. For many of the studies reported in the recognised findings that human ECCs must literature (e.g., Morant, 1927), there is un- be viewed: certainty as to whether water-standardisa1. Measurements of the endocranial cation was performed. Unless the estimation methods are fully described and standard- pacities of a wide variety of populations of ised, comparisons must be treated with cau- Homo sapiens supiens average about 1,420 tion. For example mustard seed estimations cm3 for males and 1,270 cm3 for females of the ECCs of 9 Western Australian crania, (Tobias, 1981). However, there is very conreported by Woodward (1901), were mea- siderable inter- and intra-population variasured in the present study and were found to tion. Using a very large sample, including average 61.2 cm3 larger than by our water- many different populations, Tobias (1971) gives a range for males of 900-2,000 cm3. standardised procedure. Individuals appear to be able to exhibit all of INITIAL DATA ANALYSIS the complex behavioural patterns considProcedures ered characteristic of anatomically modern Because Western Australia covers over 2.5 humans within these very wide limits. 2. It is widely acce ted that endocranial million km’, the material was initially subdivided into four samples: north, central shape, foramina and Essures make estimacoastal, south coastal, and east inland, as tions of ECC difficult and that different techwas done in the three previous studies of this niques may give values which are not compa-
v
f:
401
AUSTRALIAN ABORIGINAL ENDOCRANIAL CAPACITIES
TABLE 1. Endocranial capacities (cm3)of regional samples of Western Australian male and female crania’ Statistic Male N X SD Female N X SD
North
Central
South
East
?
9 1,251.3 88.9
16 1,261.1 93.5
10 1,210.0 87.8
6 1,238.3 80.5
10 1,222.4 110.2
4 1,098.0 91.5
5 1,100.0 114.2
9 1,132.3 71.6
1 1,097
3 1,141.7 18.5
-
‘Data on ECCs, cranial lengths, breadths, and heights on individual crania are available from the Department of Anthropology, Western Australian Museum, Perth.
TABLE 2. Australian Aboriginal male and female endocranial cauacities lcm”i (a) Major previous studies Morant (1927) Male State
N
Northern Australia/ Territory Queensland New South Wales Victoria South Australia Western Australia Tasmania Unknown Total
18
-
X
N
1,224.2
19 38 18 39 20 33 30 215
Wagner (1937) Female
1,287.7 1,287.7 1,311.3 1.319.6 1,255.5 1,264.3 1,310.9 1.286.8
1
1
Male
-
Female
-
X
X
N
10
1,142.5
10
1,261
27
1,141.4
19 12
39
1.149.9
15 22 7
1,277 1,320 1,388 1.278 1,231
25 9 110
1,153.8 1,149.4 1.148.0
98
1.294
N
\j
--
X
11
1.103
11
1.103
(b) Present study State
N
x
Western Australia
51
1.239.1
Male SD 92.3
rable (Todd, 1923; Keen, 1951). The repeat measurements, a t different times and with different observers using our standard procedure, lead us to believe that our method is acceptably reproducible (Appendix A). As noted, our water-standardised technique includes a subtraction of 20 cm3from the mustard seed estimate. Comparisons with data in the literature are, however, hazardous due to varying techniques of estimation. Having become fully cognizant with the difficulties involved, we agree with Wagner (1937)that differences of less than 50 em3between population figures from authors using different or less well standardised methods, cannot be considered significant. Australian comparisons There are scattered records of Australian Aboriginal ECCs which have been estimated
~
Range
N
X
1.467-1.050
22
1.118.5
Female SD 77.5
Range 1.248-942
by a variety of techniques for more than 100 years. In 1927 Morant collected together the early published data. The 18 studies he summarised were made between 1865 and 1918 and included the capacities of 215 males and 110 females (Table %a). He noted that: “. . . all capacities were accepted, although the different methods used to determine them might well have led to substantially different results . . .” (p. 421). The next, and apparently only other, large study was by Wagner (1937). He assessed 98 male and 11 female ECCs from mainland Australia and quotes Morant’s data for the Tasmanians (Table 2a). The mainland sample he measured included previously undescribed material but apparently also some crania included by Morant (1927). Com arisons between the regional samples o Morant (1927) and Wagner (1937) show mean differences as great as 76.7 cm3 and the difference for their Western Austra-
P
402
L. FREEDMAN ET AL 14 12
4
8
Fig. 2. Histogram of male and female Western Australian Aboriginal endocranial capacities (cm3).
lian male samples is 24.5 cm3(Table 2a). The because the extra volume of the menin es, male mean values for the present study of about 50 cm3,is balanced by the amount t at the Western Australian males (Table 2b) is a skull shrinks, 50-60 cm3, when it is dried similar to that of Wagner but 16.4 cm3 less (Todd, 1923). than that of Morant. Compared with the On mean values, the Danish male and Australia-wide totals, the present Western female sample ECCs (above)are each considAustralian male value is about 50 cm3 less eater than those for Western Australia ( able 2bbmales 167.5 cm3, females than those of both Wagner and Morant. The erably female Western Australian mean ECC is 153.6 cm3. The degree of sexual dimorphism is simi29.5 cm3 less than that of Morant’s Australia-widefigure and sexual dimorphism in the lar in the Western Australian and Danish present sample (9.7%)is also lower than that samples, 9.7% and 9.6%, respectively. In a ofMorant’s sample (10.8%). major study of brain evolution, Tobias (1975) has found the mean ECC sexual dimorphism of 67 worldwide populations to be 10.57%, Sex and population variation in ECC with a range from 5.2% for Chinese to 18.4% In order to investigate sexual dimorphism for Singhalese. So sexual dimorphism in and population differences in ECC, the both the Australian and Danish Sam les is Western Australian data were compared just slightly lower than the mean va ue for with those of the large Danish Sam le of 502 Tobias’world range. Brain size and stature: Pakkenberg and males and 165 females (Pakken erg and , 1964), recently fully re-analysed by Voigt (19641, in a linear regression analysis vOi? Holoway (1980). This was the only large of Danish brain weight, body weight, and sample we could locate with which compari- stature, showed that “. . . brain weight desons could be made. As these data_were for pends significantly on height. . .” (p. 303). brain weight in grams (males-X 1,457.2, Re-analysing the Danish data, Holloway S.D. 119.8; females-X 1,317.9, S.D. 109.81, (1980) using partial correlation statistics, to make the data comparable, the weights reported a significant (P < .001) zero-order were converted to volumes using the average correlation (0.35) for males between brain specific gravity of brain tissue, 1.036 (Vo- weight and stature but the female value neida, 1966). Thus the Danish male brain (0.07) was not significant ( P = 0.19); the volume (cm3)is estimated as X 1,406.6, S.D. combined male-female correlation (0.47)was 115.6 and that of the females as X 1,272.1, si ificant (P < .001). %en examining the relationships beS.D. 106.0.These brainvolumes may be used directly for comparison with ECCs of the tween stature and ECC, the question of aldried Western Australian crania. This is lometry needs to be considered. In order to
7-l
F
P
E
403
AUSTRALIAN ABORIGINAL ENDOCRANIAL CAPACITIES
maintain a balance between biomechanical and hysiological requirements with body size c an es, “. . . a different growth rate in different irections for different structures” (Gans, 1974,p. 144)is usually required. If all linear dimensions of a population increased by a similar percentage, a 5% increase in stature and in cranial length, breadth, and height would result in an ECC increase of 15.8%. Using Abbie’s(1976)Australia-wide mean stature data (males 168.4 cm, females 157.2 cm), the male excess of stature over that of the females is 7.1%.If each of the Western Australian cranial dimensions (length, breadth, and height) were greater by a similar percentage (7.1%),the male ECC would be greater by28.2%.The differences in ECCs actually found was 10.6%. When the three basic cranial measurements of the Western Australian Aborigines (Margetts and Freedman, 1977) are examined, it is found that, on mean values, the male excesses over those of the female are length 5.8%(male 186.5 cm, female 176.3 cm), breadth 2.5% (male 131.4 cm, female 128.2 cm), and hei ht 5.3%(male 132.5 cm, female 125.8 cm). hus all three ercentage differences are less than those or stature (7.1%)and the cranial breadth excess is only about one-third. However, in considering these figures in relation to ECC, bone thickness and cranial contours need to be considered, and it becomes clear that breadth, with only relativelythin bone on each side, is most closely related to ECC. The marked development of the Australian Aboriginal browridges, especiallyin the male, greatly adds to the external length measurement. As the height dimension is made from basion, because of the contour shape of the cranial base, the cranial height component of internal volume is exaggerated. Calculations show that a 3.4%increase in each of the three linear cranial measurements would be enough to produce the 10.6% difference in ECC which was found. The allometricreduction in the percentage difference in each of these three dimensions has resulted in a 40 cm3smaller ECC than would have been expected if each dimension was greater by the same ercentage. It would seem that there has een an appropriate adaptive, evolutiona adjustment for the new balance required etween ECC and the greater body size in males. In a comparison of the Western Australian
% %
8
g
t 73
Aboriginal data with Holloway’s (1980) data on Danes, in each case the difference between males and females in mean ECC is about 125 cm3 (Western Australians 120.6 cm3, Danes 134.5 an3)and, in stature (using Abbie’s 1964 data for the former), it is about 10 cm (11.2 cm and 10.0 cm, respectively). Pakkenberg and Voigt (1964) roduced a common (for Danish males and emales) regression slope of 4.56 of brain weight on stature. If, as an ap roximation, this regression, which takes a lometry into account, is applied to those two PO ulations, the 10 cm stature difference coul account for almost 50 cm3(40%)of the ECC difference between the Western Australian and Danish samples. Further, for each sex, the mean ECCs of the Danes were found to be about 160 cm3 lar er than those ofthe Western Australians (Ta le 2b) but, ain using the AustraliaY Iie, on average, Danish wide data of Ab males and females are each also about 5 cm. taller than the Australian Aborigines (Australian Aboriginal males 168.4 cm, females 157.2 cm; Danish males 173.2 cm, females 163.0 cm). If the Pakkenberg-Voigt common regression slope (4.56) is applied to these data, the Western Australian 5 cm lower mean height in both males and females could account for 23 cm3of the difference in ECC. Bruin size and latitude: In addition, latitude has been shown by Beals et al. (19841,to have a significant correlation with ECC; for example, larger ECCs are more often found in very cold regions of the world. Their broad assessment for 82 ethnic groups is that ECC increases by 2.5 cm3 per degree of latitude away from the e uator. The mean latitude for the Western ustralian sample can be taken as 25”s (15”535”S),while for Denmark it is about 56“N. This 31° latitude se aration would lead to an expected 77 cm3 E&C! difference, which is about half of that observed between the male and between the female Western Australian and Danish samples. To ether with the ECC differences ascribed a ove to stature, two-thirds of the mean ECC differences between each sex of the two population samples could be accounted for. The influence of latitude may be due to climate, i.e., to its thermodynamic effects. In the past, for mammals including the Primates, the allometric exponent for bod wei ht-brain weight was generally consii ere to be about two-thirds (e.g., Gould,
P
P
%
1
%
f
B
404
L. FREEDMAN ET AL
1975) and a possible relationship with surface area was proposed. More recently, the exponent has been revised, using a larger number of mammalian species, t o threequarters and an association with basal metabolic rate suggested (Harvey and Bennett, 1983).Thus the Western Australian Aboriginal-Danish comparison, although being based on stature, would seem to suggest that stature and climate could well be associated with ECC through factors such as surface area and metabolic rate. SUMMARY
In this paper we report on the endocranial capacities (ECCs) of a series of male and female Western Australian Aboriginal crania. Sexual dimorphism is about 9.7%. We found no significant differences between local samples, and the mean male and female values for the whole sample appear to be slightly lower than those for the rest of Australia. A large, well-analysed sample of Danish brains (Pakkenberg and Voigt, 1964; Holloway, 1980)was used in combination with the Western Australian Aboriginal data to analyse sex and population differences in ECC. Differences in mean stature and in latitude (the latter reflecting the thermodynamic effects of climate), plus allometric adjustments, appear to account for 40% of the sex differences and two-thirds of the population differences. ACKNOWLEDGMENTS
We thank Dr. Matt Cartmill and three anonymous readers for some most useful comments on a previous version of this paper. Ms. Rita Bonjour patiently typed a number of drafts of this paper and Mr. Martin Thompson drew the diagrams. LITERATURE CITED Abbie AA (1976) Studies in Physical Anthropology, Vol. 11. Canberra: Australian Institute of Aboriginal Studies, p. 77. Beals KL, Smith CL, and Dodd SM (1984) Brain size, cranial morphology, climate and time machines. Curr. Anthropol. 25:301-330. Brown P (1981) Sex determination of Aboriginal crania from the Murray River Valley: A re-assessment of the Larnach and Freedman technique. Archaeol. Oceania 16:53-63. Davivongs V (1963) The pelvic girdle of the Australian Aborigine. Sex differences and sex determination. Am. J. Phys. Anthropol. 21:443-455. Freedman L, and Lofgren M (1981) Odontometrics of Western Australian Aborigines. Archaeol. Oceania 16:87-93.
Gans C (1974)Biomechanics. An Approach to Vertebrate Biology. Ann Arbor: University of Michigan Press, p. 144. Gould SJ (1975) Allometry in Primates, with emphasis on scaling and the evolution of the brain. In F Szalay (ed.): Approaches to Primate Biology. Contributions to Primatology Basel: Karger, Vol. 5, pp. 244-292. Harvey PH, and Bennett PM (1983) Brain size, energetics, ecology and life history patterns. Nature 306:314, 315. Holloway RL (1980) Within-species brain-body weight variability: A reexamination of the Danish data and other primate species. Am. J. Phys. Anthropol. 53: 109-121. Keen JA (1951) Standardization of the technique of cranial capacity determination. S. Afr. J. Clin. Sci. 2:170-182. Larnach S, and Freedman L (1964) Sex determination of Aboriginal crania from coastal New South Wales, Australia. Records Aust. Mus. 26:295-310. Margetts BM, Freedman L (1977) Morphometrics of Western Australian Aboriginal Skulls. Records West. MUS.6.63-105. Milne N. Schmitt LH. and Freedman L (1983) Discrete trait variation in ' Western Australian Aboriginal Skulls. J. Hum. Evol. 12:157-168. Morant GM (1927)A studv of the Australian and Tasmanian skulls based on "previously studied measurements. Biometrika 19:417-440. Pakkenberg H, and Voigt J (1964) Brain weight of the Danes. Acta Anat. (Basel) 5:297-307. Tobias PV (1971) The distribution of cranial ca acity values among living hominoids. Proc. 3rd int. 8ongr. Primat., Zurich 1970, vol. 1, pp. 18-35. Tobias PV (1975) Brain evolution in the Hominoidea. In RH Tuttle (ed.): Primate Functional Morphology and Evolution. The Hague: Mouton Publishers, pp. 353392. Tobias PV (1981) Evolution of the Human Brain, Intellect and Spirit. South Australia: University of Adelaide, p. 19. Todd TW (1923) Cranial capacity and linear dimensions in White andNegro. Am. J. Phys. Anthropol. 6:97-194. Voneida TJ (1966) Central nervous system. In JAAnson (ed.): Morris' Human Anatomy, 12th Ed. New York: McGraw Hill Book Company, p. 928. Wagner K (1937) The Craniology of the Oceanic Races. Norske Videnskaps-akademi i Oslo. I Mat.-Naturv. Klasse No. 2. Woodward BH (1901) Catalogue of skulls of Natives of Western Australia in the Museum. Perth, W.A.: Western Australian Museum. APPENDIX A
In the present study we used mustard seeds to assess ECC. The crania were preared by blocking all of the externally visible oramina and fissures through which seeds might be lost when in the inverted position. The seeds were kept under stable conditions for the duration of the study. In filling the endocranium, the inverted cranium with its anterior end tilted slightly upward was placed on a stand. Five hundred cubic centimeters of mustard seed were poured into a plastic funnel held with a
P
405
AUSTRALIAN ABORIGINAL ENDOCRANIAL CAPACITIES
finger blocking its outlet. The spout of the funnel was placed just inside the foramen magnum, and the seeds allowed to drop into the cranium under their own weight. The cranium and its stand were then sharp1 jerked left, right, forward, and backward: each movement being done twice. The funnel was refilled with seed and the procedure was repeated. Using a beaker, the cranial cavity was then filled to the internal lip of the foramen magnum, and the jerking was repeated, this time more gently. Any space that appeared after this procedure was refilled. Because of the contour of the bone just inside the foramen magnum, a heap of seeds commonly accumulated below the foramen and prevented flow sideways. This heap was eased out with a finger just before a final filling all the way up to the external lip of the foramen magnum. For volume assessment, the mustard seeds were poured from the cranium through the same funnel into a 2 litre measuring cylinder. They were gently agitated and tamped to a level meniscus before reading. Reliability of the technique was tested by inter- and intra-observer error experiments. A typical result of an intra-observer error test for 10 successjve estimations (cm3) of one cranium was X 1,320, S.D. 7.82, range
1,310-1,330. For the study, all assessments were made by the same observer (W.F.C.B.) and each cranium was measured three times. If a range greater than 15 cm3 was found, testing was continued until that degree of consistency was obtained for three successive tests. The technique was standardised using water on a specially prepared cranium of which the calotte was initially removable. All relevant foramina, fissures, and sutures were internally and externally sealed with plasticine. The whole internal surface was then coated with rubber latex (to ensure waterproofing), followed by a coat of pol urethane (to prevent the mustard see s adhering). The calotte was then re-attached using “Bedacryl”.ECC was estimated using mustard seed and then water, five times each. A typical result ofsne such test was (in cm2 mustard seed -X 1,304, S.D. 9.6; water-X 1,284, S.D. 2.2. This, and subsequent tests, showed that our mustard seed technique, compared to water, overestimated ECC by about 20 cm3. Variations in seed compacting in either the cranium, or in the measuring cylinder, can significantly vary the volume. Hence all our data are given as water-standardised values.
B
